Sea-ice and snow interaction revealed by combined retrieval of sea-ice thickness and snow depth with CryoSat-2 and SMOS

Shiming Xu, Lu Zhou

Corresponding author: Shiming Xu

Corresponding author e-mail: xusm@tsinghua.edu.cn

Sea ice and its snow cover is an integral component of the polar and global climate system. Sea ice is a direct indicator of polar air–sea interaction through its volume, and snow enforces important control over the sea ice through processes such as thermal insulation and change of surface albedo. By physical synergy of satellite data including altimetry (CryoSat-2) and passive radiometry (SMOS), the sea-ice thickness and snow depth can both be retrieved for Arctic winter seasons. Through analysis of winter ice growth and thermodynamic atmospheric and snow-induced forcings, we demonstrate quantitative results of snow’s active roles during the winter months. Projections about future change of sea ice and snow are also made through climate modeling studies and Coupled Model Inter-comparison Project (CMIP5) data.


Development of phosphorus forms in soil chronosequence of Nordenskioldbreen glacier, Svalbard

Adel Allaberdina

Corresponding author: Adel Allaberdina

Corresponding author e-mail: adelallaberdina92@gmail.com

Nordenskioldbreen is located between Dickson Land and Bunsow Land, Svalbard. The glacier flows roughly southwest and is 25 km (16 mi) long and 11 km (6.8 mi) wide. It has its terminus in Adolfsbukta, a branch of Billefjorden. After the expedition to the glacier in 2015, soil samples were brought to the Czech University of Life Sciences, Prague. We have 42 samples from the south and north of the glacier. With each of them we worked in the laboratory during several months, made a filtration and with the help of robotics acquired the complete chemical composition of each sample. In the process, we determined the amount of P-form and other soil properties in soils below Nordenskioldbreen. The soil properties after the retreat of Nordenskioldbreen vary with the age of the soil. The amount of accessible P decreases with the age of the soil and the amount of crystalline Fe forms increases. All results of the study are presented in spreadsheet format.


Restoring Arctic ice with strategically placed, albedo-enhancing reflective glass beads

Leslie Field

Corresponding author: Leslie Field

Corresponding author e-mail: leslie@ice911.org

Over the past few decades, since 1979, the Arctic has lost 95% of its bright, reflective multiyear ice. Now each winter, instead of the season adding to and maintaining the Arctic’s historic reservoir of bright multiyear ice, thin, transparent first-year ice forms over much of the Arctic. The ice in the Arctic has historically acted like a global heat shield against 24-hour-per-day incoming solar radiation, reflecting incoming energy that would otherwise warm the ocean and destabilize global weather. This loss of the Earth’s reflective heat shield has dramatically altered the jet stream and mid-latitude weather. Professor Peter Wadhams has said, ‘The overall change in ice/snow albedo in the Arctic could add as much as 50% to the direct global heating effect of CO2’. This drastic change in reflectivity could be a key lever to mitigating climate change, and one that could be solved with reflective materials. Restoring Arctic ice could therefore be the largest single safe lever we could use to lessen climate change impacts while the world gets a handle on decarbonization. When dispersed on vulnerable ice in a strategic location, environmentally benign reflective materials could slow the melt, helping younger, thinner ice survive longer into the summer and adding to the store of multiyear ice. With 10 years of field testing, including three seasons of field testing on an Arctic lake, expert climate modeling, and peer-reviewed publication, Ice911’s mission is to restore Arctic sea ice.


The impacts of increased Southern Ocean freshwater fluxes in CCSM4

Inga J. Smith, Andrew G. Pauling, Cecilia M. Bitz, Katherine Lilly, Patricia J. Langhorne, Christina L. Hulbe

Corresponding author: Inga J. Smith

Corresponding author e-mail: inga.smith@otago.ac.nz

To examine potential future impacts on sea-ice area and feedback effects on climate (e.g. surface air temperatures in both hemispheres) we ran three climate model scenarios of changes in outflow from Antarctic ice sheet and ice shelves in CCSM4 starting in 1980. We then ran two idealized experiments starting at 1850 for comparison. We chose to run from a base year of 1980, which we assumed to be when the Antarctic ice sheet was in mass balance, and then increased freshwater fluxes to give the approximate equivalent of 3 m of sea level rise over a 150-year period. Sea-ice-area behaviour ‘turned around’ from increasing after approximately 78 years. Two shorter branched runs were carried out to test the persistence of the effects, one where the additional fresh water and latent heat effects were switched off and the other where the freshwater was held constant at the point where sea-ice-area behaviour ‘turned around’. Another two shorter runs were carried out, starting in 1850, with ramped freshwater and with latent heat effects included under historical greenhouse gas forcings and with greenhouse gas forcings held constant at 1850 levels, respectively. This was to separate out the effects of greenhouse gas forcings on the ramped freshwater and latent heat effects over long time periods.


Sea-ice thickness records from altimetry over Antarctica

Florent Garnier, Sara Fleury, Kévin Guerreiro, Antoine Laforge, Benoit Meyssignac

Corresponding author: Sara Fleury

Corresponding author e-mail: sara.fleury@legos.obs-mip.fr

The main difficulties in retrieving sea-ice thickness (SIT) in the Southern Ocean come from the lack of in-situ observation and knowledge related to sea ice. For instance, whereas polar expeditions and in-situ observations over the Arctic have enabled the construction of snow-depth climatologies (e.g the Warren climatology), there are no equivalent data available over Antarctica. By consequence, except for a few studies such as those of Zwally et al (2008) or Kurtz et al (2012), based on ICESat data, sea-ice thickness estimations over Antarctica remain virtually nonexistent and no valid sea-ice volume estimations have yet been drawn up (SI-CCI-2015 report). The objective of this presentation is to review our recent developments leading towards sea-ice thickness estimations over Antarctica. First, we detail the methodology used to derive sea-ice freeboard from altimetric power echo measurements and we present a 2002–17 Envisat/Cryosat-2 sea-ice radar freeboard time series. The continuity between these two satellites is ensured by a re-calibration of the Envisat low-resolution mode (LRM) on the Cryosat synthetic aperture radar (SAR) mode we describe in this presentation. Thereafter, the computation of sea-ice thickness from sea-ice freeboard requires snow-depth fields. Recently, the ESA CryoSeaNICE project demonstrated that snow depth is one of the most important obstacle to computing sea-ice thickness. Meanwhile, Guerreiro et al (2016) showed the ability to retrieve snow depth from CryoSat-2 Ku and Saral/AltiKa Ka radar frequencies. Based on this approach, we present in this second part the first bi-frequency altimetric snow depth product over Antarctica computed from the Cryosat-2 Pseudo-LRM ESA product. The relevancy of the solution will be assessed by comparison with ICESat and Operation Ice Bridge (OIB) laser measurements. From these results, a 2013–18 sea-ice thickness time series will be presented and assessed. Finally, we briefly explain how these sea-ice products will be used to better understand recent sea-ice variations and to derive sea-ice volume estimations.


Wind, the Beaufort Gyre and survival of perennial ice

Jennifer Hutchings, Ben Lewis

Corresponding author: Jennifer Hutchings

Corresponding author e-mail: jhutchings@coas.oregonstate.edu

Wind-stress transfer into the ocean is modulated by the sea-ice cover. Variability in ice drift and geostrophic currents is controlled by both interannual variability in winds and a changing ice cover. The Beaufort Sea has experienced a large reduction in perennial ice cover since 2006, together with accumulation of fresh water in the Beaufort Gyre. The largest changes in the ice pack, faster drift speeds and decreased ice cover are in September through November. We do not see significant changes to ice drift in winter. While the ice pack is more predominately first-year ice that is presumably weaker now than in the past, we do not find evidence that the ice pack is responding more to wind stress. Twenty years of AVHRR imagery were used to identify opening fractures in the Beaufort Sea ice pack. The majority of these fractures form under tension and in ice–coast interaction. We do not find any trends in their occurrence and can relate ice motion associated with these fractures to the passage of weather systems across the central Arctic, Chukchi and Beaufort Seas. Typical seasonality in position of anticyclones tends to promote transport of perennial ice into regions of the Beaufort Sea that have experienced enhanced melt in the recent decade. Survival of perennial ice in the Beaufort Gyre now requires sea ice to transit the southern part of the gyre in a single winter. This requires anomalous seasonality in the location of the Beaufort High. Indices such as the Arctic Oscillation do not capture the detail in the position of the Beaufort High that can be related to enhanced loss or retention of perennial ice in the following summer.


Spatial and temporal patterns of Arctic and Antarctic sea-ice leads, 2002–18

Sascha Willmes, Fabian Reiser, Günther Heinemann

Corresponding author: Sascha Willmes

Corresponding author e-mail: willmes@uni-trier.de

The occurrence of leads represents a key feature of the polar sea-ice cover. Leads promote the flux of sensible and latent heat from the ocean to the cold winter atmosphere and are thereby crucial for air–sea-ice–ocean interactions and feedbacks. We use the thermal signature of leads in the MODIS ice-surface temperature product and a subsequent cloud artefact filter to infer daily and monthly composite lead maps for the sea-ice area in both hemispheres during wintertime, 2002–18. For the Arctic, our results highlight the marginal ice zone in Fram Strait and in the Barents Sea as the primary region for lead activity. However, the spatial distribution of the average pan-Arctic lead frequencies also reveals distinct and hitherto less known patterns of predominant fracture zones along the Arctic Boundary Current. For the Antarctic, we present results for the most active sea-ice regions with respect to lead occurrences and their connection to bathymetric features. Hemispheric and regional peculiarities as well as methodological improvements are introduced and discussed. Results are compared to microwave lead products and the potential of this dataset for an operational high-resolution sea-ice monitoring is investigated.


Seasonal evolution of light transmission through Arctic summer sea ice

Christian Katlein, Stefanie Arndt, H. Jakob Belter, Marcel Nicolaus

Corresponding author: Christian Katlein

Corresponding author e-mail: ckatlein@awi.de

Light transmission through sea ice has been identified as a critical process for energy partitioning at the polar atmosphere–ice–ocean boundary. Transmission of sunlight influences direct sea-ice melting by absorption, heat deposition in the upper ocean, and in particular primary productivity. While earlier observations relied on a limited number of point observations, the recent years have seen an increase in spatially distributed light measurements underneath sea ice using remotely operated vehicles (ROV). These measurements allow us to reconstruct the seasonal evolution of the spatial variability in light transmission. Here we present measurements of sea-ice light transmittance from 6 years of polar ROV operations. The dataset covers the entire melt cycle of central Arctic sea ice. Data is combined into a pseudo timeseries describing the seasonal evolution of the changing spatial variability of sea-ice optical properties. Snow melt in spring increases light transmission continuously, until a secondary mode originating from translucent melt-ponds appears in the histograms of light transmittance. This secondary mode persists long into autumn, before snow fall reduces overall light levels again. Comparison to several autonomous time series measurements from single locations confirms the detected general patterns of the seasonal evolution of light transmittance variability. These results allow for the evaluation of two different light transmittance parameterizations, implying that light transmission is overestimated in current ice–ocean models.


New tools for optical measurements in sea ice

Christian Katlein, Simon Lambert Girard, Christophe Perron, Raphaël Larouche, Yasmine Alikacem, Simon Thibault, Pierre Marquet, Marcel Babin, Lovro Valcic

Corresponding author: Christian Katlein

Corresponding author e-mail: ckatlein@awi.de

The quantity and quality of sunlight transmitted into and through sea ice is a crucial key necessary to understand the thermodynamic development of the ice cover, upper ocean heat and freshwater budget, as well as the associated primary production. Due to its solid impenetrable nature, most optical measurements so far have been conducted above and underneath the sea ice covering our polar oceans. Only very limited measurements have been carried out inside the ice cover itself. This strongly limits our current knowledge of the vertically varying inherent optical properties (IOP) of sea ice, as well as the geometric shape of the in-ice light field. Both factors currently limit our abilities to reliably model radiative transfer in sea ice. Here we present multiple new tools that can fill this observational gap and provide comprehensive optical measurements within the ice: This includes a chain of multispectral light sensors for seasonal long-term monitoring. It is derived from the proven design of the newest generation of ice-mass-balance buoys with digital thermistor strings and enables a non-destructive measurement with flexible geometry. We present data from a first prototype deployed together with an array of drifting ice observatories at the North Pole in September 2018. These vertically resolved in-ice light profiles are compared to in-ice measurements with a newly designed in-ice optical profiler system based on the well-proven TriOS Ramses hyperspectral radiometers. Combining expertise from photonics, medical and sea-ice science enables the ongoing development of a set of endoscopic probes allowing optical studies in sea ice with minimum disturbance of the ice. This includes in-ice microscopy for in-situ ice algal investigations, a UV-spectrometer to observe brine nitrate concentration in situ, a reflectance probe for high-resolution direct determination of inherent optical properties, as well as a radiance camera for quantification of the angular radiance distribution. Here we present data from the first field tests during the Arctic field season 2018. First ruggedized prototypes could be available to the scientific community soon.


Impact of model resolution on Arctic sea ice and North Atlantic Ocean heat transport

David Docquier, Jeremy P. Grist, Malcolm J. Roberts, Christopher D. Roberts, Tido Semmler, Leandro Ponsoni, François Massonnet, Dmitry Sidorenko, Dmitry V. Sein

Corresponding author: David Docquier

Corresponding author e-mail: david.docquier@uclouvain.be

Arctic sea-ice area and volume have substantially decreased since the beginning of the satellite era. Concurrently, the poleward heat transport from the North Atlantic Ocean into the Arctic has increased, partly contributing to the loss of sea ice. Increasing the horizontal resolution of general circulation models (GCMs) improves their ability to represent the complex interplay of processes at high latitudes. Here, we investigate the impact of model resolution on Arctic sea ice and Atlantic Ocean heat transport (OHT) by using different state-of-the-art coupled GCMs that include dynamic representations of the ocean, atmosphere and sea ice. The models participate in the High Resolution Model Intercomparison Project (HighResMIP) of the sixth phase of the Coupled Model Intercomparison Project (CMIP6). Model results over the period 1950–2014 are compared to different observational datasets. In the models studied, a finer ocean resolution drives lower Arctic sea-ice area and volume and generally enhances Atlantic OHT. The representation of ocean surface characteristics, such as sea-surface temperature (SST) and velocity, is greatly improved by using a finer ocean resolution. This study highlights a clear anticorrelation at interannual time scales between Arctic sea ice (area and volume) and Atlantic OHT in all models. However, the strength of this relationship is not systematically impacted by model resolution. Sea ice in the Barents/Kara and Greenland–Iceland–Norwegian (GIN) Seas is more strongly connected to Atlantic OHT than other Arctic seas.


Teleconnections between the Weddell polynya and the Filchner–Ronne Ice Shelf cavity

Kaitlin Naughten, Adrian Jenkins, Paul Holland, Ruth Mugford, Keith Nicholls, David Munday

Corresponding author: Kaitlin Naughten

Corresponding author e-mail: kaight@bas.ac.uk

Open-ocean polynyas in the Weddell Sea of Antarctica are the products of deep convection, which transports warm deep water (WDW) to the surface and melts sea ice. These polynyas occur only rarely in the observational record, but are a near-permanent feature of many climate and ocean simulations. A question which has not previously been considered is what impact the Weddell Polynya has on the nearby Filchner–Ronne Ice Shelf (FRIS) cavity. Here we assess this impact using regional ocean model simulations of the Weddell Sea and FRIS, where deep convection is imposed with varying extent, location and duration. In these simulations, the idealized Weddell polynyas consistently cause an increase in WDW transport onto the continental shelf, due to density changes above the shelf break. This leads to saltier, denser source waters for the FRIS cavity, which then experiences stronger circulation and increased ice-shelf basal melting. It takes approximately 14 years for melt rates to return to normal after the deep convection ceases. The Weddell polynya of 2017 was likely too small and short-lived to have a discernible impact on the FRIS cavity, but should a larger Weddell polynya recur in observations, the response of the cavity should be monitored. Our results also suggest that ocean models with excessive Weddell Sea convection may not be suitable boundary conditions for regional models of the Antarctic continental shelf and ice-shelf cavities.


Sea-ice–ocean feedbacks in the Antarctic shelf seas

Rebecca Frew, Daniel Feltham, Paul Holland, Alek Petty

Corresponding author: Daniel Feltham

Corresponding author e-mail: d.l.feltham@reading.ac.uk

Observed changes in Antarctic sea ice are poorly understood, in part due to the complexity of its interactions with the atmosphere and ocean. A highly simplified, coupled sea-ice–ocean mixed layer model has been developed to investigate the importance of sea-ice–ocean feedbacks on the evolution of sea ice and the ocean mixed layer in two contrasting regions of the Southern Ocean: the Amundsen Sea, which has warm shelf waters; and the Weddell Sea, which has cold and saline shelf waters. Modelling studies where we deny the feedback response to surface air temperature perturbations show the importance of feedbacks on the mixed layer and ice cover in the Weddell Sea to be smaller than the sensitivity to surface atmospheric conditions. In the Amundsen Sea the effect of surface air-temperature perturbations on the sea ice are opposed by changes in the entrainment of warm deep waters into the mixed layer. The net impact depends on the relative balance between changes in sea-ice growth driven by surface perturbations and basal driven melting. The changes in the entrainment of warm water in the Amundsen Sea were found to have a much larger impact on the ice volume than perturbations in the surface energy budget. This creates a net negative ice albedo feedback in the Amundsen Sea, reversing this feedback’s usual sign.


Evaluation of sea-ice datasets with CMIP6 multi-model datasets

Arun Rana, François Massonnet

Corresponding author: Arun Rana

Corresponding author e-mail: arunranain@gmail.com

We investigate differences, similarities and uncertainties in available CMIP6 historical Arctic and Antarctic sea-ice simulations and compare them to observational references (namely, National Snow & Ice Data Centre (NSIDC) and Ocean and Sea Ice Satellite Application Facility (OSISAF)). We focus our analyses on sea-ice concentration and extent. The said observational products differ in the retrieval methods that were employed as well as in their horizontal resolution (varying from about 0.1° to 0.25°). We apply several univariate (climatologies, trends, standard deviations) and multivariate (per-pair RMSE, principal component analysis, and Taylor diagram, cluster analysis) statistical diagnostics and metrics in order to evaluate the quality of the simulations, but also the impact of observational uncertainty on these indices. Primary results have indicated that the observational uncertainties (read spread) are considerable and would affect the comparison. The goal of the study is to i) evaluate the ability of each of the CMIP6 models to simulate the observed climatologies and trends in Arctic and Antarctic sea ice areal properties, ii) analyze similarities and contradictions between observational datasets and output from multiple models, and iii) estimate the level of observational uncertainty and gauge its impact in climate model evaluation studies.


Arctic sea-ice biogeochemistry and air–ice–water CO2 exchange in the coldest period in February to April

Agneta Fransson, Melissa Chierici, Daiki Nomura, Mats Granskog, Eva Leu, Clara Hoppe, Svein Kristiansen

Corresponding author: Agneta Fransson

Corresponding author e-mail: agneta.fransson@npolar.no

The sea ice in the Arctic region is changing from thicker, multiyear ice to thinner, younger and more first-year ice, which will have consequences for the biogeochemical processes in sea ice and the air–ice–water CO2 fluxes. In the coldest period, several studies in sea ice show the indications of formation of ikaite, a form of solid calcium carbonate, where CO2 is released to the brine. Ikaite can escape from the ice to underlying water, adding alkalinity to the underlying water. With time, CO2 becomes depleted in the sea-ice brine due to ice–water brine–CO2 rejection and ice–air CO2 exchange. Frost flowers, formed on new ice, are shown to mediate CO2 transfer from the ice to the atmosphere. In winter sea ice, CO2 is released due to bacterial respiration, which is an important process for the net change in carbonate chemistry when primary production is limited by insufficient light. Here, we present the development and changes of sea-ice properties of salinity, temperature, carbonate chemistry, nutrients and isotopic ratio of oxygen (δ18O) at three field sites with different sea-ice types (fjord ice, drift ice, landfast, old or newly formed) in the coldest Arctic sea ice of the year, from February to April. We discuss the importance of different processes driving the air–ice–water CO2 fluxes in winter. The results show that sea-ice types investigated north of Svalbard (Nansen Basin), and in the Svalbard fjords Van Mijenfjorden and Kongsfjorden, showed increased salinity, total alkalinity, and total dissolved inorganic carbon, with coldest ice, from the ice–air interface to 10 cm in the ice. Below this top layer, the sea-ice properties remained relatively constant throughout the cold period, decreasing towards the end of the period. In Kongsfjorden and north of Svalbard, frost flowers and brine skim acted as mediators for air–ice CO2 exchange, where CO2 in new frost flowers disappeared to atmosphere after 24 hours exposure. Ikaite played a role at all sites, in the change of carbonate chemistry in the sea ice, in underlying water, and the CO2 exchange with the atmosphere.


Effects of sea-ice and biogeochemical processes and storms on under-ice water fCO2 from winter to spring in the high Arctic Ocean: Implications for sea–air CO2 fluxes

Agneta Fransson, Melissa Chierici, Ingunn Skjelvan, Are Olsen, Philipp Assmy, Algot K. Peterson, Gunnar Spreen, Brian Ward

Corresponding author: Agneta Fransson

Corresponding author e-mail: agneta.fransson@npolar.no

The ice cover in the Arctic Ocean has decreased during the last decades, manifested in particular as an extensive transition from thicker multiyear ice to thinner first-year ice. As the summer sea-ice cover is decreasing, larger areas with open water will be exposed to the atmosphere. This will have implications for the carbonate chemistry and sea–air carbon dioxide (CO2) exchange. We present measurements of CO2 fugacity (fCO2) and estimates of the effects biogeochemical processes in the surface water under Arctic sea ice, driving the sea–air CO2 fluxes. The data was obtained from January to June 2015 during the Norwegian young sea ICE (N-ICE2015) expedition, where the ship drifted with four different ice floes and covered the deep Nansen Basin, the slopes north of Svalbard, and the Yermak Plateau. This unique winter-to-spring dataset includes the first winter under-ice water fCO2 observations in this region. The observed under-ice fCO2 ranged between 315 μatm in winter and 153 μatm in spring, hence was undersaturated relative to the atmospheric fCO2. Although the sea ice partly prevented direct CO2 exchange between ocean and atmosphere, frequently occurring leads and breakup of the ice sheet promoted sea–air CO2 fluxes. The CO2 sink varied between 0.3 and 86 mmol C m–2 d–1, depending strongly on the open-water fractions (OW) and storm events. The maximum sea–air CO2 fluxes occurred during storm events in February and June. In winter, the main drivers of the change in under-ice water fCO2 were dissolution of CaCO3 (ikaite) and vertical mixing. In June, in addition to these processes, primary production and sea–air CO2 fluxes were important. The cumulative loss due to CaCO3 dissolution of 0.7 mol C m–2 in the upper 10 m played a major role in sustaining the undersaturation of fCO2 during the entire study. The relative effects of the total fCO2 change due to CaCO3 dissolution was 38%, primary production 26%, vertical mixing 16%, sea–air CO2 fluxes 16%, and temperature and salinity insignificant.


ICESat-2 over sea ice: retrieved heights and freeboards

Ron Kwok, Glenn Cunningham, Nathan Kurtz, Alek Petty, Thorsten Markus, Tom Armitage

Corresponding author: Ron Kwok

Corresponding author e-mail: ron.kwok@jpl.nasa.gov

One of the science objectives of NASA’s ICESat-2 altimetry mission is to provide observations to quantify changes and to add to previous satellite and airborne records of freeboard, thickness and sea-surface height of the ice-covered Arctic and Southern Oceans (e.g. from ICESat, Operation IceBridge and CryoSat-2). ATLAS, a multi-beam photon-counting lidar, the sole instrument on the ICESat-2 observatory, launched in September 2018, provides a rich altimetric dataset of profiles of the ice and ocean surfaces. In this talk, we will show the capabilities of the multi-beam instrument based on data acquired thus far (4 months at the time of writing) over the Arctic and Antarctic ice covers. In particular, we will show the precision in the retrieved surface heights over a relatively flat surface, the spatial resolution of the height estimates, the time-varying freeboard estimates and sea-surface height anomalies over a seasonal cycle, and assessments of the retrievals when compared with airborne and field acquisitions.


Snow depth and ice thickness on Arctic sea ice from ICESat-2 and CryoSat-2 freeboards

Ron Kwok, Sahra Kacimi, Melinda Webster, Nathan Kurtz, Alek Petty, Thorsten Markus

Corresponding author: Ron Kwok

Corresponding author e-mail: ron.kwok@jpl.nasa.gov

Prior to the launch of ICESat-2 (IS-2), the potential to produce basin-scale estimates of snow depth by differencing freeboard heights from CryoSat-2 (CS-2) and IS-2 was examined by Kwok and Markus (2018). With CS-2 freeboards and lidar freeboards now available in IS-2 data products, we have been able to construct monthly fields of snow depth estimates over Arctic sea ice. The monthly fields clearly show the development of the snow cover since mid-October 2018, the beginning of the IS-2 science mission, to its melt onset in spring 2019. Here, we show comparisons of the snow-depth estimates with reconstructions from ERA-interim and ERA5, and available estimates from airborne and field measurements. Also, we discuss the variability in these estimates associated with the: (1) lack of time-space coincidence of the freeboard datasets, (2) uncertainties stemming from the snow density used in the conversion to snow depth, and (3) potential impact of snow salinity on radar freeboard.


Recent extreme events in the Arctic: collapsing ice arches and new polynyas

Kent Moore, Axel Schweiger, Jinlun Zhang, Mike Steele

Corresponding author: Kent Moore

Corresponding author e-mail: gwk.moore@utoronto.ca

One of the most dramatic indicators of climate change is the reduction in the extent and thickness of Arctic sea ice that has resulted in an increase in wind-driven sea ice mobility. A number of unusual events have occurred in the Arctic over the past few years that appear to be associated with this increase. Included was the early collapse of the Lincoln Sea ice arch, which forms at the northern end of Nares Strait, during May 2017, as well as the formation of a polynya over the Wandel Sea to the north of Greenland, a region not known for polynya development, during February 2018. In this talk I will describe the impacts of these events on the climate system and will use satellite data, in-situ weather data, atmospheric reanalyses and a coupled sea-ice–ocean model to document the role that atmospheric forcing and sea-ice characteristics played in these events. In particular, we show that the May 2017 ice-arch collapse occurred during record low sea-ice thicknesses in the Lincoln Sea in the presence of a polynya in northern Nares Strait and an unusual wind regime characterized by a strong northerly flow down Nares Strait. The Wandel Sea polynya occurred during a period of warm enhanced southerly flow associated with the sudden stratospheric warming that occurred around the same time. However, unlike ice-arch collapse, thinning sea ice did not appear to play a role in the development of the polynya and, for the foreseeable future, the polynya will only develop under similar strong southerly flow conditions. These events demonstrate that not everything of scientific interest in the Arctic falls neatly into the climate change canon, and unusual and noteworthy events will contribute to provide natural experiments that advance our understanding of the complex interactions between the ocean, the atmosphere and sea ice that are a characteristic of the region’s climate.


High spatial resolution and rapid temporal repeat retrievals of sea-ice motion and melt-onset timing from multi-sensor Sentinel–1 and RADARSAT-2 backscatter

Stephen Howell, David Small, Mike Brady

Corresponding author: Stephen Howell

Corresponding author e-mail: Stephen.Howell@Canada.ca

The European Space Agency Sentinel–1 satellites, combined with the Canadian Space Agency RADARSAT-2 mission, deliver new capabilities for high-spatial-resolution and rapid-temporal-epeat pan-Arctic sea-ice retrievals from C-band synthetic aperture radar (SAR) data. The upcoming launch of the RADARSAT Constellation Mission will add yet more C-band SAR sensors to the existing satellite data pool. Combining SAR imagery from these satellites provides a new opportunity to construct high-temporal-resolution (i.e. dense time series) datasets across a pan-Arctic domain, analogous to what is possible from passive microwave brightness temperatures but at much higher spatial resolution (i.e. 20–100 m compared to 25 km). In this presentation, we will highlight the development and application of two multi-sensor sea ice products to illustrate the utility of combining data from multiple C-band satellite sensors. First, we present composite normalized gamma-nought backscatter products from Sentinel–1 and RADARSAT-2, which provide complete coverage at daily revisit covering the northern Canadian Arctic and north coast of Greenland, followed by an evaluation of melt-onset timing estimates over Arctic sea ice compared to coarser-resolution sensors. Second, we present the results of an automated approach for utilizing all available C-band SAR imagery to produce high-spatial-resolution estimates of sea-ice motion for the pan-Arctic. Both applications illustrate that a multi-sensor approach offers a robust approach to retrieve sea-ice variables at high spatial resolution over the entire pan-Arctic.


Sea-ice detection with GNSS-R data from TechDemoSat-1

Jessica Cartwright, Christopher J. Banks, Meric Srokosz

Corresponding author: Jessica Cartwright

Corresponding author e-mail: jc1n15@noc.soton.ac.uk

Observations of sea ice are essential in order to monitor the effects of a changing climate as well as understanding potential changes in the future. As a highly dynamic component of the cryosphere, sea ice plays an important role in these changes. Due to the remote nature of the study area and the large area of sea-ice extent, satellite remote sensing is the only viable approach to this task. The use of reflected navigation signals, such as those from GPS satellites presents a cost-effective means of making sea-ice observations. At present such techniques are used primarily for the monitoring of ocean winds, but when applied to sea ice, GNSS-R (Global Navigation Satellite Systems-Reflectometry) would allow a reduction in the cost associated with these measurements, thus enabling an increase in the spatio-temporal resolution and coverage. Here we present a new method for the detection of sea ice applied to 33 months of GNSS-R data from the UK TechDemoSat-1 satellite. This method of sea-ice detection shows the potential for a future GNSS-R polar mission, attaining an agreement of over 98% and 96% in the Antarctic and Arctic respectively, when compared to ESA’s CCI (European Space Agency’s Climate Change Initiative) sea ice concentration product. The algorithm uses a combination of two parameters derived from the delay-Doppler maps (DDMs) to quantify the spread of power in delay and Doppler. Threshold application on a small training dataset then allows sea ice to be distinguished from open water. Almost 50 different parameters are derived from the DDMs throughout this study and the two best performing parameters are then applied to the dataset as a whole. Differences between the sea-ice detection from this method and comparison datasets are explored. Previous studies into sea-ice detection from GNSS-R have largely focussed on a single track or season. The application of sea-ice detection to the entire TDS-1 dataset provides information on the seasonal and multiyear changes in sea-ice distribution of both the Arctic and Antarctic.


Mechanisms controlling asymmetry in the annual cycle of Antarctic sea ice

Clare Eayrs, Daiane Faller, Rajesh Kumar, David Holland

Corresponding author: Clare Eayrs

Corresponding author e-mail: clare.eayrs@nyu.edu

Sea ice is a critical component of the climate system, but the drivers of the strong temporal and regional variability of Antarctic sea ice are poorly understood. Nearly 16 million square kilometres of sea ice grows and subsequently melts each year in the seas surrounding Antarctica, a sixfold increase in sea-ice extent that effectively doubles the size of the continent each winter. The modest overall increase in Antarctic sea-ice cover (1.5% per decade, 1981–2010) masks substantial interannual and regional variability and, in recent years, record maximum (2012–14) and minimum (2016–18) extents have been observed. Despite this considerable interannual variability, Antarctic sea ice melts much faster than it grows (5 months compared to 7 months) in each year of the satellite record. Twice-yearly changes in the position and intensity of the zonal winds circling Antarctica are believed to drive the system by working with/against the evolving sea- ice edge to slow the autumn advance and hasten the spring melt. Divergence during the melting season leads to open-water regions through which heat input drives bottom and lateral melt and deterioration and fragmentation of ice floes. We explore these processes using the ERA5 reanalyses (1979–2018), satellite sea-ice observations (1979–2018) and the 40 historical ensembles (1920–2005) from the CESM-LENS project. Understanding the role of natural variability is key to improving model prediction of Antarctic sea ice and a better understanding of the processes governing the seasonal cycle of Antarctic sea ice can provide insight into its variability on longer timescales.


The spatio-temporal patterns of landfast ice in Antarctica during 2006–11 and 2016–17 using high-resolution SAR imagery

Fengming Hui, Xinqing Li, Xiao Cheng, Mohammed Shokr

Corresponding author: Fengming Hui

Corresponding author e-mail: huifm@bnu.edu.cn

Landfast ice is an important component of the Antarctic sea ice. It affects the Antarctic climate and ecological system. In this study, the first high-resolution, long-term series of landfast ice edge during 2006–11 and 2016–17 is presented. The dataset is produced based on the improved net gradient difference algorithm by using 2470 SAR scenes from ENVISAT and Sentinel–1A/B as well as manual depiction using MODIS data to fill the SAR data gaps. Results show that landfast ice area in November is about (49.49 ± 3.25) × 104 km2, accounting for about 3–4% of the total Antarctic sea-ice area. The area in West Antarctica is about 40% of that in East Antarctica. During the study period the landfast-ice area decreased at a linear rate of 12033 km2 a–1. The distribution of landfast ice in Antarctic shows significant regional differences. The extent in the Indian Ocean sector is the greatest, a mean value of (16.49 ± 1.1) × 104km2, but the ratio of landfast-ice area to sea-ice area is highest in the Pacific Ocean. Twenty-four landfast-ice zones with groups of small grounded icebergs were identified, most of which are locatde in East Antarctica, especially along Wilkes Land and Oates Land. Two typical cases have demonstrated how giant icebergs influence the existence and development of landfast ice. This implies that small grounded icebergs play a significant role in formation and development of landfast ice in Antarctica. The data from this study have the potential to promote studies of the Antarctic climate and ecology system.


Arctic sea-ice classification using microwave scatterometer and radiometer data during 2002–17

Zhilun Zhang, Yining Yu, Xinqing Li, Fengming Hui, Xiao Cheng, Zhuoqi Chen

Corresponding author: Fengming Hui

Corresponding author e-mail: huifm@bnu.edu.cn

Temporal and spatial variation of sea-ice type in the Arctic is an indicator of regional and global change. Arctic sea ice can be classified into two major categories: multiyear ice (MYI) and first-year ice (FYI). In this study, a classification method based on machine learning is established and applied to produce a daily sea-ice classification dataset during winter (November–April) from 2002 to 2017 using active microwave data from QuikSCAT and ASCAT scatterometer as well as passive microwave data from AMSR-E, SSMI/S and AMSR-2 radiometer. First, the open-water area is flagged out using brightness temperature from the PM sensor. Then the K-means algorithm is applied to identify the clusters of the two ice types in the brightness–temperature/backscatter parameter space and finally to assign pixels to each class. Two optimization methods based on the movement of MYI and the marginal ice zone are used to correct the misclassification of MYI. Results have shown a decrease of MYI in winter from 2002–17, especially in 2008 and 2013, with a remarkable recovery in 2014. The classifications are consistent with results vfromisual interpretation of SAR images in the Canadian Arctic Archipelago with an overall classification accuracy of more than 93%. Comparison against classifications from previous studies and products shows that our method could reflect more differences in the MYI declining trend interannually and less anomalous fluctuations in certain years.


Photoacclimation state of an Arctic under-ice phytoplankton bloom

Hanna M. Kauko, Alexey K. Pavlov, Geir Johnsen, Mats A. Granskog, Ilka Peeken, Philipp Assmy

Corresponding author: Hanna M. Kauko

Corresponding author e-mail: hanna.kauko@npolar.no

Recently, several reports on Arctic under-ice phytoplankton blooms have improved our understanding of Arctic primary production. However, such under-ice blooms cannot be detected from space. Innovative use of autonomous optical under-ice measurements can extend observations beyond ship-based surveys. In this study, we concentrate on the photoacclimation state of a Phaeocystis pouchetii dominated under-ice bloom in the Arctic Ocean using ratios of photoprotective (PPC) to photosynthetic carotenoids (PSC), and the possibilities of using in-situ absorption measurements to estimate the photoacclimation state. The photoacclimation state of an under-ice bloom can provide clues about the light conditions the algae were growing in, and therefore of the origin of the bloom: high-light acclimation indicates advection from open-water areas with high light conditions, whereas low-light-acclimated cells were possibly growing in the typically low-light conditions under sea ice. The under-ice bloom was found to be low-light acclimated, which indicates local growth under the ice pack. We could also confirm the applicability of a method using in-situ light absorption measurements to estimate the PPC:PSC ratio. The slope of in-situ phytoplankton absorption between 488 and 532 nm, affected by both PPC and PSC, had a significant linear relationship to the PPC : PSC ratio, indicating that absorption profiles can be used for prediction of the photo-acclimation state. Future work is needed for example to assess the impact of pigment packaging on the relationship between PPC : PSC and absorption-measurement slopes. Our study shows the potential of in-situ absorption measurements to study phytoplankton physiology below sea ice.


Spring progression of spectral light transmission through landfast sea ice and implications on Arctic primary production

Lisa C. Matthes, C. J. Mundy, Simon Lambert Girard, Marcel Babin, Jens K. Ehn

Corresponding author: Lisa C. Matthes

Corresponding author e-mail: matthesl@myumanitoba.ca

The availability of photosynthetically active radiation (PAR; 400–700 nm) is a key factor for under-ice phytoplankton growth in a seasonally sea-ice-covered water body. The increase towards sufficient light levels for positive net photosynthesis occurs concurrently with the melt progression in late spring as ice-surface conditions shift from a highly reflective snow cover to a less reflective mosaic of bare ice and melt ponds. However, spatial variability in ice-surface characteristics, i.e. snow thickness or melt pond distributions, and the subsequent impact on transmitted PAR makes estimates of light-limited primary production difficult at this time of year. To quantify the increase of spectral light transmission as a function of melting sea-ice surface and changing ice-algae concentration in the ice bottom, a remotely operated vehicle (ROV) equipped with hyperspectral radiometers was deployed during the GreenEdge ice-camp campaign in June–July 2016 in Baffin Bay. Horizontal transects of ~130 m length at 2 m depth and vertical profiles, to a depth of 50 m, were repeatedly performed beneath the ice cover with changing quantities of snow, ice, melt ponds and ice algae. In this poster, we present a detailed dataset on the spatial and temporal progression of transmitted spectral irradiance and discuss the availability of PAR on average in the upper water column in relation to the changes in the aforementioned variables.


First measurements of spring primary production in central Hudson Bay

Lisa Matthes, Laura Dalman, Jens K. Ehn, Jean-Éric Tremblay, Janghan Lee, Simon Bélanger, C. J. Mundy

Corresponding author: Lisa Matthes

Corresponding author e-mail: matthesl@myumanitoba.ca

Strong surface stratification and a seasonal sea-ice cover lasting up to 9 months, combined with predominately post-bloom autumnal observations of the region, have contributed to a characterization of central Hudson Bay as a region of relatively low annual primary production. Remote sensing of surface chlorophyll a (chl-a) concentration have suggested a strong pulse of phytoplankton production in early summer while the ice cover rapidly ablates; however, no study to date has directly observed primary production, neither of ice algae nor of phytoplankton, in central Hudson Bay during this period. During the 6-week scientific cruise on board the CCGS Amundsen in June–July 2018 as part of the BaySys project, we sought to fill this gap by collecting a dataset on algal biomass, species composition, and primary production in the pack ice and water column along transects across the bay, contrasting ice-covered with ice-free conditions. Simultaneously, light and nutrient availability was measured at the ice bottom and along vertical profiles. Results show large patches of high chl-a concentration in the pack ice in northeastern Hudson Bay, in comparison to low ice-algal chl-a concentrations at the bottom of thick (>2 m), sediment-laden ice in southern Hudson Bay. Under-ice chl-a concentrations were consistently low throughout most of the bay. However, a large phytoplankton bloom with a pronounced subsurface chlorophyll maximum (SCM) was observed in northwestern Hudson Bay, which was ice-free during the sampling period. The location of the SCM in the water column, which was shallower near shore and at the ice edge compared to the center of the ice-free region, indicated a nutrient depletion in the surface water layer. Here, we present initial results, seeking to place an estimate of central Hudson Bay primary production in the bottom ice and water column for this early summer period.


Role of winter storms on the evolution of sea ice in the Atlantic sector of the Arctic

Mats Granskog, Polona Itkin, Robert M. Graham, the N-ICE team

Corresponding author: Mats Granskog

Corresponding author e-mail: mats@npolar.no

Regionality in sea-ice and snow conditions in the Arctic Ocean are often overlooked. Recent work from the Norwegian young sea-ice (N-ICE2015) expedition in the area north of Svalbard showcases how sea ice in this region is frequently affected by passing winter storms, which affect the sea-ice system in a number of ways that are likely unique to this region. The oceanic conditions, especially with relatively shallow warm Atlantic water, affects the sea-ice system and sets the region apart from other regions of the Arctic. Here we make a synthesis of the first comprehensive dataset of winter observations from the N-ICE2015 expedition, which took place in a thin first-and second-year sea-ice regime in the ‘stormy’ Atlantic sector. The multidisciplinary dataset includes atmosphere, snow, sea-ice, ocean and ecosystem observations from a drifting ice station from winter to spring (Jan–Jun). We use these observations to illustrate the mechanisms through which winter storms affect the coupled Arctic sea-ice system. These short-lived and episodic synoptic-scale events transport pulses of heat and moisture into the Arctic, which temporarily reduces radiative cooling and hence ice growth. Cumulative snowfall from each sequential storm acts to deepen the snow pack, being thicker than in other regions of the Arctic, which insulates the sea-ice and inhibits ice growth for the remaining winter season. In addition, strong winds fracture the ice cover, enhance ice drift and thus also increase ocean heat fluxes, which also can reduce thermodynamic ice growth. The heavy snow load induces flooding of the ice, and snow-ice formation, which likely is much more widespread in this part of the Arctic than elsewhere, due to the combination of heavy snowfall and thinning of the sea-ice cover. Flooding also induced phytoplankton growth at the bottom of the snow pack, which is a widespread phenomenon in the Antarctic but might become more prevalent in the Arctic with thinning ice, especially in the Atlantic sector. In spring, an early under-ice phytoplankton bloom developed, despite average snow depths of the order of 0.3–0.5 m. The broken-up ice pack and prevalence of leads in the region allowed enough light to penetrate to the ocean. Thus the legacy of Arctic winter storms for sea-ice and the ice-associated ecosystem in the Atlantic Sector lasts far beyond their short lifespan.


The seasonal ice zone domain of the Arctic Ocean: an overview

Paul Wassmann

Corresponding author: Paul Wassmann

Corresponding author e-mail: paul.wassmann@uit.no

Over the last decades the seasonal ice zone domain (SIZD) has developed into the most dominating of the contiguous domains in the Arctic Ocean. It comprises the cumulative area that is temporarily ice-covered at any given time within a year, now comprising about two-thirds of the total area of the Arctic Ocean (similar to the size of Europe). Ice and variable amounts of snow limit the radiation and thus photosynthesis. Ice and stratification by meltwater reduce the impact of wind on vertical mixing and support the frequently explosive ice-edge bloom. Global climate change will in decades to come have immense consequences in the SIZD, with climate feedbacks impacting the human living conditions of the entire northern hemisphere. As the knowledge base for physical and biogeochemical SIZD dynamics is limited, the lack ofinformation is particularly discomforting. Outside the landfast ice zone is the seasonal ice zone of the pack ice that is free-floating and not connected to land and that moves generally northward with the melting season. Its circumference of far more than 10 000 km is now rapidly decreasing so that it it possible to circumnavigate and synoptically investigate it during a single cruise. The marginal ice zone is the biologically especially active fringe of the SIZD and as the ratio of plankton and sea-ice algae and the bloom phenology changes so do the zooplankton and pelagic–benthic cycles. To better comprehend the dynamic nature of the SIZD and some of the time and space variability across the Arctic Ocean, various conceptual models reflecting ice over and thickness, light and plankton blooms are presented. Particular emphasis will be placed upon the biological significance of the role of the SIZD during winter. The phenology of ice and phytoplankton varies significantly across the Arctic Ocean with the largest blooms, the greatest ice-melt and increase in new production on the Eurasian side.


Representation of Antarctic climate in CESM2: sea-ice variability and atmosphere interactions

Marilyn Raphael, Marika Holland, Laura Landrum, Erick Kim

Corresponding author: Marilyn Raphael

Corresponding author e-mail: raphael@geog.ucla.edu

The ability of climate models to represent the temporal and spatial variability in Antarctic sea ice and its links to the atmospheric circulation is fundamental to the advancement of our understanding of this important component of the Antarctic climate. Research has shown that Antarctic sea ice, a key component of the Southern Hemisphere climate system, is strongly influenced by several large-scale modes of atmospheric circulation. Antarctic sea-ice variability is spatially heterogeneous, suggesting differences in strength, timing and region of influence of these atmospheric circulation modes. We examine the ability of the CESM2 to reproduce the observed regions of variability in sea-ice concentration around Antarctica and use spectral analysis to evaluate the timing of the advance and retreat of sea ice in these regions. These regions of sea-ice variability are then linked statistically to known atmospheric modes of circulation – the Southern Hemisphere Annular Mode, Zonal Wave Three and the Amundsen Sea Low. The results are compared with observations as well as with previously published work.


A new passive microwave melt-onset retrieval approach and comparison to existing methods

Stephen Marshall, K. Andrea Scott, Randall Scharien

Corresponding author: K. Andrea Scott

Corresponding author e-mail: ka3scott@uwaterloo.ca

The determination of melt onset (MO) in the Canadian Arctic Archipelago (CAA) presents unique challenges, as the islands and narrow bays of the region present an obstacle to the passive microwave MO methods currently available. Current passive microwave MO retrieval algorithms for the Arctic utilize daily averaged 19 GHz and 37 GHz data from the multi-channel microwave radiometer (SMMR) and/or the special sensor microwave/imager (SSM/I). The development of a new passive microwave MO method capable of using higher-resolution data is desirable. The MO method described here uses higher-resolution data from the 37 GHz vertically polarized channel on the advanced microwave scanning radiometers (AMSR-E and AMSR-2). The MO detection methodology differs from those presented previously for the Arctic primarily in that it does not use a fixed threshold of a brightness temperature parameter. Instead, it determines the MO date based on the distribution of dates corresponding to when a range of brightness temperature variability thresholds are exceeded. The method also uses swath data instead of daily averaged brightness temperatures, which is found to lead to improved melt detection. Two current passive microwave MO methods are compared and evaluated for applicability in the CAA alongside the new passive microwave MO method. The new method provides MO dates at a higher spatial resolution than earlier methods in addition to higher correlation with MO dates from surface air-temperature reanalyses.


Pancake sea-ice dynamic and kinematic properties observed using stereo video

Madison Smith, Jim Thomson

Corresponding author: Madison Smith

Corresponding author e-mail: mmsmith@uw.edu

Surface waves in the marginal ice zone drive the convergence and collision of floes with each passing wave. This motion is responsible for the formation of pancake-ice floes, and more broadly the transfer of kinetic energy from the wave field into the ice and ocean. This work uses shipboard stereo video from the western Arctic to provide the first in-situ observations of phase-resolved ice motion under wave forcing. A number of previous studies have developed theoretical models to describe the wave-driven motion of ice floes and the resulting ice growth and energy transfer, but these models have not been validated in the field. Here, stereo video is used to obtain wave motion following methods developed for use in open water. Rectified images are then used to determine statistical estimates of floe velocities from the change in concentration as floes converge and diverge with wave motion. These observations of ice velocity across a range of wave conditions are compared with predictions from three different theoretical models. In the conditions observed, where floe radii are much smaller than wavelengths, average floe velocities are well predicted by all three models. However, floe motion typically lags behind wave orbital motion, which is only well predicted by one of the models. This quantification of model skill enables improved estimates and parameterizations of pancake-floe motion and collision, including pancake formation. As an example, we present an updated parameterization of the lateral growth of pancake floes based on the water temperature and wave conditions. These results improve understanding of small-scale dynamics of the marginal-ice zone under conditions that are likely to continue expanding in the Arctic Ocean.


Impact of a floe size distribution on the fragmentation and melting of the seasonal Arctic sea-ice cover

Adam Bateson, Daniel Feltham, David Schröder, Lucia Hosekova, Jeff Ridley, Yevgeny Aksenov

Corresponding author: Adam Bateson

Corresponding author e-mail: a.w.bateson@pgr.reading.ac.uk

The sea-ice cover of the Arctic ocean is composed of discrete areas of sea ice called floes. Climate models currently assume that these floes all have a uniform diameter of 300 m. This simplification is particularly likely to affect model simulations of the marginal ice zone, a region at the edge of the sea-ice cover with between 15% and 80% sea-ice cover. In this region floes are heavily fragmented by several processes, including incoming ocean waves. Floe size impacts lateral melt rates, momentum transfer between the sea ice, atmosphere and ocean, and sea-ice rheology. Observations generally show that a truncated power law produces a good fit to floe-size distribution data. In this study, the floe-size distribution is represented as a truncated power law defined by three key parameters: minimum floe size, maximum floe size, and exponent. The exponent and minimum floe size are fixed; however, the maximum floe size evolves independently within individual grid cells in response to lateral melting, wave induced fragmentation and during freeze-up of the Arctic Ocean. This distribution is implemented within the CICE sea-ice model coupled to a prognostic ocean mixed layer. We present results to show that the use of the power-law-derived floe-size distribution has a spatially and temporally dependent impact on the sea ice, in particular increasing the role of the marginal ice zone in seasonal sea-ice loss. This feature is important in correcting existing biases within sea-ice models. In addition, we show a stronger model sensitivity to floe-size distribution parameters than other parameters used to calculate the volume of lateral melt, justifying the focus on floe-size distribution in model development. We also explore floe-size distribution–mixed layer interactions within a coupled CICE-NEMO model. Finally, we compare two approaches to representing the floe-size distribution within models: the approach described here and a prognostic floe size–thickness distribution.


Albedo and surface energy balance of melt ponds during early autumn in the central Arctic

Tao li, Jialiang Zhu

Corresponding author: Tao Li

Corresponding author e-mail: litaoocean@ouc.edu.cn

Arctic sea ice has been declining remarkably over the past few decades, which is attributed to the dynamical and thermodynamical processes in the Arctic climate. Radiation and the energy balance of the melt ponds on the Arctic sea ice has been considered as a major component of the Arctic ice climate. In order to better understand the thermodynamical process on the melt-pond surface, in-situ observations of the melt-pond radiation in early autumn have been carried out in the central Arctic during the Chinese Arctic surveys in 2012, 2014 and 2016. Based on the surface state of the melt pond during the freezing season, we categorized the melt ponds into five types: partially frozen melt pond (PMP), grey-lid melt pond (GMP), blue-lid melt pond (BMP), white-lid melt pond (WMP) and snow-covered melt pond (SMP). The albedos of the five types of melt pond are 0.19 ± 0.04 for the PMP, 0.28 ± 0.08 for the GMP, 0.33 ± 0.10 for the BMP, 0.50 ± 0.05 for the WMP and 0.72 ± 0.14 for the SMP. In contrast with the melt-pond albedo in melting season, freezing speed, meltwater depth, and thickness of the ice underlying the melt pond, as well as the snow cover on the ice lid, strongly affected the albedo during the freezing months. Meanwhile, the spatial distribution of the albedo of melt ponds in the autumn has shown a significant heterogeneity as higher albedo occurred in the far northern area, which was partly attributed to snow falling while the melt ponds were freezing. The thickness of the ice lid in the surface of the melt pond is another important factor affecting albedo, which is even more complicated due to the ice lid growing. Previous studies from ML96 have shown that the albedo of melt ponds with an ice-lid thickness of 0–0.05 m decreased with the lid thickness while that in ponds with a lid thickness of 0–0.01 m declined faster In contrast with ML96, the albedo of the melt ponds with ice-lid thickness of more than 0.05 m in our study increased with the thickness. The difference in albedo variation with the various ice-lid thickness was mainly attributed to the contribution of the underlying ice of melt ponds by which the solar shortwave radiation was reflected and scattered.


An image-processing method for retrieving ice type and thickness from ship-borne camera observations

Yasuhiro Tanaka

Corresponding author: Yasuhiro Tanaka

Corresponding author e-mail: tanaka.yasuhiro@jaxa.jp

Sea ice is a prerequisite factor to understand the Earth’s climate system. Satellite remote-sensing observations are effective for estimating Arctic sea-ice conditions (such as concentration, type and thickness). To evaluate the satellite-derived sea-ice condition retrieval better, ground-truth observations are needed. Ground-truth observations from ships in polar regions have been performed under Antarctic Sea Ice Processes and Climate (ASPeCt) protocols. However, the sea-ice parameters are caused by different subjective retrievals of the sea-ice conditions around the ship. Tanaka et al. (2016) proposed an image-processing method for retrieving the sea-ice (and melt-pond) concentrations using images obtained from a forward-looking ship-borne optical camera. To separate three possible surface conditions, they found peaks in the curves fit to histograms. Additionally, Tanka (2019) classified the sea-ice conditions as open water, thin ice, first-year ice and old ice using images obtained from a downward-looking camera the ship-borne optical camera in the Arctic Ocean during late summer and early autumn in 2014. However, an image-processing method for estimating sea-ice thickness along with ice type is yet to be developed. In this study, we have introduced an image-processing method for classifying ice type using images obtained from a downward-looking camera during the Joint Ocean Ice Study in 2014 on the Canadian Coast Guard ship Louis S. St-Laurent. In the ice-type classification, open water was identified based on the average level of the red channel, whereas the thin-ice and thick-(first-year or old) ice types are identified based on the average hue. Furthermore, differences between the red and blue levels as well as between green and blue levels can be used to distinguish between first-year ice and old ice. Additionally, we introduce an image-processing method for estimating ice thickness along with ice type using the downward-looking camera images.


Characterizing the snow distribution over sea ice: analysis of in-situ, airborne and satellite remote-sensing techniques

Marissa Dattler, Sinéad L. Farrell

Corresponding author: Marissa Dattler

Corresponding author e-mail: marissadattler@gmail.com

The distribution of snow cover over sea ice has several important effects on the Arctic climate system. Thicker snow provides greater insulation for sea ice from cold air, inhibiting winter ice growth. Additionally, snow that persists into the summer season can delay melt onset due to its high albedo relative to bare ice. The snow-depth distribution is currently challenging to measure or approximate using either remote sensing or modeling techniques. This leads to large uncertainties in snow depth across Arctic sea ice and hence significant errors in freeboard-derived sea-ice thickness. In this research, we analyze the winter snow distribution over a range of sea-ice types using in-situ snow depth measurements from several field campaigns over a 20-year period. Dividing the snow-depth datasets by ice roughness and ice age, we compare snow-depth probability density functions to distributions found commonly in nature (e.g. normal, lognormal, log-logistic and Weibull). Across all field study sites, snow on level ice was lognormally or log-logistically distributed, whereas the snow cover on rougher, hummocky ice tended to be better fitted by a Weibull distribution. We extend the snow-depth distribution analyses to the regional scale using IceBridge airborne snow radar observations. When classified by ice type, we find that snow-depth distributions derived from the airborne measurements are consistent with the in-situ results. Finally, using coincident satellite remote-sensing data, from CryoSat-2 and ICESat-2, we investigate the feasibility of deriving the snow-depth distribution from satellite altimetry. We compare the physical properties of the ice floe to its snow-depth distribution. In this process, we attempt to identify remote-sensing signatures that might be indicative of snow depth or the snow-depth distribution. Developing these relationships between in-situ, airborne and remote-sensing measurements can provide insight into the snow-depth distribution over sea-ice floes without the need for widespread in-situ observations.


Impact of CryoSat-2 ice-thickness initialization on seasonal Arctic ice prediction

Richard Allard, Neil Barton, Nathan Kurtz, Li Li, E. Joseph Metzger, Michael Phelps, Ole Martin Smedstad

Corresponding author: Richard Allard

Corresponding author e-mail: richard.allard@nrlssc.navy.mil

A series of twin experiments are performed to predict the September 2018 minimum ice extent using the fully coupled US Navy Earth System Prediction Capability (ESPC), which consists of the Navy Global Environmental Model (NAVGEM), the Hybrid Coordinate Ocean Model, and the Community Ice Code. In the control run, ensemble forecasts are initialized from the operational US Navy Global Ocean Forecasting System (GOFS) 3.1 for the ocean and sea ice using the Navy Coupled Ocean Data Assimilation (NCODA) system that assimilated SSMIS and AMSR2 sea ice concentration products. Atmospheric initial conditions are from the NAVGEM using the Naval Research Laboratory Atmospheric Variational Data Assimilation System. Another set of forecasts are initialized with ice thickness derived from CryoSat-2 (CS2) for early May 2018. Both sets of forecasts are performed with 10 time-lagged ensemble members beginning 1–10 May 2018. We present results for the predicted versus observed September sea-ice minimum extent to examine the impact of satellite-derived ice-thickness initialization. We find an improvement in the September minimum ice extent for the Beaufort/Chukchi, Laptev and Kara Seas versus the control run without CS2 initialization. The Navy ESPC ensemble mean September 2018 minimum sea-ice extent initialized with GOFS 3.1 ice thickness was over-predicted by 0.68 Mkm2 (5.27 Mkm2) versus the ensemble set of forecasts initialized with CS2 ice thickness, which had an error of 0.40 Mkm2 (4.99 Mkm2), a 59% reduction in error. Comparison against WHOI Upward Looking Sonar (ULS) ice thickness reveals significant improvement at mooring ‘A’ showing predictive skill for more than 75 days. Ice concentration from AMSR2 and SSMIS also show improvement at this location; however, the open water as shown by the ULS beginning in mid-August is not captured by either of the ensemble forecasts.


UAV observations of the influence of sediment on sea-ice surface deformation and melt rates during the summer melt period

Madison Harasyn, Dustin Isleifson, Ryan Galley, David Barber

Corresponding author: Madison Harasyn

Corresponding author e-mail: harasynm@myumanitoba.ca

Sediment-laden sea ice has been previously observed in the Chukchi and Beaufort Seas, Fram Strait, Foxe Basin and southern Hudson Bay. Sediments can become entrained in sea ice during formation in coastal regions where required conditions exist (i.e. turbidity, strong winds, fine sediments), spreading and reaching areas hundreds of kilometers offshore under the influence of winds. Sediments influence the surface albedo of sea ice, decreasing shortwave reflectance in all optical bands. The decrease in surface albedo is related to the total sediment concentration on the surface, with higher sediment concentrations having a lower surface reflectance. Here, we present results from an experimental study, in which we investigated how the total sediment loading on a sea-ice surface affects the melt rate and small-scale surface topography. This past spring, an experiment was conducted at the Sea Ice Experimental Research Facility (SERF) at the University of Manitoba to replicate sea-ice surface melt rates under the influence of sediment presence. Sediments were distributed across the ice surface to replicate varying sediment concentrations. An unmanned aerial vehicle (UAV) was flown at a low elevation over the ice surface throughout the melt period to capture high-resolution optical imagery of the surface and to generate a time series of digital elevation models of the surface. Supervised classification was used to identify areas of varying sediment concentration to map the dispersion of surface sediments throughout ice melt. Rates of surface melt were determined using raster subtraction methods, which were compared to corresponding sediment concentrations. We will present in depth methods of data analysis as well as preliminary results from this analysis.


‘Snow ice’ contribution to the structure of sea ice in the Amundsen Sea, Antarctica

Lijun Tian, Yongli Gao, Blake Weissling, Stephen Ackley

Corresponding author: Stephen Ackley

Corresponding author e-mail: stephen.ackley@utsa.edu

When snowpack on sea ice is heavy and thick enough to depress the top surface below sea level, a slush layer or slurry is formed through the mixture of seawater or brine and snow. In the Antarctic, overall thinner ice cover and higher snow accumulation rates result in the widespread occurrence of surface flooding and ‘snow-ice’ formation. The snow ice is incorporated into the ice cover after subsequent periods of freezing, and it becomes exceedingly difficult to distinguish such snow ice from similarly fine-grained consolidated frazil ice. However, the stark contrast in the stable-isotope signatures of snow, which is greatly depleted in the heavy stable isotopes, and ice grown from sea water with its undepleted composition, can help determine the contributions of snow ice to the total ice thickness. This work investigates the stable-isotope composition, salinity and ice texture of pack ice in the Amundsen Sea. A total of eight ice cores were obtained from the Amundsen Sea during the Oden Southern Ocean 2010/11 expedition from late December 2010 and January 2011. These ice cores vary in depth, the shallowest being 60 cm and the deepest 190 cm. Each sea-ice core was cut into 10 cm contiguous subsections. Stable-oxygen and hydrogen-isotope measurements of 92 sea-ice samples were conducted on a Picarro L2130-isotopic water analyzer (cavity ring-down laser spectroscopy technology). The total contribution of snow ice to the pack ice in the Amundsen Sea was 20%. However, this procedure to calculate the percentage of core length that contains meteoric water might be biased due to the normally low-resolution isotopic measurements for sea-ice cores. High-resolution isotopic measurements within one 10 cm subsection demonstrated there are probably both snow ice and frazil ice portions within one subsection. A meteoric water fraction in these sea-ice cores based instead on an updated isotope-mixing model will also be presented. Depth profiles of salinity and ice texture were also described to serve as illustrations of the structures of these pack-ice cores in the Amundsen Sea. These results will be compared to previous investigations of snow-ice occurrence around Antarctica.


The role of ice kinematics in the state of Antarctic sea ice

Tian Tian, Petra Heil, Alexander Fraser, Maxim Nikurashin

Corresponding author: Tian Tian

Corresponding author e-mail: tian.tian@utas.edu.au

Sea ice is a crucial component of the Earth system due to its roles in controlling energy and moisture transfer between the polar ocean and atmosphere, the role of the ice-albedo feedback for the polar amplification, and its impact on marine ecosystems and bio-geochemical activity. Due to the different geographical settings, distributions in Arctic and Antarctic sea-ice motion and deformation differ considerably. In previous works, these ice kinematics has been derived from satellite-borne and in-situ buoy data. However, few cross-sensor studies have been conducted to derive improved knowledge of ice kinematics and how to represent it in numerical sea-ice models. Here we seek to characterize and quantify sea-ice motion and deformation in the Southern Ocean, with a view to determining the drivers of these processes and estimating their susceptibility to change under predicted atmospheric and oceanic change. Three types of data with differing resolution have been compared in our research. These are: 1) sea-ice buoy information in the Weddell Sea from 2013 onwards; 2) synthetic-aperture-radar-based sea-ice motion data (5–150 m resolution), and 3) passive-microwave-radiometer-derived sea-ice motion data with resolution about 30–60 km (daily; circumpolar). The merged ice-motion information will be a valuable dataset for validation of high-resolution sea-ice models. From intercomparison of these three observational datasets, we deduce that the temporal and spatial variability of Antarctic sea-ice strength is anisotropic. This data intercomparison provides an impetus to derive an updated climatology of Antarctic sea-ice motion and deformation.


The impact on Arctic sea ice of increased ice–ocean drag caused by ocean internal waves

Daniela Flocco, Daniel Feltham, David Schroeder, Yevgeny Aksenov, Antony Siahan, Michel Tsamados

Corresponding author: Daniela Flocco

Corresponding author e-mail: d.flocco@reading.ac.uk

The phenomenon of dead waters was first observed by Nansen in 1893 when navigating through polar waters; it is caused by the ship’s hull inducing internal waves in the ocean that radiate momentum away from the ship, effectively increasing the ocean’s drag on the ship. The rough topography of the underside of sea ice also generates internal waves as sea ice drifts over the stratified ocean, increasing the total ice–ocean drag. A parameterization of the impact of internal waves on momentum transfer at the sea-ice–ocean interface has been developed and implemented in a sea-ice model (CICE) for the first time. The parameterization comes from a previous study by McPhee, which we have adjusted to account for the presence of keels deeper than the mixed layer depth. The extra ice–ocean drag from internal waves is stronger for shallow mixed-layer depth and large density jump at the pycnocline, and is a function of the strength of the stratification beneath the ocean mixed layer and geometry of the ice interface. We consider the contribution to internal wave drag from both ridged and non-ridged ice. We present results from a coupled sea-ice–ocean model (NEMO-CICE) where the internal wave drag has been implemented. Simulations were run from 1980 to 2016. We show results demonstrating the regional effect of internal wave drag on emergent Arctic sea-ice characteristics such as thickness, motion and deformation. In particular, we observe up to 15% increase in sea-ice thickness and extent in the Canadian Arctic due to an overall slow-down of the ice drift by 10% causing decrease of the ice to ocean heat flux.


Sea ice and atmosphere interactions and predictability: preliminary results from CMIP6

Daniela Flocco, Ed Hawkins, Daniel Feltham

Corresponding author: Daniela Flocco

Corresponding author e-mail: d.flocco@reading.ac.uk

Arctic sea-ice extent has declined in the past 30 years. Aside from the global impact on climate change, there is a growing interest in regional information on sea-ice presence and its impact on oceanic and atmospheric patterns. There is a growing need for seasonal-to-decadal timescale climate forecasts to help inform local communities and industry stakeholders. The APPLICATE project seeks to better understand the role of natural climate fluctuations in producing recent Arctic sea-ice changes on these timescales, and whether they are predictable. Here we present preliminary results from a CMIP6 control run to investigate relationships between sea-ice state variables and atmospheric patterns to understand their reciprocal influences. The understanding of these relationships will help constrain sea-ice projections and inform stakeholders.


Sensitivity of atmospheric parameters on snow and sea-ice mass balance during spring and melting season in the central Arctic

Bin Cheng, Timo Vihma, Timo Palo, Marcel Nicolaus, Sebastian Gerland, Laura Rontu, Jari Haapala, Donald Perovich

Corresponding author: Bin Cheng

Corresponding author e-mail: bin.cheng@fmi.fi

Snow depth and sea-ice thickness were observed by an ice mass-balance buoy (IMB) along the French schooner Tara ice-drift station during the International Polar Year in 2007. High-quality in-situ meteorological parameters and fluxes were observed during spring and melting season (May–August). The operational analyses and short-term forecasts from two numerical weather prediction (NWP) models (ECMWF and HIRLAM) were obtained in the domains covering the Tara drift trajectory and meteorological parameters were extracted at positions of the Tara ice-drift station. The results of the NWP models were compared with observations. A one-dimensional thermodynamic snow/ice model (HIGHTSI) was applied to calculate snow and ice mass balance applying in-situ observations as well as results of NWP models as external forcing. The modelled snow accumulation, controlled by NWP-based precipitation, was in line with observed snow depth. The HIGHTSI reproduced snowmelt well: the modelled snowmelt onset was only 1–2 days later than observed, and the first snow-free day was within 2–3 days of the observed one. The simulated and observed basal ice growth was almost identical during freezing conditions. Under cold conditions, the evolution of the vertical temperature profile in snow and ice was better when the model was forced by in-situ observations instead of NWP results. During the melting period, the nonlinear ice temperature profile was successfully modelled with both forcing options. During spring and melting season, HIGHTSI sensitivity modelling indicated that the order of most influential atmospheric forcing factors on snow and sea-ice mass balance are downward longwave radiative flux followed by air temperature, downward shortwave radiative flux, wind speed and moisture. The HIGHTSI model forced by in-situ observations yielded better results for surface temperature than the ECMWF and HIRLAM models themselves.


On-ice Arctic sea-ice thickness archive – first analysis

Benjamin Holt, Axel Schweiger, Christian Haas, Ryan Avila

Corresponding author: Benjamin Holt

Corresponding author e-mail: benjamin.m.holt@jpl.nasa.gov

In this study, we will present analysis of on-ice sea-ice thickness measurements of the Arctic Ocean (On-Ice archive), recently archived at the National Snow and Ice Data Center. This archive consists of measurements primarily made by ice auger, ice coring devices and surface-based electromagnetic induction (EM). To date, the archive starts with the USS Jeanette cruise in 1879–1881, includes a handful of measurements from the Fram Expedition in 1894/95, skips to the Maud Expedition in 1922–1924, when the data spans through the decades up to the most recent record from 2015. These methods of ice-thickness measurement are the most accurate available compared to upward-looking sonar and remote-sensing derivations from aircraft and satellites. Also, these measurements are the only source of ice-thickness information prior to 1958 when ice-draft measurements from US submarines began. Currently, the archive includes over 66 000 individual measurements of sea-ice thickness obtained within the Arctic Ocean. When available, the archive includes associated freeboard and snow-depth measurements, and often sea-ice type. The analysis will describe annual and decadal mean values of sea-ice thickness and snow depth, in addition to seasonal analysis. To a limited degree, we will compare these results with other data collections, including the Unified Ice Thickness Climate Data Record (ThickCDR). While comparatively sparse, this focused collection of on-ice historical and contemporary data is of particular value in examining sea-ice variability over time periods prior to the rapid decline observed over the last four decades. By extending the sea-ice thickness record over more than a century, these measurements add more relevancy to the more recent observations from satellite and airborne measurements being used for the analysis of Arctic variability over timescales of multiple decades. In addition, they are useful for the validation of sea-ice models and remote sensing techniques by providing data for areas and times that have little other information.


Estimates of net community production during the spring ice-edge bloom in Baffin Bay

Tonya Burgers, Jean-Éric Tremblay, Tim Papakyriakou

Corresponding author: Tonya Burgers

Corresponding author e-mail: tonya.burgers@umanitoba.ca

Net community production (NCP) is a measure of the metabolic balance of an ecosystem, defined as the gross primary production minus community respiration. As such, rates of NCP represent a maximum constraint on carbon export to the deep ocean. Here we present NCP estimates from a series of east–west transects in Baffin Bay, ranging from pre-bloom to post-bloom conditions. Three different approaches were employed to estimate NCP; (1) oxygen mass balance based on continuous surface O2/Ar measurements, (2) seasonal inorganic carbon drawdown from continuous surface pCO2 measurements, and (3) seasonal drawdown of nitrate from discrete samples. Our observations capture the presence of a significant ice-edge bloom and associated NCP in the ice-free waters of eastern Baffin Bay. However, in western Baffin Bay persistent sea-ice cover and strong stratification result in very low NCP. The comparison of NCP estimates from various approaches provides additional insight into the spatial and temporal progression of the bloom across Baffin Bay. Comparing residence time integrated NCP(O2/Ar) against seasonally integrated NCP(pCO2) we find the slope of their relationship to be a function of days of open water, with measurements within the marginal ice zone being closest to a 1 : 1 relationship.


Hudson Strait inflow: transport, variability and source waters

Natasha Ridenour, Fiammetta Straneo, James Holte, Yves Gratton, Paul Myers, David Barber

Corresponding author: Natasha Ridenour

Corresponding author e-mail: ridenour@ualberta.ca

Hudson Strait is the main pathway of heat, mass and fresh water between Hudson Bay, the Arctic and the North Atlantic. Flow along the southern coast, a low-saline, baroclinic jet directed towards the North Atlantic, has received more attention due to its potential impact on deep convection in the Labrador Sea. However, details about the westward, barotropic flow along the northern coast of Hudson Strait remain unknown due to the lack of observations. It is thought that the inflow is comprised of waters from Baffin Bay, via the Baffin Island Current, as well as waters from the Labrador Sea. Hudson Strait inflow waters affect the physical and biogeochemical systems of the bay, as well as the marine ecosystem, which supports the livelihoods of many indigenous communities surrounding the Hudson Bay complex. Additionally, identifying the inflow waters will help our understanding of water mass transformations occurring in the bay due to sea-ice growth and melt, and continental runoff, as well as tidal mixing. Here, we present data from two synoptic surveys of the circulation and properties across Hudson Strait from 2008 and 2009, as well as data from four moorings deployed in the strait. Three moorings were deployed on the northern side of the strait to map the inflow, and one was deployed on the southern side of the strait to map the outflow. We investigate the seasonality of the inflow and identify its source waters. Based on these data and knowledge of the outflow from previous moored deployments, we present the first year-round estimates of heat, mass and salt transport in the Hudson Strait inflow.


The contributions of ice algae and phytoplankton to the total primary production at the seasonal sea-ice zone in Dease Strait, Canadian Arctic Archipelago

Kwanwoo Kim, Sang Heon Lee, Sun-Yong Ha, C.J. Mundy

Corresponding author: Sang Heon Lee

Corresponding author e-mail: sanglee@pusan.ac.kr

The primary production of ice algae and phytoplankton is an important carbon and energy source for higher trophic levels in the ice-covered Arctic regions. To identify the relative contribution of ice algae and phytoplankton to the total primary production, 13C enrichment experiments were conducted to determine carbon uptake rates of ice algae and phytoplankton in Dease Strait, Canadian Arctic. In addition, the bottom 10 cm segment of sea ice and underneath water samples were collected for analysis of chlorophyll a (chl-a) concentration and macromolecular composition of ice algae and phytoplankton. Total chl-a concentration in the bottom 10 cm ice section ranged from 9.0 μg L–1 to 65.3 μg L–1 with a mean of 32.0 ± 16.1 μg L–1. In the case of the water samples, significantly low chl-a values were observed during our study period, ranging from 0.29–6.72 μg L–1 with an average of 1.27 μg L–1 (± 1.55 μg L–1). The average concentrations of carbohydrates, proteins and lipids constituting particulate organic matter (POM) in the bottom 10 cm segment were 1695.3 μg L–1 (± 643.8 μg L–1), 1676.4 μg L–1 (± 630.1 μg L–1) and 3606.3 μg L–1 (± 1521.4 μg L–1), respectively. In comparison, the average concentrations of carbohydrates, proteins and lipids of phytoplankton were 355.1 μg L–1 (± 181.0 μg L–1), 76.1 μg L–1 (± 35.3 μg L–1) and 156.9 μg L–1 (± 84.2 μg L–1), respectively. The carbon-uptake rate of ice algae in the bottom 10 cm section ranged from 0.571–4.114 mg C m–3 h–1 with an average of 1.703 mg C m–3 h–1 (± 1.024 mg C m–3 h–1) from the productivity stations. In comparison, the carbon-uptake rates of phytoplankton showed relatively low values (0.074–1.508 mg C m–3 h–1).


Melt ponds and sea-ice albedo; comparison of MIZMAS model with CLARA-A2 satellite observations

Aku Riihelä, Jinlun Zhang

Corresponding author: Aku Riihelä

Corresponding author e-mail: aku.riihela@fmi.fi

Alongside sea-ice concentration, melt ponds are an important driver of the area-average surface albedo of Arctic sea ice, and thus play a key role in the regulation of the surface radiative energy budget during each melting season. Recent advances in modeling have enabled estimates for melt-pond coverage and property estimates with decadal coverage, such as those from the MIZMAS model. On the other hand, concurrent advances in satellite remote-sensing techniques have enabled the creation of stable, multidecadal and spatially comprehensive estimates of the Arctic Ocean’s surface albedo, such as the CLARA-A2 SAL dataset, covering 1982–2015 from the intercalibrated AVHRR family of optical imagers. The objective of this study is to first investigate the sea-ice-albedo trends from the latest CLARA dataset, and then to compare these findings with MIZMAS modeled melt-pond coverages and pond radiative properties to delineate and quantify the impacts of decreasing sea-ice concentration and increasing melt-pond coverage on the surface albedo of the ice. The veracity of CLARA-A2 SAL for these studies is ensured by intercomparison with the MEDEA dataset, which contains carefully reconstructed estimates for ice concentration and melt-pond fraction at multi-kilometer spatial scales from declassified reconnaissance satellite imagery. The results demonstrate that CLARA sea-ice albedo agrees well with the MEDEA data and that both sea-ice concentration and melt-pond coverage are necessary parameters to explain the observed albedo. In the comparison between CLARA and MIZMAS, we will show that systematic relationships exist between observed albedo and modeled melt-pond coverage, and that observed sea-ice albedo decreases are in line with increasing melt-pond-area coverage over the high-concentration sea-ice zone. The connection between mean melt-pond depths and observed surface albedo will also be explored.


Reconstructing NCP from O2 and Ar concentrations in winter sea ice of the Ross Sea,Antarctica

Sarah Wauthy, Bruno Delille, Stephen Ackley, Ted Maksym, Sharon Stammerjohn, Jean-Louis Tison

Corresponding author: Jean-Louis Tison

Corresponding author e-mail: jtison@ulb.ac.be

Measuring the net community production (NCP) – the balance between O2 production by primary producers and the respiration of the entire community – in sea ice is very challenging due to its heterogeneous nature (mixture of pure ice, brines, gas bubbles and salts). NCP measurements are also scarce in sea ice, especially in winter. Here we present a reconstruction of the NCP levels by measuring the concentrations of O2 and Ar at high resolution using gas chromatography in sea-ice cores collected during the PIPERS (polynyas, ice production and seasonal evolution in the Ross Sea) field project. This is one of the rare projects to be conducted in the Ross Sea during the austral winter. Two main conclusions can be drawn from this study: it has been possible to dissociate the biotic and abiotic controls on O2 concentrations in sea ice and therefore to reconstruct the levels of NCP during this winter period. The discrimination of abiotic and biotic controls is based on the use of O2/Ar ratios because O2 concentration is modified by physical processes (temperature and salinity changes, brine convection) and by biological activity (photosynthesis and respiration), whereas Ar is only influenced by physical processes. A dominance of the physical processes was highlighted in most of the stations analyzed, with a contrast between the polynya stations, where negligible traces of biological activity could be identified, and non-negligible biological activity observed in the center of the Ross Sea and along the coast. To reconstruct the levels of NCP, a multidisciplinary approach was used because of the spatial and temporal variability of the cores sampled during the mission. Estimates of the age of the cores (and thus of the period of biological activity) required the use of satellite data and of a thermodynamic model. The concentrations measured in the ice made it possible to calculate the deviation from the saturation ratio (Δ(O2/Ar)), the equilibrium O2 concentration in the brines and, eventually, the O2 concentration due to biological activity. From the time evolution of the latter, the NCP allowed us to determine the general regime (auto- or heterotrophic) in place. These NCP values are compared to the literature for the winter period and the good correspondence confers reliability to our results. Finally, we compare and contrast NCPs calculated in cores located along the coast and in the marginal ice zone and in cores from the central Ross Sea.


Seasonality of snow depth on Weddell Sea sea ice from snow buoy observations and SNOWPACK simulations

Leonard Rossmann, Stefanie Arndt, Marcel Nicolaus, Louisa von Hülsen, Mahdi Jafari, Michael Lehning, Nander Wever, Lars Kaleschke, Nina Maaß

Corresponding author: Leonard Rossmann

Corresponding author e-mail: leonard.rossmann@awi.de

The snow cover on Antarctic sea ice impacts the energy, mass and momentum balance of the sea-ice cover, which in return strongly influence fluxes between ocean, sea ice and atmosphere. Despite the fact that snow depth is considered to be an essential climate variable, knowledge about snow-cover distribution and properties is still very limited. Here, we present measurements of snow depth and physical snow properties obtained along drift trajectories from autonomous snow buoys. These buoys were deployed during several RV Polarstern cruises in the Weddell Sea from 2014. In addition, we provide insights into internal processes such as snow-ice formation based on the newly developed 1-D multi-layer thermodynamic snow model SNOWPACK sea ice version. The model has two ways of forcing. Either the snow surface height is prescribed or it is forced by precipitation from reanalysis data. Both scenarios are presented in this study. Snow-surface-height-forced simulations reveal that bottom sea-ice melt is the main driver of snow-ice formation, once a critical threshold of 60 cm snow depth is reached. Up to 40% of the sea ice, depending on ocean heat flux, consists of snow ice after 1.5 years of drift. Precipitation-re-analysis-forced simulations expose severe discrepancies between simulated and in-situ snow surface heights. A comparison of the reanalysis products Era-Interim and Era5 from the European Centre for Medium-Range Weather Forecast with Autonomous Weather Stations were performed to find a suitable forcing dataset and its appropriate tuning. This revealed that Era-Interim overestimates precipitation by 25% and underestimates temperature during winter. Era5 overestimates the incoming solar short-wave radiation by 25–50 Wm2 during summer (Nov–Feb). Nevertheless, Era5 qualifies as an appropriate forcing data source if short-wave radiation is reduced during austral summer. Model simulations with tuned atmospheric forcing are within 15 cm of the observed surface heights. The combination between the newly developed model and in-situ snow=height observations from snow buoys enables us to retrieve realistic snow depths on sea ice. The tuned precipitation-forced model provides a solid reference for other snow applications (e.g. space-borne retrievals). Finally, we will show how to proceed with the combination of the different datasets and the model towards a Weddell-Sea-wide snow-depth product.


13 years of sea-ice-draft observations in the Laptev Sea from moored ADCPs and upward-looking sonar: changes, variability and comparison to Earth Observation data

H. Jakob Belter, Thomas Krumpen, Markus Janout, Robert Ricker, Stefan Hendricks, Christian Haas

Corresponding author: H. Jakob Belter

Corresponding author e-mail: jbelter@awi.de

Moored upward-looking acoustic Doppler current profilers (ADCPs) can be used to observe sea-ice draft. While previous studies relied on the availability of auxiliary pressure sensors to measure the instrument depth of the ADCP, we present an adaptive approach that infers instrument depth from the ADCP’s default bottom- track mode measurements of error velocity and range. We demonstrate that this method can be used to obtain daily mean sea-ice draft time series with an estimated uncertainty of 0.1 m. The ADCP-derived ice-draft time series are validated with data from adjacent upward-looking sonar moorings in the Laptev Sea. This new approach provides a low-cost opportunity to derive daily mean ice-draft time series accessing existing ADCP data. Applying this method to ADCP data from the sparsely sampled Laptev Sea allows an extension of mooring-based in-situ ice-draft measurements from 2 to about 13 years. The Laptev Sea is an important region for net ice production and a major contributor to the transpolar drift system. Recent studies show that sea-ice area and volume exports from the Laptev Sea are increasing. This increased export accelerates summer sea-ice retreat in the Laptev Sea and has far-reaching consequences for the entire Arctic sea-ice balance. The-newly acquired Laptev Sea ice-draft data archive is used to analyse seasonal and interannual changes in sea ice thickness from 2003 to 2016. In addition, it provides unique data for the comparison with sea-ice-thickness data records derived from ENVISAT, CryoSat-2 and SMOS satellite measurements in an area where large-scale validation datasets are currently unavailable.


High-resolution Arctic sea-ice modeling in a coupled climate model

Shiming Xu

Corresponding author: Shiming Xu

Corresponding author e-mail: xusm@tsinghua.edu.cn

High-resolution simulation with coupled models is becoming a new norm in both climate studies and operations. Here, we introduce recent activities on high-resolution model development, focusing on sea ice and higher latitudes. The model development is based on a hierarchy of grids, with grid spacings in the Arctic basin ranging from about 30 km down to finer than 3 km. Through the use of massively parallel supercomputers and tailored computational performance tuning, we carry out both ice–ocean and atmosphere–ice coupled experiments. Initial model evaluation of sea- ice simulations is presented, including recent Arctic changes and sea-ice kinematics.


Ocean dynamics and the latitude of the sea-ice edge

Jake Aylmer, David Ferreira, Daniel Feltham

Corresponding author: Jake Aylmer

Corresponding author e-mail: j.r.aylmer@pgr.reading.ac.uk

Projections of sea-ice extent in comprehensive general circulation models exhibit large inter-model spread, leading to large uncertainties in estimates of past and future climate conditions. Simulated sea-ice extent and ocean heat transport (OHT) are correlated across climate models, and various studies provide evidence that OHT is a leading-order constraint on the latitude of the ice edge on climatic time and spatial scales. This suggests that model biases in OHT may be a significant contributor to model spread in sea-ice extent. This highlights the need for an improved physical understanding of how the oceans impact the latitude of the sea-ice edge, from which we may determine how much of the spread in sea-ice extents may be attributed to biases in ocean forcings. We have developed a zonal-average energy-balance model of the coupled atmosphere, ocean and sea-ice system, to explore dynamical constraints on the latitude of the sea-ice edge. This highly idealized model improves upon existing work by better representation of the OHT, partitioning it between a dynamic surface mixed layer and a prescribed, conservative deep-ocean forcing. We show that our model reproduces the qualitative behaviour of CMIP models. The sensitivity of the ice-edge latitude to changes in OHT and how this sensitivity depends on other climate processes is presented.


Canadian scatterometers (CanScats) for snow-covered sea-ice process studies during MOSAiC

John Yackel, Randall Scharien, Claude Duguay, Gunnar Spreen

Corresponding author: John Yackel

Corresponding author e-mail: yackel@ucalgary.ca

The Multidisciplinary Drifting Observatory for the Study of Arctic Climate (MOSAiC) will be a continuous, full-year scientific expedition in the central Arctic ocean, bringing together leading researchers from an international consortium of polar-research institutions committed to understanding the consequences of Arctic climate change, enhancing our knowledge of climate-related processes, and addressing research needs in Earth observation and global climate modelling. The centerpiece platform of MOSAiC is the dedicated operation of the German research vessel RV Polarstern, serving as a central observatory and drifting with the Arctic pack ice for an entire annual cycle from September 2019 to September 2020. MOSAiC represents an unprecedented opportunity to examine and link detailed atmosphere–sea-ice–ocean processes and snow/ice geophysics with multi-frequency and multi-scale remote-sensing data and to improve our ability to utilize space-based Earth-observing technologies for mapping and predicting sea-ice conditions. The Canadian Scatterometers (CanScats) project unites three experienced research teams from the Universities of Calgary, Victoria and Waterloo. CanScats will feature the simultaneous deployment of proven, surface-based, multi-frequency, microwave scatterometer systems (L-, C-, X-, and Ku-bands) on the snow-covered sea ice for most of the MOSAiC campaign. Scatterometer observations will be collected by Canadian and MOSAiC HQP coincident with detailed in-situ and airborne acquired snow and sea-ice geophysical property measurements, complementary microwave radiometer, and multi-sensor satellite observations. This project aims to collect a first-of-its-kind, multi-frequency, continuous-season, microwave backscatter dataset in support of snow and sea-ice geophysical and remote sensing process studies and MOSAiC science objectives.


Methods for predicting partitioning and fate of petroleum hydrocarbons and heterocyclic compounds in a sea-ice environment

Durell Desmond, Georg Schreckenbach, James Xidos, Diana Saltymakova, Dustin Isleifson, David Barber, Gary Stern

Corresponding author: Durell Desmond

Corresponding author e-mail: umdesmod@myumanitoba.ca

Decreases in Arctic sea-ice extent and thickness have led to more open ice conditions, encouraging both ship traffic and oil exploration within the northern Arctic. As a result, the increased potential for accidental oil releases of crude oil or fuel into the Arctic environment threatens the pristine marine environment, its ecosystem, and local inhabitants. Thus, there is a need to develop oil-spill detection and mitigation techniques suitable for ice-covered waters. To this end, more research is required to understand the microscopic behaviors of oil constituents within a sea-ice environment and the resulting impact on the oil’s density and dielectrics, the latter being an essential parameter for remote sensing of oil-contaminated sea ice. Partitioning of compounds found in crude oil or fuel is expected to occur to a variable degree within an Arctic setting. Partitioning through sea ice towards the surface (air–ice interface), or the subsurface (water–ice interface), is dependent upon the physiochemical properties and porosity of the ice, the composition of the subsurface water, air temperature, and the properties of the individual compounds found in crude oil. Both scenarios are feasible, allowing for dissolution into the water column or evaporation from the top of the ice into the atmosphere even at lower temperatures. As such, it is essential to investigate the role that these weathering processes play in altering the oil’s composition and the resulting impact on the dielectrics of the oil itself and the sea ice. Computational quantum chemistry was used to simulate the effects of evaporation, dissolution and partitioning within sea ice. Vapor pressures, solubilities, densities, octanol–water partition coefficients, and dielectric constants were calculated using quantum chemistry and thermodynamics for pure liquid solutes (oil constituents) of interest. These calculations incorporated experimentally measured temperatures and salinities taken throughout oil-in-ice mesocosm experiments conducted at the University of Manitoba during 2016–18. The absolute and relative accuracies of these models, as well as their potential for predicting the relative movements of oil constituents, were assessed. Conclusions were drawn as to the impact of Arctic weathering on oil density and dielectrics and resulting implications for remote-sensing detection.


Evaluating the aggregate-scale thermal conductivity of Arctic sea-ice snow cover

Chris Polashenski, Nicholas Wright, Glen Liston

Corresponding author: Chris Polashenski

Corresponding author e-mail: chris.polashenski@gmail.com

Snow’s insulating and reflective properties substantially influence Arctic sea-ice growth and decay. Here we present the results of two winter-long field experiments conducted in Utqiaġvik, Alaska, USA. The experiments seek to better quantify the role of snow in insulating the ice and controlling growth during winter. We present detailed observations of the evolving snow stratigraphy and underlying ice growth over ~1 km2 domains through the duration of the winter. Terrestrial lidar is used to observe snow-surface morphology at centimeter-scale as the snowpack evolves. Thousands of observations of snow properties, including density, grain size/structure, thermal conductivity and hardness, are used to track properties for each accumulated layer. Data sets are combined to produce a centimeter-scale model of evolving snowpack properties over the two ~1 m2-scale domains. We find that meaningful changes in the snowpack properties are primarily event-driven, with only a handful of snowstorms, wind events and melt/rain events exerting the majority of control over the snowpack state. The outsize impact of individual events suggests that inter-annual variability in snow pack properties may be large. We also evaluate the impact of wind-driven spatial redistribution of snow into dunes and drifts. The distribution of a snow cover governs its aggregate-scale thermal resistance. We integrate the data from this experiment with prior studies and a resolved-scale snow model to address the question: How much variability in ice growth is controlled by inter-annual variability in snowpack redistribution?


Bio-physical sea-ice processes in the Lincoln Sea: an overview of the Multidisciplinary Arctic Program (MAP) – Last Ice

Christine Michel, Benjamin Lange, Joannie Charette, Karley Campbell, Steve Duerksen, Pascal Tremblay, Cody Carlyle, Pierre Coupel, Hauke Flores

Corresponding author: Christine Michel

Corresponding author e-mail: christine.michel@dfo-mpo.gc.ca

The Multidisciplinary Arctic Program (MAP) – Last Ice investigates bio-physical processes associated with the presence of old multiyear ice in the Lincoln Sea, in support of conservation initiatives in the Arctic. The Lincoln Sea is one of few regions of the Arctic Ocean where multiyear ice is predicted to persist over the next decades. This region is poorly characterized, largely due to accessibility constraints. The first MAP-Last Ice field campaign took place from 25 April to 7 June 2018 in the Lincoln Sea, off Ellesmere Island. Sea ice, water column and zooplankton sampling, as well as continuous atmospheric and oceanographic measurements, was carried out at an ice camp 9 km offshore of CFS (Canadian Forces Station) Alert until 26 May, when the ice cover broke up. Aerial marine mammal surveys were also conducted. Sea-ice and water samples were regularly collected and processed for a suite of biochemical analyses including nutrients, protist abundance and composition, pigmented biomass and primary production (sea ice only). Sea-ice biomass, zooplankton and fish larvae collected under the ice were also analyzed for fatty acid biomarkers. A remotely operated underwater vehicle provided in-situ bio-physical measurements at the under-ice interface along repeated transects. A total of 258 ice cores were collected, of which 131 were multiyear ice. Ice thickness ranged between 1.37 and 4.58 m. Preliminary results show higher biomass in first-year compared to multiyear ice, likely related to habitat space availability. Ice algal species composition and photo-physiology also differed between first-year and multiyear ice, with very low saturating irradiance and high sensitivity to photo-inhibition in multiyear ice. We propose that the diversity of sea-ice conditions and habitats created by the presence of multiyear ice supports heightened biodiversity associated with distinct niches, benefiting highly specialized species. The loss of thick multiyear ice is expected to directly impact these species and, as a corollary, the structure and function of the Arctic sea-ice ecosystem.


A pan-Arctic kilometre-scale record of sea-ice thickness and surface roughness from 2010–19 through the application of a numerical SAR altimeter echo model to Cryosat-2

Jack Landy, Michel Tsamados, Randy Scharien, Alek Petty, Julienne Stroeve

Corresponding author: Jack Landy

Corresponding author e-mail: jack.landy@bristol.ac.uk

We present a newly developed numerical model designed to simulate delay-Doppler SAR altimeter echoes from snow-covered sea ice, such as those detected by ESA’s Cryosat-2. Waveforms are simulated from virtual tetrahedral models of the sea-ice surface topography, with backscattering properties modelled directly from the geophysical properties of the snow cover and ice (for instance, temperature, brine content, roughness and presence of water). An analysis of Operation IceBridge ATM data shows that sea-ice surface topography is more realistically represented by a lognormal height distribution, rather than a Gaussian distribution as often assumed. So we have performed several hundreds of thousands of echo simulations to generate lookup tables of waveforms for a range of sea-ice roughness scenarios, assuming lognormal or Gaussian topography. We use these lookup tables to retrack the entire winter 2010–19 Cryosat-2 SAR and SARIn record, by optimizing the fit of the modelled echo to each L1b observed waveform. Our results demonstrate that it is crucial to account for varied sea-ice roughness when retracking altimeter observations for estimating sea-ice thickness. Techniques employing threshold retrackers likely overestimate the thickness of smooth first-year ice, whereas physical retracking techniques assuming Gaussian roughness likely underestimate the thickness of deformed and multiyear ice. We will contextualize this new surface roughness record with existing airborne and terrestrial observations to demonstrate that the sea ice roughness power spectral density can be accurately characterized with a two-scale Lorentzian model. The multi-scale (1–100 km) monthly record of sea-ice freeboard and surface roughness can be shared on request by the authors.


Estimation of turbulent heat flux over leads using satellite thermal images

Meng Qu, Xi Zhao, Xiaoping Pang, Jinlun Zhang

Corresponding author: Xi Zhao

Corresponding author e-mail: xi.zhao@whu.edu.cn

Sea-ice leads are an important feature in pack ice in the Arctic. Even covered by thin ice, leads can still serve as the prime window for heat exchange between the atmosphere and the ocean, especially in winter. Lead geometry and distribution in the Arctic have been studied using optical or microwave remote-sensing data. But turbulent heat flux over lead areas has only been measured on site during a few special expeditions. Due to insufficient resolution of satellite images and difficulties in parameterization, fetch-limited models haven’t been applied with remote-sensing temperature field yet. In this study, we derive turbulent heat flux through leads at different scales using both bulk aerodynamic formulae and a fetch-limited model with a combination of surface temperature and lead distribution from remote-sensing images and meteorological parameters from a reanalysis dataset. Firstly, ice-surface temperature was calculated from Landsat-8 thermal infrared sensor (TIRS) and MODIS thermal images using a split-window algorithm at 30 m and 1 km scales, respectively. Then lead pixels were segmented from colder ice. Heat flux over the lead area was estimated using two empirical models, including bulk aerodynamic formulae used in Coupled Large-scale Ice Ocean (CILO) and a fetch-limited model proposed by Andreas and Murphy (1999), with lead-width derived from Landsat-8. Results show that, even though the lead area from MODIS is slightly larger, the length of leads is underestimated by 72.9% in MODIS data compared to that from TIRS due to its inability to resolve small leads. Heat flux estimated using bulk formulae from Landsat-8 TIRS data is 42.33% larger than that from MODIS data. When the fetch-limited model is applied, turbulent heat flux calculated from TIRS data is 31.87% higher than that from bulk formulae. In both cases, small leads account for more than a quarter of total heat flux over leads, mainly due to their large area. More contribution from small leads is expected with larger air-surface temperature differences and stronger winds.


Investigating the robustness of Antarctic sea-ice reconstructions from ice cores

Shweta Mayekar, Mark Curran, Tessa Vance, Christopher Plummer, Andrew Moy, Jason Roberts, Lenneke Jong, Chelsea Long, Jan Lieser

Corresponding author: Mark Curran

Corresponding author e-mail: mark.curran@aad.gov.au

The trend of Antarctic sea-ice extent (SIE) cannot with any degree of certainty be determined at any timescales longer than decadal. Antarctic SIE reached an all-time low (in the satellite period) in spring 2016, after a generally increasing trend since 1979. 2012, 2013 and 2014 had previously recorded consecutive maximum SIE. The IPCC 5th Assessment stated Antarctic SIE was expanding significantly, with the magnitude of the increase since 1979 being one-third as large as the SIE retreat observed in the Arctic over the same period. In contrast, most climate model experiments simulate a reduction in Antarctic SIE since 1979 and out to 2100. The short satellite record means that discerning long-term trends above interannual variability is not possible, prompting researchers to reconstruct SIE using proxy information from ice-core records. Marine biological activity, which is closely linked to sea-ice extent and behaviour, is the single known source of methanesulphonic acid (MSA) in ice cores. MSA preserved in ice cores can be a proxy for SIE, provided SIE (rather than atmospheric transport) is the dominant control on MSA delivery to the ice sheet. Such records show SIE decreases in summer in all sectors, and in winter since the 1960s in the Amundsen, Bellingshausen and East Antarctic sectors. Bellingshausen Sea SIE has steadily declined throughout the 20th century, with the rate increasing since 1958. In contrast, SIE has increased in the Ross and Weddell Seas over this same interval. MSA concentrations from the Law Dome ice core have declined since the 1950s, while the South Orkney Islands observational sea-ice record showed a decline from 1930 to 1960. Generally, the extended information provided by pre-satellite data such as MSA records suggests that a decline in Antarctic sea ice has occurred over the 20th century, albeit with distinct regional differences. A more spatially complete record of Antarctic sea-ice proxies such as MSA may help resolve the discrepancies between satellite data and climate simulations of Antarctic sea-ice trends. Here we present an assessment of the relationship between MSA and SIE during the instrumental era from the Mount Brown South and Law Dome ice cores from coastal East Antractica, including new data from recently drilled ice cores at these sites. Investigating the robustness and variability of SIE reconstructions from these new ice cores is an important step toward understanding regional patterns of long-term Antarctic sea-ice trends.


The SIKU.org platform and mobile app: social-media tools and services for sea-ice safety and Inuit self-determination in research and stewardship

Joel Heath, Lucassie Arragutainaq

Corresponding author: Joel Heath

Corresponding author e-mail: joelheath@arcticeider.com

SIKU.org is a social-media online platform and mobile app designed with and for Inuit. It provides a wide variety of tools and services towards Inuit self-determination in research, education and environmental stewardship. Winner of the 2017 Google.org Impact Challenge in Canada, over the last year substantial progress has been made in developing, conducting workshops, consultation, piloting in northern communities and obtaining design input and feedback from Inuit hunters, youth including Nunavut Sivuniksavut and community-driven research programs. Far too often, Inuit knowledge and observations have been considered anecdotal by academic communities. The unique research tools available on the SIKU platform and mobile app prove a means for Inuit to document their land-use observations on an ongoing basis, providing a detailed quantitative dataset that can be used to bolster their reports and analysis of observations that were previously considered qualitative. Through the tools and services of the SIKU platform, communities and Indigenous organizations can define and implement their own research programs, as well as stewarding and analyzing their own results for their own purposes. As such, it is an important distinction that the platform is designed to facilitate Inuit self-determination in research and stewardship, rather than citizen science (i.e. where the public helps crowd-source collection of data towards academic endeavour). In addition to the tools and services provided by the platform, an important component of aligning the approach of SIKU with Inuit Tapiriit Kanatami’s National Inuit Strategy on Research involves consultation on approaches to data stewardship, intellectual property rights and tools and permissions that support the unique needs of Inuit, towards refining a Terms of Use and Privacy Policy for the platform, as well as defining logic for sharing, permissions and other features. This presentation will highlight the approach and logic defined to date through ongoing consultation, and seek additional input moving forward towards how SIKU can best facilitate the parallel needs of individual contributors, projects, communities and Indigenous organizations, towards the long-term benefit of Inuit self-determination.


Dynamic and thermodynamic winter sea-ice growth in a changing Arctic

Robert Ricker, Stefan Hendricks, Stephan Paul, Frank Kauker, Hiroshi Sumata

Corresponding author: Robert Ricker

Corresponding author e-mail: Robert.Ricker@awi.de

In contrast to land ice, the Arctic sea-ice cover is highly variable due to horizontal advection and the seasonal freeze–melt cycle, leading to significant seasonal and interannual changes. The near-surface air temperature is the main controlling factor for thermodynamic growth in the Arctic, while wind is the main driver for dynamic growth. The sea-ice thickness distribution is a result of the interaction between dynamic and thermodynamic processes. Moreover, they are mutually dependent. Divergent motion of sea ice stimulates thermodynamic ice growth, while convergence retards further thermodynamic growth. On the other hand, thinner/thicker sea ice tends to be more/less mobile. Since the last decade, satellites allow to observe sea-ice thickness from space. Recently, a consistent and calibrated climate data record of sea-ice thickness derived from satellite radar altimetry has been developed for both hemispheres, based on the 15-year (2002–17) monthly retrievals from Envisat and CryoSat-2. This 15-year period is also characterized by a significant increase in Arctic winter mean surface air temperatures. For the interpretation of observed changes in sea-ice volume in the context of changing atmospheric parameters, it is crucial to discriminate between dynamic and thermodynamic ice growth. Here, we use the ESA Climate Change Initiative ice-thickness data record together with satellite ice concentration and drift products to estimate winter ice-volume changes and ice-volume fluxes in order to discriminate between dynamic and thermodynamic ice growth in key regions of the Arctic. We find that the mean thermodynamic winter ice-volume growth in the Barents Sea has decreased by more than 50% between 2002 and 2017, while the mean dynamic growth is unchanged. In contrast, we find an increase in thermodynamic ice growth in the Laptev Sea, accompanied by an increase of ice-volume export, equivalent to a decrease in the mean dynamic ice-volume growth. We also compare these results with numerical simulations by a coupled sea-ice–ocean model, NAOSIM (the North Atlantic Ocean Sea Ice Model).


Merged operational satellite ice-thickness retrievals to inform about the present state of Arctic sea ice

Robert Ricker, Xiangshan Tian-Kunze, Stefan Hendricks, Lars Kaleschke, Antonio de la Fuente

Corresponding author: Robert Ricker

Corresponding author e-mail: Robert.Ricker@awi.de

Sea-ice thickness on a global scale is derived from different satellite sensors using independent retrieval methods. Due to the sensor and orbit characteristics, such satellite retrievals differ in spatial and temporal resolution as well as in their sensitivity to certain sea-ice types and thickness ranges. Satellite altimeters, such as CryoSat-2 (CS2), sense the height of the ice surface above sea level, which can be converted into sea-ice thickness. However, relative uncertainties associated with this method are large over thin-ice regimes. Another retrieval method is based on the evaluation of surface brightness temperature in L-band microwave frequencies with a thickness-dependent emission model, as measured by the Soil Moisture and Ocean Salinity (SMOS) satellite. While the radiometer-based method loses sensitivity for thick sea ice (>1 m), relative uncertainties over thin ice are significantly smaller than for the altimetry-based retrievals. In addition, the SMOS product provides global sea-ice coverage on a daily basis, unlike the altimeter data. In the framework of the ESA project ‘SMOS & CryoSat-2 Sea Ice Data Product Processing and Dissemination Service’, we present the first operational merged product of complementary weekly Arctic sea-ice thickness data records from the CS2 altimeter and SMOS radiometer. We use an optimal interpolation scheme to produce weekly Arctic-wide sea-ice thickness fields. The data product is publicly available and has been already used in several scientific studies that investigate sea-ice cover in the context of climate change. Moreover, it supports operational use of remotely sensed sea-ice thickness information for international sea-ice monitoring programs and climate forecasting systems. Here, we present sea-ice thickness and volume changes from last winter 2018/19, informing about the present state of the Arctic sea- ice cover.


Assessing stability and precision of sea-ice thickness retrievals from satellite altimetry by a cross-over analysis

Robert Ricker, Stefan Hendricks, Stephan Paul

Corresponding author: Robert Ricker

Corresponding author e-mail: Robert.Ricker@awi.de

A climate data record of sea-ice thickness derived from satellite radar altimetry that has been developed for both hemispheres, based on the 15-year (2002–17) monthly retrievals from Envisat and CryoSat-2 and calibrated in the 2010–12 overlap period. In addition, a sea-ice thickness retrieval from ICESat laser altimetry measurements is available from 2003–08. In order to assess the stability and precision of these three different ice thickness products, we present a novel approach using cross-over analysis over sea ice. As an advantage over other performance indicators, this method is independent of validation data. In order to minimize the impact of random noise of single measurements, we calculate differences between ice- thickness retrievals derived from two crossing satellite orbits within a 12.5 km radius around the cross-over. Moreover, we only consider cross-overs within a 24 h window to minimize the impact of sea-ice drift. This study will be a step forward to further constrain the precision of recent altimeter missions used to derive sea-ice thickness retrievals.


Seasonal and climatological changes in the thickness distribution of Arctic sea ice

Srikanth Toppaladoddi, Woosok Moon, John Wettlaufer

Corresponding author: John Wettlaufer

Corresponding author e-mail: john.wettlaufer@yale.edu

A mathematically and physically consistent extension of our 2015 theory of sea-ice thickness distribution, g(h), that includes open water is made. The Fokker–Planck equation for g(h) is coupled to a modified version of the observationally consistent sea-ice growth model of Eisenman and Wettlaufer (2009) to study the seasonal evolution of g(h) under climate forcing. We find that g(h) transitions from a single- to a double-peaked distribution in spring, which is in agreement with recent satellite observations. To understand the cause of this transition, we construct a simpler description of the system using the equivalent Langevin formulation and solve the resulting stochastic ordinary differential equation numerically.


Verification of AMSR-2 sea-ice concentration in the Arctic Ocean

Nodoka Ono

Corresponding author: Nodoka Ono

Corresponding author e-mail: ono.nodoka@jaxa.jp

Mitsui O.S.K. Lines, Ltd plans to transport liquefied natural gas (LNG) produced in Yamaru via the Arctic Ocean for 28 years starting in 2018 using the Northern Sea Route. The Japan Aerospace Exploration Agency (JAXA) offers sea-ice products (sea-ice concentration, etc.) from the Advanced Microwave Scanning Radiometer-2 (AMSR-2) carried by the Global Change Observation Mission–Water (GCOM-W) through the ship navigation support service ‘VENUS’ developed by the Japanese National Institute of Polar Research. In addition, JAXA has installed interval cameras developed by Weathernews Inc. on the LNG ship to acquire sea-ice information in the Arctic Ocean year-round. Using these field data, I calculated sea-ice concentrations using the image analysis system developed by Tanaka et al. (2015) and verified the results against products from the AMSR-2. In the Northern Sea Route and the Russian coastal area, these are mostly unverified because, for political reasons, there are almost no field data. I used the Moderate Resolution Imaging Spectroradiometer (MODIS) carried by the Terra/Aqua satellite to fill in the difference of spatial resolution between the field data and AMSR-2 data. In this study, I describe the verification results of AMSR-2 data.


Estimation of sea-ice production in the Bering Sea from AMSR-E and AMSR2 data, with special emphasis on the Anadyr polynya

Kay Ohshima, Naoya Tamaru, Sohey Nihashi, Kazuki Nakata, Katsushi Iwamoto

Corresponding author: Kay Ohshima

Corresponding author e-mail: ohshima@lowtem.hokudai.ac.jp

In the Bering Sea, coastal polynyas often occur off the coast of Russia, Alaska and St Lawrence Island, because of the prevailing northerly wind, and dense shelf water (DSW) can be formed there. It has been generally thought that DSW from the Bering coastal polynyas is not dense enough for bottom/intermediate water formation, but dense enough to intrude and maintain the cold halocline layer, which is a major subsurface water mass in the Arctic Ocean. On the other hand, a high chlorofluorocarbon concentration was observed in the abyssal layer of the Bering Sea, suggesting the possibility of temporary bottom water formation. The paleo-oceanographic studies proposed that intermediate/deep water formation occurred in the Bering Sea during glacial periods. Therefore, estimation of sea-ice production is critical for understanding the maintenance of and variability in ocean stratification and re-consideration of bottom/intermediate water formation in the Bering Sea. We provide the mapping of sea-ice production in the Bering Sea for the first time, from the AMSR-E and AMSR2 passive microwave data and heat flux calculation. We modified the thin-ice thickness algorithm developed for the whole Arctic Ocean by Iwamoto et al. (2014) and then used it for the estimation of ice production. We estimated the ice production in the whole Bering Sea from the 2002/03 season. It is found that the Anadyr polynya (AP) has by far the highest ice production among the Bering coastal polynyas, making up about one-third of total ice production in the Bering Sea. The AP is found to be the second or third highest ice-production polynya in the Northern Hemisphere. The high ice production is caused by the strong offshore prevailing wind and the cold air temperature from the continent. Ice production in the AP shows very large interannual variability, having the highest coefficient of variation among major coastal polynyas in the world. This is because the wind direction over the AP strongly depends on the position of the Aleutian Low. This variability in ice production corresponds well with that of winter salinity increase in the Bering Strait. This suggests that ice production in the Bering Sea, particularly from the AP, significantly affects water mass formation and stratification in the Arctic Ocean through the inflowing dense water.


Surf’s up! Sea-ice loss, ocean swell and Antarctic ice-shelf disintegrations

Rob Massom, Ted Scambos, Luke Bennetts, Phil Reid, Vernon Squire, Sharon Stammerjohn

Corresponding author: Rob Massom

Corresponding author e-mail: rob.massom@aad.gov.au

Understanding the causes of catastrophic disintegrations of the Larsen A and B and Wilkins ice shelves on the Antarctic Peninsula since 1995 is a key step to improving models of the Antarctic ice-sheet system and assessing the vulnerability of the remaining ice shelves. This in turn is crucial to enabling more accurate prediction of the future state of the ice sheet and its contribution to sea-level rise. Here, we examine a climate-related causal factor and trigger mechanism that has been largely overlooked to date – namely regional sea-ice loss (involving both pack and fast ice). Based upon analysis of satellite, wave hindcast and model output data, we propose that increased seasonal absence of a protective sea-ice buffer offshore exposed the vulnerable outer margins of the Larsen and Wilkins ice shelves to enhanced flexure by ocean swells. Over time, this weakened outer-margin crevasse and rift systems to the point of calving, which then precipitated rapid runaway disintegration of extensive ice-shelf areas weakened (preconditioned) by thinning, surface flooding and glaciological factors. These are common essential prerequisites for disintegration in the cases examined. The new results highlight sea ice (change/variability) as an important additional player affecting ice-shelf stability, depending on the region and ice shelf. They also underline the highly coupled and complex nature of the ice-shelf system undergoing change, and the need to better understand and quantify the cross-cryospheric interactions involved.


Biogeochemical linkage of the sea-ice, pelagic and benthic systems and its role on the Arctic ecosystem functioning

Déborah Benkort, Ute Daewel, Richard Hofmeister, Corinna Schrum

Corresponding author: Déborah Benkort

Corresponding author e-mail: deborah.benkort@hzg.de

Primary production in the Arctic system is principally supported by pelagic phytoplankton production. However, sea-ice algae play an important role in the total Arctic primary production, and therefore represent a crucial element in the entire Arctic food web dynamic. With the rapid changes ongoing in the Arctic, especially in the sea-ice cover, what will become of the Arctic marine ecosystem is still uncertain. A proper representation of sea-ice algae phenology and the linkage with the pelagic and benthic ecosystems, taking into account sea-ice structural changes, appears essential to understand the dynamics of the Arctic ecosystem and its future changes. In order to study the dynamics of the Arctic ecosystem in the Barents Sea area, we have further developed the biogeochemical model ECOSMO by implementing a sea-ice algae group in the model formulation. In a first attempt to investigate and solve the scientific and technical challenges related to the coupling between the pelagic ecosystem and the sea ice biology, the model was implemented in a 1-d application of the General Ocean Turbulence Model. Here we present results from this numerical framework, aiming specifically at understanding the linkage between pelagic and benthic with the sea-ice ecosystem. The results already indicate how the physical environment and the projected changes in sea-ice coverage impact the entire Arctic ecosystem dynamic.


ARC3O: an Arctic Ocean observation operator for 6.9 GHz

Clara Burgard, Dirk Notz, Leif Toudal Pedersen, Rasmus Tage Tonboe

Corresponding author: Clara Burgard

Corresponding author e-mail: clara.burgard@mpimet.mpg.de

The observational uncertainty in sea-ice-concentration estimates from remotely sensed passive-microwave brightness temperatures is a challenge for robust climate-model evaluation and reliable climate-model initialization. To address this challenge, we introduce the Arctic Ocean Observation Operator (ARC3O). ARC3O allows us to simulate brightness temperatures at 6.9 GHz, vertical polarization, from output of a sea-ice model with minimum complexity. We compare our simulated brightness temperatures to brightness temperatures measured by the Advanced Microwave Scanning Radiometer Earth Observing System (AMSR-E) and find that they compare with a difference on the order of 5 K in all seasons except summer, where they differ on the order of 10 K. As a possible application of the observation operator, we investigate the sensitivity of the simulated brightness temperatures to different parameters and find that the sea-ice concentration is the main driver for brightness-temperature variations in marginal regions and the surface temperature is the main driver in the central Arctic. The pattern depends on the time of the year. We use this knowledge to compare simulated brightness temperatures from three different sea-ice concentration products assimilated into the MPI Earth System Model. The comparison with the observed brightness temperature uncovers limitations of the sea-ice concentration products, especially in the marginal seas and in summer. In the case of summer brightness temperatures, we can uncover biases related to melt ponds in both model and observations. ARC3O can therefore be used for a better process-understanding of the drivers of brightness temperatures and uncover uncertainties and biases in both retrieved sea-ice concentration and simulated climate state.


Using a mesocosm for snow-covered sea-ice-ridge characterization and detection using remote sensing: implications for ringed-seal breeding habitat

Veronica Coppolaro, Mark Christopher Fuller, Madyson Harasyn, Dustin Isleifson, Julienne Stroeve, David Barber

Corresponding author: Veronica Coppolaro

Corresponding author e-mail: coppolav@myumanitoba.ca

Sea ice plays a major role in the Arctic as it represents the boundary between ocean and atmosphere. It is also of great importance for animal species that rely on sea ice, such as ringed seals (Phoca hispida), known to use snow-covered sea ice as a breeding habitat. Sea ice morphological formations such as pressure ridges influence snow accumulation and redistribution processes on the sea-ice surface, thus driving ringed-seal habitat selection. Climate change is amplified in the Arctic, causing rapid increases in air temperature, early sea-ice break-up and late freeze-up that could alter ringed-seal habitat, affecting their breeding success. Mesocosm-scale studies of artificial sea ice can improve our fundamental understanding of natural processes. The first experiment of this project took place in January 2019 at the Sea-ice Environmental Research Facility (SERF). The aim was to investigate the parameters and the processes that influence winter microwave passive radiation emitted by first-year ice ridges with and without a snow cover for remote sensing. The experiment involved the construction of a sea-ice ridge in the sea-ice-covered SERF pool, in order to mimic a natural first-year sea-ice ridge. The study was run by using, among other instruments, a light detection and ranging scanner, a surface-based radiometer, and a forward-looking infrared radiometer in order to collect an exhaustive dataset to analyze the ongoing electro-thermo-geophysical processes both in snow and in sea ice. The data were collected for different ridge heights, light polarizations, frequencies, incident angles and snow depth, and in different meteorological conditions. The results will be used to study emissivity and brightness temperature for several sea-ice-ridge configurations and to better identify and predict ringed-seal breeding habitat driven by sea-ice topography and snow depth. The mentioned parameters together with the combined use of both microwave and infrared signals represent a promising dataset for a comprehensive study of snow-covered sea-ice ridges.


Spatial scales of seasonal snow-property variations on Antarctic sea ice

Stefanie Arndt, Stephan Paul, Nicolas Stoll, Arttu Jutila, Joshua King

Corresponding author: Stefanie Arndt

Corresponding author e-mail: stefanie.arndt@awi.de

Snow on sea ice alters the properties of the underlying ice cover as well as associated exchange processes at the interfaces between atmosphere, sea ice and ocean. As Antarctic snow cover persists during most of the year, it contributes significantly to the sea-ice mass and energy budgets due to comprehensive physical (seasonal) transition processes within the snowpack. However, field studies reveal not only a strong seasonality but especially spatial variations from local to regional scales. It is therefore necessary to quantify seasonal snow processes, such as internal snowmelt, snow metamorphism and snow-ice formation at multiple spatial scales on Antarctic sea ice. Doing so, we present here in-situ observations of physical snow properties from point measurements (snow pits) and transect lines (SnowMicroPen, SMP) during recent expeditions in the Weddell Sea from 2013–19, covering summer and winter conditions. Results from a case study of snow-pit analyses in the Weddell Sea during austral winter reveal a high spatiotemporal variability of snow parameters, highlighting the need to distinguish between seasonal and perennial snow regimes. Also, it is shown that snow-grain size dominates the spatial variability of the snowpack while snow-density variability can be neglected. In order to extend local snow-pit analysis towards the description of snow-layer evolution on small scales (up to 500 m), SMP measurements are added. A layer-tracking algorithm applied to the vertical density profiles throughout the snowpack allows us to quantify length-scale variabilities of snow properties in different ice regimes. Overall, results will improve our understanding of seasonal processes in the snowpack and will guide us towards upscaling approaches of vertical snow layers on Antarctic sea ice.


Seasonal changes in snow properties from passive and active microwave satellite observations: a conceptual model

Stefanie Arndt, Christian Haas

Corresponding author: Stefanie Arndt

Corresponding author e-mail: stefanie.arndt@awi.de

Snowmelt processes on sea ice are the key drivers determining seasonal sea-ice energy and mass budgets. Around Antarctica, snowmelt on pack ice is weak and very different from in the Arctic, with most snow surviving the summer. It is therefore important to understand the mechanisms that drive snowmelt, both at different times of the year and in different regions around Antarctica. To do this, we compile time series of snowmelt-onset dates on perennial Antarctic sea ice from 1992–2014 using active microwave observations from the European Remote Sensing Satellite (ERS-1/2), Quick Scatterometer (QSCAT) and Advanced Scatterometer (ASCAT) radar scatterometers. Describing snow melt processes, we define two transition stages: a weak backscatter rise indicating the initial warming and metamorphism of the snowpack (pre-melt), followed by a rapid rise indicating the onset of thaw–freeze cycles in the interior snowpack (snowmelt). We compare these with pan-Antarctic temporary snowmelt-onset dates in the uppermost snowpack retrieved from diurnal variations in the brightness temperatures from passive microwave (PMW) observations. Results show that QSCAT Ku-band (13.4 GHz signal frequency) derived pre-melt and snowmelt onset dates are earlier by 25 and 11 days, respectively, than ERS and ASCAT C-band (5.6 GHz)-derived dates. Snowmelt-onset dates from the shortwave PMW observations (37 GHz) are later by 13 and 5 days than those from the scatterometers, respectively. Based on the observed successive timing of melt events retrieved from different sensors and microwave bands, we developed a conceptual model of the temporal evolution of snow temperature and metamorphism and their effect on different microwave wavelengths during the spring/summer transition. These results suggest that future multi-frequency microwave satellite missions could be used to resolve melt processes throughout the vertical snow column. Overall, results show that the magnitude and timing of seasonal and diurnal variations in Antarctic snow on sea ice are highly dependent on latitude, with earlier and more frequent snowmelt in the north. All retrieved melt-onset dates show large interannual variability but no significant decadal trends.


Seasonal and interannual variability of landfast sea ice in Atka Bay, Weddell Sea, Antarctica

Stefanie Arndt, Marcel Nicolaus, Mario Hoppmann, Holger Schmithüsen

Corresponding author: Stefanie Arndt

Corresponding author e-mail: stefanie.arndt@awi.de

Landfast sea ice (fast ice) attached to the Antarctic coast is a critical element of the local physical and ecological systems. Through its direct coupling with the atmosphere and ocean, fast ice and its snow cover are also a potential indicator of processes related to climate change. However, in-situ fast-ice observations in Antarctica are extremely sparse because of logistical challenges. Since 2010, a monitoring program, which is part of the Antarctic Fast Ice Network, has been conducted on the seasonal evolution of the fast ice of Atka Bay. The bay is located on the north-eastern edge of the Ekström Ice Shelf in the eastern Weddell Sea, close to the German wintering station Neumayer III. A number of sampling sites have been regularly revisited between annual ice formation and breakup each year to obtain a continuous record of snow depth, freeboard, sea-ice- and sub-ice platelet layer thickness across the bay. Here, we show the results of these measurements, combining them with observations from the nearby meteorological observatory at Neumayer Station as well as satellite images to relate the seasonal and interannual fast-ice cycle to the factors that influence its evolution. On average, the annual fast-ice thickness at the end of the growth season is about 2 m, with a platelet layer accumulation of 4 m beneath. Due to the substantial snow accumulation on the ice, a characteristic feature is frequent negative freeboard and associated flooding of the snow/ice interface. Results highlight the predominately seasonal character of the fast-ice regime in Atka Bay without a significant trend in any of the observed variables over the 9-year observation period. Also, no changes are evident when comparing with measurements in the 1980s and 1990s. However, strong easterly winds in the area govern the year-round snow redistribution and also trigger the breakup events of the bay during summer months. An enhanced knowledge of the seasonal and interannual variability of fast ice in this region will improve our understanding on local atmosphere/sea-ice/ocean interactions, which can then be transferred to pan-Antarctic fast-ice regimes.


Angles between conjugate linear kinematic features with sea-ice viscous–plastic rheologies

Damien Ringeisen, Martin Losch, L. Bruno Tremblay, Nils Hutter

Corresponding author: Damien Ringeisen

Corresponding author e-mail: damien.ringeisen@awi.de

Sea-ice observations and models show zones of high deformation typical of granular medium (linear kinematic features (LKFs)). Recent high-resolution simulations feature fractures that mimic the observed pattern but with wider intersection angles. Motivated by this, we investigate the dependence between conjugate faults intersection angles and different viscous–plastic rheologies. Using an idealized uniaxial setting, the ice fracture is modeled with different confinement ratios and two different VP rheologies: one with an elliptical yield curve and a normal flow rule, and one with a Coulombic yield curve and a normal flow rule that applies only to the elliptical cap. Modeling fracture angles smaller than 30° is not possible with elliptical yield curve in a pure compression setting. Further several modeled behaviors are inconsistent with the granular nature of sea ice : (1) the fracture angle increases with ice shear strength; (2) the divergence along the fracture lines (or LKFs) is uniquely defined by the shear strength of the material with divergence for high shear strength and convergenge forlow shear strength; (3) the angle of fracture depends on the confining pressure with more convergence as the confining pressure increases. With Mohr’s circle, this behavior is shown to be linked to the convexity of yield curve. The Coulombic yield curve is able to model smaller angles but the solution is unstable because of non-differentiable corners between the straight limbs of the Coulombic yield curve and the elliptical cap. The results shows that, although the fracture patterns at first appear realistic, the yield curve should be revised to take into account the nature of sea ice as a pressure-sensitive and dilatant granular material.


Sea-ice and open-water classification using historical multi-sensor synthetic aperture radar satellite datasets, applied to two Svalbard fjords

Malin Johansson, Sebastian Gerland, Anca Cristea, Eirik Malnes, Anthony Doulgeris, Dmitry Divine, Olga Pavlova, Tom Rune Lauknes

Corresponding author: Malin Johansson

Corresponding author e-mail: malin.johansson@uit.no

Satellite data records covering the Arctic region date back to 1978 and now span over 40 years. For synthetic aperture radar (SAR) data the records date back to 1991. The records are therefore long enough to enable monitoring of changing sea-ice conditions over several decades. Here we propose a sea-ice segmentation and classification method that separates open water from sea ice and that can utilize ERS-1/2, Envisat ASAR, Radarsat-2 and Sentinel–1 ScanSAR images. The segmentation algorithm is generic and only requires information about the backscatter intensity values and the incidence angle, and the same method can therefore be employed on images from each of these different sensors. One advantage of such a method is the possibility of establishin consistent decadal records of sea-ice situations. To demonstrate our concept, we studied the sea-ice conditions in Kongsfjorden, a west-facing fjord on Svalbard with a seasonal fjord ice cover. Our monitoring has weekly to daily records from 2002 until now, and additionally there are less frequent records between 1991 and 2002. Overlap in space and time between Envisat ASAR and Radarsat-2 as well as between Radarsat-2 and Sentinel–1 is investigated to ensure consistency in reported sea- ice cover between the different sensors. The classification results have also been compared with high-resolution optical and radar satellite data as well as in-situ sea-ice measurements and web-camera images from Ny Ålesund. An additional study is performed using the same sensors and time-interval covering Rijpfjorden, a north-facing Svalbard fjord with a longer sea-ice season than Kongsfjorden. Challenges such distinguishing among open water with variable wave conditions, sea ice and icebergs, growlers and bergy bits originating from nearby glaciers are discussed. From the records we observe that for Kongsfjorden the length of the sea-ice season has shortened since 2002 and that the maximum sea-ice coverage is significantly lower after 2006.


Variability in algal production between first-year and multiyear sea ice habitats

Karley Campbell, Christine Michel, Ben Lange, Philipp Anhaus, Joannie Charette, Steve Duerksen, Pascal Tremblay, Cody Carlyle, Pierre Coupel

Corresponding author: Karley Campbell

Corresponding author e-mail: kc17823@bristol.ac.uk

The production and speciation of ice algal communities is dependent on the physical–geochemical characteristics of sea-ice habitats, including the availability of light and nutrients. As a result, variability in such properties with the age of sea ice have the potential to cause differences in ice algal production and species composition between younger first-year and older multiyear ice habitats. In this study we investigate the potential for such differences by characterizing the biogeochemical conditions of adjacent first-year and multiyear sea-ice floes in the Lincoln Sea of the Canadian high Arctic during the Last Ice – Multidisciplinary Arctic Program 3–23 May 2018. Experimental data on nutrient uptake and algal photophysiology have shown consistent limitation of production in the region by nitrogen, but differing speciation and capacity for photoacclimation in the algal communities in the respective ice habitats. We combine these results with floe-scale measurements of spectral transmitted irradiance collected using a remotely operated vehicle to further contrast the overall productive potential of ice floes in the region. This work is a critical step in understanding the impact of an increasingly younger sea-ice cover in the Arctic Ocean with ongoing climate change. It also represents the first field season of the Changing Arctic Ocean project Diatom-ARCTIC.


Landfast ice diminishes shelf–ocean interaction over northeast Greenland

Igor Dmitrenko, Sergei Kirillov, David Babb, Leif Toudal Pedersen, Soeren Rysgaard, David Barber

Corresponding author: Igor Dmitrenko

Corresponding author e-mail: igor.dmitrenko@umanotoba.ca

The Wandel Sea in northeast Greenland is the only place in the high Arctic where the landfast ice can extend over the shelfbreak and upper continental slope. The Wandel Sea outer shelf is covered by the multiyear landfast sea ice all year around. Starting in summer 2016, the multiyear landfast ice became unstable, and a sizeable portion collapsed in August 2017. The landfastice edge to the east roughly delineates the Wandel Sea continental shelf break, which is where a coastal polynya opens in response to southerly winds. The landfast ice-tethered oceanographic mooring was deployed over the southeast Wandel Sea outer shelf from May 2015 to April 2016. The mooring located ∼18–20 km from the landfast ice edge carried an ice tethered profiler recording salinity–temperature–depth (CTD) and colored dissolved organic matter (CDOM) fluorescence profiles every 3 hours for a year. This was accompanied by the ADCP velocity observations. The satellite imagery shows that since mid-December 2014 the landfast ice edge, controlled by northerly winter winds through a surface Ekman onshore transport, was gradually extending eastward and in mid-March 2015 it was finally stabilized over the Wandel Sea upper continental slope. During this time, Ekman transport of the Pacific-derived Arctic water to the Wandel Sea shelf was observed. For the upwelling-favourable summer wind forcing, the Atlantic water on-shelf inflow is expected along with outflow of the Pacific-derived Arctic water through the overlying water layer. In fact, however, the on-shelf Atlantic water flow was observed following the upwelling-favourable storms in June–July 2015 by about one month. We suggest that this delay is attributable to the landfast ice extending eastward beyond the shelfbreak. This implies that upwelling is sensitive to the sea-ice conditions over the continental slope. Once the outer part of the landfast ice area had collapsed, and the landfast ice edge was positioned onshore by∼17 km beyond the shelf break, the Atlantic water on-shelf flow is established in response to the upwelling-favourable southerly wind forcing. Thus, the landfast ice extending during winter over the Wandel Sea shelfbreak and upper continental slope diminishes upwelling, which starts developing in summer once the outer portion of the landfast ice has collapsed.


The influence of sea-ice floe-size distribution on the area-averaged values of surface fluxes

Marta Wenta, Agnieszka Herman

Corresponding author: Marta Wenta

Corresponding author e-mail: marta.wenta@phdstud.ug.edu.pl

The atmospheric boundary layer (ABL) response to sea-ice fragmentation has only recently attracted attention of the scientific community. So far, most of the studies focused on the ABL modelling and observations over leads, rarely considering multiple, differently oriented cracks between sea-ice floes. Furthermore, the large-scale effects of submesoscale processes taking place on the level of individual floes are still not fully understood and not taken into account in mesoscale numerical weather prediction (NWP) model parameterizations. In the research presented the idealized version of Weather Research and Forecasting (WRF) model is launched for a series of high-resolution simulations with different spatial sea-ice distributions. The simulations are divided into several groups with identical initial conditions and sea-ice concentrations but different sizes and spatial arrangements of ice floes. At present, in many global climate models the values of moisture and energy fluxes are calculated from the grid-cell averaged values of temperature, wind speed, humidity, etc. Our analysis indicates that this may lead to substantial errors and underestimation of the value of the flux. To study the problem further, the values of moisture flux estimated from area-averaged quantities forming the flux (QFX1) are compared with those calculated for every grid cell of 100 × 100 m and then averaged (QFX2). Due to the nonlinear character of the described dependencies the order of averaging has a significant effect on the result. We assume that our model domain of 20 × 20 km represents a single grid cell of the global climate model. On the basis of these results, the ratio of QFX1/QFX2 is computed for every dataset with different floe-size distribution and initial ambient wind speed. Obtained values are further analysed in order to find empirical relationships in the form of a best-fit equation between the size and distribution of the floes, wind speed and the values of QFX1/QFX2. It is a preliminary step to further research, the aim of which is to formulate parametrizations of the influence of the spatial arrangement of sea-ice floes and their size on the values of QFX. We plan to expand our research with the analysis of observational data from UAV flights over fragmented sea ice along with with real WRF model simulations. They will certainly give an additional insight into studied processes and contribute to the parametrization development and improvement of NWP models.


Feature-based comparison of sea-ice deformation in lead-resolving sea-ice simulations

Nils Hutter, Martin Losch

Corresponding author: Nils Hutter

Corresponding author e-mail: nils.hutter@awi.de

The sea-ice modelling community progresses towards pan-Arctic simulations that explicitly resolve leads in the simulated ice cover. This creates new challenges of proper model evaluation. We introduce a feature-based evaluation of simulated deformation fields and compare the results to a scaling analysis of sea-ice deformation. Leads and pressure ridges – combined as linear kinematic features (LKF) – are detected and tracked by an algorithm that uses deformation and drift data. LKFs in two pan-Arctic sea-ice simulations with a horizontal grid spacing of 2 km are compared with an LKF dataset derived from the RADARSAT Geophysical Processor System (RGPS). One simulation uses a 5-class ice-thickness distribution (ITD). The simulated sea-ice deformation is described by multi-fractal spatial and temporal scaling as observed from RGPS for both simulations. Interannual and seasonal variations of the number of LKFs, LKF densities and LKF orientations in the ITD simulation are found to be in line with RGPS observations. The heavy-tailed distribution of LKF lengths and the scale invariance of LKF curvature is reproduced by the model and points towards the self-similar nature of sea-ice deformation fields. The model overestimates the intersection angle of LKFs, which is attributed to the use of the elliptical yield curve. In addition, the lifetime and growth rates of LKFs are found to be described by an exponential tail. In conclusion, our analysis of LKF statistics is a useful tool for a comprehensive description of deformation features and as such complements the previously used scaling analysis. The ITD simulation is shown to reproduce LKFs sufficiently well to be used for studying the effect of directly resolved leads in climate simulation.


Leads and ridges in Arctic sea ice from RGPS data and a new tracking algorithm

Nils Hutter, Lorenzo Zampieri, Martin Losch

Corresponding author: Nils Hutter

Corresponding author e-mail: nils.hutter@awi.de

Leads and pressure ridges are dominant features of the Arctic sea-ice cover. Not only do they affect heat loss and surface drag, but they also provide insight into the underlying physics of sea-ice deformation. Due to their elongated shape they are referred to as linear kinematic features (LKFs). We introduce two methods that detect and track LKFs in sea-ice deformation data and establish an LKF dataset for the entire observing period of the RADARSAT Geophysical Processor System (RGPS). Both algorithms are available as open-source code and applicable to any gridded sea-ice drift and deformation data. The LKF detection algorithm classifies pixels with higher deformation rates compared to the immediate environment as LKF pixels, divides the binary LKF map into small segments, and reconnects multiple segments into individual LKFs based on their distance and orientation relative to each other. The tracking algorithm uses sea-ice- drift information to estimate a first guess of LKF distribution and identifies tracked features by the degree of overlap between detected features and the first guess. An optimization of the parameters of both algorithms, as well as an extensive evaluation of both algorithms against handpicked features in a reference dataset, is presented. An LKF dataset is derived from RGPS deformation data for the years 1996–2008 that enables a comprehensive description of LKFs. LKF densities and LKF intersection angles derived from this dataset agree with previous estimates. Further, a stretched exponential distribution of LKF length, an exponential tail in the distribution of LKF lifetimes, and a strong link to atmospheric drivers, here Arctic cyclones, are derived from the dataset. Both algorithms are applied to the output of a numerical sea-ice model to compare the LKF intersection angles in a high-resolution Arctic sea-ice simulation with the LKF dataset.


An inter-comparison of the mass budget of Arctic sea ice and snow in CMIP6 models

Ed Blockley, Ann Keen

Corresponding author: Ed Blockley

Corresponding author e-mail: ed.blockley@metoffice.gov.uk

Sea ice is a key component of the global climate system and a very visible indicator of climate change. Arctic summer sea-ice cover has declined at a rate of over 13% per decade since satellite observation began, and there is much interest in how this decline will continue in the future. Global coupled models are arguably the best tool we have for making future predictions of Arctic sea ice, but generate a wide spread of projections of 21st-century decline. Comparing integrated quantities such as sea-ice extent and volume is not sufficient to understand the reasons for these differences in model projections. It is also necessary to consider, compare and evaluate the underlying processes causing ice growth and decline, and how they are likely to change in a warming world. This inter-comparison will use sea-ice budget diagnostics from the SIMIP data request to evaluate the mass budget of the Arctic sea ice and overlying snow over a defined region of the Arctic Ocean for the period 1960–2100. Area-weighted monthly-mean budget terms representing the dynamic and thermodynamic processes causing ice formation and loss will be calculated by participating modelling centres. We will compare the mean seasonal cycle of these budget terms for the 30-year reference period 1960–89 to determine how closely models agree on the relative importance of the processes causing the seasonal growth and decline of sea ice. Where possible, budget components will be compared with observations during the historical period. We shall then compare the evolution of these terms as the ice declines during the 21st century, to establish the dominant changes in the budget and investigate to what extent the changes are robust across models (or how they differ!). To date, 10 modelling centres have expressed interest in joining this inter-comparison, which is being coordinated through the budget sub-group of the Sea Ice Model Inter-comparison Project (SIMIP). Here we will provide an overview of the proposed model inter-comparison activity and present early results from those modelling centres that have completed the appropriate model integrations.


Towards a global sea-ice thickness climate data record from radar altimetry

Stefan Hendricks, Eero Rinne, Stephan Paul, Stefan Kern, Henriette Skourup, Robert Ricker, Thomas Lavergne

Corresponding author: Stefan Hendricks

Corresponding author e-mail: stefan.hendricks@awi.de

Sea-ice thickness (SIT) is one key indicator to understand the causes and consequences of Arctic change and differences between Arctic and Antarctic sea ice trends. Climate data records (CDRs) of sea-ice thickness in both hemispheres with sufficient length and accuracy are therefore a high priority dataset for climate research. While the CryoSat-2 mission, a dedicated satellite radar altimeter mission for the cryosphere, was pivotal for establishing routine SIT retrieval, its current data record of 9 years is of itself too short to separate climate trends from inter-annual SIT variability. Significant algorithm development has been made in the ESA Climate Change Initiative (CCI) to extend the SIT CDR by using its predecessor Envisat (2002–12). The production of the SIT CDR was then operationalized within the Copernicus Climate Change Service (C3S). The evolution from the pulse-limited radar altimeter RA-2 on Envisat to the SAR altimeter SIRAL onboard CryoSat-2 and the subsequent improvement in footprint size, however, poses a significant challenge for maintaining stability of the decadal SIT CDR. A separate challenge is the quality of auxiliary parameters, such as snow depth on sea ice and sea-ice density, that are required for the conversion of the freeboard estimates of the altimeter into SIT. For the actual SIT retrieval, snow-depth climatologies are applied for both hemispheres. In addition, SIT retrieval in the southern hemisphere is already particularly challenging at the level of freeboard estimation due to the complex snow on sea ice conditions and its interaction with radar signals. Here we show sea-ice thickness changes observed by the SIT CDR and its error characterization and stability estimations. The development of strategies to mitigate intermission biases and improve data record quality in both hemispheres is the objective of the ESA CCI+ Sea Ice project (2019–21). Based on user request, a high-priority activity is the extension of the CDR back to 1993 with data from the ERS-1/2 satellites. Continuity of the CDR is secured by the Copernicus Sentinel-3 constellation, although with lesser geographical coverage than CryoSat-2. Algorithm development in the CCI+ project will benefit from the joint mission phase of CryoSat-2 radar and ICESat-2 laser altimeter observations, for example to better quantify biases for different snow-depth and surface-roughness scenarios.


Sea-ice conditions in the Last Ice area: trends, comparisons and implications for ocean–sea-ice–atmosphere interface processes

Monika Pućko, Benjamin Lange, Steve Duerksen, Joannie Charette, Pascal Tremblay, Christine Michel

Corresponding author: Monika Pućko

Corresponding author e-mail: monika.pucko@dfo-mpo.gc.ca

The Last Ice area (LIA) is the only Arctic region that is expected to retain summer sea ice until 2050. Due to the importance of this region, Fisheries and Oceans Canada has initiated the Multidisciplinary Arctic Program – Last Ice Science Program, to characterize the ecosystem of the LIA. Here, we present the analysis of sea-ice conditions within the LIA region (1999–present) focusing on open-water period duration, multiyear-ice (MYI) versus first-year-ice (FYI) distribution, sea-ice break-up and summer-melt onset dates. We contrast the sea-ice conditions recorded since 1999 in the LIA region with those in adjacent Canadian regions and present trend analysis results. Finally, we discuss macro- and micro-scale implications of changing sea-ice conditions in the LIA for the sea ice habitats’ characteristics and biological processes that occur at the ocean–sea-ice–atmosphere interface.


Retrieval of surface and atmospheric parameters over sea ice from multi-frequency microwave radiometers: AMSR2 compared with the CIMR candidate mission

Raul Scarlat, Gunnar Spreen, Georg Heygster, Marcus Huntemann, Catalin Paţilea, Leif Toudal Pedersen, Roberto Saldo

Corresponding author: Gunnar Spreen

Corresponding author e-mail: gunnar.spreen@uni-bremen.de

At microwave frequencies both the Earth’ surface and its atmosphere influence the measurements of satellite radiometers. A special characteristic of the Arctic sea-ice-covered regions at microwave frequencies is the high emissivity of sea ice, with and without snow. Since the satellite-observed signal contains contributions from both surface and atmosphere, retrievals of atmospheric parameters require some information about the surface emissivity, and vice versa. Here we present a retrieval method that takes advantage of the multispectral capabilities of imaging radiometers in order to retrieve seven geophysical parameters: (1) sea-ice concentration, (2) multiyear-ice fraction, (3) ice-surface temperature, (4) total water vapor, (5) liquid-water path and, for open ocean areas, also (6) wind speed and (7) sea-surface temperature. A constrained optimal estimation technique is used to invert the forward model and extract the ensemble of seven parameters that optimally match the observed brightness temperatures. A priori information from climatological and meteorological sources is used to constrain the method to the natural variability of each parameter. The method can use observations from the current generation of space-borne radiometers such as AMSR2. In the near to medium future the Copernicus Imaging Microwave Radiometer (CIMR) candidate mission promises to offer high-resolution, low-uncertainty observation capabilities for the L-, C-, X-, Ku- and Ka-bands. To assess the potential impact of CIMR, a comparison is made between retrievals based on AMSR2 observations and a retrieval using CIMR-equivalent observations over a dataset of validated sea-ice concentration values. Individual channels or channel combinations can be used as input for the optimal estimation retrieval, which allows for flexibility in selecting frequency-specific parameter sensitivities. An information content analysis expands the comparison between AMSR2 and CIMR to all retrievable surface and atmospheric parameters. This analysis quantifies individual parameter contributions to the observed signal and highlights the differences between different input-channel combinations. The higher resolution of the low-frequency CIMR channels would allow for unprecedented detail to be achieved in Arctic passive-microwave retrievals.


On the role of atmospheric forcing and ice drift driving interannual variability in sea-ice thickness within Hudson Bay

Sergei Kirillov, David Babb, Igor Dmitrenko, Jack Landy, Jennifer Lukovich, Jens Ehn, David Barber

Corresponding author: Sergei Kirillov

Corresponding author e-mail: sergei.kirillov@umanitoba.ca

Sea-ice processes are a central part of the Hudson Bay System Study, which focuses on understanding the role of freshwater–marine coupling in Hudson Bay (HB). From this perspective, a continuous time series of sea-ice thickness obtained from three moored upward-looking sonars in 2016–18 represents a unique dataset for evaluating the interannual variability in the seasonal growth of the HB ice cover. During winter 2016/17, sea ice in western HB grew steadily to a peak modal thickness of 1.2–1.3 m in April, whereas during winter 2017/18 the ice only grew to a peak thickness of 0.7–0.9 m. In southeastern HB, the pattern was the opposite – thinner ice of 0.9 m was observed in April 2017, while the ice thickness reached 1.3 m in April 2018. We attribute this interannual east–west asymmetry to the different atmospheric forcing that prevailed during these two winters. Winter 2016/17 was characterized by low atmospheric pressure with relatively weak NNW winds and numerous reversals in the typical west-to-east ice transport. Conversely, winter 2017/18 was characterized by a high SLP gradient across HB that led to strong NW winds and across-bay transport of ice cover. Considering that NW winds are favorable for opening the polynya in northwestern HB, we suggest that during winter 2017/18 the ice cover over both western moorings was continually replaced by new ice formed upstream in the polynya. Furthermore, the numerous reversals in surface winds during winter 2016/17 also resulted in stronger deformation of the ice cover, which is evident in the ice-draft time series. The difference in atmospheric forcing regimes between these two winters not only contributed to changes in the ice cover at the mooring locations, but also impacted the regional sea-ice cover and the regional pattern of break-up. Using remotely sensed fields of ice drift (OSI SAF) and ice thickness (Cryosat-2) we further examine the difference between these two winters and examine the spatial distribution of remnant ice during the melt season. Overall, stronger NW winds during winter 2017/18 caused faster ice drift towards eastern Hudson Bay and dynamically thickened the ice cover, thereby delaying break-up and impacting the summer shipping season. In contrast, weaker atmospheric circulation during 2016/17 reduced dynamic ice growth in eastern Hudson Bay and left thicker ice types in the central part of HB where remnant ice cover melted out in southern HB, which is typical of previous years.


Beluga use of estuary habitats and effects of noise pollution

Emma Ausen, Marianne Marcoux, David Barber

Corresponding author: Emma Ausen

Corresponding author e-mail: ausene@myumanitoba.ca

The western Hudson Bay beluga populations return to estuaries near Churchill, Manitoba every summer. Beluga habitat associations within these estuaries are not fully understood, but theories include protection from predators, metabolic benefits from warm water, and rich habitat. Baseline studies have assessed the Churchill River beluga call types, beluga habitat associations, behavior and shipping traffic. Increased shipping traffic into the port of Churchill is likely as warming temperatures allow for greater access to the Arctic Ocean. This has unknown consequences for the belugas occupying the Churchill and Seal estuaries. This poster will outline methods and give initial findings on beluga distribution and shipping traffic in the estuaries. An aerial photographic survey of the Nelson River, Seal River, Knife River and Churchill River was completed during the BaySys summer cruise in June and July 2018. Georeferencing these photos using track logs from the flight in ArcMap will allow beluga locations to be inputted, counts obtained and the distribution compared with habitat features, including bathymetry, sea ice and sediment concentration. These data will reveal beluga habitat associations in the estuary that will provide information on critical habitat. In addition to the aerial survey data, acoustic recordings of ship traffic in and around the port of Churchill will be compared with beluga call characteristics and beluga distribution to determine if their behavior changes in response. The purpose of this study is to improve understanding on beluga estuary use and their response to increased shipping traffic, providing important information for management decisions.


Changing storm-surge regime as a result of sea-ice decline in the Canadian Beaufort Sea

Matthew Asplin, David Atkinson, Laura Eerkes-Medrano

Corresponding author: Matthew Asplin

Corresponding author e-mail: masplin@aslenv.com

Declining Arctic sea-ice cover and increases in fetch are increasing the risk of damaging storm surges along the Canadian Arctic coastline. Storm-driven changes in water levels can result in coastal flooding, increased wave erosion and low-water levels (negative surge). Extensive storm-induced flooding occurs under northwesterly winds mainly during the fall before sea ice has formed. Delayed freeze-up attributed to climate change maintains fetch in October when strong storms and winds can occur, and will likely increase the likelihood of storm flooding and the frequency of over-bank flooding. Although coastal ecosystems rely on sedimentation and salinization from small floods, large storm inundations can cause salinization of freshwater ponds and non-saline meadows, damage vegetation along the margins of permafrost plateaus, and melt subterranean permafrost, causing underground hollows subject to collapse (thermokarst). This will accelerate erosion rates in coastal areas, and will likely introduce subsequent hazards and challenges to local communities and ecosystems. This work presents results from an extensive analysis of meteorological and ocean variables. DFO/CHS water-level-gauge data from a number of water-level stations were extracted and reduced to identify surge events. Identification of synoptic meteorological drivers for extreme events were assessed using a synoptic climatology based upon principal components analysis and k-means clustering of gridded NCEP-NCAR II mean sea-level pressure data. High-resolution gridded wind–wave reanalysis data from the Meteorological Service of Canada Beaufort Sea Wind and Wave Reanalysis (Oceanweather Inc.) was retrieved at a 2.5 km resolution (0.1°) for the continental shelf area near Tuktoyaktuk, Northern Territories. Wind data were processed to determine the characteristics of wind divergence and convergence associated with each synoptic type. These parameters are then used to investigate meteorological drivers of ‘worst case’ storm-surge events, as well as negative surge events in coastal areas under varying synoptic meteorological and sea-ice conditions. Results are presented for all seasons, with emphasis on events near Tuktoyaktuk. Northern indigenous community knowledge is also incorporated into the study to better understand consequences on coastal communities, infrastructure and traditional socioeconomic activities.


Light in the water column beneath a melting Arctic sea-ice cover: observations and modeling

Bonnie Light, Regina Carns, Karen Frey

Corresponding author: Bonnie Light

Corresponding author e-mail: bonnie@apl.washington.edu

Partitioning of solar radiation by sea ice modulates light availability for primary productivity and the accumulation of heat in the ocean. Melting Arctic ice covers typically exhibit strong horizontal inhomogeneity: snow patchiness, melt ponding, leads, cracks, ridges, ice thickness, and floe size. Such variability can make measurements of the under-ice light field difficult to generalize. Light fields, water temperature, chlorophyll concentration, and CDOM absorption were measured beneath melting ice in the Chukchi Sea during the 2011 NASA-sponsored ICESCAPE campaign. Spectral downwelling irradiance (Fdn) and upwelling radiance (Lup) from just below the ice to 50 m depth were recorded at 19 wavelength channels with a C-OPS optical profiling system (Biospherical Instruments, Inc.) beneath both bare and ponded surface conditions. Measurements were made in three distinct water masses: (i) relatively clean, (ii) under-ice bloom (chlorophyll concentrations up to 10 mg m–3), and (iii) CDOM-enriched. As surface melt ponds caused strong horizontal inhomogeneity in the optical properties of the ice cover, measured light-field profiles consistently misrepresented water-column radiative transport. Examples include profiles taken beneath bare ice where the downwelling irradiance increased with depth owing to proximal ponds entering the field of view as the profiler depth increased, suggesting negative extinction coefficients in the water column. In this context, standard 1-D radiative-transfer modeling techniques are of little utility in determining where light is absorbed and how heat is distributed. A full 3-D model could be useful, but unnecessarily complicated and difficult to generalize. To address this problem, we propose a ‘bootstrap’ model to treat radiative transport in this horizontally inhomogeneous domain. The column is broken into layers of finite thickness and observed downwelling irradiance is assigned at the top of each layer. Layer-specific attenuation coefficients are calculated based on the absorption properties of water, observed chlorophyll and CDOM, and a multiple scattering correction derived from the observed Lup/Fdn ratio. A Beer’s Law approximation is applied for each layer. Results yield vertical light field patterns and heating profiles representative of each of the sampled water masses and estimates of heat accumulation in the water column are compared with observed in-situ temperatures.


Sea-ice thickness in a dynamic flaw polynya in southwestern Hudson Bay

David Babb, Greg McCullough, Jens Ehn, Sergei Kirillov, Ryan Galley, Madison Harasyn, John Iacozza, Klaus Hochheim, David Barber

Corresponding author: David Babb

Corresponding author e-mail: David.Babb@umanitoba.ca

Helicopter-based ice thickness surveys conducted in southwestern Hudson Bay on two days during March 2009 are used to examine sea-ice thickness in a highly dynamic sea-ice environment. Surveys covered the coastal band of landfast sea ice, the deformed pack ice and areas of new sea ice that formed within polynyas and the tidally driven flaw polynya in the area. Overall the landfast ice cover was very thick (mean 4.73; mode 3.6 m) and highly deformed with large ridges, broad rubble fields and pronounced stamukhi. Surveys across the landfast ice cover are compared to the seasonal evolution of the ice edge and show that large ridge features correspond to previous ice edges. Beyond the landfast ice, both surface winds and large tides drive a dynamic ice cover and promote continued formation of new ice. Prior to the first survey, westerly winds flushed the pack ice out of the study area, creating a vast polynya where new sea ice formed, giving the survey a modal thickness of 0.6 m and mean of 1.61 m. However, prior to the second survey onshore winds advected the thicker pack ice back into the study area, increasing the modal thickness to 1.6 m with a mean of 3.29 m. While surface winds drive this process on the scale of days to weeks, it also occurs at a smaller diurnal scale within the coastal flaw polynya that is driven by the large tides in the region. Essentially twice daily the low tide exposes an area of open water where new ice forms before being subsequently crushed during the next high tide. This repetitive cycle of ice growth and deformation contributes to the highly deformed ice cover of the area and promotes deformation along the ice edge. Using complimentary data from a pair of ice beacons, an on-ice weather station, satellite imagery, satellite-derived fields of sea-ice concentration and drift, atmospheric re-anlayses and complementary observations collected during another survey in 2017, we present an in-depth analysis of a highly dynamic ice cover in a coastal flaw polynya in southwestern Hudson Bay.


Where is the bottom of the sea ice?

Maren E. Richter, Greg H. Leonard, Inga J. Smith, Pat J. Langhorne, Andrew R. Mahoney

Corresponding author: Maren E. Richter

Corresponding author e-mail: maren.richter@postgrad.otago.ac.nz

A precise knowledge of the thickness of sea ice is important for many different reasons. It allows the volume of sea ice to be estimated, determines the mechanical strength of the sea-ice cover, influences the amount of light available under the ice and gives information on the growth and decay of the ice cover. However, defining the underside of the ice is not as trivial as it may appear, for example where there is a sub-ice platelet layer under the ice. Sea ice in McMurdo Sound, Antarctica is influenced by supercooled water exported from a nearby ice shelf, which allows the formation of a sub-ice platelet layer. This friable layer consists of ice platelets of seemingly random orientation under the consolidated sea ice and forms an important habitat for sea ice organisms as well as influencing the sea-ice growth rate. The transition, when there is a sub-ice platelet layer present, from solid ice to water, via a matrix of loose or semi-consolidated ice crystals, complicates the definition of sea-ice thickness. Here, we present an analysis of sea-ice thicknesses calculated from thermistor strings deployed in McMurdo Sound over two decades. We compare different methods of processing thermistor string data found in the literature to determine the ice–ocean interface with a view of comparing robustness, precision and accuracy of these methods in the presence of platelet ice. The results are compared to other instruments that use mechanical or acoustic properties to determine sea-ice thickness and to results from thermistor strings deployed in the Arctic, in a location unaffected by platelet ice. The resulting thickness time series can be used to study interannual variability and existing or emerging trends in McMurdo sea ice and the sub-ice platelet layer over the past 20 years. Although sub-ice platelet layers are rarely found outside the vicinity of deep-draft ice shelves, our results are of relevance to other situations in which the ice bottom is similarly ill-defined, such as the very early stages of growth or advanced stages of melt.


Identifying climatic drivers of Arctic sea-ice internal variability

Alexandra Jahn

Corresponding author: Alexandra Jahn

Corresponding author e-mail: alexandra.jahn@colorado.edu

Previous work has shown that internal variability has likely contributed significantly to the observed Arctic sea-ice loss over the satellite era. But the climatic mechanisms that lead to larger or smaller sea-ice loss within the possible spectrum of internal variability have not been identified. To address this gap, I investigate the underlying primary mechanisms driving the internal variability of Arctic sea-ice change over 1979–2017, using the Community Earth System Model Large Ensemble (CESM LE). This analysis will provide a physical understanding of the significant CESM LE model spread of Arctic sea-ice simulations. The insights gained in this work will allow a more process-based assessment of climate models against observations in the future, as we can assess the sea-ice evolution under climatic conditions occurring in the real world with periods of similar conditions in the climate models. It may also help to understand why internal variability in different climate model simulations differs, for example CMIP5 and CMIP6 models, based on the representation of the underlying mechanisms.


New observations of late summer bio-physical ice and snow conditions in the northwestern Weddell Sea

Christian Haas, Stefanie Arndt, Ilka Peeken

Corresponding author: Christian Haas

Corresponding author e-mail: chaas@awi.de

Summer sea-ice extent in the Weddell Sea has increased overall during the last four decades, with large interannual variations. However, the underlying causes and the related ice and snow properties are still poorly known. Here we present results of the interdisciplinary Weddell Sea Ice (WedIce) project carried out in the northwestern Weddell Sea on board the German icebreaker RV Polarstern in February and March 2019, i.e. at the end of the summer ablation period. This is the region of the thickest, oldest ice in the Weddell Sea, at the outflow of the Weddell Gyre. Measurements included airborne ice-thickness surveys and in-situ snow and ice sampling of mostly second- and third-year ice. Preliminary results show mean ice thicknesses between 2.6 and 5.4 m, increasing from the Antarctic Sound towards the Larsen B region. The ice had mostly positive ice freeboard. Mean snow thicknesses ranged between 0.05 and 0.46 m. Snow was well below the melting temperature on most days and was highly metamorphic and icy, with melt–freeze forms as the dominant snow type. In addition, as a result of the summer’s thaw, an average of 0.14 m of superimposed ice was found in all ice cores drilled during the cruise. Although there was rotten ice below a solid, ~30 cm thick surface-ice layer, pronounced gap layers typical for late summer ice in the marginal ice zone were rare, and algal biomass was patchily distributed within individual sea-ice cores. Overall, there was a strong gradient of increasing ice algal biomass from the Larsen B to the Antarctic Sound region. The presented results show that sea-ice conditions in the northwestern Weddell Sea are still severe and have not changed significantly since the last observations carried out in 2004 and 2006. The presence of relatively thin, icy snow has strong implications for the ice and snow mass balance, for freshwater oceanography, and for the application of remote-sensing methods. Overall sea-ice properties strongly affect the biological productivity of this region and limit carbon fluxes to the seafloor in the northwestern Weddell Sea.


Formation and variability of the Cape Darnley polynya, derived from ice-type and ice-production data by AMSR-E passive microwave observations

Kazuki Nakata, Kay Ohshima

Corresponding author: Kazuki Nakata

Corresponding author e-mail: nakata_kazuki@restec.or.jp

Coastal polynyas are thin-ice and/or low-ice-concentration areas formed by offshore winds. In Antarctic coastal polynyas, high production of sea ice occurs due to a huge heat loss to the atmosphere, resulting in the formation of dense shelf water, precursor of Antarctic bottom water (AABW). The Cape Darnley polynya (CDP), the second largest sea-ice production area, is the fourth formation area of AABW, with high biological productivity. Thin-ice thickness algorithms using satellite passive microwave radiometers have been developed to detect coastal polynyas and to estimate thin-ice thickness and sea-ice production. However, these algorithms cannot discriminate the type of thin ice, such as frazil/grease ice or nilas. Recently we have developed a new AMSR-E algorithm that can discriminate two categories of thin-ice type: one is ‘active frazil’, comprising frazil and open water, and the other is ‘thin solid ice’, an area of relatively uniform thin ice. This discrimination greatly improves the estimate of thermal ice thickness and sea-ice production. We investigate the formation and variability of the CDP, using the ice-type data derived from the new algorithm and ice-production data calculated from heat-budget analysis, with temporal and spatial resolution being daily and 12 km, respectively. The correlation analysis shows that the polynya (active frazil plus thin solid ice) area is not well correlated with the offshore wind, as also shown in the previous studies. However, the area of active frazil has quite a high positive correlation with offshore wind, while the area of thin solid ice has a significant negative correlation with the wind. For time series of ice-type data from the AMSR-E, we applied a simplified polynya model in which the active frazil area is determined by a balance between ice production and offshore ice drift, where the ice drift is estimated from the wind data. Our time-series analysis clearly shows that the model represents the expansion of the active frazil region well (explained variance of 70%). Further, combined analysis of the model and AMSR-E ice-type data suggests that the reduction of the active frazil region is achieved by replacement with a region of thin solid ice, which can explain its negative correlation with the wind. These analyses demonstrate the following polynya cycle for the CDP. The active frazil region is quickly expanded by the strong offshore wind and subsequently transformed into thin solid ice under calm conditions.


Developing an on-demand service module for mining geophysical properties of sea ice from high spatial resolution imagery

Dexuan Sha, Mengchao Xu, Chaowei Yang, Xin Miao, Hongjie Xie, Alberto Mestas-Nuñez

Corresponding author: Chaowei Yang

Corresponding author e-mail: cyang3@gmu.edu

Cloud computing and big data brought innovative methods for developing an open data platform. For sea-ice observation, geophysical properties extracted from high-spatial-resolution (HSR) imagery are important for coupled sea-ice and climate modeling and verification. But HSR imagery has been largely ignored as compared to moderate- or low-resolution satellite images and products because of its complex and heterogeneous nature in both space and time. We introduce an Arctic Cyberinfrastructures (ArcCI) portal to enable researchers to upload, collect, manage, analyze and explore HSR imagery of sea-ice observations. The ArcCI portal is a domain-specific open-source web portal based on cloud computing. The small-size imagery and processed-output dataset are stored in a distributed file system and a hierarchy index system is developed to efficiently manage and operate datasets. he ArcCI portal provides on-demand data-driven batch processing and modular analysis for image classification, geophysical parameters of sea-ice extraction and other specific services. The spatiotemporal visualization module creates 3-D virtual globes and allows fast interactive data exploration of extracted sea-ice features. Our aim is to provide an on-demand platform that focus on data sharing, visualization and analysis of sea ice. We expect the portal, with continuous accumulation of HSR data, to aid in polar sea-ice investigations. The portal developed for Arctic sea ice should be easily transferable to Antarctic sea ice.


Characterization of under-ice habitats in the Antarctic pack-ice zone: preliminary results from a summer voyage to the Weddell Sea

Klaus Meiners, Giulia Castellani, Fokje Schaafsma, Hauke Flores

Corresponding author: Klaus Meiners

Corresponding author e-mail: klaus.meiners@aad.gov.au

Sea ice plays a key role in Southern Ocean physics and biogeochemical cycles, and is a major driver of Antarctic marine ecosystem processes. In terms of supporting marine food webs, sea ice provides a substrate for ice algae, which serve as a food source for pelagic herbivores, e.g. Antarctic krill (Euphausia superba). The under-ice environment also provides a spatially complex refuge from predators. Coincident measurements of sea-ice parameters and krill under the sea ice are key to better understanding the habitat utilization of this Antarctic keystone species, and to assess the vulnerability of pelagic herbivores to changing sea-ice conditions. During a summer voyage to the Weddell Sea (RV Polarstern PS117, December 2018–February 2019) we combined classical ice coring methods with the deployment of novel instrumented under-ice observing platforms to collect the first concomitant measurements of ice algal biomass and the abundance of Antarctic krill at the sea-ice–water interface, under different types of sea ice. In particular, we deployed horizontally profiling platforms (remotely operated vehicle(ROV) and surface and under-ice trawl (SUIT)) to measure ice-algal biomass and krill abundance along 100–1000 m long transects. Algal biomass was estimated from transmitted under-ice irradiance data, and cross-calibrated with point measurements from ice-coring surveys. Krill abundance data were determined from the SUIT net catches as well as from images of the ice–water interface taken with an upward-looking stills camera mounted to the ROV. Our preliminary data show high small-scale spatial variability in both ice- algal biomass and krill abundance. First analyses indicate that the Weddell Sea marginal ice zone, particularly areas characterized by small ice floes separated by a high amount of brash ice, harbour high ice-algal standing stocks with associated high abundances of krill dwelling directly at the sea-ice–water interface during summer.


Sea-ice thickness and volume in the Sea of Okhotsk estimated on the basis of ICESat and CryoSat-2 data

Sohey Nihashi, Nathan T. Kurtz, Takenobu Toyota

Corresponding author: Sohey Nihashi

Corresponding author e-mail: sohey@tomakomai-ct.ac.jp

Sea-ice thickness in the Sea of Okhotsk is estimated from freeboard derived from satellite altimeter data (ICESat, 2004–08; CryoSat-2, 2011–18). Comparison with ice thickness in the southern Sea of Okhotsk observed hourly aboard icebreaker Soya by visual observation suggests that ice thicknesses derived from ICESat and CryoSat-2 are almost consistent. Total ice thickness (snow depth plus ice thickness) in February and March, when the sea-ice area is maximum, averaged over the entire sea-ice zone ranges from 77.5 cm (2008) to 123.0 cm (2017). The mode of total ice thickness ranges from 50–60 cm (2007; 2008; 2017) to 90–100 cm (2013). Sea-ice volume in the Sea of Okhotsk is estimated from the total ice thickness and ice concentration derived from AMSR-E (2004–11), SSMI (2012), and AMSR2 (2013–18). Ice volume is estimated by interpolating these sea-ice datasets onto an NSIDC polar stereographic grid at a spatial resolution of about 12 km. The maximum ice volume is 8.3 × 1011 m3 (2016), while the minimum is 5.4 × 1011 m3 (2015). The interannual variability is shown to be mostly determined by the sea-ice area. Multiple regression analysis on the ice volume using atmospheric and oceanic data reveals that the ice volume is mainly determined by winter air temperature near coastal polynya and autumn sea-surface temperature around the East Kamchatka Current where the upstream area for the Sea of Okhotsk. We reproduce ice volume in the Sea of Okhotsk since 1958 when the satellite data does not exist by using the regression line. The ice volume can be divided into three regimes of every 20 years; for 1958–78, the sea-ice volume increased; for 1978–98, the ice volume decreased significantly; for 1998–2018, the ice volume didn’t change much. We also try to predict future ice volume based on some scenarios.


Assessment and improvement of sea-ice processing for dissolved inorganic carbon analysis

Yubin Hu, Feiyue Wang, Søren Rysgaard, David Barber

Corresponding author: Yubin Hu

Corresponding author e-mail: yubinhu@sdu.edu.cn

Dissolved inorganic carbon (DIC) is an important parameter for characterizing the biogeochemical processes in sea ice and across the ocean–sea-ice–atmosphere interface. The main challenge in bulk sea-ice processing for DIC analysis is to melt the ice core without exposure to the air, which otherwise might contaminate the sample. A common practice is to seal the ice core in a gas-tight plastic bag and remove the air gently using a syringe or a hand pump. However, this procedure is time-consuming and the uncertainty in DIC concentration processed in this way has not been fully accessed. In this study, we modified the method by using a vacuum sealer and evaluated this procedure by examining the impact of ice sample processing, biological activity, gaseous CO2 initially present in sea ice, and the presence of ikaite (CaCO3.6H2O) crystals. The results show that no loss or gain in DIC occurs during the evacuation and ice-melting process and that it might not be necessary to pre-poison the ice samples during the ice-melting process. In addition, gaseous CO2 initially present in sea ice has a negligible impact on DIC analysis. If detectable ikaite crystals are present in sea ice, the measurement results should be referred to total inorganic carbon rather than DIC. The field test at Station Nord in Greenland demonstrates that the modified method is simple and quick to use even under the most remote and extreme environments.


Use of high-resolution satellite products with the neXtSIM sea-ice model

Einar Olason, Pierre Rampal, Véronique Dansereau, Anton Korosov

Corresponding author: Einar Olason

Corresponding author e-mail: einar.olason@nersc.no

In this presentation, I will outline the main use of high-resolution satellite products within the Sea-ice Modelling Group at the Nansen Center, Bergen, Norway. Our model evaluation methods are centred on the fact that sea-ice deformation displays scaling invariance properties in both the spatial and temporal domains. I will show that the model is able to reproduce the observed properties of these scalings in both the spatial and temporal domains over a wide range of scales and their multi-fractality. I will then also show how modelled and observed lead fraction displays mono-fractal scaling in space, which can be related to the scaling properties of deformation. Finally, I will demonstrate the assimilation of deformation rates into neXtSIM and how this allows the model to capture the location of deformation events, in addition to their statistics.


Radiative transfer model of sea ice and its validation with field measurement of the spectral albedo of sea ice at Saroma Lagoon, Japan

Tomonori Tanikawa, Toru Hirawake, Takenobu Toyota, Teruo Aoki, Masashi Niwano, Masahiro Hosaka, Masahiro Hori

Corresponding author: Tomonori Tanikawa

Corresponding author e-mail: tanikawa@mri-jma.go.jp

A radiative transfer (RT) model for coupled atmosphere–snow–sea-ice systems is developed to compute the spectral albedo of sea ice. In the model, we consider sea ice’s inherent optical properties (IOPs: single-scattering albedo, extinction optical depth, and scattering asymmetry parameter) for any wavelength between 300 and 2500 nm as functions of sea ice’s physical parameters, including the ice refractive index, brine-pocket concentration and size, air-bubble concentration and size, ice-impurity concentration (e.g. ice algae) and sea-ice thickness. On the snow over the sea ice in the model, snow IOPs as functions of snow-grain size, snow density, snow depth and light-absorption snow-impurity concentration are taken into account. This RT model generates spectral albedo and transmittance of the sea ice as functions of sun-sensor geometry, angles of illumination, and snow and sea-ice IOPs. In order to validate this RT model, field measurement of spectral albedo was made together with snow-pit work and sea-ice sampling on a flat sea-ice field in Saroma lagoon, Japan. We confirmed that the RT model is able to simulate spectral albedo with good agreement by comparing theoretical albedos with the measurements. Introducing the effects of ice impurities with chlorophyll concentration into the sea-ice optical properties can lead to significant improvements in the visible spectral albedo. This RT model would be useful for remote-sensing applications to monitor snow/sea ice parameters and for accurate climate-change predictions by regional and global climate models.


Application of ALOS-2/PALSAR-2 for detecting sea-ice-surface features in the seasonal ice zone

Takenobu Toyota, Junno Ishiyama

Corresponding author: Takenobu Toyota

Corresponding author e-mail: toyota@lowtem.hokudai.ac.jp

To improve our understanding of the deformation processes of sea ice, monitoring the temporal evolution of a deformed ice area is quite important. For this purpose, satellite L-band SAR is expected to be a useful tool because of its high spatial resolution, relatively wide coverage and longer wavelength relative to surface roughness compared with C-band SAR. As our analysis using PALSAR and Radarsat-2 for the Sea of Okhotsk also confirmed this property, we derived the algorithm that extracts deformed ice area by plotting the backscatter coefficients (BS) of PALSAR (HH) against incidence angle. However, since the data were limited, further validation is needed to confirm it. Since 2014, PALSAR has been followed by PALSAR-2, adding the functions of wider coverage with wider incidence angles (8–70°) and dual polarization (HH, HV). Thus, the purpose of this study is to examine the applicability of our algorithm to PALSAR-2 and the usefulness of dual polarization to detection of deformed ice, based on the field data obtained during the PV Soya cruise in the southern Sea of Okhotsk over 3 years (2016–18). Since the ice conditions were significantly different among these 3 years, with mean ice thickness being 19, 38 and 46 cm in 2016, 2017 and 2018, respectively, it may facilitate the interpretation of the properties through comparative analysis. To examine the BS produced by the surface features of sea ice, we took the cross-section of BS along a line with the same incidence angle lying within the sea-ice area in each year. The results show that, although our algorithm reasonably predicts the ice type, much care is needed for the incidence angle <20° or >45°, that BS data are affected not only by surface roughness but also, significantly, by floe-size distribution, and that this effect works more effectively at HV than at HH polarization. From these results, it is likely that high BS values along the ice edge found in PALSAR images are attributable mainly to smaller ice floes, produced by wave–ice interaction, rather than to surface roughness. In the light of less sensitivity to incidence angle at HV, it seems that, whereas BS at HH is more reflected by deformed ice in the compact ice region, the HV data are more useful to detect the marginal ice zones.


Effects of small-scale processes on air–sea-ice interactions in East Antarctica

Pierre-Vincent Huot, Thierry Fichefet, Nicolas Jourdain, Pierre Mathiot, Christoph Kittel, Clément Rousset, Xavier Fettweis

Corresponding author: Pierre-Vincent Huot

Corresponding author e-mail: pierre-vincent.huot@uclouvain.be

The current generation of global climate models exhibits large biases in the Southern Ocean. While key processes governing heat and freshwater exchanges – such as polynyas or ocean–ice-shelf interactions – take place in coastal Antarctica, the resolution of global models is too coarse to adequately represent them. Parameterizations are used to simulate their effect, but they rely on sparse observations and might be too crude to catch the complexity of air–sea-ice interactions. Knowing whether or not small-scale processes are relevant for the study of the Antarctic climate and if current parameterizations are adequate to represent them is yet unclear. Here, we propose to evaluate the sensitivity of air–sea-ice interactions to the representation of small-scale processes. To do so, we developed a very high-resolution model of the ocean and sea ice off Adélie Land, East Antarctica. We use it to evaluate the source of variability in sea-ice growth and ice-shelves basal melt, from hourly to seasonal time scales. First, tidal forcing and ice-shelf cavities are removed, with the aim of understanding how they affect ocean–ice interactions. Tides strongly decrease winter sea-ice growth above shoals, where the amplitude of tidal velocities is found to be large. Here, tides increase vertical mixing, which warms the surface layer and limits surface freezing. A similar increase in heat supply is found beneath the ice shelves, but this effect differs from one glacier to the other. This spatial variability is due to substantial modification of ocean circulation in coastal seas due to tides. Then, we focus on the role of interactions with the atmosphere, and on the importance of forcing resolution. A set of simulations is performed where outputs from an atmospheric regional model at different resolution are used as forcing. Our aim is to better represent katabatic winds and assess their impact on ocean circulation, sea-ice production and ice-shelves basal melt. With this set of experiments, we propose to evaluate the relative importance of local interactions due to small-scale processes over large-scale variability in governing the regional Antarctic climate.


Influence of the initial ocean–sea-ice state on the predictability of the Antarctic sea ice at the seasonal timescale: a study with NEMO3.6-LIM3

Sylvain Marchi, Thierry Fichefet, Hugues Goosse

Corresponding author: Sylvain Marchi

Corresponding author e-mail: sylvain.marchi@uclouvain.be

The Southern Hemisphere sea-ice extent has experienced an overall positive trend over the last 30 years. However, after a record high in 2014, the sea-ice extent in 2017 decreased to its lowest value since the beginning of satellite measurements in 1979. Due to the unprecedented melting rate in December 2018, the Antarctic sea-ice extent headed for a new record summer minimum in 2019 before stabilizing. These rapid sea-ice fluctuations exemplify the high seasonal and year-to-year variability of Antarctic sea ice. The reasons for these recent changes are still the subject of active research. Predicting these anomalies several months in advance is of prime importance for multiple activities, including the organization of scientific field campaigns. Besides, exploring the sources of Antarctic sea-ice predictability at sub-seasonal to interannual timescales certainly helps refine our understanding of Southern Ocean variability and the way the Southern Ocean interacts with sea ice. Although there is evidence that recent changes in Antarctic sea ice have been triggered by atmospheric forcing, the Southern Ocean has been pointed out as a source of predictability of sea-ice cover. In this study, we explore the influence on sea ice predictability over 1 year of a biased initial ocean state. In this respect, a control simulation covering the period 1980–2016 was performed using the ocean–sea-ice model NEMO3.6-LIM3. The model was driven by atmospheric fields derived from the JRA-55 reanalysis. For one specific control year, the perturbation of the initial ocean state was achieved by simply selecting the 36 other control ocean states simulated over the period 1980–2016 by the model at the same time. Repeating this procedure for the 37 years of the control simulation, i.e. for the 37 different atmospheric forcing years, led us to create 37 × 37 simulations. We demonstrate how unsatisfactory it might be to attempt to produce the best possible sea-ice forecasts from an uncertain ocean state, even if the atmospheric conditions were perfectly known, which is never the case in reality.


Evidence of freezing pressure in sea ice discrete brine inclusions and its impact on aqueous–gaseous equilibrium

Ryan Galley, Lionel Mercury, Bruno Delille, jean-Louis Tison, Soren Rysgaard, Odile Crabeck

Corresponding author: Odile Crabeck

Corresponding author e-mail: O.Crabeck@uea.ac.uk

Sea ice in part controls surface-water properties and the ocean–atmosphere exchange of greenhouse gases at high latitudes. In sea ice, gas exists dissolved in brine and as air bubbles contained in liquid brine inclusions or as bubbles trapped directly within the ice matrix. Current research on gas dynamics within the ocean–sea-ice–atmosphere interface has been based on the premise that brine with dissolved air becomes supersaturated with respect to the atmosphere during ice growth. Based on Henry’s law, gas bubbles within brine should grow when brine reaches saturation during cooling, given that the total partial pressure of atmospheric gases is above the implicit pressure in brine of 1 atm. Using high-resolution light-microscopy time-series imagery of gas-bubble evolution inside discrete brine pockets, we observed bubbles shrinking during cooling events in response to the development of freezing pressure above 3 atm. During warming of discrete brine pockets, existing bubbles expand and new bubbles nucleate in response to depressurization. Pressure variation within these inclusions has direct impacts on aqueous–gaseous equilibrium, indicating that Henry’s law at constant pressure of 1 atm is inadequate to assess the partitioning between dissolved and gaseous fractions of gas in sea ice. This new evidence of pressure build-up in discrete brine inclusions controlling the solubility of gas and nucleation of bubbles in these inclusions has the potential to affect the transport pathways of air bubbles and dissolved gases within the sea ice–ocean–atmosphere interface and modifies brine biochemical properties.


A four-component coupled configuration for assessing decadal predictability in Antarctica

Charles Pelletier, Hugues Goosse, François Klein

Corresponding author: Charles Pelletier

Corresponding author e-mail: charles.pelletier@uclouvain.be

Within a context of global climate warming and polar amplification, observations of the Antarctic climate conceal a unique spatial pattern: some of its regions have warmed less than the global average with some sea-ice advance; others have warmed significantly and displayed sea-ice loss. While the ability to predict decadal climate variations has been confirmed in various parts of the world, polar regions have been given little attention so far, mostly because, in these areas, observational data is sparser than elsewhere and global climate models suffer from systematic biases. Moreover, the Antarctic climate has been shown to strongly depend on complex feedback mechanisms between several distinct components. Finally, global climate models involved in current decadal prediction experiments run at a rather coarse (~1°) resolution and are therefore likely missing a whole variety of smaller-scale processes holding predictability. For these reasons, performing relevant decadal predictions in polar regions is only possible if the coupling of the climate components is properly taken into account, and if cutting-edge models are used on local configurations, allowing higher resolution, while making full use of the ever-increasing observational database at the disposal of the community. We introduce a new fully coupled (ocean, sea ice, land, atmosphere, ice sheet) Antarctic model configuration which aims at fulfilling the conditions described above.


Investigating the impact of ocean-thermal-content variability on the 2016 sea-ice-extent events

Charles Pelletier, Hugues Goosse, François Klein

Corresponding author: Charles Pelletier

Corresponding author e-mail: charles.pelletier@uclouvain.be

The Southern Ocean (SO) sea ice extent showed unprecedented and unanticipated properties in 2016. In that year, the SO sea ice extent minimum was significantly lower than usual, while its winter maximum occurred much earlier than what recent records were showing. A large part of these events’ causes can be attributed to special atmospheric conditions over the peri-Antarctic region, reverberating on the SO through air–sea fluxes. However, the ocean thermal content also needs to be taken into account for enhancing future predictions of such extreme sea-ice events. Here we investigate this aspect using several simulations obtained from a new NEMO-LIM eddy-permitting (1/4°), ice-shelf-cavities-including Southern Ocean configuration. A particular focus will be given to the vertical heat fluxes in the ocean and how the conditions in summer and fall have an impact on the ocean and sea-ice state at the winter maximum of the ice extent.


Effects of ice stessors and pollutants on the Arctic marine cryosphere

Rui Shen, Ralf Ebinghaus

Corresponding author: Rui Shen

Corresponding author e-mail: rui.shen@hzg.de

It has been widely recognized that global climate change is leading to substantial changes in Arctic ecosystems. Sea ice is often quoted as a key indicator of these changes. Continuing losses of multiyear sea ice (MYI) across the Arctic are causing first-year sea ice (FYI) to dominate the Arctic ice pack. Seasonal freezing and thawing of FYI affect the biogeochemical cycling and the behavior of organic contaminants, especially the ‘legacy’ and ‘emerging’ organic contaminants. In this work, we place our emphasis on the behavior of per- and polyfluoroalkyl substances (PFASs) in FYI, due to the serious health and environmental concerns caused by PFASs. The aim of this work is to improve our understanding of the dynamics of PFASs in response to the evolution of the Arctic sea ice under a warming climate. We reframe the question of the extent of temporal and spatial storage capacity of FYI and whether this is associated with the evolution of FYI. An array of experiments are conducted at the Roland von Glasow Air-Sea-Ice Chamber (RvG-ASIC) facility of the University of East Anglia, UK. These experiments address the ice–ocean interactions and focus on simulated inclusion and melting scenario with respect to PFASs. The observed variability in inclusion of PFASs during seasonal growth is quantitatively characterized in comparison with baseline data on physical properties regarding the evolution of sea ice. Different types of FYI including columnar congelation ice and artificially desalinated columnar-textured ice are investigated. We represent analytic results of the evaluation on the temporal and spatial storage capacity of the Arctic sea ice. The influences of each parameter on the behavior of PFASs associated to the evolution of Arctic sea ice are discussed. As the ice cover goes through seasonal warming and melting, redistribution of PFASs occurs in the ‘warm’ ice. In addition, the release of PFASs from sea ice back into the marine environment establishes a predictable pattern. We describe a novel and reproducible experimental design on the sea-ice chamber. Results in this work can be used for experimental guidance on how to make the chamber method efficient and fidelity. This study helps to refine the role of cryospheric compartments in global biogeochemical cycling of ‘legacy’ and ‘emerging’ organic pollutants.


The spatial distribution and temporal evolution of sunlight under Arctic sea ice

Don Perovich, Bonnie Light

Corresponding author: Don Perovich

Corresponding author e-mail: donald.k.perovich@dartmouth.edu

The Arctic sea-ice cover has undergone a significant decline in recent decades. The melt season is starting earlier, ice is thinner, and first-year ice dominates. Here we examine the effects of these changes on the light field under the ice. On a large scale, satellite observations of ice concentration, melt-onset date, ice age, and melt-pond coverage are combined with modeled ice thickness to calculate the contributions of leads, bare ice and ponded ice to the solar heat input to the upper ocean. We examine the spatial distribution of the solar input and examine how that input has varied seasonally and from year to year. While the light input to the ocean is most sensitive to the open-water fraction, the contribution from light transmittance through the ice has been increasing in recent years. This is due to earlier onset of melt, more first-year ice, and thinner ice. On a small scale in the Chukchi Sea, we explore the interplay of the changing physical properties of the sea-ice cover with the reflection, absorption and transmission of sunlight using a radiative transfer model, inherent optical properties for sea ice derived from previous field studies, and observations of snow depth, ice thickness and melt ponds. The impact of the variegated summer surface conditions on light transmission into the ice and upper ocean and the effects of algae in the ice and phytoplankton in the upper ocean are explored. As the snow melts, transmitted light increases by an order of magnitude in just a few days. When ponds form and develop, transmitted light levels further increase, as does the contrast between light levels under bare and ponded ice. Ice algae layers are highly absorbing and greatly limit transmission to the ocean.


Discrete-element modeling of wave-induced floe–floe collisions and wave attenuation in the marginal ice zone

Agnieszka Herman

Corresponding author: Agnieszka Herman

Corresponding author e-mail: oceagah@ug.edu.pl

Among mechanisms potentially contributing to wave-energy attenuation in fragmented sea ice are wave-induced floe collisions. At present, little is known about collision patterns and their phase-averaged effects under different combinations of sea-ice properties and wave forcing. In this work, selected aspects of interactions between sea-ice floes and waves are analyzed numerically with a discrete-element sea-ice model. The model is based on momentum equations for an arbitrary number of ice floes, with source terms computed by integrating local forcing (wave-induced dynamic pressure, surface drag, etc.) over the surface area/volume of each floe (and not by specifying the forcing at the center of each floe, as in similar earlier models). In the first set of simulations, the model is run with prescribed wave forcing for different combinations of parameters: ice concentration, wave steepness, relative floe size, as well as restitution coefficient ε and ice–water drag coefficient Cd. Different collision/motion patterns of ice floes are identified for different regions of the parameter space. In particular, it is shown that an important consequence of collisions, apart from generating stress within the ice, is their contribution to enhanced ice–water velocity differences and thus to enhanced drag at the bottom of the ice. This effect is especially strong when the floes are large and/or when the ice concentration is high enough so that the floe–floe contacts are prolonged and span a substantial fraction of the wave period. In general, high Cd and low ε are favorable for regular collision patterns with prolonged floe–floe contact, whereas low Cd and high ε tend to produce more energetic and irregular collisions. Crucially, different collision ‘regimes’ strongly influence phase-averaged effects of collisions, including those that are relevant for rheology of this type of sea ice: kinetic and contact stress, granular temperature, and work done by forces acting on the ice. In the second set of simulations, instead of prescribing the wave amplitude, wave attenuation is computed based on energy dissipation due to ice–water drag. It is shown that the quadratic drag formulation used in the model produces non-exponential attenuation; the dependence of attenuation coefficient on wave frequency changes with changing floe size. The results are compared with laboratory observations of wave propagation in broken ice.


A discrete element sea-ice model for climate applications

Adrian Turner, Kara Peterson, Dan Bolintineanu, Andrew Roberts, Min Wang

Corresponding author: Adrian Turner

Corresponding author e-mail: akt@lanl.gov

The current sea-ice component of the US Department of Energy’s Energy Exascale Earth System Model (E3SM) approximates the sea-ice cover as a continuous material rather than as a series of discrete floes and assumes that sufficient cracks exist within each model to ensure an isotropic distribution of crack orientations. Such models were developed for grid resolutions of ~100 km, whereas current models, including E3SM, are routinely applying these physics at much higher model resolutions of ~5 km. Evidence from both remote sensing and in-situ observations suggest that ~10 km represents a transition scale below which the dynamics of individual floes dominate the dynamics of sea ice. To correct the deficiencies of the current E3SM sea-ice model, we are developing a new sea-ice dynamic core based on the discrete element method (DEM). In this method collections of floes are explicitly modeled as discrete elements, contact forces between the elements are determined and equations of motion for individual elements are integrated in time. This new model uses the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) model for its dynamic core, and the IcePack library, provided by the CICE consortium, for its column physics. Here, we describe the development of the model, including the development of element contact models suitable for sea ice at climate scales, attempts to improve model performance by using the Kokkos framework to allow efficient computation on heterogeneous computing architectures, and progress on methodologies to ameliorate the effect of element distortion during deformation.


Distribution of volatile hydrocarbons within sea ice: a mesocosm study

Katarzyna Półćwiartek, Diana Saltymakova, Nolan Snyder, Durell Desmond, Gary A. Stern, Feiyue Wang

Corresponding author: Katarzyna Polcwiartek

Corresponding author e-mail: polcwika@myumanitoba.ca

The Canadian Arctic is currently experiencing significant increases in vessel traffic due to an extended open-water season caused by the warming climate. This has given rise to a greater risk of accidental spills of fuel and other transportation-related contaminants. Environmental risk assessment, oil-spill contingency planning, and oil-spill mitigation operations in the Arctic environment can only be properly assessed when interactions occurring between oil and sea ice are understood. These interactions depend on the physical–chemical characteristics of oil, the thermodynamic and physical properties of sea ice, and the environmental conditions. This study has been designed to better understand the vertical distribution of toxic volatile and water-soluble aromatic hydrocarbons in sea ice. Compounds including benzene-toluene-ethylbenzene-xylene compounds (BTEX) and polycyclic aromatic hydrocarbons (PAHs) were investigated using a mesocosm-scale experiment that took place at the Sea-ice Environmental Research Facility (SERF) of the University of Manitoba between January and April 2018. The experiment involved two pools filled with formulated seawater. Once the ice thickness reached approximately 20 cm, ~6 L of a light sour crude oil (provided by Tundra Oil & Gas Partnership) was injected into the water column from the bottom of each tank. Ice-core and water samples were collected at various times throughout the experiment; they were then subsampled for chemical and microbiological analyses. In the course of the experiment, sea-ice physical properties such as salinity, temperature and brine volume fraction were measured concurrently. Aromatic hydrocarbons were analyzed using gas chromatography with static headspace injection on an Agilent QQQ GC-MS. Preliminary results revealed increased presence of the volatile compounds in the surface layers of the ice cores. This suggests a greater propensity for these compounds to penetrate and migrate up through the ice brine channels independently of the bulk oil behaviour. Their composition and concentration within the ice cores shifted towards the heavier compounds within all targeted groups over time, which could be caused by concurrent processes such as chemical partitioning and/or microbial activity. However, further analyses are necessary to establish precisely to what extent each of these processes contributes to the distribution of targeted compounds.


Low sea ice concentration in the central Arctic and its impact factors

Jie Su, Cheng Li, Jinping Zhao, Hongjie Liang

Corresponding author: Jie Su

Corresponding author e-mail: sujie@ouc.edu.cn

The central Arctic has experienced low sea-ice concentration (SIC) in recent years. To study the occurrence of this unusual phenomenon and its impact factors, the Low Sea Ice Concentration in Central Arctic (LCCA) index is defined by using several pieces of SIC data. Although results from each dataset give a different index value, it is obvious that the LCCA process exists after 2010. The analysis results show that the leading factor in low SIC is not the local air temperature. Dynamically, the drifting pattern of sea ice and the location where the low SIC occurred respond consistently to the atmospheric circulation. In particular, cyclones used to be found north of 70° N before the LCCA index reached peak value. These cyclones moved northwards with warm air from lower latitudes causing sea-ice divergence and rapid melting of sea ice. Frequently, cyclones were accompanied by the dipole anomaly atmospheric circulation pattern. The LCCA index correlates positively with northward heat advection across the circle of 84° N as well as the divergence of central Arctic sea ice.


High-fidelity sea-ice simulation: some reflections on resolution and realism

Daniel Feltham

Corresponding author: Daniel Feltham

Corresponding author e-mail: d.l.feltham@reading.ac.uk

I reflect on some of the assumptions built into the structure of contemporary, continuum sea-ice models. These models are now being used in a variety of applications by a wide range of users. Numerical implementations of these models typically discretize space and time. For many applications it would be useful if the model solutions were accurate at fine resolutions not anticipated during model development. For example, accurate simulation of fluxes and feedbacks between sea ice, the ocean and the atmosphere may require resolution of the local, baroclinic ocean Rossby radius (~2–5 km). However, interpreted most literally, these length and timescales violate the continuity assumption upon which the sea-ice models are built, i.e. that each grid cell, over a single timestep of simulation, contains a statistically representative sample of sea ice. Does this matter? To gain insight into this question, I discuss modelling work that has attempted to relate the continuum behavior of statistically representative samples to underlying discrete modes of failure, and discuss scale dependence, or independence, of relevant processes. This will touch upon sea ice rheology, state dependence, and anisotropy. Awareness of these areas of uncertainty may be useful when it comes to the use of, and perhaps especially the interpretation of results from, our current generation of sea-ice models.


Sea-ice freeboard and thickness retrievals from 2016 and 2017 Ross Sea airborne lidar

Liuxi Tian, Hongjie Xie, Stephen F. Ackley, Kirsty Tinto, Robin E. Bell, Christopher J. Zappa, Yongli Gao, Alberto M. Mestas-Nuñez

Corresponding author: Hongjie Xie

Corresponding author e-mail: Hongjie.Xie@utsa.edu

As part of the Polynyas and Ice Production in the Ross Sea (PIPERS) project in November 2016 and 2017, the LDEO IcePod system onboard the NSF C-130 aircraft based at McMurdo Station was flown over the Ross Sea, Antarctica, with the purpose of repeating the same lines that NASA’s IceBridge aircraft flew over in 2013. A particular emphasis was to repeat the line along the FluxGate, the line separating the continental shelf from the deep ocean (shelf–slope break). The IcePod lidar and digital camera systems used are part of an externally mounted modular sensor and data acquisition system that can be attached and detached relatively easily from a C-130. The IcePod is ~2.6 m long by 0.6 m wide with the interior divided into three bays of ~0.6 m each. Lead detection is the key step to retrieve sea-ice freeboard, as leads represent the local sea level to compare to adjacent sea-ice elevation and derive ice thickness. Two methods to classify leads from IcePod LiDAR data were explored in this study. As leads have lower reflectivity than ice, using reflectivity values is the first way to differentiate them. Using some percentage of lowest-elevation values to represent the lead elevation is the second method. Results from both methods are compared in this study. IcePod camera data is used to validate the results from the LiDAR analyses of leads. The original LIDAR data is then resampled into different pixel sizes to assess the impact of spatial aliasing on sea ice thickness. The best estimation of the sea ice freeboard and thickness will be provided. Most of the IcePod data are over the same flight lines taken by IceBridge in 2013, so the thickness changes from 2013 to 2016 and 2017 will be calculated. Combining with the ICESat (2003–08) and the ongoing ICESat-2 data (available since 2018), we will be able to get a better picture of sea-ice thickness and its interannual variability in the Ross Sea.


Contrasting the sea-ice-algae chlorophyll a and snow-depth-dependent irradiance relationships between Arctic MYI and FYI

Benjamin A. Lange, Christian Haas, Joannie Charette, Christian Katlein, Karley L. Campbell, Steve Duerksen, Pierre Coupel, Philipp Anhaus, Arttu Jutila

Corresponding author: Benjamin A. Lange

Corresponding author e-mail: benjamin.lange@dfo-mpo.gc.ca

Multiyear sea ice (MYI) is continuing to disappear, yet our understanding of how the Arctic ecosystem will change with the replacement of MYI by first-year sea ice (FYI) is poorly understood. Contrasting bio-physical properties of MYI with FYI from the same region can provide key insights into the future Arctic Ocean. Here we present observations from the 2018 Multidisciplinary Arctic Program – Last Ice in the Lincoln Sea, where we sampled 45 multiyear (MYI, range: 2.1–4.6 m) and 34 first-year ice (FYI, 1.4–1.8 m) cores, combined with snow depth, ice thickness and transmittance surveys from adjacent FYI and MYI sites. Our results of FYI support sea-ice-algae chl-a biomass–snow-depth de-coupling patterns documented from lower latitude studies. MYI, however, showed a coupled ice algae chl-a biomass–snow-depth relationship. We also observed a spatio-temporal change in the pattern of percent PAR transmittance in FYI based on under-ice ROV surveys but no change in MYI. The more stable light field under MYI drives the coupling of the chl-a–snow-depth relationship due to the undulating surface of MYI, which creates a consistent snow-redistribution pattern of snow removal from high-elevation regions and deposition into low-elevation regions. On the other hand, the more variable light field under FYI explains the de-coupling of the chl-a–snow-depth relationship due to the continuously varying snow cover, which is the result of wind-driven re-distribution of snow over a more level surface. Although our results indicated that the overall mean chl-a biomass was higher for FYI compared to MYI, MYI provides a more predictable and stable light environment for ice algae, which may be an important factor in terms of access and reliability of a food source for ice-associated organisms. Lastly, this comparison provides key insights for the parameterization of ice-algal biomass and irradiance relationships to support predictions for Arctic change with the impending replacement of MYI by FYI.


Arctic sea-ice classification and validation using a multi-frequency fully polarimetric and interferometric airborne F-SAR system

Suman Singha, Marc Jäger

Corresponding author: Suman Singha

Corresponding author e-mail: Suman.Singha@dlr.de

Along with increasing scientific interest in sea ice, the operational aspect of high-resolution ice charting is becoming more important due to growing navigational possibilities in an increasingly ice-free Arctic, especially through the marginal ice zone. Despite proven sea-ice classification achievements on single polarimetric SAR data, a fully automated, general-purpose classifier for single-polarimetric data has not been established due to large variation and incidence-angle dependencies of SAR backscatter. Recently, through the advent of polarimetric SAR sensors, polarimetric features have moved into the focus of ice-classification research. The higher information content of four polarimetric channels promises to offer a greater insight into the sea-ice scattering mechanism. While airborne and shipborne radar cannot always be used during adverse weather conditions, it provides us with unique simultaneous multi-frequency and fully polarimetric and interferometric observations, which is not possible at the moment using space-borne sensors. In this study, fully polarimetric data in L-, S- and X-band simultaneously acquired by DLR’s FSAR system are investigated. The specific dataset was acquired in the framework of DLR-DALO ARCTIC’15 campaign over west Greenland. The proposed supervised classification algorithm consists of two steps: The first step comprises a feature extraction, the results of which are ingested into a neural network classifier in the second step for training and validation. The usefulness of different polarimetric features at different frequency bands is investigated using mutual information analysis along with quantitative comparison of classification results at different frequency bands. In this study we also investigated for the first time single-pass Across Track Interferometry (XTI)–derived sea-ice-freeboard measurement validation of our classification results with XTI-derived freeboard measurements.


Assessing snow and sea-ice properties using an unmanned aerial vehicle: a comparison with ground-, airborne- and satellite-based observations

Benjamin Lange, Steve Duerksen, Pascal Tremblay, Arttu Jutila, Stefan Hendricks, Robert Ricker, Christian Haas, Christine Michel

Corresponding author: Benjamin Lange

Corresponding author e-mail: Benjamin.Lange@dfo-mpo.gc.ca

The logistical difficulties of polar research in ice-covered regions have resulted in spatial and temporal knowledge gaps of key physical and biological processes. Airborne and satellite-based approaches to remotely map the marine environment provide synoptic observations at regional to basin-wide scales, which have been crucial for our current understanding of climate-change impacts. This large spatial coverage typically comes at the expense of better spatial resolution. Bridging the gap between ground-based local measurements and larger-scale airborne or satellite-based surveys is required to resolve the appropriate scales of variability on the order of hundreds to thousands of meters. However, ground-based operations in polar regions can be unsafe, logistically demanding and unfeasible. With the rapid advances in unmanned aerial vehicle (UAV) technologies and the increased affordability that comes with widespread usage and application, UAVs have become an obvious choice for research applications to improve the spatial coverage of observations in polar environments. Here, we present a case study of UAV surveys conducted on landfast first-year sea ice and multiyear sea ice in the Lincoln Sea, as part of the 2019 Multidisciplinary Arctic Program (MAP) – Last Ice. We used an Indro Robotics M210C RTK model quad-copter style UAV. The UAV was flown in systematic overlapping grids (~200 × 200 m), with visible (DJI Zenmuse X5S), thermal infrared (DJI Zenmuse XT Flir) and multi-spectral cameras (MicaSense RedEdge). The UAV surveys were followed by snow-depth measurements within the grids and were collocated at three locations along airborne surveys conducted with the Polar 6 aircraft. The Polar 6 was equipped with an electromagnetic (EM) induction sounding ice-thickness instrument (EM bird), a Snow Radar and a laser scanner to survey total ice thickness, snow depth and surface roughness. We present a comparison of the UAV-derived snow and ice-surface properties (digital surface models and surface optical properties) with airborne observations of snow and ice thickness and surface properties. In addition, we compare all results with RADARSAT-2 SAR imagery.


Enhanced bottom-ice algal biomass across a tidal strait in the Kitikmeot Sea, Canadian Arctic

Laura Dalman, Brent Else, David Barber, Eddy Carmack, Bill Williams, Karley Campbell, Patrick Duke, Sergei Kirillov, C.J. Mundy

Corresponding author: Laura Dalman

Corresponding author e-mail: laura.dalman@umanitoba.ca

Sea-ice algae are an important contributor of primary production in the Arctic ecosystem. Within the bottom-ice environment, access to nutrients from the underlying ocean is a major factor controlling production, phenology and taxonomic composition of ice algae. Previous studies have demonstrated that tides and currents play an important role in driving the flux of nutrients to bottom-ice algal communities when biological demand during the spring bloom is high. In this study we investigate how surface currents under landfast first-year ice influence nutrient supply based on stoichiometric composition, algal chlorophyll a (chl-a) biomass and species composition during spring 2016, in Dease Strait, Nunavut, Canada. Stronger water dynamics over a shoaled and constricted strait dominated by tidal currents (tidal strait) supported turbulent flow more than 85% of the deployment duration in comparison to outside the tidal strait in an embayment where turbulent flow was only evidenced a small percentage (<15%) of the time. The system appeared to be nitrate-depleted with surface-water concentrations averaging 1.3 μmol L–1. Increased currents were significantly correlated with a decrease in ice thickness and an increase in ice-algal chl-a. Furthermore, pennate diatoms dominated the ice-algal community abundance with greater contribution within the strait where currents were greatest. These observations all support the existence of a greater nutrient flux to the ice bottom where currents increased towards the center of the tidal strait, resulting in an increase of bottom-ice chl-a biomass by 5–7 times relative to that outside the strait. Therefore, expanding beyond the long-identified biological hotspots of open water polynyas, this paper presents the argument for newly identified hotspots in regions of strong sub-ice currents but persistent ice covers, so called ‘invisible polynyas’.


Arctic sea-ice thickness and volume over the 20th century from measurements and reconstructions

Axel Schweiger, Kevin Wood, Jinlun Zhang

Corresponding author: Axel Schweiger

Corresponding author e-mail: schweig@uw.edu

Changes in Arctic sea ice are a fingerprint of natural and anthropogenic climate change. The dominant signal in sea-ice variability from 1979 to the present is the reduction of sea-ice extent, area and thickness. Prior to 1979, the state of our knowledge about sea-ice variability is limited to information about sea-ice extent and concentration assembled mostly from shipping logs, and very little is known about the variability of sea-ice thickness and total sea-ice volume. Here we use the Panarctic Ice and Ocean Modelling and Assimilation system (PIOMAS) to generate a sea-ice reconstruction from 1901 to 2010 (PIOMAS-20C). PIOMAS-20C is generated by forcing PIOMAS with atmospheric reanalysis data from the ERA-20C project. We present results that include validation of atmospheric forcing parameters over sea ice from the ERA20C project and sea-ice thickness from PIOMAS-20C. The PIOMAS-20C sea-ice thickness is generally in good agreement with available observations before and after 1979. We specifically investigate patterns of sea-ice thickness and volume variability in the early 20th century and compare them with changes over the more recent period.


Response of biological communities to a seasonal freshwater gradient in southwestern Hudson Bay, Canada

Laura Dalman, Lisa Matthes, David Barber, Zou Zou Kuzyk, Jean-Eric Tremblay, Janghan Lee, C.J. Mundy

Corresponding author: Laura Dalman

Corresponding author e-mail: laura.dalman@umanitoba.ca

Riverine input to the marine system can modify the surrounding hydrography and chemistry, and subsequently shape the marine ecosystems in coastal regions. Freshwater can have indirect effects on biological communities by influencing sea=ice thermodynamic and dynamic processes, nutrient transport, turbidity and direct effects by osmotic and physiological impacts. Increased discharge from regulated rivers in winter arrives in Hudson Bay during the annual ice-algal spring bloom and is expected to have a direct effect on its production. In this study, we investigate the role of regulated rivers on the bottom ice-algal communities and phytoplankton production located along two spatial gradients from the estuary to marine system in southwestern Hudson Bay, Canada, during the winter–spring transition in 2017 and spring–summer in 2018, respectively. The influence of a salinity gradient on ice-algal biomass and phytoplankton production was examined using water-column structure, nutrient concentrations, chlorophyll-a concentration, production, and species composition within the sea-ice and water column. Preliminary results show that salinity significantly influenced ice-algal biomass where the horizontal distribution of ice algae was positively associated with the salinity of the underlying water column. The gradient in surface-water salinity likely influenced the structure of the ice and thus suitability of habitat for bottom communities in addition to osmotic and physiological effects. Phytoplankton production followed a similar pattern to ice-algal biomass, increasing productivity with increasing salinity along the transect away from the estuary. These observations complement previous findings which documented ice-algal biomass and production increasing along a horizontal gradient with increasing salinity in southeastern Hudson Bay.


Arctic sea-ice floe-length distributions from CryoSat-2, Operation IceBridge and ICESat-2

Rachel Tilling, Nathan Kurtz, Alek Petty, Andy Ridout, Andrew Shepherd

Corresponding author: Rachel Tilling

Corresponding author e-mail: r.tilling@leeds.ac.uk

The size of sea-ice floes impacts the vulnerability of the ice pack to dynamic and thermodynamic forcing. Whilst the Arctic sea-ice floe-size distribution has been characterized in a number of modelling studies, no basin-scale observational dataset currently exists. Sea-ice floe lengths can be estimated from the ESA CryoSat-2 radar and NASA ICESat-2 laser altimeter satellites as they discriminate between the ice and ocean surface along each orbit. However, the discrimination of discrete surfaces in sea-ice-covered regions is limited by the satellite resolution, which is ~250 m along-track for CryoSat-2 and ~20–100 m for ICESat-2. In this presentation we show an initial comparison of Arctic sea-ice floe-length distributions from CryoSat-2 and ICESat-2 over the 2018/2019 growth season. Particular focus will be on the marginal ice zone, where ICESat-2 provides unprecedented satellite resolution over these areas of high lead density and small ice floes. Our floe-length distributions are compared with spring airborne data from NASA’s Operation IceBridge mission. In cases where sea-ice floes are unresolved by satellite altimeters, ice-thickness estimates are likely to be biased high. Therefore, understanding the nature of geometric sampling differences between airborne and satellite missions will help to reconcile their ice-thickness estimates.


Remote sensing of the sea-ice floe-size distribution using satellite altimetry

Christopher Horvat, Lettie Roach, Rachel Tilling, Cecilia Bitz, Baylor Fox-Kemper, Andy Ridout, Andy Shepherd

Corresponding author: Christopher Horvat

Corresponding author e-mail: christopher_horvat@brown.edu

Sea ice is a heterogeneous material of constituent pieces known as floes. These floes may be identified with their horizontal extent, or ‘size’, which may span orders of magnitude. The variability of floe sizes, and the statistical floe-size distribution (FSD) in sea-ice-covered areas, is an important factor in many processes affecting the entire coupled system. To date, there remain very few observations of the FSD, and no knowledge of the seasonal and decadal evolution of floe size regionally and at the climate scale. We exploit a new mathematical technique to infer moments of the sea-ice FSD from altimetric data, and apply this methodology to the CRYOSAT-2 record, covering the period from 2010–18. We show the first Arctic seasonal cycle and climatology of sea-ice floe-size statistics, and discuss geographic and temporal variability of mean floe size and fragmentation. We also perform a rigorous test of the power law hypothesis across time and geographic scales, and find limited support in all cases except for in limited regions above about 10 km in size.


Distinguishing sea-ice types in the Antarctic using microwave satellite observations

Christian Melsheimer, Gunnar Spreen, Yufang Ye, Mohammed Shokr, Stefanie Arndt, Stefan Kern

Corresponding author: Christian Melsheimer

Corresponding author e-mail: melsheimer@uni-bremen.de

Sea ice is classified into three major types, namely, young ice (YI; newly formed and thin), first-year ice (FYI; formed during one freezing season) and multiyear ice (MYI; having survived at least one melt season). As the physical properties of sea ice differ significantly for the different ice types, knowledge of the sea-ice type is essential for properly modeling the ice–ocean–atmosphere system. Here we apply a new satellite-based retrieval of sea-ice types in the Antarctic which has originally been developed for the Arctic, where it can distinguish YI, FYI and MYI during the freezing season. Applying this retrieval in the Antarctic is useful for a number of reasons. The MYI in the Antarctic is less in extension, limited in geographic location and most importantly has different physical and radiometric properties from that in the Arctic. The spatial and annual distribution of this ice type has not yet been investigated much in the Antarctic. Moreover, as the environmental conditions in the Antarctic differ from those in the Arctic, there are sea-ice types that are much more important in the Antarctic than in the Arctic. For example, rougher FYI, more polynyas with thin ice and remarkably larger areas of fast ice are particular features in the Antarctic. The retrieval uses input data from radar scatterometer and microwave radiometers and in addition corrects for the effects of melt–refreeze cycles during the transition seasons, snow metamorphosis and sea-ice drift. In order to distinguish the ice types, the algorithm needs information on their typical emission and backscatter behavior in the channels used. Therefore it can in principle be trained to distinguish ice types other than the main types YI, FYI and MYI. This flexibility is useful because of the possible importance of other ice types in the Antarctic, which is mentioned above. We will present and discuss results of the new retrieval applied to Antarctic sea ice for recent years (2012 to date). However, the needed satellite data have been available since 1999 with daily coverage; spatial resolution is about 25 km. Establishing a time series of Antarctic MYI from 1999 until now is therefore a possible long-term goal of this study.


Impact of ice covers on diel vertical migration of zooplankton in Arctic marine environments

Vladislav Petrusevich, Igor Dmitrenko, Sergei Kirillov, David G. Barber, Jens K. Ehn

Corresponding author: Vladislav Petrusevich

Corresponding author e-mail: vlad.petrusevich@umanitoba.ca

Here, we discuss the impact of ice cover on diel vertical migration (DVM) of zooplankton at three locations: Young Sound fjord in northeast Greenland, northeast of Churchill in Hudson Bay and the southeastern Beaufort Sea in the Canadian Arctic. At all three locations, we deployed ice-tethered or bottom-anchored moorings equipped with acoustic Doppler current profilers (ADCP), and conductivity and temperature sensors. At three locations the backscatter-intensity and vertical-velocity time series from the mooring ADCPs revealed a typical pattern for zooplankton DVM, even under sea ice during winter. Using existing models for solar and lunar illuminance, and the transmission of this light through the sea-ice and snow covers, we estimated under ice illuminance and compared it with the known light sensitivity of Arctic zooplankton. From the acquired data we observed the interaction of vertical migration with lunar light, tides, water and sea-ice dynamics. In all three locations, we observed DVM modification or completely disruption during highly energetic current and spring tide events. In Young Sound, our modelled analysis suggests that the zooplankton in question have an outstanding sensitivity to low illuminance levels of lunar light attenuated by sea ice and snow cover.


On the entrainment, transport and biogeochemical processes associated with sediment-laden, desalinated sea ice in southwest Hudson Bay

David Barber, Madison Harasyn, David Babb, Greg McCullough, Sergei Kirillov, Laura Dalman, David Capelle, Tim Papakyriakou, Soeren Rysgaard

Corresponding author: David Babb

Corresponding author e-mail: David.Babb@umanitoba.ca

Hudson Bay is a large, shallow inland sea that is seasonally covered by a dynamic first-year sea-ice cover. During a research cruise onboard the research icebreaker CCGS Amundsen in June 2018 vast areas of thick, sediment-laden sea ice were observed in southwestern Hudson Bay. Navigation, by the Amundsen, through this ice was surprisingly difficult due to its thickness and low salinity. While sediment-laden sea ice has previously been observed in James Bay and Foxe Basin the vast presence and thickness of the sediment-laden ice was surprising. Within this paper we present in-situ observations of thickness, surface topography, sediment provenance and characteristics, oxygen isotopes, salinity, temperature and microwave properties of this unusual form of sea ice. We provide results from remote-sensing data which illustrate the annual and interannual presence and spatial extent of the sediment-laden desalinated sea ice relative to more traditional first-year sea ice in Hudson Bay. We explore the relationship between hydroelectric regulation, tidal forces and winter to spring transitions in freezing degree days as possible mechanisms controlling the formation of this type of sea ice. We conclude with an overview of the physical, biological and geochemical significance of this unusual form of sea ice in southwest Hudson Bay.


Arctic sea-ice conditions in the early melt season from diurnal variability in satellite passive-microwave swath brightness temperatures

Angela Bliss

Corresponding author: Angela Bliss

Corresponding author e-mail: angela.bliss@oregonstate.edu

Daily satellite passive-microwave brightness temperature (Tb) observations have long been used to identify the seasonal changes in Arctic sea ice such as ice extent, type, and the dates of melt and freeze onset. Following the initiation of melting, rapid fluctuations in radiance are attributed to changes in the scattering properties of snow as freeze/thaw cycling and snow metamorphosis occur. As the melt season progresses, melt-pond formation and drainage further affect spatial and temporal variability of Tbs. In this work, a period of enhanced diurnal Tb variability during the early melt season is identified and used to assess the evolution of sea-ice surface conditions following melt onset. Daily time series of temporal variability are computed along buoy tracks and at static grid cell locations from a level-2 SSM/I and SSMIS Tb climate data record. A regime-change detection method is then used to define the melt-onset date and the duration of the early spring high-variability period. Within this period, sea-ice conditions from ancillary satellite data products, such as ice concentration, skin temperature and melt-pond fraction, and in-situ buoy observations are used to interpret the day-to-day changes in diurnal Tb variability. This analysis utilizes the sub-daily temporal resolution of satellite passive-microwave observations to investigate the melting conditions of Arctic sea ice with an eye towards expanding and improving retrievals of regional sea-ice melt-condition status on a daily basis.


An accelerated implicit solver for sea-ice dynamics

Philippe Blain, Jean-François Lemieux, Abdessamad Qaddouri, Frédéric Dupont

Corresponding author: Philippe Blain

Corresponding author e-mail: philippe.blain@canada.ca

Sea-ice models account for the movement of the ice by solving the sea-ice momentum equation, which relates the ice velocity to different terms such as the ocean and atmosphere stresses and the ice-pack internal stresses. Despite the fact that new sea-ice rheologies were recently proposed, most sea-ice models are still based on the Hibler viscous–plastic (VP) rheology or the elastic–viscous–plastic (EVP) framework. The latter allows the sea-ice momentum equation to be solved explicitly by using a subcycling iteration at each time step. Previous attempts at solving the momentum equation implicitly with a VP formulation have concentrated on two methods: Picard iteration and Newton-type methods. Picard iteration is a simple fixed-point iteration method in which the equations are linearized using a previous estimate, resulting in a linear system. Due the highly nonlinear character of the VP rheology, this approach requires a very high, often prohibitive number of iterations in order to reach a suitable solution. Newton-type methods usually offer far better convergence properties, but they require the formation of the Jacobian matrix, which can be complicated to formulate and extremely costly to compute and store. In this regard, the Jacobian-free Newton–Krylov method, which completely eliminates the need to form this matrix, is very attractive, but it suffers from robustness issues at high resolution. We present a new, implicit VP solver based on an accelerated fixed-point iteration method. The Anderson acceleration algorithm uses the residuals of the previous iterates to find a new, optimal iterate by a minimization procedure. The solver, implemented in the CICE sea-ice model, shows great promises, achieving a 12-fold decrease in the number of iterations needed to reach a given precision when compared to a Picard solver. Its performance in terms of robustness and computational efficiency will also be discussed.


Sea-surface dimethylsulfide hotspots linked to sea-ice dynamics and solar radiation in a fine-scale study of the Canadian Arctic Archipelago

Joanie St-Onge, Martine Lizotte, Guillaume Massé, Maurice Levasseur, Jean-Éric Tremblay, Michel Gosselin

Corresponding author: Martine Lizotte

Corresponding author e-mail: Martine.Lizotte@qo.ulaval.ca

The sources and strength of oceanic emissions of dimethylsulfide (DMS), a climate-active biogenic gas, could be modified in the Arctic due to reductions in snow cover, sea-ice extent and thickness. Understanding the impacts of climate change on DMS dynamics is crucial since DMS-derived sulfate is thought to be the main precursor of secondary marine aerosols that lead to cloud formation and therefore contribute to moderate solar-energy input in the Arctic. Using a novel automated instrument (ACT-MIMS), DMS samples were collected at high frequency in the surface waters of the Canadian Arctic Archipelago during the summer of 2017 (July–August) and 2018 (July) aboard the Canadian Coast Guard Ship Amundsen. More than 3500 DMS measurements were collected alongside ancillary measurements of sea-surface salinity, sea-surface temperature, fluorescence (chlorophyll a proxy), photosynthetically active radiation, sea-ice concentration and the algal precursor of DMS, dimethylsulfoniopropionate (DMSP). DMS concentrations ranged from ~0.2 to 43.0 nmol L–1 (average of 8.1 nmol L–1) in 2017 and from ~0.8 to 55.0 nmol L–1 (average of 13.7 nmol L–1) in 2018 over an area covering a wide range of contrasting marine environments from coastal to open-ocean ice-free waters, as well as under-ice waters. Surface-water DMS hotspots were measured at the ice edge and in marginal ice zones, as well as in ponded first-year ice areas. This suggests the synthesis of DMSP by sea-ice algae in response to environmental stressors such as solar radiation or large variations of salinity or sea-surface temperature due to ice melt. Furthermore, nighttime increases and daytime decreases in DMS concentrations were observed in the northern Labrador Sea and Davis Strait. The relationship between DMS concentrations and diurnal solar radiation variations suggests the involvement of photobiological processes. Overall, our results strengthen the view that the cycle of marine DMS in the Arctic is closely related to sea-ice dynamics and physiological responses to light. As such, future changes in the seasonality of the Arctic cryosphere are likely to play an important role in shaping DMS emissions, although the signs and magnitude of this change remain highly uncertain.


The influence of sea-ice conditions on crude-oil spill behavior

Diana Saltymakova, Durell Desmond, Thomas Neusitzer, Nariman Firoozy, Katarzyna Polcwiartek, Nolan Snyder, David Barber, Gary Stern

Corresponding author: Diana Saltymakova

Corresponding author e-mail: Diana.Saltymakova@umanitoba.ca

The development of the Canadian Arctic is closely tied to the petroleum industry. Increasing volumes of oil – both as a source of fuel and commercial cargo – pass through the region, thus significantly elevating the risk of uncontrolled oil spills. Spill detection and remediation need to account for the harsh environmental conditions, the presence of sea ice, and the remoteness of oil release sites. The presence of sea ice brings additional uncertainties, as sea-ce structure depends on the temperature as well as the conditions during ice formation. Three oil-spill scenarios were tested at the University of Manitoba Sea-ice Environmental Research Facility. Mesocosm experiments were conducted in 2016, 2017 and 2018, from January to March each year. The experiments were carried out in a 7 m3 tank filled with artificial seawater, prepared from groundwater mixed with sea salts, to a salinity of 32 ppt. Twenty liters of light crude oil provided by Tundra Oil & Gas Partnership was used in each experiment. A thermocouple string was used to measure the ice-temperature profile at intervals of 2.5 cm. Scenarios 1 and 2 demonstrate a crude oil spill under sea ice. Once the ice thickness reached 6 cm crude oil was injected from the tank bottom. Scenario 3 represents a crude oil spill in open water followed by ice formation. In all the experiments sea ice continued to grow up to 20–28 cm for 5–10 days when it was sampled for chemical analysis of oil and sea-ice morphology. Scenario 1 simulates crude oil migration through fractured ice. The experiment demonstrated that the majority of compound losses were associated with evaporation due to the fast penetration of crude oil to the sea ice surface. Scenario 2 simulates a crude oil spill under solid ice, which illustrates crude-oil migration through brine channels. This scenario was additionally tested on an initial ice thickness of 20 cm. Bulk crude-oil migration and distribution of its constituents within sea ice is a function of oil concentration, in which its lower contents produced higher contact between the oil and brine, allowing for more effective dissolution. In scenario 3 crude oil became encapsulated 2 cm below the surface in a pancake-ice formation and behaved similarly to scenario 2. Each of these scenarios demonstrates the ability of crude oil to partition within sea ice, snow, the water column and the atmosphere, resulting in changes to its composition and sea-ice properties (i.e. temperature, salinity, dielectrics).


Dive behaviour of Beaufort Sea beluga whales (Delphinapterus leucas) in relation to oceanographic covariates and prey distribution

Luke Storrie, Lisa Loseto, Shannon MacPhee, John Iacozza, Nigel Hussey, Greg O’Corry-Crowe, David Barber

Corresponding author: Luke Storrie

Corresponding author e-mail: storriel@myumanitoba.ca

The Beaufort Sea is thought to be an important summer foraging ground for one of the largest populations of beluga whales (Delphinapterus leucas). Recent effects of climate change, including alterations to the date of sea-ice freeze-up, total sea-ice extent, and a shift in the prey assemblage, may be impacting the movements and dive behaviour of this population. Satellite-linked transmitters were used to monitor the horizontal movements of this population during the 1990s and mid 2000s; however, there is little high-resolution information on dive behaviour available despite this being an important factor to determine the energetic consequences of diving for beluga whales. Recent improvements in the technology of satellite-linked transmitters now enable fine-scale questions about habitat use, diving and physiology to be answered. In July 2018, satellite tags were deployed on 10 beluga whales in the Mackenzie Estuary, a traditional summering area for Beaufort belugas. Tags were programmed to collect time-series data on dive depth and water temperature every 75 s, and location (fastloc GPS) during surfacing events. This high-resolution data enables in-depth analysis of dive profiles, including time spent at each depth layer, and on the surface, and dive shape in relation to geographic location, bathymetry, temperature and sea-ice cover. Initial analyses indicates that depths between 400 and 600 m were targeted most frequently once whales left the shallow estuary. Whales were diving almost exclusively to the seafloor in shelf and slope habitats, and to the mid-water column over the deep Arctic basin. Viscount Melville Sound, part of the Northwest Passage, appeared to be an important feeding ground for several of the whales between late July and mid-August. These results will be evaluated in context with results from a summer fish survey to identify potential foraging activities on key species. The information from this study will be important in contributing to our understanding of the ecological significance of the various habitats used by beluga in summer through fall. The data provided here will also be used in habitat-suitability and bioenergetic modelling of Beaufort Sea beluga, which will allow prediction of their response to climate change.


The regional evaluation of heat uptake and release in the Arctic and impacts on sea-ice concentration

Meghan Helmberger, Mark Serreze

Corresponding author: Meghan Helmberger

Corresponding author e-mail: meghan.helmberger@colorado.edu

As the Arctic Ocean loses its sea-ice cover, there is a larger oceanic heat gain from the net surface flux throughout the spring and summer; meaning that there is more energy to transfer from the ocean to the atmosphere and outer space in the autumn and winter. Recent work has shown that the increased oceanic heat content at the end of summer delays autumn ice growth, with implications for marine shipping and other economic activities. Depending on patterns of seasonal sea-ice retreat and weather conditions, the spring–summer heat uptake and autumn–winter heat loss can be highly variable from year to year and regionally. Here, we examine how the seasonality in upper-ocean heat uptake and release has evolved over the period 1979–2018 using three different atmospheric reanalyses (MERRA-2, ERA-5, CFSR) and the relationships between this seasonal change and the evolution of sea-ice cover. We determine which regions have seen the largest increases in total seasonal heat uptake and how variable this uptake can be. What changes have been observed in the rates of seasonal heat uptake and release? Do multiple reanalyses agree with each other with respect to patterns and magnitudes of heat gain and loss? Are there different relationships between surface fluxes and sea-ice concentration depending on which reanalysis is used? Do these relationships change when evaluated regionally? In future work, downward shortwave flux data collected across the Chukchi Sea in August 2018 will be compared to flux estimates from these reanalyses.


The Greenland Ice Sheet’s influence on the Arctic and sub-Arctic North Atlantic Ocean

Laura Gillard, Helen Johnson, Juliana Marson, Paul Myers

Corresponding author: Laura Gillard

Corresponding author e-mail: gillard2@ualberta.ca

The fresh-water budget of the Arctic and sub-Arctic North Atlantic has been changing in recent decades. The altered freshwater budget may impact sea ice, the surrounding oceans, and therefore large-scale ocean circulation. The Greenland Ice Sheet (GrIS), the largest storage of fresh water in the Northern Hemisphere, has been assessed regarding its contribution to these changes in the fresh-water budget. Marine-terminating glaciers are one of the most influential components of the GrIS for releasing fresh water into the ocean. A common approach to the present generation of ocean models is to inject Greenland fresh water at the surface. How does changing the vertical distribution of glacial meltwater impact sea ice and the surrounding oceans? To assess the impact of the GrIS’s fresh-water fluxes we use a regional eddy-permitting coupled ocean–sea-ice general circulation model. We set up a suite of experiments in a 0.25° Arctic and Northern Hemisphere Atlantic configuration of NEMO v3.6, forced with realistic estimates of Greenland’s meltwater and icebergs. We compare two different forcing sets of the GrIS fresh water, and consider different distributions of the fresh water in the vertical. This study will assess the importance of both liquid and solid fresh-water discharge for the large-scale ocean circulation, impact on sea ice, and the renewal of warm water towards the GrIS.


Implementation of an updated snow cover and adjustment for snow salinity in sea-ice-thickness retrievals from CryoSat-2 ice freeboard

Hoi Ming Lam, Torsten Geldsetzer, Vishnu Nandan, Stephen Howell, John Yackel

Corresponding author: Hoi Ming Lam

Corresponding author e-mail: hoiming.lam@ucalgary.ca

Many established ice-thickness retrievals from CryoSat-2 freeboard data assume that the radar-scattering horizon is at the snow–ice interface regardless of the properties of the snow cover. In these retrievals, a modified climatological snow depth from Warren et al. (1999) is often used, which may not characterize the present icescapes in the Arctic. This study introduces two modifications to these assumptions. We assess both multiyear sea ice (MYI) and first-year sea ice (FYI) within the Canadian Arctic Archipelago. First, we compare monthly-averaged modelled snow depth from the Canadian Regional Ice Ocean Prediction System (RIOPS) to the Warren et al. snow climatology, to illustrate the differences between the climatological data and the present climate scenarios. We then substitute the Warren snow climatology with RIOPS snow depths in the conversion of CryoSat-2 ice freeboard to sea0ice thickness, to investigate the outcome of using a more recent and spatially representative snow cover. Second, we implement a radar-scattering horizon-adjustment factor for snow salinity on FYI in CryoSat-2 ice-freeboard retrievals. We test the adjustment factor on both the climatology-based and the RIOPS-based FYI thickness retrievals. The results are validated with field measurements available from CryoVEx aerial and ground campaigns and Operation IceBridge flights. Preliminary results indicate that by applying the scattering horizon adjustment for salinity, the resulting FYI thickness resembles the validation datasets more than by using snow-cover substitution. The former method decreases the retrieved ice freeboard and effectively reduces a positive bias in CryoSat-2 FYI thickness calculations. Our analysis suggests that the role of snow properties in ice-thickness retrievals requires further study, and that snow salinity measurements should be included in field validation measurements.


From the land to the table: learning about Aklavik’s declining beluga whale harvest through community-driven research

Elizabeth Worden, Tristan Pearce, Jill Oakes, Lisa Loseto

Corresponding author: Elizabeth Worden

Corresponding author e-mail: elizabeth.worden@umanitoba.ca

As species harvested on the land feed family members at the dinner table in Aklavik, Inuvialuit observations of the land fuel decision-making tables of regional co-management boards. When pronounced societal and environmental changes are culminating to seriously impact Inuvialuit interaction with the land and its species, it is imperative to structure research initiatives in response to community concerns. The aim of this research is to improve understanding of human–beluga-whale relations over time and the implications of change for the beluga-whale harvest in Aklavik. The community’s beluga-whale harvest has experienced a rapid decline in the last four decades, with 35 whales harvested in 1980 compared to an average of a couple per year since 2010. Concerns from local representatives were expressed at the 2016 Beluga Summit in Inuvik, and this project was proposed to address their priorities. Three research objectives were developed in collaboration with the Aklavik Hunters and Trappers Committee (AHTC) and through local input. These objectives aimed to (1) document peoples’ memories of Aklavik’s beluga harvest in its prime; (2) examine local understanding of social and ecological changes affecting beluga hunting; and (3) assess the implication of these changes for the future of the beluga hunt. The project is guided by community-based participatory research. An early visit in March 2017 allowed for research expectations to be discussed with the AHTC. From June to August 2017, data was collected with the help of two local research assistants in Aklavik. Research methods included: semi-structured/open-ended interviews (n = 32), experiential learning, and verification of results in summer 2018. Results of this research demonstrate that several Inuvialuit from Aklavik can hunt and prepare beluga whale due to intergenerational knowledge transmission. However, a continuous spectrum of social and environmental change is reducing opportunities for success in the beluga harvest. Climate change is limiting access to preferred whaling camps and social dynamics at the current coastal camp are not conducive to the whale harvest. Shifting Inuvialuit values are resulting from the passing of elders and the ever-increasing influence of southern culture. Despite these changes, the level of interest in the beluga hunt is high. It is hoped that this research will help AHTC identify opportunities to revive the hunt.


Estimating early-winter Antarctic sea-ice thickness from deformed surface morphology

M. Jeffrey Mei, Ted Maksym

Corresponding author: M. Jeffrey Mei

Corresponding author e-mail: mjmei@mit.edu

Accurate estimates of sea-ice thickness in the Antarctic are challenging, particularly over smaller scales, as estimates or measurements of snow depth and snow and ice density are needed. Current linear models of the relationship between freeboard and ice thickness produce errors in sea-ice thickness of up to 50% at the floe scale. This can be significantly reduced by averaging over large scales, but these relationships may not be regionally or seasonally consistent, and produce a significant bias against thin ice. Here, we examine the potential of higher-order surface morphological information to reduce the error in local sea-ice thickness estimates by analyzing high-resolution, co-located three-dimensional surveys of surface elevation, snow depth, and ice thickness obtained in the Ross Sea in May–June 2017. Using a convolutional neural network with laser altimetry profiles of sea-ice surfaces it is possible to estimate sea-ice thickness with lower error (~15%) than current methods, and without bias for any thickness classes. Moreover, the neural network does not require snow depth to be inferred. The learned filters appear to correspond to basic morphological features such as pressure ridges and snow dunes. This method may be extended to lower-resolution, larger-footprint data such as IceBridge and ICESat2, which may help create low-error and high-resolution estimates of sea-ice thickness.


Sea-ice properties from enhanced-resolution passive-microwave data

Walter N. Meier, J. Scott Stewart

Corresponding author: Walter N. Meier

Corresponding author e-mail: walt@nsidc.org

Passive-microwave sensors have been one of the most valuable sources for sea-ice information, including concentration, extent, melt, age and drift. They provide a continuous, consistent and near-complete record of sea-ice characteristics that is now over 40 years in length. These records are some of the longest satellite-derived climate records and the significant decline in Arctic sea-ice extent over that period is one of the most iconic indicators of climate change. A significant limitation of passive-microwave remote sensing is the low spatial resolution of the sensors. For much of the record, the gridded resolution is limited to 25 km, but sensor footprint resolution is as low as ~45 × 70 km. This means that small-scale features such as leads and polynyas are not observable. A new passive-microwave brightness-temperature product is now available that uses resolution-enhancement techniques to obtain gridded resolutions of roughly 3–6 km. The product uses a reconstruction method that goes beyond a simple interpolation to use multiple observations to effectively reconstruct the microwave emission signal at the size of the enhanced-resolution grid cell. This allows the product to obtain much finer-scale detail than has been previously been available from passive-microwave instruments. Also, while previous gridded products yield a simple daily average field or twice-daily composites of ascending and descending satellite passes, the new product grids products into twice-daily fields based on local time of day, with evening and morning fields. This allows better investigation of diurnal effects. Here, we apply the enhanced-resolution brightness-temperature data to derive new estimates of sea-ice concentration and motion. Results show the potential to improve the precision of the ice-edge location, show more spatial detail within the ice pack, and yield more accurate ice motions. Polynyas and large leads are more accurately captured and there is the potential to improve estimates of ice-formation rates, salinity fluxes, and sensible-heat transfer estimates. In addition, the morning and evening composites yield interesting information on diurnal changes in the ice cover, particularly near the equinoxes. The results point to the potential for an improved long-term passive-microwave record of key sea-ice climate indicators.


Sensitivity of L-band signature to detect sea-ice-melt onset in the Arctic

Mallik Mahmud, Stephen Howell, John Yackel

Corresponding author: Mallik Mahmud

Corresponding author e-mail: msmahmud@ucalgary.ca

High-resolution synthetic aperture radar (SAR) has proved its utility in sea-ice melt-onset detection in the Arctic. Although C-band SAR has been exclusively used for operational sea-ice monitoring, L-band SAR has demonstrated improved ice-type separability due to its larger penetration depth owing to the longer wavelength compared to C-band SAR. To that end, L-band SAR is considered an optimal choice for sea-ice monitoring but its utility beyond ice-type classification has received less attention. Here, we investigate dense time-series L-band SAR signatures from ALOS PALSAR and explore their sensitivity to detect sea-ice-melt onset (MO) for both FYI and MYI. The mean MO date was year-day (YD) 162 (for FYI) and YD 164 (for MYI). L-band-detected melt-onset dates were 3 days and 2 days later compared to Ku-band scatterometer (QuikSCAT) and C-band SAR (RADARSAT-2), respectively. The snow thickness on sea ice and the progression of air temperature dictate the timing of MO detection from all sensors. However, data availability of L-band SAR imagery during winter to melt transition can also have an effect on the MO detection variation as daily-resolution L-band SAR imagery was not available during winter to spring transition. We also investigate the sensitivity of the 3 dB threshold up to  ± 1 dB with a 0.5 dB increment. Our results suggest that only a 2 dB threshold resulted in the earlier melt (e.g. YD 159 and YD 162 for FYI and MYI, respectively). Comparing the result with Mahmud et al. (2016), we conclude that the L-band melt-onset detection threshold is less sensitive compared to the C-band SAR threshold. The distinct melt signature at L-band minimizes false alarm events for early melt detection, which provides more accuracy to the melt detection algorithm. This melt algorithm will be invaluable for current and upcoming L-band SAR missions.


Systematic analysis of Arctic cyclone impacts on sea-ice concentration from 1979 to present

Erika Schreiber, Mark Serreze

Corresponding author: Erika Schreiber

Corresponding author e-mail: erika.schreiber@colorado.edu

Sea ice in the Arctic is subjected to dynamic and thermodynamic forces when cyclones pass. The cyclonic winds promote ice divergence and have been associated with greater ice extent, but strong cyclones preceded the record low sea-ice extents in 2012 and 2016, calling this relationship into question. With the ice cover declining in extent, concentration and thickness, the response of the sea ice to cyclones may be changing. Lower ice concentrations in the marginal ice zone, particularly in summer, allow for freer drift, while ice is subject to greater internal forces at higher concentrations. Lesser extent and lower concentrations also permit greater energy absorption by the ocean surface, increasing potential for melt. Thinner sea ice is more susceptible to melting out as well, and has weaker internal forces to inhibit the divergent pressure of cyclonic winds. We examine cyclone-induced concentration changes across the Arctic from 1979 onward using passive microwave data from SMMR/SSMI and AMSR-2. Cyclone areas are defined using a tracking algorithm applied to the ERA-Interim atmospheric reanalysis, and dynamic response is determined using sea-ice motion data. We see a dominant pattern of increased sea-ice concentration following cyclone passage in all seasons, following historic expectations, with the strongest impacts in the marginal ice zone.


A decade of late-winter Arctic sea-ice surface-roughness observations

Kyle Duncan, Sinéad Farrell, Laurence Connor, Jennifer Hutchings, Jacqueline Richter-Menge, Roseanne Dominguez

Corresponding author: Kyle Duncan

Corresponding author e-mail: Kyle.Duncan@noaa.gov

The surface roughness of the Arctic sea-ice cover can provide numerous insights into the processes experienced since ice formation. One of the main indicators of sea-ice dynamic processes that is captured within the ice pack is the formation of pressure ridges. Pressure ridges are a dominant feature of the sea-ice cover, impacting the mass, energy and momentum transfer budgets for the Arctic Ocean, and presenting an impediment to travel across the ice surface. Obtaining observations of pressure-ridge sail heights at high spatial and temporal resolutions is necessary for improving the representation of sea-ice dynamics in high-resolution numerical sea-ice models. With the advent of dedicated airborne surveys of the Arctic Ocean, such as NASA’s Operation IceBridge (OIB) mission, such high-resolution observations have become available. Here we use high-resolution digital mapping system (DMS) imagery, collected during NASA’s OIB airborne missions, to derive ridge sail heights and assess results across regional scales. Using derived sail heights from ~64 000 DMS images, we present the spatial characteristics of ridge-sail heights for the Beaufort/Chukchi Seas region and compare these with results for the central Arctic Ocean and Canada Basin. We examine the interannual variability in the sail-height distributions for the period 2010–18, placing the results in the context of the parent ice type. We show that the maximum sail height for new ridges formed in first-year ice (FYI) has a mean of 2.4 m and a mode of 1.8 m, whereas in multiyear ice (MYI) the mean is 3.1 m and the mode is 2.7 m. The Airborne Topographic Mapper (ATM) provides sea-ice surface roughness from airborne laser-altimeter elevations across all OIB flights. While maximum sail heights are not consistently determined, the ATM data does measure a wide range of surface features ranging from level ice and sastrugi to hummocks and ridge sails. We use the standard deviation of ATM elevations over these surface features to characterize the surface roughness, both along and across-track. We examine the interannual variability in surface roughness over the 10-year period 2009–18 in the context of the parent ice type. We find evidence that the surface roughness at the end of winter was highest in 2015 and lowest in 2018, due to changes in the MYI pack. Our results from airborne sensors will be extended to the basin scale in the future using high-resolution ICESat-2 elevation data.


Bridging temporal and spatial scales of sea-ice deformation: lessons from N-ICE2015

Polona Itkin, Jari Haapala, Anton Korosov, Annu Oikkonen, Gunnar Spreen

Corresponding author: Polona Itkin

Corresponding author e-mail: polona.itkin@gmail.com

In the changing Arctic where the sea ice is getting thinner and at the same time moving faster, sea-ice deformation processes are also increasing. The magnitude of deformation rates are, however, difficult to compare between different sources as they depend on the spatial and temporal scales of the measurements. Here we present three sea-ice deformation datasets collected during N-ICE2015, an expedition to the ice-covered region north of Svalbard in the first half of 2015. The deformation rates calculated from the autonomous buoy drift, ship radar images and sea-ice drift from satellite remote-sensing (Sentinel-1 Synthetic Aperture Radar (SAR)) images span across the temporal scales from 10 min to 3 d and across the spatial scales from 200 m to 100 km. This wide range of scales allows us to test the scaling laws of deformation and explore the detection limits of the three methods. Our results show that the deformation rates measured by all methods coincide if compared for the same time window and spatial extent. Likewise, we show that the SAR-derived deformation rates can be used to detect deformation at very short spatial and temporal scales.


Impact of remnant brine channels on percolation blockage and melt-pond drainage through sea ice

Andrew Wells, James Parkinson, Daniel Martin

Corresponding author: Andrew Wells

Corresponding author e-mail: andrew.wells@physics.ox.ac.uk

Melt ponds form on the surface of Arctic summer sea ice from the accumulation of melted snow and ice. They provide a major feedback on the energy budget and climatological evolution of the sea-ice cover. Drainage of meltwater through porous sea ice is driven by the hydrostatic head of a pond surface above sea level, and provides a key control on the depth and areal coverage of melt ponds, and their resulting albedo. Recent field experiments have suggested an important role for so-called percolation blockage in reducing sea-ice permeability, and allowing the formation of melt ponds early in the melt season on first-year ice, in conditions that would otherwise favour complete pond drainage. The ice is initially warm and relatively permeable at the onset of surface melt. This allows rapid initial drainage and flushing of salt from the ice-pore space, with subsequent internal solidification to maintain phase equilibrium. The corresponding reduction in porosity and permeability caps the drainage flow, and allows a surface melt pond to develop. We use enthalpy-method simulations to study the physical controls on percolation blockage in porous sea ice, exploiting a newly developed code for simulating coupled fluid flow and phase change in mushy layers using the Chombo framework. To build dynamical insight, we consider flow and solidification in a thin 2-D Hele–Shaw cell. By cooling salt water from above, we initially simulate growth of a layer of ice with fully resolved brine channels. A surface melt pond is then added, with flow driven down through the underlying ice by an imposed hydrostatic pressure head. Our simulations confirm the previously hypothesized mechanism for development of a percolation blockage by flushing of the ice. We quantify how the variation of net permeability and the timescale for blockage vary with the salinity, temperature and hydrostatic head applied in the surface melt pond. We compare simulations with remnant brine channels from earlier growth, versus simulations with an initial horizontal averaging of the porosity. Despite having the same horizontally averaged porosity, experiments with remnant brine channels allow significantly larger initial drainage than those with horizontally uniform porosity. This highlights the importance of accurately accounting for inhomogeneous brine channel dynamics in models of permeability and flow through porous sea ice.


Microwave remote sensing of oil in-and-on snow and sea ice in a controlled, purpose-built mesocosm

Mark Christopher Fuller, Dustin Isleifson, John Yackel, David Barber

Corresponding author: Mark Christopher Fuller

Corresponding author e-mail: chris.fuller@umanitoba.ca

Continued warming in the Arctic, resulting in a reduced extent, thickness and seasonality of sea ice, is increasing access for shipping, tourism, and resource extraction, in turn increasing the risk of oil spills and environmental contamination. A system for detection and tracking, in often dark and remote regions, is necessary to minimize spill impacts. Microwave remote sensing, used in the Arctic since 1972, is an all-weather, all-season tool to detect and track differences in the electrothermophysical properties of snow and sea ice and its spatiotemporal variability. For a given thermal state or event, the thermodynamic impact of brine volume fraction and phase change is governed by snow and sea-ice geophysical properties manifested in dielectric permittivity, which allows for delineation based on the microwave normalized radar cross-section (NRCS) response. Sub-ice oil moves up through cracks and brine channels in sea ice, displacing brine, thereby changing the bulk dielectric and thermodynamic properties in space and time. Recent experiments inverted C-band NRCS at a single incidence angle for oil volume and fraction estimates in sea ice made from a synthetic mix of water and salt; however, the 3 m diameter pool constrained parameterizations of the radar, sea ice and oil. Therefore, novel experiments are required to further quantify responses of microwave energy over various frequencies, polarizations and incidence angles, and for a wider variety of controlled oil in-and-on sea ice scenarios. The Churchill Marine Observatory (CMO), a set of purpose-built facilities including the Oil in Sea Ice Mesocosm (OSIM) and the environmental observing system, is designed to investigate the impact of contaminants on the ocean/sea-ice/atmosphere environment. The OSIM is a 60 × 30 × 10 ft (18 × 9 × 3 m) pool divided into two sub-pools (to accommodate both contaminant and control experiments), in which sea ice can be grown using local sea water and weather (via a retractable roof) to create controlled oil in-and-on snow and sea ice experiments. Operational in the winter of 2020, the CMO will facilitate a more detailed quantification of the impacts of oil spills and other contaminants on the dielectrics and thermodynamics of a variety of sea-ice formation, thickness and melt scenarios, employing a suite of polarimetric C- and Ku-band microwave scatterometers, sensing over a full range of incidence angles, for linkage and upscaling to current and future satellite observations.


Airborne measurements of the sea-ice thickness distribution in the southwest Ross Sea

Pat Langhorne, Christian Haas, Wolfgang Rack, Greg Leonard, Natalie Robinson, Gemma Brett, Dan Price, Mike Williams

Corresponding author: Pat Langhorne

Corresponding author e-mail: pat.langhorne@otago.ac.nz

The southwest Ross Sea is a unique region with a highly variable sea-ice cover including thin pack ice formed in polynyas and extremely thick pack ice where it is deformed and exported into the Ross Sea proper. The coast of Victoria Land is fringed by landfast sea ice which interacts with ice shelves and floating ice streams, resulting in the presence of a sub-ice platelet layer (SIPL) as an indicator of supercooled ice shelf meltwater at the ocean surface. Airborne electromagnetic induction (AEM) sounding can remotely characterize these different ice types and is sensitive to the presence and thickness of platelet ice. Here we present results from extensive AEM surveys of pack ice and landfast ice thickness in November 2017, off the coast of Victoria Land. Four surveys with a total profile length of about 4000 km were carried out with a fixed-wing DC3T aircraft. For calibration and context, the AEM surveys were supported by in-situ measurements of snow, sea ice and SIPL thickness, and ocean salinity and temperature at accessible locations on the fast ice of McMurdo Sound. Surveys were aligned with CryoSat-2 satellite altimeter tracks for validation of satellite observations. Results show strong thickness gradients from the McMurdo Sound and Terra Nova Bay polynyas towards the central Ross Sea, with surprisingly thick, deformed sea ice (up to 4 m in places) in a prominent convergence zone where ice exported from both polynyas meets. Fast ice between Terra Nova Bay and the Adare Peninsula was more than 2 m thick and heavily deformed by onshore drift of pack ice. There were indications of a SIPL only very close to some floating ice tongues. In contrast, the fast ice in front of the Hell’s Gate Ice Shelf was level and more than 2 m thick, with a SIPL between 0.8 and 2.5 m thick. The landfast ice in McMurdo Sound was also mostly level and more than 2 m thick. It was underlain by a well-documented SIPL, with thicknesses of more than 6 m near the ice shelf edge and thinning towards the north, away from the ice shelf. The latter results are in good agreement with in-situ measurements. Our results have important implications for understanding and modelling ice-shelf melt and interactions with sea ice, for sea-ice remote sensing, and for studies of the biological productivity of the fast-ice/platelet-ice system and its role for the ecosystem and carbon fluxes in the region.


Response of winter sea-ice velocities to wind forcing in the Canadian Beaufort Sea: large spatial and temporal variability over scales of 20–120 km and several hours

David Fissel, Matthew Asplin, Keath Borg, Alex Slonimer, Humfrey Melling

Corresponding author: David Fissel

Corresponding author e-mail: dfissel@aslenv.com

Continuous measurements of ice velocity, ice draft and near-surface ocean currents were obtained from an extensive array of ten upward looking sonar (ULS) moorings operated from September 2009 through September 2011. The moorings are in the continental margin of the Canadian Beaufort Sea, at water depths ranging from 55 m to 1010 m, separated by distances of up to 120 km. In autumn, ice motion is dominated by its response to episodic wind events, especially those with mean wind speeds exceeding 4 m s–1 with ice-to-wind-speed ratios ranging from 2% to 5%. However, by late winter and early spring, the response factor of ice to wind speed is much reduced from that of the autumn, to average values of 0.7% and 0.4% for easterly and westerly winds, respectively. In most of these moderate to large winter wind events, there are intervals when ice motion nearly ceases (i.e. ice speeds of <0.5 cm s–1) at some or all mooring sites. This highly suppressed ice motion is attributed to high values of internal ice stress during winter. The spatial variability of the ice speeds and response ratios is very large in winter with coefficient of variance values of typically 100–1000% vs 15–50% in autumn. Among the measurement sites, the differences in the timing of a few to several hours of the large reductions and cessation of ice movement is examined among the wind episodes. A different pattern is noted for easterly winds vs westerly wind episodes. Distinct spatial patterns occur in the ice-to-wind-response factors and these differ between the two winters; these spatial patterns are examined in relation to the mesoscale patterns in sea-ice concentrations and stages of development within the mooring area and over the larger spatial scales within the Canadian Beaufort Sea continental margin.


Arctic Ocean surface mixed-layer evolution in the Community Earth Systems Climate Model Large Ensemble

Erica Rosenblum, Robert Fjaber, Julienne Stroeve, David Barber

Corresponding author: Erica Rosenblum

Corresponding author e-mail: erica.j.rosenblum@gmail.com

Arctic sea-ice retreat is a well-known indicator of climate change, yet climate models struggle to accurately simulate this feature. One potential source of this bias could be related to simulated Arctic Ocean surface mixed-layer evolution, which has undergone drastic changes over the past few decades. Mixed-layer properties are closely related to upper-ocean heat storage and are therefore thought to have a direct impact on sea-ice evolution. Here we examine how well Arctic mixed layers are simulated in the Community Earth Systems Climate Model Large Ensemble (CESM-LE) compared to observations from the Monthly Isopycnal and Mixed-Layer Ocean Climatology (MIMOC), Ice-Tethered Profilers (ITPs), and previous studies using hydrographic data. Specifically, we examine simulated and observed seasonal and decadal mixed-layer evolution across each Arctic Ocean basin. A simplified, one-dimensional framework is then used to investigate the cause of differences between observed and modeled summer mixed-layer properties in the Canada Basin. Specifically, we explore how simulated sea-ice cover, summer sea-ice melt, and mixed-layer dynamics might explain differences between observed and modeled summer mixed-layer properties. The results may provide insight regarding what ice–ocean processes are necessary to accurately simulate Arctic mixed-layer processes in climate models.


Probabilistic forecasting of the Arctic sea-ice edge

Hannah Director, Adrian Raftery, Cecilia Bitz

Corresponding author: Hannah Director

Corresponding author e-mail: direch@uw.edu

Reduced sea-ice cover in the Arctic has increased commercial shipping, tourism and search-and-rescue needs in the region. Since sea ice is costly and time-consuming to traverse, this has created a need for forecasts of the sea-ice edge at seasonal and sub-seasonal time scales. Probabilistic forecasts are well suited for this task, since forecast uncertainty must be accounted for to accurately assess the risk of a particular navigation route. I will introduce a statistical technique for directly modeling the sea ice edge built from historical observations. I will then extend this technique to also incorporate information from bias-corrected output from dynamic sea-ice models. Finally, I will compare the performance of these new techniques with existing dynamic and statistical approaches in a probabilistic framework. This will allow for evaluation of both mean sea-ice edge forecasts and the uncertainty around them.


Snow-depth and air-temperature evolution on sea ice derived from snow-buoy measurements

Marcel Nicolaus, Mario Hoppmann, Stefanie Arndt, Stefan Hendricks, Christian Katlein, Anja Nicolaus, Leonard Rossmann, Martin Schiller

Corresponding author: Marcel Nicolaus

Corresponding author e-mail: marcel.nicolaus@awi.de

Snow depth is one of the sea-ice essential climate variables, because it dominates the energy and momentum exchanges across the atmosphere–ice–ocean interfaces and actively contributes to sea-ice mass balance. Yet, snow depth is one of the least known and most difficult to observe parameters of the Arctic and Antarctic sea-ice cover, mainly due to its exceptionally high spatial and temporal variability and the lack of accurate methods to retrieve snow depth by remote sensing. Here, we present 5 years of snow-depth and atmospheric measurements, recorded by a new buoy type. ‘Snow Buoys’ are designed for easy deployment and minimal impact on the surface. Time series from 52 buoys reveal characteristic regional differences in the annual cycle of snow depth on sea ice with high data quality and reliability. On Antarctic sea ice, almost no reduction in snow depth was observed over summer, allowing an annual net snow accumulation of 0.2–0.9 m. On Arctic sea ice, snow accumulated through the entire fall, winter and spring until a strong summer surface ablation removed the entire snow pack again. Arctic air temperature measurements revealed above-freezing temperature events in winter that likely impacted snow stratigraphy and the preceding spring snow cover and properties. Continuing the deployment program, Snow Buoy data will help to advance studies of snow on sea ice, in particular through links with numerical simulations and remote-sensing techniques.


Arctic sea-ice melt-pond-onset detection from dual sensor C-band co-polarized SAR and scatterometer data

Syeda Shahida Maknun, Torsten Geldsetzer, John Yackel, Randall Scharien, Vishnu Nandan

Corresponding author: Syeda Shahida Maknun

Corresponding author e-mail: syedashahida.maknun@ucalgary.ca

The presence of melt ponds dramatically changes the thermal and radiative properties of sea ice. The albedo of melt ponds is much lower than that of snow or ice, and thus affects the surface radiation balance through higher shortwave absorption. This process tends to accelerate melting of adjacent snow cover and the ice beneath. This ‘snow/ice-albedo’ feedback mechanism has been examined extensively and is thought to be a vital feedback mechanism operating within the ocean–sea-ice–atmosphere interface. Understanding these processes and mechanisms associated with the timing of melt-pond formation has become a subject of interest in a multidisciplinary context. This study aims to develop a novel and reliable method to detect sea-ice melt-pond onset (PO) from C- band SAR backscatter (RADARSAT-2, Sentinel–1A and B) and spaceborne scatterometer data from ASCAT, in conjunction with meteorological data. We produce daily dual-sensor co-polarized (co-pol) backscatter ratio (VV/HH) values for the Canadian Arctic Archipelago, based on daily-coincident ASCAT VV and wide-swath SAR HH data. We normalize all data to an incidence angle of 40°. We investigate the seasonal progression from late winter to PO for 3 years (2016–18). Time series of the co-pol ratio are analyzed in relation to wind speed, precipitation and air temperature. PO is observed when the co-pol ratio exceeds ~2.5–3.0 dB. We validate PO observations with available cloud-free optical imagery such as Sentinel-2A during the transition from melt onset to ponding. Full daily coverage of the Canadian Arctic Archipelago depends on SAR data availability. Therefore, daily co-pol data are not always available for all locations, occasionally reducing the precision of PO detection in those locations. Results demonstrate the efficacy of using dual-sensor C-band co-pol microwave backscatter for detecting pond onset over large spatial scales.


Winter N2O dynamics in pack ice and underlying water in the Ross Sea

Bruno Delille, Fanny van der Linden, Marie Kotovitch, Gauthier Carnat, Célia Sapart, Jeroen de Jong, Florian Deman, François Fripiat, Frank Dehairs

Corresponding author: Bruno Delille

Corresponding author e-mail: bruno.delille@uliege.be

Nitrous oxide (N2O) is a potent greenhouse gas with a high global-warming potential. N2O is also strongly involved in stratospheric ozone depletion. The Southern Ocean has been considered as one of the dominant oceanic sources of nitrous oxide for the atmosphere, due to the release of the N2O excess accumulated as a result of organic matter remineralization along the deep oceanic pathway and ultimately ventilated in the surface waters of the Southern Ocean. However, actual data are scare and reveal that both undersaturation and oversaturation conditions occur in Southern Ocean surface waters. Undersaturation in N2O of polar surface waters has frequently been ascribed to melting of sea ice that is presumably undersaturated in N2O. During the 2017 PIPERS cruise in the Ross Sea, we carried out the first winter measurements of N2O in winter in both in sea ice and in the water column. Comparison of these new winter surface-water measurements to available summer measurements reveal contrasting results and may challenge the current view of the Southern Ocean being a source of N2O for the atmosphere. In addition, we observed a build-up of N2O in the ice interior, probably being produced by the sympagic microbial community, and a subsequent release to the atmosphere during sea-ice formation. Surprisingly, surface waters, in contrast, appear to act as a sink of N2O for the atmosphere. While we confirm that melting of sea ice decreases the N2O concentration of surface waters, this impact is limited and unable to explain the level of undersaturation previously reported in N2O during summer. Further processes are therefore required to explain this level of undersaturation.


Sea-ice and ocean responses to increasing Antarctic mass loss

Shona Mackie, Jeff Ridley, David Stevens, Inga Smith, Pat Langhorne

Corresponding author: Shona Mackie

Corresponding author e-mail: shona.mackie@otago.ac.nz

Antarctica is losing mass at an increasing and spatially variable rate. Research shows a spatially variable climate response to increased freshwater fluxes entering the Southern Ocean, which suggests sensitivity to the spatial distribution of melt fluxes from icebergs and ice shelves. However, climate projections traditionally assume a constant and spatially uniform rate. The HadGEM3-GC3.1 model includes spatial variability by using glaciology estimates to distribute mass loss in a realistic way around the coast. Basal melt and icebergs are produced at each ice-shelf front: melt enters the ocean immediately using an improved parameterization for its vertical distribution, while icebergs are transported and melt according to ocean properties. The total mass loss rate, however, remains temporally constant. We have investigated sea-ice and ocean sensitivity to this assumption in order to understand the implications for climate projections that omit it, and to gain insights into possible future climate effects that are not captured by the current generation of climate models. To our knowledge, this is the first evaluation of climate response to an increasing rate of Antarctic mass loss with melt realistically distributed using a coupled model to capture feedbacks.


Improving the representation of grease ice processes in an earth system model

Shona Mackie, Inga Smith, Pat Langhorne

Corresponding author: Shona Mackie

Corresponding author e-mail: shona.mackie@otago.ac.nz

In the current standard configuration of the sea-ice module in the New Zealand Earth System Model (NZESM), ice crystals form in the ocean in response to supercooling, and rise to the surface where they freeze instantly into a layer of sea ice with uniform, prescribed, thickness. If this layer overflows the open water area of the grid cell, the ‘extra’ frazil thickens the existing ice, and the new layer. In reality, when frazil rises to the ocean surface, it mixes with the ambient water to form a slushy layer of grease ice, with a greater volume than the volume of the frazil alone. Rather than forming a uniform layer, it becomes wedge-shaped as wind drives it against the sea-ice edge. Instead of freezing immediately to become new sea ice, it freezes (or melts) primarily in response to atmospheric heat flux. In the real world, therefore, leads may remain open for longer, increasing ocean–atmosphere heat exchange and decreasing surface albedo. We have extended a scheme developed at Reading University, UK, to represent the behavior of grease ice at the ocean surface more realistically in the NZESM and will use a fully coupled model run to investigate the ocean and sea-ice response to the improvement.


Analysis of heteroatomic species in biodegraded crude oil using electrospray ionization ion mobility time-of-flight high-resolution mass spectrometry

Nolan Snyder, Feiyue Wang, Jake Ritchie, Diana Saltymakova, Katarzyna Polcwiartek, Durell S. Desmond, Alastair F. Smith, Casey Hubert, Gary A. Stern

Corresponding author: Gary A. Stern

Corresponding author e-mail: Gary.Stern@umanitoba.ca

Increasing industrialization of the Arctic has raised concerns regarding oil spills from activities such as marine shipping and calls for the development of effective oil-spill remediation. Due to the Arctic’s remote location, microbial degradation could play a role in Arctic oil-spill remediation. Our purpose is to better understand the significance of microbial degradation in the Arctic marine environment. A mesocosm was carried out at the University of Manitoba, Winnipeg, Canada, as a part of the Microbial Genomics for Oil Spill Preparedness in Canada’s Arctic Marine Environment (GENICE) project. This study involved two large (7 m3 and 12 m3) outdoor, above-ground pools filled with artificial seawater mixed with nutrients. One of the pools was inoculated with indigenous bacteria collected from Cambridge Bay during the CCGS Amundsen 2017 summer field season. Light crude oil from Tundra Oil & Gas Ltd was released under the ice in both pools after 20 cm of ice had been grown under natural ambient temperature. Ice cores were taken from both pools every 3 weeks, and extraction of crude oil was followed by direct infusion to our mass spectrometry (MS) system. Chemical analysis was performed using the Waters Synapt G2-Si electrospray ionization (ESI), and ion mobility – time of flight – high-resolution mass spectrometry (IM-TOF-HRMS) system. We took advantage of ESI coupled with IM to minimize fractionation while using collision cells before and after the IM chamber to conduct daughter ion analysis. Before direct infusion to MS, oil samples were prepared in a solution of 20 mL 1 : 1 (v : v) MeOH : toluene, and 500 μg mL–1 oil. 0.5% formic acid was added to each 30 mL Nalgene bottle for protonation, and a duplicate was prepared with 1.5% NH4OH for deprotonation. DriftScope software performs peak detection and calibration, while PetroOrg software produces a calculated molecular formula and double-bond equivalence, based on the accurate masses from the imported DriftScope peak list. Preliminary data using positive ESI mode shows a higher diversity of nitrogen-containing compounds in the pool inoculated with Arctic bacteria, while the abundance is higher in uninoculated samples. While it has been difficult to separate and analyze heteroatoms in petroleum oils using ESI-MS in the past, IM has added another dimension to the MS data providing structural identification of heteroatomic and polar species in resinous biodegraded oil in the Arctic.


Advances in modelling interactions between sea ice and ocean surface waves

Lettie Roach, Cecilia Bitz, Christopher Horvat, Samuel Dean

Corresponding author: Lettie Roach

Corresponding author e-mail: lettie.roach@niwa.co.nz

Sea ice is composed of floes that range in size from meters to many kilometers. Sea-ice models currently do not resolve individual floes, nor parameterize many important features related to the granular nature of the sea-ice mosaic. We have previously developed a process-based model that captures some key characteristics of the sea-ice floe-size distribution (FSD), which incorporates the sub-grid-scale variability of floe size and evolves subject to lateral melt, lateral growth, fracture by ocean surface waves, floe welding and new ice growth. Here, we build upon this earlier work and demonstrate a coupled ocean-wave–sea-ice model including new physical parameterizations for the prognostic evolution of the sea-ice FSD in CICE6, combined with preliminary model validation against pan-Arctic measurements. We discuss how such a floe-sensitive sea-ice model can be used to investigate changes in sea-ice fragmentation as the Arctic climate has changed, with implications for the future of Arctic sea-ice cover. We further examine potential new process improvements and observations that would improve model fidelity. Accurate simulation of the FSD opens up new opportunities for understanding sub-grid-scale variability in polar systems.


Change and variability in Antarctic coastal exposure (lack of sea ice offshore) since 1979

Phil Reid, Rob Massom

Corresponding author: Rob Massom

Corresponding author e-mail: rob.massom@aad.gov.au

Sea ice is a ubiquitous though seasonally variable feature of the Antarctic coastal zone, where it interacts strongly with the floating ice-sheet margins that are particularly vulnerable to changing ocean conditions (including warming). Indeed, there is also mounting evidence of potentially strong relationships between coastal sea-ice distribution and the characteristics and stability of glacier tongues and ice shelves, suggesting that sea ice is an indirect (though poorly understood) player in regulating sea-level rise. One aspect of this, as highlighted in a recent article in Nature, is that change in the presence/seasonality of a protective sea-ice ‘buffer’ offshore may enhance flexure and fatigue of exposed outer ice-shelf margins by storm-generated ocean swells that are normally damped by a sea-ice cover. This can in turn promote outer-margin calving, which in extreme cases (such as Larsen B in 2002) can precipitate rapid large-scale disintegration of ice-shelf regions weakened by multiple melt and glaciological factors. Here, we introduce and present findings from two complementary new algorithms designed to quantify broad-scale change and variability in Antarctic coastal exposure to more open-ocean conditions (lack of sea ice offshore), by exploiting the satellite passive-microwave sea-ice concentration record dating back to 1979. These are: (1) an ‘Antarctic Coastal Exposure Index’ and (2) a more detailed ‘Coastal Exposure Length’ method. Initial examination of temporal and spatial patterns of occurrence and trends for 1979–2016 shows that West Antarctic coastal regions are largely dominated by an increase in coastal exposure, particularly in the West Antarctic Peninsular region but also in the Amundsen Sea. In contrast, areas of increasing coastal exposure area are confined to smaller pockets only in East Antarctica, with the general trend being towards decreasing coastal exposure. This is further characterized by a distinct westward progression around the region, largely in summer. The new algorithms and findings complement the more widely used sea-ice concentration, extent and seasonality time series.


IceNode: a buoyant sensor pod for persistent in-situ measurement beneath ice shelves

Evan Clark, Justin Schachter, Daniel Limonadi, Rebecca Castano

Corresponding author: Evan Clark

Corresponding author e-mail: evan.clark@jpl.nasa.gov

Lack of understanding about how warming climate will influence Antarctic ice-shelf collapse remains the single largest reason for uncertainty in sea-level-rise projections over the 21st century. Predictive numerical ice-shelf models require better constraining ground-truth data, but the field suffers from a dearth of in-situ measurements because these environments are physically extreme, difficult to access and operate in, and cut off from communications with the outside world. Notable efforts have succeeded in deploying underwater gliders, long-range autonomous underwater vehicles, free-floating sensor platforms, and platforms tethered to the ice ceiling, but none of these approaches have been able to achieve sufficient long-duration, well distributed measurements directly at the melt interface to satisfactorily constrain numerical melt models. Here, we present IceNode, a buoyant sensor pod being developed at the NASA Jet Propulsion Laboratory (JPL) for gathering persistent in-situ environmental measurements at the melt interface underneath ice shelves. IceNode may be distributed beneath an ice shelf by drifting on currents starting either at the shelf edge or through a borehole, then land buoyantly against the underside of the shelf to take several months of measurements. Once the science phase is complete, IceNode detaches from the ceiling and rides outgoing currents to open water, then comes to the surface and transmits its data home using Iridium. We describe the design, development and mission operations of IceNode, and why we expect IceNode to facilitate key scientific measurements not easily achievable by traditional methods in these challenging but crucially important environments below the ice.


Worldwide compilation of air–sea ice CO2 flux with the enclosure method: similar amplitudes to open-ocean measurements

Daiki Nomura, Nicolas-Xavier Geilfus, Jean-Louis Tison, Brent Else, Kristina Brown, Lisa Miller, Gauthier Carnat, Sebastien Moreau, Tim Papakyriakou

Corresponding author: Daiki Nomura

Corresponding author e-mail: daiki.nomura@fish.hokudai.ac.jp

A compilation of air–sea-ice CO2 fluxes from the Arctic Ocean, the Sea of Okhotsk and the Southern Ocean, measured with the enclosure method, shows a clear seasonal cycle in air–sea-ice CO2 fluxes, with sea ice acting as a CO2 source to the atmosphere during the cold season (up to +12 mmol m–2 d–1) and a sink during the warm season (down to nearly –6 mmol m–2 d–1). This seasonal variation is correlated with air temperature, mainly because the partial pressure of CO2 in the near-surface upper part of sea ice is a function of the ice temperature. This temperature dependence of the air–sea-ice CO2 flux was also observed in tank experiments. In addition, air–sea-ice CO2 fluxes are strongly dependant on the ice surface conditions (such as presence/absence of snow). The magnitude of the air–sea-ice CO2 fluxes is comparable to those of open ocean air–sea CO2 fluxes, suggesting that sea ice could play a significant role in the global carbon cycles.


Using BGC-Argo to understand Southern Ocean marginal ice zone bloom dynamics

Sebastien Moreau, Philip W. Boyd, Peter G. Strutton

Corresponding author: Sebastien Moreau

Corresponding author e-mail: sebastien.moreau@npolar.no

We have developed a new framework using biogeochemical (BGC) sensors on Argo floats to differentiate the main fates of primary production. We focused on the sea-ice zone of the Southern Ocean. Our multi-faceted analysis involves the cumulative accounting of phytoplankton loss terms. When combined with conventional observations (chl-a, particulate organic carbon (POC), mixed-layer dynamics and sea-ice cover), this stepwise analytical approach allows us to begin to discriminate between export mechanisms: particle detrainment, sinking ice algae, ice-edge bloom export, settling aggregated senescent phytoplankton, marine snow and sinking faecal pellets. After constraining export, we can estimate how much chl-a has been grazed by zooplankton, and hence begin to develop regional and seasonal patterns into the fate of primary production in the geographically remote Southern Ocean.


Analysis of the Airborne Topographic Mapper response over melt ponds on sea ice

Ellen M. Buckley, Sinéad L. Farrell, Kyle Duncan, Laurence N. Connor, John M. Kuhn, Roseanne T. Dominguez

Corresponding author: Ellen M. Buckley

Corresponding author e-mail: buckley@umd.edu

During the summer melt season, melt ponds form as snow melts and collects in low-lying topographic features on the sea ice surface. These melt ponds allow for additional heat uptake as their blue color has a lower albedo than the surrounding snow-covered sea ice. As the snow and ice continue to melt, the pond darkens and becomes more absorptive, enhancing the positive ice-albedo feedback. The exact albedo of a pond is determined by the melt-pond depth and by the properties of the underlying ice. By observing melt-pond parameters on sea ice, we can better understand the sea-ice processes that occur during the summer melt season. Also, melt-pond observations may improve sea-ice forecasting, since it has been shown that the inclusion of melt-pond parameters in sea-ice models increases the accuracy of predicting the summer sea-ice minimum extent. This study utilizes airborne observations collected during NASA Operation IceBridge (OIB) summer sea-ice flights to extract new parameters describing melt ponds. In 2016 and 2017 the Airborne Topographic Mapper (ATM) and Digital Mapping System (DMS) were flown onboard the OIB aircraft during two Arctic summer melt campaigns, which surveyed thousands of kilometers of sea ice. The DMS instrument captures RGB spectral values at 0.1 m resolution. The ATM laser altimeter measures sea-ice surface elevation. The records of the received waveform strength and shape information are also available for the 2017 flights. The ATM laser operates at 532 nm, a wavelength that is able to penetrate water. Applying our novel sea-ice surface-classification algorithm to the DMS images, we locate melt ponds and assess the coincident ATM elevations and waveform returns to characterize the ATM instrument response over the melt ponds. Using the ATM elevation measurements, we analyze relationships between the sea-ice topographic features and melt-pond location. We evaluate melt-pond formation and depth on first-year and multiyear ice where we expect to see differences due to their distinctive ice topographies and thickness. We evaluate the feasibility of using the ATM data to determine melt-pond depth. Our goal is to assess the airborne laser altimetry measurements in the summer melt season to provide insight into the ICESat-2 returns, a sensor that obtains spaceborne observations operating at the same wavelength. This study enhances our understanding of melt-season processes and develops new techniques to remotely measure sea-ice melt ponds.


Uncertainty in projections of freshwater supply to Hudson Bay

Andrew A.G. Tefs, Tricia Stadnyk, Stephen Déry, Scott Pokorny, Kristina Koenig

Corresponding author: Andrew A.G. Tefs

Corresponding author e-mail: umtefs@myumanitoba.ca

Through the BaySys group of projects, daily projections of freshwater discharge delivered to the Hudson Bay complex (HBC) are modelled, to be supplied to a regional NEMO sea-ice model to quantify the effects of hydroelectric regulation and climate change on sea-ice formation, bay-wide dynamics, and the subsequent biogeochemical processes of this biologically complex and important region. Terrestrial discharge projections are generated using the Hydrological Predictions for the Environment (HYPE) hydrological model, forced by an ensemble of CMIP-5 climate models. Previous studies have generated a robust uncertainty study examining input, parameter, model-structure and output uncertainties over the period 1981–2070 for the largest single contributor of freshwater to Hudson Bay, the Nelson River. The work presented here introduces a proposed methodology for transferring these Nelson River daily uncertainty bounds to all the remaining rivers (397 in total) of the Hudson Bay drainage basin. We also present new, modelled, uncertainty bounds created for nearby regulated (Churchill River) and natural (Hayes River) watersheds. Through a more sophisticated quantification of the types and magnitude of uncertainty in freshwater modelling, this work aims to improve confidence in the results of modelled sea-ice dynamics and ultimately, confidence in the integration of modelled results and observed biogeochemical data into a complete portrait of the future health of Hudson Bay in a changing climate.


The evolution of sea ice from CMIP3 to CMIP6

Julienne Stroeve, Dirk Notz, Dirk Olonscheck, Erica Rosenblum

Corresponding author: Julienne Stroeve

Corresponding author e-mail: j.stroeve@ucl.ac.uk

Arctic sea-ice retreat is one of the most striking indications of climate change in recent decades. Historically, climate models have struggled to simulate an Arctic sea-ice retreat that is as fast as in the observations. Previous studies found that CMIP5 models simulated a faster Arctic sea-ice retreat that was more consistent with the observations than CMIP3 models, though the reason for this apparent improvement was unclear. A number of follow-up studies have since drawn upon the close relationship between sea-ice area, global mean surface temperatures, and CO2 emissions to examine how simulated sea-ice retreat might be related to climate sensitivity and sea-ice sensitivity. The results of these studies indicate that both CMIP3 and CMIP5 models have Arctic sea-ice cover that is similarly sensitive to global warming and CO2 and that this sensitivity is significantly smaller than the observations. Moreover, the faster sea-ice retreat in CMIP5 models may be due to their overestimating the level of global warming. Here we examine how well 1979–2018 Arctic sea-ice retreat is simulated in the last three generations of climate models compared to the observations using simulations from CMIP3, CMIP5 and CMIP6. Specifically, we evaluate simulated sea-ice concentration, area and ice thickness and their sensitivity to global warming and CO2 across all months. Lastly, we estimate the extent to which differences in simulated sea-ice retreat could be related to differences in simulated climate sensitivity, sea-ice sensitivity, global temperatures, CO2 emissions, or internal variability.


Assessing relative climate change and regulation impacts in the Hudson Bay complex: on the development of an integrated observational-modeling framework

Jennifer V. Lukovich, Shabnam JafariKhasragh, Andrew Tefs, Natasha Ridenour, Sergei Kirillov, Inge Deschepper, Tricia Stadnyk, Kathleen Munson, Karen Wong

Corresponding author: Jennifer V. Lukovich

Corresponding author e-mail: Jennifer.Lukovich@umanitoba.ca

BaySys, a collaborative project between the University of Manitoba and Manitoba Hydro, is designed to understand the relative impacts of climate change and regulation on freshwater-marine coupling in the Hudson Bay complex (HBC). Central to this understanding is the development of an integrated observational-modeling framework that combines sea-ice, oceanographic, hydrological and biogeochemical considerations. This presentation describes the components of such a framework using the Arctic and Northern Hemisphere Atlantic configuration of the Nucleus for European Modelling of the Ocean (NEMO) model as well as observations from the 2016–18 BaySys field program. In particular, consistent and standardized diagnostics are presented for i) a description of atmospheric and discharge conditions that drive the NEMO model and govern physical processes in the HBC for the BaySys 2016–18 timeframe, and corresponding rankings relative to the 1981–2010 atmospheric and discharge climatology, ii) inter-model and model-observational comparison of ice, oceanographic and biogeochemical variables for the BaySys 2016–2018 timeframe, and iii) an assessment and synthesis of relative climate change and regulation impacts on oceanographic, sea-ice, hydrological and biogeochemical processes in the HBC by comparing physical variables estimated from naturalized and regulated historical and future climate modeling scenarios. These diagnostics will quantify the relative contribution of climate and regulation on conditions in Hudson Bay, illustrate spatial and temporal differences and sensitivity in freshwater-marine processes to climate change and regulation impacts, and contribute to the establishment of indices relevant to industry, academia and communities.


Future sea-ice decline predicted to bring the Arctic nations closer together

Patricia DeRepentigny, Bruno Tremblay, Robert Newton, Stephanie Pfirman, Alexandra Jahn

Corresponding author: Patricia DeRepentigny

Corresponding author e-mail: patricia.derepentigny@colorado.edu

Over the past decades, Arctic sea ice has declined in thickness and extent and is shifting towards a seasonal ice regime, with accelerated ice drift and an increasingly larger seasonal ice zone. These rapid changes have widespread implications for ecological and human activities as well as for the global climate, and accurate predictions could benefit a wide range of stakeholders, from local residents to governmental policy makers. Specifically, the changing Arctic ice cover will impact the trans-border exchange of sea ice between the exclusive economic zones (EEZs) of the Arctic nations, with important consequences for ice-rafted contaminant transport. To investigate current and projected changes to transnational ice exchange, we use the Lagrangian Ice Tracking System (LITS) to follow ice floes from the location of their formation to where they ultimately melt. We apply this tool to output from two ensembles of the Community Earth System Model (CESM): the Large Ensemble uses a high emission scenario that leads to over 4°C of global warming by 2100, and the Low Warming ensemble uses reduced emissions resulting in a stabilized warming of 2°C by 2060. We also use the National Snow and Ice Data Center Polar Pathfinder and Climate Data Record products to evaluate the fidelity of the CESM present-day tracking simulations and find that transnational ice exchange in the CESM compares well with observations. The CESM projects that by mid-century, transnational ice exchange will greatly expand, with a large increase in the fraction of transnational ice originating from Russia and the Central Arctic. By the end of the 21st century, we see a large impact of the emission scenario on ice exchange: consistent ice-free summers under the high-emission scenario act to reduce the total fraction of transnational ice exchange compared to mid-century, whereas the low-emission scenario continues to see an increase. Under both scenarios, average transit times are predicted to decrease to less than 2 years by 2100, compared to a maximum of 6 years under present-day conditions and 2.5 years by mid-century, effectively bringing the Arctic nations closer together.


Modelling the impact of sea-ice biogeochemical processes on dimethylsulfide and carbon fluxes in the Arctic

Nadja Steiner, Hakase Hayashida, Eric Mortenson, Tessa Sou, Adam Monahan

Corresponding author: Nadja Steiner

Corresponding author e-mail: nadja.steiner@dfo-mpo.gc.ca

A coupled sea-ice–ocean biogeochemistry model with pelagic and sea-ice ecosystem components including carbon and sulfur cycling is applied to the Arctic. The model has been run for recent time periods (1979–2015) and evaluated with respect to sympagic and pelagic primary production, carbon fluxes and oceanic dimethylsulfide (DMS). The model indicates high interannual variability in ice-algal production as well as long-term trends. Simulated DMS shows good correspondence with satellite-derived DMS observations and highlights significant shortcomings of the coarsely interpolated commonly used DMS climatology of Lana et al. (2011) in the Arctic region. The carbon system shows significant regional variability in the Arctic. Including the sea-ice carbon pump shows only a small change to the total uptake of carbon but a marked decrease in the seasonal variability of both DIC and TA, as well as an offset in the summertime saturation state in the surface Arctic Ocean. Including sea-ice algae has a cumulative effect on upper-water-column nutrients and carbon uptake, suggesting that inclusion of the sympagic ecosystem has a growing importance for longer-term model runs. The model results suggest this to be a useful tool to assess future projections of DMS emissions and carbon state variables.


Effects of sea ice, runoff, and carbon cycling on CO2 flux and ocean acidification in Hudson Bay

David Capelle, Lisa Miller, Tim Papakyriakou, Zou Zou Kuzyk, Robie Macdonald

Corresponding author: David Capelle

Corresponding author e-mail: david.capelle@umanitoba.ca

Air–sea exchanges of CO2 and ocean acidification in high-latitude coastal seas are influenced by a wide range of biogeochemical processes, including i) sea-ice formation and melt; ii) river runoff; iii) the formation and remineralization of organic carbon; and iv) the formation and dissolution of calcium carbonate. Previous studies in Hudson Bay have shown a link between fresh water (from runoff and sea-ice melt) on air–sea CO2 exchange and aragonite saturation, but lacked a quantitative evaluation of these relationships. We use a box model to quantitatively evaluate the changes in CO2 partial pressure and aragonite saturation resulting from each of the above processes in different regions of Hudson Bay, Canada. Our findings support the field data, which show links with river runoff and ice melt. This work builds upon previous work that evaluated the relative impacts of terrestrial carbon remineralization, dissolution of terrestrial calcium carbonate, and export production on pCO2 and aragonite saturation in Hudson Bay. These processes affect CO2 flux and ocean acidification not only in Hudson Bay but also in the Labrador Sea, which is influenced by Hudson Bay waters.


Challenges and advancements in sea-ice biogeochemical modeling for Earth-system applications

Nicole Jeffery, Mathew Maltrud, Shanlin Wang, Elizabeth Hunke, Scott Elliott, Jon Wolfe

Corresponding author: Nicole Jeffery

Corresponding author e-mail: njeffery@lanl.gov

We show results from the E3SM.v1 (Energy Exascale Earth System Model) 1850 control and historical simulations with active biogeochemical components in ocean and sea ice. In fully coupled Earth-system modeling, determining the fidelity of the sea-ice BGC component requires an assessment of the simulated polar climate as it pertains to ice algae. By defining a maximal bound for algal growth based on simulated and observed surface nutrients and incident shortwave, we are better able to identify consequential biases in the simulated polar climate. In addition, we find that simulated ice-algal primary production in nutrient-limited regions is strongly correlated with the maximal growth bound. For the Arctic, this provides a means to better estimate annual ice primary production. In contrast, simulated Southern Ocean sea-ice primary production is less sensitive to ocean surface nutrients, not correlated with the maximal growth bound, and controlled to a greater extent by sea-ice processes. We conclude by contrasting the role of snow and ice melt on simulated ice-algal chlorophyll concentrations in the two poles.


FAMOS multi-model intercomparison of ice-algal productivity in the Arctic Ocean on seasonal, interannual and decadal timescales

Eiji Watanabe, Meibing Jin, Hakase Hayashida, Jinlun Zhang, Nadja Steiner

Corresponding author: Eiji Watanabe

Corresponding author e-mail: ejnabe@jamstec.go.jp

Seasonal, interannual and decadal variations in the Arctic ice-algal productivity for 1980–2009 are investigated using daily outputs from five sea-ice‒ocean ecosystem models participating in the Forum for Arctic Modeling and Observational Synthesis (FAMOS) project. The spatial distribution of the simulated ice-algal productivity is commonly characterized by a sharp shelf–basin contrast. The analysis focuses on four sub-regions: the Chukchi Sea, Canada Basin, Eurasian Basin and Barents Sea. The FAMOS models simulate reasonable seasonal cycles of snow, sea-ice and ocean properties related to ice algae, whereas the amplitudes of most properties are broadly different in each sub-region. The simulated annual total ice-algal productivity has no common decadal trend at least for 1980–2009 among the five models, although snow depth and sea-ice thickness in spring mostly decline. The model intercomparison indicates that an appropriate balance of stable ice-algal habitat (i.e. sea-ice cover) and enough light availability is necessary to retain the productivity. The multi-model averages show that the ice-algal bloom timing shifts to an earlier date and the bloom duration shortens in the four sub-regions. However, both the positive and negative decadal trends in the timing and duration are simulated owing to different types of ice-algal blooms: long–massive, short‒massive, long‒gentle, and short‒gentle bloom characteristic for the five models. The simulated uncertainties on the pan-Arctic and decadal scales are expected to inform for further studies and future projections.


Arctic sea ice in the CESM2 over the 20th and 21st centuries

Patricia DeRepentigny, Alexandra Jahn

Corresponding author: Patricia DeRepentigny

Corresponding author e-mail: patricia.derepentigny@colorado.edu

In this study, we explore the current and future states of Arctic sea ice within the Community Earth System Model Version 2 (CESM2), the CESM contribution to the Coupled Model Intercomparison Project Phase 6 (CMIP6). We investigate projected changes in Arctic sea-ice cover in order to establish the main features of the CESM2 simulations and understand any model biases. Specifically, we look at the evolution of September and March sea-ice extent and associated trends, changes in sea-ice thickness, multiyear ice loss and volume tendencies, the timing of an ice-free Arctic, rapid ice-loss events and the mechanisms behind them based on existing theories, changes in ice drift, and regional patterns of sea ice loss and associated atmospheric and oceanic drivers. We will show results from two different atmospheric model components of the CESM2, namely the Community Atmosphere Model Version 6 (CAM6) and the Whole Atmosphere Community Climate Model (WACCM), as those two sets of experiments show large differences in their simulation of sea ice over the historical period. The goal of this study is to provide an overview of Arctic sea ice to guide future investigation of the transient polar climate response within the CESM2 and CMIP6.


Using models to assess climate-change stressors on marine habitats and species distribution in the Arctic

Nadja Steiner, Tessa Sou

Corresponding author: Nadja Steiner

Corresponding author e-mail: nadja.steiner@dfo-mpo.gc.ca

Local observations in the western Canadian Arctic indicate clear changes in marine species, e.g. reduced catches of Arctic char, increased appearance of salmon species, shifts in marine-forage fish species with less Arctic cod and more sandlance, temporal and spatial shifts in beluga and bowhead whales’ appearances. Here we evaluate a regional sea-ice–ocean–biogeochemical model representing sea-ice and pelagic ecosystems with respect to changes in the last 30 years to understand if changes in the sea-ice–ocean environment can shed more light on observed shifts in species. We will particularly focus on changes in temperature, sea-ice stratification and as pelagic versus sea-ice algal production, and discuss additional stressors on the marine ecosystem, such as ocean acidification.


The sensitivity of CICE to the choice of ice-shelf–ocean coupling algorithms

Stefan Jendersie, Alena Malyarenko

Corresponding author: Stefan Jendersie

Corresponding author e-mail: s.jendersie@mailbox.org

To quantify Antarctic ice-mass loss and the subsequent sea-level rise the geophysical modelling community is pushing towards frameworks that fully couple increasingly complex models of atmosphere, ocean, sea ice, and ice sheets and shelves. One particular hurdle remains the accurate representation of the vertical ocean–ice interaction at the base of ice shelves. Parameterizations that are tuned to particular datasets naturally perform best in comparable ice-shelf-cavity environments. This poses the challenge in continental scale ocean–ice-shelf models to chose one melt parameterizaton that performs sufficiently well in diverse cavity environments. However this adds significant uncertainty in ice-shelf-induced ocean freshening, which crucially affects modeled sea-ice growth. The magnitude of the impact of ice-shelf-supplied melt water on growth rates, thickness and extent of sea ice in the open ocean is currently debated in the literature. We present results of two studies. First we review, compare and discuss 16 commonly utilized melting/freezing parameterizations in coupled ocean–ice-shelf models. Melt rates differ hugely, in an idealized ISOMIP configurations from less than 0.1 m a–1 to 5 m a–1, depending on the chosen parameterization. Secondly, by use of a realistic circum-Antarctic coupled ice-shelf, sea-ice and ocean model (CICE, ROMS) we look at the effects of the chosen ice-shelf-melt parameterization on modeled sea-ice growth, regionally and globally.


Sensitivity analysis of a simple wave tracker for Arctic environments using a combination of empirical and numerical models

Yanique Campbell, Sergei Kirillov, Jens Ehn, David Barber

Corresponding author: David Barber

Corresponding author e-mail: David.Barber@umanitoba.ca

Ocean waves have become an important consideration when mapping how changes in Arctic sea ice will affect, and are affecting, the marine and coastal systems. While advances in wave modelling have allowed for improved forecasting of waves, the ability to regress in time during a wave-growth event can provide valuable insight into the origin of wave groups, as well as their propagation and growth, particularly in sea-ice environments. This sensitivity analysis considers a simple wave tracker that begins with the end result: observations of significant wave heights and peak periods of wind–sea (growing waves), and uses both numerical and empirical models to delineate the path over which the wave growth occurred in time, the growth limitation caused by sea-ice cover and the wind duration that allowed for wave development. The tracker uses wind speed and direction, sea-ice concentration, wave peak periods, wave direction and bathymetry data to determine the environment that was necessary for the resulting wave group to develop. Once these conditions are determined, the modified JONSWAP spectrum was used to calculate the incremental growth of waves as they travel along a specified fetch. The calculated wave heights were compared to observations recorded by an Ice Profiling Sonar. Wind and wave data were obtained from the ERA5 reanalysis, ice-concentration data from the Canadian Ice Service charts as well as the OSI SAF products, and bathymetry from the ETOPO1 dataset for September–October 2009 in the Beaufort Sea. The sensitivity of this tracker to the variations in wind speed and direction along a wave-growth fetch and throughout time, as well as with different wind-fetch-duration test cases, was analyzed. The tracker showed good results in determining wave generation and propagation environments as well as wave growth in space and time, particularly in steady wind conditions, where there are no large or erratic shifts in the wind direction over the duration of wave growth.


Ross Sea ice production and fast-ice edge using Sentinel–1 SAR images

Liyun Dai, Hongjie Xie, Stephen Ackley, Alberto Mestas

Corresponding author: Hongjie Xie

Corresponding author e-mail: hongjie.xie@utsa.edu

High sea-ice production (SIP) causes high-salinity-water formation, and the sinking of dense water in polar regions drives the global thermohaline circulation and plays important roles in the bio-related material cycle and ecosystem. Studies reported that Ross Sea ice production is believed to take place in the coastal polynya and wind-driven ice export from the Ross Sea polynya area. However, we did not know how much SIP in the coastal polynya is from wind. In this study, we used Sentinel–1 SAR data to calculate the SIP driven by katabatic winds in the Ross Sea and determine the edges of fast ice in Ross Bay. The result showed that the fast-ice edge arrived at its most southerly point on 30 March 30 2017; from then on, it expanded northward. The edge reached a maximum on 3 May 2017 and maintained it to 30 January 2018. The edge started retreating on 6 February 2018; from then on, it moved south quickly and reached the most southerly point on 14 March 2018. During 2018/19, the change of fast-ice edge was similar to that in 2017/18. Wind-driven SIP from the Ross Ice Shelf polynyas and McMurdo Sound polynyas was also estimated based on GRD images for the 2017/18 season. Wind-driven sea ice events in these two polynyas during 2017/18 occurred 34 and 38 times, respectively, and the total areas were 746 594 km2 and 75 931 km2, respectively. During 2018/19, the number of occurrences were 66 and 26, and the cumulative wind-driven sea-ice areas of these two polynyas were 939 531 and 77 521km2, respectively. According to the observation of PIPERS, the thickness of the wind-driven fresh ice is ~20–30 cm, and thus accumulated sea-ice production was 149.319 and 15.186 k m3 during 2017/18, and 187.906 and 15.504 k m3 during 2018/19, respectively.


Evaluation of two CMIP6 sea-ice models using an altimetric satellite emulator

Andrew Roberts, Alice DuVivier, Adrian Turner

Corresponding author: Andrew Roberts

Corresponding author e-mail: afroberts@lanl.gov

Separating sea-ice-model bias from skill in coupled models is a difficult task and has often been confined to comparison with gridded satellite retrievals in previous uses of Coupled Model Inter-comparison Project (CMIP) data. In this study, we have developed a method that determines model bias and skill of CMIP models using an altimetric emulator to match modeled sea-ice thickness against satellite measurements along ground tracks. Rather than evaluating ice thickness directly, we use sea-ice freeboard as a metric in historical CMIP6 ensembles of the Exascale Energy Earth System Model (E3SM) and the Community Earth System Model (CESM) using spring and autumn measurements of the surface topography of the Arctic Ocean by the Geoscience Laser Altimeter System (GLAS) aboard the Ice, Cloud, and Land Elevation Satellite (ICESat). We confine this study to analysis of the simulated Arctic Ocean for the years 2003–08 and demonstrate that model performance derived from modeled freeboard in the same spatiotemporal proximity as GLAS-derived freeboard samples provides an efficient method for quantifying regional and pan-Arctic model bias, as well as correlation- and variance-weighted skill. Our method places clear confidence bounds on the results based on observational uncertainty and model ensemble spread and it also clearly identifies the contribution of snow to model bias and variance. The method is equally applicable to structured (CESM) and unstructured (E3SM) grids and provides distinctly different and more accurate assessments of bias and skill than do gridded sea-ice thickness and extent products using our skill scores. We demonstrate the ability of the emulator to indicate the severity of an erroneous Beaufort Gyre in E3SM version 1, and to tease out the difference in skill between two versions of CESM that differ in the atmospheric component, with very different Arctic sea-ice thickness patterns. The method we present is also applicable to CryoSat-2 and ICESat-2.


A new method for retrieving sea-ice surface elevation and total freeboard height using ICESat-2 data

Jiakui Tang, Hongjie Xie, Alberto M. Mestas-Nuñez, Stephen F. Ackley, Liuxi Tian

Corresponding author: Hongjie Xie

Corresponding author e-mail: Hongjie.Xie@utsa.edu

Sea ice plays a critically important yet highly dynamic role in global climate change. ICESat-2 (launched on 15 September 2018) with the onboard Advanced Topographic Laser Altimeter System (ATLAS) instrument is the most advanced, highest-resolution altimetry instrument ever placed in Earth orbit. ICESat-2 uses the state-of-the-art photon-counting technology, enabling elevation measurements of the highest accuracy. Based on published literature and our experience in processing ICESat-2’s airborne simulator (MABEL) data, a new method is proposed to retrieve sea-ice surface elevation and total freeboard height from the ICESat-2 ATL03 product. The improved three key steps include (1) a new surface-finding method for separating signal photons from background-noise photons; (2) a new sea-surface-type classification algorithm for classifying the sea surface into different types, especially water or leads as the reference to compute the freeboard height; and (3) a wavelet-based surface-topography fitting algorithm for acquiring the accurate surface fluctuations of sea ice. Retrieval experiments have been implemented using MABEL data and an ICESat-2 preliminary sample dataset, and will also be implemented using the soon to be released ICESat-2 data. We will present some results over sea ice off Greenland in the Arctic and the Ross Sea, Antarctica. Previously collected data based on the polynyas, ice production, and seasonal evolution in the Ross Sea (PIPERS) field experiment funded by the US National Science Foundation, and our experience, will be used for a qualitative cross validation.


Mapping the Antarctic marginal ice zone: a physical basis

Alexander Fraser, Jessica Cartwright, Alison Kohout, Alberto Alberello, Nicola Ramm, Alessandro Toffoli, Guy Williams, David Holland

Corresponding author: Alexander Fraser

Corresponding author e-mail: adfraser@utas.edu.au

The Antarctic marginal ice zone (MIZ) is defined as the region of ice where the ice properties are modified by interaction with ocean waves. Operationally, the MIZ is typically given as the region between 15% and 80% sea-ice concentration (SIC), however this definition is primarily based on convenience, not evidence of wave–ice interaction. Using wave-sensing buoys in the Ross and Weddell seas as validation data, we demonstrate that particular scatterometer-based parameters are a much better proxy of Antarctic MIZ than SIC. Using this new proxy, we map the Antarctic MIZ from 2007 until 2018, and present the results. The Antarctic MIZ occupies nearly 6 million km2 at maximum extent. The circumpolar mean MIZ width for April to October (328 km) is found to be wider than when using the SIC-based proxy (277 km), in line with recent work undertaken in East Antarctica. Regional trends are presented in the context of overall sea-ice-extent trends. We also discuss potential applications for Arctic sea ice MIZ mapping.


First-year-sea-ice thickness in the Weddell Sea from microwave backscatter

Alexander Fraser, Nicola Ramm, Jessica Cartwright, Noriaki Kimura, Robert Massom, Takenobu Toyota, Petra Heil

Corresponding author: Alexander Fraser

Corresponding author e-mail: adfraser@utas.edu.au

Large-scale and accurate retrieval of sea-ice volume remains the grand challenge for Antarctic sea-ice remote sensing. Recent work has illustrated the utility of microwave backscatter as a proxy for thickness in regions where ice dynamics is the primary thickness driver (e.g. the Sea of Okhotsk). Backscatter is a proxy measurement for thickness because, in those regions, thickening and (Rayleigh-scale) roughening occur in tandem, and backscatter is particularly sensitive to roughness. Until now, most studies have used synthetic aperture radar imagery to determine backscatter, enabling high-resolution snapshots of this thickness proxy. Here we extend the use of backscatter-related proxies both spatially and temporally by using lower-resolution (but far broader-coverage) backscatter measurements from a scatterometer (the C-band Advanced Scatterometer (ASCAT)), and validate this technique in the Weddell Sea. We use standard regression techniques to statistically relate two backscatter parameters to thickness/freeboard. We find high correlations between these scatterometer parameters and a variety of thickness/freeboard validation datasets including a) ICESat freeboard measurements; b) Operation Icebridge freeboard measurements; c) upward-looking-sonar-equipped moorings and d) ice-parcel-divergence history derived from Lagrangian passive-microwave back-trajectories, indicating its validity in this region. This proxy is found to be particularly accurate for thicker ice, and as such presents a natural complement to established techniques for estimating thin-ice thickness. Further, the ASCAT dataset spans the gap between ICESat1 and ICESat2, enabling validation against both instruments and potentially filling the gap.


Simulating the bi-variate thickness distribution of a sea-ice ridge using discrete elements

Travis Davis, Andrew Roberts, Adrian Turner

Corresponding author: Travis Davis

Corresponding author e-mail: tjdavis1@nps.edu

This paper presents results from a three-dimensional discrete element model of sea-ice ridging that has been developed to investigate the evolution of macro-porosity in first-year pressure ridges. Our work follows an analytic derivation of the sea-ice thickness redistribution function by Roberts et al. (2019; https://doi.org/10.1029/2018MS001395). That work indicated that the thickness distribution g(h) of sea ice needs to be expanded to a bi-variate distribution g(h,Φ) that includes the macro-porosity Φ of sea-ice ridges so as to correctly account for ice-mass evolution of the pack. In the current work, we use the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) to conduct a numerical investigation of one component of Roberts et al. (2019); we investigate the evolution of g(h,Φ) for a single ridge. Observational studies have found that the macro-porosity of ridges can account for as much as 30% of deformed ice volume, independent of the micro-porosity of sea ice from brine and air pockets in the crystalline fabric. The discrete element model we have designed accounts for both the variation in macro-porosity and ice thickness along a ridge line and approximates the dynamics of sea ice by modeling two floes as a series of bonded spheres in a regular lattice. The model identifies three distinct dynamic forcing mechanisms: pairwise interactions of un-bonded elements, bonded element interactions, and ‘macro’-scale field forces, validated against an idealized Timoshenko cantilever. Results from our floe collision experiments indicate a close qualitative approximation of the evolution of a pressure ridge. Our analysis of the model includes a characterization of the ridge properties against submarine observations. We also compare simulated ridges with previous two-dimensional discrete-element models of ridges prevalent in sea-ice literature. With our three-dimensional model, we demonstrate that bi-variate thickness distribution along a ridge line mimics observed Φ values in the 10–30% range along a ridge line and we compare evolution of the bivariate results with the theory of Roberts et al. (2019).


A new method to estimate Arctic sea-ice thickness from CryoSat-2 based on least squares

Feng Xiao, Fei Li, Shengkai Zhang, Jiaxing Li, Tong Geng, Yue Xuan

Corresponding author: Shengkai Zhang

Corresponding author e-mail: zskai@whu.edu.cn

Sea-ice thickness is an important parameter of the polar cryosphere, where changes in its seasonal cycle may cause significant negative feedbacks. Satellite altimeters have been used to monitor Arctic sea-ice thickness since the early 2000s. The altimetric sea-ice thickness retrieval is based in measurements of freeboard, the height of the ice surface above the local sea level, which can be used to calculated ice thickness. The conversion of freeboard to sea-ice thickness depends on the correct knowledge of snow depth and the densities of sea ice and snow. However, those parameters are not very well constrained by observations at basin scale. In this study, we proposed a novel method to convert freeboard to sea-ice thickness using CryoSat-2 data. Pulse peakiness (PP) and stack standard deviation (SSD) parameters were used to distinguish leads from ice floes, and then freeboard was estimated. We separated the freeboard values with a 5 × 5 km grid. We modeled the freeboard as a quadratic function of local ice-surface terrain. The model coefficients in each grid cell were retrieved using a least squares method. By using the model-fit method, we generated estimates of sea-ice thickness for winter (January–March) 2017 and 2018. The Arctic sea-ice thickness during winter 2017 and 2018 is 2.171 m and 2.038 m, respectively. We compared our results with two CryoSat-2 sea-ice-thickness products, namely sea-ice thickness from AWI-CryoSat-2 and near-real-time sea-ice thickness from CryoSat-2 processed by CPOM (Center for Polar Observation and Modelling). The difference in sea-ice thickness between our results and the AWI and CPOM products is 0.402 ± 0.252 m and 0.389 ± 0.235 m, respectively. Good agreements between our results and the two products demonstrate the potential of the proposed method to produce accurate sea-ice thickness estimates.


Exploratory data analysis of first-year sea-ice ridge properties derived from an airborne optical survey of the Beaufort Sea in May 2018

Parnian Rezania, Randall Scharien

Corresponding author: Randall Scharien

Corresponding author e-mail: randy@uvic.ca

Pressure ridges are one of the most dominant topographical surface features of the sea-ice cover. Ridges play a critical component in drag formation and seasonal sea-ice extent, as well as climate change, ecology, shipping and navigation, and human-related activities. Scientific research and observation of sea ice ridges has been developed since the first polar explorations in the 16th century. One consistent source of information on sea-ice ridges is very high spatial resolution (VHR) airborne imagery. VHR images can provide a detailed distribution of sea-ice features and improve the treatment of sea-ice dynamics in high-resolution numerical models. Structure from motion (SfM) is a low-cost, simple and flexible technique for obtaining three-dimensional sea ice topography, including pressure ridges, with an accuracy equivalent to other photogrammetric techniques. In this study, preliminary results from an evaluation of sea-ice pressure-ridge sail heights, derived using SfM technique applied to VHR airborne images acquired from the Beaufort Sea in May 2018, is presented. The SfM technique for sail-height determination is evaluated against other techniques, such as using the length of ridges’ shadows, with the aim of improving techniques for understanding the regional and seasonal distribution of ridges, as well as establishing a realistic model for ridge properties and distribution on the first-year sea ice for the treatment of its dynamics.


Antarctic landfast sea-ice and sub-ice platelet-layer growth rates from a sea-ice mass-balance station

Greg Leonard, Gemma Brett, Inga Smith, Maren Richter, Anne Irvin, Wolfgang Rack, Pat Langhorne

Corresponding author: Greg Leonard

Corresponding author e-mail: greg.leonard@otago.ac.nz

Predicting the response of Antarctic sea ice to a warming world is a complex undertaking as the linkages between enhanced basal melt from cold cavity ice shelves and the thickness and extent of proximal sea ice are not fully understood. The relatively buoyant meltwater is known to flow out from beneath ice shelves at shallow depths where it interacts with a thickening sea-ice cover over the winter growth season. The meltwater is supercooled and can sustain populations of frazil crystals that are deposited at the sea-ice/water interface, thus acting as a conveyor belt that transfers mass from the base of an ice shelf to the surrounding sea ice. The path of the meltwater is influenced by the basal slope of the ice shelf, the stratification of the mixed layer and ambient currents. Hence, the distribution of frazil crystals is not uniform and is known to vary substantially over scales of tens of kilometres. Here we present results from a 2018 winter (8 August–19 November) deployment of a first-of-its-kind sea-ice mass-balance station designed to monitor both sea-ice and sub-ice platelet-layer thickness changes over winter in McMurdo Sound, Antarctica. The station integrates standard sea-ice mass-balance componentry (sea-ice thermistor string and ultrasonic snow sensor) with a single-frequency Geonics EM31 electromagnetic sensor that measures the apparent conductivity of the sea ice and underlying sub-ice platelet layer. Twenty thermistors, the snow sensor and the EM31 are sampled at 10-minute intervals and the data are autonomously sent to New Zealand via a radio telemetry/ftp link in near-real time. The thermistor data are processed to remove sensor drift and then smoothed using a local weighted scatterplot-smoothing (LOWESS) technique. Sea-ice vertical temperature gradients are calculated and used to estimate the thickness of the sea-ice layer. Thickness changes are then tracked over time to determine ice-growth rates. Three-layer (sea ice/snow, sub-ice platelet layer and seawater) forward and inverse modelling techniques are used to simultaneously derive sea-ice and sub-ice platelet-layer thicknesses from the in-phase and quadrature components of the EM31 signal. The novel integration of the EM31 sensor allows, for the first time, estimates of snow depth, sea-ice thickness and sub-ice platelet-layer thickness to be made autonomously at a single site over a winter growth season.


Impact of sea-ice thickness on the predictability of the Arctic sea ice based on the APPOSITE data with climate model MIROC

Jun Ono, Yoshiki Komuro, Hiroaki Tatebe

Corresponding author: Jun Ono

Corresponding author e-mail: jun.ono@jamstec.go.jp

To clarify the mechanisms contributing to skillful predictions of Arctic sea ice with seasonal-to-interannual timescales, we have conducted research using the Model for Interdisciplinary Research on Climate (MIROC) 5.2. The horizontal resolution of the atmospheric component is a T42 spectral truncation (about 300 km), and there are 40 vertical levels up to 3 hPa. The oceanic component had 1° longitudinal grid spacing in the spherical coordinate portion south of 63° N. The meridional grid spacing varies from about 0.5° near the equator to 1° in the mid-latitudes. There are 63 vertical levels, the lowermost level of which is located at the 6300 m depth. The sea-ice component of MIROC5.2 implements one-layer thermodynamics, elastic–viscous–plastic rheology, subgrid ice-thickness distribution with five-category, and linear-remapping. Snow cover on sea ice in the model affects the thermodynamical processes through changes in the surface albedo and vertical heat fluxes. In the present study, we mainly focus on the role of sea-ice thickness on the predictability of various variables, including sea-ice extent, based on the perfect model ensemble experiments under the Arctic Prediction and Predictability on Seasonal-to-Interannual Time Scales (APPOSITE) project. Although this study includes preliminary results, there are three key points, as follows: (1) Initialized sea-ice thickness in April significantly contributes to skillful forecast of the September sea-ice extent, but not to sea-ice volume. (2) The Pacific sector of the Arctic Ocean is thought to be the key area for skillful forecasts. (3) Near-surface ocean and atmosphere variables are partly influenced in the first several lead months by the initialized sea-ice thickness. The results for additional experiments will be introduced during the presentation.


Investigation of coastal landfast sea ice in the East Siberian Sea for winter 1995/96–2013/14

Mengxi Zhai, Bin Cheng, Matti Leppäranta, Xiao Cheng, Fengming Hui, Xinqing Li, Yawen Huang

Corresponding author: Xiao Cheng

Corresponding author e-mail: polecx@163.com

The average sea-ice thickness in the Arctic Ocean has been reducing considerably in the past 30 years. The most dramatic consequences of the Arctic warming have been the decrease in sea-ice extent, sea-ice concentration, and sea ice thickness, as well as the duration of the ice season. In this study, we investigate thermodynamic growth of landfast sea ice in the coastal zone of the East Siberian Sea. The study period covers winter seasons 1995/96–2013/14. The MODIS images from 2002–14 were applied to detect the ice fracturing process. A one-dimensional snow and ice thermodynamic model was employed to calculate snow depth and ice thickness. The model was forced by meteorological observations from the Russian weather station at Kotel’ny Island. We focus on interannual variations, in particular the role of precipitation on snow, sea-ice mass balance, onset of snow and ice melting, and ice breakup. The melting process is the most challenging question since that is sensitive to the evolution of the optical properties of ice and snow. The modelled interannual average maximum ice thickness was 2 m and showed a decreasing trend of about 0.1 m per decade. The measured snow depth showed large internal variability with no significant trend during the investigation period. The modelled onset of snowmelt was in good agreement with observations. Landfast ice decay dates can vary by as much as 1 month. The total ice-breakup process can be explained by a combination of the ice fracturing and melting process. The results show that the breaking up of landfast ice is a coupled thermal–mechanical process and warrants particular concern in large-scale sea-ice modelling in the Arctic Ocean.


The sub-ice platelet layer in a one-dimensional thermodynamic sea ice model

Pat Wongpan, Martin Vancoppenolle, Pat Langhorne, Inga Smith, Gurvan Madec, Alex Gough, Andrew Mahoney, Tim Haskell

Corresponding author: Pat Wongpan

Corresponding author e-mail: wongpan@lowtem.hokudai.ac.jp

Landfast sea ice grows due to conductive heat losses to the atmosphere. Near continental ice shelves, where ice shelf water (ISW) exists at the ocean surface, fast ice also thickens because of interaction with the supercooled water column. This ISW is a result of the ice-shelf–ocean interaction and creates specific sea-ice forms: the porous, friable sub-ice platelet layer and incorporated platelet ice. However, the large-scale distribution and seasonality of platelet ice are not well documented, which is where model representations may help to progress the understanding of their role in the functioning of the Southern Ocean. In this work, we introduce a representation of platelet-ice processes by analysing mushy-layer physics emerging from a one-dimensional sea-ice model (the one-dimensional Louvain-la-Neuve Sea Ice Model: LIM1D). We evaluate the approach by forcing LIM1D with meteorological observations and prescribed oceanic heat flux based on observations from an over-winter study in 2009 on the landfast sea ice of McMurdo Sound, Antarctica. We also evaluate the response of the simulated sub-ice platelet layer to oceanic heat flux, by comparison with observation-based retrievals from a ~20 km transect in November 2009. Sub-ice platelet layers several meters thick are observed and simulated. Analysis of model results suggests that the high liquid fraction of the sub-ice platelet layer implies a low thermal conductivity and a high specific heat. If the imposed heat loss to the ocean becomes large, the conductive heat flux through the sub-ice platelet layer decreases, which thermodynamically decouples the ocean from the sea ice. Sensitivity experiments revealed that deep snow intensifies this effect. In summary, a realistic sub-ice platelet layer emerges in a one-dimensional thermodynamic sea ice model when suitably forced. Key factors to formation are sufficient heat loss to underlying water and thermal insulation provided by the ice and snow. The sub-ice platelet layer is stabilized by its liquid content, yielding low thermal conductivity and high heat capacity. Ultimately, the model will not only help us to understand the coupled heat and salt transfer of the sub-ice platelet layer and how this controls the biogeochemical specifics of Antarctic fast ice, but also allow upscaling of these processes to the scale of the Southern Ocean.


Using under-ice spectra to determine landfast sea-ice-algal biomass in Saroma-ko lagoon, Hokkaido, Japan

Pat Wongpan, Daiki Nomura, Takenobu Toyota, Tomonori Tanikawa, Tomomi Ishino, Tetsuya Tamura, Manami Tozawa, Toru Hirawake, Atsushi Ooki

Corresponding author: Pat Wongpan

Corresponding author e-mail: wongpan@lowtem.hokudai.ac.jp

Landfast sea ice is a key component of coastal ecosystems in polar regions, providing a habitat for ice-algal communities. The first-year to multiyear ice ratio in the Arctic is increasing and sea ice is thinning towards the predicted summer ice-free Arctic in this century. To date the estimation of algal biomass by satellites has only applied to the unfrozen ocean. This study examines the relationships between the normalized difference indices calculated from under-ice hyperspectral measurements, and ice-algal biomass for landfast first-year ice in Saroma-ko lagoon, Hokkaido, Japan where sea ice is thin (~0.5 m). We analyze physical properties of snow and ice supporting our 27 paired in-situ optical and biological measurements along transect lines across multiple scales covering over 250 × 250 m in February 2019. Our new observation-based algorithms can be applied to non-invasively estimate landfast ice-algal biomass which will fill the gap of monitoring algal biomass under the ice cover during winter and winter–spring transition for thin first-year ice. Together with the ocean color remote sensing, our algorithms will help improve the understanding of the temporal and spatial variability of algal biomass using moorings and underwater vehicles focusing with the thin Arctic sea-ice scenario.


Snow-related variability of spectral light transmittance of Arctic first-year ice in the Lincoln Sea

Philipp Anhaus, Christian Katlein, Marcel Nicolaus, Arttu Jutila, Christian Haas

Corresponding author: Philipp Anhaus

Corresponding author e-mail: Philipp.Anhaus@awi.de

Light transmittance through Arctic sea ice and snow has an important impact on both the ocean heat content and the ice-associated ecosystem. The partitioning of the radiation is a key factor of the mass and energy balance of Arctic sea ice. It is therefore crucial to measure sea-ice transmittance and understand which parameters determine its variation on temporal and spatial scales. Ice and snow imprint characteristic features in the spectral shape of transmitted light. Transmitted spectral irradiance was recorded at the underside of levelled landfast first-year ice (FYI) in a refrozen lead using a hyper-spectral radiometer mounted on a remotely operated vehicle (ROV) during the Last Ice Area campaign off Alert in the Lincoln Sea in May 2018. The main benefits of using the ROV are large spatial coverage in comparably short survey times and non-destructive measurements under sea ice. Snow depth was obtained using a Magna probe and a terrestrial laser scanner measured the surface topography. The total ice thickness was recorded with a ground-based electromagnetic induction sounding device while an upward-looking single-beam sonar also mounted on the ROV recorded ice draft. This unique co-located dataset enables us to categorize groups of spectral transmittances. Due to the relatively constant FYI thickness it was possible to separate the spectral effect of snow depth on the light transmittance. Further we discuss how to retrieve snow depth and ice thickness based only on spectral transmittance data by developing a new observation-based inverse algorithm. Three methods are envisioned: First, to fit a multiplicative exponential function to the spectra which includes wavelength-dependent extinction coefficients of snow and sea ice. Second, to follow a statistical approach using normalized difference indices (NDIs) to construct spectral correlation coefficients between the NDIs with snow depth and ice thickness. Third, to generate synthetic spectra from snow depth and ice thickness using the radiative transfer model AccuRT and compare those with the observed spectra. Expected results are accurate snow depth and sea-ice thickness (as well as melt-pond depth and coverage).


High production going along with high respiration: impact of biofilm formation for sea-ice biogeochemistry

Florian Deman, Arnout Roukaerts, Jean-Louis Tison, Bruno Delille, Frank Dehairs, François Fripiat

Corresponding author: Florian Deman

Corresponding author e-mail: florian.deman@vub.be

While representing less than 5% of the total ice cover around Antarctica, landfast sea ice is nevertheless an important habitat known to exhibit high biomass levels at the ocean/ice interface, with particulate organic carbon (POC) concentrations easily reaching 2000 μmol C L–1 during spring bloom. Surprisingly, together with the POC increase in bottom ice, fieldwork measurements performed in East Antarctica (Adélie Land 2011, McMurdo Sound 2012, Prydz Bay 2015) of nitrate and phosphate concentrations report a simultaneous increase with concentrations exceeding those of underlying seawater, suggesting an intense remineralization and nitrification processes within the ice. This goes against the classic view of nutrients being consumed during the growth season and regenerated after the height of the bloom. Regardless of the high nitrate levels available in the ice, increasing total nitrogen concentrations also suggest still more nitrogen from the underlying seawater was brought into the ice. Results of a NPZD-model indicates that a second nutrient pool, in addition to the brine pool, is essential to successfully model and reproduce field observations. The presence of a biofilm attached to the ice walls could act as a water-retaining substrate forming microenvironments with chemical gradients within the brine channels. The effect of biofilm on nitrogen dynamics (concentration and isotopic composition) in sea ice will be discussed as well as potential implications for other parameters (phosphate, carbon, oxygen). This calls for the integration of the biofilm concept into the current view of sea-ice biogeochemistry.


What is the impact of brine conditions on biogenic DMSP and DMSO production? A cell-culture approach

Boris Wittek, Gauthier Carnat, Jean-Louis Tison, Nathalie Gypens

Corresponding author: Boris Wittek

Corresponding author e-mail: boris.wittek@outlook.com

Sea ice is an extreme environment known to host microbial communities which produce dimethylsulfoniopropionate (DMSP) and dimethylsulfoxyde (DMSO), two biogenic precursors of the climate cooling gas dimethylsulfide (DMS). Despite decades of research, drivers and pathways of the sea-ice DMS cycle remain largely unknown. This study quantifies for the first time the production of the metabolites DMSP and DMSO by the diatom Fragilariopsis cylindrus and the prymnesiophycea Phaeocystis antarctica under changes of temperature and salinity typically encountered in the sea-ice brine habitat. Salinity 75 and salinity 100 experiments suggest an osmolyte function of both dimethyled sulfur compounds in the diatom cell. A stronger salinity shift to 150 induces osmotic shock and ultimately cell death. Decreases in temperature combined with increases in salinity reveal similar trends and suggest that the cryoprotectant function of DMSP and DMSO is not relevant in our cultures. Through this study, we improve our knowledge and the modelling capabilities of the sea-ice DMS cycle.


Towards pan-Antarctic polynya monitoring using MODIS through improved nighttime cloud-cover detection

Stephan Paul, Marcus Huntemann

Corresponding author: Stephan Paul

Corresponding author e-mail: stephan.paul@lmu.de

Coastal polynyas are recurring areas of thin ice and open water with substantial impact on sea-ice production, deep-water formation and gas ventilation of the ocean, which are generally formed by divergent ice motion due to strong offshore winds or ocean currents. Several studies highlight the capability of MODIS-derived polynya monitoring to complement and even extend existing passive-microwave-derived estimates in both the Arctic and Antarctic due to their high spatio-temporal resolution. By utilizing a combination of normalized multi-channel thermal-infrared MODIS data, information from image texture analysis and swath-to-swath change detection, a machine-learning algorithm was trained to identify cloud cover over Antarctic sea ice. For the initial training of the classifier, we use a mix of manually classified visible/near-infrared training data from daytime conditions as well as nighttime data. Hence, the identification works despite the low thermal contrast between polar clouds and the surface as well as the lack of visible and near-infrared channels during nighttime conditions. Validation with Sentinel 1A/B SAR data highlights the substantial underestimation of thin-ice and open-water areas for thin-ice thickness estimates when compared to estimates based on the standard MOD/MYD29 ice-surface-temperature product. This has substantial implications for subsequent calculation of polynya area and ice-production rates. Furthermore, a more accurate classification of polar cloud cover offers new possibilities for, for example, year-round fast-ice monitoring through persistent absence of leads, and ice-motion measurements during nighttime through better initial detection and separation of clouds as well as lead detections under non-opaque clouds during daytime conditions, in addition to the established long-term monitoring of polynyas. For the future, this approach might also be applicable to continuation missions such as the European Commission’s Copernicus programme Sentinel 3A/B SLSTR sensor.


Mooring observation of underwater interaction between frazil ice and sediment in an Arctic polynya

Masato Ito, Kay Ohshima, Yasushi Fukamachi, Daisuke Hirano, Andrew Mahoney, Joshua Jones, Toru Takatsuka, Hajo Eicken

Corresponding author: Masato Ito

Corresponding author e-mail: itoh-m@ees.hokudai.ac.jp

Incorporation, transport and release of sediment by sea ice potentially have an important role in the bio-related material cycle in polar oceans. In the Arctic Ocean, sea ice containing much particulate matter (dirty ice) is widespread from coastal to offshore regions. Although the materials inside sea ice are considered to be the ocean bottom sediment, the process of sediment incorporation into sea ice has not been well understood. To reveal this process, we analyze the mooring data obtained in the polynya region in the northeastern Chukchi Sea, with the help of the marine radar and meteorological data. In the northernmost part of the polynya, which is formed along the Alaskan coast, mooring observations have been conducted since 2009 through a collaboration between Hokkaido University and University of Alaska Fairbanks. For the observation of 2014/15, the mooring was deployed at a point with the water depth of 45 m and equipped with an ice-profiling sonar (IPS), conductivity-temperature recorder, an acoustic Doppler current Profiler (ADCP) and a turbidity sensor. For this winter, episodic polynya events lasting for 1 week were identified by the IPS and radar data. During these periods, in turbulent conditions, potential and in-situ supercooling occurred at the depth of 32 m. At these timings, the ADCP detected the surface-intensified acoustic signals from the surface down to 30 m. These signals were enhanced with strong winds. All the facts above strongly suggest that those signals were caused by underwater frazil ice. Frazil ice was continuously formed in the water column over 1 week, and then the open water was maintained. This is the most efficient ice production in a polynya. At the same time of frazil-ice detection, the ADCP also showed another acoustic signal intensified near the ocean bottom. These signals were detected under the strong current of >0.4 m s–1. Moreover, for these periods, turbidity at a depth of 35 m was clearly increased. Thus, those acoustic signals were likely caused by the re-suspended ocean bottom sediment. Frazil ice and sedimentary particles were detected simultaneously in the overlapping range, implying their direct contact. This fact indicates that frazil ice captures re-suspended sediment through underwater interaction and incorporates it associated with consolidation. Thus, we propose that suspension freezing in polynyas can be a major process of sediment incorporation into sea ice.


Decaying sea ice and its capacity for physical and biogeochemical influence at the ocean–ice interface off East Antarctica

Matthew Corkill, Alex Fraser, Petra Heil, Eva Cougnon, Cristina Genovese, Julie Janssens, Sebastien Moreau, Alessandro Silvano, Delphine Lannuzel

Corresponding author: Matthew Corkill

Corresponding author e-mail: matthew.corkill@utas.edu.au

The decline to minimum Antarctic sea-ice extent during austral summer coincides with decay of individual floes. The state of decay of a sea-ice floe determines its local physical and biogeochemical influence as well as its potential for future influence along its trajectory. In East Antarctica, few observations have been made of decayed, rafted and therefore relatively inhospitable ice floes; thus little is understood about the influence of these floes on chemical and biological processes. Seven sea-ice stations were undertaken on decaying ice (including ex-fast ice) off the Adélie/George V Land coast, East Antarctica in austral summer 2016/17. Temperature and salinity data from these stations were used to assess porosity. Porosity was used here to determine sea-ice permeability; i.e. how connected the atmosphere, sea ice and the underlying water column are. Stratigraphy, isotopic composition (δ18O) and Antarctic Sea Ice Processes and Climate (ASPeCt) underway observations were combined with satellite derived sea-ice motion to evaluate the inner structure, dynamic growth, origins and trajectories of these floes. Our data show that all cores were highly permeable, well above the thresholds conventionally used for columnar and granular ice. The advanced seasonal stage of these ice floes is further evidenced by the extremely low nutrients concentrations and high level of heterotrophic activity. Growth patterns indicated some dynamic growth via rafting as well as the presence of platelet ice originating from ice-shelf meltwater. Back and forward trajectories showed a general pattern of greater movement further offshore with some anomalous movement around the Mertz Glacier. These results provide valuable insights into the structure of decayed East Antarctic sea ice and its capacity to act as a vector of physical and biogeochemical influence. The data are of particular value for modelling-study validation, to evaluate how the predicted physical changes to the sea-ice scape are likely to affect the biogeochemistry and associated ecosystems of the East Antarctic sector.


Using biomass accumulation to estimate Antarctic sea-ice primary production

Florian Deman, Arnout Roukaerts, Vancoppenolle Martin, Jean-Louis Tison, Bruno Delille, Frank Dehairs, François Fripiat

Corresponding author: Florian Deman

Corresponding author e-mail: florian.deman@vub.be

With an extent varying between a maximum of 19 × 106 km2 in late winter and a minimum of 3 × 106 km2 in late summer, Antarctic sea ice is one of the largest ecosystems on Earth, most of which consists of annual pack ice. Primary production in-situ measurements in Antarctic sea ice, using either oxygen-based or tracer incubation methods, are relatively tricky to achieve and remain scarce. Thus, to estimate large-scale Antarctic sea-ice primary productivity, two approaches have been used. First, the use of sea-ice biogeochemical models suggest that Antarctic pack ice contributes to a small but significant fraction (10–28%) of the primary production in the ice-covered area of the Southern Ocean. Second, accumulation of organic matter trapped within sea ice during the growth season is likely to be representative of the net community production. More than 20 years ago, Legendre et al. (1992) used the few available observations to infer Antarctic sea-ice primary productivity. We believe that it is time to revisit this estimation by accounting from a much larger compilation of data (historical to present). Here, we present the first results using an updated dataset of historical ice cores sampled between 1989 and 2017 (± 400 pack-ice cores). These allow us to provide an updated estimation of the sea-ice primary production based on in-situ data, and its contribution to the SIZ and Southern Ocean. A comparison between pack and fast ice (± 110 fast- ice cores) will be also briefly discussed.


Arctic warming interrupts the transpolar drift and affects long-range transport of sea ice and ice-rafted matter

Thomas Krumpen, Jakob Belter, Antje Boetius, Ellen Damm, Christian Haas, Stefan Hendricks, Marcel Nicolaus, Eva-Maria Noethig, Ilka Peeken

Corresponding author: Thomas Krumpen

Corresponding author e-mail: tkrumpen@awi.de

Sea ice is an important transport vehicle for gaseous, dissolved and particulate matter in the Arctic Ocean. Due to the recently observed acceleration in sea-ice drift, it has been assumed that more matter is advected by the transpolar drift from shallow shelf waters to the central Arctic Ocean and beyond. However, this study provides first evidence that intensified melt in the marginal zones of the Arctic Ocean interrupts the transarctic conveyor belt and has led to a reduction in the survival rates of sea ice exported from the shallow Siberian shelves (–15% per decade). As a consequence, less and less ice formed in shallow water areas (< 30 m) has reached Fram Strait (–17% per decade), and more ice and ice-rafted material is released in the northern Laptev Sea and central Arctic Ocean. Decreasing survival rates of first-year ice are seen all along the Russian shelves, but significant only in the Kara Sea, East Siberian Sea and western Laptev Sea. Identified changes affect biogeochemical fluxes and ecological processes in the central Arctic: reduced long-range transport of sea ice alters transport and redistribution of climate-relevant gases, and increases accumulation of sediments and contaminants in the central Arctic Ocean, with consequences for primary production and the biodiversity of the Arctic Ocean.


The color of ice

Chris Petrich

Corresponding author: Chris Petrich

Corresponding author e-mail: christian.petrich@norut.no

Color is an artistic application of the optical properties of ice. Color drives science. There is ample anecdotal evidence that people interacting with ice hypothesize about ice properties based on visual appearance, including color. The color of ice can be understood as a combination of incoherent scattering and absorption. In homogeneous systems color can therefore be related to geophysical properties. It is shown that the characteristic color of ice is blue and how hue varies with ice properties. Patterns and challenges arising from inhomogeneity are illustrated. The color of ice lies within the gamut of conventional display systems and is therefore readily conveyed for appreciation by a broad audience.


Combining satellites and models to map pan-Arctic under-ice light availability

Julienne Stroeve, Glen Liston

Corresponding author: Julienne Stroeve

Corresponding author e-mail: j.stroeve@ucl.ac.uk

Sea-ice thickness is a critical variable, both as a climate indicator and for forecasting sea ice conditions on seasonal and longer time scales. The lack of snow depth and density information is a major source of uncertainty in current thickness retrievals from laser and radar altimetry. Snow and ice thickness in turn influence how much light can enter the Arctic Ocean, impacting primary productivity. In this study we present our newly developed long-term daily snow depth and density product, discussing its accuracy and uncertainties. We then apply the new dataset to improve sea-ice-thickness retrievals and estimate how combined thickness and snow-depth changes are impacting the under-ice light climate. Implications on under-ice algae and phytoplankton are discussed.


Analysis of changes in the Arctic marginal ice zone, using observations and models

Rebecca Rolph, Daniel Feltham, David Schroeder

Corresponding author: Rebecca Rolph

Corresponding author e-mail: r.rolph@reading.ac.uk

We have developed a sea-ice–ocean model with a floe size distribution and use this to interpret and analyse changes in the seasonal marginal ice zone (MIZ) within the last three decades. Here, we define the MIZ as being that region of the Arctic Ocean covered in sea ice where the sea-ice-area fraction lies between lower and upper limits, rather than the more traditional definition based on the role of ocean waves in sea-ice evolution. We quantify monthly-averaged changes in the extent of the MIZ and the sea-ice area within the MIZ from 1980 until the present. We use this, and other data including sea-ice thickness from CryoSat-2, to calibrate and assess our model. We use our model to estimate the proximate causes of the changes we are seeing in the seasonal MIZ in order to gain insight into which processes, such as wave-induced breakup, melting, and advection, may become more dominant in controlling MIZ dynamics in the coming decades.


IceBird: a pan-Arctic airborne sea ice observation system

Stefan Hendricks, Thomas Krumpen, Christian Haas, Gerit Birnbaum

Corresponding author: Stefan Hendricks

Corresponding author e-mail: shendricks@awi.de

We present an Arctic sea-ice observation system that focuses on unique direct observations of sea-ice-plus-snow thickness and other sea-ice-surface properties in both winter and summer. The program is named the IceBird campaign series and key activities in the field include the use of long-range polar-research aircraft and helicopter operations from research icebreakers and bases on land. Data collections are based on electromagnetic induction sounding and consistent time series are available in key regions of the Arctic Ocean since 2001. The increased use of polar-research aircrafts in recent years has resulted in several initiatives that aim for long-term observations of ice thickness during seasonal minimum and maximum sea-ice extent in the Arctic. The scientific payload of the research aircraft of type Basler BT-67 and its capability to fly low-altitude surveys makes it an ideal tool for the validation and on-going verification of various satellite remote sensing products. The availability of airborne sea-ice-thickness information spans the periods of generations of satellite altimeter missions, from Envisat and ICESat-1 to CryoSat-2, the Copernicus Sentinel-3 mission and ICESat-2. Wherever possible, the airborne surveys are accompanied by in-situ observations on the ice surface to compile a hierarchy of validation data from local to basin scales. Results of the observation network have found broad use for studying inter-annual variability and changes of sea-ice thickness as well as the validation of satellite data products. We identify a gap of observations over the multiyear sea ice zone during the melt season and early freeze-up. We also stress the need for the continuation of a coordinated observational program that has produced a time series of sea-ice thickness only paralleled by submarine ULS observations.


Joint MOSIDEO/CIRFA tank experiments on behavior and detection of oil in ice

Chris Petrich, Megan O’Sadnick, Camilla Brekke, Marianne Mynes, Sönke Maus, Martina Salomon, Sofie Woelk, Tom Grydeland, Rolf-Ole Jenssen

Corresponding author: Chris Petrich

Corresponding author e-mail: christian.petrich@norut.no

In the Arctic, the presence of sea ice presents a challenge to safe and sustainable operations. To optimize planning and minimize impact of inadvertent oil spills, oil-in-ice experiments were performed at the HSVA Arctic Environmental Test Basin (AETB). Following an under-ice spill and simulated springtime warming, the microscopic movement and distribution of oil in the sea-ice pore space, as well as the detectability of oil as it approaches the surface, were investigated. Two ice types were studied simultaneously, i.e. columnar ice with and without a granular ice surface layer. Granular ice was generated by blowing air against a barrier. Among the detection techniques were electromagnetic (radar, tomographic SAR) and optical (fluorescent, hyperspectral, thermal) sensors, and microscopic distribution of oil in sea ice was determined through X-ray computed tomography (CT). It was found that the movement of oil differed considerably between the investigated ice types. Predicting the behavior of oil in ice based on environmental conditions will help optimize the approaches used in spill detection and response.


Mapping of sea-ice types from Sentinel–1 considering the surface-type-dependent effect of incidence angle

Johannes Lohse, Anthony P. Doulgeris

Corresponding author: Johannes Lohse

Corresponding author e-mail: johannes.p.lohse@uit.no

As a consequence of declining sea-ice area and extent, operational activity and marine traffic in the Arctic are increasing. High-resolution sea-ice charts can assist in ship route planning and tactical navigation, and thus help to ensure the safety of Arctic operations. Furthermore, ice-chart information can be assimilated into numerical sea-ice and climate models to improve forecasts. Because of its independence of daylight and weather conditions, space-borne synthetic aperture radar (SAR) provides an excellent tool for sea-ice observations at large scales with sufficiently high spatial resolution. Today, sea-ice charts in operational services are largely produced manually by ice analysts. With new satellite missions being launched and an increasing number of images available, this manual approach needs to be supplemented by reliable methods for automatic or semi-automatic mapping of sea-ice conditions. However, automatic pixel-based sea-ice type classification from Sentinel–1 wide-swath images faces issues, e.g. the incidence-angle dependence of the backscattered signal (1) or the variable noise-floor in the data (2). In this study, we present an approach to overcome the incidence-angle dependence by using a Bayesian classifier with multi-variate Gaussian probability density functions with a non-stationary mean. The constant mean value in the traditional Gaussian classifier is replaced with a linearly varying mean value for the backscatter intensities in decibels. The slope of this linear variation differs between classes and can be directly estimated from the training data or prescribed according to values from previous studies. A projection of the pixels along the estimated slopes results in Gaussian distributions with smaller variance and better between-class separability. To generate training data, we have used 85 Sentinel–1 Extra Wide mode images with overlapping optical remote-sensing data. We have trained up to eight different ice classes and classified Arctic-wide imagery. The resulting charts show open water and different ice types at the pixel-level. Achieving such reliable ice-type maps from Sentinel–1 SAR data will then allow secondary products, such as sea-ice concentration or lead fraction, which are highly relevant to science and industry.


Evaluation of six atmospheric reanalyses over Arctic sea ice from winter to early summer

Robert Graham, Lana Cohen, Nicole Ritzhaupt, Benjamin Segger, Rune G. Graversen, Annette Rinke, Von P. Walden, Mats A. Granskog, Stephen R. Hudson

Corresponding author: Mats A. Granskog

Corresponding author e-mail: mats@npolar.no

This study evaluates the performance of six atmospheric reanalyses (ERA-Interim, ERA5, JRA‑55, CFSv2, MERRA‑2 and ASRv2) over Arctic sea ice from winter to early summer. The reanalyses are evaluated using observations from the Norwegian young sea-ice campaign (N‑ICE2015); a 5-month ice drift in pack ice north of Svalbard. N‑ICE2015 observations include surface meteorology and vertical profiles from radiosondes, as well as radiative and turbulent heat fluxes. The reanalyses simulate surface analysis variables well throughout the campaign, but have difficulties with most forecast variables. Winter (January–March) correlation coefficients between the reanalyses and observations are above 0.90 for the surface pressure, 2 m temperature, total column water vapor, and downward longwave flux. However, all reanalyses have a positive winter 2 m temperature bias, ranging from 1–4°C, and negative (i.e. upward) net longwave bias of 3–19 W m–2. These biases are associated with poorly represented surface inversions and are largest during cold-stable periods. Notably, the recent ERA5 and ASRv2 have some of the largest temperature and net longwave biases, respectively. During spring (April–May), reanalyses fail to simulate observed persistent cloud layers. Therefore they overestimate the net shortwave flux (5–79 W m–2) and underestimate the net longwave flux (8–38 W m–2). Promisingly, ERA5 provides the best estimates of downward radiative fluxes in spring and summer, suggesting improved forecasting of Arctic cloud cover. All reanalyses exhibit large negative (upward) residual heat flux biases during winter and positive (downward) biases during summer. Turbulent heat fluxes over sea ice are simulated poorly in all seasons.


Subseasonal sea-ice prediction at both poles

Lorenzo Zampieri, Helge F. Goessling, Thomas Jung

Corresponding author: Lorenzo Zampieri

Corresponding author e-mail: lorenzo.zampieri@awi.de

With retreating sea ice and increasing human activities comes a growing need for reliable sea-ice forecasts up to months ahead. We exploit the subseasonal-to-seasonal (S2S) prediction database and provide a thorough assessment of the skill of operational forecast systems in predicting the location of the Arctic and Antarctic sea-ice edges on these time scales. This study employs the Spatial Probability Score, a probabilistic verification metric specifically designed to capture the correctness of the sea ice edge position. Our verification methodology goes beyond the classical sea-ice extent and area, and tries to provide a sea-ice forecast description that could be valuable for planning shipping operations. We find large differences in skill between the systems, with some showing a lack of predictive skill even at short weather time scales, and the best producing skillful Arctic forecasts more than 1.5 months ahead. We assess the forecast skill in both hemispheres, thereby showing that prospects for subseasonal sea-ice predictions are promising, especially for Arctic late summer forecasts. To fully exploit this potential, it will be imperative to reduce systematic model errors and develop advanced data assimilation capacity. Furthermore, the relatively long time-span of the S2S prediction database – which is more than 20 years for some of the considered forecast systems – allows us to present some considerations about the changes in predictive skills as the sea-ice extent and volume decreases.


Impact of sea-ice thickness distribution on summer ice melt: comparing CryoSat-2 data and simulations with the ocean–sea-ice model NEMO-CICE

David Schroeder, Danny Feltham, Michel Tsamados, Rachel Telling

Corresponding author: David Schroeder

Corresponding author e-mail: D.Schroeder@reading.ac.uk

Estimates of Arctic sea-ice thickness are available from the CryoSat-2 radar altimetry mission during the ice-growth seasons since 2010. We derive the sub-grid-scale ice-thickness distribution (ITD) with respect to five ice-thickness categories used, for instance, in the sea-ice component CICE of HadGEM3 climate simulations: (1) ice thickness h < 60 cm, (2) 60 cm < h < 1.4 m, (3) 1.4 m < h < 2.4 m, (4) 2.4 m < h < 3.6 m, (5) h> 3.6 m. This allows us to verify the simulated cycle of mean ice thickness and of each thickness category. A default simulation with the ocean–sea-ice model NEMO-CICE underestimates the mean ice thickness in the central Arctic. We can identity the underestimation of winter ice growth being responsible and show that increasing the ice conductive flux for lower temperatures (bubbly brine scheme) and accounting for the loss of drifting snow results in the simulated sea-ice growth being more realistic. However, large differences remain regarding the annual cycle of ITD. According to CryoSat-2, most of the thick ice (h > 3.6 m) present in April does not survive the summer in the central Arctic, as shown by low fraction values in October and November. In the simulations, a large fraction of the thick ice does survive the summer, resulting in a weaker annual cycle. Calculating ocean–ice heat-transfer coefficients individually for each ice-thickness category has the potential to narrow the gap between CryoSat-2 and our model. This has an impact on summer ice melt and concentration and thus on the strength of the positive albedo feedback mechanism that is crucial for climate projections.


Snow on Arctic first-year sea ice: a geophysical complexity to accurately estimate sea-ice thickness from radar altimetry

Vishnu Nandan, Torsten Geldsetzer, Randall Scharien, John Yackel

Corresponding author: Vishnu Nandan

Corresponding author e-mail: vishnu@uvic.ca

The Arctic is on the path to a new climate regime influenced by thinner first-year ice (FYI). With recent Arctic amplification of warming, Arctic sea ice has been experiencing increasing atmospheric moisture transport, frequent rain on snow, and/or melt/refreeze events. These changes have affected the annual snow-covered FYI thermodynamic regime, with the traditional late-winter Arctic snow on sea ice becoming increasingly warmer, and becoming more saline and complexly layered. Such snow, with highly-dense, compacted wind slabs, ice lenses and crusts, will significantly impact the accuracy of FYI freeboard and thickness retrievals from Ku-band radar altimeters such as the ERS-1/2 RA, ENVISAT RA-2, CryoSat-2 and Sentinel-3. A vertical shift in the Ku-band radar scattering horizon caused by the presence of warm, saline and complexly ;ayered snow cover can lead to an overestimate of FYI thickness from radar-altimeter retrieval approaches utilizing that signal to determine the FYI freeboard and invert thickness. In this study, snow property measurements from selected locations in the Canadian and the Norwegian Arctic are used to evaluate the impact of warm/saline/complexly layered snow on Ku-band propagation through snow-covered FYI. Estimates of the main radar scattering horizon, using a semi-empirical approach, are then used to assess potential errors in Ku-band-derived FYI freeboard and thickness. Using the hydrostatic equilibrium condition, we calculate the difference between the Ku-band modeled FYI thickness and in-situ drill-hole FYI thickness measurements. We find that saline snow under both cold and warm conditions shifts the Ku-band main scattering horizon from the snow/ice interface by ~7 cm, causing thickness-etrieval errors. The largest error of up to 150% is found from warm, saline snow covers overlaying thin sea ice observed in the Canadian Arctic regions; the error decreases with an increase in sea-ce thickness. Similar errors are observed from sites in the Norwegian Arctic, where thicker snow cover induces negative sea-ice freeboard and causes highly saline snow-ice formations. Results indicate a potential overestimation in FYI thicknesses by up to 60%, where snow has high density (~450 kg m–3) and is complexly layered, with ice lenses and crusts. Our study recommends incorporating snow microstructure in the radar-scattering horizon models to accurately retrieve FYI freeboard and thickness from radar altimetry.


Variability and trends of polynya ice production in the Arctic between 2002 and 2018

Andreas Preußer, Kay I. Ohshima, Sascha Willmes, Günther Heinemann

Corresponding author: Andreas Preußer

Corresponding author e-mail: preusser@uni-trier.de

Polynyas in the Arctic shelf seas are an integral part of the winter sea-ice cover, as the presence of open water and thin ice promotes intense heat fluxes and consequently the formation of new sea ice. This process represents an important aspect of atmosphere–sea-ice–ocean interactions. An accurate quantification of sea-ice production requires detailed knowledge about the thin-ice thickness distribution within a polynya, which can be achieved by using different satellite remote-sensing approaches. This study features a high-resolution (2 km) MODIS thermal infrared satellite dataset with spatial and temporal characteristics of 17 coastal polynya regions over the entire Arctic basin for 2002/03–2017/18. A similar dataset based on lower-resolution (6.25 km) AMSR-E passive-microwave satellite data is used for comparisons over a 9-year overlapping period (up to 2010/11). The MODIS approach employs a 1-D energy-balance model, where quasi-daily thin-ice thickness composites (≤20 cm) are directly calculated from ice-surface temperature swath data and ERA-Interim atmospheric reanalysis data. Dedicated cloud screening and spatial/temporal interpolation techniques are applied to effectively account for sensor-specific drawbacks. The AMSR-E approach derives thin-ice thicknesses empirically by making use of a characteristic polarization-ratio–ice-thickness relationship. In general, both datasets allow for a (quasi-)daily pan-Arctic mapping of thin-ice thickness distributions and, based on that, a long-term derivation of important properties such as polynya area (POLA) and potential thermodynamic ice production (IP). For the overlapping 9-year period, we show that the average POLA (average accumulated IP) for all Arctic polynyas combined is 1.99 × 105 km2 (1.34 × 103 km3) in case of MODIS and 2.23 × 105 km2 (1.29 × 103 km3) for AMSR-E. Although the two datasets are independently derived and despite all methodical differences, they are to a large degree coherent in terms of capturing the general spatial and temporal characteristics of Arctic polynyas. Interestingly, emerging positive trends in POLA and IP over the long 16-year period of the MODIS dataset are mainly visible in the eastern Arctic and the eastern part of the Canadian Arctic Archipelago. We suspect that at least the former observation could be related to changing large-scale atmospheric modes and/or changing characteristics of the Transpolar drift system.


Contrasting snow-covered sea-ice geophysical properties and active microwave signatures from the Weddell and the Bellingshausen Seas, Antarctica

Vishnu Nandan, Randall Scharien

Corresponding author: Vishnu Nandan

Corresponding author e-mail: vishnu@uvic.ca

The Antarctic region has experienced an increase in winter sea-ice thickness in the inner Weddell Sea sector and thinning in the Bellingshausen Sea sector over the past four decades. The regional-scale contrast in ice thicknesses between these two sectors has been attributed to significant differences in snow thicknesses and the associated impact of this on sea-ice thermodynamics. This, in turn, may significantly affect satellite-scale active- and passive-microwave signatures used to accurately retrieve critical snow and sea-ice state variables and monitor sea-ice conditions. However, few studies have investigated the snow-covered sea-ice properties and processes from these two sectors, and their associated impact on active-microwave signatures at high spatial resolutions. This study explores the mechanisms influencing the variability in C-band polarimetric radar signatures of sea ice in the Weddell and Bellingshausen Seas during the spring period, by merging surface- and ship-based C-band polarimetric radar scatterometer observations with in-situ snow and sea-ice geophysical measurements. Data were collected during the 2010 IceBell cruise, which took place from 11–30 November. High spatial and temporal resolution scatterometer measurements are used in a sector-based comparison, as well as to examine the range of variability in scattering caused by spring-period diurnal changes within each sector. Preliminary results indicate an overall difference of ~4 dB in microwave backscatter between the two sectors. This strong difference has been found to be attributed to characteristic regional-scale variations in snow thicknesses and the associated thermodynamic impact of this on underlying ice floes, observed in the two sectors. The presence of slushy layers and snow ice (~10–30 cm thick) under thick snow (~70 cm) over the Bellingshausen Sea floes, contrasts with thin snow (~10 cm) over the Weddell Sea floes, leading to differences in surface and volume scattering mechanisms and strong differences in microwave backscatter. Ultimately, we aim to contribute to the effective utilization of C-band SAR as an observational tool for understanding changing Antarctic sea-ice conditions and resolving discrepancies between models and observed ice conditions, as well as for improving passive-microwave-based sea-ice geophysical information retrievals.


Acquiring sea-ice data under high probability of loss of equipment

Chris Petrich, Megan O’Sadnick, Øystein Kleven, Irina Sæther

Corresponding author: Chris Petrich

Corresponding author e-mail: christian.petrich@norut.no

Regionally, an ice cover in fjords of mainland Norway may form and break up repeatedly during winter. Performing geophysical measurements in this environment carries a high risk of loss of or damage to equipment. One way to address this challenge is to reduce equipment costs below the pain point while ensuring remote transfer of data during the acquisition period. We developed a buoy that logs GPS/GLONASS coordinates and ocean, ice and air temperature, and transmits data through the cell-phone network. Experience from the first season of multiple deployments showed that the concept is working but the physical design of the buoy could be improved to withstand forces in open water.


Opportunity-based fast-ice-thickness observations in Rektangelbukta, off Dronning Maud Land, Antarctica

Sebastian Gerland, Mats A. Granskog, Tore Hattermann

Corresponding author: Sebastian Gerland

Corresponding author e-mail: gerland@npolar.no

Observational data on Antarctic sea-ice thickness is sparse. For estimating Antarctic sea-ice volume and its changes, knowledge of sea-ice thickness is crucial, and in improving new satellite-based sea-ice thickness products, in-situ data are important for validation and calibration. Here, we give an overview on sea-ice observations close to the Fimbul ice shelf. Between 2005 and 2019, ice-shelf-fast sea-ice in Rektangelbukta off Dronning Maud Land (70°7′ S, 5°20.5′ E) has been observed in 10 out of 14 years once a year in November or December when people were present for logistical work related to the Norwegian Antarctic wintering base ‘Troll’. Measurements include ice thickness, snow thickness and freeboard, along with qualitative description and documentation of the ice situation. Despite the relatively basic observational setup, the data collected give interesting insights into the nature of the fast ice in this sector of the Antarctic coast. The current observation setup contains measurements along a cross-shaped pattern (two lines each 200 m long), with in total 17 thickness holes and 81 snow-thickness measurement points. At the thickness holes, ice thickness, snow thickness and freeboard is measured. Thickness levels of fast ice were observed commonly in the range 1.5–2.0 m. Snow thickness showed larger variability between years, leading to negative freeboard when snow thicknesses were rather large, versus positive freeboard with more moderate snow-thickness levels. Beyond the quantitative observations, other relevant features were also recorded, such as the occurrence of platelet ice and icebergs, both features that can affect fast-ice formation. Future work will extend the observational program by use of autonomous technology and connect the observations at Rektangelbukta to fast-ice observations at other Antarctic coastal sites, e.g. through the Antarctic Fast Ice Network (AFIN). We also consider developing multidisciplinary approaches, connecting the fast-ice processes to ocean, atmosphere, ice-shelf and ecosystem processes.


Evaluating the Arctic energy and water budgets in a climate model

Jeff Ridley

Corresponding author: Jeff Ridley

Corresponding author e-mail: jeff.ridley@metoffice.gov.uk

The energy and water budgets for land, ice, atmosphere and ocean across the Arctic domain, within the CMIP6 model HadGEM3-GC3.1, are evaluated. We look at seasonal changes in energy and water budget, driven by internal variability, and climate change, and assess using an ensemble of historical simulations. Riverine outflow to the Arctic Ocean remained constant until 1990 when it increased, particularly in autumn and winter. This is associated with less precipitation falling as snow. Consequently, the Arctic Ocean freshened, particularly in the Siberian shelf seas, leading to an earlier onset of sea-ice freeze. Atmospheric transports of heat and water into the Arctic remain unchanged despite rising global temperature. It is investigated to what extent the Arctic sea-ice decline is influencing the regional energy budget.


A new model of melt-pond refreezing

Lucia Hosekova, Daniel Feltham, David Schroeder, Daniela Flocco

Corresponding author: Lucia Hosekova

Corresponding author e-mail: l.hosekova@reading.ac.uk

The response of climate models to the heat, momentum and moisture transfer between sea ice, ocean and atmosphere depends on the parameterization of sub-grid-scale sea-ice physics. While melt ponds occur on the scales of tens to hundreds of meters, they influence sea ice and its couplings to components across a range of scales. The effect of melt ponds on the albedo feedback mechanism has been shown to have a leading-order impact on sea-ice simulations, leading to significant skill in sea-ice forecasts. In addition to albedo, melt ponds play a role in delaying winter basal freezing of sea ice during, and after, they refreeze. This is because basal growth cannot occur until a negative temperature gradient is established in the ice, drawing heat out of the ocean, and this cannot occur until a melt pond has refrozen. The impact of melt pond refreezing on basal growth is estimated to last up to a couple of months. Since this effect is not currently accounted for in climate sea-ice models (they only account for the impact of melt ponds on albedo), they are expected to overestimate winter ice growth. We have developed a new model of melt pond refreezing that accounts for melt ponds as a separate component in the thermodynamic model accounting for the local heat and salt balances. We have implemented a new radiative–thermodynamic scheme in CICE that represents concurrent phases during the entire pond cycle: the open/refreezing melt pond, the ice beneath the pond and the ice lid. We investigate the impact of our new model on the mean state and variability of Arctic sea ice, and explore sensitivity of the results to the model parameters.


Sampling strategies for snow depth on sea ice studies

David Clemens-Sewall, Chris Polashenski, Don Perovich

Corresponding author: David Clemens-Sewall

Corresponding author e-mail: david.w.clemens-sewall.th@dartmouth.edu

Snow depth on sea ice, which is spatially and temporally heterogenous, is a key parameter impacting the growth and melting of the sea-ice cover. Yet, there is not a consensus upon an optimal strategy for sampling snow depth on sea ice. Rigorous intercomparison studies have been limited by the logistics of collecting sufficient data. By using Terrestrial LIDAR Scanning of snow on level ice, we can generate exhaustive datasets of snow depth to decimeter-scale resolution. We use these datasets to evaluate the performance of numerous sampling strategies for direct snow-depth measurements that have been proposed, for example straight lines, multiple straight lines at various angles, two dimensional grids, spirals, and random walks. Additionally, we evaluate the different heuristics for choosing areal extent (e.g. length of the line or diameter of the spiral) and measurement spacing. The utility of sampling strategies is evaluated at three levels which are chosen to suit the needs of different researchers. The first level is the accuracy of a measure of central tendency (e.g. mean or median). The second level is the accuracy of the snow-depth distribution, and the third is the spatial covariance. In addition to evaluate the performance of commonly used sampling strategies; we also present novel sampling strategies targeted for each of the levels.


Snow depth on Arctic sea ice derived from airborne radar measurements

Arttu Jutila, Robert Ricker, Stefan Hendricks, John Paden, Joshua King, Chris Polashenski, Benjamin Lange, Christine Michel, Christian Haas

Corresponding author: Arttu Jutila

Corresponding author e-mail: arttu.jutila@awi.de

The snow layer on sea ice has high importance for polar climate as it affects heat, radiation, and fresh-water budgets. Additionally, snow loading is a critical parameter for the sea-ice freeboard-to-thickness conversion for satellite radar and laser altimeters. Despite its importance, there is a lack of snow observations spanning different spatial and temporal scales, thus introducing a significant source of uncertainty to altimetric sea-ice thickness retrievals. The ultra-wideband microwave radar (UWBM) Snow Radar, a 2–18 GHz airborne frequency-modulated continuous-wave (FMCW) radar developed by the Center for Remote Sensing of Ice Sheets (CReSIS) at the University of Kansas, can accurately detect the air/snow and snow/ice interfaces to measure snow thickness. Since 2009, an airborne Snow Radar has been operated onboard NASA’s Operation IceBridge (OIB) campaigns. In 2017, the UWBM Snow Radar was operated for the first time on an Alfred Wegener Institute (AWI) research aircraft, together with an airborne laser scanner for surface topography and freeboard measurements and an electromagnetic induction sounding instrument (EM Bird) to measure total ice thickness. The AWI airborne surveys operate at a low survey altitude (60 m a.g.l.) and slow aircraft speed, enabling fine-resolution mapping of the snow layer. Furthermore, the unique instrument setup on board the AWI research aircraft and the concurrent measurements of snow freeboard, total sea-ice thickness and snow depth allow us to directly investigate the freeboard-to-thickness conversion on regional scales for the first time. Here, we evaluate the performance of the radar installation and present radar-derived snow depth retrieved with a wavelet technique from recent airborne campaigns, PAMARCMiP2017 and IceBird winter 2019, over Arctic sea ice in the Greenland, Lincoln, Beaufort and Chukchi Seas and the central Arctic Ocean in March–April of the respective years.


Excess carbon induces polyhydroxyalkaonate production in Antarctic sea-ice bacteria

Eeva Eronen-Rasimus, Jenni Hultman, Igor Pessi, Hai Tran, Sirja Viitamäki, David Thomas, Peter Golyshin, Anne-Mari Luhtanen, Harri Kuosa

Corresponding author: Eeva Eronen-Rasimus

Corresponding author e-mail: eeva.eronen-rasimus@ymparisto.fi

Sea-ice bacteria are subjected to fluctuating and harsh environmental conditions such as low temperature, high salinity and intermittent substrate supply. Sea-ice bacteria possess different strategies to survive in the ice, i.e. compatible solutes and EPS-production; however, these mechanisms are not thoroughly understood. Polyhydroxyalkaonates (PHAs) are polyesters that serve as a pool for carbon storage and are readily available for different cellular processes. Commonly, PHA is produced in conditions with nutrient imbalance such as high carbon accompanied by low nitrogen or phosphorus. PHA granules and phaC synthase genes have been detected from sea-ice bacteria, however, the production mechanism and ecological significance is not known. Our aim was to investigate whether or not sea-ice bacteria are capable of PHA production and to elucidate the genetic mechanism behind the production. Two bacterial isolates, Paracoccus 392 (Alphaproteobacteria) and Halomonas 363 (Gammaproteobacteria), isolated from Antarctic sea ice, were grown in nitrogen-limited media at +4°C. Samples were collected for RNA extractions (transcriptomes), bacterial abundance (flow cytometry) and Nile blue A microscopy. In addition, genomes from both bacterial strains were sequenced (MiSeq) and the PHA was extracted and analysed with GC-MS. The results demonstrate that sea-ice bacteria are capable of producing PHAs. In addition to nitrogen limitation, PHA production was also induced in high nitrogen concentrations when excess carbon was available. We hypothesize that PHA production in ice is related to the ephemeral feature of labile dissolved organic carbon availability (i.e from the initial freezing and spring bloom). Thereafter there is carbon storage in PHA granules for enhanced survival in extreme and fluctuating environmental conditions.


Convection, phase change and solute transport in mushy sea ice

James Parkinson, Dan Martin, Rich Katz, Andrew Wells

Corresponding author: James Parkinson

Corresponding author e-mail: james.parkinson@physics.ox.ac.uk

Sea ice is a porous material composed of ice crystals and interstitial brine; a mushy layer. The dense brine tends to sink through the ice, driving convection. Downwelling at the edge of convective cells leads to the development of narrow, entirely liquid channels, through which cold saline brine is efficiently rejected into the underlying ocean. This brine rejection provides an important buoyancy forcing on the ocean, and can have important consequences for the internal structure and properties of sea ice. We consider 2-D numerical simulations of ice formation and convective brine rejection in a narrow Hele–Shaw cell. Adaptive mesh refinement, implemented via the Chombo framework, allows us to resolve narrow brine channels while integrating over many months of ice growth. The convective desalination of sea ice promotes increased internal solidification, and we find that convective brine drainage is restricted to a narrow porous layer at the ice–ocean interface, which evolves as the ice layer grows thicker over time. Away from this interface, stagnant sea ice consists of a network of previously active brine channels which retain higher solute concentrations than the surrounding ice. We investigate the response of these remnant brine channels to changes in atmospheric and oceanic conditions. Surface warming can increase the porosity in the upper layers of the ice, allowing blocked brine channels to drain into the ocean. Basal melt leads to the formation of a buoyant layer of fresh meltwater, which initially drives convective desalination within the ice before ultimately restricting the transport of salt into the underlying ocean. We consider the potential implications for nutrient transport, sea ice ecology, and parametrizations of ice–ocean brine fluxes.


Snow-ice potential formation over the Arctic Ocean

Ioanna Merkouriadi, Glen Liston, Mats Granskog, Robert Graham

Corresponding author: Ioanna Merkouriadi

Corresponding author e-mail: ioanna_merkouriadi@hotmail.com

Snow on sea ice is an important factor for sea-ice evolution. Snow can contribute to the sea-ice thickness via snow-ice formation. Snow-ice forms when seawater floods and refreezes at the ice/snow interface, due to excessive snow load that submerges the ice surface below sea level. Even though snow-ice is widespread in seasonally ice-covered seas and in the Southern Ocean, it is not considered to be prevalent in the Arctic, where thick perennial sea ice once dominated. However, snow-ice was recently observed north of Svalbard. It is unclear whether that was a one-off event, or becoming a more widespread phenomenon due to the recent thinning of the sea ice. The aim of this work is to examine the regional patterns and trends of snow-ice potential on a pan-Arctic scale, over the period 1980–2015. For this we use daily sea-ice concentration data, ice motion vectors, and ice-parcel-trajectory data, which are derived from NASA remote-sensing products. We apply a 1-D snow/ice thermodynamic model (HIGHTSI) for these ice parcels, and look into the snow-ice potential formation over the Arctic Ocean. HIGHTSI is forced with MERRA-2 and ERA-I reanalyses products. We look at the potential snow-ice contribution to the sea-ice mass balance, separately for first-year and second-year ice. The maximum snow-ice potential is commonly found at the Atlantic sector of the Arctic Ocean, but there is potential for snow-ice formation over the entire Arctic.


Desalination of sea ice in a fluid dynamics model

Chris Petrich, Pat Langhorne, Hajo Eicken

Corresponding author: Chris Petrich

Corresponding author e-mail: christian.petrich@norut.no

Sea-ice bulk salinity is relatively constant through much of the lifetime of an ice cover and therefore an extremely useful input parameter to model other sea-ice biogeophysical properties. In this study we solved the two-dimensional Navier–Stokes equations numerically for a binary salt–water system cooled from above while undergoing phase transition, i.e. for the growth of sea ice depressed to zero freeboard. Volume expansion of ice during freezing is accounted for and mass and energy conservation are ensured by solving the equations with the fnite volume method. The solid phase is a porous medium with microscopic permeability and porosity, and solid and liquid phases assume thermodynamic equilibrium within each grid cell at each time step. Simulations were performed with grid sizes from 100 μm to 100 mm. The model appears to be able to reproduce realistic ‘stable’ bulk salinity profiles whether the pore space is assumed to disconnect at low porosity or not. The model reproduces many of the observations from detailed, laboratory-scale investigations of the past half century, including the growth of brine channels, feeder channels, and the closing of channels in response to flow reversal. More or less periodic flow reversal through the pore space appears to be a universal phenomenon in the simulations, and a critical depth needs to be reached for desalination to begin. It further simulates a significant rate of fluid motion into the ocean that is balanced by fluid moving from the ocean back into the sea-ice pore space. A heat flux from the ocean to the ice results as a direct consequence of this. Ice–ocean fluid exchange may be significant. The presented modeling approach has its limitations. To maintain thermodynamic equilibrium, the model tends to form an interface with infinitesimally low porosity. To be realistic, this has to be limited by increasing the ocean temperature or by introducing parameterizations that enforce some minimum amount of consolidation at the interface. While the development of bulk salinity is largely governed by permeability at moderate porosities, fluid flow rates at the interface depend significantly on the permeability at high porosities, which is notoriously difficult to measure. Aspects related to the onset of desalination depend on some form of perturbation, which is numerical noise in the presented cases and thus poorly defined. Domain size limitations may result in incorrect spatial periodicity.


Seasonal evolution of fast ice off northeast Greenland (79° N) from an ice mass-balance buoy

Caixin Wang, Mats A. Granskog, Sebastian Gerland, Jean Negrel, Dimitry Divine

Corresponding author: Caixin Wang

Corresponding author e-mail: Caixin.Wang@npolar.no

Fast ice is a type of sea ice anchored to the coast, anchored to the sea floor or locked in place between grounded icebergs. It can be affected by tides, waves and swell. Icebergs originating from the Greenland ice sheet are often grounded on the shallow Belgica Bank off the northeast coast of Greenland in Fram Strait. This allows fast ice to form in a region called the Norske Øer Ice Barrier (NØIB). Fast ice in the NØIB is not always attached to the coast but may be kept in place solely by grounded icebergs. An ice mass balance buoy (IMB) records snow and sea-ice thermodynamic changes with temperature profiles from air through snow/ice into upper ocean, the snow and ice thickness, atmospheric pressure, and buoy position. IMBs were deployed on fast ice in Fram Strait in August or September 2012, 2014 and 2016. The buoy deployed in 2012 covered nearly a full annual cycle, while the buoys in 2014 and 2016 only functioned well for some weeks at the deployment sites and then some buoy components malfunctioned, likely due to ice dynamics, polar bears, or the buoys drifting out of the region with the breakup of the fast ice. In this study, we show the atmospheric conditions at the 2012 IMB site, examine the snow and sea-ice evolution during a full year, and derive the oceanic heat flux at the deployment site. According to the 2012 IMB record, the air temperature was lowest in late February, –52°C, and snow reached a maximum thickness of 0.73 m in late May. According to the ice changes at the ice base, we divide the period into five phases: I (29 August–27 September 2012), II (27 September–18 November 2012), III (18 November–29 December 2012), IV (29 December 2012–27 June 2013), and V (27 June–15 August 2013). The derived oceanic heat flux was large during phases I, II and III when ice thickness decreased at the ice base, and nearly zero in phase IV when the ice thickness started to increase, and became large again in phase V when ice melting took place both at the ice surface and bottom. The mean oceanic heat flux was around 10 W m–2 at the deployment site. With a 1-D sea-ice model, we examine the evolution of this fast sea ice and examine the sensitivity of ice growth to snow accumulation, and atmospheric and oceanic forcing.


Advection diffusion in the polar sea-ice cover

Huy Dinh, Noa Kraitzman, Rebecca Hardenbrook, Benjamin Murphy, Elena Cherkaev, Kenneth Golden

Corresponding author: Kenneth Golden

Corresponding author e-mail: golden@math.utah.edu

Over short length and time scales sea floes often exhibit Brownian-like dynamical behavior, but are also influenced by larger-scale advective forcing by the ocean and atmosphere. Similarly, thermal coupling of the ice-covered ocean to the atmosphere is controlled not only by diffusive heat transport through the sea ice but by advective contributions as well from the flow of brine inside the sea ice. In this presentation we give mathematical models of the effective behavior of advection diffusion processes in sea ice. In particular, we describe a numerical model of interacting floes under various advective forcings and pack conditions that yield observed regimes of anomalous floe diffusion and Hurst exponent values characterizing floe dispersion. We also discuss a related theory for the effective thermal conductivity of sea ice which rigorously accounts for the contribution from a fluid flow field and yields bounds on the thermal conductivity of sea ice for a simple model of brine convection.


Freshwater–marine interactions in the greater Hudson Bay marine region

Greg McCullough, Zou Zou Kuzyk, Paul Myers, Jens Ehn, David Babb, David Barber

Corresponding author: Greg McCullough

Corresponding author e-mail: Greg.McCullough@umanitoba.ca

The greater Hudson Bay marine region (HBMR; i.e. Hudson Bay, James Bay, Foxe Basin and Hudson Strait) comprises one of the largest inland seas in the world (1.3 Mkm2). With surface salinities ranging from 10‰ in James Bay to 33‰ in Hudson Strait, it is also the freshest marine region of similar magnitude. It is a major contributor of freshwater to the North Atlantic Ocean. Freshwater in the HBMR derives mainly from terrestrial runoff, which has been estimated by various authors at 850–950 k m3 a–1. Marine inflow contributes about 90 k m3 a–1 (baseline salinity = 33.1‰). Estimates of meteoric inputs range very widely, from net evaporation to 330 km3 a–1 net precipitation. Freshwater outflow at the eastern extremity of Hudson Strait is estimated to be about 1300 km3 a–1, and this value supports the larger estimate of meteoric inputs. Seasonal ice formation plays an important internal seasonal role by releasing about 1300 km3 of fresh water at the surface each spring and by brine rejection through the winter. More saline marine water enters the HBMR through Hudson Strait, but some is created locally by brine rejection during the formation of sea ice. This has been documented in Foxe Basin, as have mechanisms to transport relatively saline deep water from Foxe Basin into Hudson Bay, but it is believed that the process also occurs within Hudson Bay. In this presentation, we review the literature on fresh-water circulation and fresh-water–marine interactions in the HBMR, and supplement this with new information on volume discharges, variability and trends derived from the NEMO (Nucleus for European Modelling of the Ocean) ocean model, coupled with the LIM2 (Louvain-la-Neuve Ice Model) and forced with historical data.


Investigating the occurrence of newly formed sea ice in the Barents Sea using archived Sentinel–1 SAR data

Malin Johansson, Camilla Brekke, Anthony Doulgeris

Corresponding author: Malin Johansson

Corresponding author e-mail: malin.johansson@uit.no

Newly formed sea-ice areas provide safe routes for ship traffic and cost-effective passage through the sea ice, increasing the chance of marine pollution in these areas. Synthetic aperture radar (SAR) images are used for regular monitoring of Arctic sea ice and also for oil-spill detection. Oil spill and newly formed sea ice are both low-backscatter phenomena and may exhibit a dark signature in a radar image due to wave damping on the ocean surface; hence, they may be may be difficult to separate. It is therefore of interest to both i) detect newly formed sea ice on the ocean surface and ii) separate it from slicks caused by oil pollution. Medium-resolution wide-swath products such as Sentinel–1A and -1B are normally preferred in operational surveillance of Arctic waters, where dual-polarimetric data (HH/HV) in extra-wide mode (EW) with a spatial coverage of 400 × 400 km is most frequently acquired. Here we use a fully automatic segmentation algorithm that accounts for the large-scale variation in the wide-swath images to segment Sentinel–1A/B images from 2015 onwards covering the Barents Sea. Dark spots are subsequently identified and separated with respect to their size, shape and internal variation. As the dark spots are low-backscatter phenomena the distance to the noise floor is of importance and the areas will subsequently be compared to the nominal noise floor. A statistical analysis is performed to establish the possibility of separating newly formed sea ice from oil slicks using operationally available Sentinel–1 images. Thereafter we quantify the time of the year where newly formed sea ice, oil spills or look-alikes are most frequently observed. Using a grid-scale distribution, the spread through the year is broken down into regional differences, considering ship lanes and active oil-production sites, with respect to the longitude/latitude grid.


Sensitivity of two melt-pond schemes to the uncertainties in atmospheric reanalyses for global climate models

Jean Sterlin, Thierry Fichefet, François Massonnet

Corresponding author: Jean Sterlin

Corresponding author e-mail: jean.sterlin@uclouvain.be

Melt ponds appear during the melt season in the Arctic, when the surface melt water collects in the depressions of the ice field. The albedo of the ponds is lower than the surrounding ice and snow areas, and for this reason the ponds are hot spots for the ice-albedo feedback. There are two main approaches to represent the melt ponds in global climate models. The first approach is empiric and relies on observations to determine the available water capacity of the ponds from the sea-ice state. Then, a fraction of the surface meltwater accumulates in the ponds. The second makes use of the ice-thickness distribution to infer the surface topography of the sea ice and distribute the meltwater among the ice categories. Although the role of melt ponds has been extensively studied, less is known on the response of the ponds to climate change. Insights can be gained from using different reanalyses of the atmospheric surface state to force the ocean and ice components. Because of a lack of observations in remote areas, reanalyses still suffer from biases, notably in the polar regions. The choice of a reanalysis has a strong influence on the representation of the sea-ice state of the Antarctic. We expect similar deviations in the Northern Hemisphere. To evaluate the effect of the melt-pond schemes on the sea ice when subject to uncertainties in the atmospheric state, we have run the empiric and topographic schemes forced with JRA-55, DFS 5.2 and NCEP/NCAR atmospheric reanalyses. From the simulations, we expect to see the degree of difference between the pond schemes and the influence of the forcing onto their climatic response. We will be able to assess the importance of the melt ponds for the climate and check the consistency of the parameterizations. This will allow us to formulate a recommendation on the use of melt ponds in climate models.


Encouraging agreement between snowfall from CloudSat and reanalysis over the Arctic Ocean and Arctic sea ice

Alex Cabaj, Paul Kushner, Chris Fletcher

Corresponding author: Alex Cabaj

Corresponding author e-mail: acabaj@physics.utoronto.ca

The amount of snow that falls on Arctic sea ice is currently not well characterized. Reanalyses produce gridded snowfall rate products with good geographical coverage, but the reported snowfall is inconsistent between different reanalyses. Conversely, many snow observations are limited in spatial and temporal extent, due to the logistical challenges of making observations in such a remote region. Snowfall rate measurements from the CloudSat satellite can provide an observational reference to help assess some of the biases between reanalysis products. The 94 GHz cloud profiling radar instrument on CloudSat produces vertical profiles of radar reflectivity, from which surface snowfall rates can be retrieved. The CloudSat orbit extends to latitudes of up to 82° N, and the satellite has a 16-day repeat cycle. Thus, the instrument provides repeated measurements of high-latitude snowfall which can be compared to reanalysis. In this work, regionally averaged surface snowfall rates from CloudSat are compared with similarly averaged surface snowfall rate products from three reanalyses: ERA-Interim, ERA5 and MERRA-2. The products are averaged over the ocean in a region spanning 68–82° N. The effect of sampling discrepancies between the products is examined, and relative biases between the products are assessed. After the averaging is performed, an encouraging level of agreement is found between the CloudSat and reanalysis products, and a constant relative bias is found to account for much of the discrepancy between them. CloudSat is then used as a basis for calibration for the reanalysis products. Finally, the products are also compared over sea ice, and ongoing work using CloudSat to better quantify snow on sea ice is discussed.


The Polar Sea Ice Topography Reconstruction System

Scott Sorensen, Vinit Veerendraveer Singh, Wayne Treible, Andrew R. Mahoney, Chandra Kambhamettu

Corresponding author: Scott Sorensen

Corresponding author e-mail: sorensen@udel.edu

The Polar Sea Ice Topography Reconstruction System, or PSITRES, is a 3-D camera system designed to continuously monitor an area of ice and water adjacent to an ice-going vessel. Camera systems aboard ships in the polar regions are common, but the application of computer vision techniques to extract high-level information from the imagery is infrequent. Many of the existing systems are built for humans in the loop and lack the automation necessary for round-the-clock use. The PSITRES system was designed with computer vision in mind and can capture images continuously for days on end with limited oversight. We have developed a number of sample applications in the role of ice observation that can be accessed and evaluated on the web. We have deployed the PSITRES system on three research expeditions in the Arctic and sub-Arctic, and present applications in measuring ice concentration, melt-pond fraction and presence of algae. Systems such as PSITRES and the computer vision algorithms applied represent steps towards automatically observing, evaluating and analyzing ice and the environment around ships in ice-covered waters. We additionally provide a series of web applications and web APIs for the community to process the data collected from the PSITRES system.


On the relationship between Arctic sea-ice variability and precipitation over the Brazilian northeast and Amazon rainforest

Enoil de Souza Júnior, Eder Maier, Jefferson Simoes, David Barber

Corresponding author: Enoil de Souza Júnior

Corresponding author e-mail: enoil@myumanitoba.ca

A recent mode of climatic variability in the Northern Hemisphere mid-latitudes has been shown to be connected to the rapid warming of the Arctic and the associated reduction in the summer sea-ice extent. In this work, we go a step further and investigate whether variations in precipitation in equatorial South America are correlated to variations in the Arctic sea-ice extent from January 1979 to December 2015. Using sea-ice data provided by the National Snow and Ice Data Center (NSIDC) and precipitation monthly means reanalysis data for the same period provided by NCEP/NCAR, we correlate precipitation anomalies in equatorial South America (ESA) with Arctic sea-ice extent anomaly series. A precipitation center with the highest correlation with Arctic sea-ice-extent anomalies was identified in the eastern part of equatorial South America (in the Amazon rainforest). The time series of this region was then compared to the sea-ice extent anomaly using a linear regression dependence analysis. An r2 of 20% (a = 0.95) was found; in other words, the sea-ice variability might influence one-fifth of the precipitation volume anomaly in some parts of equatorial South America. This time-series analysis supports the hypothesis that Arctic warming forces the North Atlantic Oscillation (NAO) to its negative phase, causing changes in the pressure gradient over the North Atlantic mid-latitudes, influencing the trade winds. A higher/lower pressure gradient in the mid-latitudes over the North Atlantic Ocean intensifies/deintensifies the northeast trade winds, which intensify or shift the inter-tropical convergence zone to the Southern/Northern hemisphere, generating positive/negative precipitation anomalies in the eastern part of the South American equatorial region.


Is it really getting younger? Sea-ice type and age in a model simulation and satellite remote-sensing products

Polona Itkin, Pierre Rampal, Einar Ólason, Timothy Williams

Corresponding author: Polona Itkin

Corresponding author e-mail: polona.itkin@gmail.com

Sea-ice type and age are two of the basic indicators of Arctic sea-ice state. For a sea-ice model to simulate sea-ice type or age faithfully, both sea-ice dynamics and thermodynamics need to be represented well. In contrast to sea-ice thickness, ice age and type have been able to be retrieved from satellite observations relatively reliably for more than a decade. In this study we are using neXtSIM (‘next generation sea ice model’), which uses Maxwell–elasto–brittle rheology to simulate sea-ice motion and a thermodynamical model that accounts for healing of damaged ice through freezing. Previous studies have demonstrated that neXtSIM represents the seasonal cycle of Arctic sea ice well. It has also been shown that the model represents the statistical characteristics of sea-ice deformation. Here we are presenting the first long–multi-seasonal run with neXtSIM in a stand-alone mode. Our simulation spans 2000–17 and encompasses the recent rapid decline of the summer and winter sea-ice cover in the Arctic. Our comparison with the satellite observations offers an assessment of the multitude of conditions that need to be satisfied to obtain realistic simulations (e.g. realistic sea-ice drift, melt and export out of Arctic) and reliable remote-sensing products (e.g information about degree of ridging and snow-cover depth). Our simulations confirm the results from the satellite remote-sensing observations about the current prevalence of the first-year and second-year ice and new equilibrium state regarding sea-ice thickness and age. We also explore the relationship between the sea-ice age detectable from space (surface sea-ice age) and the true sea-ice age (whole column).


Towards improving snow-depth retrievals on Arctic sea ice from satellite microwave radiometers

Philip Rostosky, Gunnar Spreen, Sebastian Gerland, Marcus Huntemann, Mario Mech

Corresponding author: Philip Rostosky

Corresponding author e-mail: prostosky@iup.physik.uni-bremen.de

Snow on sea ice strongly influences the ocean–ice–atmosphere heat exchange and snow reflects most of the incoming solar radiation. Consequently, knowledge about snow depth and snow properties is important for understanding the Arctic climate system. In addition, information about snow depth is needed for sea-ice-thickness retrievals based on altimetry. In coupled climate models, snow and its processes relevant for the Arctic climate system are poorly described. One major issue is the lack of an Arctic-wide snow-depth product suitable for evaluating the model results. Snow depth on Arctic sea-ice retrievals based on passive-microwave radiometer satellite observations are capable of providing daily snow-depth observation over the whole Arctic sea ice independent of sunlight and (mostly) cloud cover. However, the retrievals are strongly influenced by ice and snow properties and by atmospheric variability, leading to errors in the retrieved snow depth. Quantifying these errors is one of the major challenges to be solved before the retrieved snow depth can be used for model evaluation. In this study, we investigate the influence of uncertainties in the properties of ice, snow and atmosphere on passive-microwave snow-depth retrievals based on a Monte-Carlo modeling approach. The models used here are the microwave emission model for layered snowpack MEMLS, the snow-evolution model SNOWPACK and the radiative transfer model PAMTRA. All simulations are based on ice, snow and atmospheric measurements obtained during the N-ICE2015 campaign north of Svalbard in 2015. Our results show that the mean error in the retrieved snow depth due to unknown snow properties and atmospheric state is 11–19%, depending on the frequencies selected and used for the retrieval. For higher frequencies (18.7 GHz and 36.5 GHz), clouds and unknown snow properties are the main error sources. For lower frequencies (6.9 GHz and 18.7 GHz), clouds have no impact on the retrieved snow depth. However, lower frequencies are less sensitive to changes in snow depth. In the case of multiyear ice, the ice properties too have a large influence on the retrieved snow depth, especially at high frequencies. However, the implementation of multiyear ice in MEMLS is simplified and therefore our results might underestimate the influence of the ice properties on snow-depth retrievals. Thus, further improvements of the snow and ice emissivity models will be needed.


Monitoring thin-sea-ice thickness with Sentinel-3

Øystein Rudjord, Rune Solberg, Øivind Due Trier, Nick Hughes

Corresponding author: Øystein Rudjord

Corresponding author e-mail: oystein.rudjord@nr.no

Along with the retreating sea ice, the human presence in the Arctic is increasing, and this development is likely to continue in the future. However, more traffic in ice-infested waters also increases the risk of accidents. Search-and-Rescue missions and clean-up operations in the Arctic are potentially very challenging and expensive to deal with. Harsh weather conditions and the presence of sea ice make such events difficult and costly to handle. Careful monitoring of sea ice is therefore important in order to ensure safe and efficient navigation in the Arctic. With the reduction of the sea-ice area along with the changing climate, an increasing proportion of the sea ice consists of young and first-year ice. A remote-sensing-based service monitoring the thickness of thin, seasonal sea ice would therefore be valuable. The Sentinel-3 satellites are highly suitable for operational monitoring. The frequent coverage from the dual A and B satellites provide frequent information about the rapidly changing sea ice. Using the thermal bands of the Sentinel-3 SLSTR sensor, it is possible to build an automated algorithm to retrieve the thickness of thin sea ice. For cold winter conditions, the heat balance on the upper ice surface is strongly influenced by the heat transfer from the water under the ice, and therefore by the ice thickness. The method was previously developed by Yu and Rothrock (1996) and is based on modelling the various heat fluxes on the ice surface. The surface temperature is assumed to be stable, meaning that the sum of the component heat fluxes is zero. This condition gives an equation, which is solved to estimate the ice thickness. The algorithm is fully automatic and takes as input the surface temperature of the ice, found from the thermal bands of SLSTR, as well as atmospheric observational or model data (air temperature and wind speed.) The spatial resolution of the ice-thickness products are the same as for the SLSTR thermal data, 1 km. We will present the algorithm and discuss the implementation of this method. We will also show some sea-ice-thickness products, and evaluation results.


CO2 transfer in landfast sea ice: impact of processes at the interfaces

Fanny van der Linden, Sebastien Moreau, Jean-Louis Tison, Willy Champenois, Marie Kotovitch, Gauthier Carnat, François Fripiat, Florian Deman, Frank Dehairs

Corresponding author: Fanny nan der Linden

Corresponding author e-mail: fvanderlinden@uliege.be

Sea ice is a biome actively participating in the regional cycling of CO2 both as a source and a sink at different times of the year depending on ice physics, ice chemistry and ice trophic status (autotrophic vs heterotrophic). The porous sea ice provides a dynamic habitat hosting diverse communities of microorganisms (algae, bacteria, heterotrophic protists, fungi and viruses), particularly concentrated at the bottom of the ice at McMurdo Sound, Antarctica. Bacterial and algal productions affect the CO2 dynamics by releasing or consuming CO2, which in turn impacts concentrations of dissolved inorganic carbon (DIC) and total alkalinity (TA) – key parameters to describe the ocean–sea-ice–atmosphere CO2 fluxes. The balance between photosynthesis and respiration of both algae and bacteria, expressed as the net community production (NCP), determines the net trophic status of the ice. NCP relates directly to the biogenic contribution of sea ice to CO2 uptake or release. During the YROSIAE project, which took place at Cape Evans in McMurdo Sound from November 2011 to December 2012, we carried out the first long-term monitoring of pCO2 and CO2 fluxes at sea-ice interfaces. The seasonal pattern of air–ice CO2 fluxes was consistent with pCO2 changes, i.e. brine pCO2 over-saturation during late winter (brine concentration of DIC and upward brine expulsion), leading to CO2 degassing, and under-saturation during spring (brine dilution and DIC depletion), leading to atmospheric CO2 uptake. However, diurnal cycles of air–snow–ice CO2 fluxes were superimposed on seasonal changes and appeared to be controlled by the diurnal cycle of basal snow and ice skin temperatures. Though the ice trophic status is likely to affect CO2 fluxes, it appeared that seasonal and diurnal changes at the sea-ice surface were decoupled from the succession of autotrophic and heterotrophic phases observed in the ice interior. At the bottom of the ice, a large biomass build-up was associated with high remineralization and heterotrophic activity. Such condition is likely due to the presence of a biofilm (microbial assemblages embedded in extracellular polymeric substances). The biofilm may further promote CaCO3 precipitation in parallel with an increase of salinity-normalized TA. Such a sea-ice system, where significant heterotrophic activity is maintained independently of the biomass build-up and which supports CaCO3 precipitation jointly with increasing alkalinity, challenges previous insights.


Accelerated melting of sea ice in the Arctic central region and its possible mechanism

Fei Huang, Zhihong Sun, Ruichang Ding

Corresponding author: Fei Huang

Corresponding author e-mail: huangf@ouc.edu.cn

In recent years, with global warming, the rapid melting of Arctic sea ice and the Arctic amplification (AA) mechanism have become hot issues in the international arena. The maximum variability of Arctic sea-ice concentration (SIC) mainly occurs in the marginal ice zone (MIZ) of multiyear ice. Therefore, variation of seasonal sea ice in the MIZ has been widely studied. However, our preliminary study found that sea ice in the central area of the Arctic (north of 80° N) has been characterized by accelerated melting in recent years. The SIC average in the Arctic central region (ACR) always appears extremely low, lower than 1.5 times standard deviation in the decade after 2006. The lowest SIC in ACR occurred in autumn and winter 2016 and there were three extremely low SIC events in September, November and December, with the record lowest SIC value lower than 50%. Spectral analysis shows that the time series of the SIC in ACR north of 85° N has amore significant intraseasonal oscillation instead of annual or semi-annual oscillation than that north of 80° N. The extremely decrease of SIC in ACR and its frequent occurrence in recent years both suggest that the sea ice in ACR has been experiencing accelerated melting in the past decade. Previous studies have suggested several possible mechanisms for accelerated sea-ice melting or AA, including ice albedo feedback, temperature feedback, water vapor feedback, cloud feedback, and ocean forcing. But so far no mechanism can fully explain the extremely low SIC in ACR. The physical mechanism of this new phenomenon may be different from the feedback mechanism of the MIZ. We found that increasing of melting ponds on the multiyear ice in ACR in summer, which is induced by albedo feedback and divergence of sea ice, could be a good reason for the ACR sea ice melting. The divergence of sea ice in ACR driven by surface wind stress plays an important role in decreasing the SIC in the ACR. Super-strong storms in the atmosphere frequently moving into the ACR in recent years may contribute to the surface wind and bring much more water vapor and cloud, which could enhance the water vapor and cloud feedback.


The 2018 north Greenland polynya observed by a newly introduced merged optical and passive-microwave sea-ice-concentration dataset

Valentin Ludwig, Gunnar Spreen, Larysa Istomina, Frank Kauker, Dmitrii Murashkin

Corresponding author: Valentin Ludwig

Corresponding author e-mail: vludwig@uni-bremen.de

Sea-ice concentration, the fraction of an area covered by sea ice, is highly relevant for physics, biology and the safety of shipping routes. Passive-microwave satellite data allow sea-ice-concentration retrieval all year round and almost independently of the state of the atmosphere. They have resolutions of down to 5 km. Thermal infrared measurements allow sea-ice-concentration retrieval at a resolution of 1 km, but are only available for cloud-free scenarios. We present a merged sea-ice-concentration product from thermal infrared and passive-microwave measurements. The merged product benefits from the high resolution of the thermal infrared data while preserving the spatial continuity of the passive microwave data. We demonstrate its benefits towards the single-sensor sea-ice concentration by observing a polynya that formed north of Greenland in February 2018. A sea-ice-concentration dataset derived from SAR measurements serves for comparison. The merged product resolves leads as sea-ice concentration between 60% and 80%, which would have been missed by the coarse-resolution passive-microwave data. Next, we analyse why the polynya opened. An unusual distribution of air-pressure systems during which the sea ice north of Greenland drifted in the opposite direction from usual is identified as the reason.


Sea-ice-pressure verification using ship movements

Angela Cheng, Jean-François Lemieux, Bruno Tremblay, Adrienne Tivy, Barbara Casati

Corresponding author: Angela Cheng

Corresponding author e-mail: angela.cheng2@canada.ca

Pressured ice poses a significant threat to ships that traverse the Canadian Arctic: it can restrict a ship’s movement, leading to increased travel time and economic losses. At worst, it can drive a beset ship into shallow waters and cause a hull breach. Previous verification studies on sea-ice pressure in models have relied on ship-besetting events, where known besetments were compared against forecasted sea-ice pressure to measure the accuracy of the model in capturing pressure. However, this relies on readily available data on ship-besetting events. In our study, we consider the alternative scenario: we compare reported ship movements against forecasted ice pressure to seek areas where the sea-ice model is inconsistent with observations. These events highlight areas where the sea-ice model is incorrect and requires refinement. We also use this data to determine the amount of pressure required to cause a ship to be beset. This will assist us with downscaling modelled sea ice at 5 km resolution to ship scales at 150 m. We present a plan and preliminary results of a comparison of the Regional Ice Ocean Prediction System (RIOPS), developed and run by Environment and Climate Change Canada, with Automated Identification System (AIS) mandatory ship-navigation data for January 2017 to August 2018. Our proposed method allows for verification of sea-ice-pressure models in the absence of known ship-besetting events.


Estimation of sea-ice concentration using signals of a global navigation satellite system

Maximilian Semmling, Dmitry Divine, Sebastian Gerland, Jens Wickert, Harald Schuh

Corresponding author: Maximilian Semmling

Corresponding author e-mail: maxsem@gfz-potsdam.de

Signals from global navigation satellite systems (GNSS) can be received at any location on Earth with high reliability and independently of daylight or visibility. The prior use of GNSS for positioning and navigation has been extended during the last two decades by various remote-sensing applications. Atmosphere parameters are estimated from GNSS signal refraction and are nowadays widely used in weather forecast and climate research. There is a growing science community that exploits Earth-reflected GNSS signals to characterize water, ice and land surfaces. Altimetric and scatterometric reflectometry methods have been demonstrated for ocean and sea-ice remote sensing. The sea-ice evolution that goes along with permittivity changes can be investigated by polarimetric methods. These methods rely on a polarization-dependent measurement of the signal’s phase and amplitude. A calibration-free polarimetric approach is considered here for estimation of sea-ice concentration (SIC). The GFZ working group for GNSS reflectometry gathered experience with the GORS (GNSS Occultation Reflectometry Scatterometry) receiver, which allows simultaneous records at left- and right-handed polarization using up to four antenna links. Polarimetric measurements with the GORS receiver were conducted in co-operation with the Norwegian Polar Institute at Kongsfjorden, Svalbard, in 2014/15 and during a cruise of R/V Lance to Fram Strait in 2016. The Fram Strait dataset, in particular, allowed us to develop a method for SIC retrieval. The method covers three data levels from the receiver’s raw data, to power estimates of separated direct and reflected links, to the SIC estimates. The Fram Strait dataset yielded a SIC time series with 3-hour resolution covering all 13 days when sea ice occurred. For validation the results are compared to regular sea-ice observations from the ship. The comparison shows a correlation of 0.75 (0.67) and a concentration RMSE of 25% (31%) for cross-polar (cross-to-co-polar) data. Concluding the results, the use of reflected GNSS signals can be of interest to retrieve independent data on regional sea-ice conditions.


Changes in sea-ice coverage in Norwegian fjords since 2002 and relationship to climatic, oceanographic and topographic variables

Megan O’Sadnick, Chris Petrich, Camilla Brekke, Jofrid Skarðhamar

Corresponding author: Megan O’Sadnick

Corresponding author e-mail: megan@tek.norut.no

Fjords along the northern Norwegian coast have similarities to other regions of the Arctic, including long periods of temperatures below freezing and the possible input of fresh water throughout the winter. Through studies of this region we can improve our understanding of variations in ice conditions in other coastal areas of the Arctic. Such work is useful as the number of ships and boats transiting the Arctic increases and, relatedly, the risks associated including the potential for oil spills. The latter can be greatly impacted by not only the presence of ice but its properties, for example, porosity. Observations of ice coverage in Norwegian fjords is limited largely to those found in the Norwegian pilot guide, which provides short descriptions of where ice may be present for use by ship and boat captains. A measurement campaign from 2017–19, however, examining ice extent, thickness and properties in seven northern Norwegian fjords, indicates the potential for substantial variations between fjords and from year to year. Such results provide motivation to further examine ice conditions along the coast through time and in relation to climatic, oceanographic and topographic variables. Here we present results quantifying variations in the ice area along the entire Norwegian coastline based primarily on analysis of MODIS images and supplemented by other satellite imagery. We focus on several fjords in northern Norway that show large variation, to better understand possible causes. Through enhancing our understanding of ice conditions in one Arctic region, we can better prepare for variable conditions throughout the Arctic as a whole. Such knowledge helps to ensure the safety of those traveling through and the protection of the Arctic environment.


Evaluating CryoSat-2 snow-freeboard retrievals in the Antarctic using data from ICESat-2

Steven Fons, Nathan Kurtz

Corresponding author: Steven Fons

Corresponding author e-mail: steven.w.fons@nasa.gov

Snow on Antarctic sea ice can exhibit a variable vertical structure brought on by the flooding of seawater into the snowpack. This complex stratigraphy can obscure the snow–ice interface in Ku-band radar returns, resulting in a higher than expected dominant scattering surface and thicker than expected freeboard estimates. To counteract the retracking uncertainty introduced by the snow layer, we utilize a waveform-fitting algorithm to retrieve the air–snow interface elevation of Antarctic sea ice. This routine employs a forward model and a least-squares approach to fit a modeled waveform to CryoSat-2 level 1B data, from which the air–snow interface elevation and snow freeboard are computed. Here, we present improvements to earlier versions of the algorithm, which include initializing the physical model with ICESat-2 freeboard data. We also show results of Antarctic snow freeboard using this method and comparisons with ICESat-2, with the goal of reconciling CryoSat-2 and ICESat-2 measurements of sea-ice freeboard and thickness.


Cold and dry vs warm and wet: how two winters of differing weather impacted sea-ice conditions and properties in seven northern Norwegian fjords

Megan O’Sadnick, Chris Petrich, Øystein Kleven, Camilla Brekke, Jofrid Skarðhamar

Corresponding author: Megan O’Sadnick

Corresponding author e-mail: megan@tek.norut.no

Measurements of ice extent, thickness and properties including temperature, bulk salinity and δ18O were obtained in March 2018 and March 2019 from seven northern Norwegian fjords. Results from 2018 show ice generally having a structure characteristic of sea ice including elongated pores and brine channels, with only one sample differing from this structure with far fewer pores and lower salinity. In 2019, ice conditions differed substantially not only in extent and thickness but properties and structure. Ice cores revealed layers, and in several instances entire cores, with few pores that were largely disconnected, characteristics of freshwater ice. Weather conditions between 2018 and 2019 varied, with the former having a greater number of days with below-freezing temperatures and very little snowfall. In 2019, temperatures were generally warmer with several days above zero. In addition, both snowfall and rainfall were greater. From measurements of bulk salinity and δ18O in combinations with observations of weather conditions, it is surmised that fresh water largely contributed to the variations in ice conditions observed. These variations are significant given their impact on the structure of the ice. Sea ice found in the open ocean of the Arctic is generally columnar in structure, having brine pores and channels that connect and widen as temperature increases. This process allows for biota living within and at the ice–ocean interface to move and live throughout the ice. In addition, in studies of the interaction of oil and sea ice, a connected pore space enables the upward migration of oil to the ice surface once certain temperatures are reached. The layers of freshwater ice observed in samples obtained in 2018 and 2019 may create a block to the movement of both biota as well as to pollutants such as oil in the ice. Such an occurrence could potentially disrupt further biological processes and spill response respectively. The data presented here can assist in improving our understanding of why ice extent, thickness and properties vary between fjords and years. Such knowledge is useful as the Arctic experiences changing weather patterns and interest in the region for its shipping routes and resources increases.


Assessing internal variability of Arctic sea-ice thickness and volume in model simulations and satellite data

Alexandra Jahn, David Hall, Kerry Dochen

Corresponding author: Alexandra Jahn

Corresponding author e-mail: alexandra.jahn@colorado.edu

Satellite-based remote-sensing measurements of Arctic sea-ice thickness have provided important new constraints on the sea-ice changes in the Arctic for the last 1.5 decades. However, ICESat and CyroSat-2 campaigns did not overlap, they (so far) provide only a short time series in terms of climate records, and they have sampled an unprecedented period of rapid sea-ice decline in the Arctic. Given all these constraints, how do we best use these sea-ice thickness products for model assessment? By analyzing the variability of sea-ice thickness from the CESM large ensemble, CMIP5 and CMIP6 models, and the satellite data in novels ways, we will show how we can interpret these short but crucial records of Arctic sea-ice change in the light of natural variability, and which insights they allow about sea-ice thickness and volume climate-model biases and their sources. In particular, we use a novel concept called decorrelation length scale analysis to assess the spatial and temporal variability of sea-ice thickness in models and observations.


Modelled effect of Arctic sea-ice geoengineering

Andrew Pauling, Cecilia Bitz

Corresponding author: Andrew Pauling

Corresponding author e-mail: apauling@uw.edu

Sea-ice geoengineering has recently been proposed as a way to offset the rapid decline in Arctic sea-ice extent and thickness. The proposed method involves deploying wind-powered pumps to draw seawater onto the surface of Arctic sea ice when the surface temperature is below freezing. This seawater would flood the snow layer and subsequently freeze, thus thickening the sea ice and prolonging the lifetime of the sea-ice cover on the Arctic Ocean and subpolar seas. We use both the Icepack column sea-ice model and the Community Earth System Model version 4 (a fully-coupled Earth System Model) to investigate the mechanism, the optimum timing of flooding during the seasonal cycle, and in which year the geoengineering must commence to be most effective. Our integrations show that flooding the snow increases Arctic sea-ice thickness by the additive effects of enhancing the sea-ice growth rate and snow-to-sea-ice conversion. The thickness is only reliably increased if flooding occurs when the surface is below freezing, so the latent heat of freezing can be conducted to the air above. The sea-ice growth rate is enhanced most by preventing the snow from accumulating through flooding, which is accomplished by flooding the snow layer when snowfall rates become appreciable in autumn. If the surface is too warm, flooding may leave the snow wet and may cause melt ponds to form. Therefore, flooding must be halted during summer, when decreasing the surface albedo would increase shortwave absorption and accelerate sea-ice melt. In tests with uniform flooding year-round, enhanced melt during summer dominated over any thickness gains during winter. Even when flooding is limited to the cold season, there is a diminishing benefit to flooding with greater amounts of seawater since the amount of snow-to-sea-ice conversion is limited by the snow depth available to be flooded. Sea-ice flooding can only increase the thickness where sea ice is present, and the method is only optimal where sea ice exists early in the cold season. To be most effective, the flooding must begin almost immediately, and the flooding must be continued for decades to have an appreciable effect on sea-ice thickness. Even when implemented at the optimal time of year and as soon as possible, geoengineered flooding of Arctic sea ice is not sufficient to prevent decline in Arctic sea ice in moderate-to-high future warming scenarios.


Surface-albedo change trends over the Antarctic sea-ice region during 1982–2015

Chunxia Zhou, Teng Zhang, Lei Zheng

Corresponding author: Chunxia Zhou

Corresponding author e-mail: zhoucx@whu.edu.cn

Based on a long-time series of remote sensing data, we analyzed surface-albedo (SAL) change during summer (December–February) for the entire Antarctic sea-ice region (ASIR) and five longitudinal sectors around Antarctica: the Weddell Sea (WS), Indian Ocean, Pacific Ocean (PO), Ross Sea and Bellingshausen–Amundsen Sea (BS). Empirical mode decomposition was used to extract the trend of the original signal, and then a slope test method was utilized to identify a transition point. The SAL provided by the CM SAF cloud, albedo, and surface radiation dataset from AVHRR data was validated at Neumayer station. Sea-ice concentration (SIC) and sea-surface temperature (SST) were also analyzed. The trend of the SAL/SIC was positive during summer over the ASIR and five longitudinal sectors, except for the BS. Moreover, the largest increasing trend of SAL and SIC appeared in the PO at approximately 3.781% and 3.358% per decade, respectively. However, the decreasing trend of SAL/SIC in the BS slowed down, and the increasing trend of SAL/SIC in the PO accelerated. The trend curves of the SST exhibited a crest around 2000–05; thus, the slope lines of the SST showed an increasing–decreasing type for the ASIR and the five longitudinal sectors. The evolution of summer albedo decreased rapidly in the early summer and then maintained a relatively stable level for the whole ASIR, and its change mainly depended on the early melt of sea ice during the entire summer. The change of sea-ice albedo had a narrow range when compared with composite albedo and SIC over the five longitudinal sectors, and reached a stable level earlier. The transition point of SAL/SIC in several sectors appeared around the year 2000, whereas that of the SST for the entire ASIR occurred in 2003–05. A high value of SAL/SIC and a low value of the SST existed in the WS which can be displayed by the spatial distribution of pixel average. In addition, a low latitude indicates a low SAL/SIC and a high SST. A transition point of SAL appeared in 2001 in most areas of West Antarctica. This transition point could be illustrated by anomaly maps. The spatial distribution of the pixel-based trend of SAL demonstrated that the change of SAL in East Antarctica has exhibited a positive trend in recent decades. However, that in West Antarctica presented a decreasing trend before 2001 and transformed into an increasing trend afterward, especially in the east of the Antarctic Peninsula.


Reappearance of Weddell Sea offshore polynyas driven by Southern Hemisphere climate anomalies

Ethan Campbell, Earle Wilson, Kent Moore, Stephen Riser, Casey Brayton, Matthew Mazloff, Lynne Talley

Corresponding author: Ethan Campbell

Corresponding author e-mail: ethancc@uw.edu

The austral winters of 2016 and 2017 saw vast sea-ice openings in the Weddell Sea, Antarctica, over the Maud Rise seamount. These offshore polynyas were the largest to emerge from complete ice cover since the major polynya events of 1974–76. While deep convective mixing of the water column within polynyas near Maud Rise is thought to have far-reaching impacts on ocean circulation, carbon sequestration, regional meteorology, and possibly global climate, the formation mechanisms of offshore polynyas are not well understood. Here we present in situ observations of this rare phenomenon, collected by SOCCOM under-ice Argo biogeochemical profiling floats near Maud Rise, in combination with records of the daily evolution of sea ice and atmosphere from passive-microwave remote sensing and atmospheric reanalysis. Intense ocean heat loss drove deep overturning within the openings, which we find were initiated and modulated by the passage of severe storms, likely through turbulent mixing and ice divergence. Idealized 1-D model simulations demonstrate that the ice–ocean system near Maud Rise is particularly susceptible to ice thinning from storm episodes. Additionally, wind-driven upwelling of record strength generated low upper-ocean stratification that favored destabilization in 2016 and 2017. We show that previous polynyas in the Weddell region likely developed under similarly anomalous conditions, which are associated with a strengthening mode of Southern Hemisphere climate variability. While many climate models tend to generate polynyas following an accumulation of subsurface heat, which weakens stratification from below, our analysis suggests that concurrent upper-ocean preconditioning and meteorological perturbations are instead responsible for their appearance.


Snow depth on sea ice from altimetry for 2013–18 Arctic and austral winters

Sara Fleury, Florent Garnier, Antoine Laforge, Kévin Guerreiro, Frédérique Rémy

Corresponding author: Sara Fleury

Corresponding author e-mail: sara@toloza.net

Snow depth at the top of sea ice is a key parameter of climate change due to its isolation and albedo properties, which condition sea-ice growth and melt, heat absorption and primary production under the ice. Moreover, for several reasons that will be described in this presentation, lack of knowledge about snow depth can impact sea-ice thickness retrieval using altimetry, introducing an error that can reach 100% in the worst cases. Nevertheless, there exists currently no reliable snow-depth product over Arctic and austral sea ice. Indeed, the Warren (1999) climatology, frequently used to convert freeboard to sea-ice thickness, has been built with data obtained decades ago, before the first impacts of climate change, and the meteorological re-analyses fail to faithfully reproduce snowfall in polar regions. A recent study has shown that the difference of the scattering properties of the Ku-band and the Ka-band can provide a good proxy of the snow depth using an adapted processing chain, Alti Snow Depth (ASD). As a first demonstrator, the ASD chain has been applied to the Ka measurements derived from Saral/AltiKa, a French–Indian satellite, and to the corresponding Ku measurements derived from the European satellite altimeter CryoSat-2/SIRAL. These two altimeters are based on different techniqus and the ratio between the surface areas illuminated by each radar is about one order of magnitude. In order to make the AltiKa LRM measurements comparable to the CryoSat-2 SAR measurements, these latter have been pre-processed by the CNES in a degraded PLRM mode. The evaluation of the ASD product with regard to the Operation Ice Bridge snow radar data shows a good correlation (R = 0.67), much better than all previous solutions. Nevertheless the results were limited to two winter campaigns. We will present the results for the period extended to the 5 CryoSat-2/Saral common years (2013–18) and to austral sea ice, the GOP product dedicated to ocean from ESA. Their performance in regards with several in-situ datasets will be analysed. This work has been supported by the CryoSeaNICE ESA project and TOSCA SICKAyS CNES project.


Processes driving the Arctic sea-ice open-water period

Abigail Smith, Alexandra Jahn, Muyin Wang

Corresponding author: Abigail Smith

Corresponding author e-mail: abigail.l.smith@colorado.edu

The duration of the Arctic sea-ice open-water period has important climatic, ecological and societal implications. To define the open-water period, previous studies have either used sea-ice melt and freeze onset or sea-ice break-up and freeze-up, and sometimes these terms are even used interchangeably. However, the time between the two sets of dates is not spatially homogeneous and changes from year to year. Here we assess the processes that determine the differences in timing between melt onset and break-up (the melt period) and between freeze onset and freeze-up (the freeze period), to better understand the ongoing changes in the Arctic sea-ice cover. For this analysis, we are using the Community Earth System Model Large Ensemble (CESM LE) as well as CMIP5 models. We also use Bootstrap sea-ice concentrations from Nimbus 7 SMMR and DMSP SSM/I-SSMIS from 1990–2014. By utilizing multiple data sources (satellite observations, the CESM LE and CMIP5 models), we are able to better assess the causes of model biases, as well as the processes driving sea-ice melt and freeze period variability. Over the period 1990–2014, the CESM LE produces break-up and freeze-up dates comparable to those derived from satellite observations, but there is substantial internal variability in sea-ice break-up and freeze-up. In both the CESM LE and satellite observations, the melt period is longer than the freeze period over the satellite era. Using the CESM LE and CMIP5 models, we evaluate how and where across the Arctic Ocean the melt and freeze periods are projected to change during the 21st century, and how these changes may affect sea-ice sensitivity to global warming.


Detailed spatially explicit model simulations of the snow covering sea ice

Nander Wever, Katherine Leonard, Ted Maksym, Leonard Rossmann, Marcel Nicolaus, Michael Lehning, Jan Lenaerts

Corresponding author: Nander Wever

Corresponding author e-mail: nander.wever@colorado.edu

Sea ice is an important component of the global climate system. The presence of a snow layer covering sea ice modifies its thermodynamic behavior, due to the low thermal conductivity and high albedo of snow. The snowpack can be strongly stratified and change properties (density, water content, grain size and shape) throughout the seasons. Wind redistribution of snow is an important control on the spatial distribution of snow thickness as a function of the sea-ice topology, but also influences snow density by erosion and deposition. The snow load can cause flooding of the snow layer by saline ocean water, which can strongly impact the ice mass balance and freezing point of the snow. To capture these complex dynamical snowpack processes, the physics-based, multi-layer SNOWPACK model was recently modified to simulate the snow–sea-ice system. In addition to modifications to the model thermodynamics, SNOWPACK now also describes water and salt transport through the snow–sea-ice system by coupling the transport equation to Richards equation. These modifications allow the snow microstructure descriptions developed in the SNOWPACK model to be applied for sea-ice conditions as well. We show model performance for three sea-ice floes visited in austral winter 2013 in the Weddell Sea, Antarctica. These flows were measured by terrestrial laser scanning and spatially distributed snow-depth measurement and ice thickness was determined using electromagnetic measurements. This dataset is used to force distributed simulations of the ice floe, including the wind redistribution of snow and its influence on the thermodynamics of the floe. SnowMicroPenetrometer measurements from the ice floes show good agreement with simulated snowpack stratigraphy. The model framework can be used to assess the snow microstructure on sea ice, which is relevant for satellite retrieval algorithms of snow and ice thickness.


Spring distribution of bacteria and viruses in first-year and multiyear sea ice of the Lincoln Sea

Constance Duffaud, Michel Gosselin, Karley Campbell, Pierre Coupel, Steve Duerksen, Joannie Charette, Benjamin Lange, Pascal Tremblay, Claude Belzile

Corresponding author: Constance Duffaud

Corresponding author e-mail: Constance.Duffaud@uqar.ca

The reduction in the extent and thickness of sea-ice cover in the Arctic Ocean and the gradual replacement of multiyear ice (MYI) by first-year ice (FYI) significantly alter the habitats of sea-ice microbial communities. Due to their importance in biogeochemical cycles and microbial loop, there is an urgent need to understand the distribution of sea-ice bacteria and viruses in a changing Arctic. The Lincoln Sea, north of Ellesmere Island, is currently characterized by both FYI and MYI types during the spring season. It is a unique region to understand the MYI ecosystem as it is one of the last in the Arctic Ocean. The objective of the study was to determine the vertical distribution of bacteria and viruses in FYI and MYI in this rapidly changing region. Ice cores and water column samples were collected in the coastal region of Lincoln Sea, near Alert, from 3–23 May 2018. Bacteria and viruses were not homogeneously distributed in FYI and MYI cores. Bacterial abundances ranged from 4.75 × 103 mL–1 to 0.86 × 106 mL–1 and viral abundances ranged from 0.16 × 106 mL–1 to 4.53 × 106 mL–1 both of which peaked in the bottom centimeters of the ice and were higher than abundances measured in the underlying water column (mean of 0.11 × 106 mL–1 and 0.87 × 106 mL–1 for bacteria and viruses, respectively). Abundances in full ice core were low compared to values reported in the literature, whereas those measured in surface water under ice were similar to previous studies. As the season progresses, the virus-to-bacteria ratio (VBR) tended to decrease as bacterial abundances increased. The VBR values ranged from 3 to 119 in sea ice and from 6 to 10 in the water column. Inside the sea ice, the vertical distributions of bacteria and viruses followed similar patterns to those of chlorophyll a and nutrients (i.e. NO3 + NO2 and PO4). These results suggest that the vertical distribution of bacteria and viruses is influenced by biological and environmental factors. This study will improve our knowledge on the mechanisms controlling the distribution of microbial communities in sea ice, notably in MYI.


First-year and multiyear ice algal community in the Last Ice Area

Joannie Charette, Benjamin A. Lange, Karley Campbell, Pierre Coupel, Steve Duerksen, Pascal Tremblay, Cody Carlyle, Christine Michel

Corresponding author: Joannie Charette

Corresponding author e-mail: joannie.charette@dfo-mpo.gc.ca

The Last Ice Area (LIA), north of Ellesmere Island, is of high ecological significance because it is the only Arctic region expected to retain summer sea ice until 2050. The presence of first-year and multiyear sea ice in this region provides a unique opportunity to study both ice types under similar environmental and oceanographic conditions. In this context, water column, first-year and multiyear sea ice were sampled during a 30-day ice camp, set up in the coastal region of the Lincoln Sea, 9 km from CFS (Canadian Forces Station) Alert. In this study, we present results from 45 first-year and 72 multiyear ice cores collected between 3 and 23 May. Ice thickness varied between 1.37 and 1.76 m and between 2.08 and 4.58 m for first-year and multiyear ice, respectively. Chl-a concentrations ranged from <0.01 to 2.56 mg m–2 and from <0.01 to 1.29 mg m–2 in first-year and multiyear ice, respectively. Algae were nitrogen-limited in the bottom and mid sections of both ice types, while nitrogen, phosphate or silicate were limiting in the top sections. Pennate diatoms were generally the dominant taxonomic group. However, the taxa found in the bottom sections differed, with Fragilariopsis cylindrus and Pseudo-nitzschia delicatissima group dominating first-year ice, while Nitzschia spp. (20–100 μm) and pennate diatoms (5–20 μm) were the main taxa present in multiyear ice. Resting cells were found in the top and middle sections of first-year and multiyear ice, mainly Chaetoceros spores and Chrysophyceae stomatocysts, respectively. These results highlight the differences found in the algal community othe f first-year and multiyear ice of the LIA, but also the variability within the sea-ice sections.


Understanding Arctic sea ice variability and predictability with regional- to global-scale process-oriented evaluation

Cecilia Bitz, Wei Cheng, Aaron Donohoe, Edward Blanchard-Wrigglesworth, Mitchell Bushuk

Corresponding author: Cecilia Bitz

Corresponding author e-mail: bitz@uw.edu

Sea ice is a key source of climate predictability on weekly to decadal timescales. The ability of models to represent important sea-ice processes that give rise to sea-ice predictability is influenced by model bias. Yet a long-standing feature of sea-ice predictions is the very large intermodel spread that is pervasive on many time scales. Part of this spread appears to be irreducible, but a considerable portion is due to model errors in the sea-ice physics and coupled interactions of the sea ice, atmosphere and ocean. We are analyzing CMIP5 and -6 simulations and observations to identify errors that are critical to simulating sea-ice variability with a focus on quantifying model processes that influence persistence. One might expect that excessive persistence leads to overestimating predictability. We find that anomalies in the Arctic sea-ice cover (local concentration and area/extent by region) in CMIP models are too persistent from year to year and month to month compared to observations. Such excessive persistence means that the anomalies last too long, drive variance higher and create bias. We also find that the standard deviation of monthly sea-ice area is too high in nearly all models. Unsurprisingly, sea-surface temperature anomalies are similarly overly persistent in the Arctic in CMIP models. We will present results of new metrics that associate sea-ice variability and persistence to oceanic and atmospheric heat transport, ocean stratification, sea-ice albedo feedbacks, and teleconnections from the midlatitudes and tropics.


Investigating the role of the coupling between the ocean, the sea ice and the atmosphere in the biogeochemical cycle of CH4

Caroline Jacques, C.J. Sapart, B. Thornton, B. Delille, G. Carnat, P. Crill, T. Gkritzalis, T. Röckmann, C. van der Veen

Corresponding author: Caroline Jacques

Corresponding author e-mail: caroline.jacques@ulb.ac.be

The contribution of the ocean to the atmospheric methane (CH4) budget is poorly understood. In polar regions, this contribution is further influenced by the sea-ice cover. Sea ice has long been considered as an inert and impermeable barrier, but recent studies have highlighted the existence of gas fluxes at the atmosphere–sea-ice and sea-ice–seawater interfaces. These fluxes are to date poorly characterized and quantified, so that the role of sea ice as a net sink or source of CH4 is still unclear. The PIPERS (Polynyas, Ice Production, and seasonal Evolution in the Ross Sea) expedition on the icebreaker Nathaniel B. Palmer provided a unique opportunity to investigate the complex coupling between the ocean, the sea ice and the atmosphere, at the beginning of winter 2017. As the season was progressing, we could study how sea-ice formation influences the biogeochemical cycle of CH4. We performed continuous measurements of dissolved CH4 in surface waters together with CH4 mixing ratio in the atmosphere. Discrete samples were also collected both to calibrate continuous systems and to carry out subsequent stable isotope analyses. A total of 17 ice cores dedicated to CH4 concentration measurements were drilled in the Ross Sea pack ice and in the vicinity of the Terra Nova Bay polynya. Additional analyses, such as characterization of ice texture and physical properties, were conducted on these cores. This multiparametric dataset will allows us to determine the distribution of CH4 between the ocean, the sea ice and the atmosphere during sea-ice formation and help us to unravel the net role of sea ice in the CH4 budget.


A parallel multigrid finite-element framework to solve viscous–plastic sea-ice models at high spatial resolution

Carolin Mehlmann, Thomas Richter

Corresponding author: Carolin Mehlmann

Corresponding author e-mail: carolin.mehlmann@ovgu.de

The subject of this presentation is the mathematical challenges and the numerical treatment of sea-ice models at high spatial resolutions. The model under consideration goes back to Hibler (1979) and is based on a viscous–plastic description of the ice as a two-dimensional thin layer on the ocean surface. The development of fast robust and converging solvers for the nonlinear momentum equation at high resolutions is still a big issue. While simple iterations such as the Picard solver call for a vast number of iterations, Newton linearization entails a very stiff and unstructured Jacobian, where no efficient linear solver is available up to date. We introduce a parallelized geometric multigrid method as preconditioner to the GMRES iteration accelerating the solution of the linear problems. We show that the convergence rate of the multigrid method is robust with respect to mesh refinement. This makes it an appealing method for high-resolution simulations. We validate the robustness of the linear solver and show that, by using a multigrid preconditioner, the linear iteration count is reduced by 80% compared to an ILU preconditioner. As smoothing operator in the multigrid method we present a parallelized Vanka relaxation and an ILU method.


Investigating ocean and atmosphere anomalies on and off the Ross Sea continental shelf to help explain persistent low sea ice in the Ross Sea since spring 2016

Sharon Stammerjohn, Stephen Ackley, Ted Maksym, Phillip Reid, Rob Massom, Douglas Martinson

Corresponding author: Ted Maksym

Corresponding author e-mail: tmaksym@whoi.edu

Several mechanisms have been previously proposed to explain the record-breaking low sea-ice extent in the Ross Sea in spring 2016, including the high-latitude response to tropical forcing, combined with internal and regional atmospheric variability (e.g. zonal wave 3, SAM), as well as persistent ocean thermal anomalies operating at seasonal to decadal time scales. This record low spring sea-ice extent in the Ross Sea was particularly noteworthy since it stood in stark contrast to the strong positive sea-ice extent trends previously observed in all seasons in the Ross Sea (though strongest in autumn and spring). Remarkably, since spring 2016, the Ross Sea has continued to experience anomalously low sea-ice extent. Here we explore autumn–winter ocean data acquired during PIPERS (Polynyas, Ice Production and seasonal Evolution in the Ross Sea) in 2017 to help distinguish ocean versus atmospheric anomalies that may have contributed to the anomalously late autumn–winter ice-edge advance and thickness evolution in 2017. We also highlight potentially different mechanisms that may be operating on and off the continental shelf as well as seasonally. Finally, the 2017 ocean and atmospheric conditions are compared to available historic data to explore potential explanations for the multiyear persistence of low sea ice in the Ross Sea since spring 2016.


New global sea-ice results from ICESat-2

Nathan Kurtz, Alek Petty, Ron Kwok, Thorsten Markus, Steven Fons

Corresponding author: Nathan Kurtz

Corresponding author e-mail: nathan.t.kurtz@nasa.gov

The successful launch of NASA’s ICESat-2 laser altimeter in September 2018 has ushered in a new era for global altimetry measurements of sea ice. ICESat-2 operates year-round with unprecedented spatial resolution enabling new measurements of sea-ice freeboard, thickness, surface roughness, melt ponds, and linkages to other systems such as clouds and ice sheets. In this presentation we give an overview of the ICESat-2 mission and a summary of the data products and results to date. We demonstrate the improvements seen in the dataset compared to the predecessor ICESat mission due to the much higher spatial sampling of the six-beam laser configuration and the statistical approach afforded by the photon-counting laser system. We then demonstrate new retrieval processes for sea-ice thickness and present new results for both the Arctic and Southern Ocean. Finally, we discuss ongoing efforts to reconcile the ICESat-2 record with CryoSat-2 and ICESat, through comparison with airborne data from NASA’s Operation IceBridge. Particular focus will be towards reconciling the seasonal evolution of sea-ice thickness over the growth season.


Sea-ice production, snow and ice-type variability in the Ross Sea

Ted Maksym, Jean-Louis Tison, Steve Ackley, Sharon Stammerjohn

Corresponding author: Ted Maksym

Corresponding author e-mail: tmaksym@whoi.edu

Sea-ice production and thickness in the Ross Sea is determined by the balance of three processes – frazil ice production at the ice edge and in coastal polynyas, snow-ice production, and columnar basal growth. To assess the impact of variability in ice-advance phenology on these processes, we present ice-core structural data and drifting-buoy observations acquired on the PIPERS (Polynyas, Ice Production and its seasonal Evolution in the Ross Sea) cruise during the anomalous April–June ice advance of 2017, and compare to prior sea-ice core data obtained during winter cruises in May–June 1995 and 1998. Using a simple, observationally based model of ice production driven by satellite ice drift and sea-ice buoy observations, we examine the relative roles of large-scale ice dynamics and thermodynamics in the polynyas, in the central pack, and at the ice edge in governing sea-ice type and production and their interannual variability. Despite high rates of ice production in the polynyas compared to prior years, the ice remained thin throughout the pack due to rapid export and northward drift, and the delayed onset of ice advance in 2017. The amount of snow ice observed (~6% of the thickness) was less than half that previously observed due both to the younger ice cover and anomalously low precipitation. In contrast to prior cruises, the percentage of columnar ice increased while that of frazil ice decreased from the Ross Sea polynya northward to the marginal ice zone. This is consistent with a melting ice edge and reduced pancake-ice production in the marginal ice zone in the western Ross Sea in 2017. These data and analyses suggest that (1) previous assessments of the role of snow ice in thickening the Ross Sea pack may have been biased high due to anomalously high precipitation in 1995, and (2) the structural development of the Ross Sea pack is determined by the balance of polynya ice production and export and ice–ocean interactions driving ice-advance variability at the ice edge.


The role of sea-ice-derived carbon in food-web processes of Eclipse Sound, Canadian Arctic

Doreen Kohlbach, Steven H. Ferguson, Thomas A. Brown, Cody Carlyle, Christine Michel

Corresponding author: Doreen Kohlbach

Corresponding author e-mail: doreen.kohlbach@dfo-mpo.gc.ca

Changes to the sea-ice habitat due to climate warming and potentially increased shipping activities in the Canadian Archipelago not only might have dramatic consequences for the ice-associated ecosystem but will also have a large impact on pelagic and benthic food-web dynamics due to the close connectivity between these ecosystems. Changes at the base of the food web will affect the energy flow from lowest to highest trophic level, with uncertain implications for top predators. Thus, it is crucial to identify to what extent the food web depends on carbon produced by sea-ice algae (sea-ice-derived carbon) in comparison to carbon produced by pelagic phytoplankton. Based on state-of-the-art lipid and stable-isotope analyses, we investigated the trophic linkage between the sea-ice system and the associated food web in Eclipse Sound, Nunavut, Canada. Our results reflect the critical importance of sea-ice-derived carbon for benthic food-web processes during the spring ice-covered period, and give insights into the utilization of sea-ice-derived carbon by marine mammals. In Eclipse Sound where landfast ice is present for most of the year, changes in timing, magnitude and quality of sea-ice-derived carbon under alterations of climatic conditions and possibly increased shipping in this region might disrupt the important linkage between ice-associated, pelagic and benthic food webs. Consequently, the adaptive capacity of sea-ice-dependent species is anticipated to play a key role in structuring future food-web dynamics.


Insights into the functioning of sea-ice biogeochemistry: the BEPSII inter-comparison of 1-D sea-ice biogeochemical models

Letizia Tedesco, Martin Vancoppenolle, Pedro Duarte, Eric Mortenson, Giulia Castellani, Meibing Jin, Sebastien Moreau, Ben Saenz, Nadja Steiner

Corresponding author: Letizia Tedesco

Corresponding author e-mail: letizia.tedesco@environment.fi

Observations over the past few decades reveal active biochemical processes in the sea-ice zone. At the same time, large sea-ice changes are observed and projected for the future. These changes in the sea ice have the potential to affect the biogeochemical cycles and associated marine ecosystems, yet by what means, to what extent, and with what consequences are not well understood. These unknowns have motivated the implementation of model representations of sea-ice impact on biogeochemical processes in the past decades and the current need of an intercomparison among these models. The most common model type is process models, i.e. 1-D deterministic biogeochemical models that focus on the microbial ecosystem. They differ in, for example, their complexity in terms of the number of functional groups and processes, type and number of limiting nutrients, radiation scheme, vertical resolution, the coupling to an interactive sea-ice physical and/or ocean biogeochemical model. In this study, we investigate the capabilities of all of the participating models in the BEPII intercomparison experiment to reproduce very different biogeochemical conditions at two Arctic sites: a time series of landfast ice data collected at a very productive site in Resolute, Canada (2017), and a drift experiment in the Svalbard area (N-ICE, 2016) with a lead opening and refreezing in the middle of winter. Our driving questions are: What type of model is suitable for a certain scope? Can one size fit all purposes? How the vertical resolution of the model affect the goodness of the results? How can models help to integrate among sparse observations and time series? We attempt to answer these questions by running a set of experiments where we learn the models’ strengths and weaknesses by running the models as they are, and then by tuning them to possibly reproduce the observed variability.


Arctic sea-ice surface roughness from C- and L-band-frequency synthetic-aperture radar

Randall Scharien, Silvie Cafarella, Christian Haas

Corresponding author: Randall Scharien

Corresponding author e-mail: randy@uvic.ca

Sea-ice surface roughness modifies energy and momentum exchanges between the atmosphere and ocean, and controls snow and melt-pond distributions and subsequent spring melting rates. Surface-roughness information is needed to accurately retrieve sea-ice-freeboard and ice-thickness estimates from laser and radar altimeters. This study addresses the use of C- and L-band-frequency synthetic aperture radar (SAR) backscatter and polarimetric parameters for retrieving sea-ice surface roughness information at metre-scale spatial resolution. An airborne laser scanner and airborne topographic mapper (ATM) measured first-year sea-ice (smooth and deformed) and multiyear sea-ice surface-topography datasets, collected in the Canadian Arctic Archipelago (CAA) during late winter over two seasons (April 2016 and April 2017), facilitate the assessment of radar backscatter and polarimetric parameters from coincident ALOS-2/PALSAR-2 (L-band) and RADARSAT-2 (C-band) SAR systems for surface-roughness retrieval. Included is an examination of sea-ice evolution in the Victoria Strait portion of the CAA, where dynamic ice conditions occur in the predominantly landfast ice environment until late January, leading to significant deformation and greater surface roughness. Optimal SAR parameters for retrieving surface roughness are presented in two ways: (i) in the context of overall retrieval accuracy; and (ii) in the context of utilizing current mission data for pan-Arctic mapping in standalone mode, or in tandem with other sensors. Tandem sensors include the optical Multi-Angular Imaging SpectroRadiometer (MISR) and the Cryosat-2 radar altimeter. To elucidate surface-roughness controls on melt-pond distributions, derived roughness is compared to spatially coincident melt-pond fraction information from airborne-survey and high-resolution WorldView-2 satellite imagery collected during advanced melt conditions.


Storm-track responses to sea-ice loss impacted by ocean coupling

Paul Kushner, Alexandre Audette, Stephanie Hay, Russell Blackport

Corresponding author: Paul Kushner

Corresponding author e-mail: paul.kushner@utoronto.ca

In order to assess the broad impacts of Arctic sea-ice loss on global circulation and climate in the context of global warming, it is critical to understand which of these impacts are related to processes internal to the atmosphere and which rely on coupling between the ocean, atmosphere and sea ice. Recent modelling studies have highlighted the key role that dynamical coupling to an ocean model plays in shaping the climatic response to sea-ice loss. In response to imposed Arctic ice loss, ocean dynamical coupling amplifies warming of the Arctic troposphere, drives patterned warming of the mid latitude and tropical oceans, and promotes increased planetary-wave amplitudes. What these effects imply for storm-track responses Is complicated because of competition between lower and upper tropospheric changes and the interplay between storm tracks and stationary waves. A new technique of pattern scaling that combines information from greenhouse-warming simulations and targeted sea-ice loss simulations allow us to separate the Arctic and tropical drivers of the coupled response to ice loss. This separation shows how storm tracks over the Pacific and Atlantic basins shift equatorward and generally weaken in response to ice loss, with a degree of competition from tropical driving. The effect of midlatitude ocean warming in response to ice loss can also be isolated. This is accomplished by extracting the midlatitude sea-surface temperature (SST) response to sea-ice loss from atmosphere–ocean GCM (AOGCM) simulations and imposing it as a separate boundary forcing in an atmospheric general circulation model (AGCM). This midlatitude ocean warming drives significant Arctic heating and its impact on storm tracks and weather extremes is now being assessed. Preliminary expectations are that the midlatitude ocean-warming signal is responsible for a considerable fraction of the storm-track response, and results of this analysis will be reported on. Important questions remain about the dynamical drivers of all these responses, and the possibility of observing such forced signals in nature, starting from greenhouse gas forcing and leading to predicted changes in storm tracks.


Sea-ice thickness with ICESat-2

Alek Petty, Nathan Kurtz, Ron Kwok, Thorsten Markus

Corresponding author: Alek Petty

Corresponding author e-mail: alek.a.petty@nasa.gov

NASA’s new ICESat-2 mission is now providing us with continuous, very high-resolution surface-elevation data over the polar oceans, enabling us to characterize the entire polar sea-ice thickness distribution in unprecedented detail. In this presentation we demonstrate our current processing chain for converting the official along-track ICESat-2 freeboard product (ATL10) into sea-ice thickness. We primarily make use of snow depth and density data from the NASA Eulerian Snow on Sea Ice Model (NESOSIM) forced by ERA-I snowfall/winds and satellite-derived concentration/drift products, and compare this with thicknesses derived from other snow reconstructions and modified versions of the Warren climatology. The coarse resolution (~100 km) snow data are redistributed onto the high-resolution (~10–100 m) ICESat-2 footprint using the along-track freeboard and the large-scale (~100 km) snow depth/freeboard using relationships obtained from data collected by NASA’s Operation IceBridge mission. We include thickness uncertainty estimates and attempt to fully account for the uncertainties introduced across the data-processing chain, while also acknowledging the challenges associated with our limited knowledge of these underlying uncertainties. We present regional sea-ice thickness distributions and highlight their seasonal evolution through our first winter of data collection and show comparisons of these data with thickness estimates obtained from Operation IceBridge and ESA’s CryoSat-2 mission. Finally, we highlight the generation of a gridded thickness product and the availability of along-track and gridded ICESat-2 sea-ice-thickness datasets.


A persistent, long-range hybrid autonomous vehicle for under-ice observation

Richard Camilli, Ted Maksym, Angelos Mallios, Brian Claus

Corresponding author: Ted Maksym

Corresponding author e-mail: tmaksym@whoi.edu

Autonomous underwater vehicles (AUVs) offer an attractive means of acquiring otherwise difficult-to-obtain observations under ice, but at present are limited by several key challenges, particularly limited endurance and the need for significant logistical support for longer-range vehicles. A new, inexpensive hybrid AUV is being developed for long-range, multi-month missions under sea ice. Built on a Slocum G3 platform, the hybrid AUV incorporates a low-power thruster, thus combining the long endurance of a glider with the speed and mission capability of a traditional AUV. Through adaptive buoyancy- and thruster-driven propulsion, very efficient transit can be achieved in areas ranging from shallow coastal to deeper shelf-break regions across tuneable depth bands ranging from 10 to 1000 m. The vehicle will be capable of load-dependent ranges of 2000–5000 km. With an integrated low-power ice-observing sonar and the capability to efficiently operate at relatively constant range from the sea ice afforded by the thruster, the hybrid AUV offers the potential for both large-scale ocean observation and sea-ice survey. Vehicle development is built around an operational strategy of low-cost unsupported missions by minimizing the need for vessel support or external navigational aids. Accurate navigation can instead be maintained under ice without surfacing for extended periods through a combination of Doppler bottom tracking and terrain-aided navigation. These capabilities are demonstrated in recent ice-free missions mapping the seafloor off Costa Rica. With these capabilities, a future fleet of these vehicles offers the potential for year-round under-ice operation over significant expanses of the polar sea-ice cover, and with the prospect of launching and recovering from more accessible locations near shore, with low resource and logistical costs.


Impact of atmospheric forcing on the interannual variability of polynyas in landfast sea ice off Baffin Island, 1985–2017

Marzena Marosz-Wantuch, Christian Haas, J. Alec Casey

Corresponding author: Marzena Marosz-Wantuch

Corresponding author e-mail: marzenam@yorku.ca

Sensible heat polynyas in landfast sea ice are a widespread phenomenon in the Canadian Arctic and are sensitive indicators of climate change. They develop due to locally enhanced ocean heat flux and exhibit spatio-temporal variability resulting from variations in oceanic and atmospheric forcing. This study assesses the influence of atmospheric boundary conditions on the interannual variability of polynya dynamics at Qikiqtarjuaq, Nunavut, between 1985 and 2017, observed with time series of optical satellite imagery. Small polynyas, persistent in the study area and other regions of the Canadian Arctic Archipelago, are difficult to discern by commonly used, satellite optical and microwave sensors with coarse spatial resolution. However, Landsat and Sentinel-2 optical sensors, which have finer spatial resolution, allow us to extract open-water areas through image-classification techniques, and subsequently to derive polynya dynamics. We compare the variability of derived opening dates, growth rates and duration of polynyas with the variability of atmospheric forcing such as incoming shortwave radiation, air temperature, wind speed and snow depth. While these small polynyas respond to atmospheric forcing at all stages of their development, their interannual variability cannot be solely explained by any single atmospheric factor. Multiple regression analysis indicates that atmospheric boundary conditions affect polynya dynamics, but the ocean heat flux is the dominant factor contributing to their interannual variability. The processes governing polynya dynamics are intricate and interdependent; they therefore require detailed, long-term observations, which are becoming increasingly possible with the growing number of remote-sensing sensors characterized by fine temporal and spatial resolution.


Distinct pelagic and sympagic zooplankton energy flows under multiyear sea ice determined using fatty-acid and stable-isotope biomarkers

Steven Duerksen, Doreen Kohlbach, Anke Reppchen, Pascal Tremblay, Benjamin Lange, Joannie Charette, Ron ten Boer, Jana Hildebrand, Philipp Anhaus

Corresponding author: Steven Duerksen

Corresponding author e-mail: steve.duerksen@dfo-mpo.gc.ca

Old, multiyear sea ice has been a constant feature of the high Arctic for millennia. However, due to climate change, the minimum ice extent is at historic lows and the Arctic Ocean may be ice-free during the summer within a few decades. Little is known about the roles that multiyear ice may play in marine food webs, and what effect its loss will have across the ecosystem. We collected pelagic and sympagic zooplankton using vertical tows and a net mounted on a remotely operated vehicle during the 2018 Multidisciplinary Arctic Program Last Ice field season in the Lincoln Sea and analyzed their stable-isotope and fatty-acid signatures. These signatures were compared with primary-producer fatty acids collected from first-year ice, multiyear ice and phytoplankton. Calanoid copepods represented the bulk of our pelagic samples, although ostracods, amphipods and gelatinous zooplankton were also common. The amphipods Apherusa glacialis and Gammarus wilkitzkii were the dominant sympagic species sampled. While some coupling between the ice and the water column is evident, the energetic basis of pelagic and sympagic zooplankton appears to be very distinct. Dinoflagellate-related biomarkers were more prevalent in pelagic organisms compared to ice-associated zooplankton. The similarities of ice-based herbivore biomarker signatures and their predators were much greater than those of the pelagic system, indicating greater dietary specialization. Our results provide insight into the energetic contribution of multiyear sea ice to Arctic marine food webs, and will be critical in determining what the effects of continued ice loss could mean for these systems.


Characterizing Arctic cyclone precipitation and the effects on sea ice in a changing climate

Chelsea Parker, Melinda Webster, Linette Boisvert, Priscilla Mooney

Corresponding author: Chelsea Parker

Corresponding author e-mail: chelsea_parker@brown.edu

Arctic cyclones are rapidly intensifying low-pressure systems of variable size (~200–1500 km radius). They can be associated with increased longwave downward radiation, mixed-phase clouds, reduced shortwave downward radiation, strong winds, and heavy, persistent precipitation. Arctic cyclones are therefore a major mechanism for heat and moisture transport into the central Arctic from lower latitudes. The effects on sea ice and Arctic energy balance are complex and vary with the cyclone characteristics, their geographic setting and the time of year. Arctic cyclones affect both the dynamics and thermodynamics of the sea-ice pack. Surface winds can influence sea-ice motion; and temperature advection can result in sea-ice melt, reduced growth and altered freeze-up/melt-onset timings. However, Arctic cyclone precipitation as snowfall represents a significant input to the snowpack on sea ice, with important implications for sea-ice mass balance. Conversely, cyclone rainfall can inhibit snowpack development over sea ice and encourage melt. Observational data remain very sparse at these high-latitude, maritime locations; and global models and reanalysis datasets are often too coarse to capture Arctic cyclone dynamics. This study uses high-resolution numerical weather modeling techniques, in concert with observations (e.g. in situ ice-mass-balance buoy and remotely sensed CloudSat data), to provide insight on the precipitation phase, rate and accumulation footprint over sea ice during Arctic cyclones. We utilize the US National Center for Atmospheric Research Weather Research and Forecasting model in a regional two-domain configuration of 36/12 km horizontal grid spacing to diagnose the precipitation dynamics over the lifecycle of cyclones from the Atlantic and Pacific basins, specifically during summer and winter seasons. The effects of climate change on Arctic cyclone activity, and thus the Arctic sea-ice pack, are complex and still unresolved. This work employs the regional, high-resolution modeling framework and the Climate Model Intercomparison Project future climate projections to gain new insights into these multifaceted relationships and feedbacks, and explore the relative change in cyclone precipitation over sea ice. Importantly, this study examines the range in sensitivity of the Arctic cyclone precipitation by contrasting the response to climate forcing from the cases in disparate geographic regions and seasons.


From science to policy in the Greater Hudson Bay Marine Region: an integrated regional impact study of climate change and modernization

Zou Zou Kuzyk, Lauren Candlish, Michelle Kamula, David Barber

Corresponding author: Lauren Candlish

Corresponding author e-mail: lauren.candlish@umanitoba.ca

The Greater Hudson Bay Marine Region, comprising Hudson Bay, James Bay, Foxe Basin, Hudson Strait and Ungava Bay, is the largest inland sea in the world to experience a seasonal sea-ice cover. Fresh water enters the Marine Region from its large watershed, which covers a third of the Canadian landmass. Its ecosystems are broad and varied, with both year-round presence and seasonal abundance of fish, birds and marine mammals. There are forty coastal communities, mostly Inuit and Cree, who have depended on the waters and icescapes for their food, culture, mobility and livelihoods for millennia. The significance of environmental change in the Marine Region is thus most profound for these people; yet the results of scientific studies have not been widely disseminated outside academia nor influential with regard to policy. Inuit and Cree priorities such as changes in coastal sea ice affecting safety of travel have escaped scientific agendas. To integrate research results and communicate them to communities, northern organizations and other interested parties, ArcticNet initiated an Integrated Regional Impact Study (IRIS) process. For this process, the north was divided into four, broadly-defined regions: Western and Central Arctic, Eastern Arctic, Greater Hudson Bay Region, and Eastern Subarctic. The regions do not reflect land-claim boundaries but rather similarities or commonalities in important aspects of the environment. We have recently completed the first Greater Hudson Bay Marine Region IRIS report. This IRIS incorporates results from scientific studies, published compilations of traditional knowledge, insights of Inuit and Cree represented through the IRIS steering committee, input from community meetings at which components of the IRIS were presented, and comments from a variety of stakeholders who contributed to the editorial team. The IRIS consists of two parts: a large report of science-based knowledge, and a synthesis of this knowledge along with resultant policy-related recommendations. However, the best way to understand the Hudson Bay IRIS is not as an end, but as a substantial step in the continual process of bringing together knowledge to inform decision-making. In this presentation, we will summarize the process, which was an example of cooperation among the regional, national and local partners and academics. We will give an overview of the report and the key recommendations, including future research priorities.


The contribution of sympagic algae to past and future primary production in the Arctic Ocean

Letizia Tedesco, Marcello Vichi, Enrico Scoccimarro

Corresponding author: Letizia Tedesco

Corresponding author e-mail: letizia.tedesco@environment.fi

The Arctic sea-ice decline is among the most emblematic manifestations of climate change and is occurring before we understand its ecological consequences. Some work has been done on the observed and predicted changes in pelagic primary production in the Arctic Ocean as a consequence of sea-ice decline. More recently, new insights have been provided on the increase in extent and frequency of sub-ice and fall phytoplankton blooms. However, little attention had been paid to: i) the sympagic (sea-ice-associated) primary production that has been lost as a consequence of the physical loss of habitat in large areas of the Arctic Ocean that became permanently ice-free, and ii) the potential increase in sympagic production in the other areas of the Arctic Ocean as a consequence of the thinning, warming and rejuvenating of Arctic sea ice. Combining sea-ice physical drivers from an ensemble of 18 CMIP5 climate models with a state-of-the-art sea-ice biogeochemical model for sympagic algae, we investigate here: i) the past (since 1961) and future (until the end of the century, under an extreme scenario of no mitigation measures) losses of Arctic sea-ice extent, and we estimate the relative losses of sympagic primary production and any associated trend; ii) the future changes (until the end of the century with respect to 1961, under an extreme scenario of no mitigation measures) in FYI algal productivity along latitudinal gradients. Model projections indicate diverse latitudinal responses, pointing to the fact that the impact of declining sea ice on Arctic sympagic production is large and also complex, as well as the expected trophic and phenological cascades on the rest of the food web. Are we losing an important pulse of energy for the functioning of the Arctic Ocean ecosystem?


New coordinated simulations to help explain Arctic–midlatitude linkages mediated by moist overturning atmospheric circulation responses

Alexandre Audette, Paul Kushner, Robert Fajber, Yannick Peings, Yannick Peings

Corresponding author: Alexandre Audette

Corresponding author e-mail: alexandre.audette@utoronto.ca

An important advance in our understanding of the atmospheric general circulation is to recognize the role of large-scale moisture transport in setting the stratification of the midlatitudes and polar regions and in determining the pole-to-equator temperature gradient. Thus the interplay of large-scale moist transport with regional feedback-driven processes in the Arctic under global warming and sea-ice loss is critical to understanding Arctic–midlatitude linkages. This paper focuses on how the moist overturning circulation, which connects the midlatitudes to the Arctic atmosphere, responds to sea-ice loss in the latest-generation Earth-system models. This will enable identification of real-world observational targets and socio-economic impacts on precipitating systems and storm tracks that would be related to changes in the moist circulation. This work is being carried out within the framework of the Polar Amplification Model Intercomparison Project (PAMIP), which is a contribution to the Sixth Coupled Model Intercomparison Project (CMIP6). PAMIP is carrying out a series of coordinated numerical model experiments to address, among other issues, the relative roles of Arctic sea-ice loss and remote-ocean warming in driving polar amplification, and the global climate system response to Arctic ice loss. Coordinated experiments with multi-model simulations forced with different combinations of sea ice and/or sea-surface temperatures (SSTs) representing present day, pre-industrial and future conditions are being carried out. The principal purpose of this paper is to identify primary responses in the moist overturning circulation and its poleward heat transport across the model set in PAMIP. Using a diagnostic technique known as ‘STEM’, two effects can be readily separated: the impact of PAMIP forcing of sea-ice loss or SST boundary-condition changes on stratification and on local heat transport. Ice loss is generally expected to weaken poleward heat transport and the moist overturning circulation, but how strong this effect is versus remote SST warming and changes to stratification is poorly understood. First PAMIP results will be reported at a workshop in the UK in June 2019, based on initial simulations being generated in spring 2019 from the PAMIP collaboration. These results will be reported here with a focus on moist-circulation diagnostics using STEM and comparing SST and Arctic ice-loss changes.


Historic and future trends of simulated pan-Arctic freshwater discharge: changes in seasonality and volumes due to climate change

Marie Broesky, Tricia Stadnyk, Stephen Dery

Corresponding author: Marie Broesky

Corresponding author e-mail: wredem@myumanitoba.ca

The pan-Arctic domain is observed to be one of the areas most significantly impacted by recent changes in the global climate. Oceanic circulation and sea-ice dynamics are significantly impacted by climate-induced changes in terrestrial freshwater discharge entering the Arctic Ocean from the pan-Arctic domain. Freshwater changes, and the anticipated variability in both their timing and magnitudes, are often inadequately represented in sea-ice/ocean models. This study identifies trends in freshwater discharge for the total freshwater contribution into the Arctic Ocean, as well as for the top 12 (by volume) Arctic rivers. Freshwater discharges simulated by the HYPE model from 1981–2070 were studied. Simulated flows reflect the average conditions of two different climate-change scenarios and three different climate models from the BaySys group of projects. The Mann–Kendall test was used to analyse trends in freshwater discharge and their significance for nine decadal periods, and using a decadal moving- window approach over the entire study period. Changes in seasonality were investigated by examining statistically significant shifts in daily annual average hydrographs evolving over the historic period. This study aims to further inform the marine community about the projected dynamic changes in terrestrial freshwater supply received by the Arctic, with the express intent of providing valuable insights and gridded runoff datasets for use in sea-ice/ocean models. Understanding the possible changes to freshwater discharge can aid in the preparation for drastic future alterations to the Arctic system that humanity is likely to face.


Temporal and spatial evolution of Foxe Basin sea ice (1971–2018)

Slawomir Kowal, William Gough, Ken Butler

Corresponding author: Slawomir Kowal

Corresponding author e-mail: kowal@utsc.utoronto.ca

Previous research has found Hudson Bay seasonal sea ice particularly sensitive to climate change with a strong signal towards an earlier break-up, later freeze-up and a longer ice-free season. The work presented here focuses on the nearby region of Foxe Basin in order to determine whether the same phenomenon, and its observed patterns, holds true for this region during the 1971–2018 time period for all three of the metrics, by utilizing the same methodologies applied to the three Hudson Bay datasets pertaining to the three metrics. These methodologies currently focus on the temporal trends and spatial patterns of sea ice through the use of a superimposed grid of 24 sampling locations spread uniformly across Foxe Basin. The analysis of the temporal trends reveals that 12 locations show statistically significant trends for the break-up metric, 20 for the freeze-up metric and 18 for the ice-free metric. The average magnitude of the trend for all 24 locations for an earlier break-up is 0.43 days per year, for a later freeze-up it is 0.54 days per year and for a longer ice-free period it is 1.01 days per year. Lastly, accelerating temporal trends have been observed in specific regions of the Basin for all three metrics. Three points for the break-up period, three points for the freeze-up period, and five points for the ice-free season demonstrate accelerating temporal trends. The spatial analysis was performed using clustering statistics by employing two clustering methods (Ward’s and K-means). The two clustering methods produced consistent results with some small variations and the clusters of sea-ice data points were found to be consistent with observed patterns of sea-ice break-up and freeze-up. Clustering along the coasts for all three metrics had the greatest coherency, suggesting that the connection with the shoreline provided a constraint that enabled the ice within a cluster to act in a coherent fashion. There was less consistency in clustering of points in the central regions and the two methods differed in this regard, but not substantially.


Encouraging K-12 math interest through sea-ice dynamics

Jennifer Hutchings, Jay Well, Ignatius Rigor, Amy Solomon

Corresponding author: Jennifer Hutchings

Corresponding author e-mail: jhutchings@coas.oregonstate.edu

Many of us are requested to broaden our research to education and public outreach. The International Arctic Buoy Program and many of our colleagues have participated in school outreach through ‘Adopt a Buoy’ and ‘Float your Boat’ school activities. During the MOSAiC campaign we will be expanding upon these programs. The Oregon State SMILE program presents lesson plans for elementary-, middle- and high-school students that engage students in their grade-appropriate mathematics curriculum. Navigation and forecasting are introduced through tracking and forecasting buoy trajectories, with math skills from fractions and the concept of differentiation explored in the context of their use in our research. In 2019/20 the lessons will be tested in Oregon schools and SMILE clubs. Lesson plans will become freely available, after trials, to extend your IABP/IPAB and research activities. Elementary- and middle-school pupils will send drifters to the Arctic and track them. High-school students will forecast drifter projects, developing persistence and rule of thumb models whose performance will be compared to model-derived sea-ice drift forecasts. We invite you to join these students in providing a sea-ice drift forecast. Can you do better than persistence?


Mechanical response of sea ice to atmospheric wind stress

Björn Erlingsson, Cathleen Geiger, Jennifer Lukovich

Corresponding author: Björn Erlingsson

Corresponding author e-mail: bjorn@uw.is

In this presentation it is shown that a generalized yield condition of sea ice as a function of isotropic pressure and shear (maximum shear stress) gives rise to the notion of (angle of) internal friction. The far-field mechanical forcing is identified as a residual stress component of thermally induced stress records from SHEBA (1997/98) and SEDNA (2007). Analysis of the far-field mechanical forcing gives rise to information on the sea-ice stress and drift response to atmospheric forcing and provides information on large-scale mechanical properties (yield condition) of the sea ice. It is shown that analysis of stress-gauge data and ice-motion kinematics yields information on: 1) internal friction and its relation to wind stress forcing; 2) cohesion and its seasonal evolution of the ice field and; 3) impact of rate of deformation (scales) from kinematic data. This experimental approach sheds light on the constitutive relationships for sea-ice dynamical modeling and its response to atmospheric loads and conditions.


Winter and summer cyclone impacts on Arctic sea ice

Jennifer V. Lukovich, Julienne Stroeve

Corresponding author: Jennifer V. Lukovich

Corresponding author e-mail: Jennifer.Lukovich@umanitoba.ca

In this study we investigate the impact of individual storms on Arctic sea ice in winter and summer. Examined in particular are relative thermodynamic and dynamic contributions to the sea-ice concentration and effective ice-volume budgets in the vicinity of Arctic summer cyclones in 2012 and 2016. Results from investigation of cyclone impacts on Arctic sea ice in summer illustrate sea-ice loss associated with thermodynamic processes (melting) in the Pacific sector of the Arctic in 2012, and both thermodynamic and dynamic processes in the central Arctic in 2016 in the vicinity of cyclone trajectories. Comparison of both years further suggests that the Arctic minimum sea-ice extent is influenced by cyclone timing and location relative to the sea-ice edge. Examination of sea-ice effective volume changes in the vicinity of Arctic cyclones in winter in 2011/12 and 2015/16 shows that sea-ice volume decreases in the vicinity of cyclones located in the central and Pacific sector of the Arctic, and increases in the vicinity of cyclones located near the sea-ice edge. Central to a characterization of cyclone impacts on Arctic sea ice from the perspective of thermodynamic and dynamic processes is an evaluation of societal implications including navigation, shipping, conservation and designation of protected and ecologically sensitive marine zones, in addition to non-local responses. We present as a contribution to this characterization an index describing relative residual (thermodynamic and ridging) and dynamic contributions to sea-ice concentration and volume changes to quantify and improve our understanding of sea-ice state and dynamical responses to cyclones in a rapidly changing climate and warming Arctic.


Characterization of the spatial and temporal patterns of the northwest polynya in Hudson Bay

Jennifer Bruneau, David Babb, Wayne Chan, David Barber, Geoffrey Gunn

Corresponding author: Jennifer Bruneau

Corresponding author e-mail: Jennifer.Bruneau@umanitoba.ca

Polynyas play an important role in atmosphere–sea-ice–ocean interactions in polar regions as they provide a source of heat and water vapour to the atmosphere and promote mixing and upwelling in the underlying ocean. They influence local weather and climate, affect biogeochemical cycles, and provide a biological hotspot for marine wildlife. Within the dynamic seasonal ice cover of Hudson Bay, there is a large latent-heat polynya that forms throughout winter in the northwest as a result of strong northwesterly surface winds. Due to its role in atmosphere–ocean feedback mechanisms and subsequent influence on the marine ecosystem in Hudson Bay, this research seeks to examine the interannual variability of the northwest polynya’s spatial-temporal characteristics throughout the ice season. Using a 17-year record of high-resolution sea-ice concentration data from the Advanced Microwave Scanning Radiometer (AMSR-E and AMSR-2) we explore the spatial-temporal characteristics of the polynya. Subsequently, we link polynya events and changes in the polynya’s activity to atmospheric forcing and local sea-ice conditions. In this way, the atmosphere’s influence on the production and maintenance of the northwest polynya can be determined.


Characterization of 18O and D fractionation during sea-ice growth and melt

Heather Kyle, Feiyue Wang

Corresponding author: Heather Kyle

Corresponding author e-mail: umkyleh@myumanitoba.ca

The thickness and extent of Arctic sea ice is rapidly changing, so monitoring ice growth and melt rates in the region is of major concern. In first-year Antarctic columnar sea ice, growth-rate estimates using 18O fractionation factors have been fairly well studied. However, differences between Arctic and Antarctic sea ice mean that ice-growth-rate estimates are less well established in the Arctic Ocean. In addition, δ18O in sea ice has been used to determine the contribution of snow and snow-ice to ice mass balances. Until recently, snow and snow-ice were thought to be fairly minor components of Arctic sea ice, although this is beginning to be re-evaluated with increased precipitation in the Arctic. Historically, δ18O has been the main stable water isotope used to estimate sea-ice growth rates and the contribution of snow to sea-ice mass balances. To refine the existing knowledge of water-isotope fractionation in sea ice and to examine how micro-scale processes may affect the fractionation, a series of ice cores were collected from the Sea-ice Environmental Research Facility to determine variations in δ18O, δD, temperature, and salinity profiles during sea-ice growth and melt between January and March 2019. In addition, the underlying seawater, sea-ice brines from various depths, and snow cover were sampled throughout the experiment to examine variations in fractionation factors throughout the sea-ice growth–melt cycle. Due to difficulties with sampling brine, in the past the contribution of ice-brine fractionation to growth-rate models has generally been estimated. By sampling brine from various depths at SERF, it will be possible to more accurately incorporate the effects of brine in sea ice into ice-growth-rate estimates.


A spring barrier for regional predictions of summer sea ice

Mitch Bushuk, David Bonan, Michael Winton

Corresponding author: Mitch Bushuk

Corresponding author e-mail: mitchell.bushuk@noaa.gov

How far in advance can skillful regional predictions of summer Arctic sea-ice extent (SIE) be made? In this work, we present evidence for a springtime predictability barrier, which imposes a relatively sharp limit on the potential skill of regional summer SIE predictions. Using pre-industrial control runs and perfect model predictability experiments performed with global climate models (GCMs), we find that summer SIE prediction skill drops significantly for forecasts initialized prior to a given springtime barrier date. This barrier date is model-dependent and ranges from 1 June to 1 May. The existence of a predictability barrier is found to be remarkably universal across GCMs, with nearly all models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5) displaying a predictability barrier structure. We perform a sea-ice mass-budget analysis with daily temporal resolution to explore the physical mechanisms underlying the spring predictability barrier. Finally, we utilize large ensembles of historical and future scenario GCM simulations to investigate how the structure and characteristics of the predictability barrier are expected to evolve under a changing climate. We discuss the implications of these results for future observational networks and seasonal prediction systems.


Modeling the sea-surface temperature over the Hudson Bay complex

Shabnam Jafarikhasragh, Jennifer V. Lukovich, Xianmin Hu, Paul G. Myers, Kevin Sydor, David G. Barber

Corresponding author: Shabnam Jafarikhasragh

Corresponding author e-mail: jafariks@myumanitoba.ca

Understanding the mechanisms by which an ocean general circulation model (OGCM) produces sea surface temperature (SST) is of fundamental interest for both the modelling and surface climate communities. Sea surface temperature (SST) from four NEMO (Nucleus for European Modelling of the Ocean) model simulations is analysed to study the bulk flux parameterization used to compute SST over the Hudson Bay complex for the summer months (August and September) from 2002–09. NEMO was forced with two atmospheric forcing sets of different resolutions, the Coordinated Ocean-ice Reference Experiment, version 2 (CORE 2) as the lower-resolution atmospheric forcing, and the Canadian Meteorological Centre’s Global Deterministic Prediction System Reforecasts (CGRF) as the higher-resolution atmospheric forcing. The CGRF forcing is also implemented in the third and fourth runs using different runoff data (integrated Hydrological Predictions for the Environment (HYPE) versus Dai and Trenberth) and different NEMO resolutions (1/12° versus 1/4°). Results from these different sensitivity experiments demonstrated the low impact of the model resolution and freshwater runoff on the simulated SST in the HBC. Long-wave fluxes for different atmospheric forcings were, however, shown to be the main contributor to heat flux and thus simulated SST differences in HBC in summer when compared to observational satellite SSTs.


On the origin and composition of surface fresh water in winter around the Belcher Islands, southeast Hudson Bay

Rosemary Eastwood, Rob Macdonald, Jens Ehn, Joel Heath, Lucassie Arragutainaq, David Barber, Zou Zou Kuzyk

Corresponding author: Zou Zou Kuzyk

Corresponding author e-mail: zouzou.kuzyk@umanitoba.ca

Hudson Bay is undergoing climate-driven changes in sea-ice and freshwater inflow and seasonal redistribution of river inflow due to regulation. In recent winters, Inuit have observed unprecedented rapid freezing of leads and polynyas around the Belcher Islands in southeast Hudson Bay. To investigate potential connections between runoff and winter oceanographic conditions in this area, we initiated the first winter study of these shallow waters by collecting conductivity temperature–depth (CTD) profiles, water samples and ice cores during four campaigns between January 2014 and March 2015. Tandem measurements of salinity and δ18O were made for the water and ice samples to discriminate between river runoff and sea-ice melt. The observations under sea ice in winter show that southeast Hudson Bay has an offshore and a coastal domain, the latter distinguished by the presence of excess river water and strong stratification. The amount of river water around the Belcher Islands increased significantly from fall through to late winter, counter to expectations for a system in which the greatest river inflows occur during spring freshet (May–June). δ18O records of ice show the increased presence of river water in January and its persistence throughout winter. The river water in surface waters during the winter is associated with brine, which considerably exceeds the capacity of local ice to produce. We infer that brine is advected into the study area together with river water, and that interplay between these properties establishes the thickness of the mixed layer. The river water in the coastal domain in winter acts to reduce the thickness of the surface mixed layer relative to that in the offshore domain (~20 m vs 40 m). We propose that, before hydrological development, winter river inflow to James Bay was less capable of sustaining a freshwater surface layer along the coast, with the result that the brine produced by rapid sea-ice growth in the leads led to mixing in the water column, destroying stratification and, thus, hindering plume transport. A shift in runoff to winter is thus likely a key driver of transport of both river water and brine, with consequences for deep convection in the flaw lead. Delayed freeze-up in Hudson Bay under a warmer climate would reduce brine early in the winter, promoting stratification and river plume transport.


Physical and biological properties of Antarctic winter sea ice: Ross Sea vs Weddell Sea

Jean-Louis Tison, Ted Maksym, Stephen Ackley, Sharon Stammerjohn, Sarah Wauthy, Fanny van der Linden, Gauthier Carnat, Célia Sapart, Johannes de Jong

Corresponding author: Jean-Louis Tison

Corresponding author e-mail: jtison@ulb.ac.be

Physical and biogeochemical properties of Antarctic winter pack ice are typically under-sampled. Here, we present the results of physical and biological investigations on early-winter pack ice in the Ross Sea (April–June 2017). Ice textures, temperature, bulk salinity, brine volume, brine salinity and Rayleigh number help us characterize the physical environment in which biological activity will be able to develop in early winter, with contrasted behavior between polynyas, the marginal ice zone (MIZ) and the central Ross Sea. As expected, the very dynamic Terra Nova Bay polynya and the MIZ are dominated by granular ice, while columnar ice takes over the central Ross Sea, with decreasing proportions as one progresses north from the Ross Sea polynya. Snow is present in negligible amounts (a few centimeters) almost everywhere, except in the northern section of the central Ross Sea (10–20 cm), which has a longer and more spatially extended growth history, as shown by satellite imagery. The latter also clearly shows that the southern section has a much shorter growth history, due to a very late ice-growth onset, a remarkable feature of year 2017. The result is a globally low chl-a standing stock, as compared to previous Ross Sea winter cruises (1995–98). The Terra Nova Bay polynya, the MIZ and the southern central Ross Sea show low internal chl-a (≤1 μg L–1). A bottom community, however, develops with time and ice growth, showing higher chl-a concentrations, both from the outer skirt of the Terra Nova Bay polynya northward, and from the MIZ southwards. The highest chl-a levels (>30 μg L–1) are found in rafted coastal floes within the transition zone to the central Ross Sea. Finally, the northern central Ross Sea floes, with a longer growth history and thicker ice cover, show an increase in the internal community chl-a, potentially triggered by brine-tube development. A recent update on the biogeochemical impact of snow cover and cyclonic intrusions on the winter pack ice in the Weddell Sea has shown that winter 2013 (June–August) conditions were favorable to high ice permeability and cyclic events of brine movements within the sea-ice cover (via well-developed brine tubes), favoring relatively high internal chlorophyll-a (chl-a) concentrations. We will discuss if this contrast with the Ross Sea conditions in 2017 is solely due to the time difference in the season, or to contrasted initial conditions of sea-ice growth.


The third-generation ice mass-balance buoy: improvements, adaptations and use cases

Cameron Planck, Don Perovich

Corresponding author: Cameron Planck

Corresponding author e-mail: cameron.j.planck.th@dartmouth.edu

Sea-ice mass-balance buoys have been a modality for in situ determination of sea-ice mass since the 1990s. By directly measuring the time series of surface and bottom ice/snow position and the vertical profile of ice temperature, these buoys are able to resolve thermodynamic changes to local floe sea-ice mass balance to centimeter-level precision. Here we present a top-to-bottom redesign of the commonly used CRREL seasonal-ice mass-balance buoy (SIMB) platform. The buoy was designed using input from a comprehensive set of stakeholders, including the principal scientist, collaborating scientists, and the deployment and production teams. Emphasis was placed on reducing cost while maximizing reliability, adaptability, data quality and ease of installation. Several variations of this platform are presented, including models modified with additional sensors for ‘ocean’, ‘atmosphere’, and ‘distributed hub’ use cases. Data from recently deployed buoys are shown, and the reported precisions, accuracies, platform limitations, and future development plans are also discussed.


Weighing the A68 iceberg with ICESat-2

Sinéad L. Farrell, Helen Amanda Fricker, Kelly M Brunt, Maya Becker, Christopher Jackson

Corresponding author: Sinéad L. Farrell

Corresponding author e-mail: sineadf@umd.edu

In July 2017, a large tabular iceberg calved from the Larsen C ice shelf. A68A, which was over 7000 km2 in size, and larger than the US state of Delaware, reduced the area of the ice shelf by over 12%. Although its movement was initially limited, A68A rotated by approximately 90° after 1 year, and is currently drifting northeastward in the Weddell Sea, away from the Larsen C. Now under the influence of winds, tides and ocean currents, A68A continues to slowly rotate in the Weddell Gyre, and its southern portion has started to disintegrate. Here we track the initial calving, drift and disintegration of A68A using synthetic aperture radar (SAR) imagery acquired by the European Sentinel–1A and -1B satellites. Normalized radar cross section (NRCS) measurements allow us to accurately measure the iceberg’s size and trajectory, which we compare with seafloor bathymetry. The launch of NASA’s Ice, Cloud, and land Elevation Satellite-2 (ICESat-2), in September 2018 provides a new capability to precisely monitor A68A’s height above local sea level (freeboard). The Advanced Topographic Laser Altimeter System (ATLAS), carried onboard ICESat-2, is a photon-counting lidar comprising three pairs of laser beams, with each pair separated by ~3000 m, and a footprint size of ~17 m, sampled at 0.7 m in the along-track direction. The ICESat-2 mission requirement for height measurements precise to ±0.1 m allows ATLAS to not only obtain precise measurements of freeboard, but also build a detailed picture of iceberg surface topography for the first time. Initial assessment of ICESat-2 data suggests a freeboard of 37.5 m. We combine the iceberg’s area with estimates of its freeboard to estimate its total volume and mass. We estimated iceberg mass at ~2 trillion tonnes, enough ice to fill Lake Erie five times. This new estimate of A68A volume based on updated remote-sensing data is twice as large as initially reported.


Early assessment of ICESat-2 capabilities for Arctic sea ice

Sinéad L. Farrell, Ellen Buckley, Nathan Kurtz, Ron Kwok, Stefan Hendricks, Henriette Skourup, Tania G.D. Casal

Corresponding author: Sinéad L. Farrell

Corresponding author e-mail: sineadf@umd.edu

One of the high-priority science goals for the ICESat-2 mission is to estimate sea-ice thickness to examine ice–ocean–atmosphere exchanges of energy, mass and moisture. The mission requirement is to measure sea-ice freeboard, monthly, to an uncertainty of 0.03 m over 25 km length scales. After the successful launch of ICESat-2 on 15 September 2018, we have conducted initial analysis and quality assessment of Advanced Topographic Laser Altimeter System (ATLAS) photon-counting elevation data over Earth’s polar regions. We have exploited early ATLAS height measurements to derive sea-ice freeboard, the height of the sea ice above local sea level. Early results suggest excellent agreement in freeboard across all six ATLAS laser beams. The observations reveal very fine details of sea-ice topography including rough, multiyear sea-ice floes, leads, new ice growth, and pressure ridges within the ice cover, the first time such high-fidelity measurements have been acquired from a space-borne platform. Here we present an overview of the first assessments of ATLAS quality, and the utility of the data for measuring sea-ice surface height to estimate sea-ice thickness. We exploit synthetic aperture radar (SAR) acquired by the European Sentinel–1A and -1B satellites to validate ICESat-2 discrimination of leads and ice floes, and examine correlations with ice type. We describe a variety of community efforts to validate ICESat-2 sea-ice observations in Spring 2019 over sea ice in the central Arctic, Canada Basin and Beaufort Sea. We present details of three airborne campaigns – NASA IceBridge, the European Space Agency/DTU-Space CryoVEx mission, and the Alfred Wegener Institute IceBird experiments – that include flights designed specifically to validate ATLAS height measurements.


Modelling seasonal dynamic and thermodynamic contributions to sea-ice thickness variations in the Hudson Bay complex

Shabnam Jafarikhasragh, Jennifer V. Lukovich, Xianmin Hu, Paul G. Myers, Kevin Sydor, David G. Barber

Corresponding author: Shabnam Jafarikhasragh

Corresponding author e-mail: jafariks@myumanitoba.ca

The Hudson Bay complex (HBC), comprising Hudson Bay (HB), Hudson Strait (HS) and Foxe Basin (FB), is a sub-Arctic shallow inland sea, which is fully ice-covered during cold seasons in contrast to other water masses at the same latitude (e.g. the Labrador Sea, Baffin Bay). In this study the seasonal cycle of sea-ice-thickness variability in the HBC, together with the dynamic and thermodynamic processes that control it, are examined using the NEMO (Nucleus for European Modelling of the Ocean) numerical framework and its ice component, the Louvain-la-Neuve Ice Model version 2 (LIM2) from 2002–09. Results from this analysis show dominant thermodynamic contributions to changes in sea-ice thickness in the HBC compared with dynamic processes. Energy-budget considerations further demonstrate that the surface-energy flux governs ice-thickness changes, with secondary contributions from ocean heat flux in fall. In winter larger heat-flux release in northwest HB, north of FB and around the coastlines of HS, provides a signature of open water and polynya formation and maintenance due to prevailing winds and therefore the sources of escaping sensible heat flux. Sensible and longwave heat fluxes are also shown to have the most impact on the total surface heat flux in winter.


Zooplankton ascent and reproduction below drifting sea ice after winter in the Arctic Ocean

Haakon Hop, Anette Wold, Janne E. Søreide, Maja K.V. Hatlebakk, Amelie Meyer, Mikko Vihtakari, Allison Bailey, Philipp Assmy, C.J. Mundy

Corresponding author: Haakon Hop

Corresponding author e-mail: Haakon.Hop@npolar.no

The seasonal development of the zooplankton community was examined during the Norwegian young sea ICE (N-ICE2015) expedition in the Nansen Basin and over the Yermak Plateau from January–June 2015. During the expedition, the hydrography was dominated by northward-flowing warm saline Atlantic water at intermediate depth and southward-flowing cold Arctic water in the upper 100 m. This region is characterized by high inter-annual variability in the formation and break-up of sea ice, causing large variation in the timing of the ice-algae and phytoplankton blooms. Consequently, the zooplankton living there need to adapt to variable environmental conditions. The most pronounced change in the zooplankton community throughout the season was the increased abundance of the large herbivorous copepods Calanus finmarchicus and Calanus glacialis. Adults started to appear below sea ice in mid-May and nauplii at the end of May and beginning of June. Calanus hyperboreus was present in lower numbers than the two other Calanus species in surface water but occurred throughout the water column and contributed substantially to total zooplankton biomass due to its relatively large size. The species composition and the abundance of zooplankton in the two deepest layers (200–600 m, 600–1000 m) did not change much throughout the season. The deep samples were dominated by smaller species such as Oithona similis, Triconia borealis and Microcalanus spp. as well as typical deep-water species such as Heterorhabdus compactus, Gaetanus brevispinus and Paraeuchaeta norvegica. The largest increase in Calanus nauplii and copepodids was observed just after a rise in temperature and reduced salinity in surface waters in late May, due to the presence of Atlantic water at depth and a shallower mixed layer followed by a peak in chlorophyll a standing stock at the end of May. Subsequently the zooplankton biomass increased to a total of 149 mg dry weight m–3 in the upper 50 m in mid-June of which the Calanus species comprised approximately 90%. The springtime increase in biomass of Calanus spp. in surface waters was due to 1) the presence of relatively zooplankton-rich Atlantic water, 2) new recruitment, and 3) overwintering individuals ascending below sea ice. The Atlantic water inflow to the Arctic has led to reduced ice cover north of Svalbard and decreased surface salinity, as well as a northward expansion of Atlantic zooplankton, such as C. finmarchicus.


The Petroleum Environmental Research Laboratory: target and non-target analysis for characterization of petroleum products and their weathering in the environment

Jake Ritchie, Kim Hoang Luong, Ashish Sarker, Zou Zou Kuzyk, Feiyue Wang, Gary Stern

Corresponding author: Jake Ritchie

Corresponding author e-mail: jake.ritchie@umanitoba.ca

The Petroleum Environmental Research Laboratory (PETRL) at the Centre for Earth Observation Science, University of Manitoba, was created to conduct state-of-the-art mass-spectrometric chemical characterization and quantification of petroleum hydrocarbons and heteroatomic hydrocarbon compounds. As oil and its by-products contain thousands of compounds, each analyte requires different specifications for accurate characterization and quantification. PETRL instrumentation can separate and analyse extremely complex matrices to ultra-low concentrations of petroleum products from vastly different origins and undergoing various extents of weathering. The facility is equipped with one gas chromatography (GC) flame ionization detector (FID), two GC mass spectrometers (MS), a two-dimensional GCxGC-MS, and a liquid chromatography (LC) high-definition ion mobility MS. These instruments are configured to allow for targeted and non-targeted analysis of petroleum products and their weathering products in the environment. In addition to petroleum compounds, these instruments provide us with the capacity to analyse a wide range of compound classes such as lignin phenols, fatty acids, pesticides, industrial chemicals, natural products and many more.


Investigating sea-ice-type detectability and incidence angle effect using simulated compact polarimetric radar data during early and advanced melt conditions

Aikaterini Tavri, Randall Scharien

Corresponding author: Randall Scharien

Corresponding author e-mail: randy@uvic.ca

Large seasonal and regional variability in Arctic sea-ice conditions requires daily monitoring at vast spatial scales and regardless of weather conditions. Synthetic Aperture Radar (SAR) data have been traditionally utilized for operational monitoring and sea-ice charting. Single-, dual- and quad-polarimetric SAR data offer information for relationships between scattering mechanisms and sea-ice surface properties. However, detecting sea-ice types during the melt season remains a challenge due to the presence of meltwater on the sea-ice surface which interferes with radar signal. Compact polarimetric (CP) SAR represents a compromise between dual-polarized and quad-polarized SAR, reducing the complexity of system design and maintenance, while providing a larger swath coverage. Investigating CP parameters for sea-ice-type detection can provide extended spatial information for complex sea-ice backscatter signatures during melting conditions. Additionally, examining the incidence-angle effect is important for SAR datasets and has not been well studied for sea-ice detectability when using CP data. This study examines incidence angle and seasonal effects on simulated CP parameters from first year (FY) and multiyear (MY) ice targets, based on acquired RADARSAT-2 scenes of the Canadian Arctic Archipelago across early melt, melt onset and advanced melt regimes. Backscatter analysis of 32 CP parameters from FY and MY ice types is performed and the detectability potential of each parameter is evaluated. Spearman’s correlation and statistical tests are implemented for CP parameter selection and grouping. In addition, comparison between data acquired at steep and shallow incidence angles is presented for understanding the role of incidence angle on CP parameter behavior within the context of ice-type separability. The aim of this work is to provide a complete set of indispensable CP parameters for sea-ice-type classification over seasonal regimes, considering the influence of incidence-angle variability. The examined CP parameters are simulated to the 50 m medium-resolution (MR50) mode, which will be the operational mode of the upcoming Radarsat Constelation Mission (RCM), to evaluate the utility of the CP data for sea-ice classification and charting.


Higher-order statistical moments to analyze sea-ice drift patterns

Satwant Kaur, Jennifer V. Lukovich, Jens K. Ehn, David G. Barber

Corresponding author: Satwant Kaur

Corresponding author e-mail: kaurs34@myumanitoba.ca

In statistical analyses of geophysical phenomena, we often assume that observations will have Gaussian probability density functions (PDFs). However, geophysical systems are not necessarily Gaussian, and deviations from Gaussianity can shed light on the underlying dynamics. Deviations from Gaussian behaviour have been interpreted as signatures of nonlinear behaviour, which is a key issue in meteorology and sea-ice dynamics. In this study we use higher-order moments to identify patterns that distinguish sea-ice circulation in the Beaufort Gyre (BG) from transport associated with the transpolar drift (TPD) and changes in each as a function of space and time. Higher-order moments (skewness and kurtosis) of ice-drift speeds are examined over the winter period of 2006–17 to describe the persistence of features such as the BG and TPD, and their variation over time. Zonal and meridional components of ice drift are evaluated to determine the relative contributions of each to the BG and TPD. The shift in the centre of the gyre over time is also observed. The indexing patterns indicate that the periphery of the BG can be identified by a combination of high positive skewness and high positive kurtosis. Sea-ice concentration, mean sea-level pressure and surface winds are further analysed using higher-order moments to identify correspondence with, and physical mechanisms responsible for, distinctive patterns in sea-ice drift fields.


Sea-ice–atmosphere–ocean–biogeochemistry interactions in the winter Ross Sea: an overview of the PIPERS project

Stephen Ackley, Sharon Stammerjohn, Ted Maksym, John Cassano, Peter Guest, Jean-Louis Tison, Brice Loose, Peter Sedwick

Corresponding author: Stephen Ackley

Corresponding author e-mail: stevejacackley@gmail.com

For the Ross Sea, satellite data has shown increases in sea-ice extent, duration and concentration but it is not known how ice thickness, and thus volume, has changed. If sea-ice production has increased, where and why?, as answers to these questions bear on the large role that sea-ice formation also plays in water mass transformation in the Ross Sea region. The PIPERS (Polynyas and Ice Production and its seasonal Evolution in the Ross Sea) cruise on Nathaniel B. Palmer into the early winter Ross Sea took place between 11 April and 4 June 2017. The main objectives were to provide in situ data, previously lacking, to understand the role of the sea-ice–atmosphere–ocean–biogeochemistry interactions in the Terra Nova Bay and Ross Ice Shelf coastal polynyas and ice production on the continental shelf, in the determination of the sea-ice cover formation in the Ross Sea sector. Multidisciplinary and complementary activities were conducted in several different phases of the cruise (and thus at several different phases of the season). These activities were: ARGO and SOCCOM float deployments north of the sea ice; marginal ice zone wave buoy deployments and sampling activities; along-track sampling for physical oceanography, trace metals, and trace and noble gases; ice thickness, roughness and physical-property sampling at ice stations using lidar and an autonomous underwater vehicle; gas fluxes (e.g. CO2, dimethylsulfide, methane) and sea-ice-biogeochemistry sampling at ice stations and in polynyas; boundary-layer profiling using drone aircraft and met towers at ice stations; and atmosphere, ice and ocean investigations in the coastal polynyas. Highlights of the main results from the cruise using these newer technologies will be described here for these multiple phases of the cruise. Other presentations will be identified for additional specific details and results.


A probabilistic damped-persistence forecast of the sea-ice-edge location: a challenging benchmark for dynamical forecast systems

Bimochan Niraula, Helge Goessling

Corresponding author: Bimochan Niraula

Corresponding author e-mail: bimochan.niraula@awi.de

Increased presence of marine traffic and tourists in the Arctic in recent years has highlighted the necessity for meaningful prediction of sea-ice conditions at sub-seasonal to seasonal time scales. In this regard, reference forecasts based on present and past observations of the ice-edge location are important to benchmark the added value of dynamical forecast systems. However, the simplest types of reference forecasts – persistence of the present state and climatology – do not exploit the observations optimally and thus lead to overestimation of forecast skill. For spatial objects such as the ice-edge location, the development of damped-persistence forecasts that combine persistence and climatology in a meaningful way poses a challenge. We have developed a probabilistic reference forecast method that combines the climatologically derived probability of ice presence with initial (present) anomalies of the ice edge. We have tested and optimized the method based on minimization of the spatial probability score, using observed as well as idealized model data. The resulting reference forecasts provide a challenging benchmark to assess the added value of dynamical forecast systems.


The role of Greenland ice streams in discharge of fresh water in a warming climate

Dorthe Dahl-Jensen, Nicholas Rathmann, David Lilien

Corresponding author: Dorthe Dahl-Jensen

Corresponding author e-mail: dorthe.dahl-jensen@umanitoba.ca

Greenland is losing mass at an accelerating rate and will very likely be the largest contribution to sea-level rise in the next few decades. Mass loss from Greenland is from melt near the margin and from ice discharge from the many ice streams surrounding Greenland. The North Greenland Ice Stream (NEGIS) is the largest ice stream in Greenland, reaching from the central ice ridge of the Greenland Ice Stream to the margin, where it discharges ice through three ice streams: Nioghalvfjerdsfjorden glacier (NG), Zachariæ ice stream (ZI) and Storstrømmen glacier (SG). The NEGIS ice stream discharges 30 km3 a–1 of ice (27.5 × 109 m3 a–1 of fresh water) to the Fram Strait and through the last year an acceleration of ice streams followed by a retreat of the floating ice tongue of NG has been observed. For ZI the ice melange influences the discharge. The EGRIP ice camp is placed on the NEGIS ice stream and an international team is drilling a 2550 m deep ice core through the ice stream to understand the flow of ice in an ice stream. During the 2019 field season we expect to reach the basal fast-flowing ice. Results from the deformation of the EGRIP deep borehole will be presented together with the studies of the role of the shear margins in the area around the EGRIP camp. The role of the Greenland ice streams in a warming climate will be discussed based on the NEGIS findings.


A novel altimetry data processing technique for sea-ice freeboard estimation

Alejandro Egido, Laurence Connor

Corresponding author: Alejandro Egido

Corresponding author e-mail: alejandro.egido@noaa.gov

Satellite altimetry provides the means to measure sea-ice freeboard, an essential parameter to estimate ice thickness. Synthetic aperture radar (SAR) altimeters, such as the European Space Agency’s CryoSat-2 (2010–present) and Sentinel-3 (2016–present), have proved tremendously valuable for that task, particularly given their improved resolution compared with older satellite missions. However, observations of ice thickness are still challenging; in summer due to the presence of melt ponds, and in regions with numerous leads, which contaminate waveforms and complicate surface discrimination. Fully docused SAR (FF-SAR) is a novel data-processing technique that can further improve the resolution of observations over sea ice by using phase information from the standard SAR data to focus echoes along the satellite track. This allows a better representation of radar backscatter changes on the surface and therefore a better discrimination between leads and floes, which could in turn improve freeboard and thickness estimates compared with standard data-processing techniques. In this paper we describe the method of FF-SAR processing of CryoSat-2 and Sentinel-3A data over Arctic sea ice, and we present an evaluation of the data with coincident observations from NASA’s Operation IceBridge (OIB) airborne mission. We verify sea-ice lead and floe delineations from Sentinel-3A data with both standard SAR and FF-SAR processing methods, and assess the accuracy of Sentinel-3A surface-elevation measurements for both techniques via a comparison with independent OIB freeboard measurements.


Estimating spectral light attenuation coefficients for different types of Arctic sea ice

Ilkka Matero, Philipp Anhaus, Christian Katlein, Marcel Nicolaus

Corresponding author: Ilkka Matero

Corresponding author e-mail: ilkka.matero@awi.de

Transmittance of short-wave radiation through and inside Arctic sea ice varies with the stage of melt and type of the surface. The high heterogeneity of the surface due to varying snow depth and, for example, fractions of refrozen leads and melt ponds makes representing the attenuation challenging in models due to scales of variability being smaller than the model resolution. The fraction of short-wave radiation transmitted through sea ice is of importance for both biological and physical processes, as the heat can cause substantial internal and bottom melting and sustain productivity in algal communities adapted to low-light conditions. This work estimates characteristic floe-scale attenuation coefficients for different ice-surface types including first-year, multiyear, ponded and ridged ice with varying snow depth and ice thickness. These values are based on 12 Arctic field campaigns, and cover a multitude of primarily spring and summer sea-ice conditions from 2007–19. The measurements were obtained using a setup of RAMSES-ACC/ARC hyperspectral radiometers measuring downwelling shortwave irradiance both over the surface and under ice on different mobile platforms. Using one of the latest developments of such platforms, the remotely operated vehicle allowed for spatial coverage of up to several hundred meters, as well as estimating the ice thickness based on ice draft. The data of the under-ice light conditions is combined with colocated measurements of the surface conditions and snow depth to provide extinction coefficients for a variety of ice types in four spectral bands (320–480 nm, 480–580 nm, 580–740 nm and 740–920 nm). Earlier studies suggested that ice thickness might be a reliable predictor of under-ice light conditions when combined with knowledge of aspects such as ice type and pond coverage. Combining knowledge of the spectral transmissivity of different types of sea ice with Arctic-wide remote-sensing datasets of ice thickness, albedo and estimated ice age could allow development of a knowledge of internal and bottom heating, as well as availability of photosynthetically active radiation under ice. These parameterizations can thus help improve both physical and biogeochemical modeling approaches to sea ice.


Should recent observation-based sea-ice thermodynamic developments be incorporated into large-scale sea-ice models?

Martin Vancoppenolle

Corresponding author: Martin Vancoppenolle

Corresponding author e-mail: martin.vancoppenolle@locean-ipsl.upmc.fr

Thermodynamic processes are fundamental to the response of sea ice to climate change, control many sea-ice chemical processes and are instrumental in the suitability of sea ice for microbial life. New and revisited observations have recently fostered progress on sea-ice phase composition and enthalpy equation. Specific processes are also now better described, such as snow-ice, platelet-ice and melt-pond formation, as well as the motion of liquid brine. These results have been feeding an improved, physically more realistic, and increasingly consistent theoretical framework – combining the mushy-layer theory and a Gibbs-function-based equation of state, compatible with international thermodynamic standards for pure ice and seawater. However, climate models suggest that the state of sea ice in climate simulations is rather insensitive to such details of the thermodynamic representation of sea ice. Therefore, there is little chance that incorporating full-blown thermodynamic representations would affect the current answers to climate and sea-ice questions. Yet an advantage of upgrading sea-ice thermodynamics would be to enable a more precise description of changes in sea-ice scape – and of their impacts on sea-ice chemistry and ecology. Another benefit would be to decrease the exposure of sea-ice models to the objection of overly simple physics raised by our peers. Regardless of these modelling developments, current climate models repeat to us that the most pressing need to reduce climate-related sea-ice uncertainties, is to determine the contemporary volume of sea ice.


Arctic ice, freshwater–marine coupling and climate change

Dorthe Dahl-Jensen, David Barber

Corresponding author: Dorthe Dahl-Jensen

Corresponding author e-mail: dorthe.dahl-jensen@umanitoba.ca

The Arctic is warming and as a consequence sea ice and glaciers are losing mass. There is an urgent need to develop knowledge, tools and models that will improve our understanding of how freshwater fluxes (solid and liquid phase) from glaciers, ice caps and the Greenland Ice Sheet are delivered to the adjacent marine system and what impacts this fresh water has on physical, biological and geochemical processes in the marine system. The Canada Excellence Research Chair will focus on studies of fresh water discharge in the Baffin Bay, with investigations of both Canadian and Greenlandic glaciers exporting fresh water to the Bay. Process studies will also include in situ studies of the northeast Greenland Ice Stream (NEGIS), Petermann Fjord, Upernavik ice streams and the Mueller ice cap. An Inuit-led community-based monitoring (CBM) program is also being developed through a unique partnership with the Inuit Circumpolar Council (ICC), focused on the Pikialasorsuaq, also called the North Water (NOW) polynya, area of northern Baffin Bay. The vision is to improve understanding of the freshwater discharge, understand the key processes affecting overall glacial fluxes of fresh water and the impact of the freshwater discharge on the climate in the past, present and future.


In situ autonomous measurements of light transmission through Arctic first-year ice

Victoria Hill, Bonnie Light, Micheal Steele

Corresponding author: Victoria Hill

Corresponding author e-mail: vhill@odu.edu

The WARM project uses autonomous ice-tethered buoys to collect data on seasonal light penetration through the ice pack. A 20–50 m string supports sensors both within and below the ice for the collection of hourly estimates of vertically resolved downwelling irradiance (412, 443, 555 nm and PAR), temperature, chlorophyll a (chl-a) backscatter, and dissolved organic material (DOM) fluorescence. Since 2014, eight buoys have been deployed in seasonal ice on the Beaufort shelf and Canada basin. The buoys provide data from later winter (March) through the ice-melt period and into the summer open-water period. Buoys have observed repeated under-ice algal blooms on the shallow Chukchi shelf. Using the observations as input to a light-limited primary production proves that there is sufficient light under the ice in May and June to stimulate and sustain under-ice algal blooms. In 2018, sub-meter spatial scale satellite imagery (WorldView2) was available throughout the lifetime of the buoys and was used to classify the areal coverage of snow, melt ponds and open water. Maps of light transmission generated from this classification will be used to quantify the integrated light exposure of the water column under moving ice cover.


The role of sea ice in determining air–sea CO2 fluxes in the Canadian Arctic Archipelago

Mohamed Ahmed, Brent Else, Tonya Burgers, Tim Papakyriakou

Corresponding author: Mohamed Ahmed

Corresponding author e-mail: mohamed.ahmed3@ucalgary.ca

Shelf seas cover more than 50% of the Arctic marine system, and they are known to exhibit significant spatiotemporal variability in air–sea CO2 exchange. To properly estimate the Arctic marine carbon sink, CO2 flux estimates for these shelves must account for seasonal and spatial variations. Using 6 years of high-resolution underway measurements in the Canadian Arctic Archipelago (CAA), we develop empirical relationships that connect dissolved CO2 (pCO2), sea-surface temperature, and salinity to the timing of ice break-up date. Since ice break-up date is easily obtained from ice charts, we exploit this relationship to estimate a contemporary open-water CO2 flux budget for the CAA. Our current estimates show the CAA acting as a net oceanic sink, with an average sea–air CO2 flux of –2.3 Tg&thinC a–1. This sink is significantly smaller than previous estimates for the CAA, emphasizing the importance of properly accounting for spatiotemporal variability in Arctic seas. We also apply our empirical relationships to historical ice charts, allowing us to estimate how this CO2 sink has changed since 1980.


Greenland is melting: how much of this fresh water ends up at the surface?

Fiamma Straneo, Isabela Le Bras, James Holte, Nicholas Beaird, Donald Slater

Corresponding author: Fiamma Straneo

Corresponding author e-mail: fstraneo@ucsd.edu

Mass loss from the Greenland Ice Sheet has quadrupled over the last few decades, discharging additional fresh water, in the form of iceberg and glacier melt, surface runoff and subsurface discharge, into the Arctic and North Atlantic Oceans. Future projections are for continued and enhanced mass loss, implying an even larger freshwater discharge into the ocean. Given the potential impact of this additional fresh water on sea-ice formation, the regional circulation, dense-water formation and the marine ecosystem, it is important to understand its fate and its pathways, and to include it in past model reconstructions and future model projections. Here we use observations collected both in the glacial fjords and around the coastal margins of Greenland to quantify and describe the timing and spatial distribution for the spreading of the meltwater discharged from the ice sheet. Next, building on these observations, we propose a framework to account for the different meltwater components and illustrate how it can be applied to different systems around Greenland.


A novel application of eddy covariance to study seasonal CO2 exchange processes in the Arctic marine environment

Brian Butterworth, Brent Else

Corresponding author: Brent Else

Corresponding author e-mail: belse@ucalgary.ca

The Arctic marine environment plays an important role in the global carbon cycle. However, there remain large uncertainties as to how sea ice affects air–sea fluxes of carbon dioxide (CO2), partially due to disagreement between the two main methods (enclosure and eddy covariance) for measuring CO2 flux (FCO2). The enclosure method has appeared to produce more credible FCO2 than eddy covariance (EC), but is not suited for collecting long-term, ecosystem-scale flux datasets in such remote regions. Here we describe the design and performance of an EC system to measure FCO2 in landfast sea-ice regions that addresses past problems. The system is installed on a small rock outcrop in Dease Strait near Cambridge Bay, Nunavut, in the Canadian Arctic Archipelago. The system incorporates recent developments in the field of air–sea gas exchange by measuring atmospheric CO2 using a closed-path infrared gas analyzer (IRGA) with a dried sample airstream, thus avoiding the known water-vapor issues associated with using open-path IRGAs in low-flux environments. A description of the methods and the results from 2 years of flux measurements are presented, highlighting seasonal variability related to sea-ice formation, melt, and open water. We show that the dried, closed-path EC system greatly reduces the magnitude of measured FCO2 compared to simultaneous open-path EC measurements, and for the first time reconciles EC and enclosure flux measurements over sea ice. We also show that the full CO2 flux budget for an ice-affected region cannot be determined without accounting for the role of sea ice during the melt and freeze-up seasons.


Spatial variability of the properties of suspended sediment in Hudson Bay

Masoud Goharrokhi, David Lobb, Phil Owens

Corresponding author: Masoud Goharrokhi

Corresponding author e-mail: umgoharr@myumanitoba.ca

Documenting suspended-sediment characteristics provides a wealth of information on the nature, source and magnitude of particulate substances in surface-water systems. In this study, two high-flow-rate suspended-sediment samplers were used to collect a significant quantity of such materials along approximately 2000 km of Hudson Bay from 26 September (Churchill) to 10 October (Salluit) as a part of the 2016 BaySys field program. In total, more than 20 sets of large-quantity (i.e. >10 g) samples of suspended sediment were collected. Organic matter, spectral-reflectance (colour) coefficients, particle-size distribution and fallout-radionuclide concentrations (caesium-137 (137Cs), radium-226 (226Ra), and unsupported lead-210 (210Pb)) for all samples were determined to gain insight into the spatial variability of such properties in different parts of the Bay. The ultimate goal for this study is to investigate the utility of the sediment-source-fingerprinting approach to discriminate terrestrial versus marine sources of sediment in the Hudson Bay system.


Optical delineation and assessment of the Nelson River plume, Hudson Bay

Atreya Basu, Greg McCullough, Anirban Mukhopadhyay, Simon Bélanger, Jens Ehn, David G. Barber

Corresponding author: Atreya Basu

Corresponding author e-mail: basua@myumanitoba.ca

Having the largest drainage area and discharge volume, the Nelson River (NR) runoff significantly contributes to the physical, biological and biogeochemical processes of southern Hudson Bay. This study analyzes the NR-plume dynamics in response to different tidal conditions, using ocean color data. An empirical band ratio of Rrs 667 ⁄ (Rrs 488) was applied on NASA-MODIS level 2 product for the months August–September 2006, to estimate absorbance of color dissolved organic matter (CDOM) at 412 nm (aCDOM 412). River waters are the carriers of terrestrial CDOM, so (aCDOM 412) values in the coastal waters proved to be essential in studying the plume-dispersion pattern. A correlation coefficient of 0.7 was calculated for validation measurements using the developed band ratio. Combined use of the optical data of CDOM absorbance to total absorption ratio (aCDOM/aTotal 412) and suspended sediment (SS) led to the delineation of the NR-plume. It’s flow direction is towards the east, where the main plume hugs the eastern shoreline. Plume spreading is most prominent in ebb tide/low tide conditions, when the northeastern extent of the dispersed plume is a linear distance of 250 km from the river mouth. The coastal anticyclonic circulation is responsible for freshwater dispersal in southern Hudson Bay, during flood/high-tide conditions, carrying fresh water from the north. Both in-situ and satellite-retrieved CDOM values were found to be in the range of 1–5 absorption units with a linear-negative dilution curve along the increasing salinity gradient. The suspended sediment mixing curve indicated a zone of resuspension along the deep channel of the Nelson estuary.


Community sea-ice records show changes in shore-fast-ice formation

Alice Bradley, Emily Sun

Corresponding author: Alice Bradley

Corresponding author e-mail: acb8@williams.edu

Shore-fast sea-ice growth and melt seems to be changing from year to year across the Arctic, but limited remote-sensing options make it a difficult environment to monitor. Local records from several coastal Alaskan communities collated through ELOKA and the SIZOnet projects provide an opportunity to investigate the variability of the shore-fast-ice season since 2007. This information is recorded by local residents who work and travel in the ice environment, and tens of thousands of entries have been recorded from 12 communities. Community records are sporadic, with entries more likely around unusual or extreme events. This means that the community records can be used to investigate changes in event frequency over time. We use this data to investigate the changes in number and timing of shore-ice breakout events during the freeze-up season before shore-fast ice stabilizes for the winter season, as well as the timing of ice break-up in the spring.


Investigating the annual cycle and decadal variability of landfast sea ice in the Canadian sub-Arctic: a Hudson-Bay-wide study

Kaushik Gupta, Anirban Mukhopadhyay, Jens Ehn

Corresponding author: Kaushik Gupta

Corresponding author e-mail: guptak1@myumanitoba.ca

Hudson Bay and James Bay combined form a major section of the large and relatively shallow Canadian inland seas. This sub-Arctic basin experiences a seasonal event of landfast-sea-ice formation and melt. Landfast sea ice is freezing of seawater with a continuous extent from the shore, extending towards the offshore region. Formation of this feature is dependent on the regional temperature and oceanographic regimes, with a close relationship also to shore-zone geomorphology. Arctic and sub-Arctic coastlines are more dynamic than their counterparts in the tropics and sub-tropics; one of the primary contributors to this attribute is the event of ‘fastend ice’ formation along the coasts. Variations in the presence of ice coverage and break-up influences the coastal geomorphology. In this study we attempt to investigate the annual cycle of landfast-sea-ice formation and melt in the Hudson Bay and James Bay region by estimating the period of coverage, stages of development and offshore extent of this ice type. Through this study we also investigate the variation in the time of landfast-ice break-up and extent of the seaward landfast-ice extent  (SLIE) from a period between 2000 and 2018. The results of this study were achieved through the use of over 2000 ice charts produced by the Canadian Ice Service (CIS) and satellite coverage observations from NASA Worldview. The Canadian Ice Service publishes ice concentration and stage-of-development charts of Hudson and James Bays on a monthly, weekly and daily scale. We provide the variation in landfast-ice dynamics by digitally extracting information from the daily and weekly ice charts produced by the CIS and satellite observation coupled with mean surface temperature throughout the period of study. We conclude the study with a description of the multiyear variability of landfast sea ice under a changing temperature regime over the Canadian sub-Arctic.


Modeling and prediction of Arctic sea ice and its sensitivity to varying rheology in the Regional Arctic System Model

Wiesław Masłowski, Anthony Craig, Younjoo Lee, Robert Osinski, Andrew Roberts, Mark Seefeldt, John Cassano, Bart Nijssen

Corresponding author: Wiesław Masłowski

Corresponding author e-mail: maslowsk@nps.edu

Results from the Regional Arctic System Model (RASM) are presented to investigate the sensitivity of Arctic sea ice to varying rheology. RASM is a fully coupled limited-domain ice–ocean–atmosphere–land hydrology model. Its domain is pan-Arctic, with the ocean and sea-ice components configured on rotated sphere mesh at 1/12° (~9.3 km) in the horizontal space and with 45 or 60 ocean vertical layers. The sea-ice model uses the latest version 6 of the CICE Consortium model, which includes two rheology options: (i) elastic–visco–plastic (EVP) or (ii) elastic–anisotropic–plastic (EAP). The atmosphere and land-hydrology components are configured on a 50 km grid. As a regional climate model, RASM requires boundary conditions along its lateral boundaries and in the upper atmosphere, which are derived from the National Centers for Environmental Prediction (NCEP) global Climate Forecast System Reanalysis (CFSR) for simulations of the past to present. This allows comparison of RASM results with observations in place and time to diagnose potential improvements of the model physics and to reduce model biases, which is a unique capability not available in global Earth system models. In addition, RASM has been also evaluated for potential gains from dynamical downscaling of NCEP CFS version 2 forecasts. Sea-ice results from near-identical RASM simulations, with the rheology being the only difference, are compared against observations for 1979–2018. Intra-annual ensemble predictions, out to 6 months in the future, are also evaluated to demonstrate the RASM predictive capability beyond seasonal as well as gains of dynamical downscaling of global forecasts.


Arctic sea-ice response to early-20th-century warming

M. Kathleen Brennan, Gregory Hakim

Corresponding author: Gregory Hakim

Corresponding author e-mail: ghakim@uw.edu

Arctic sea ice (ASI) concentrations have undergone rapid declines in recent decades. Many factors have been shown to contribute to this decline, but much of it has been attributed to greenhouse gas forcing and natural variability. In order to understand the relative roles of these factors on ASI decline, a longer record of spatially complete data is needed. Currently, few such records exist prior to the satellite era, so model simulations, proxy records and sparse historical observations have been used to study ASI on longer timescales. This project employs data assimilation to combine these different types of data to reconstruct past climate fields using the Last Millennium Reanalysis (LMR) framework, resulting in spatially complete gridded fields with annual resolution over the last two millennia. Within this longer record, this study will focus in particular on how ASI responded to early-20th-century warming. Temperature observations indicate that there was a period of anomalously warm conditions throughout the northern hemisphere in the early 20th century. The magnitude of this warming was similar to that observed in the late 20th century; however, existing pan-Arctic sea-ice records indicate that the total sea-ice extent did not decline as dramatically during this period as it did later in the century. This discrepancy implies a non-linear relationship between ASI and temperature despite recent studies that indicate a linear relationship in satellite observations and models on longer timescales. The LMR reconstructions of surface air temperature compare well with observations during this period, but the reconstructed total ASI extent shows changes that were larger and longer lasting than in previous records. These reconstructions show a period of rapid total ASI extent decline between 1920 and 1940, followed by an increase between 1940 and 1970. The low period, between 1930 and 1955, in the LMR reconstructions show total ASI extent values similar to those observed in the 1990s. The discrepancy between these and other records is investigated as well as the spatial distribution of the declines.


Untangling the near-field thermal and far-field dynamic components of sea-ice mechanical behavior

Cathleen Geiger, Bjorn Erlingsson, Jennifer Lukovich

Corresponding author: Cathleen Geiger

Corresponding author e-mail: cgeiger@udel.edu

Since the 1990s, telemetry buoys configured with autonomous stress gauges embedded into sea ice have provided point measurement, in-situ, time-series records of relative changes in principal stresses. These time-series records provide a unique insight into the evolution (history) of sea-ice mechanical behavior over time scales of hours up to seasonal and interannual variability. A seminal finding from these records is a strong correlation between the minor stress component (principal shear) and changes in local temperature (near-field thermal). These repeatable findings continue to advance the hypothesis that thermal-induced mechanical stresses in sea ice are isotropic. Subsequently, the non-isotropic dynamic component can be isolated from a two-component plane-stress signal through a simple point-by-point subtraction of minor principal (shear) stress magnitude from the major principal stress (pressure). Unfortunately there is a catch. There is a third component involved. This third component is the isotropic far-field dynamic component, which is entwined with the near-field thermal-induced stress in the isotropic signal as captured by the principal shear. Here, we explore a solution to untangle the two isotropic components of near-field thermal-induced and far-field dynamic-induced isotropic mechanical behavior. Our approach involves filtering techniques and spectral analysis with local temperature records serving as the critical third signal to separate three distinct components. New innovative visualizations – including animations – are used to demonstrate and explain these mechanical components so as to promote cross-disciplinary understanding and interpretation.


Metagenome-assembled genomes of Bacteria, Archaea, viruses and Picoeukaryotes from Arctic sea ice

R. Eric Collins

Corresponding author: R. Eric Collins

Corresponding author e-mail: recollins@alaska.edu

We conducted shotgun metagenomic sequencing on 45 samples of Arctic sea ice. Here we will present a comparison of metagenome-assembled genomes from those samples to genomes of traditionally cultured sea-ice microbes and metagenome-assembled genomes from seawater.


Development of a low-cost Arduino-based iceberg and ice-island drift-tracking beacon

Adam Garbo, Derek Mueller

Corresponding author: Adam Garbo

Corresponding author e-mail: adam.garbo@carleton.ca

Recent atmospheric and oceanic warming has resulted in dramatic increases in the rates of iceberg calving at high latitudes. These ice hazards pose threats to marine vessels and infrastructure at a time when demand for access to Arctic and Antarctic waters is also increasing. There is a growing demand for in-situ iceberg-tracking data to a) monitor their drift trajectory, b) improve drift models used by researchers and operational stakeholders and c) improve iceberg-detection algorithms in satellite imagery. However, the high cost of commercially available tracking devices often prevents monitoring at optimal spatial and temporal resolutions. Here, we describe the design of the Cryologger, a low-cost, robust and user-friendly iceberg and ice-island drift-tracking beacon based on the open-source electronic Arduino platform. Constructed using off-the-shelf components, the Cryologger does not require specialized tools or expertise to build and can be easily modified to meet individual application requirements. The design is planned for extended deployments of 1 year or more and is capable of providing long-term measurements of multiple parameters, including GPS position, temperature, pressure, pitch, roll, tilt-compensated heading and battery voltage. Data are transmitted over the Iridium satellite network at specified intervals, which can be remotely updated based on the desired sampling frequency. In August 2018, six Cryologgers were successfully deployed from the CCGS Amundsen on icebergs and ice islands along the coasts of Ellesmere and Baffin islands. Since deployment, all six beacons have achieved over 225 days of continuous operation, transmitting a total of 13 500 GPS positions and travelling a combined distance of 2200 km. With a total cost of $650 in materials each, initial results have demonstrated that inexpensive, open-source hardware and software can provide a reliable and cost-effective alternative to proprietary commercial equipment for use in monitoring the drift patterns of icebergs and ice islands.


One year of daily high-resolution sea-ice kinematics from Sentinel–1A/B

Tom Armitage, Ron Kwok, Shirley Pang

Corresponding author: Tom Armitage

Corresponding author e-mail: tom.w.armitage@jpl.nasa.gov

Sea-ice kinematics and deformation play an important role in determining the sea-ice-thickness distribution through the opening of leads and growth of new ice and by the formation of thick ice in pressure ridges. Leads generated by ice divergence or shear are responsible for a significant ocean–atmosphere heat flux, and brine rejection into the ocean mixed layer. Hence, observation of sea-ice motion at time and spatial scales suitable for studying deformation is important to understand the evolution of the sea-ice pack. SAR imagery collected by the Sentinel–1A and -1B satellites are processed using the Radarsat Geophysical Processor System (RGPS) to produce daily maps of sea-ice motion (displacement) in the central Arctic Ocean. The data are posted on a 5 km stereographic grid, and in the area surveyed 50% of the region has data coverage 75% of the time and 30% of the region has data coverage 90% of the time. We evaluate the satellite-derived ice-drift vectors against high-resolution GPS buoy trajectories. We compute daily fields of ice deformation (divergence, shear, vorticity), and examine its seasonal cycle and variability over the course of a full annual growth/melt cycle. The quick repeat time of the Sentinel–1A/B imagery is ideal for studies of Arctic sea-ice motion and deformation, and we anticipate it will play an important role in the remote-sensing component of the upcoming MOSAiC cruise, due to embark during freeze-up in 2019.


Measurements of floe size from ICESat-2 laser altimetry

Tom Armitage, Ron Kwok, Glenn Cunningham

Corresponding author: Tom Armitage

Corresponding author e-mail: tom.w.armitage@jpl.nasa.gov

ICESat-2 was launched in September 2018 and provides height measurements in three pairs of laser beams, with each pair of beams separated by 3 km across-track and each pair separated by 90 m across-track. The ATL07 sea-ice height product provides the along-track height at variable along-track resolution; for high-signal photon return rates an along-track resolution of ~22 m is achieved, degrading to up to ~67 m for low-signal photon return rates (for the strong beams). We utilize the spacing between openings in the sea-ice pack to obtain a one-dimensional measurement of the floe size along ICESat-2 height transects, with greater spatial coverage at a higher resolution than has previously been possible. We examine the spatial and temporal evolution of this quantity over the season, and identify limitations to its physical interpretation. The areal distribution of sea-ice floes and openings between them is an important quantity for studying atmosphere–ocean heat transfer. Further, sea-ice models typically use statistical models to simulate the distribution floe sizesand ICESat-2 data could be a valuable new tool for evaluating floe-size distribution parameterizations.


Cryo-reactions and sea-ice biogeochemistry

Feiyue Wang

Corresponding author: Feiyue Wang

Corresponding author e-mail: feiyue.wang@umanitoba.ca

Most chemical reactions in aqueous solution slow down as temperature decreases. In frozen media, however, certain reactions are known to proceed very differently from their aqueous counterparts, some being accelerated in rate while others yield unusual products. Although the mechanisms of most of these ‘cryo-reactions’ remain poorly understood, their importance in many stratospheric and tropospheric processes has been long recognized. Evidence is mounting that the sea-ice environment is also a much more chemically active environment than previously thought. This presentation will provide a synopsis of the current understanding of cryo-reactions and their role in sea-ice biogeochemical processes. Three types of cryo-reaction will be discussed: type I reactions, which can be fully accounted for by the freeze-concentration effect; type II reactions, which are accelerated compared to those in non-frozen solutions and cannot be explained by the freeze-concentration effect; and type III reactions, which yield different products from those in non-frozen solutions and cannot be explained by the freeze-concentration effect. In the sea-ice environment, many of these cryo-reactions are photolytically initiated or microbially mediated. Emphasis will be placed on the critical knowledge gaps and how they affect our current understanding of the cycling of greenhouse gases and chemical contaminants in polar regions in a changing climate.


From the Sea-ice Environmental Research Facility to the Churchill Marine Observatory: controlled mesocosm-scale sea-ice research at the University of Manitoba

Feiyue Wang, David Barber, Søren Rysgaard, Tim Papakyriakou, C.J. Mundy, Gary Stern, David Binne

Corresponding author: Feiyue Wang

Corresponding author e-mail: feiyue.wang@umanitoba.ca

Controlled, mesocosm-scale sea-ice research at the University of Manitoba started in 2012 with the opening of the Sea-ice Environmental Research Facility (SERF). The main features of SERF include an outdoor sea-ice pool with a movable roof, two large sea-ice tubs, numerous in situ sensors and instruments, and an on-site trailer laboratory. Sea ice can be created from formulated seawater under various controlled conditions (e.g. seawater chemistry, snow cover, heating) with the additions of chemical, isotopic and/or microbiological tracers. Over the past 7 years, SERF has provided an unprecedented platform where process-oriented sea-ice studies are carried out under controlled or semi-controlled conditions. Building upon the success of SERF, a new major research facility, the Churchill Marine Observatory (CMO), is currently under construction at the port of Churchill, Manitoba, adjacent to North America’s only Arctic deep-water port. The core CMO infrastructure includes 1) an outdoor Oil-in-Sea-Ice-Mesocosm with two pools, which is designed to simultaneously accommodate mesocosm-scale contaminated and control experiments on various scenarios of oil and related contaminants in ice-covered waters; 2) the Environmental Observatory system, which is a network of state-of-the-art sensors and equipment located in the Churchill estuary and along the main shipping corridor across Hudson Bay to Baffin Bay; and 3) a coastal research vessel William Kennedy. CMO will be fully operational by early 2020. In this presentation, we review the motivation and research capacity of SERF and CMO, highlight major discoveries from geophysical, chemical and geochemical experiments so far, and provide a critical analysis of the pros and cons of mesocosm-scale studies of sea ice and their relevance to field and modeling studies.


The role of sea ice as an ocean fertilizer

Delphine Lannuzel

Corresponding author: Delphine Lannuzel

Corresponding author e-mail: delphine.lannuzel@utas.edu.au

The availability of iron has been shown to limit biological production in large areas of the Southern Ocean, and changes in the supply of iron to the surface ocean are evoked in explaining glacial–interglacial changes in atmospheric CO2 concentrations. Artificial iron fertilization experiments conducted in the Southern Ocean confirmed that iron addition stimulates phytoplankton growth. The focus eventually moved away from the large-scale artificial experiments conducted in the 1990s to now studying naturally fertilized areas, with an emphasis on continental shelf sediments, atmospheric dust and hydrothermal vents. Particularly high rates of productivity have been reported in the Antarctic marginal ice zone, suggesting that sea ice could be an overlooked source of iron. This was later confirmed by several studies reporting considerable enrichment of iron in sea ice. Given the vast area that sea ice covers annually, iron-rich meltwaters constitute the dominant and most widespread source of iron in polar waters during spring. Fast ice has a particularly high iron content, because of its proximity to eroded continental rocks, sediments and dust. Pack ice, on the other hand, relies on biological activity as its iron source. The processes driving these high concentrations are not well understood, but the co-occurrence of high concentrations of Fe and organic matter in the ice suggests a coupling leading to their enrichment. This process likely involves the complexation of iron with exopolymeric substances, which can act as binding agents. While most studies have focused on the origin and magnitude of the iron sources, a growing body of work is now dedicated to quantifying how accessible iron is for marine microorganisms, the so-called bio-availability. Iron bio-availability can be estimated by looking at the size (dissolved versus particulate), redox state, and complexation to organic ligands. Other metals, such as cobalt and manganese, may also co-limit phytoplankton growth and should become a growing focus of future work. This presentation will revisit research on trace metals in sea-ice, with a particular emphasis on iron. As the scientific community recognizes that iron-rich and climate-sensitive environments such as sea ice will be subject to rapid changes in the near future, there is an urgent need to predict how changes in the sea-icescape will affect the capacity of the Southern Ocean to absorb CO2.


The ESA CryoVEx airborne sea-ice freeboard campaigns – with early results from 2019 CryoSat and IceSat-2 underflights

Henriette Skourup, Alessandro diBella, Rene Forsberg, Sine Hvidegard, Tania Casal

Corresponding author: Rene Forsberg

Corresponding author e-mail: rf@space.dtu.dk

DTU Space has carried out a large number of satellite-validation airborne campaigns over sea ice since 1993, both in the Arctic and lately in Antarctica. The ESA CryoVEx campaigns (along with associated EU and national campaigns) have given a unique dataset for insight in sea-ice dynamics, free-board, ridging and related processes, as well as data for validating geoid and ocean topography models. The primary instruments flown include scanning lidar, various radars (Ku- and Ka-band), as well as GNSS and IMU equipment; in connection with other campaigns gravity, magnetic and radiometer sensors have also been flown. The main objectives of the campaigns have been to understand retrieval of CryoSat-2 sea-ice-thickness measurements, and also CryoSat-2 inter-calibration with other satellite missions such as AltiKa and in 2019 IceSat-2. The campaigns have frequently been coordinated with AWI EM-bird ice-thickness measurements, as well as NASA IceBridge and ground team activities. In the presentation we will show examples of campaign results, including snow-depth estimation from lidar/radar combinations, and validation of Arctic Ocean CryoSat thickness results (and associated mean sea level/geoid surfaces), as well as showing first results of 2019 CryoSat/IceSat-2 underflights.


New high-resolution sea-ice floe-size distribution results from US Geological Survey LIDP imagery

Byongjun Hwang, Takenobu Toyota

Corresponding author: Byongjun Hwang

Corresponding author e-mail: B.Hwang@hud.ac.uk

Declassified images from US intelligence satellites (National Technical Means) have been made available to sea-ice researchers by the group called MEDEA, and then released to the public by the US Geological Survey (USGS) as LIDPs (Literal Image Derived Products) at 1 m image resolution. In the Arctic, six fixed sites known as ‘fiducial sites’ were selected to provide baseline data to understand recent and future changes in the Arctic. In this research, we derived and analysed the floe-size distributions from the LIDP images acquired at three fiducial sites: the Chukchi Sea, East Siberian Sea and Fram Strait. The data span over about 10 years between 1999 and 2014, which provides a unique insight into temporal and spatial variability of high-resolution floe-size distributions. From over 200 images available in total, we have selected about 45 characterized as pre-ponding (no evident melt ponds visible in the images). The preliminary results show quite significant interannual variation in floe-size distribution at the same site. We are currently analysing more data to examine whether any long-term changes in floe-size distribution have actually occurred, and how the temporal evolution of floe-size distribution differs between the selected sites.


Diversity and distribution of microbial eukaryotes in the Churchill–Nelson River Basin

Loïc Jacquemot, Carlee Morency, Marianne Potvin, Connie Lovejoy

Corresponding author: Connie Lovejoy

Corresponding author e-mail: Connie.Lovejoy@bio.ulaval.ca

Freshwater within the Greater Hudson Bay marine region originates from rivers with large drainage basins and is modified by sea-ice formation and melting. Within the BaySys project, which aims to provide a scientific basis to separate climate-change and water-regulation impacts, our subproject aims to provide baseline data on microbial biodiversity and associated drivers. Being at the interface of freshwater and marine systems, estuaries are of great interest to understand the influence of fresh water on microbial coastal communities. Within estuaries, gradients in light, nutrients, heat and salinity through freshwater and seawater mixing, are predicted to influence microbial community structure. We sampled microbial eukaryote assemblages using rRNA 18S marker genes along the salinity gradient of two major inflows into the Hudson Bay: the Nelson and Churchill rivers. These rivers are both highly stratified but exhibit strong hydrodynamic differences under the contrasting influence of river damming and estuarine geomorphology. Data on the distribution and composition of microbial communities was compared with environmental parameters along the salinity gradient to identify spatial patterns and factors potentially controlling the distribution of the microbes. Samples were also used to identify a coastal core community and combined with earlier western Hudson Bay results. We then screened for potential unique and invasive species in all samples. These results will provide a baseline to compare western and eastern Hudson Bay coastal communities and will be of use to managers by suggesting whether new invading organisms will be more likely to replace current species.


Examining the potential impact of increased vessel traffic noise on marine mammals in the proposed Tallurutiup Imanga (Lancaster Sound) National Marine Conservation Area

Zuzanna Kochanowicz, Jackie Dawson, William Halliday

Corresponding author: Zuzanna Kochanowicz

Corresponding author e-mail: zkoch086@uottawa.ca

Recent reductions in sea-ice extent in the Canadian Arctic, as a result of global climate change, has led to an increase in maritime navigability and it is expected that there will be an increase in shipping activities related to tourism, fisheries and trade in the future. There is already increased interest in the commercial viability of the Northwest Passage, as evidenced by the recent sailings of cargo vessels such as the Nordic Orion (2011) and the Nunavik (2014) and the unprecedented number of yachts, cruise ships and new research vessels such as the Chinese research icebreaker Snow Dragon (2017) and cruise ship Crystal Serenity (2016, 2017). The navigation season in the Canadian Arctic is also observed to be increasing in length with the season starting earlier for some vessel types and ending later for others. Tallurutiup Imanga (Lancaster Sound) a marine area that will soon become a National Marine Conservation Area (NMCA) spanning approximately 110 000 km2, is located at the eastern entrance of the Northwest Passage. The area is rich in ecological and cultural significance and the protection of Tallurutiup Imanga has been in progress for decades. With the official boundaries set in August of 2017, the future protection and management of the new NMCA will be crucial to protecting its integrity, especially as commercial shipping is expected to increase in precisely this area. Our ongoing research project aims to: 1) utilize an existing ship-track database developed using Canadian Coast Guard (NORDREG) data, to evaluate shipping trends in Tallurutiup Imanga from 1990 to 2016; 2) project future shipping traffic in the region based on best available data; 3) model the current and potential future impact of vessel noise on relevant marine mammal species in the NMCA using Automatic Identification System (AIS) data; and 4) propose spatial management options for reducing noise impacts from increased marine traffic.


Sea-ice-lead detection using Sentinel SAR data in the Arctic Ocean

Jullian Williams, Hongjie Xie, Alberto Mestas-Nuñez, Stephen Ackley

Corresponding author: Jullian Williams

Corresponding author e-mail: jullian.williams@utsa.edu

The atmosphere, ocean and coastal boundaries exert forces onto sea-ice surfaces which cause the formation of linear kinematic features (leads) where open water is exposed. Leads are the principal locations for vertical air–sea fluxes. The leads are a regulating factor of the net turbulent heat flux in the Arctic and are almost two orders of magnitude larger than those over the pack ice. Leads are responsible for approximately half the surface sensible-heat flux, and virtually all the surface moisture flux that the atmosphere over the Arctic Ocean receives during winter. Still, open water gives the largest heat flux while thin ice provides the second largest heat flux over the Arctic region. As a result, thick multiyear ice is no longer dominant but rather a thin, first-year ice, which can quickly form in the cold open water areas. Though the thickness of this thin ice can increase rapidly, ice drift and wind stresses have also gradually accelerated in the Arctic in the last five decades. This suggests a marked increase in the resultant sea-ice mean strain rate; a factor that is at the forefront of lead production. In tandem with this, there is widening of the marginal ice zone of approximately 39%. This suggests further that an increasing surface area of total ice in the Arctic is broken ice. The study of ice leads is dependent on this factor since leads play an essential role in regulating temperature in the Arctic. The objective of this research is to use Sentinel synthetic-aperture radar (SAR) data to successfully identify and quantify lead features in sea-ice cover in the Arctic Ocean. Satellite data used for lead detection in previous studies, such as those derived from AMSR-E and MODIS, were restricted in their lead detection capabilities due to their coarse resolution. However, SAR data derived from Sentinel are now employed in high-resolution lead detection to identify leads of geometries narrower than 3 km. The correctness of the method employed is essential since the lead parameters, such as lead width and orientation, govern the ability of the heat-conductive plumes to penetrate higher into the atmosphere. The results derived from this new algorithm will be used to inform energy-balance and climate-research arenas in the cryosphere.


Distribution and impacts of microplastic incorporation within sea ice

Nicolas-Xavier Geilfus, Kathleen Munson, João Sousa, Yaroslav Germanov, Saamia Bhugallo, David Babb, Feiyue Wang

Corresponding author: Nicolas-Xavier Geilfus

Corresponding author e-mail: Nicolas-Xavier.Geilfus@umanitoba.ca

Since their mass production in the 1950s, plastics have become the most common and persistent marine pollutants. Of emerging concern are microplastics, which are particles <5 mm in size. While recent indications suggested that the Arctic Ocean could be sensitive to microplastic accumulation, microplastic concentrations in the Arctic Ocean are expected to increase rapidly due to increases in freshwater input and shipping and resource-development activities. However, processes and rates of microplastic incorporation within sea ice, as well as their impacts on sea-ice properties, are largely unknown. We designed a microcosm experiment at the Sea-ice Environmental Research Facility, University of Manitoba, Winnipeg, to determine the distribution of microplastics in sea ice and their effects on the growth, melting rates, and albedo of the ice. Microplastic in the ice was analysed by a modified Nile red dye method. Our results suggested that sea ice concentrated particles within its structure and, once the particles are incorporated into the ice, they seem to be stationary in the ice matrix. Although the presence of microplastic particles seems not to affect sea-ice growth, high salinity was associated with a large concentration of particles, which could affect physical properties of sea ice such as brine volume content and permeability. Our results also reveal distinct changes in sea-ice albedo in response to additions of light-absorbing microplastic impurities, a process that could impact the absorption of incident solar radiation and thus sea-ice melt. In addition, microplastics could affect the light penetration depth, altering photochemical and photo-biological processes that occur in sea ice.


From oil-in-sea-ice research to spill response and damage assessment support

Marc Oggier, Hajo Eicken, Christian Petrich, Scott Pegau

Corresponding author: Hajo Eicken

Corresponding author e-mail: heicken@alaska.edu

Petroleum development continues to account for a significant part of human activities in the maritime Arctic, both directly through exploration and production and indirectly through support activities in coastal and offshore waters. Alaska’s Beaufort Sea has seen an increase in exploration activities in recent years, including planned offshore gravel island construction and projected expansion of subsea pipeline installations to bring oil onshore. At the same time, shipping has increased in Arctic regions as well. In the event of a spill, the presence of coastal landfast and drifting sea ice is a major factor in spill response and assessment of damages to ecosystems and habitats in the region. Drawing on a combination of ice-tank experiments, community-based monitoring, and sea-ice field and remote sensing data collected in northern Alaska, we discuss the relevance of recent research findings in the context of spill response decision support and damage assessments. Full-scale ice tank experiments at the Cold Regions Research and Engineering Laboratory (CRREL), Hamburgische Schiffsbau-Versuchsanstalt (HSVA) and University of Alaska Fairbanks (UAF) helped us develop a conceptual model for sequential infiltration and upward mobility of oil in sea ice. Acoustic tracking of the water–oil–ice interface served to quantify oil fluxes into the sea ice over the course of a simulated seasonal cycle. Stratigraphic analysis of oil distribution revealed the importance of sea-ice textural units, in particular granular ice, in impeding the surfacing of oil, relative to surface pooling in columnar ice. This work also revealed the importance of oil-layer thickness and ice-permeability evolution in determining the fraction of pore space occupied by oil and its pervasion through the ice matrix. A simple oil-distribution model drawing on this data can provide guidance on the timing and manner of oil surfacing, as well as the relative proportion of surfaced oil vs oil retained within the sea-ice pore space over the course of the season. The latter information is also critical in assessing damages and impacts to sea-ice microbial communities and habitat. Remote-sensing data on the seasonal distribution of different ice types (in particular landfast-ice extent) and observed trends in ice-season duration help inform present-day and future oil-spill scenarios and response options in coastal environments.


Antarctic landfast sea-ice distribution and variability, and its influence on coastal polynyas

Alexander Fraser, Robert Massom, Kay Ohshima, Sascha Willmes, Peter Kappes, Jessica Cartwright, Simon Wotherspoon, Kazuya Kusahara, Takeshi Tamura

Corresponding author: Alexander Fraser

Corresponding author e-mail: adfraser@utas.edu.au

Antarctic landfast ice (fast ice) forms in embayments, around and upstream of grounded icebergs/coastal promontories, and between grounded icebergs and the coast. It is a key element of the coastal cryosphere, is closely related to coastal polynya location, has important ecological and biogeochemical roles, and can either facilitate or impede logistical operations. Accurate and high-resolution characterization of the circumpolar distribution and variability of fast-ice coverage and seasonality (and their drivers) is required for more complete understanding of the Antarctic coastal environment and its response to climate change/variability. Until now, this level of characterization has been confined to East Antarctica, and spans only 2000–08. Here, we introduce a new, high-resolution (1 km; 15 day) time series of complete circum-Antarctic fast-ice extent based on cloud-free Moderate Resolution Imaging Spectroradiometer (MODIS) imagery from 2000–14. We present fast-ice climatologies and extent trends for each of the five Southern Ocean sectors, but also focus on specific regions to highlight the ‘black swan’/stochastic response of fast-ice extent to large iceberg-grounding events. Using this new baseline dataset, we also highlight the roles fast ice plays in determining coastal polynya size and modulating the total amount and location of sea-ice production on the continental shelf. We conclude with a discussion of Antarctic fast-ice representation in next-generation high-resolution models capable of simulating Antarctic fast ice explicitly.


Dynamics of elemental mercury across the seawater–ice–atmosphere interface in a controlled mesocosm environment

Zhiyuan Gao, Kathleen Munson, Feiyue Wang

Corresponding author: Feiyue Wang

Corresponding author e-mail: feiyue.wang@umanitoba.ca

Mercury is a contaminant of global concern. Albeit far away from its anthropogenic sources, mercury can be found in elevated concentrations in the Arctic marine cryosphere and in tissues of Arctic marine mammals. Mercury evasion, usually in the form of elemental mercury, is a major sink for cryospheric and oceanic mercury. This is due to the high mobility of elemental mercury, which is readily redistributed between the atmosphere and seawater. However, it is influenced by the presence of a sea-ice cover, which is hypothesized to act as a ‘porous’ physical barrier. Previous studies focused on total mercury and methylmercury distributions in sea ice, while the distribution of elemental mercury was much less studied. When ice starts to grow, dissolved gaseous mercury (DGM) may accumulate in intermediate water underneath the sea-ice cover or be trapped within the ice layer; in either case, air–sea gas exchange will be limited. As sea ice grows thicker and decreases its permeability, DGM concentration in the sea-ice cover will likely decrease due to the weakening of air–sea gas exchange and potential cryospheric reactions. Meanwhile, DGM in intermediate water and gaseous elemental mercury (GEM) in the atmosphere above the ice can also change during the growth of sea ice. To study elemental mercury dynamics under natural cryospheric conditions, here we report a controlled mesocosm study carried out over a sea-ice grow–melt cycle at the Sea-ice Environmental Research Facility (SERF), University of Manitoba, Winnipeg, Canada. A system (12 ft (3.7 m) in length; 5 ft (1.5 m) in width) consisting of 12 mesocosms was planted in an outdoor seawater pool and froze into ice during sea-ice growth. Individual mesocosm was covered with clear/dark poly sheeting to create a constrained headspace of at least 25 L above the ice surface for atmospheric GEM sampling. DGM in ice-intermediate seawater and ice-core samples were measured discretely over the sea-ice grow–melt cycle. Preliminary results show that DGM in intermediate water decreases gradually when the sea-ice cover is present; within the ice core, DGM is found to be enriched in granular sea ice compared to columnar sea ice. Our results indicate that sea-ice cover acts as a physical barrier to mercury flux between the seawater–ice–atmosphere interface and the redistribution of elemental mercury over a sea-ice grow–melt cycle is controlled by both physical and chemical processes in the simulated cryospheric environment.


Will sea-ice leads in the Beaufort Sea increase as ice becomes thinner?

Xi Zhao, Meng Qu, Xiaoping Pang

Corresponding author: Xi Zhao

Corresponding author e-mail: xi.zhao@whu.edu.cn

The spatial pattern of sea-ice leads reflects the process of ice dynamics in the Arctic. Open-water leads or refrozen leads with a thickness of less than half a meter allow a large conductive heat flux, which affects local cloud cover and precipitation, even the opening date of sea ice for the coming spring. The average sea-ice thickness in the Arctic Ocean has been reducing considerably in the past 30 years. The composition of the sea ice in the Arctic is expected to change continuously in the future. The mature and stable multiyear ice will probably be replaced by thinner and more mobile young and first-year ice, which may therefore produce more ice leads in some regions like the Beaufort Sea. This hypothesis needs to be tested by collecting observations over a long period. Since in-situ observations on thin ice are difficult, further development of remote-sensing methods is required. This study is to monitor the spatial-temporal pattern of sea-ice leads in Beaufort Sea by utilizing multisensory remote-sensing data over the past decades. Optical reflectance and thermal brightness temperature images from MODIS, Landsat 8 and Sentinel-2A, and dual-polarization backscatter images from Radarsat-2, were used in the detection, classification and validation. In addition, spatial statistics were applied on derived lead parameters, including surface temperature, width, area, orientation and frequency. Their seasonal and annual change is also discussed.


Manitpiatuk hiku or manniqtuk hiku? Reducing on-ice trafficability uncertainty using satellite remote-sensing-based roughness maps

Rebecca Segal, Randall Scharien, Joel Heath

Corresponding author: Rebecca Segal

Corresponding author e-mail: rasegal@uvic.ca

Residents in Cambridge Bay and Kugluktuk, in the western Canadian Arctic, are interested in accessing remotely sensed image data and enhanced image products to help plan travel and activities on sea ice. Interviews helped determine what types of information residents are interested in obtaining, in relation to experienced sea-ice features and conditions. An emergent theme is the need for sea-ice surface-roughness information, due to its impact on landfast-sea-ice trafficability and safety. Consequently, we are investigating the use of two satellite remote-sensing datasets with the goal of optimizing the identification of three roughness classes: smooth ice (manniqtuk hiku); moderately rough ice (manitutun hiku); and rough ice (manitpiatuk hiku; dialect is Inuinnaqtun). The first, Sentinel–1, is a synthetic-aperture radar (SAR) that provides high-resolution (metre-scale) backscatter independently of sunlight and cloud cover. SAR backscatter is related to surface- and volume-scattering mechanisms, with the former linked closely to ice roughness and topography. The second, the multi-angle imaging spectroradiometer (MISR), is an optical sensor with multi-angle cameras. MISR is used to map sea-ice topography through the creation of a normalized differential angular index that takes advantage of the difference in surface reflectance between forward- and backward-facing cameras. Airborne lidar-derived roughness data collected in the Canadian Arctic Archipelago in April 2017 is used as validation. Roughness patterns observed by Sentinel–1 and MISR are similar for first-year ice. MISR provides roughness information in sea-ice areas influenced by freshening and enhanced volume scattering (multiyear ice and river-outflow areas). Despite the influence of volume-scattering from multiyear ice, Sentinel–1 exhibits a stronger overall relationship with roughness than MISR (R2 = 0.61 vs 0.26 respectively). Community experts describe SAR-based maps as highly accurate, with benefits including improved safety, travel time, route planning and fuel efficiency. We are working to make validated roughness products available to sea-ice users in electronic format via SIKU.org, the Inuit Knowledge Social Media platform. Through SIKU.org, sea-ice data is integrated into a suite of tools that enable users to make informed decisions related to safe navigation and ice-based activities. The SIKU.org platform will help scale the availability of this information in the future.


Using barium concentration to assess freshwater sources in southeast Hudson Bay

Celine Gueguen, Joshua Creppin, Zou Zou Kuzyk

Corresponding author: Celine Gueguen

Corresponding author e-mail: celine.gueguen@usherbrooke.ca

The need for reliable tracers for freshwater circulation is particularly important in Arctic seas, where rapid changes are occurring. This study assesses the potential to use barium concentrations to track the sources and distribution of freshwater in Hudson Bay. Surface-water samples were collected as part of a 3-year collaborative project with the Inuit and Cree communities in southeast Hudson Bay and northeast James Bay. The Ba survey was complemented with salinity δ18O and CDOM measurements. We used a three-endmember mass balance approach, employing salinity and barium concentration as tracers. A difference of <1% was found in the fraction of meteoric water determined using either Ba or δ18O as tracer, suggesting that Ba can used as a reliable tracer in southeast Hudson Bay. A significant difference in Ba concentrations between the Great Whale and La Grande rivers, the main river inputs in the study area, was used to distinguish their contribution to the surface freshwater composition. Preliminary results suggest that the freshwater signature on a regional scale in southeast Hudson Bay, including around the Belchers, was dominated by the La Grande River whereas the influence of the Great Whale River was only found near the river mouth. Together, these results suggest that barium concentration could complement already- established tracers in tracing fresh water in Hudson Bay.


The impact of the 2017 reversal of the Beaufort Gyre on the regional ice cover

David Babb, Jack Landy, Ryan Galley, Christian Haas, Jennifer Lukovich, David Barber

Corresponding author: David Babb

Corresponding author e-mail: David.Babb@umanitoba.ca

During the winter of 2017 the semi-permanent Beaufort High broke down and caused a reversal of the anticyclonic Beaufort Gyre. While the cause of this event has been traced to negative sea-ice anomalies in the Barents Sea, its impact on the ice cover of the western Arctic has yet to be discerned. Within this research we utilize a variety of remote-sensing and in situ observations to analyze how the anomalous transport of sea ice during winter 2017 impacted the regional ice cover. Specifically we focus on the impact of this event on sea-ice thickness and multiyear-ice presence in the Beaufort Sea. Typically, the anticyclonic Beaufort Gyre transports thicker, multiyear sea ice into the Beaufort Sea during winter and exports a mix of multiyear and seasonal sea ice westward into the Chukchi sea. However, the reversal of 2017 changed the patterns of ice transport and precluded the replenishment of multiyear sea ice into the Beaufort Sea. This led to the lowest multiyear-ice area in the Beaufort Sea in the observational record and fostered a more seasonal ice cover. Theoretically the seasonal ice cover should be thinner; however, convergence during the reversal deformed the seasonal ice cover and limited the negative ice-thickness anomalies, which in turn influenced the subsequent melt season. Looking to the future we discuss how increasingly frequent reversals of the Beaufort Gyre, coupled with an increasingly mobile, thinner ice pack, may impact future conditions in the region.


Iceberg production from the Prince of Wales Icefield and drift patterns through Canadian waters

Abigail Dalton, Luke Copland, Wesley Van Wychen

Corresponding author: Abigail Dalton

Corresponding author e-mail: adalt043@uottawa.ca

Tidewater glaciers drain a significant proportion of the Greenland Ice Sheet, and the ice caps of the Canadian Arctic Archipelago (CAA), and are the primary source of icebergs and ice islands (large tabular icebergs) in Canadian waters. The Canadian Ice Service (Environment and Climate Change Canada) produces daily charts which identify the presence of icebergs, but currently has little knowledge about the sources and sinks of icebergs in Canadian waters. Recent studies have shown that, as of 2016, Trinity and Wykeham glaciers on the Prince of Wales Icefield (southeast Ellesmere Island) are responsible for ~62% of all iceberg production from the CAA, compared to 22% in 2000. This study uses synthetic aperture radar (SAR) data and optical imagery to identify the most active iceberg-producing glaciers from the Prince of Wales Icefield over the last ~20 years. Results show a clear relationship between the presence of sea ice and the production of icebergs, with ~49% of total iceberg-plume events occurring during the 3–4-month-long summer open-water season and ~51% of events when sea ice was present the remaining 8–9 months of the year. Using the CCGS Amundsen from 2016–18, a total of 39 iceberg-tracking beacons have been deployed to monitor the movement of icebergs from northern Baffin Bay through Canadian waters. Helicopter-deployed satellite tracking beacons have provided near-real-time (hourly) movement of these icebergs, and reveal that the icebergs can be highly mobile, with a maximum drift of >5600 km over a span of ~1 year between Jones Sound and Hudson Strait. These results illustrate patterns of iceberg production and movement, and the interactions between iceberg-drift patterns and primary shipping routes along the east coast of Canada.


Pikialsorsuaq: the confluence of science and community

Chris Debicki

Corresponding author: Chris Debicki

Corresponding author e-mail: cdebicki@oceansnorth.ca

Shared by Canada and Greenland, the Pikialasorsuaq (the Western Greenlandic name for the North Water) is the world’s largest Arctic polynya. This vast sea-ice opening is the most biologically productive region north of the Arctic Circle, and fosters primary production and critical habitat for migratory species that local Inuit depend on (e.g. seabirds, narwhal, Arctic cod and seals). Oceans North is a Canadian not-for-profit organization with a history of working with Inuit and northern communities. Through partnerships, we strive to advance policy solutions to many of the pressing challenges facing Canada’s Arctic. Oceans North has collaborated with the Inuit Circumpolar Council to establish the Pikialaorsuaq Commission, which was mandated to consult with coastal communities in Greenland and Nunavut that are connected to Pikaialasorsuaq. In its 2017 report, The People of the Ice Bridge: The future of the Pikialasorsuaq, the Commission makes three recommendations: 1. Establish a management regime led by Inuit representatives from communities in the Pikialasorsuaq region. 2. In consultation with communities adjacent to the Pikialasorsuaq, identify a protected area comprising the polynya itself and a larger management zone that reflects the connection between communities, their natural resources and the polynya. These areas would be monitored and managed by Inuit in agreement with all parties and formally recognized by governments. 3. Establish a free travel zone for Inuit across the Pikialasorsuaq region. While no structures presently exist for international co-management of this region, formal intergovernmental dialogue has been initiated to this end. Just as important, a collaborative approach between the university research community, non-governmental organizations and communities is taking root. The Commission recommendations and subsequent cooperation between ICC, the University of Manitoba and Oceans North are helping to foment a paradigm shift in terms of how high-Arctic ocean research will unfold in ensuing years. Critical to the success of these new research partnerships is a research approach that is adaptable to community input and removes barriers to local and indigenous participation.


Development of a reflectance probe to measure sea-ice inherent optical properties

Christophe Perron, Simon Lambert Girard, Christian Katlein, Edouard Leymarie, Pierre Marquet, Marcel Babin

Corresponding author: Christophe Perron

Corresponding author e-mail: christophe.perron.1@ulaval.ca

More detailed characterization of the spatially and temporally varying inherent optical properties (IOPs) of sea ice is necessary to better predict sea-ice energy and mass balance and under-ice primary production. Here we present the development of an active optical in situ probe for measuring IOPs of a small volume of ice (mm3 to cm3) non-destructively and within a short time. The precision, efficiency and ruggedness of the concept allows obtaining sea-ice IOPs values within a 2 in (5 cm) hole directly in the field within minutes. It provides high-resolution vertical profiles of sea-ice IOPs and allows the investigation of IOPs’ relationship to other physical sea-ice properties. The probe is based on the diffuse-reflectance technique used to measure IOPs of human tissues. Conceptually, the instrument emits light into the ice via an optical fiber. Backscattered light is measured at multiple distances from the in-situ source using other fibers. Measured reflectance vs distance curves are compared to values derived from Monte Carlo simulations of radiative transfer. A pre-computed look-up table and an inverse algorithm allows inference of the absorption coefficient, the reduced-scattering coefficient and the asymmetry parameter of the scanned sea ice. Here we present the design of the instrument, as well as the first results from field tests made in the Canadian Arctic in 2018 and 2019. This includes the first vertically resolved in-situ measurements of sea-ice IOPs.


Implications of the wave climate of the Southern Hemisphere on polar ship design

Sally Garrett

Corresponding author: Sally Garrett

Corresponding author e-mail: s.garrett@dta.mil.nz

The waves of the Southern Ocean and Ross Sea are largely unstudied. The New Zealand Defence Force (NZDF) routinely operates in these areas and is currently engaged in a shipbuilding program that requires detailed understanding of the wave climate for ship design. Unlike other areas, the Southern Ocean and Ross Sea has limited ship traffic and therefore limited wave observations from volunteer observing ships. Moreover, due to the difficult conditions and remote location, limited scientific measurements of waves have been completed. In 2017, the NZDF deployed a moored wave buoy in open ocean 11 nautical miles south of Campbell Island. In addition, 23 free-floating buoys were also deployed between 42° S and 67° S. This array has provided an understanding of wave characteristics across the Southern Ocean and Ross Sea. The initial findings of the wave climate measured by the array will be compared to similar Northern Hemisphere locations used to develop shipbuilding guidelines. A summary of the implication of the differences for future shipbuilding will be presented.


Impedance at the sea-ice/atmosphere interface: implications for surface temperature

Chaincy Kuo, Daniel Feldman, Daniel Martin

Corresponding author: Chaincy Kuo

Corresponding author e-mail: CKuo@lbl.gov

The IPCC Fifth Assessment Report presented a CMIP5 model mean wintertime surface temperature bias of approximately –4°C in the Arctic Sea region as compared to ERA-Interim reanalysis, over the period 1986–2005 (Figure 9.40 of IPCC AR5). On the other hand, summertime surface-temperature bias over the same region is reported to be approximately –0.7°C. Considering that the conducted ocean heat flux at the sea-ice/atmosphere interface makes a larger contribution to surface energy balance in the winter than in the summer, the climate-model implementation of the surface energy-balance equation for the Arctic sea-ice by Maykut and Untersteinter (1971) (MU71 hereafter) is examined here from a process-level theoretical perspective. An analytic, theoretical expression for sea-ice surface temperature at the sea-ice/atmosphere interface will be presented based on the MU71 surface energy-balance expressions, but where sea-ice temperature in the solid material is governed by the heat diffusion equation. In this theoretical framework, the conductive heat-flux expressions at the sea-ice/atmosphere interface have terms that include impedance effects. A preliminary 1-D mathematical model shows that the impedance effect can have considerable implications for a predicted surface temperature in winter with warmer temperatures by a few degrees Celsius, but have a negligible impact in summer. To quantify the significance of the impedance effect, we compare the theoretical derivations of surface temperature to numerical implementations of the surface energy balance in the sea-ice thermodynamics model of CICE. Sea-ice heat-conductive flux calculated from in-situ temperature measurements in the Surface Heat Budget of the Arctic Ocean Experiment (SHEBA) campaign and prior Arctic expeditions have not considered this additional term of which the magnitude is on the order of their predicted ocean heat flux (up to 5 W m–2). One development of the theoretical model is an improved understanding and accounting of energy fluxes at the sea-ice/atmosphere interface derived from observations, processes and trends. This could provide potential to improve temporal predictive fidelity for surface melt and freeze. With this perspective, we develop predictions for the relationships between the radiative fluxes and the sea-ice-temperature profiles that will be observed as part of the MOSAiC campaign in 2019/20.


Connections between floating-glacier-tongue and sea-ice losses on northern Ellesmere Island, Canada

Luke Copland, Adrienne White

Corresponding author: Luke Copland

Corresponding author e-mail: luke.copland@uottawa.ca

There have been rapid recent decreases in glacier and ice-shelf extent across northern Ellesmere Island, with eight floating glacier tongues shrinking in area by >85% in the Yelverton Bay region since 1959. To understand the causes of these losses this study undertakes the first examination of the factors driving ice-tongue changes in this region, including an assessment of changes in surrounding sea ice, sikussak and melange, and atmospheric and oceanographic forcings. From 1959–2017 the total ice-tongue area decreased by 49.07 sq km, with >70% of losses occurring between 2005 and 2009. The loss of ice tongues since 2005 has occurred when open water replaced multiyear landfast sea ice and first-year sea ice in the regions adjacent to the glacier tongues. These changes were accompanied by an increase in 100 m and 200 m depth ocean temperatures from –0.29°C from 1999–2005 to 0.67°C from 2006–2012. Despite the recent return of ocean temperatures to below pre-2006 levels, atmospheric summer temperatures have continued to rise (+0.15°C per decade between 1948 and 2016), with open water continuing to occur. Without the sustained presence of multiyear landfast sea ice in this region the ice tongues are unable to stabilize, making it unlikely that they will re-form in the current climate.


Direct measurements of the radiance distribution beneath Arctic landfast sea ice during the spring transition

Simon Lambert Girard, Edouard Leymarie, Sabine Marty, Lisa Matthes, David Antoine, Jens Ehn, Marcel Babin

Corresponding author: Simon Lambert Girard

Corresponding author e-mail: Simon.Lambert-Girard@takuvik.ulaval.ca

Improved modelling of the interactions between sun photons (radiative transfer) and snow, sea ice and the upper water column is required to better understand energy deposition and primary production in ice-covered oceans. The basic requirement for this – and the focus of many former studies – is measurement of spectral surface albedo and transmittance for the various ice types and surface conditions present. In addition to irradiance information, measurements of the radiance angular distribution (the light field) are needed to facilitate the inversion of inherent optical properties and sea-ice structural properties. This presentation details the deployment of a radiance camera beneath first-year landfast sea ice near Broughton Island, southern Baffin Bay, during the Green Edge ice camp in June/July 2016. The design of the experiment allowed data to be gathered along four dimensions to capture the variability due to the variegated nature of sea ice: space (40–150 m horizontal transmittance transects and 25 m vertical profiles under the sea ice), time (from dry snow to melt pond and bloom seasons), spectral (6 bands across the visible region) and angular (downwelling radiance distribution). Relationships between these dimensions and, for example, ice/water inherent optical properties, ice/water chlorophyll-a concentrations, sea-ice surface conditions, sea-ice thickness and salinity will be presented.


On the design of an optical sensor to measure nitrate in sea ice

Yasmine Alikacem, Simon Lambert Girard, Denis Boudreau, Simon Thibault, Marcel Babin

Corresponding author: Marcel Babin

Corresponding author e-mail: marcel.babin@takuvik.ulaval.ca

The intricate balance between nutrients, specifically nitrate, in the Arctic Ocean and light availability has been shown to control the dynamics of primary production. The amount of available nutrients for ice algae depends on the flux nutrients from the upper ocean to sea ice, and transport within sea ice. In a context where the Arctic icescape is profoundly changing together with sea-ice biology, this research project aims at developing an optical sensor for measuring in situ sea-ice nitrate concentration at small space and time scales, to better understand how the flux of nutrients from the upper ocean to sea ice and their transport within sea ice are controlled by the physical properties of the two media. Current in situ oceanographic nitrate measurements are performed with ultraviolet absorption spectroscopy. Unfortunately, conventional instruments are bulky, and require tricky signal deconvolution and correction. Two new technological approaches are being assessed for measuring nitrate concentration in sea ice while perturbing the environment as little as possible: transmission dip fiber probe and surface-enhanced Raman spectroscopy. A transmission fiber dip probe consists of an excitation and a collection fiber which transmit light through an open cavity to interact with the medium in order to measure the absorption spectrum. Surface-enhanced Raman spectroscopy is a promising new approach to amplify weak Raman signals by the use of nanoparticles assembled together to form a plasmonic surface, which generates a highly localized and intense electric field when excited by a laser. The selected new nitrate sensor is intended to be integrated onto a sea ice endoscopic (SIE) platform, a non-destructive multimodal probe which will be used to characterize sea-ice optics, physics, biology and biogeochemistry.


Impacts of climate change on choke points for shipping in the Canadian Arctic

Alison Cook, Luke Copland, Jackie Dawson

Corresponding author: Luke Copland

Corresponding author e-mail: luke.copland@uottawa.ca

In the Canadian Arctic the recent reduction in sea-ice extent and thickness, along with an increase in mobility of hazardous sea ice, creates navigational challenges for Arctic ship operations. The operational risks vary depending on the polar class (i.e. level of ice strengthening) of the vessel and on the extent to which dynamic and mobile sea ice is prevalent and changing in the regions where shipping operates. Locations where sea ice is frequently present throughout a shipping season, impeding travel along routes that are otherwise largely ice-free, have been called ‘choke points’. In this study we use a comprehensive temporal and spatial inventory of historic shipping traffic by polar class (1990–2018) to examine the location of choke points in the Canadian Arctic, how they’re changing over time, and how the strength of ships navigating through these locations is evolving. Results reveal that since 1990 there has been a marked reduction in the voyages of highly strengthened Polar Class 3 ships, but large increases in the number of voyages of ships with medium ice strengthening (Polar Class 7) and little or no ice strengthening (Polar Class 1B). In addition, there are many more voyages by non-ice-strengthened ship types throughout the Northwest Passages in the 2010s than in the 1990s. This has occurred during a period when an analysis of historical sea-ice charts indicates that ice navigability for vessels with medium or little ice strengthening has greatly eased.


A multimodal endoscopic approach for characterizing sea-ice optics, physics, biology and biogeochemistry at small scale

Marcel Babin, Simon Lambert Girard, Christian Katlein, Yasmine Alikacem, Raphaël Larouche, Christophe Perron, Jean-Marie Trudeau, Éric Bharucha, Guislain Bécu

Corresponding author: Marcel Babin

Corresponding author e-mail: marcel.babin@takuvik.ulaval.ca

Sea ice is a complex and heterogeneous medium that hosts a rich community of microbial organisms and small invertebrates. This ecosystem is shaped by a network of inhabitable spaces where the upward and downward fluxes of solutes and light support primary production, and ultimately the whole sea-ice trophic network. Describing the optical, physical, biological and biogeochemical processes that drive the functioning of the sea-ice ecosystem at the appropriate, i.e. small scale (micro- to centimeter), is very challenging. This medium is solid, fragile and highly heterogeneous. Traditional sea-ice sampling methods based on coring are most often coarse and destructive. Not only do they not allow the small scale to be explored, they generally alter the material to be analyzed. Here, we present a new approach for measuring relevant variables of the sea-ice ecosystem at small scale and, as much as possible, non-destructively. Inspired by medical endoscopes, the custom-built platform is intended to carry various types of miniaturized optical sensors for radiometry, chemistry and high-resolution imaging of the sea-ice interior. In this presentation, we will describe the concept and present the progress made to date.


Sea-ice-algal biomass response to snow metamorphism during the melt season

Gauthier Verin, Philippe Massicotte, Marti Gali Tapias, Ghislain Picard, Laurent Arnaud, Marcel Babin

Corresponding author: Gauthier Verin

Corresponding author e-mail: Gauthier.Verin@takuvik.ulaval.ca

Primary production in the Arctic Ocean is strongly affected by the limitation of light due to the presence of a reflective sea-ice cover. The aim of this study is to investigate potential relationships between light transmitted through snow-covered sea-ice and bottom-ice algal biomass. Intensive field sampling was conducted on typical landfast sea ice near Qikiqtarjuaq, Baffin Island, over two melting periods in 2015 and 2016 as a part of the GreeenEdge project. In this study we report temporal changes in snow properties, including snow-specific surface area (SSA) and density and in photosynthetically active radiation (PAR) transmitted through the ice that were derived from C-OPS irradiance profiles performed in the water column. Our results show an important inter-annual variability in both the observed environmental conditions and the amount of biomass measured within bottom sea ice. Daily mean snowpack depths, SSA and densities were reconstructed using the snow radiative-transfer model TARTES, extended to the ice layer, combined with temporal changes in C-OPS-erived transmittances. This reconstruction makes it possible to evaluate the absolute irradiance at any depth of the sea ice, representative of the average spatial conditions, and finally to investigate the impact of the evolution of snow-properties on ice-algal biomass. Time evolution of irradiance within the ice was represented by contour lines of equal light intensity called isolums. In both 2015 and 2016, changes in ice-algal biomass and irradiance through sea ice followed similar trends, suggesting a strong control of light changes on sea-ice biomass. The effect of light changes on sea-ice algae was particularly important in 2015, when one week of heavy snowfall resulted in a four-fold decrease in light availability under sea ice (<0.4 μEm–2 s–1) after 9 May. Following this episode, snow metamorphism and surface melting resulted in a rapid increase of irradiance through sea ice that was closely matched with peaking algal biomass.


Accuracy assessment of sea-ice concentration derived from newly released FY3D-MWRI

Xi Zhao, Ying Chen, Xiaoping Pang

Corresponding author: Xi Zhao

Corresponding author e-mail: xi.zhao@whu.edu.cn

The Microwave Radiation Imager (MWRI) aboard on the Chinese FengYun-3D satellite (FY3D) was launched in November 2017. The MWRI receives radiation from land, atmosphere and ocean, including 10 channels in five different frequencies from 10–89 GHz at both horizontal and vertical polarization. Spatial resolution of the individual measurements varies from 10 km at 89 GHz to 73 km at 10.65 GHz. The MWRI Level 1 data contains 15 scientific datasets such as latitude, longitude, EARTH_OBSERVE_BT_10_to_89GHz, DEM in HDF5 format. The aim of this study is to produce a daily sea-ice-concentration product using FY3D-MWRI and to assess its accuracy in the Arctic. Based on MWRI Level 1 brightness temperature data, daily sea-ice concentration (SIC) at 12.5 km resolution was derived by the Arctic Radiation and Turbulence Interaction Study (ARTIST) sea-ice (ASI) algorithm using both constant tie points and dynamic tie points. The correlations between MWRI brightness temperature and Advanced Microwave Scanning Radiometer 2 (AMSR2) brightness temperature are highest in the sea-ice region (0.952 at 89H GHz and 0.985 at 89V GHz), and are lowest at the ice edge (0.814 at 89H GHz and 0.919 at 89V GHz). The dynamic tie points of sea ice have substantially low values in summer, whereas those of open water have a fluctuating seasonal pattern. Yearly-averaged dynamic tie points were estimated to be 48.8 K and 8.6 K, whereas the constant tie points of open water and sea ice were set to be 47 K and 11.7 K respectively. Since the FY3D-MWRI dataset is newly released, we test the accuracy of the daily SIC for the year 2018. Intercomparisons were made among MWRI-ASI, AMSR2-ASI, AMSR2-NT2 and AMSR2-Bootstrap using the results from both dynamic tie points and constant tie points. The dynamic-tie-points-based SIC gave the largest difference 15% to the three comparator products, while the constant-tie-points-based SIC has 10% maximum difference. The correlations between MWRI-ASI and the comparator products are 0.90/0.99, 0.82/0.92 and 0.87/0.94 for the SIC using dynamic tie points/the SIC using constant tie points, respectively. Obviously the constant-tie-points-based SIC shows a closer result to the comparator products. Finally, ten scenes of MODIS data were used to make detailed validations in core ice area and marginal ice zones in cloud-free conditions. We conclude that FY3D-MWRI provides comparable brightness temperature datasets and MWRI-ASI can serve as a good data source for SIC monitoring.


Satellite-derived first-year landfast-ice albedo evolution during the Green Edge project

Julien Laliberte, Eric Rhem, Marcel Babin

Corresponding author: Marcel Babin

Corresponding author e-mail: marcel.babin@takuvik.ulaval.ca

Arctic marine ecosystems are fueled by the production of algal biomass. While the growth of phytoplankton (single-celled algae suspended in seawater) was believed to be largely limited to the period when Arctic Ocean seasonal ice cover was decreasing (Jul–Oct), massive blooms of phytoplankton occurring under sea ice in the spring were recently documented. It is currently impossible to determine the extent of this phenomenon and its contribution, perhaps major, to annual marine primary production, as the mechanisms controlling the dynamics of phytoplankton blooms under sea ice are poorly understood. Recent observations to understand this phenomenon suggest that phytoplankton growth under sea ice is largely conditioned by access to underwater light, which is mainly controlled by the amount of light reflected by snow, bare ice, leads and melt ponds. Many environmental components were measured at a coastal Baffin Bay location during the Green Edge 2015 and 2016 field campaign. Using in situ and satellite observations, we evaluate how the snow, bare ice and melt ponds reflect visible light at a very local scale for the spring and summer seasons.


Hatching time of Boreogadus saida and under-ice river plumes; further testing of the freshwater winter refuge hypothesis

Sarah Schembri, Inge Deschepper, Caroline Bouchard, Paul G. Myers, Louis Fortier

Corresponding author: Sarah Schembri

Corresponding author e-mail: sarah.schembri.1@ulaval.ca

The eggs of Arctic cod (Boreogadus saida) start hatching from January to early July depending on the Arctic or sub-Arctic region. Individuals that hatch early at low temperatures under the ice face poor feeding and survival conditions, but benefit from a longer growing season than late hatchers. These early hatchers reach a larger size in the fall than late hatchers, which presumably provides some survival advantage during the following winter. Winter hatching has been observed in seas strongly influenced by fresh water such as the Laptev Sea, the Beaufort Sea and Hudson Bay. While winter temperatures in saline waters falls below freezing, the temperature in under-ice river plumes stays around 0°C, thus potentially providing a thermal refuge for egg survival, fast development and eventually feeding success. The notion that Arctic cod spawn in brackish water to enhance larval survival is called the freshwater winter refuge hypothesis. The young fish used in previous studies to test the hypothesis were collected in mid to late summer and therefore there is no direct evidence that B. saida larvae hatch in and occupy under-ice brackish water in winter. In this study we look at the hatching time of B. saida larvae captured in early spring on the 2018 CCGS Amundsen BaySys mission, some of which were found among sediment-laden freshwater ice. We compare the hatching time of larvae in the context of their location in or around estuaries in the Hudson Bay system. In collaboration with BaySys Team 6 we also model the backward planktonic movement of the B. saida larvae to estimate their hatching location and whether this could potentially occur in under-ice river plumes.


Rheological parameter estimation in sea-ice models through variational data assimilation

Gleb Panteleev, Max Yaremchuk, Jacob Stroh, Oceana Francis, Richard Allard

Corresponding author: Gleb Panteleev

Corresponding author e-mail: gleb.panteleev@nrlssc.navy.mil

Sea-ice models that allow for deformation are primarily based on rheological formulations originally developed by Hibler (1979). In both the original visco-plastic (VP) and elastic-VP (EVP) schemes, the internal sea-ice pressure term is modeled as a function of variable sea-ice thickness and concentration with empirical parameters for ice strength and ice–ocean coupling prescribed as constants throughout the domain. This work considers a spatially variable extension of those parameters in the two-dimensional EVP sea-ice formulation of Hunke and Dukowicz (1997); Lemeaux et al., (2015) and one-dimensional VP sea-ice formulation of Konig Beatty and Holland (2010). Feasibility of optimization of the rheological parameters was assessed by applying time-dependent variational data assimilation to the synthetic sea-ice observations. We found that for areas covered by ice (i.e. pack ice zones), the optimization of the variable rheological parameters provides better hindcast/forecast of the sea-ice state than equivalent experiments using spatially homogeneous parameters for partially ice-covered regions where internal ice stresses are negligible. Overall, the conducted observing system simulation experiments suggest that basic rheological parameters (sea-ice strength, ellipse ratio) and parameters responsible for landfast and ‘arching’ sea-ice phenomena can be successfully recovered through the assimilation of sea-ice thickness, concentration and velocity observations. This results in an improved hindcast/forecast of the sea-ice condition.


Measurement of in-ice angular radiance distributions

Raphaël Larouche, Simon Lambert Girard, Christian Katlein, Simon Thibault, Marcel Babin

Corresponding author: Raphaël Larouche

Corresponding author e-mail: raphael.larouche@takuvik.ulaval.ca

To better understand sea-ice thermodynamics and the growth of ice algae, the propagation of solar radiation within the medium needs to be studied in more detail than isusually achieved. Previously, most sea-ice light measurements were made at the upper and lower boundaries in order to derive spectral albedo and transmittance for different ice conditions. These bulk apparent optical properties do not allow inferring vertical partitioning of light at small scale within sea ice. A few studies have tried to measure in-ice irradiance vertical profiles using rather bulky instruments that induced shadow, destroyed the medium and were limited in vertical and angular resolutions. The aim of this project is to design, build, characterize and field test a miniaturized probe for measuring radiance angular distributions within sea ice. This instrument is to be included on a sea-ice endoscopic platform which will gather other probes for documenting sea-ice structural, optical and biogeochemical properties. In addition to this prototype, we investigate the possibility of using commercial miniature 360° cameras as an affordable and easy solution to get radiance angular distributions within sea ice. The goal is to be able to use these measurements to develop strong structural–optical relationships that will help to develop better light-transport models.


Understanding the productivity of the sympagic and pelagic ecosystems of the Hudson Bay complex using biogeochemical modelling

Inge Deschepper, Diane Lavoie, Paul Myers, Tim Papakyriakou, Frédéric Maps

Corresponding author: Inge Deschepper

Corresponding author e-mail: inge.deschepper.1@ulaval.ca

Ice-algal growth is an important source of biological carbon for the pelagic and benthic systems within the Arctic and sub-Arctic regions and its role in the biological carbon pump needs to be re-evaluated in the context of a rapidly changing ice-scape in Arctic seas. The sympagic ecosystem can provide a large draw-down of carbon, but its contribution to carbon burial through the biological carbon pump is still not well known, due in particularlto the Arctic being difficult to observe and sample. The contribution of nutrients from freshwater river runoff to the sea-ice ecosystem and ice-algae growth is a further source of uncertainty, particularly for regions that experience high river-runoff discharge. A better understanding of these complex interacting processes is crucial to improve our predictive capabilities in areas that are undergoing an unprecedented acceleration of changes in sea-ice conditions. The Hudson Bay complex is a sub-Arctic system that receives high levels of freshwater runoff from over 65 rivers draining into the bay and experiences seasonal ice cover. Understanding bay-wide productivity and what impacts the timing, distribution and size of the ice algae and pelagic blooms can give us insight into how productive the system can be. The BioGeoChemical Ice Incorporated Model (BiGCIIM) representing the carbon and nitrogen cycles within the pelagic and sympagic ecosystems is coupled to the Arctic Northern Hemisphere Atlantic (ANHA4) configuration of the ocean-circulation model NEMO v3.6 and the sea-ice model LIM2 to try and understand the spatial and temporal productivity of the Hudson Bay complex.


Modeling virus–host dynamics to assess the potential of a viral shunt within sea ice

Max Showalter, David Talmy, Jody Deming

Corresponding author: Max Showalter

Corresponding author e-mail: gshowalt@uw.edu

In marine environments, nutrient cycling is understood through the framework of the microbial loop, in which bacteria recover dissolved organic carbon (DOC) from the environment and incorporate it into biomass, facilitating upward trophic transfer. Viruses ‘short-circuit’ this loop in a process known as the viral shunt, whereby viruses lyse bacteria, returning DOC to the environment. We have limited understanding of how such microbial/viral DOC cycling may proceed through wintertime sea ice. However, observations suggesting both high concentrations and production rates of viruses within sea-ice brines indicate that a viral shunt may be present and important in this environment. Here, we assess the existence and potential impact of a viral shunt within sea ice using a mathematical model of virus–host population dynamics within sea-ice brines. Data from both field and laboratory isolates were used to achieve most likely parameter distributions for in situ communities, constraining the model to observed dynamics. Deeper understanding of the viral impact on sea-ice communities and DOC cycling within the ice will help further elucidate how these largely heterotrophic communities maintain activity through winter, and further constrain Arctic carbon budget through the season.


Spatial distribution of leads and lead parameters from the high-resolution optical imagery of the Operational IceBridge digital mapping system

Hongjie Xie, Dongqin You, Jinlong Gao, Alberto Mestas, Steve Ackley

Corresponding author: Hongjie Xie

Corresponding author e-mail: Hongjie.Xie@utsa.edu

Leads are critical features in the polar regions as they are involved in the exchange of mass and energy between the ocean and atmosphere, especially during winter. This study extracts the leads and parameters from 12 904 optical images with a high resolution of 0.1 m, acquired on 20 April 2016 by the Operational IceBridge digital mMapping system (DMS). The leads are automatically extracted based on the grey value distribution of each DMS image. Post classification has been processed to ensure the classification quality. Lead statistics including number of leads, dimension of leads, fraction of leads, and dominant orientation of leads are used to describe their spatial distribution and characteristics. Minimum boundary geometry, a specified minimum bounding geometry enclosing each input feature, is employed to specify each lead’s orientation and width/length. According to leads’ different orientations, the main orientation of each image is accumulated in angular bins with an interval of π/6. Finally, the spatial distribution characteristics of leads are obtained in detail with data of such high spatial resolution.


Snow–ice interface detection from high-resolution vertical temperature profiles measured by unmanned ice mass-balance buoys in an Arctic lake

Yubing Cheng, Bin Cheng, Fei Zheng, Timo Vihma, Qinghua Yang, Zeliang Liao

Corresponding author: Bin Cheng

Corresponding author e-mail: Bin.Cheng@fmi.fi

A snow layer on a floating ice floe may result in: 1) a slush layer forming at the snow–ice interface and later refreezing to form snow-ice; 2) the snowmelt water, sleet or even rain refreezing to form superimposed ice at the snow–ice interface before onset of ice melting. Therefore, the snow–ice interface is a dynamic moving boundary between the snow depth and ice thickness, particularly, if a heavy snow load exists on a thin ice floe. Snow depth and ice thickness in remote areas are largely observed by unmanned ice mass-balance (IMB) buoys. The acoustic sounders are used, applying the snow–ice interface as a reference level, to detect the evolution of surface and bottom in order to obtain snow depth and ice thickness. For a moving snow–ice interface, the acoustic-sounder-derived snow depth and ice thickness maybe erroneous. We applied an existing thermistor-string-based IMB (SIMBA: Snow and Ice Mass Balance Array) algorithm to investigate the evolution of the snow–ice interface. The temperature gradients are applied to identify air–snow and ice–water interfaces as given in the algorithm before. A clear differentiation of the second derivative of temperature in snow and ice is used to identify the moving snow–ice interface. Our method was applied to retrieve snow depth and ice thickness from SIMBA measurement in an Arctic lake, Orajärvi, Finland, during the 2017/18 and 2018/19 boreal winters. An upward-moving snow–ice interface was evident during these two seasons. The snow depth and ice thickness calculated from the algorithm were compared with several in situ observations.


The fate and behaviour of aromatic fractions of crude oil in icy conditions: critical review and impact on remediation

Katarzyna Półćwiartek, Monika Pućko, Kathleen Munson, Feiyue Wang

Corresponding author: Feiyue Wang

Corresponding author e-mail: Feiyue.Wang@umanitoba.ca

The Arctic region is home to some of the largest unexplored reserves of crude oil and natural gas. Extended ice-free periods in the Arctic have seen increased interest in regional petroleum exploration and extraction and increased activity both in terms of commercial shipping and tourism. Associated with these industrial activities are risks of oil spills as a result of drilling operation mishaps and shipping accidents, with potentially severe environmental and economic consequences. Focusing on the Canadian Arctic, this paper presents a comprehensive and critical review of current knowledge about the fate and behaviour of crude oil, especially on its volatile aromatic fraction partitioning in multi-phase systems in the presence of sea ice. Particular emphasis is placed on studies of uncontrolled oil spills, such as the Exxon Valdez spill in Prince William Sound in 1989, as these provide invaluable insights into the understanding and improvment of effective oil-spill remediation and prevention techniques and regulations. One major conclusion of this review is that more consideration should be given to the interactions between aromatic crude oil fractions and sea ice, as partitioning of these fractions can differ considerably, depending on particular sea-ice conditions. Furthermore, aromatic hydrocarbon fractions directly impact the fate and behaviour of bulk crude oil due to their physical and chemical properties, such as high volatility, solubility in water and high toxicity. Thus, aromatic crude oil fractions have implications for environmental risk assessment, oil-spill contingency planning, and oil-spill mitigation operations in the Arctic environment.


Mid-latitude and tropical drivers of recent Arctic climate change

Kunhui Ye, Tido Semmler, Thomas Jung

Corresponding author: Thomas Jung

Corresponding author e-mail: Thomas.Jung@awi.de

The impact of processes in mid-latitudes and the tropics on recent Arctic climate is studied using seasonal hindcast experiments with the coupled ECMWF model in which parts of the atmosphere are relaxed towards ERA-Interim data. From these experiments it is argued that the ‘western Arctic’ (‘eastern Arctic’) is more strongly impacted by the tropics (mid-latitudes). Our result suggest that Arctic climate change has a strong remotely driven component, especially during winter. These results are supported by climate-change experiments in which CO2 is enhanced locally (e.g. Arctic vs non-Arctic).


High-resolution sea-ice–ocean modelling with FESOM2

Nikolay Koldunov, Sergey Danilov, Thomas Jung

Corresponding author: Thomas Jung

Corresponding author e-mail: Thomas.Jung@awi.de

The Finite Volume Sea-ice-Ocean Model (FESOM2) is the first mature global sea-ice–ocean model formulated on unstructured meshes. Here, we show results from recent high-resolution simulations with FESOM2 in which leads in the Arctic sea ice emerge. A comparison with satellite data suggests that these features are quite realistic. Furthermore, we present results from a thorough scalability analysis of FESOM2, which shows very promising scalability for the 3-D part of the model, whereas sea ice turns out to be a bottleneck for highly parallel problems. An adaptive version of EVP (mEVP) is proposed that increases the performance of high-resolution simulations significantly.


Antarctic sea-ice decline delayed well into the 21st century in high-resolution climate projections

Thomas Rackow, Sergey Danilov, Thomas Jung

Corresponding author: Thomas Jung

Corresponding author e-mail: Thomas.Jung@awi.de

Despite ongoing global warming and strong Arctic sea-ice decline, the sea-ice extent around the Antarctic continent as estimated from satellite observations is still stable and even slightly increasing in September. In contrast, the average of all current climate models shows a strong negative sea-ice trend and the confidence in projected Antarctic sea-ice changes is therefore low. Here we show, using a locally eddy-resolving global model system integrated towards the end of the 21st century, that equatorward ocean heat transport favours periods of stable September sea ice as observed over the satellite era, and that the sea-ice decline could be further delayed for more than a decade. Mixed-resolution experiments attribute this behavior to the use of the high-resolution ocean that better represents the residual overturning circulation in the Southern Ocean and shows a different response to global warming, independent of the atmospheric resolution above. The low-resolution ocean configurations fully parameterize the effect of eddies and simulate a weaker northward branch of the residual overturning that is less efficient at moderating warming in the Southern Ocean. This entails an erroneous decrease of Antarctic sea-ice extent similar to what current coarser-resolution climate models predict. Our results provide guidance for modeling centres trying to resolve this long-standing issue and they reconfirm that eddy-resolving ocean components are necessary for faithful projections of the Southern Ocean.


Development of an optical probe for continuous porosimetry of snow: a theoretical breakthrough

Félix Lévesque-Desrosiers, Simon Lambert Girard, Quentin Libois, Florent Domine, Simon Thibault

Corresponding author: Félix Lévesque-Desrosiers

Corresponding author e-mail: felix.levesque-desrosiers@takuvik.ulaval.ca

Snow density is currently measured by weighing a known volume. This process requires the presence of a scientist in the field, which is complex and expensive given the Arctic logistics constraints. Then, this property can not be characterized throughout winter and on a large area. This project focuses on developing an optical method to measure snow density without human intervention. Light propagation in the snow, a complex medium composed of air and ice grains, depends on its density and on the size and the shape of the grains that compose it. Here, we demonstrate that snow density can be optically measured using time-resolved radiative transfer. We developed a theoretical framework, supported by numerical simulations, to relate the effective refractive index of snow to its density. We then show that snow density can be optically obtained by measuring the absorption coefficient of snow and its effective refractive index. A method to measure the inherent optical properties that can be used to measure the snow density and specific surface will also be presented. This paves the way for an autonomous, fast and non-destructive method to measure snow physical properties in the Arctic.


Under-ice freshwater and nutrient dynamics in coastal northeast James Bay and southeast Hudson Bay

Alessia Guzzi, Jens Ehn, Michelle Kamula, Chris Peck, Jean-Eric Tremblay, Zou Zou Kuzyk

Corresponding author: Alessia Guzzi

Corresponding author e-mail: guzzia@myumanitoba.ca

In Arctic settings, inorganic nutrient stocks needed for primary production in spring get recharged during the ice-covered winter through microbial recycling and physical processes. Introduction of river water in winter and changes in sea ice may affect these physical processes, with implications for nutrient recharge. However, winter nutrient data from Arctic areas are extremely scarce. Hudson Bay and James Bay collectively make up the largest inland sea in the Arctic. In recent decades, these areas have experienced earlier sea-ice melt and later freeze-up, together with increased winter river inflows due to hydroelectric development. In particular, the La Grande River, in northeast James Bay, peaks in outflow during the winter period, as opposed to during the typical spring peak. In this study, we present new nutrient and freshwater tracer data for water samples collected during both ice-covered and open-water conditions from sites along the northeast coast of James Bay and northward into coastal southeast Hudson Bay. Fieldwork was conducted in partnership with the Arctic Eider Society and through a community-driven research network. Using oxygen isotope ratios (δ18O) in conjunction with salinity, we calculated the contributions of freshwater components (river runoff and sea-ice melt) of samples. Then relationships between these components and inorganic nutrient concentrations and stocks were examined. In winter, the La Grande River plume in northeast James Bay is nitrate-rich relative to the surrounding coastal waters. The marine waters that circulate into northeast James Bay in winter have higher phosphate but lower nitrate than the river-sourced water. In southeast Hudson Bay, where less of a river influence is present in winter, recharge produces higher nutrient concentrations than in northeast James Bay. However, throughout the study area, winter nutrient maxima are much lower than the concentrations observed in Hudson Bay below the winter mixed layer, suggesting generally poor winter nutrient recharge in northeast James Bay waters.


Modelling sea-ice influence on iceberg distribution

Juliana Marson, Paul Myers

Corresponding author: Paul Myers

Corresponding author e-mail: pmyers@ualberta.ca

Greenland icebergs populate Baffin Bay and the east coast of Canada, often reaching the Grand Banks region. Those icebergs are not only a threat to navigation and other offshore activities, but also a source of fresh water to the North Atlantic and of nutrients – such as iron – to primary productivity. Studying iceberg distribution is, nevertheless, a difficult task since observations are scarce. In this context, numerical models become useful tools and their improvement is paramount to obtain a better estimate of how Greenland icebergs are distributed across the subpolar North Atlantic. The Nucleus for European Modelling of the Ocean (NEMO) v3.6 introduced an iceberg module where the user can provide snow accumulation rates which will feed a calving process and subsequent tracking of icebergs throughout the model integration in time. Currently, the iceberg momentum equation in this module represents the sea-ice influence on the iceberg as a drag force. However, observations and other modelling studies have shown that, at high concentration and thickness, the sea-ice pack is able to capture the iceberg, which then moves with the pack. In this study, we compare the distribution of Greenland icebergs in two simulations. One uses only the original sea-ice drag force on the iceberg momentum equation, while the other assigns sea-ice velocity to the icebergs which are in the middle of thick and concentrated sea-ice pack. Preliminary results show that, while the differences in iceberg speed between the two simulations are small in compact sea-ice situations, this small difference is able to redistribute icebergs from the boundary currents (such as the Baffin Island current) to the middle of Baffin Bay.


Modelling the long-term fate and transport pathways of pollutant in the Canadian Arctic

Ran Tao, Paul Myers

Corresponding author: Paul Myers

Corresponding author e-mail: pmyers@ualberta.ca

Sea ice has been decreasing in recent years, especially in summer. From a shipping perspective, this means larger areas of open water in the summer, thinner and less compact ice all year round, and longer operating seasons. The Arctic Ocean has three main transpolar shipping routes, two of which have been successfully traversed by large cargo vessels in the early 2010s. The transpolar shipping routes significantly shorten the distance travelled via the traditional Suez or Panama Canals, and their depth allows more cargo to be carried. Furthermore, the Aarctic Ocean is the reservoir for one of the world’s largest untapped hydrocarbon resources, the development of which has begun in the Beaufort Sea since 2018. Increasing commercial activities promote the possibility of accidental spills of pollutant, especially considering the more mobile sea ice, harsh environment and lack of detailed navigation charts. Compared to the Northern Sea Route along the Russian coast, the Northwest Passage in the Canadian Arctic is unprepared for the increasing commercial activities, especially considering its complex waterway and lack of supportive ports, search-and-rescue vessels and spill recovery schemes. If an accidental spill occurred, the pollutant would be less likely be recovered within the same year, before sea ice makes it inaccessible. Therefore, we were motivated to examine the transport and fate of pollutant if spilt into the Arctic Ocean along the Northwest Passage, in order to provide comprehensive data and scientific guidance for future regulations and development. We use a high-resolution ocean model, the Arctic and Northern Hemisphere Atlantic configuration of NEMO (Nucleus for European Modelling of the Ocean), run at 1/12°, and a Lagrangian particle-tracking tool Ariane to simulate the spread of pollutants with oceanic advection. We will examine the transport of pollutant spilt at various sites along the Northwest Passage every 10 days during the operational season (June–October), and determine its trajectory at the end of the year of the spill as well as the end of the following year, marking the long-term fate of pollutant. From the results, we expect to highlight the importance of oceanic and atmospheric circulation to the spread of pollutant, and determine over which site and when the spill would have the most severe outcome, and over which area the pollutant tends to accumulate.


Exchange through the Gulf of Boothia and Fury and Hecla Strait based upon historical data

Paul Myers, Samantha Roch, Sophie Rohrl

Corresponding author: Paul Myers

Corresponding author e-mail: pmyers@ualberta.ca

This talk involves examining the physical oceanography of the Canadian Arctic Archipelago, with a focus on the region of Fury and Hecla Strait and the Gulf of Boothia. The Canadian Arctic Archipelago is an important route for fresh water from the Arctic Ocean to the North Atlantic Ocean. The Gulf of Boothia and Fury and Hecla Strait connect the central archipelago to the Labrador Sea via Foxe Basin and Hudson Strait. As they are narrow and ice-choked, the total flux of water, as well as the freshwater transport through this region is not well known. Using historical ocean observations stored in the ICES and MEDS databased, as well as those from more recent ArcticNet cruises, we examine the different water masses in the region. Baroclinic geostrophic velocities and thus transports are determined, with the reference velocity taken from a high-resolution numerical model (the Arctic and Northern Hemisphere Atlantic configuration of NEMO). The different historical sections, in combination with the model, are thus used to examine the exchange through this region, as well as the drivers of transport variability.


Genomic signatures of virus-mediated gene transfer relevant for microbial life in subzero, hypersaline brines

Josephine Z. Rapp, Zhiping Zhong, Dean Vik, Matthew Sullivan, Shelly D. Carpenter, Jody W. Deming

Corresponding author: Josephine Z. Rapp

Corresponding author e-mail: jzrapp@uw.edu

In frozen environments, such as sea ice and permafrost, microbes reside within hypersaline liquid inclusions, sometimes at very high densities (up to 108 cells mL–1 brine). Microbial activity in these subzero brines requires adaptations that enable core cellular processes and membrane fluidity for solute transport under extreme temperature and salinity conditions (e.g. –2 to –35°C and 35–270 ppt in sea-ice brines through an annual cycle). Viruses have been detected in similarly high concentrations in subzero brines, leading to calculations of much higher contact rates between virus and host, and thus greater potential for virally mediated transfer of genetic material, compared to seawater scenarios. Several surface-seawater studies have provided examples of virally mediated gene transfers that alter functional potential and boost fitness of the host, but the relevance of this process to life in subzero brines is poorly known. We used metagenomic and metatranscriptomic sequencing to examine this question in two different types of subzero brine: short-lived brines within landfast sea ice (near Utqiagvik, Alaska, USA), subject to strong T/S seasonality; and ancient cryopeg brines within ~40 000-year-old permafrost, under comparatively stable conditions over time. We gained insights into the adaptive strategies of the distinctive community structures found in each system, as well as early support for virally mediated gene transfer. For example, a higher relative abundance and expression of genes involved in osmo-protection was found in the saltier cryopeg brines (–6°C, 115–140 ppt) compared to the sampled sea-ice brines (–3°C, 75–78 ppt). Genes indicating past viral infections (CRISPR spacers), as well as mobile genetic elements and prophage-like regions, were present in both brine types, assessed at the community level and within individual metagenome-assembled genomes. These gene categories showed higher relative abundance and expression in cryopeg brines; however, the variety of CRISPR spacers detected in sea-ice brine was greater, suggesting that these communities had been exposed to a greater diversity of viruses. Several incidences of long sequences of genes relevant to cold adaptation, including genes for the alteration of membrane-fatty-acid composition, were preceded and terminated by putative viral genes. Ongoing analyses include cross-comparison of putative virus-mediated insertions to corresponding virome datasets.


Observational study of the radiative effect of clouds on the melting of subarctic sea ice

Aura Diaz, Tim Papakyriakou, Jens Ehn

Corresponding author: Aura Diaz

Corresponding author e-mail: umdiaza@myumanitoba.ca

Hudson Bay, an inland sea in the Canadian Arctic, experiences full seasonal ice cover annually that moderates radiative and turbulent heat transport between the atmosphere and ocean following a seasonal cycle. However, synoptic-scale weather events trigger changes in the sea-ice-surface energy budget, which can either advance or delay the seasonal progression of sea-ice melt during spring. Key energy-budget processes to understand the seasonal progression of sea ice are 1) diurnal variations in the incident radiative forcing, 2) cloud radiative forcing, and 3) physical changes in the sea ice and snow pack. Here, we present preliminary results from time-series observations of temperature, salinity and water/ice stable oxygen isotope ratios (δ18O), in conjunction with surface energy-balance observations, of sea ice near Sanikiluaq, southeast Hudson Bay, during the spring melt season of 2019. This is the first year of surface energy-balance observations as a part of a community-based monitoring program with a scientific goal to understand how the radiative forcing of clouds either positively or negatively affects the snow/ice surface temperature and progression of sea-ice melting in a sub-Arctic climate zone.


Wave attenuation in the marginal ice zone due to under-ice turbulence: observations from the BicWin field campaigns

Luc Barast, Peter Sutherland, Dany Dumont

Corresponding author: Luc Barast

Corresponding author e-mail: luc.barast@ifremer.fr

Surface waves and sea ice strongly interact in marginal ice zones (MIZs), but many of those interactions are still lacking quantitative physical explanations. In particular, the effects of turbulence, generated by the interplay between waves and ice, on wave attenuation in the MIZ remain uncertain. In this work, direct measurements of under-ice turbulence and surface-wave attenuation in the MIZ, taken during several field deployments in a natural laboratory in the St Lawrence Estuary, Canada, are used to quantify the contribution of turbulent kinetic energy (TKE) dissipation to wave energy dissipation. Water-column turbulence was measured using a pulse-coherent acoustic Doppler profiler, and wave attenuation was measured using an array of wave buoys deployed on the ice. The distribution of wave-buoys across the MIZ allows observation of a typically exponential decay of energy of up to 3 orders of magnitude below the level of the incoming wave field. Co-located measurements of TKE dissipation in the under-ice boundary layer provide estimates of the quantity of wave energy that is lost to turbulence. The fraction of wave attenuation directly attributable to TKE dissipation can then be calculated, and is found to vary significantly with ice characteristics. The implications of this work, as well as future planned work on this subject, will be discussed.


Retrieval of lake and sea-ice information from Radarsat-2 Quad-pol SAR

Mohammed Shokr, Mohammed Sabboor

Corresponding author: Mohammed Shokr

Corresponding author e-mail: mo.shokr.ms@gmail.com

This study examines three sets of parameters obtained from Radarsat-2 quad-pol data to assess their capability in retrieving geophysical information from lake ice and landfast sea ice in the Resolute Bay area, Nunavut, Canada. The first includes the three orthogonal backscatter coefficients. The second includes the three parameters from Cloude–Pottier polarimetric decomposition, entropy (H), anisotropy (A) and (a). The third includes the three scattering mechanism parameters from Yamaguchi decomposition, single-bounce, double-bounce and random. Data were obtained during the early freezing period from 20 September until end December 2017. Meteorological and ice-climatological observations from the weather station in Resolute were used to facilitate this task. The ice thickness and salinity were obtained from established empirical models. Scattering mechanisms were identified for both lake and sea ice at different freezing stages and used to interpret H/A/a parameters. Results show that backscatter parameters can be used to determine the thickness of lake ice at any growth stage. On the other hand, both backscatter and H/A/a decomposition parameters are related to the thickness of thin sea ice but up to 30 cm thick. A physical explanation is offered in each case. Polarimetric decomposition parameters are found to be too sensitive to scattering mechanisms from the snow cover over lake ice to be useful. On the other hand, H and a have proved to be useful in identifying new sea ice. The orthogonal backscatter coefficients and polarimetric decomposition parameters provide different information. They both offer clues to investigate further questions outlined in the Recommendations section.


Frazil-ice growth and ice production during katabatic wind events in Ross Sea polynyas, Antarctica

Lisa De Pace, Brice Loose, Madison Smith, Sharon Stammerjohn, Jim Thomson, Stephen Ackley

Corresponding author: Lisa De Pace

Corresponding author e-mail: lmdepace@my.uri.edu

During katabatic wind events in the Terra Nova Bay and Ross Sea polynyas, wind speeds exceeded 20 m s–1, air temperatures were below –25°C, and the mixed layer extended as deep as 600 m. Yet temperature and salinity profiles were not perfectly homogeneous, as would be expected with vigorous convective heat loss. Instead, the profiles revealed bulges of warm and salty water starting at the ocean surface and extending to the top tens of meters. Considering both the colder air above and the colder water below, we propose that the increase in temperature and salinity reflects latent heat and salt release during unconsolidated frazil-ice production throughout the upper water column. We use a simplified salt budget to analyze these anomalies to estimate in-situ frazil-ice concentration of 332 and 9.6 g m–3. Contemporaneous estimates of vertical mixing by turbulent kinetic energy dissipation reveals rapid convection in these unstable density profiles, and mixing lifetimes from 2–30 minutes. The corresponding ice production rates median value of 14 cm d–1 compares well with previous empirical and model estimates. However, our individual estimates of ice production of up to 358 cm d–1 reveal the intensity of short-term ice production in the windiest sections of the Terra Nova Bay polynya.


Indicators of Arctic change from passive-microwave dates of sea-ice seasonal evolution

Angela Bliss, Michael Steele, Ge Peng, Walter Meier

Corresponding author: Angela Bliss

Corresponding author e-mail: angela.bliss@oregonstate.edu

Data from satellite passive-microwave sensors have been used to quantify changes in Arctic sea-ice properties for the last 40+ years. Rapid changes in Arctic sea ice that have been observed over this period (e.g. reductions in September sea-ice extent and lengthening of the summer melt season) are a clear indicator of ongoing climate change. The passive-microwave data record is used to quantify changes in the seasonal evolution of Arctic sea-ice cover using a new set of climate-change indicators that describe the timing of dates representing key sea-ice concentration (SIC) thresholds within the sea-ice seasonal cycle. Dates provided in this new dataset include: dates of melt and freeze onset, sea-ice opening (SIC <80%) and retreat (SIC <15%), and sea-ice advance (SIC >15%) and closing (SIC >80%). These dates can be further used to define periods of time during the annual sea-ice growth cycle such as the melt-season length, the ice-loss period, the open-water period, and the ice-growth period. In some cases, there is ambiguity where dates overlap, for example, when the melt-onset date is coincident with the date of retreat or when the freeze-up date is not coincident with the date of sea-ice advance. Temporary SIC changes due to dynamic ice motion can also introduce ambiguity in the dates and the periods of ice loss and growth computed from the dates. In light of these ambiguities, we examine the interoperability of the dates of seasonal sea-ice change to further refine characterization of key stages of the sea-ice melt and growth cycle and to better resolve the temporal progression of the dates through the melt season within the seasonal ice zone.


Object-based image classification of Arctic sea ice and melt ponds through high-spatial-resolution imagery

Xin Miao, Hongjie Xie, Chaowei Yang

Corresponding author: Xin Miao

Corresponding author e-mail: xinmiao@missouristate.edu

The last 10 years have marked the lowest Arctic summer sea-ice extents in the modern era, with a new record summer minimum (3.4 million km2) set in 2012. It has been predicted that the Arctic could be free of summer ice within the next 25 30 years. The loss of Arctic summer ice could have serious consequences, such as higher water temperature, more powerful and frequent storms, rising sea levels, diminished habitats for polar animals, and more pollution due to fossil-fuel exploitation. In these processes, melt ponds play an important role in Earth’s radiation balance since they strongly absorb solar radiation rather than reflecting it as snow and ice do. Therefore, it is necessary to develop the ability to predict sea-ice/melt-pond extents and space–time evolution, which is pivotal to prepare for the variation and uncertainty of the future environment. A lot of effort has been put into Arctic sea-ice modeling to simulate sea-ice processes. However, these sea-ice models were initiated and developed based on limited field surveys, aircraft or satellite image data. Therefore, it is necessary to collect high-spatial-resolution images in a systematic way to tune up, validate and improve models. Currently there are many sea-ice aerial photos available. However, manually delineating sea ice and melt ponds from these images is time-consuming and labor-intensive. In this study, we use the object-based remote-sensing classification scheme to extract sea ice and melt ponds efficiently from high-spatial-resolution images through open source Python programming. The algorithm includes three major steps, as follows: (1) Image segmentation groups the neighboring pixels into objects according to the similarity of spectral and texture information; (2) random forest ensemble classifier can distinguish the following objects: water, submerged ice, shadow, and ice/snow; and (3) polygon neighbor analysis can further separate melt ponds from submerged ice according to the spatial neighboring relationship. Our results illustrate the spatial distribution and morphological characters of melt ponds in different latitudes of the Arctic Pacific sector. This method will be integrated into ArcCI, a polar cyberinfrastructure for spatial–temporal analysis.


N-ICE2015: observational study on drifting Arctic sea ice north of Svalbard from winter to summer

Mats Granskog, the N-ICE2015 Team

Corresponding author: Mats Granskog

Corresponding author e-mail: mats@npolar.no

Between January and June 2015 the Norwegian Polar Institute’s research vessel Lance served as a research station in the drifting sea ice in the Arctic Ocean north of Svalbard (80–83° N) during the international Norwegian young sea ice (N-ICE2015) expedition. The main objective of the campaign was to understand the effects of the shift to a younger and thinner sea-ice regime in the Arctic on energy fluxes, ice dynamics and the ice-associated ecosystem. Perhaps surprisingly, observations over/of/from sea ice in this part of the Arctic, especially in winter, are still scarce. Due to rapid ice drift and ice break-up the expedition consisted of four drifting ice camps, with RV Lance moored to to an ice floe. The ice pack was dominated by relatively thin (<1.5 m) first-year and second-year ice. Here we report on the layout of the study and the main work conducted during the campaign, and summarize major findings. Observations highlight that sea in the Atlantic sector has regionally some unique characteristics, largely due to atmospheric and oceanic forcings that are unique to this region. Frequent storms and warm Atlantic water entering the Arctic in this sector are among the defining factors that govern the sea system in this part of the Arctic.


Towards better understanding of the correlation between undeformed sea-ice thickness and ridge keel draft

Ilija Samardžija, Knut V. Høyland

Corresponding author: Ilija Samardžija

Corresponding author e-mail: ilija.samardzija@ntnu.no

The most important geometrical variables of ice ridges for load calculations on offshore structures are the consolidated layer thickness and the ridge-keel draft. However, the thickness of the surrounding undeformed sea ice is also a vital variable. It is required for estimating the forcing of the surrounding ice on a ridge interacting with an offshore structure. Probabilistic assessment of the ice ridge loads on offshore structures must include correlations between the relevant variables. This study aims to analyse the correlation between undeformed sea-ice thickness and ridge-keel draft. A method is proposed for the analysis of this correlation in a way that the produced results are readily applicable in the probabilistic assessment of ice-ridge loads. Long-term measurements of sea-ice draft from the Beaufort Sea obtained by upward-looking sonars are utilized for this purpose. The data are divided into weekly subsets. A semi-novel approach is used for identifying the draft of undeformed sea ice and the weekly deepest ridge-keel draft is recorded. The correlation between the undeformed sea-ice draft (transferrable to thickness) and the ridge-keel draft is then analysed. A positive linear correlation is found (i.e. thick undeformed sea ice is associated with the presence of deep ridges). This finding is partially in agreement with previous studies, the only discrepancy being that the previous studies indicate a nonlinear nature of the correlation. This discrepancy can be explained by the difference in sampling. The temporal sampling technique used here (week-long fixed time frame of data subsets), in comparison to the spatial sampling technique (fixed equidistant lengths of ice transects, e.g. 50 km), is better suited for quantifying the correlation in a way that the results can be used in probabilistic assessments of the ridge loads. In addition to the proposed method for analysing the correlation of the two variables, we describe a method for probabilistic simulation of the two variables. Our analysis could be repeated for other locations and, if general trends are found, the findings could be used for estimating the ridge-keel-draft statistics for locations where only information about undeformed sea-ice thickness is available. This would help to quantify the risk caused by ice ridges on offshore installations in regions where little to no data about sea ice is available.


Humans at the interface: insights from an anthropology of sea ice

Julianne Yip

Corresponding author: Julianne Yip

Corresponding author e-mail: jcyip@mit.edu

Anthropogenic climate change demands new concepts for thinking about the world. As a global phenomenon, it escapes individual perception, national and international politics. As an anthropogenic phenomenon, it destabilizes divides between mere ‘nature’ and human beings as more than mere nature. At the same time that climate change casts into relief carbon dioxide, sea ice, ice sheets, phytoplankton and other nonhuman entities as actors, it decentres human beings – as well as concepts of ‘politics’, ‘society’ and ‘culture’ that take humans as their reference point. In view of these shortcomings, what existing or emerging vocabularies and conceptual frameworks can help us get traction on climate change? Sea ice – or, more precisely, scientific knowledge of sea ice – offers a conceptual lens for thinking the world. This poster is based on findings from my dissertation research conducting an anthropology of sea ice from 2014 until the present. It draws on fieldwork among sea-ice scientists at mid-latitudes in North America as well as northern fieldwork in Utqiagvik, Alaska, USA. Broadly speaking, I invite people – scientists and non-scientists alike – into the imaginaries opened up by sea ice and climate. Through field anecdotes and sea-ice ‘fables’, I show how sea ice enlarges the field of view beyond human beings to include the atmosphere, ocean and other components of the climate system. These components of the global climate system give rise to humans as we know them – and, in a way, are constitutive of human beings. Through sea ice, humans come into view on geologic timescales as a contingent feature of the planet, as ephemeral as sea ice at the interface between ocean and atmosphere. The challenges of predicting sea-ice conditions on seasonal to interannual timescales challenges assumptions about the timelessness of nature – and conceptions of human agency that presuppose the stability of nature. Sea ice, I argue, offers a conceptual lens to rethink being human today. Findings from this dissertation do not offer prescriptions or formalized knowledge, but a sensibility and perspective on the world. Continued research seeks to examine how to bring such insights into a more public domain in collaboration with artists and filmmakers. In addition to science communication, activism and policy, art can also be a way of shaping new imaginaries. Climate change demands not only political action but a phase transition in thinking the world.


In-situ trial of core sample drilling of extremely thick multiyear landfast ice in the Antarctic

Shuki Ushio, Akio Kobayashi, Morihiro Miyahara

Corresponding author: Shuki Ushio

Corresponding author e-mail: ushio@nipr.ac.jp

Near the edge of ice sheets and glaciers, extremely long-lived perennial sea ice, or multiyear ice (MYI), has existed. Satellite observation data show that MYI with a lifetime of above 30 years as landfast ice has been maintained in Lützow–Holm Bay, near 39° E, Antarctica. This area is characterized by heavy snow, so it is presumed that the sea ice is extremely thick due to the formation of snow ice and/or superimposed ice. Furthermore, at the bottom of the MYI sheet, it is possible that ice is growing due to refreezing of meltwater inflowing from the surrounding ice sheets and glaciers. In this way, ultra-MYI contributes to stability in the landfast-ice area and the physical mechanism of ice-sheet runoff through growing and maintaining under the influence of snow and land ice meltwater. It is considered that the growth history of sea ice is reflected in its physical and chemical characteristics, such as the crystal structure and chemical-omponent profile found in the ice-core sample, so understanding sea-ice structure can also contribute to elucidate the atmosphere–ice-sheet–ocean system. In order to collect sea-ice core samples, we developed and manufactured a mechanical drilling system and have commenced drilling in the summer of 2018/19. Near the terminal of Shirase Glacier, a core sample 3.86 m in length has been collected in very good condition. We will introduce the drill system and report on the drilling operation.


Metagenomic survey of ice adhesin genes in sea ice and seawater at the ice–seawater interface from the Canadian high Arctic

Qinhong Cai, Etienne Yergeau, Christine Michel, Julien Tremblay, Nathalie Fortin, Thomas L. King, Kenneth Lee, Charles W. Greer

Corresponding author: Charles W. Greer

Corresponding author e-mail: Charles.Greer@cnrc-nrc.gc.ca

Ice algae play a critical role in primary production and serve as part of the base of the polar food web. An emerging group of ice-binding proteins produced by bacteria has recently been discovered that enables bacteria to adhere to the surface of ice as well as photosynthetic diatoms. In terms of ecological significance, this symbiotic process involving bacteria, diatoms and ice would enhance solar-energy conversion and nutrient cycling under the ice. Although these ‘ice adhesins’ have been examined in vitro with pure cultures, they have yet to be investigated in the context of indigenous microbial communities dwelling at the ice–seawater interface. In this study, the bottom layer of sea-ice cores and their underlying seawater samples were collected at 23 locations around Cornwallis Island in the Canadian high Arctic. Following extraction and DNA sequencing, representative samples of major taxonomic clusters were selected for shotgun metagenomic sequencing, based on the results of UPGMA (unweighted pair group method with arithmetic mean) analysis of the 16S rRNA gene dataset. Genes encoding two types of ice adhesin were studied: type I with repeats-in-toxin (RTX)-like ice-binding domains; and type II, containing Domain-Of-Unknown-Function (DUF) 3494 for ice binding. To gain insights into the factors regulating this ‘strategic’ ecological process, the relative abundance of ice-adhesion encoding genes was correlated with data on chlorophyll a, bacterial abundance, particulate organic carbon, dissolved organic carbon, dissolved nitrogen, macro-nutrients (NO3 + NO2, PO4, SiOH4) and salinity.


Estimating the sea-ice compressive yield strength (P*) using NASA IceBridge Observations

James Williams

Corresponding author: James Williams

Corresponding author e-mail: jwilli@mit.edu

Using sea-ice thickness transects from NASA operation IceBridge, we are able to quantify the mean thickness gradient north of the CAA. This, along with observed surface wind fields, allows us to estimate the appropriate sea-ice yield strength for geophysical sea ice used in numerical models. It is important that P* be properly calibrated in order to accurately simulate the sea-ice volume. This is because the sea-ice compressive strength is fundamental to determining the forcing required to dynamically thicken the sea-ice.