Anisotropic radiative transfer in sea ice

Christian Katlein, Marcel Nicolaus

Corresponding author: Christian Katlein

Corresponding author e-mail: Christian.Katlein@awi.de

The optical properties of sea ice such as albedo and transmittance are governed by strong scattering of light in the sea-ice matrix. Due to its crystal structure, most physical properties of sea ice are anisotropic. Nevertheless, the optical properties of sea ice are treated as isotropic in most models although existing measurements give a contradicting picture. Understanding radiative transfer and the light-field geometry under sea ice is crucial for correct conversion of radiance data e.g. in AUV or ROV applications. We present ROV-based measurements of irradiance and radiance under summer sea ice of the central Arctic. They allow for insights into the under-ice light field under ponded sea ice and into micro-optical scattering properties of sea ice. Field measurements are interpreted along with numerical radiative transfer calculations, laboratory experiments and microstructure analysis. Our results show that the ratio of synchronous measurements of transmitted irradiance and radiance shows a clear deviation of the angular radiance distribution from the widely used assumption of an isotropic under-ice light field. We show that the angular radiance distribution under sea ice is more downward-directed than commonly assumed. This implies that assuming isotropic conditions under sea ice leads to significant errors in light-field modeling and the interpretation of radiation measurements. The under-ice light field is directly influenced by the anisotropic scattering coefficient of the bottommost sea ice, and the downward-directed radiance distribution causes a deeper light penetration into the ocean under pond-covered ice.


UAV observations of the boundary layer over an Antarctic polynya

John Cassano

Corresponding author: John Cassano

Corresponding author e-mail: john.cassano@colorado.edu

During September 2009 and September 2012 Aerosonde unmanned aerial vehicles (UAVs) were used to study the late-winter boundary layer over the Terra Nova Bay polynya, Antarctica. These flights allowed for the first in situ observations of the wintertime atmosphere over the Terra Nova Bay polynya. Thirty UAV flights, totaling just under 300 hours of flight time, were completed during these two field campaigns. Data collected during these flights include atmospheric state (wind, temperature, humidity and pressure) from 100 to 1500 m AGL, surface temperature, net shortwave and longwave radiation, and aerial photographs of the polynya. These observations allowed us to map the horizontal and vertical extent of the katabatic winds draining from the Antarctic continent into Terra Nova Bay, indicating very large horizontal wind shear at the edge of the katabatic air stream. Data from the flights are being used to analyze the dynamics of the katabatic air stream and to document the downstream modification of the katabatic air mass as it passed over the polynya. Boundary-layer profiles over the polynya allowed for the estimation of surface turbulent heat, moisture and momentum fluxes. The UAV-estimated fluxes have been compared with global atmospheric reanalysis fluxes and fluxes from the Antarctic Mesoscale Prediction System (AMPS) real-time numerical weather prediction model. This presentation will summarize the atmospheric observations collected during the two field campaigns and will also discuss the difficulties encountered while operating in the Antarctic with temperatures down to -40°C and winds up to 40 m s–1.


Comparison of pan-Arctic first-year sea ice, bare ice and melt pond albedo at late stages of melt

Mats Granskog, Stephen Hudson, Donald Perovich, Chris Polashenski

Corresponding author: Mats Granskog

Corresponding author e-mail: mats@npolar.no

The recent shift in the Arctic ice pack from thicker multi-year ice to a thinner ice pack dominated by first-year ice (FYI) has considerable consequences on the energy balance of the Arctic sea-ice cover and the upper ocean. Here we compile a number of observations taken in recent years on the new thinner FYI regime in the Arctic during late stages of melt. We compare observations from nearshore landfast sea ice (Barrow, Alaska) with observations in pack ice (Chukchi Sea and Nansen Basin). We examine the albedo for representative observations on bare ice and melt ponds in all cases. We find that the lowest visible albedos were found for the thickest FYI ice at Barrow, with a distinct drop in albedo at shorter wavelengths, in contrast to the thinner pack ice where there is little wavelength dependence in the visible. Similarly for ponds, the thinner pack ice had somewhat higher albedo with less wavelength dependence in the visible. The red color observed on the Barrow ice suggests the presence of mineral sediments (e.g. dust, soil, sand) in the ice at that nearshore location has a significant effect on both the magnitude of the albedo and on its spectral shape. At near-infrared wavelengths, the ice at Barrow becomes brighter than the pack ice as the sediments compensate for the increasing absorption by ice at these longer wavelengths. We analyse how these differences would translate into energy deposited into the ice and ocean, and thus how well lessons from shorefast ice can be used for pack ice.


Evolution of a directional wave spectrum in a 3-D marginal ice zone with random floe size distribution

Fabien Montiel, Vernon A. Squire

Corresponding author: Fabien Montiel

Corresponding author e-mail: fmontiel@maths.otago.ac.nz

A new ocean wave/sea-ice interaction model is proposed that simulates how a directional wave spectrum evolves as it travels through a realistic marginal ice zone (MIZ), where wave/ice dynamics are entirely governed by coherent conservative wave scattering effects. Field experiments conducted in the Greenland Sea generated important data on wave attenuation in the MIZ and, particularly, on whether the wave spectrum spreads directionally or collimates with distance from the ice edge. The data suggest that angular isotropy, arising from multiple scattering by ice floes, occurs close to the edge and thenceforth dominates wave propagation throughout the MIZ. Although several attempts have been made to replicate this finding theoretically, including by the use of numerical models, none have confronted this problem in a 3-D MIZ with fully randomized floe distribution properties. We construct such a model by subdividing the discontinuous ice cover into adjacent infinite slabs of finite width parallel to the ice edge. Each slab contains an arbitrary (but finite) number of circular ice floes with randomly distributed properties. Ice floes are modelled as thin elastic plates with uniform thickness and finite draught. We consider a directional wave spectrum with harmonic time dependence incident on the MIZ from the open ocean, defined as a continuous superposition of plane waves travelling at different angles. The scattering problem within each slab is then solved using Graf’s interaction theory for an arbitrary incident directional plane wave spectrum. Using an appropriate integral representation of the Hankel function of the first kind, we map the outgoing circular wave field from each floe on the slab boundaries into a directional spectrum of plane waves, which characterizes the slab reflected and transmitted fields. Discretizing the angular spectrum, we can obtain a scattering matrix for each slab. Standard recursive techniques are then used to solve the problem for the full MIZ. We set up a rectangular testing platform containing ~1000 floes to analyse the sensitivity of wave attenuation and spreading of the directional wave spectrum through the MIZ against floe size distribution properties. Ensemble averaging techniques are used to ensure that the results are representative of the random medium defining characteristics.


Distribution of algal aggregates under summer sea ice in the central Arctic

Christian Katlein, Marcel Nicolaus, Mar Fernández-Méndez, Ilka Peeken, Frank Wenzhöfer

Corresponding author: Christian Katlein

Corresponding author e-mail: Christian.Katlein@awi.de

Arctic sea ice is changing dramatically in the last decades and the consequences for the sea-ice-associated ecosystem are difficult to assess. Sea ice is observed to become thinner, younger and more pond-covered. It also allows more light to transmit into and under the ice. Intensive melting might impact the life of sea-ice algae within the brine channels. Algal aggregates underneath the sea ice of the central Arctic have been described sporadically, but the frequency and distribution of their occurrence as well as their role in the ecosystem remain unknown due to the lack of large-scale observations. During the TransArc and IceArc expedition of RV Polarstern in the late summer of 2011 and 2012, we observed ice algal aggregates with a remotely operated vehicle (ROV) underneath various ice types in the central basins. We observed different types of ice algal aggregates floating underneath, and attached to, the underside of the sea ice. Maps of the distribution of aggregate abundance could be obtained by complementing the upward-looking imagery with synchronously recorded ROV attitude and position data. Besides the analysis of spatial distribution patterns and estimates of total biomass of the algal assemblages, this data allowed for shape analysis and the extraction of size distributions. Aggregate distributions can be compared with physical properties of the habitat such as ice thickness, light availability, temperature, salinity and dissolved oxygen measured during the ROV surveys.


Variability and trends in Laptev Sea ice outflow between 1992 and 2011

Thomas Krumpen, Markus Janout, Kevin I. Hodges, Ruediger Gerdes, Fanny Girard-Ardhuin, Jens A. Hoelemann, Sascha Willmes

Corresponding author: Thomas Krumpen

Corresponding author e-mail: tkrumpen@awi.de

Variability and trends in seasonal and interannual ice area export out of the Laptev Sea between 1992 and 2011 are investigated using satellite-based sea-ice drift and concentration data. We found an average total winter (October to May) ice area transport across the northern and eastern Laptev Sea boundaries (NB and EB) of 3.48 × 105 km2. The average transport across the NB (2.87 × 105 km2) is thereby higher than across the EB (0.61 × 105 km2), with a less pronounced seasonal cycle. The total Laptev Sea ice area flux significantly increased over the last decades (0.85 × 105 km2/decade, p > 0.95), dominated by increasing export through the EB (0.55 × 105 km2/decade, p > 0.90), while the increase in export across the NB is smaller (0.3 × 105 km2/decade) and statistically not significant. The strong coupling between across-boundary SLP gradient and ice-drift velocity indicates that monthly variations in ice area flux are primarily controlled by changes in geostrophic wind velocities, although the Laptev Sea ice circulation shows no clear relationship with large-scale atmospheric indices. Also there is no evidence of increasing wind velocities that could explain the overall positive trends in ice export. The increased transport rates are rather the consequence of a changing ice cover such as thinning and/or a decrease in concentration. The use of a back-propagation method revealed that most of the ice that is incorporated into the Transpolar Drift is formed during freeze-up and originates from the central and western part of the Laptev Sea, while the exchange with the East Siberian Sea is dominated by ice coming from the central and southeastern Laptev Sea. Furthermore, our results imply that years of high ice export in late winter (February to May) have a thinning effect on the ice cover, which in turn preconditions the occurrence of negative sea-ice extent anomalies in summer.


Wintertime sea-ice–water biogeochemistry and ocean acidification in a Svalbard fjord – two contrasting years

Agneta Fransson, Melissa Chierici, Daiki Nomura, Mats Granskog, Mikael Hedblom, Svein Kristiansen, Anders Torstensson, Angela Wulff

Corresponding author: Agneta Fransson

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

In 2012 (March and April) and 2013 (April), we investigated the biogeochemistry in sea ice, brine, ice–water interface, water column and frost-flowered snow in the Svalbard fjord Tempelfjorden. The 2 years had contrasting sea-ice conditions; in 2012, the sea ice arrived relatively late and was thinner than in 2013. We sampled between three and five stations, starting at the glacier front to near the ice edge to investigate the effect of glacier ice and water masses on the carbonate system and its drivers. We measured salinity, temperature, total alkalinity (AT), total inorganic carbon (CT), nutrients, stable oxygen isotope (δ18O), bacteria biomass, fatty acids, particulate organic carbon (POC) and nitrogen (PON) in sea ice and ice–water interface. Variability in the properties was observed depending on temperature, ice thickness, and ice–water processes such as brine rejection and CaCO3 precipitation. Glacier water was traced along the transect, using measurements of salinity, AT and δ18O. In addition, we present pH, pCO2 and calcium carbonate saturation levels in the ice–water interface along the transect.


Antarctic sea-ice CO2 system and controls

Agneta Fransson, Melissa Chierici, Patricia L. Yager, Walker O. Smith Jr

Corresponding author: Agneta Fransson

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

In austral summer, from December 2008 to January 2009, we investigated the sea-ice carbon dioxide (CO2) system and CO2 controls in the Amundsen and Ross Seas, Antarctica. We sampled sea water, brine and sea ice for the measurements of total alkalinity (AT), total inorganic carbon (DIC), pH, inorganic nutrients, particulate organic carbon (POC) and nitrogen (PON), chlorophyll a, pigments, salinity and temperature. Large variability in all measured parameters was observed in time and space due to the complex sea-ice dynamics. We discuss the controls of the sea-ice CO2 system, such as brine rejection, biological processes, calcium carbonate (CaCO3) precipitation/dissolution and CO2 exchange. Most (80–90%) of the DIC loss was due to brine rejection, which suggests that the sea ice acted as an efficient DIC sink from 0.8 and 2.6 mol m–2 a–1 (9.6–31 g C m–2 a–1). The remaining change in DIC was to a large extent explained by net biological production. The AT:DIC ratio in the sea ice was higher than in the under-ice water (UIW), suggesting CaCO3 precipitation and concomitant DIC loss in the sea ice. Elevated AT:DIC ratios and carbonate concentrations were also observed in the UIW. The potential for uptake of atmospheric CO2 in the mixed layer increased by ~56 μatm due to the combined effect of CaCO3 precipitation during ice formation, and ice melt in summer.


Taking a look at both sides of the ice: comparing ice thickness and drift speed as observed from above and below sea ice near Barrow, Alaska

Andrew Mahoney, Hajo Eicken, Yasushi Fukamachi, Kay Ohshima, Daisuke Simizu, Chandra Kambhamettu, Rohith MV, Stefan Hendricks, Joshua Jones

Corresponding author: Andrew Mahoney

Corresponding author e-mail: mahoney@gi.alaska.edu

Determining sea thickness from Ice Profiling Sonar (IPS) data requires a variety of additional inputs to convert IPS sonar echo travel-time records to ice draft and ice thickness. Calculating the distance to the ice bottom requires knowledge of the temperature and salinity of the intervening water column, which can be fine-tuned by identifying open-water sections of the record. Having determined the draft of the ice, assumptions regarding the ice and snow cover are required to equate this to the total ice thickness. Moreover, since moored IPSs are stationary, it is also necessary to know the velocity of the ice drifting overhead in order to be able to determine meaningful statistics about the regional ice cover. Such ice velocity data are typically provided by an acoustic Doppler current profiler (ADCP) moored at the same location. However, for long-term moorings, ADCP data are generally recorded less frequently than IPS data and therefore need to be interpolated in time. Here we assemble data acquired near Barrow, Alaska, as part of the Seasonal Ice Zone Observing Network (SIZONet) to make unique comparisons of ice thickness and drift speed determined from above and below the ice. From below, we use ice draft and bottom track data from two moorings that deployed to cover the 2009/10 ice season. From above the ice, we use airborne electromagnetic (AEM) data to measure ice thickness and data from a coastal sea-ice radar to determine ice drift velocity. Sea ice near the first of these moorings was passed over by a helicopter during an AEM survey on 12 April 2010. While further assumptions are necessary to reconcile these two independent measures of ice thickness, we find the mean values are broadly similar, but the maximum keel depths recorded by the IPS are greater than those observed by the AEM survey. By exploring the limits of this comparison, we identify key sensitivities and investigate approaches to improve ice thickness calculations from both approaches. The second mooring lies within the footprint of a coastal ice radar providing ice-drift velocities for comparison with the ADCP bottom track data. During periods of continuous ice cover, there is good agreement between the radar- and ADCP-derived ice velocities including landfast periods and breakout events. However, when ice concentration is low, we note that the bottom track data show higher variance in drift direction and spuriously high instantaneous drift speeds.


Automated detection of melt pond distribution during 2007–13 summers in the Arctic Ocean

Yasuhiro Tanaka, Kazutaka Tateyama, Jennifer Hutchings, Takao Kameda

Corresponding author: Yasuhiro Tanaka

Corresponding author e-mail: d1171400040@std.kitami-it.ac.jp

Ice-albedo feedback is one of the most significant processes in understanding of the Arctic climate. Melt ponds contribute the ice-albedo feedback because the albedo of the melt pond water is lower than that of sea ice and snow cover. This study analyzes the melt pond, sea ice and open water fraction using the forward-looking camera image obtained from science cruises by icebreakers in the Canada Basin since 2007. Forward-looking imagery from forward-looking cameras mounted on icebreakers was analyzed, in ice-covered regions with over 90% concentration, during Beaufort Sea cruises from 2007 to 2012. Camera imagery of such ice was recorded between 78°N and 84°N in the various years. An algorithm is introduced to robustly differentiate between ice, open water and melt ponds in varying light conditions. Analysis of the fractional coverage of melt ponds and ice is presented. We observe melt ponds formed near the North Pole in 2005 and 2011. In 2008, the ice concentration increases with latitude, but high ice cover is not encounted. In 2009 and 2010, melt ponds were rarely observed because cruises occurred after freezing onset. In 2012, sea ice was observed off Barrow and the east of Banks Island. Furthermore, we investigated the frequency of the ratio of melt ponds in 2005, 2008 and 2011. The presence of melt ponds is correlated to the air temperature, which is related to freezing onset. Air temperature is one of the most important factors in the melt pond growth process. We investigated the relation between the ratio of melt pond and air temperature to the mean temperature observed over a few days. We find that the highest melt pond coverage, 20%, occurs when the mean daily temperature is greater than 0°C, and 5% coverage is observed when the mean daily air temperature is below zero. 2011 is unusual, not following this relationship, when melt ponds were observed to be thawed through. In 2011, the ratio of melt pond was found to be correlated with the mean temperature of the past 6 days. This result points to the role of ocean heat flux in delaying freeze onset.


Research highlights from the Sea-ice Environmental Research Facility (SERF)

Feiyue Wang, David Barber, Tim Papakyriakou, Soren Rysgaard

Corresponding author: Feiyue Wang

Corresponding author e-mail: feiyue.wang@umanitoba.ca

The Sea-ice Environmental Research Facility (SERF) is the first experimental sea-ice facility in Canada. Located in Winnipeg on the campus of the University of Manitoba, the main feature of SERF is an outdoor sea-water pool with a movable roof, numerous in situ sensors and instruments, and an on-site trailer laboratory. Sea ice can be created at the pool under various controlled conditions (e.g. seawater chemistry, snow cover, heating) with the additions of chemical, isotopic and/or microbiological tracers. During the first 2 years of operation (2011–13), several types of sea ice, including pancake ice and frost flowers, were successfully created at the SERF pool. Real-time monitoring was carried out on surface and optical properties and on the evolution of temperature, salinity, dissolved oxygen, pH, alkalinity, pCO2 and mercury in and across the sea-ice environment. The results demonstrate that SERF could provide a unique research platform for hypothesis-driven, mesocosm-scale studies to examine geophysical properties and biogeochemical processes in the sea-ice environment. A few SERF case studies, including remote sensing of frost flowers, pH evolution of sea ice, and the dynamics of ikaite and CO2 flux, will be highlighted in this presentation.


The role of sea ice as a natural ocean fertilizer

Delphine Lannuzel, Fanny Chever, Pier Van der Merwe, Julie Janssens, Anne-Julie Cavagna, Arnout Roukaerts, Klaus Meiners

Corresponding author: Delphine Lannuzel

Corresponding author e-mail: delphine.lannuzel@utas.edu.au

One of the most important advances in oceanography was the discovery that iron can limit the productivity of high-nutrient low-chlorophyll areas like the Southern Ocean where iron input to the sunlit layer of the ocean is low. In order to elucidate the ocean carbon cycle in the past, present and future climate scenarios, iron biogeochemistry has increasingly become a focal point in our understanding of the processes that regulate the delivery and retention of this key micro-nutrient in surface waters. It has been demonstrated that sea ice is an important reservoir of iron compared with ice-free waters, but no data currently show whether this iron is easily accessible to micro-organisms. Given the nutritive importance of iron in the sea-ice environment, small changes in the concentration or bio-availability of iron could have a large impact on productivity in the resultant ice-edge blooms during sea-ice melt. Our study quantified the spatial and temporal distribution of iron in sea ice and underlying sea water collected during an interdisciplinary Australian Antarctic fieldwork in East Antarctic pack ice in September–October 2012 (SIPEX-2). Ice samples were processed to obtain for the first time the chemical and physical speciations of iron in the sea-ice environment. The chemical speciation of iron was assessed by determining the level of complexation between iron and organic ligands in sea ice, and the physical speciation was assessed by determining the size fractionation of iron in the soluble (<0.02 μm), dissolved (<0.2 μm) and particulate (>0.2 μm) fractions. Complementary samples were collected for the determination of sea-ice salinity and temperature, sea-ice texture, and the concentrations of macro-nutrients, dissolved and particulate organic carbon, and chlorophyll a in first-year (seasonal) sea ice. Results are presented and discussed in this session.


Carbon contribution of Arctic sea-ice floes

Sang Lee

Corresponding author: Sang Lee

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

The areal extent of Arctic sea ice has rapidly decreased but the areal extent of melt ponds within sea ice recently increased during the Arctic Ocean summer. However, the biological impacts of these changes on the Arctic marine ecosystem have rarely been studied. To estimate detail contributions of particulate organic carbon (POC) as a potential food source in various environments of the Arctic sea-ice floes, intensive works were executed at two different types of sea-ice station in the northern Chukchi Sea during the second Korean Arctic cruise, 2011. The surface ice of melt ponds at ST 1 had the highest POC concentration with a mean of 148.0 mg C m–3, followed by sea-ice cores at ST 2 (mean = 125.7 mg C m–3). The POC concentrations in melt ponds ranged between 90.0 mg C m–3 and 103.9 mg C m–3 at ST 1 and ST 2, respectively. Major POC contributors for melt ponds were diatoms with a mean biovolume contribution of 48.7% (SD = ± 39.1%), which was strongly related to in situ salinity in melt ponds. Although the total POC concentration of entire sea-ice floes ranged from 2.8% to 5.3% of the POC concentration within the euphotic water column at the study locations, the carbon contribution of sea-ice floes could be important to higher trophic levels because of the concentrated POC within sea-ice floes.


Impact of variable atmospheric and oceanic form drag on simulations of Arctic sea ice and Arctic Ocean spin-up

Michel Tsamados, Daniel Feltham, David Schroeder, Daniela Flocco

Corresponding author: Daniel Feltham

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

Over Arctic sea ice, pressure ridges (sails and keels), floe and melt pond edges all introduce discrete obstructions to the flow of air or water past the ice, and are a source of form drag. For most ice types, the form drag contribution to the total drag is of comparable or greater magnitude to the surface or skin drag. In current climate models, form drag is only accounted for by tuning the air–ice and ice–ocean drag coefficients, i.e. by effectively altering the roughness length in a surface drag parameterization. The existing approach of skin drag parameter tuning, while numerically convenient, is poorly constrained by observations and fails to describe correctly the physics associated with the air–ice and ocean–ice drag. Here we combine recent theoretical developments to deduce the total neutral form drag coefficients from properties of the ice cover such as ice concentration, vertical extent and area of the ridges, freeboard and floe draft, and size of floes and melt ponds. We incorporate the drag coefficients into the CICE sea-ice model and show the influence of the new drag parameterization on the motion and state of the ice cover, with the most noticeable being a depletion of sea ice over the west boundary of the Arctic Ocean and over the Beaufort Sea. The new parameterization allows the drag coefficients to be coupled to the sea-ice state and therefore to evolve spatially and temporally. We find that the range of values predicted for the drag coefficients agree with the range of values measured in several regions of the Arctic. Finally we discuss the implications of the new form drag formulation for the spin-up or spin-down of the Arctic Ocean.


The impact of refreezing of melt ponds on Arctic sea-ice thinning

Daniela Flocco, Daniel Feltham, Eleanor Bailey, David Schroeder

Corresponding author: Daniela Flocco

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

The presence of melt ponds over the sea-ice cover in the Arctic has a profound impact on the surface albedo inducing a positive feedback leading to sea-ice thinning. When ponds freeze, the ice that forms on them insulates the pond, trapping it between the sea ice and the ice lid. The pond water also stores a certain quantity of heat that is released with time. Ponds trapped under a layer of refrozen ice have been observed in the Arctic and our model results, confirmed by observations, show that they are present for a few months after the formation of the initial ice lid. In this work we study the ice/water temperature profile in the trapped pond system and its evolution until the pond freezes and show the impact of the presence of a trapped pond on sea-ice growth. We also show the impact of the implementation of the treatment of the salt release and its consequences on the delaying of the refreezing pond.


Nitrogen biogeochemical dynamics in Antarctic pack ice as reflected in the nitrogen isotopes

François Fripiat, Daniel Sigman, Jean-Louis Tison

Corresponding author: François Fripiat

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

We report for the first time nitrate (NO3), total dissolved nitrogen (TDN) and particulate nitrogen (PN) isotope data from Antarctic pack ice in the Weddell Sea (ISPOL, December 2004) and Bellinghausen Sea (SIMBA, October 2007). By measuring NO3 and TDN δ15N, we calculate the δ15N of the remaining dissolved fixed nitrogen, which was dominated by dissolved organic nitrogen (DON), with ~20% ammonium (NH4+). The objectives were twofold: (1) to use the isotopic distributions to better constrain biogeochemical dynamics; and (2) to provide robust information with which to assess the impact of sea-ice N cycling on the sedimentary N isotopic budget. In the sea ice collected, nitrate was significantly depleted. The accumulation of organic N (the dominant form of fixed nitrogen) clearly suggested that the in situ microbial community and biogeochemical conditions were mainly inherited from an earlier period of NO3 assimilation. The concentrations of TDN and PN within the ice are higher (2×) than the underlying winter water, indicating that sea ice continues to take up nitrate from sea water. Partial nitrate assimilation into immobile organic matter at the beginning of the sea-ice bloom concomitant with high δ15N-nitrate loss via convection lowers the δ15N of total fixed sea-ice N relative to the winter-supplied nitrate, throughout the ice floe thickness. The partitioning of N isotopes between PN and DON + NH4+ suggests an initial imbalance between DON production and consumption processes (early spring), which was attenuated with time with the development of an efficient regeneration loop (late spring). The sea-ice δ15N of nitrate, which is frequently lower than sea water, provides direct evidence of the products of nitrification within sea ice, with the mass and isotopic balances suggesting that nitrification contributes substantially (up to ~65%) to NO3 assimilation. Sea-ice PN δ15N is significantly higher (3–4‰) than the δ15N of PN in sea water. It seems unlikely that sea ice alone can explain the variations in the sedimentary PN δ15N during the last ice age, from 0 to 7‰ sea ice would need to have dominated sedimentary PN during the last ice age if it were to explain most of the glacial/interglacial signal. However, given the dramatic decrease in Antarctic productivity during the last ice age and the concomitant equatorward sea-ice expansion, we cannot rule out this possibility for the more polar Antarctic.


East Antarctic sea ice in spring: spectral albedo of snow, nilas, frost flowers and slush, and light-absorbing impurities in snow

Maria Zatko, Stephen Warren

Corresponding author: Maria Zatko

Corresponding author e-mail: mzatko@uw.edu

The solar energy budget of the Antarctic Ocean is largely determined by the fractional area covered by sea ice and by the sea-ice albedo, which is highly variable depending on the surface type. The albedos of open water, nilas, nilas with frost flowers, slush, and first-year ice with both thin and thick snow cover were measured in the East Antarctic sea-ice zone during the Sea Ice Physics and Ecosystems Experiment II (SIPEX II) field campaign from September to November 2012 near 65°S, 120°E. These results augment a dataset from prior ANARE expeditions by extending the spectral coverage to longer wavelengths and by measuring some ice types that had not been encountered on the prior expeditions. Spectral albedo was measured across the ultraviolet (UV), visible and near-infrared wavelengths. The UV-visible albedo of deep snow on sea ice shows no evidence of light-absorbing particulate impurities (LAI) such as black carbon or organics, which is consistent with the extremely small quantities of LAI collected by filtering snow meltwater at the ice stations occupied during SIPEX II. At the visible and ultraviolet wavelengths, the albedo depends on the thickness of snow or ice; in the near-infrared the albedo is determined by the specific surface area. The albedo of bare nilas increases with nilas thickness, but the growth of frost flowers causes the nilas albedo to increase dramatically by about 0.2 at UV and visible wavelengths. Slush is formed when blowing snow falls into a lead of open water; its albedo increases with slush thickness. For all surface types except open water, albedos are highest in the UV and visible and decrease towards the infrared. The spectral albedos are integrated over wavelength, using both clear-sky and cloudy-sky incident solar spectra, to obtain broadband albedos for wavelength bands commonly used in climate models.


Sea-ice electromagnetic profile reconstruction through inverse scattering solution by optimization

Nariman Firoozy, Puyan Mojabi, David Barber

Corresponding author: Nariman Firoozy

Corresponding author e-mail: firoozyn@myumanitoba.ca

The Arctic is the first frontier for climate change due to its delicate thermodynamic heat exchange between ocean and atmosphere. As a consequence of global warming, it is predicted that there will be an ice-free Arctic for the month of September within the next few decades. This highlights the need for comprehensive observation and monitoring of Arctic ice. Microwave imaging is a promising method for providing information about Arctic sea ice. In this non-destructive imaging method, the dielectric profile of sea ice is to be reconstructed using measured microwave scattering data. This reconstructed dielectric profile not only provides a valid model for SAR data interpretation, but can also be related to thermodynamic properties of ice, and through proxy formulation, to its physical properties like brine volume fraction and thickness. In this paper, the complex permittivity of sea ice and its thickness are simultaneously reconstructed by solving an electromagnetic inverse scattering problem. Two global optimization methods, namely the Particle Swarm Optimization (PSO) and Differential Evolution (DE) methods, are used to solve the associated inverse scattering problem. The reconstruction results using these two methods are then presented and compared. It should be noted that these optimization algorithms require an appropriate forward scattering algorithm. Utilized electromagnetic forward solver is based on the Boundary Perturbation Theory (BPT), which is equivalent to the Small Perturbation Method of the second order with coherent and incoherent NRCS of ice in bistatic measurements and coherent NRCS in monostatic measurements. This paper also investigates the effect of different measurement configurations and the frequency of operation on the reconstruction accuracy achievable from microwave sea-ice imaging. In particular, the use of P-band for this application is considered. Based on reconstruction results, P-band is a promising frequency band which allows accurate profile reconstruction for the most common case of first-year ice. The use of this frequency band is currently being investigated for soil remote-sensing studies, but little on the use of this frequency band has been studied in the Arctic environment. Therefore, this paper also sets the guidelines for the measurement setup, which will be carried out at SERF facilities at the University of Manitoba, Winnipeg, Canada, and later in an Arctic field camp located on sea ice near Cambridge Bay, NT in 2014.


Ocean waves in the Arctic: observations and trends

Alexander Babanin, Stefan Zieger, Agustinus Ribal

Corresponding author: Alexander Babanin

Corresponding author e-mail: ababanin@swin.edu.au

Wind waves are a new physical phenomenon to the Arctic seas, which in the past were covered with ice. Now, over summer months, ice coverage retreats up to 83.5°N and waves are generated. The marginal open seas provide new opportunities and new problems. Navigation and other maritime activities become possible, but wave heights, storm surges and coastal erosion will likely increase. Air–sea interactions enter a completely new regime, with momentum, energy, heat, gas and moisture fluxes being moderated or produced by the waves, and impacting on upper-ocean mixing. All these issues require knowledge of the wave climate. We will report investigation of wave climate and its trends by means of satellite altimetry. This is a challenging but important topic. On one hand, no statistical approach is possible since in the past for most of the Arctic Ocean there was limited wave activity. Extrapolations of the current observations into the future are not feasible, because ice cover and wind patterns in the Arctic are changing. On the other hand, information on the mean and extreme wave properties, such as wave height, period, direction, on the frequency of occurrence and duration of the storms is of great importance for oceanographic, meteorological, climate, naval and maritime applications in the Arctic seas.


Arctic sea-ice forecasting service during CHINARE 2012

Chunhua Li, Ming Li, Zhongxiang Tian, Xingren Wu, Lin Zhang, Fubin Liu, Shang Meng

Corresponding author: Chunhua Li

Corresponding author e-mail: lichunhua0214@hotmail.com

In an effort to facilitate the Chinese Arctic Research Expedition in summer 2012 (CHINARE 2012), sea-ice forecasting service for the Arctic Ocean was conducted. Based on SSM/I (Special Sensor Microwave/Imager) Arctic sea-ice extent monthly data from NSIDC (National Snow and Ice Data Center), the outlook for Arctic sea ice was conducted from July to September 2012 using statistical methods. 10 days and monthly ice prediction tests were also done with NCEP (National Center for Environmental Prediction) Climate Forecast System Versions 2 (CFSv2). In addition, a coupled ice–ocean model was used to forecast Arctic sea-ice change in synoptic scale; the initial condition of sea-ice concentration was from SSMIS-derived data (http://www.iup.uni-bremen.de:8084/ssmis/) and the atmospheric forcing was from the NCEP GFS (Global Forecast System). The 72–120 h sea-ice concentration numerical forecasting products were issued. The sea-ice analysis and forecast service provided effective technical support for the navigation routing and investigation plan in the ice-covered ocean for the expedition.


Distribution of dissolved and particulate barium in Antarctic sea ice

Stephanie Jacquet, Sandrine Chifflet, Anne-Julie Cavagna, Frank Dehairs, Murence Monin, Delphine Lannuzel

Corresponding author: Stephanie Jacquet

Corresponding author e-mail: stephanie.jacquet@univ-amu.fr

Dissolved Ba (D-Ba) in oceanic waters behaves as a bio-intermediate element and presents strong similarities to nutrients and alkalinity distributions. However, the biogeochemical cycling of Ba is unique and strongly linked with the particulate biogenic Ba (P-Ba; mainly formed as barite crystals, BaSO4) dynamics. In surface waters, P-Ba forms during particulate organic matter degradation by prokaryotes activity, and D-Ba subtraction in ambient water during this process is responsible for subsurface water D-Ba depletion. In sediments, P-Ba is used as an indicator of surface (paleo)-export production. To go further in understanding the interaction between the dissolved and particulate Ba phases and controls on their dynamics, P-Ba and D-Ba were measured in Antarctic sea ice (SIPEX-II cruise, 2012) where previous studies interestingly reported barite accumulation. Results will be confronted to sea-ice biogeochemical properties and discussed in the context of the importance of the Antarctic sea-ice zone as a driver and indicator of global climate processes.


Assessing ice-induced attenuation of water waves in a directional wave basin

Alessandro Toffoli, Alberto Alberello, Luke Bennetts, Michael Meylan, Alexander Babanin

Corresponding author: Alessandro Toffoli

Corresponding author e-mail: toffoli.alessandro@gmail.com

Wave–ice interaction is a critical factor in the dynamics of the marginal ice zone (MIZ), the region between open ocean and an expanse of ice floes of varying size and shape. This interaction works both ways: while waves cause the fractures of ice floes, the presence of ice floes affects waves through scattering and various dissipative processes. In order to assess the latter, a laboratory experiment has been carried out in the coastal directional basin at Plymouth University. Sea ice has been simulated with a 1 m × 1 m plastic sheet with variable thickness of polypropylene, which holds the same density (~0.9 g cm–3) of ice, and PVC Forex, which hold the same mechanical property of ice. Experiments have been conducted using monochromatic as well as random wave fields with different steepness and wavelengths (both shorter and larger than the floe). The wave field has been monitored before and after the simulated ice floe with a number of wave probes deployed along the basin, including a six-probe array to track directional properties. On the whole, results show a substantial scattering and dissipation of the wave field, which appears to be dependent on the amount of overwash on the ice floe.


A sensitivity study of the sea-ice simulation in the global coupled climate model, HadGEM3

Jamie Rae, Helene Hewitt, Ann Keen, Jeff Ridley, John Edwards, Chris Harris

Corresponding author: Jamie Rae

Corresponding author e-mail: jamie.rae@metoffice.gov.uk

We will present some results from the first sea-ice sensitivity study to be performed with a fully coupled global atmosphere–ice–ocean climate model. Results will be presented for sensitivity to a selection of sea-ice parameters, varied within the range of observational uncertainty, and additionally for the sensitivity of the sea ice to increased resolution in the atmosphere and ocean–ice models, as well as dynamics and physics changes in the atmosphere. In the Arctic, the sea-ice thickness is most sensitive to the albedo of the overlying snow layer (because of its influence on surface melt) and the thermal conductivities of ice and snow (because of their role in regulating heat flux from the ocean to the atmosphere through the ice). The winter Arctic ice extent is found to be sensitive to an increase in resolution of the ocean–ice model, because of increased sea surface temperatures in the Labrador Sea at higher resolution. In addition, the Arctic ice extent is reduced under increased atmospheric resolution, because better representation of mid-latitude storms leads to increased poleward heat transport. In the Antarctic, the sensitivity to sea-ice parameters is weaker, and atmosphere and ocean forcing dominate; in particular, increased resolution of the atmosphere and ocean–ice models leads to the enhancement of a warm bias in the Southern Ocean, which has a large impact on sea-ice thickness and extent. Inclusion of a selection of these parameters in combination, together with changes to the atmosphere and ocean models, leads to significant improvements in depiction of sea-ice extent, thickness and volume.


Atmosphere–ice–ocean feedbacks in a fully coupled global climate model with improved atmosphere–ice interaction

Jamie Rae, David Schroeder, Daniel Feltham, Helene Hewitt, Ann Keen, Alison McLaren, Jeff Ridley

Corresponding author: Jamie Rae

Corresponding author e-mail: jamie.rae@metoffice.gov.uk

Atmosphere–ice–ocean interaction is of critical importance for the energy balance of the polar regions. In summer, the high albedo of sea ice reduces ocean warming, while in winter its low thermal conductivity acts to insulate the cold atmosphere from the warmer ocean below. In addition, the presence of melt ponds on the ice surface can have a strong influence on the ice albedo. As a consequence, a fully coupled climate model is required to describe atmosphere–ice–ocean interactions sufficiently, both in present-day climate simulations and in future projections. A number of improvements to the atmosphere–ice coupling in the coupled climate model HadGEM3 will be discussed, including making the surface fluxes fully dependent on the sub-gridscale ice thickness distribution, replacing the previous broadband albedo scheme with the four-band parameterization from the CICE sea-ice model, and introducing a detailed representation of the effect of surface melt ponds on ice albedo to replace the previous heavily parameterized scheme. We will discuss the impact of these changes on the simulation of present-day and future climate, both in the polar regions and further afield.


ALVERT: a new sea-ice albedo parameterization through successive combination of vertical layers

Keguang Wang, Caixin Wang, Sebastian Gerland, Mats Granskog

Corresponding author: Keguang Wang

Corresponding author e-mail: keguang.wang@met.no

Sea-ice albedo and its positive ice-albedo feedback play a key role in the global climate system. When sea ice becomes thin, its thickness has a significant effect on the surface albedo, forming a positive thin ice-albedo feedback. In this paper, we introduce a new sea-ice albedo parameterization (ALVERT), such that, for any two vertically combined layers, the albedo of the upper layer can be expressed by its thickness and semi-infinite albedo together with the albedo of the lower layer. This process can be repeated infinitely when new layers are successively added onto the top. To validate this parameterization, we have performed extensive comparisons with a variety of in situ albedo observations on sea ice, including cold/blue/white bare ice, as well as dry/melting snow-covered ice. The good agreement with observations shows that this new parameterization is a simple yet comprehensive means to describe the surface albedo for a broad range of vertical combinations of water, ice and snow, particularly when the ice is thin.


Microfeatures of modern sea-ice-rafted sediment: implications for paleo-sea-ice reconstructions

Kristen St John, Sandra Passchier, Brooke Tantillo

Corresponding author: Kristen St John

Corresponding author e-mail: stjohnke@jmu.edu

The analysis of grain surface microfeatures in sedimentary deposits has successfully been used as a method to infer regional glacial histories. Yet in glaciomarine settings where sea ice is an additional transport mechanism, the need to differentiate between glacial (iceberg-rafted debris; IRD) and sea-ice-transported sediments becomes important. Paleoclimate reconstructions especially depend on differentiating between glacial ice and sea ice as these have different formation and transport histories and play different roles in climate-system feedbacks. However, there is a dearth of paleo-sea-ice proxies, and the sedimentological approaches largely depend on weakly tested assumptions. Therefore, the purpose of our study was: (1) to characterize quartz grain microfeatures of modern sea-ice-rafted debris (SIRD) from the Arctic Ocean; and (2) to compare these results with microfeatures from representative Pleistocene glacial deposits to potentially differentiate SIRD from IRD, thus evaluating the use of grain microfeatures as a paleo-sea-ice proxy. SEM observations of shapes, relief and microtextures of 254 quartz grains collected from recent multi-year sea-ice floes in the Arctic Ocean, and from 104 quartz grains collected from tills in Poland, were recorded using a checklist approach. SIRD grains were largely sub-rounded, with medium relief, high relative abundance of breakage blocks and silica dissolution, whereas the Pleistocene glacial grains contained high abundances of conchoidal fractures, steps, isolated cusps, fractures and striations/gouges. Discriminate analysis of these data show a distinct difference between SIRD and glacial grains, with <7% of the SIRD grains containing typical glacial microtextures, suggesting this method is a useful means of inferring paleo-sea-ice presence in the marine record when large numbers of grains are examined. We propose that differences in microfeatures of SIRD and IRD in paleo-records may reflect differences in their transport and weathering histories. SIRD would presumably have a more complex history than that of grains rafted by icebergs; grains transported to, and deposited in, a coastal setting via fluvial or glacial processes and later rafted by sea ice would be subject to increased chemical weathering, whereas ice sheets and glaciers that reach sea level and calve icebergs would essentially bypass the periglacial and coastal marine environments, thus preserving their glacial signature in IRD.


Comparison of AMSR-E concentration product, MODIS and pseudo ship observation records at Antarctic summer sea-ice edge

Xi Zhao, Haoyue Su, Xiaoping Pang, Alfred Stein, Zian Cheng

Corresponding author: Xi Zhao

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

Previous research pointed to a low agreement in the marginal edge zone, especially during the ice melting season. According to the ASPeCt standard ship observation protocol, the sea-ice edge is defined as the northernmost occurrence of sea ice of at least 10% concentration, where the ice concentration is estimated from the ship’s bridge within a 1 km radius of the ship. In addition, a 15% ice concentration is commonly used when determining the sea-ice edge with PM concentration products. Several studies used ship records of ice edges or manually digitized ice edges from high-resolution images to assess the edges derived from this threshold. The number of ship observation records that exactly located the sea-ice edge, however, is limited. To avoid subjective digitation or using a small set of ship observations, this study proposed for assessing the usability of the 15% threshold to: (1) identify sea-ice edges from pseudo ship observation records that were artificially generated from optical satellite images; and (2) compare the results from (1) with concentrations and edges from PM products. Five pairs of synchronous AMSR-E concentration products and MODIS images from summer months in the Antarctic were co-registered. To simulate ice-edge determination following the ASPeCt protocol, we first classified each MODIS pixel into ‘ice’ and ‘non-ice’ and pixels were aggregated to a 2 km cell to simulate the 1 km radius condition of a ship record. Corresponding concentrations were calculated for each cell, thus generating pseudo ship observation records. Outlines of cell regions with 10% concentration were marked as Ship-based-sea-ice-Edges (SE) from pseudo ship observations. The 15% threshold of the AMSR-E concentration product was used to mark Image-based-sea-ice-Edges (IE) and the locations of SE and IE were compared. We sampled 100 to 400 AMSR-E pixels at a SE for each scene and conducted significance tests to find out whether the 15% threshold is pertinent. Based on 1215 pseudo ship observations, results showed that the Antarctic summer sea-ice concentrations in edge zones are inconsistent. The mean concentrations from the AMSR-E pixels locating at SE were significantly different from 15% in three out of five image pairs. The lowest and largest mean AMSR-E concentrations were far off from the acknowledged 15%. Correlation coefficients between AMSR-E and MODIS were below 0.5 and were much lower than reported in previous studies.


An assessment of the causes of open-water leads and changes in sea-ice conditions along northern Ellesmere Island, Canada

Miriam Richer McCallum, Derek Mueller, Luke Copland

Corresponding author: Miriam Richer McCallum

Corresponding author e-mail: miriamrichermccallum@cmail.carleton.ca

Remote-sensing imagery indicates that open-water leads have increased in area and duration along the northern coast of Ellesmere Island over the past decade. These now frequent leads have been associated with the calving of ice shelves and multiyear landfast sea ice, meaning that an understanding of lead formation is important to understand the causes of cryospheric losses. In this study, summer weekly Canadian Ice Service Digital Archive (CISDA) geospatial data from 1999 to 2011 were analyzed to quantify the temporal and spatial occurrence of open water. Three classes of sea-ice concentration were considered as leads: (1) open water, consisting of open water and bergy water with ice concentrations <1/10; (2) very open drift, consisting of sea-ice concentrations from 1/10 to 3/10; and (3) open drift, consisting of sea-ice concentrations of 4/10 to 6/10. Climate variables were obtained from the National Center for Environmental Prediction and the National Center for Atmospheric Research (NCEP/NCAR re-analysis). Open-water leads were first observed during the summer of 1999 and were detected every summer until 2011, with the exception of 2006. The 13 year record showed appreciable interannual variability in the extent and duration of leads, with an overall maximum in 2008. The majority of the leads were observed during the months of August and September, although the onset of leads began in June and July in recent years. Leads were most frequently observed west of the northern point of Ellesmere Island, which is consistent with westerly circulation of Beaufort Gyre. Greater open-water extent was generally following periods of strong positive winter Arctic Oscillation and during strong negative summer Arctic Oscillation. Their formation was more prevalent during and following episodes of offshore or parallel winds. Leads did not form when air temperature was below -8°C, which points to a thermal threshold that must be exceeded to pre-weaken the sea ice. The results suggest that a combination of factors is important for the development of open-water leads along the northern coast of Ellesmere Island and this may have implications for future ice-shelf attrition and sea-ice losses.


Spatial variability of sea ice and Arctic shipping activities in Canada, 1990–2012

Larissa Pizzolato, Stephen Howell, Jackie Dawson, Luke Copland, Chris Derksen

Corresponding author: Larissa Pizzolato

Corresponding author e-mail: larissa.pizzolato@uottawa.ca

Declining Arctic sea ice has attracted significant attention with the prospect of increased shipping activity in the Canadian Arctic. Canada contains two of the Arctic’s key shipping corridors, the Northwest Passage (NWP) and the Arctic Bridge. Developing a greater understanding of the relationship between shipping activity and sea-ice changes in the Canadian Arctic is therefore important for issues such as policy development, assessment of potential impacts on wildlife, and improved search and rescue capacity. Observational analysis was undertaken of sea-ice area and vessel traffic within the Vessel Traffic Reporting Arctic Canada Traffic Services Zone (NORDREG zone) encompassing the Canadian Arctic. Within the NORDREG zone between 1990 and 2011 total sea-ice area experienced a declining trend of 26 × 103 km2 a–1, while observed increases of vessel activity were between 1 and 17 vessels decade–1 depending on the month and vessel type. However, a knowledge gap exists with respect to spatial linkages between sea-ice area and observed shipping patterns. To better understand this relationship, spatio–temporal analysis of sea-ice area and vessel distribution patterns specific to shipping corridors was undertaken using ship routes generated from daily positional data provided by the Canadian Coast Guard and the Canadian Ice Service Digital Archive over the 1990–2012 time period.


Methods and madness in sea-ice biogeochemistry: intercomparisons of observational techniques

Lisa Miller, Lynn Russell

Corresponding author: Lisa Miller

Corresponding author e-mail: lisa.miller@dfo-mpo.gc.ca

Sea-ice biogeochemistry currently relies on a chaotic assortment of methodological approaches originally derived from diverse disciplines, including glaciology, oceanography, sedimentology and even tundra ecology. As the field matures, the time has come to address the disparities, commonalities, strengths and weaknesses of the methods being used, and to develop standard approaches for the most common and important analyses. Towards that end, SCOR working group 140, Biogeochemical Exchange Processes at the Sea-Ice Interfaces (BEPSII), has established a task group on methodologies and intercomparisons. This task group is compiling comprehensive information on the methods currently in use and has identified a critical need for intercalibration and intercomparison experiments for measurements of biomass, primary production, nutrients, dissolved and particulate organic matter, the carbonate system, CO2 fluxes and aerosol production in sea ice. Next, we are designing many of these required experiments, including developing expedition plans and exploring potential research sites and platforms. These efforts are proceeding with input from a large fraction of the sea-ice biogeochemistry community, and we encourage contributions from, and participation by, all interested individuals and institutions.


Sea-ice motion in the Beaufort Sea: 1997–2012

Michael Brady, Stephen Howell, Chris Derksen, Richard Kelly

Corresponding author: Michael Brady

Corresponding author e-mail: m2brady@uwaterloo.ca

The Beaufort Sea is a key area of interest for natural resource exploration. Displacement of sea ice within the region is partially governed by a unique transport mechanism known as the Beaufort Gyre. The Beaufort Sea has experienced considerable reductions in summer sea-ice extent in recent years but few studies have investigated related changes in sea-ice motion and flux within the region over long time periods. In this study, all available ScanSAR imagery from RADARSAT-1 and RADARSAT-2 were used to estimate total monthly sea-ice motion in the Beaufort Sea from 1997 to 2012. For estimation of sea-ice flux between the Beaufort Sea and other Arctic regions, three areas were selected to serve as gates of exchange: Point Barrow, M’Clure Strait and Cape Andreasen. Sea-ice motion was determined by the Canadian Ice Service-Automated Sea Ice Tracking System (CIS-ASITS) applied to RADARSAT image pairs. CIS-ASITS ice motion estimates have an error of approximately 0.43 km d–1 based on analysis in previous studies. Limitations of the algorithm include difficulty in analyzing ice floes that have some component of surface melt, and delineating between ice floes and open water in areas of anomalous surface wind activity. Monthly motion composites were produced from an average of 200 images per month. Results were validated with operational sea-ice charts from the Canadian Ice Service Digital Archive, International Arctic Buoy Program (IABP) drifting-buoy observations, and National Snow and Ice Data Center (NSIDC) ice draft and velocity datasets.


Towards improving seasonal sea-ice predictability

Julienne Stroeve, Larry Hamilton

Corresponding author: Julienne Stroeve

Corresponding author e-mail: stroeve@nsidc.org

Decline in the extent and thickness of Arctic sea ice is an active area of scientific effort and one with significant implications for ecosystems and communities in the Arctic and global climate. While global climate circulation models all predict seasonal ice-free conditions within the second half of this century, forecasting on seasonal timescales remains challenging due to the variable nature of weather and ocean behavior as well as current limits to data and modeling capabilities. Forecasting ice conditions in the summer and into the fall is of particular interest to many stakeholders since many activities that take place in the Arctic are planned over the summer months, and many species are sensitive to the behavior of summer sea ice. The SEARCH Sea Ice Outlook (SIO) has been an informal network of scientists and stakeholders to improve and communicate sea-ice prediction knowledge and tools. The next step is to move this into a more formal network aimed at: (1) coordinating and evaluating seasonal predictions, (2) integrating, assessing and guiding observations, (3) synthesizing predictions and observations and (4) disseminating predictions and engaging key stakeholders. Here we present an overview of predictions contributed to the SEARCH SIO during the last 6 years, and evaluate how well they did in particular years and efforts to improve these seasonal sea-ice forecasts.


Investigating nitrate photolysis in polar snow, snow on sea ice, and mid-latitude snow

Maria Zatko, Becky Alexander

Corresponding author: Maria Zatko

Corresponding author e-mail: mzatko@uw.edu

We have implemented the photolysis of snow nitrate into a global chemical transport model for the first time in order to calculate the flux and redistribution of nitrogen in polar snowpacks, mid-latitude snowpacks, and snow on sea ice in the Arctic and Antarctic. The calculated potential flux of NOx from polar snow is comparable with observations. Model results suggest that nitrate can be recycled many times in low snow accumulation rate regions before burial below the photolytic zone. We use the model to examine the implications of snow nitrate photochemistry for boundary layer chemistry, the recycling and redistribution of nitrogen and its spatial variability, and the preservation of nitrate in ice cores. It has been suggested that some of the NOx produced by nitrate photolysis in snow on the Antarctic plateau is transported off the continent and into the Antarctic sea-ice zone. Here we present isotopic evidence that supports this hypothesis; large negative d15N values of nitrate in surface snow on first-year sea ice were measured in the East Antarctic sea-ice zone during SIPEXII.


Numerical experiments on a polar low developed over the Barents Sea in January 2011

Taku Mitsui, Jun Inoue, Atsuyoshi Manda

Corresponding author: Taku Mitsui

Corresponding author e-mail: ria10.321baku82@gmail.com

Variability of sea-ice extent over the Barents Sea is one of the important factors in predicting Japanese winter coldness. The mechanism strongly depends on cyclone activities at the marginal ice zone over the Barents Sea. In January 2011, we conducted observations over the Barents Sea with the aid of the Norwegian R/V Johan Hjort, and succeeded in observing a polar low developed along the marginal ice zone on 22 January. Polar lows are violent storms that happen to be one of the most destructive mesoscale weather systems over the North Atlantic Ocean. In this study, we conducted numerical experiments using a regional atmospheric model for elucidating the relationship between the polar low development and sea-ice distribution. The Weather Research and Forecasting model (WRF ver3.3) has been used for the simulations. The target period is from 00 UTC on 20 January to 00 UTC on 25 January 2011. The model domain covers surrounding the Barents Sea sector (2800 km × 2000 km) with 10 km horizontal resolution and 38 vertical layers. ERA-Interim reanalysis is used as the initial and boundary conditions. We conducted two sensitivity experiments for evaluating the influence on the area of sea ice by replacing surface boundary conditions. ICE_HY (heavy ice year experiment) and ICE_LY (light ice year experiment) used the observed sea-ice concentration and sea surface temperature on 20 January in 2004 and 2006, respectively. Comparing with the observed temperature vertical profile and shipboard sea-level pressure time series, the tropopause fold and a sudden drop in sea-level pressure were well reproduced. Sea-level pressure was about 5 hPa fallen for ICE_HY with ICE_LY at 00 UTC 24 January, we elucidated the development of polar low of ICE_LY. To understand this development, we conducted dry experiments for ICE_HY and ICE_LY to consider the effect of condensate heat. As a result of comparison of the time series of eddy kinetic energy, the contribution of condensation heat is much large.


Analysis of fast-ice observation at Zhongshan station in 2012

Jiechen Zhao, Qinghua Yang, Surui Xie, Chunhua Li, Ming Li, Lin Zhang

Corresponding author: Jiechen Zhao

Corresponding author e-mail: zhaojc@nmefc.gov.cn

Fast-ice measurements off the Chinese Antarctica Zhongshan station, East Antarctica, are usually conducted every year from early August to late November. Here we compare results from manual drilling of ice thickness and snow depth with those from non-destructive methods, such as electric capacitance ice thickness sensor. This paper reported the results on automated observations, such as upward (downward) shortwave and longwave radiation sensors, sonar snow depth sensor, and infrared ice surface temperature sensor, which have been used in fast ice from 2010 onwards. The record of fast-ice observations at Zhongshan station and some more detailed results from 2012 will be presented here.


Sea-ice mineralogy: narrowing the gap between experimental and modeling approaches

Feiyue Wang, Soren Rysgaard, Nicolas-Xavier Geilfus, Marcos Lemes, Wen Xu, Mark Cooper, Giles Marion, Norman Halden, Frank Hawthorne

Corresponding author: Feiyue Wang

Corresponding author e-mail: feiyue.wang@umanitoba.ca

The discovery of ikaite in sea ice has revealed a new exciting mechanism for the exchange of CO2 between the atmosphere and ocean. The formation of ikaite, along with several other minerals, was predicted more than 50 years ago based on laboratory freezing experiments. The highly saline sea-ice brine and its sub-zero temperatures create a unique environment where minerals can be precipitated and crystallized rapidly, although their significance has only begun to be appreciated. Here, we report identification of ikaite, gypsum, mirabilite and NaBr in Arctic and experimental sea ice, and show that their dynamics can be reasonably described by FREZCHEM, a thermodynamic equilibrium model for geochemical processes at sub-zero temperatures. However, our ability to quantify the importance of these minerals in sea-ice biogeochemical processes is hindered by a lack of fundamental understanding of the kinetics of cryospheric mineralization.


A wave–ice interaction model for climate studies

Luke Bennetts, Siobhan O’Farrell, Petteri Uotila, Vernon Squire

Corresponding author: Luke Bennetts

Corresponding author e-mail: luke.bennetts@adelaide.edu.au

The Antarctic marginal ice zone (MIZ) extends for tens to hundreds of kilometres in width due to the impact of powerful Southern Ocean waves. However, sea-ice models do not yet account for the unique dynamic and thermodynamic properties of MIZ ice cover, which result from the relatively small floe sizes in the MIZ. Developing a diagnostic tool to determine MIZ width and floe size distribution within the MIZ is the first step towards incorporating a faithful model of the MIZ in next generation climate models. Modelling wave–ice interactions underpins determination of the width of the MIZ and its floe size distribution. The two key components to wave–ice interactions are: (1) the wave-induced fracture of the ice cover, which regulates the maximum size of ice floes; and (2) the attenuation of waves due to the presence of ice cover, which limits the distance waves penetrate into the ice-covered ocean and hence the distance over which fracture occurs. Wave attenuation due to ice cover and wave-induced fracture of the ice cover are coupled processes that require novel modelling techniques. A wave–ice interaction model has recently been developed for use in short- to mid-term operational forecasting models. The wave–ice interaction model is the first model to incorporate both wave attenuation and wave-induced ice fracture. High-resolution spatial and temporal grids are nested in regions of operational interest to implement the wave–ice interaction model numerically. Notwithstanding, this direct approach is far too expensive for the temporal and spatial scales used in climate models. An alternative, low-cost wave–ice interaction model has therefore been developed for use in climate models. The new wave–ice interaction model determines the floe size distribution in a given gridcell by balancing the spatial distribution of wave energy in the gridcell and the floe size distribution due to wave-induced fracture of the ice cover. No subgridding is required. An idealized version of the new wave–ice interaction model will be presented at the symposium.


SIPEX II: the impact of Southern Ocean storms on Antarctic sea ice

Alison Kohout

Corresponding author: Alison Kohout

Corresponding author e-mail: a.kohout@niwa.co.nz

Antarctic sea ice is highly influenced by the dynamic nature of the Southern Ocean. Ocean waves can propagate from tens to hundreds of kilometres into sea ice, leaving behind a wake of broken ice sheets. As global climate change intensifies, storm intensity will increase in the Southern Ocean. Increased storm intensity will bring stronger winds and bigger waves, which have the potential to travel deeper into the ice pack and increase the likelihood that ice floes break apart. To enhance our understanding of this system, our aim during the Antarctic Program’s Sea Ice, Physics and Ecosystems eXperiment (SIPEXII) was to build on the scarce Antarctic waves-in-ice dataset by collecting a set of wave observations in the marginal ice zone. Five wave sensors were deployed on Antarctic sea ice. They were spread over 200 km along a meridional transect line at 121°E. Every 3 hours, the sensors simultaneously woke and recorded their location and a burst of wave acceleration data. Three significant storm events were captured. The effect these storms had on the presence of waves in sea ice is shown.


Slight changes in timing of sea-ice retreat could have major impacts on polar benthic ecosystems

Graeme Clark, Jonny Stark, Emma Johnston, John Runcie, Paul Goldsworthy, Ben Raymond, Martin Riddle

Corresponding author: Martin Riddle

Corresponding author e-mail: martin.riddle@aad.gov.au

Seasonal snow and ice cover periodically block sunlight from reaching the ecosystems below. The effect of this on total annual light available to these ecosystems depends critically on the timing of ice cover within the annual solar cycle. At high latitudes sunlight is strongly seasonal, and ice-free days around the summer solstice receive orders of magnitude more light than those in winter. In many polar areas early melt will bring the date of ice loss closer to mid-summer, causing an exponential increase in the annual amount of sunlight reaching the underlying ecosystems. This is likely to drive ecological tipping points in which primary producers (plants and algae) flourish and outcompete dark-adapted communities. We demonstrate this principle on Antarctic shallow sea-bed ecosystems, which our data suggest are sensitive to small changes in the timing of sea-ice loss. Macro-algae (seaweeds) respond to light thresholds that are easily exceeded by a slight reduction in sea-ice duration. Earlier sea-ice loss is likely to cause extensive regime shifts in which endemic shallow-water invertebrate communities are replaced by algae, reducing coastal biodiversity and fundamentally changing ecosystem functioning. Our modeling shows that recent changes in ice and snow cover have already transformed annual light budgets in large areas of the Arctic and Antarctic, and both aquatic and terrestrial ecosystems are likely to experience further significant change in light. The interaction between ice loss and solar irradiance renders polar ecosystems acutely vulnerable to abrupt ecosystem change, as light-driven tipping points are readily breached by relatively slight shifts in the timing of snow and ice loss.


Modeling CO2 dynamics in first-year sea ice

Sebastien Moreau, Martin Vancoppenolle, Jean-Louis Tison, Bruno Delille, Hugues Goosse

Corresponding author: Sebastien Moreau

Corresponding author e-mail: sebastien.moreau@hotmail.com

Sea ice may be an active source or sink for climatically significant gases such as CO2 and CH4. The dynamics of these biogeochemically active gases within sea ice are still not well understood. Modeling can help to identify and calibrate the physical and biogeochemical processes that affect gas production, consumption, diffusion and transport within sea ice. In this study, we aim at constraining the dynamics of gases within sea ice using observation data and a one-dimensional halo-thermodynamic sea-ice model, including gas physics and biogeochemistry. The incorporation and transport of total dissolved inorganic carbon (DIC) and total alkalinity (TA) within sea ice, as well as its rejection via gas-enriched brine drainage to the ocean, are modeled following fluid transport equations through sea ice. Gases are incorporated into gas bubbles when their concentration is above saturation. These bubbles rise to the ice surface due to their own buoyancy when the brine network is connected. CO2 fluxes between sea ice and the atmosphere are formulated according to snow depth, wind speed and the differential partial pressure of CO2 between sea-ice brines and the atmosphere. The consumption and release of CO2 by primary production and respiration are modeled according to the Redfield and others (1963) ratio. Finally, the precipitation and dissolution of ikaite (CaCO3), which influences CO2 concentration within sea-ice, are modeled. Two simulations corresponding to a case study covering the seasonal growth of first-year ice at Point Barrow, Alaska, and to an experimental study at Hamburg, INTERICE IV, where sea ice was grown under controlled conditions, were run. How specific physical and biogeochemical processes affect CO2 dynamics and hence the carbon cycle within sea ice is discussed.


Using stable isotopes to unravel the role of sea-ice in the methane cycle

Célia Julia Sapart, Jiayun Zhou, Helge Niemann, Thomas Röckmann, Bruno Delille, Carina Van der Veen, Jean-Louis Tison

Corresponding author: Célia Julia Sapart

Corresponding author e-mail: c.j.sapart@uu.nl

Methane (CH4) plays an important role in the Earth’s climate system. The atmospheric CH4 concentration has increased in concert with industrialization, but since the mid-1980s the CH4 growth rate decreased to reach a near-zero level in 2000 and started to increase again from 2007 onwards. However, to date, the underlying variations in sources and/or sinks that cause these variations are not well understood. To predict future climate, it is essential to unravel the processes controlling the CH4 cycle, especially in the Arctic regions, which are highly vulnerable to climate change and contain large CH4 reservoirs. Recently, an unexpected CH4 excess has been reported above Arctic sea-ice showing that sea-ice might play a significant role in the CH4 cycle. Nonetheless, the nature of the process leading to CH4 production in or nearby sea ice has not yet been identified. We applied a new multi-proxy approach merging atmospheric chemistry, glaciology and biogeochemistry to understand and quantify the processes responsible for the CH4 excess above sea ice. We performed CH4 isotope (δ13C and δD) analyses on sea-ice samples, as well as microbial (lipid biomarkers) and geochemical measurements, to determine the possible pathways involved in CH4 production and removal in or nearby sea ice. We will present results from sea-ice samples drilled above the shallow shelf in Barrow (Alaska) from January to June 2009, as well as above deep Southern Ocean locations in 2013. These results allow investigation of the seasonality and spatial variability in methane formation and removal pathways associated with the methane enclosed in sea ice.


Role of sea ice in the biogeochemical cycling of mercury in the Arctic Ocean

Sarah Beattie, Amanda Chaulk, Debbie Armstrong, Feiyue Wang

Corresponding author: Sarah Beattie

Corresponding author e-mail: sabeattie23@gmail.com

Mercury (Hg) is one of the primary contaminants of concern in the Arctic marine ecosystem. Different from Hg cycling in other oceans, Hg transport across the air–ocean interface in the Arctic Ocean is complicated by the presence and dynamics of seasonal to perennial sea-ice cover. In recent years the Arctic environment has undergone drastic changes, the most notable being the significant decrease in summer sea-ice extent. This suggests a shift from a multi-year (MYI) to a first-year sea-ice (FYI) regime, yet the effects of this on the transport, transformation and ultimate environmental fate of Hg within the Arctic Ocean system have yet to be fully understood. The evolution of MYI through the summer melt season produces an ice type that drastically contrasts first-year sea ice both physically and chemically. The objective of this work was to investigate the transport, distribution and transformation of Hg within both FYI and MYI. A study performed at the Sea Ice Environmental Research Facility (SERF) on artificial FYI revealed that particulate-bound Hg is incorporated into the sea-ice matrix upon initial formation, and is the dominant Hg species in FYI. Based on our recent studies in the Beaufort Sea of the Arctic Ocean, salinity and chlorophyll a concentrations were shown to have the most influence on total Hg distribution within first-year and multi-year ice, respectively. Transport within FYI appeared to be controlled by brine dynamics, whereas in MYI particulate dynamics may be the more dominant transport process. Analysis of one multi-year ice core showed that the concentration of methylmercury (MeHg) was low throughout the core, suggesting MYI is not an important MeHg source to the Arctic Ocean. The annual flux of Hg into the Arctic Ocean via multi-year sea-ice ablation was estimated to investigate overall Hg loading into the system, and its considerable magnitude suggests that melting MYI is a notable seasonal source of Hg into this system.


On sea-ice dynamical regimes in the Arctic

Jennifer V. Lukovich, Jennifer K. Hutchings

Corresponding author: Jennifer V. Lukovich

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

Central to an understanding of the evolution in sea-ice characteristics in response to climate change is an understanding of sea-ice dynamics and attendant implications for ocean to atmosphere heat and moisture exchange associated with the development of leads and polynyas, freshwater transport associated with changes in ice drift, and ice drift response to relative changes in ice age, thickness and regional atmospheric circulation. In this study we investigate regional differences in ice dynamics in the Arctic using high-frequency buoy and drifter IABP data from 2006 to 2012. Examined in particular are scaling laws derived from ice beacon velocity and dispersion in the FYI, MYI and seasonal ice zone regimes on weekly and seasonal (winter and summer) timescales. We create spatial rate maps to identify whether distinct dynamical regimes can be identified with differing scaling properties, determined from absolute dispersion statistics and spectral characteristics of ice drifter data. In addition, we explore the nature of correspondence between dynamical regimes and ice type characterizations and seasonality. The results from this analysis provide an alternative characterization to changes in sea ice based on dynamics rather than concentration and thickness, and thus insight into, and improved understanding of, the connections between sea-ice drift, deformation and morphology.


Influence of the Gulf Stream on the Barents Sea ice retreat and Eurasian coldness

Kazutoshi Sato, Jun Inoue

Corresponding author: Kazutoshi Sato

Corresponding author e-mail: sato.kazutoshi@nipr.ac.jp

Abnormal winter sea-ice retreat over the Barents Sea has been considered as a leading clue to the recent mid-latitude severe winters. The Barents Sea is considered as a hot spot for the rapid Arctic climate change due to the intense air–sea interaction induced by the sea-ice decrease; however, the underlying mechanisms remain uncertain, in particular the causal relation of sea-ice retreat and atmospheric forcing and response. To understand this causality, we selected typical cases, defined as averaged warm and averaged cold years of December using the NCEP Climate Forecast System Reanalysis (CFSR). The composite analysis revealed that anticyclonic anomaly is obvious over northwestern Eurasia. The western Barents Sea and Svalbard locate at the strong pressure gradient zone, prevailing southerly winds. Over the Barents Sea, the difference in daily mean air temperature between warm and cold winters is more than 10°C, suggesting that warm advection prevails during warm years. Therefore, during warm years, decrease in sea-ice cover is induced by southerly warm advection. The positive anomalies of precipitation from the southeast of Greenland to the Barents Sea and negative anomalies of them from the Nordic Sea to western Eurasia means the poleward shift of cyclone tracks, suggesting that the moisture transport is also changed poleward. Because the cyclones tend to shift poleward in less sea-ice year over the Barents Sea, it is natural that the snow depth over the sea ice near Fram Strait shows a positive anomaly during warm winters. Here we show that the poleward shift of sea surface temperature (SST) over the Gulf Stream, where is situated upstream from the Barents Sea, modifies the horizontal distribution of tropospheric condensational heating resulting from the change in convection over the warm current, likely acting as a bridge to the Barents Sea by forcing planetary waves. This remote atmospheric response modifies cyclone tracks poleward, resulting in anomalous warm advection over the Barents Sea sector.


In situ measurements of the dielectric permittivity of Arctic landfast sea ice in the kHz and GHz range

Megan O’Sadnick, Hiroyuki Wakabayashi, Hajo Eicken, Malcolm Ingham

Corresponding author: Megan O’Sadnick

Corresponding author e-mail: meosadnick@alaska.edu

The sea-ice complex permittivity is controlled by ice temperature, salinity and microstructure. Hence, in situ permittivity measurements may help track the seasonal evolution of key ice properties. Such data are lacking, but are of use for subsurface geophysical sounding and microwave remote sensing of sea ice. In spring 2013, ice permittivity was measured at frequencies of 10 Hz to 95 kHz and 4 to 6 GHz in landfast sea ice at Barrow, AK. We examine high- and low-frequency permittivities to gain insights into the property evolution of the upper ice layers during spring warming. Measurements in the GHz range can also help tie ice property evolution to active and passive microwave remote-sensing signatures of sea ice. The low-frequency complex dielectric permittivity was measured with four electrode strings in a 1 m × 1 m array, frozen into the ice in January 2013. In March, May and June, impedance and phase were measured between combinations of electrodes in the frequency range 10 Hz to 95 kHz. The imaginary and real component of the permittivity increase as the ice warms with the greatest increase in the lowest frequency of the real component of the permittivity. This finding is attributed to space charge polarization, itself the result of changes in microstructure. Data have been fitted to a broadband regression model developed by Buchanan and others (2011). Ice temperature, salinity and microstructural data obtained from X-ray tomography are analyzed in relation to the permittivity data. At the same site, the high-frequency dielectric permittivity was measured in March and June with a vector network analyzer, horn antenna and dielectric lens. Reflected power from the observation target was found by eliminating unwanted system reflection using the rotating electric field vector method. Reflected power was calibrated using the aluminum plate’s reflection in the frequency range 4 to 6 GHz. The objective of this measurement is to understand the C-band backscatter coefficient of sea ice by considering the role of sea-ice dielectric permittivity. Our analysis shows dielectric permittivity to be strongly related to ice surface salinity. Permittivities derived from the model by Vant and others (1978) were compared with measured data.


Sea-ice CO2 flux in the Southern Ocean during mid-winter and early spring

Daiki Nomura, Bruno Delille, Gerhard Dieckmann, Jean-Louis Tison, Klaus Meiners, Mats Granskog, Takeshi Tamura

Corresponding author: Daiki Nomura

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

There seems little doubt that sea ice is permeable to CO2 and other gases although air–sea-ice gas flux is more or less inhibited at a brine volume fraction of less than 5% representing the threshold for fluid permeability of sea ice. Generally, air–sea-ice CO2 flux is at its minimum in winter due to low sea-ice temperatures and consequently reduced permeability despite the fact the partial pressure of CO2 in sea ice is usually high at that time and sea ice has therefore the potential to release CO2 to the atmosphere. Here, we present first evidence that snow-laden Antarctic sea ice can act as source for atmospheric CO2 even during mid-winter and early spring. During a mid-winter cruise to the Weddell Sea (AWECS, 2013) and an early spring cruise off east Antarctica (SIPEX-2, 2012), due to thick insulating snow covers, the bottom of the snow and the surface of the sea ice were relatively warm (>-10°C) even though air temperature was sometimes below -30°C. In addition, in both areas, sea ice was characterized by high bulk salinities, resulting in brine volume fractions that are generally higher than 5%. Automatic ‘open-closed’ chamber measurements indicated positive CO2 fluxes of up to +2.5 mmol C m–2 d–1, illustrating that sea ice acted as a source of atmospheric CO2. Higher fluxes were measured at bare-ice surfaces after removing the snow. However, generally low snow densities (mean: 339 kg m–3), indicating a permeable snow cover, facilitated degassing of CO2 at the snow–air interface. Our results therefore suggest that even in the winter and early spring, Antarctic sea ice can act as CO2 source for the atmosphere, particularly in areas with a thick insulating snow cover.


Bromoform emission over the Antarctic sea ice

Daiki Nomura, Takeshi Tamura, Atsushi Ooki, Gerhard Dieckmann, Ellen Damm, Klaus Meiners

Corresponding author: Daiki Nomura

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

Bromoform is one of the volatile organic compounds emitted from the ocean surface to the atmosphere, and it is believed to affect ozone depletion in the atmosphere through photochemical reactions. While estimates of air–sea flux of bromoform are well examined in open ocean areas, fluxes have rarely been estimated in ice-covered seas, and so far no observations have been made to evaluate the bromoform flux between the sea-ice surface and atmosphere. Here, we present the first direct measurements of the air–sea-ice bromoform flux obtained from first-year sea ice off East Antarctica. Measurements were made in early austral spring (September to November 2012) as part of the Sea Ice Physics and Ecosystem Experiment II (SIPEX-2). Vertical profiles of bromoform concentrations in snow and sea ice indicated that high concentrations were mainly found in the bottom of the snow and the surface layers of the sea ice (including slush and brine) ranging from 281 to 1360 pM. Sea-ice–atmosphere bromoform fluxes measured by the chamber method ranged from +0.3 to +7.5 nmol CHBr3 m–2 d–1 (positive value indicates the emission of the bromoform from the ice surface to the atmosphere), and flux values increased with increasing bromoform concentrations at the surface layers. The mean flux estimate (+2.4 nmol CHBr3 m–2 d–1) obtained in this study was consistent with the flux estimate for the ice-free part of the Southern Ocean (+2.6 nmol CHBr3 m–2 d–1; Quack and Wallace, 2003). Our results suggest that the bromoform emitted from the sea-ice surface to the atmosphere may account for an important fraction of the global bromine budget.


Interactions between snow and melt ponds in sea-ice models

Olivier Lecomte, Thierry Fichefet, Martin Vancoppenolle, François Massonnet

Corresponding author: Olivier Lecomte

Corresponding author e-mail: olivier.lecomte@uclouvain.be

Snow cover on sea ice at the end of the winter persists into the early part of the sea-ice melt season, and the spatial distribution of snow affects surface topography, the distribution of initial melt pond formation and its subsequent evolution, especially on first-year ice (FYI). After the initial formation of melt ponds, the low albedo of the ponds compared with snow or bare ice causes the ponds to preferentially absorb solar radiation and therefore further affects surface ice melt. A physically based melt pond model was coupled to the thermodynamic-dynamic Louvain-la-Neuve Sea-Ice Model (LIM, version 3), which recently includes a representation of snow properties and processes. In the new snow scheme, snow is represented in multiple layers with varying thermo–physical properties, and simple parameterizations for blowing snow and fresh water refreezing into the snow were implemented. Several simulations were performed using the combined snow and melt pond configuration to study the impacts of the processes described above on the Arctic sea-ice melt pond fractions. Preliminary results lead to two expected but uncorroborated model behaviors. In the simulations, blowing snow tends to redistribute the accumulation of snow from relatively young ice to the older-deformed ice or causes losses into leads, thus allowing larger pond fractions on FYI. Refreezing of water in the snow on the other hand curtails the amount of meltwater available to feed melt ponds at their onset of formation, but has limited or no impact on the pond fractions at the middle of the melt season when snow has almost entirely melted away.


On the formulation of snow thermal conductivity in large-scale sea-ice models

Olivier Lecomte, Thierry Fichefet, Martin Vancoppenolle, Florent Domine, François Massonnet, Pierre Mathiot, Samuel Morin, Pierre-Yves Barriat

Corresponding author: Olivier Lecomte

Corresponding author e-mail: olivier.lecomte@uclouvain.be

An assessment of the performance of a state-of-the-art large-scale coupled sea-ice–ocean model, including a new snow multilayer thermodynamic scheme, is performed. Four 29 year long simulations are compared against each other and against sea-ice thickness and extent observations. Each simulation uses a separate parameterization for snow thermo–physical properties. The first simulation uses a constant thermal conductivity and prescribed density profiles. The second and third parameterizations use typical power-law relationships linking thermal conductivity directly to density (prescribed as in the first simulation). The fourth parameterization is newly developed and consists of a set of two linear equations relating the snow thermal conductivity and density to the mean seasonal wind speed. Results show that simulation 1 leads to a significant overestimation of the sea-ice thickness due to overestimated thermal conductivity, particularly in the Northern Hemisphere. Parameterizations 2 and 4 lead to a realistic simulation of the Arctic sea-ice mean state. Simulation 3 results in the underestimation of the sea-ice basal growth in both hemispheres, but is partly compensated by lateral growth and snow ice formation in the Southern Hemisphere. Finally, parameterization 4 improves the simulated snow depth distributions by including snow packing by wind, and shows potential for being used in future works. The intercomparison of all simulations suggests that the sea-ice model is more sensitive to the snow representation in the Arctic than it is in the Southern Ocean, where the sea-ice thickness is not driven by temperature profiles in the snow.


Pan-Arctic sea-ice lead detection from MODIS thermal infrared imagery

Sascha Willmes, Günther Heinemann

Corresponding author: Sascha Willmes

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

Polynyas and leads are key surface elements of the wintertime Arctic sea-ice cover. They play a crucial role in surface heat loss, potential ice formation and consequently in the seasonal sea-ice budget. While polynyas are generally sufficiently large to be observationally monitored with passive microwave satellite sensors, the monitoring of narrow leads requires the use of data with a finer spatial resolution. We present daily pan-Arctic maps of the spatial distribution of sea-ice leads. The product is derived from daily composites that we compute from sea-ice surface temperatures as measured by the Moderate Resolution Imaging Spectroradiometer (MODIS). A fuzzy cloud artifact filter is implemented to mitigate shortcomings in the MODIS cloud mask and the associated potential mis-classification of leads. The daily lead product can be used to deduct the structure and dynamics of wintertime sea-ice leads, to assess seasonal divergence patterns of the Arctic Ocean and to validate lead detection for sea-ice thickness retrievals.


Is sea salt in ice cores a proxy of past sea-ice extent?

James Levine, Eric Wolff, Anna Jones, Markus Frey, Xin Yang

Corresponding author: Eric Wolff

Corresponding author e-mail: ew428@cam.ac.uk

A number of marine, coastal and ice-core proxies have been used to try to assess the past extent of sea ice. Sea salt has been proposed as a proxy for past ice extent, at least in the Southern Ocean. The idea is that the sea-ice surface itself holds a source of sea salt that is stronger than the source from the open ocean it replaces. That a sea-ice source exists is apparent from observations of the ratio of sulphate to sodium in coastal aerosol and snow samples. While the idea behind using sea salt as a proxy is attractive, and leads to sensible inferences, many doubts remain. Firstly the exact nature of the source remains uncertain, and secondly it is not clear if ice extent, as opposed to changes in atmospheric transport and lifetime, would dominate variability in the ice-core record of sea salt. Here we use a model of atmospheric transport and chemistry (p-TOMCAT) to assess the consequences of a sea-ice source, focussing particularly on a source that has been proposed to arise from the sublimation of salty blowing snow. We hope to be able to report some new observations from a winter cruise, which will allow us to comment on the likelihood that blowing snow does pose a significant source. However, for now, we have implemented a scheme based on existing data. The model has been run with seasonally and interannually varying sea-ice extent and meteorology for the Antarctic, tracking, at different ice-core sites, the concentration that arises from the open ocean and sea-ice sources. The model, after tuning, is able to reproduce the magnitude and seasonal cycle of sea salt at a range of sites globally. By varying each component separately we find that meteorological variability dominates the sea-salt concentration over Antarctica between years. This suggests that sea salt cannot be used as a sea-ice proxy on interannual timescales. However, extending the study back to much larger sea-ice extents, such as at the last glacial maximum (LGM), we find a strong increase in concentration at ice-core sites when ice extent increases: this is the first process-based model in which this observed behaviour is reproduced. This suggests that we could infer ice extents, averaged over century scales, if we could assess the role of changing climate (to be attempted at a later stage). The increase in modelled sea-salt concentration tails off sharply as ice approaches the LGM extent, so that the sensitivity of the proxy is greater at lower ice extents.


Deriving past sea-ice extent from proxy data: the PAGES sea-ice proxy working group

Eric Wolff, Anne de Vernal, Rainer Gersonde, Hugues Goosse, Marit-Solveig Seidenkrantz

Corresponding author: Eric Wolff

Corresponding author e-mail: ew428@cam.ac.uk

The rapid reduction in Arctic sea ice over recent decades makes it especially important to obtain long time series that can allow us to put recent events in context. The satellite dataset reaches only 40 years into the past, while syntheses using ship and aircraft observations have been taken back to the late 19th century. Before that, we are reliant almost entirely on proxy data from marine sediments, ice cores and coastal material, each providing evidence of past sea-ice presence or absence. However, there is generally a complex chain of processes that connect ice extent or presence to the measured proxy. The basis for each proxy depends on physical and biogeochemical processes that demand modern process studies, and each proxy may be applicable only under defined conditions. To produce spatial reconstructions requires the use of multiple proxies. Because each proxy is dependent on a different characteristic of the ice, different proxies may appear to give incompatible results. Therefore, to produce products that can be used by modellers, and that are compatible with datasets based on modern methods, requires considerable care and interdisciplinary study. The IGBP project PAGES (Past Global Changes) therefore set up a sea-ice proxy (SIP) working group, whose work is represented on this poster. The poster will first describe many of the proxies that have been used for sea ice, and give a brief summary of their advantages and limitations. This will include both biological and biomarker proxies in marine sediments, and chemical proxies in ice cores. It will then consider some of the principles that can be used to make multi-proxy reconstructions of sea ice. Finally the prospects for such reconstruction on different timeslices, for both the Southern and Northern Hemispheres, will be discussed (and maybe the first such reconstructions will be shown). The work of the SIP working group continues with a third workshop in Bremerhaven in 2014.


Constraining projections of summer Arctic sea ice

François Massonnet, Thierry Fichefet, Hugues Goosse, Cecilia Bitz, Marika Holland, Gwenaëlle Philippon, Pierre-Yves Barriat

Corresponding author: François Massonnet

Corresponding author e-mail: francois.massonnet@uclouvain.be

We examine the recent (1979–2010) and future (2011–2100) characteristics of the summer Arctic sea-ice cover as simulated by 37 Earth system and general circulation models from the Coupled Model Intercomparison Project, phase 5 (CMIP5). As was the case with CMIP3, a large intermodel spread persists in the simulated summer sea-ice losses over the 21st century for a given forcing scenario. The baseline sea-ice extent, thickness distribution and volume characteristics, as well as the multidecadal trend in September sea-ice extent (SSIE) of each CMIP5 model are discussed as potential constraints on the SSIE projections. Our results suggest first that the future changes in SSIE with respect to the model baseline state are related in a complicated manner to this baseline state, due to the large diversity of the CMIP5 population: at a given time, some models are in an ice-free state while others are still on the track of ice loss. However, in phase plane plots (that do not consider the time as an independent variable), we show that the transition towards ice-free conditions is actually occurring in a very similar manner for all models. We also find that the year at which SSIE drops below a certain threshold is likely to be constrained by the present-day sea-ice properties. In a second step, using several adequate sea-ice metrics, we effectively reduce the uncertainty as to when the Arctic could become nearly ice-free in summertime, the interval (2041, 2060) being our best estimate for a high climate forcing scenario.


Tidal effects on ice-shelf melting and freshening in the Amundsen Sea

Robin Robertson

Corresponding author: Robin Robertson

Corresponding author e-mail: r.robertson@adfa.edu.au

The ice shelves in the Amundsen and Bellingshausen Seas contribute 47% of the total meltwater for the Antarctic, despite comprising only 8% of the area. Tides have been found to contribute to the ice-shelf melting for various regions. Although ice-shelf melt and ocean circulation in ocean cavities below ice shelves is primarily density driven and controlled by the ocean temperature above freezing and topography, tides play a role and their role is complex. In the Amundsen Sea, tidal impacts were dependent on the location of the ice-shelf front with respect to the M2 effective critical latitude. The critical latitude is the latitude where the tidal frequency equals the inertial frequency. The effective critical latitude is where the tidal frequency equals the inertial frequency adjusted by relative vorticity, such as that associated with a wind-driven gyre. For ice shelves located equatorward of the M2 effective critical latitude, tides increased both mixing in front of and under the ice shelf and flow into the ice-shelf cavities by as much as 50%, despite weak tides compared with the mean flows. Tides also increased melting for these ice shelves by 1–3.5 m a–1, a 50% increase for the Dotson Ice Shelf and 25% for the Pine Island Ice Shelf. These enhancements were not a result of tidal residual flows, but instead originated from resonant effects, increases in the baroclinity of the velocities, and higher mixing, all of which are associated with critical latitude effects on internal tides. For ice shelves located poleward of the effective critical latitude, tides very slightly retarded flow into the cavity and slightly reduced melting. Additionally, tides increased the freshening of the deep waters of the Amundsen Sea along with various other impacts.


Decadal trends in Antarctic sea-ice extent and atmospheric circulation

James Renwick, Sam Dean

Corresponding author: James Renwick

Corresponding author e-mail: james.renwick@vuw.ac.nz

Total Antarctic sea-ice extent has been gradually increasing over the past few decades, even as the globe has warmed. The gradual increasing trend is a small difference between large losses near the Antarctic Peninsula and large gains in the Ross Sea region. The patterns of sea-ice change observed since the late 1970s match closely with trends in near-surface winds, suggesting that atmospheric forcing has played a key role in determining the overall trend in sea-ice extent. Moreover, the patterns of change appear to be largely associated with natural variability in the circulation, and do not project strongly onto known patterns of human-induced change (e.g. the positive trend in the Southern Annular Mode). This presentation will explore decadal variability in sea-ice and circulation trends since 1979 and will discuss the level of coupling between atmospheric circulation and sea-ice extent over that time. It will also discuss the relative roles of anthropogenically forced and natural variations in shaping observed trends in the Southern Hemisphere extra-tropical atmospheric circulation and in Antarctic sea-ice extent.


Helicopter-borne observation with portable microwave radiometer in the Southern Ocean and the Sea of Okhotsk

Takeshi Tamura, Kay Ohshima, Jan Lieser, Takenobu Toyota, Kazutaka Tateyama, Daiki Nomura, Kazuki Nakata, Alexander Fraser, Peter Jansen, Kym Newbery, Robert Massom, Shuki Ushio

Corresponding author: Takeshi Tamura

Corresponding author e-mail: tamura.takeshi@nipr.ac.jp

It has been recently recognized that sea-ice production in the polar regions is controlled by the thin sea-ice area with thickness of less than 0.2 m. Spatial distribution of thin ice area and its variability are important information to better understand the reduction of the sea-ice-covered region in a changing climate environment. We have developed a thin ice thickness algorithm for satellite passive microwave data of the Advanced Microwave Scanning Radiometer-EOS (AMSR-E) and Special Sensor Microwave Imager (SSM/I). Although the microwave skin depth of bare sea ice is about several cm at most, microwave brightness temperatures correlate with the surface salinity (brine volume fraction), which is sensitive to thin ice thickness. Here, we present in situ observations using a helicopter-borne portable passive microwave radiometer that has the same specifications as the satellite AMSR-E and AMSR-II sensors (36 GHz vertical and horizontal channels), to validate and improve our thin ice thickness algorithm. This study estimates the relationship between the microwave brightness temperatures (both satellite and helicopter-borne portable sensors) and in situ observations of sea-ice thickness. We present data from two airborne missions, one in early austral spring 2012 during the Sea Ice Physics and Ecosystem eXperiment (SIPEX-2) of the Australian Antarctic Program in East Antarctica, and one from the Sea of Okhotsk in mid-winter 2009. These microwave data are compared with the satellite AMSR-E and AMSR-II data and ice thickness estimated from Moderate Resolution Imaging Spectroradiometer (MODIS) data, helicopter-borne IR sensor data, and ship-borne downward-looking camera data. High-resolution airborne microwave brightness temperatures show good agreement with low AMSR-E and AMSR-II brightness temperatures, despite the significant resolution mismatch. In the thin ice region, the polarization ratio of 36 GHz vertical and horizontal temperatures (PR-36) is found to be well correlated with ice thickness, supporting the validity of the AMSR-E thin ice algorithm which was developed previously by our group. We also discuss the microwave characteristics of fast versus pack ice, with a view to improving a satellite fast-ice detection algorithm.


Observation and modeling the iceberg drift in the eastern Weddell Sea iceberg alley

Wolfgang Dierking, Thomas Rackow, Christine Wesche

Corresponding author: Christine Wesche

Corresponding author e-mail: christine.wesche@awi.de

The continental shelf of the eastern Weddell Sea is known as ‘iceberg alley’. Icebergs that calve at the coast of East Antarctica drift with the coastal current along the Fimbul Ice Shelf towards the Weddell Sea. In the beginning of 2006, three large icebergs passed the Ekström Ice Shelf close to the sea-ice edge. Using a series of Envisat Wide Swath images, they were monitored for nearly 1 year to investigate the influence of wind, ocean currents, sea ice and the iceberg draft on their drift patterns and to validate corresponding model simulations. The iceberg drift model is driven by the Finite Elements Sea-ice Ocean Model (FESOM). It was run with a temporal resolution of 6 hours and a spatial resolution of 10 km close to the coast and 30 km further offshore. Besides comparison between iceberg drift observations and computer simulations for improving the model, the observation of icebergs in this region is of great importance for the logistic department at Alfred Wegener Institute because of the supply of the German overwintering station Neumayer III. The drift of the three icebergs is modeled with different input configurations and parameters. Events with large differences between the modeled and the observed positions are investigated in detail. The goal of this study is to improve the accuracy of the iceberg drift model in the region of the Weddell Sea ‘drift alley’ and thus to be able to forecast iceberg positions as support for marine traffic in this region.


Predictability of the Barents sea-ice extent in early winter

Takuya Nakanowatari, Kazutoshi Sato, Jun Inoue

Corresponding author: Takuya Nakanowatari

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

The sea-ice variability of the Barents Sea in early winter (December) and its resultant dynamical atmospheric response are considered to be the triggers of a dynamic atmospheric response with severe consequences for weather conditions over the Eurasia continent including Japan. In this study, canonical correlation analysis (CCA), a linear statistical method designed to find correlated patterns between predictor and predictand fields, is applied to the Barents sea-ice concentration (SIC). The cross-validation technique is used to estimate the levels and sources of forecast skill for SIC for the 30 year time period of 1980–2009. Sea surface temperature (SST), subsurface temperature at 200 m depth (Tsub), surface air temperature, surface wind in the Barents Sea and sea-level pressure (SLP) and geopotential height at 500 hPa (Z500) in the Northern Hemisphere are used as predictor fields in an attempt to maximize the strength of the predictive relationships. The highest skill for a single-predictor model is from 1 year leading Tsub and 1 month leading meridional wind (Vsfc). Tsub explains 30% of SIC variability mainly on the eastern side on a decadal timescale. CCA diagnostics suggest that change in the subpolar North Atlantic gyre generates a temperature anomaly and the resultant thermal anomaly is advected to the Barents Sea by the Norwegian Atlantic Slope Current with 3 to 4 years. Vsfc explains 22% of SIC variability mainly on the western side on year-to-year timescale. CCA diagnostics indicate that the Vsfc is related to the teleconnection pattern from the North Atlantic. Thus, our study suggests that both atmospheric and oceanic remote effects lead to more accurate forecasting of the SIC than that using only local oceanic heat condition in the Barents Sea.


Simulations of Southern Ocean sea ice under future climate scenarios

Roger Stevens, Petra Heil

Corresponding author: Roger Stevens

Corresponding author e-mail: roger.stevens@utas.edu.au

There is evidence from several sources that the global climate is changing and that this is most likely in response to increased concentrations of atmospheric greenhouse gasses. In the polar Northern Hemisphere, climate change is most apparent in the dramatic reduction in Arctic Ocean sea ice over the last 30 years. During the same time, the overall Southern Ocean sea ice has increased its extent slightly, with some regions exhibiting an increase (Ross Sea) and others a decrease (Bellingshausen and Amundsen Seas). An important question is how will Southern Ocean sea ice respond to possible projected climates? This paper uses a moderately high resolution numerical model to simulate Southern Ocean sea ice in a number of possible future climates including one based on the IPCC FAR 2100 A1B scenario projections. The simulated trend under these warm climates is for a reduction in sea-ice extent/area, volume and concentration. In summer the places where ice remains are limited to the southern Weddell Sea and the upstream side of some of the coastal protrusions along the East Antarctic coast. Implications for the annual sea-ice cycle are investigated along with estimates of sea-ice persistence.


Modeling the vertical fine structure of ice algal production

Pedro Duarte, Phillip Assmy, Haakon Hop, Gunnar Spreen, Sebastian Gerland, Stephen Hudson

Corresponding author: Pedro Duarte

Corresponding author e-mail: Pedro.Duarte@npolar.no

The relative importance of ice algae located in the interior layer of drifting sea ice is evaluated relative to the overall ice algal production. Ice algae are reported to contribute 5–20% of overall primary production in Arctic shelf areas. The spring ice algal bloom may trigger the reproduction of some Arctic zooplankton species, such as Calanus glacialis. Combined stable isotope and fatty acid analyses confirm the importance of ice algae as a food source for ice-associated fauna. Thus, it is important to estimate the future impacts of global warming on the contribution of ice algae to Arctic primary production. Primary production models, describing the relationships between ice algal physiology and population dynamics with environmental forcing and trophic interactions involving bacteria and grazers, can be applied to quantify such impacts. One important aspect in these models is how to represent the vertical distribution of ice algae in sea ice. In most models, only the bottom skeletal ice layer is considered where most of the algal biomass tends to be concentrated. However, since ice algae are also present along the entire ice column, this may lead to the underestimation of ice algal production. Some empirical data and model results suggest that ice algae located in the surface and interior layers may be kept at lower concentrations, in spite of high growth rates, due to grazing by micro- and meiofauna. Simulating the vertical distribution of ice algae within the sea ice poses technical challenges with regard to their heterogeneous distribution. Whilst this may be feasible at small spatial scales, it is challenging in pan-Arctic models, where ice algal and hydrodynamic sub-models need to be coupled. An ice algal mathematical model will be used here to resolve the vertical fine structure of sea ice with ice algae, and results are compared between simulations with ice algae located only at the bottom skeletal ice layer and those where ice algae are distributed across the ice column. Minimum requirements for an ice algal model applicable at regional spatial scales will be assessed.


Brine dynamics and tracer transport in sea ice

Philipp J. Griewank, Dirk Notz, Ronnie N. Glud

Corresponding author: Philipp J. Griewank

Corresponding author e-mail: philipp.griewank@zmaw.de

Brine movements transport salt and other dissolved tracers in sea ice. We examine the influence of brine fluxes on the chemical composition of brine in sea ice using the 1-D model SAMSIM. This model was previously used successfully to simulate the salinity evolution of sea ice and has now been extended to look at other chemical components. We specifically focus on oxygen, as anoxic conditions in sea ice have been measured by field and laboratory measurements. Anoxia in sea ice could indicate ongoing anaerobic microbial processing like denitrification, representing an unresolved nutrient sink in sea ice. We use our model to quantify how much oxygen depletion occurs via gravity drainage during growth and via dilution during melt. From the remaining oxygen we can determine how much biologic activity is necessary to reach anoxic conditions. We also quantify the impact of bubble formation on the oxygen evolution in sea ice.


Iron and macro-nutrient concentrations in sea ice and their impact on the nutritional status of surface waters in the southern Okhotsk Sea

Naoya Kanna, Takenobu Toyota, Jun Nishioka

Corresponding author: Naoya Kanna

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

To elucidate the roles of sea ice in biogeochemical cycles in the Sea of Okhotsk, the concentrations of macro-nutrients (NO3+NO2, PO4, SiO2 and NH4) and trace elements (Fe, Al) were measured in samples of sea ice, overlying snow and sea water. The oxygen isotope ratio in the sea ice was used to distinguish between snow ice and seawater-origin ice. Except for NH4, the macro-nutrient concentrations were lower in sea ice than in surface water in the ice-covered area. A linear relationship between salinity and concentrations of NO3+NO2, PO4 and SiO2 in the sea ice indicated that these macro-nutrients originated from sea water. The Fe concentrations in sea ice were variable and several orders of magnitude higher than those in surface water in the ice-covered area. The Fe concentrations in the sea ice were positively correlated with Al concentrations, the suggestion being that the Fe contained in the sea ice originated from lithogenic mineral particles. The annual Fe flux into the surface water from sea-ice melting in the southern Sea of Okhotsk was estimated to be ~740 μmol Fe m–2 a–1. This flux is comparable to the reported annual atmospheric Fe flux (267–929 μmol Fe m–2 a–1) in the western North Pacific. In spring, sea-ice melting may slightly dilute macro-nutrient concentrations but increase Fe concentrations in surface water. These results suggest that sea ice may contribute to phytoplankton growth by release of Fe into the water column and have a large impact on biogeochemical cycles in the Sea of Okhotsk.


Sea-ice research and outreach at the Centre for Earth Observation Science (CEOS), University of Manitoba, Canada

Lucette Barber, Gary Stern, Brian Horton, Feiyue Wang, Tim Papakyriakou, Soeren Rysgaard, David Barber

Corresponding author: Lucette Barber

Corresponding author e-mail: lucette.barber@umanitoba.ca

The Centre for Earth Observation Science (CEOS) is housed in the Clayton H. Riddell Faculty of Environment, Earth and Resources at the University of Manitoba, Winnipeg (Canada). The centre began in 1994 with one faculty member, one half-time technician and two graduate students. CEOS now hosts over 100 faculty, scientists, graduate students and support staff all working on aspects of ocean–sea-ice–atmosphere processes. The centre is heavily invested in the Canadian ArcticNet Networks of Centres of Excellence (NCE) program and the international Arctic Science Partnership (ASP) between Denmark, Greenland and Canada. In addition to research and academics, CEOS also invests time and resources in outreach programs aimed at promoting scientific literacy as it pertains to the Arctic marine system, policy and industrial development challenges of the Arctic. We present an overview of our unique facilities spanning a wide range of laboratory and field sites, including the Sea-ice Environmental Research Facility (SERF) and the Nellie Cournoyea Arctic Research Facility, and the proposed Churchill Marine Observatory (CMO), which would consist of an oil in sea ice mesocosm (OSIM) and an automated Environmental Observatory (EO) in Churchill Manitoba.


Large-scale ground-truth sea-ice-edge proxy from ice charts

Penelope Wagner, Cathleen Geiger, Sean Helfrich

Corresponding author: Cathleen Geiger

Corresponding author e-mail: cgeiger@udel.edu

The sea-ice outer boundary, which defines its extent, is difficult to resolve on a daily basis to high resolution (better than 12 km) due to its amorphous structure. The unclear demarcation of the sea-ice edge boundary is defined largely from different points of view in the science and operational communities. Because of these difficulties, there is currently no established daily large-scale sea-ice-edge ground-truth resource to validate remote-sensing data. As a proxy, we consider here the United States National Ice Center (NIC) ice chart estimates of the ice edge as a possible cross-validation resource. The NIC currently creates two sea-ice-edge products with different cartographic techniques. The first is a daily product to estimate the sea-ice outermost boundary based on current location and drift estimates. This product is created primarily as a navigational aid. The second is a bi-weekly product to estimate sea-ice extent for science applications. In this work, we evaluate the quality and consistency of the daily product (May 2004 to December 2012) to determine its efficacy as a large-scale proxy for sea-ice-edge ground truth. We extract samples of the sea-ice-edge location at one degree intervals around the Southern Ocean from both products and compare them using a hierarchy of statistical tools which include tests for differences in (1) normal distribution, (2) variance, (3) level of independence and (4) central tendency. We use these tools to answer the following question: what is the difference in ice-edge location between daily and bi-weekly charts? This assessment will allow us to determine the source of these differences, as well as their ranking. The results of this study provide a measure of uncertainty for both the science and operational communities when using these charts for climatology, tuning models, forecasts and navigation.


Sea-ice hazards in the Arctic: formation mechanisms, detection and management implications

David Barber, David Babb, Alexander Komarov, Jennifer Lukovich, Gregory McCullough, Klaus Hochheim

Corresponding author: David Barber

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

Sea ice in the Arctic has changed in both extent and thickness over the past several decades. A rather counterintuitive effect of this change is an increase in first-year sea-ice deformation, increased multi-year sea-ice hazards, increased presence of marine glacial ice, and an increase in the speed at which these features circulate. These processes increase hazards to marine operations including shipping, oil and gas exploration, and tourism. We summarize sea-ice-related research conducted within the Arctic Ocean, Beaufort Sea and Baffin Bay with an outlook towards development of an ice management system used by the hydrocarbon industry. We summarize the current decline in sea-ice areal extent and thickness, describe the detection of various sea-ice and glacial ice features both in situ and via RADARSAT-2 analysis, and present an assessment of the relative velocities of these ice hazards. We conclude with an overview of how these data can be used to improve sea-ice management by the Arctic maritime industry.


Production of mycosporine-like amino acids in sea-ice-covered Arctic waters

Ashley Elliott, Christopher J. Mundy, Michel Gosselin, Michel Poulin, Karley Campbell, Feiyue Wang

Corresponding author: Feiyue Wang

Corresponding author e-mail: Feiyue.Wang@umanitoba.ca

Thinning sea ice and an earlier melt combined with the decreases in stratospheric ozone over the Arctic are among some key factors influencing the springtime light environment in the ice-covered marine ecosystem. During the Arctic-ICE (ice-covered ecosystem) program in Allen Bay, Nunavut, Canada, mycosporine-like amino acids (MAAs) were measured in ice algae, under-ice phytoplankton, and ice-melt algae communities from 6 May to 24 June 2011 to give an initial assessment of the spatial and temporal presence of MAAs in this region. MAAs are secondary metabolites produced by some organisms for protection from damaging UV radiation, although little is known about their presence in these Arctic communities. Five major UV-absorbing compounds were consistently identified in the samples taken after 31 May, showing interesting temporal and habitat-specific variations. Based on chromatographic retention times, four of the compounds were tentatively identified as shinorine, palythine, porphyra-334 and palythene. Of particular interest was the dominance of an unknown UV-absorbing compound (with an absorbance maximum at 363 nm) in all water samples associated with melting sea ice, though its structure and origin remained elusive. The presence of MAAs in certain species or communities could provide them a competitive advantage; furthermore, they could be an important indicator of change in UV-stress under a changing light environment in the Arctic system.


Profiling sea ice with an airborne ICESat-2 development system

Ron Kwok, Thorsten Markus

Corresponding author: Ron Kwok

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

The sole instrument on the upcoming ICESat-2 altimetry mission is a micropulse lidar that measures the time-of-flight of individual photons from laser pulses transmitted at 532 nm. Prior to launch, MABEL serves as an airborne implementation for testing and development. In this talk, we provide a first examination of MABEL data acquired on two flights over sea ice in April 2012: one north of the Arctic coast of Greenland, and the other in the East Greenland Sea. The phenomenology of photon distributions in the sea-ice returns is discussed. We describe an approach to locate the surface and estimate its elevation in the distributions is described, and its achievable precision assessed. Retrieved surface elevations over relatively flat leads in the ice cover suggest that precisions of several centimeters are attainable. Comparisons of nearly coincident elevation profiles from MABEL with those acquired by an analog lidar show good agreement. Discrimination of ice and open water, a crucial step in the determination of sea-ice freeboard and the estimation of ice thickness, is facilitated by contrasts in the observed signal/background photon statistics. We evaluate the efficacy of a lead detection procedure.


Buoyancy mechanisms and fate of algal aggregates below Arctic sea ice

Mar Fernández-Méndez, Frank Wenzhöfer, Ilka Peeken, Heidi Louise Sørensen, Ronnie Glud, Antje Boetius

Corresponding author: Mar Fernández-Méndez

Corresponding author e-mail: mmendez@awi.de

Primary production below Arctic sea ice is limited to summer months and is not only restricted by light but also by the stratification of the water column, which constrains nutrient supply for algal growth. Recent observations from the central Arctic in the summer of 2012, when Arctic sea ice declined to a record minimum, point towards an increase in algal aggregate accumulations in melt ponds and below the sea ice, subsequently resulting in increased sedimentation towards the deep-sea floor. Diatoms inhabiting Arctic sea ice are known to form large aggregates and sub-ice filaments, but the role and regulation of these massive microbial aggregations is not yet well understood, and may vary in relation to the fate of the Arctic sea-ice cover. During the RV Polarstern cruise to the ice-covered eastern central basins, we observed a widespread deposition of ice algal biomass to the deep-sea floor. To elucidate the mechanism controlling the formation and export of these aggregates, we investigated the characteristics and sinking behaviour of different kinds of aggregates collected from melt ponds and below the ice. Two different kinds of sea-ice algal aggregates could be identified: round, floating aggregates composed mainly of pennate diatoms, and filamentous strings mainly consisting of the sub-ice algae Melosira arctica. At the end of the productive season (August to September) both types of aggregate could be found in different degradation stages floating, frozen into the ice, in melt ponds or sedimented to the sea floor. Our results suggest that some algal aggregates form as a consequence of nutrient depletion, higher irradiance and lower salinity due to ice melt, whereas other sea-ice algae like Melosira arctica use aggregation as a permanent lifestyle by populating the underside of the ice as hanging filamentous strings. Laboratory experiments revealed that oxygen produced by photosynthesis trapped within the mucous aggregate matrix was responsible for buoyancy control. As oxygen production is mainly regulated by light availability, aggregates will stay afloat only when receiving sufficient light. Aggregation can be an advantage as floating increase their probabilities of being refrozen in newly formed ice and thus might serve as a seeding population for the next spring. This study provides new insights into sea-ice algal ecology that will help to better understand the implications of changing environmental conditions in the Arctic.


Antarctic sea-ice freeboard and estimated thickness from ICESat (2003–09)

Donghui Yi, H. Jay Zwally, John Robbins, Jeremy Harbeck, Serdar Manizade, Nathan Kurtz, Michael Studinger

Corresponding author: Donghui Yi

Corresponding author e-mail: donghui.yi@nasa.gov

ICESat completed 18 observational campaigns during its lifetime from 2003 to 2009. Data from all of the 18 campaign periods are used in this study. Most of the operational periods were between 34 and 38 days long. Because of laser failure and orbit transition from 8 day to 91 day orbit, there were four periods lasting 57, 16, 23 and 12 days. It has been demonstrated that ICESat is able to measure sea-ice freeboard height. With nominal densities of snow, water and sea ice, combined with snow depth data from AMSR-E passive microwave data over the Southern Ocean, sea-ice thickness is derived from the freeboard. Combined with AMSR ice concentration, sea-ice area and volume are also calculated. Sea-ice freeboard and thickness distributions show clear seasonal variations that reflect the yearly cycle of the growth and decay of the Antarctic pack ice. Although Arctic sea ice shrinks in both thickness and area in recent years, we found no significant trend of thickness or area for the Antarctic sea ice. We also calculate the mean sea level referenced to EGM2008 geoid for the 18 campaign periods, which can be used to improve the Geoid model in the Southern Ocean and to map dynamic ocean topography. The ICESat data used in this study are release version 633 with transmitted pulse Gaussian/centroid peak location correction (G-C) applied on a shot-to-shot basis. IceBridge sea-ice freeboard and thickness data over the Weddell Sea and Amundsen and Bellingshausen Seas will be used to compare with the ICESat results.


Photon-counting laser altimetry for the retrieval of sea-ice freeboard: pre-launch activities for NASA’s ICESat-2 mission

Kelly M. Brunt, Sinead L. Farrell, Julia M. Ruth, Kaitlin M. Walsh

Corresponding author: Kelly M. Brunt

Corresponding author e-mail: kelly.m.brunt@nasa.gov

NASA’s Ice, Cloud and land Elevation Satellite-2 (ICESat-2), scheduled for launch in 2016, is a follow-on to the ICESat mission, which operated between 2003 and 2009. ICESat-2 is a next-generation laser altimeter designed to continue key observations of the Earth’s polar regions for the sustained assessment of changes in ice-sheet mass balance and sea-ice thickness and volume. ICESat-2 will carry the Advanced Topographic Laser Altimeter System (ATLAS), which represents a new approach for space-borne determination of surface elevation by using a photon-counting detection strategy that includes a high-repetition-rate 532 nm wavelength laser. An airborne ATLAS simulator, called the Multiple Altimeter Beam Experimental Lidar (MABEL), has been deployed to gather key pre-launch data for: (1) the development of ATLAS geophysical algorithms; (2) detailed ATLAS error analysis; and (3) ATLAS model validation. Here we present an overview of the ICESat-2 mission for sea ice. We present analysis of MABEL data recently gathered over the Arctic ice pack. We demonstrate the capabilities of photon-counting laser altimetry for the accurate retrieval of sea-ice freeboard. We discuss approaches for developing a post-launch, quick-look sea-ice freeboard product of potential interest to the operational community. Finally, we briefly describe current plans for post-launch calibration/validation activities for the ICESat-2 mission.


Characteristics of sea-ice draft revealed by a moored ice-profiling sonar in the Chukchi Sea off Barrow, Alaska

Yasushi Fukamachi, Daisuke Simizu, Kay Ohshima, Katsushi Iwamoto, Hajo Eicken, Andrew Mahoney

Corresponding author: Yasushi Fukamachi

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

The recent decrease in summer sea ice in the Arctic has impacted the surrounding coastal regions with respect to shipping, oil and gas development, coastal erosion, and use of the ice as a hunting platform among others. For these activities, ice thickness information is of major importance. However, availability of ice thickness data is still quite limited due to the challenges of both in situ and remote-sensing observations. From August 2009 to July 2010, two moorings with ice-profiling sonars were deployed in the coastal region of the Chukchi Sea off Barrow, Alaska. These mooring observations are part of the Seasonal Ice Zone Observing Network (SIZONet) to complement various other sea-ice observations in the region. The study sites are located within the Chukchi Sea polynya. Each mooring consists of an ice-profiling sonar, an acoustic Doppler current profiler, and a conductivity-temperature recorder. Here we present the first time-series data of sea-ice draft in this region from one of the moorings located at 71.237°N, 157.653°W near a data gridpoint of the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). The processed draft time-series reveals a number of deep keels exceeding 20 m (with a maximum of ~25.4 m) from February to May 2010. The time series also documents the occurrence of thin-ice (polynya) periods between incidences of deep keels from late fall to early spring. Some of these thin-ice periods are also captured well in the daily ice thickness estimates based on the AMSR-E data. We plan to present detailed ice-draft characteristics and statistics.


On the validity of Hibler’s sea-ice rheology for the seasonal sea ice evaluated from ice-drift pattern and SAR-derived ice-thickness distribution

Takenobu Toyota, Noriaki Kimura

Corresponding author: Takenobu Toyota

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

Sea-ice rheology formulated by Hibler (1979), which is mainly used in the present numerical sea-ice models irrespective of regions, was examined for the Sea of Okhotsk ice, typical of the seasonal ice zones (SIZ), based on the AMSR-derived ice-drift pattern and concentration and the PALSAR-derived ice thickness distribution. This rheology is characterized by the plastic behavior of sea ice under ordinary stress, in which the yield curve is given by an ellipse with an aspect ratio of 2 and the ice pressure is calculated as a function of ice concentration and thickness. This was initially introduced to reproduce the seasonal variation of the perennial sea ice in the Arctic Ocean. As for its applicability to the SIZ, where various types of sea ice are present, it still needs validation from observational data. In this study, the validity of Hibler’s rheology was investigated from a standpoint of working rate done by deformation field, as Stern and others (1995) did for Arctic sea ice. In the analysis the ice-drift pattern was obtained with a grid spacing of 37.5 km from the 89 GHz brightness temperature of AMSR-E using a maximum cross-correlation method. The temporal evolution of ice thickness distribution was obtained with a spatial resolution of 100 m from the regression, which correlates the backscatter coefficients of PALSAR with ice thickness. The result shows that Hibler’s rheology represents the property of deformation well and the observed increase in ice thickness is consistent with his formulation on a 150 km scale. This suggests that on this scale this rheology is applicable also to the SIZ. Furthermore, to examine the dependence on the scale, the coastal radar data obtained around Hokkaido were also analyzed for validation. The grid spacing of the ice-drift pattern obtained from the coastal radar was 1.3 km, and the 26 km × 26 km area was investigated for analysis. As a result, it is found that Hibler’s rheology represents the property of deformation on this scale as well as on a 150 km scale, which indicates a scale invariance property in the sea-ice dynamics. Given that the ice rheology is determined by individual ice floes composing the entire sea-ice area, this property may be attributed to the scale invariance in floe size distribution.


Impact of resolution error as an uncertainty in sea-ice thickness retrievals

Cathleen Geiger, Jacqueline Richter-Menge, Jesse Samluk, Hans-Reinhard Mueller, Peter Wadhams

Corresponding author: Cathleen Geiger

Corresponding author e-mail: cgeiger@udel.edu

Findings from a joint US-European program in 2007 show that sea-ice thickness retrievals are influenced by error sources which cannot be corrected by calibration alone. Here, we explore a particular error source which climatologists refer to as resolution error and sonar specialists call beam-width error. Using near-coincident electromagnetic induction retrievals and upward-looking sonar results, we demonstrate how resolution error causes spatial aliasing and propagates as a power law to larger scales. We make recommendations for storing resolution error slopes with their accompanying uncertainties as a compact meta-data format for archived instrument records to support subsequent applications including regional and climate forecasting, socio–economic human activities, and monitoring of polar ecosystems.


Sea-ice floe size distribution in the Labrador Sea, 1987

Benjamin Holt

Corresponding author: Benjamin Holt

Corresponding author e-mail: Benjamin.M.Holt@jpl.nasa.gov

In March 1987, the Labrador Ice Margin Experiment (LIMEX’87) took place off the coast of Labrador. In situ measurements were obtained in this marginal ice zone (MIZ), with the intent to examine the transition of a MIZ cover through the onset of seasonal melt. Concurrently, the Labrador Extreme Wave Experiment (LEWEX) took place, which obtained wave and ocean measurements to examine how incoming wave fields might impact this seasonal sea-ice transition. Aircraft SAR and photography were obtained to support both experiments, which were developed as precursor field observations for the then upcoming SAR missions of ERS-1 and RADARSAT-1 launches. As described by Carsey and others (1989), an ~2.5 m swell entered the ice cover from 19 to 23 March. Swell was observed in SAR imagery initially in the outer zone of the MIZ, composed of unconsolidated brash and cake ice. Over time, the swell broke up the inner zone of consolidated ice and became detectable all the way to the coast. Subsequently, the loosen pack was rapidly carried southward by the coastal current where it melted rapidly. The ice extent was reduced from extending between 100 and 200 km from the coast to between approximately 10 and 50 km over a few days interval. A fine-resolution set of aerial photography was obtained on 26 March clearly showing a range of floe size distribution from ocean to coast. Qualitative analysis of floe size was undertaken previously but not a detailed analysis using a proper algorithm. This paper will describe results of the floe size distribution and will compare results with coincident wave measurements and additional ice observations of the drastically altered sea-ice cover by incoming waves that occurred during this experiment. We will also compare these results with previous results obtained by others and in varying ice covers. This analysis is in preparation for the Office of Naval Research Arctic Sea State campaign that will take place in 2015.


Modelling surge motions of ice floes

Lucas Yiew, Michael Meylan, Luke Bennetts, Giles Thomas, Ben French

Corresponding author: Michael Meylan

Corresponding author e-mail: mike.meylan@newcastle.edu.au

Ocean waves that penetrate into the marginal ice zone (MIZ) cause ice floes to surge back and forth. Surge motions in a compact MIZ result in floe–floe interactions – collisions and/or rafts. Floe–floe interactions extract energy from the wave spectrum, and thus contribute to wave attenuation, which restricts the distance over which the waves impact the ice cover. Further, the interactions exert forces on the floes that can weaken and fracture the ice cover. Moreover, the interactions alter the dynamic and thermodynamic properties of the ice cover. For instance, collisions transfer momentum through the ice cover, and rafting alters the form drag between the ice cover and the atmosphere and the ocean. Modelling surge motion of individual floes is the basis to modelling the onset of floe–floe interactions, and the properties of the interactions. Accurate predictions of both the amplitude of the surge motion and its velocity are necessary. These quantities depend on the properties of the incident wave, e.g. wavelength and amplitude, and the properties of the floes themselves, e.g. diameter and thickness. We are using a unique combination of mathematical and laboratory experimental modelling to understand the surge motions of individual floes. The experiments were conducted in the Model Test Basin, Australian Maritime College, UTas. The experimental data are being used to validate two mathematical models: a linear potential flow model, which is valid for floe diameters of the order of the wavelength; and a slope sliding theory, which is valid for floes much smaller than wavelengths. In particular, we are investigating the transition between the two theories.


Validation of GPS-derived ice thickness measurements over McMurdo Sound

Daniel Price, Wolfgang Rack, Patricia Langhorne, Christian Haas, Kelvin Barnsdale, Greg Leonard

Corresponding author: Daniel Price

Corresponding author e-mail: daniel.price@pg.canterbury.ac.nz

The conversion of sea-ice freeboard to thickness critically depends on the absolute accuracy of the freeboard measurement and as such on information of snow depth as well as on the densities of the water, snow and sea ice. Near Antarctic ice shelves, information is also required on the sub-ice platelet layer, a highly porous layer of crystals that accumulate at the base of the growing sea ice as a result of ice-shelf–ocean interaction. Here we present sea-ice freeboard measurements derived by a mobile GPS system over McMurdo Sound in November 2011. The GPS investigation was carried out alongside an extensive in situ study of sea-ice freeboard and thickness, snow depth, snow density and the sub-ice platelet layer thickness. GPS height measurements were differentially corrected and reduced to height above sea level (total freeboard) by calibrating to measured freeboard along track. The mean GPS-derived freeboard agrees to within 0.03 m of in situ validation points, the primary cause of error associated with GPS vertical accuracy, itself a function of baseline distance to the GPS base station at New Zealand’s Scott Base. GPS thickness estimates show good agreement with drillhole-measured thickness with a mean absolute deviation of 0.16 m when taking account of the sub-ice platelet layer, and 0.20 m when it is neglected. The largest deviations are observed in areas where the sub-ice platelet layer is thickest, the additional buoyancy altering the freeboard of the consolidated sea ice. We used the GPS freeboard together with field data to find a best estimate for the solid fraction of the sub-ice platelet layer. To produce a best fit between freeboard and consolidated ice thickness, we find a mean solid fraction of 0.10 is most suitable for McMurdo Sound. Our results show the sensitivity of freeboard to the sub-ice platelet layer but further support the fact that the main source of error in sea-ice thickness estimation is the accuracy of the freeboard measurement. With improved GPS accuracy we aim to reduce our ice thickness errors to less than ±0.10 m. These are important considerations for the interpretation of sea-ice conditions in McMurdo Sound from satellite freeboard measurements, and support the validation effort of satellite altimeter measurements such as CryoSat-2 in the Antarctic.


15 years of 1 km resolution sea surface temperature data over the Antarctic region

Helen Beggs, Christopher Griffin, Leon Majewski, Atiur Siddique

Corresponding author: Helen Beggs

Corresponding author e-mail: h.beggs@bom.gov.au

As part of the Integrated Marine Observing System (IMOS: http://www.imos.org.au), the Australian Bureau of Meteorology produce high-resolution satellite sea surface temperature (SST) products over the Australian and Antarctic regions, designed to suit a range of operational and research applications. All these products follow the latest Group for High Resolution Sea Surface Temperature (GHRSST: http://www.ghrsst.org) file formats, including information on bias and standard deviation for each SST value, along with other useful ancillary data. The highest spatial resolution (1.1 km × 1.1 km) data from Advanced Very High Resolution Radiometer (AVHRR) sensors on NOAA polar-orbiting satellites can only be obtained through receiving direct broadcast ‘HRPT’ data from the satellite. In Australia, HRPT data is received at ground-stations located in Darwin, Townsville, Melbourne, Hobart, Perth and Alice Springs, and in Antarctica at Casey and Davis Stations. The Bureau of Meteorology, in collaboration with CSIRO Marine and Atmospheric Research, is combining raw data from the various ground-stations and producing real-time HRPT AVHRR skin (~10 micron depth) and foundation (~10 m depth) SST data files in the GHRSST formats for both geolocated, single swath data (‘L2P’) and various gridded (level 3) files (http://imos.org.au/sstproducts.html). In addition to the Australian domain (70°E to 190°E, 70°S to 20°N) IMOS AVHRR SST products, the following new Antarctic domain (2.5°E to 157.5°E, 72.5°S to 27.5°S) 0.02 degree resolution gridded SST products are now available in netCDF4 format back-processed to 1998: ‘L3U’ – single orbit skin SST, ‘L3S-01day’ – day+night foundation SST from multiple satellites (updated daily), ‘L3S-1month’ – day+night foundation SST from multiple satellites (updated each calendar month). These very high resolution Southern Ocean SST datasets spanning a 15 year period are useful for studies of the relationship between sea temperature and sea ice, particularly in the period before global 1 km resolution SST datasets became available in 2002, following the launch of the Aqua satellite carrying the MODIS radiometer. The Australian and Antarctic SST products are available through the IMOS FTP server (ftp://aodaac2-cbr.act.csiro.au/imos/GHRSST). The presentation will describe the new Antarctic SST datasets and present validation results.


Automation of Antarctic landfast sea-ice retrieval from MODIS imagery

Alexander Fraser, Glenn Hyland, Rob Massom, Kay Ohshima

Corresponding author: Alexander Fraser

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

Moderate Resolution Imaging Spectroradiometer (MODIS) visible/thermal infrared (IR) imagery has been used to map landfast sea ice along the East Antarctic coast. In the past, this technique has involved creation of 20 day cloud-free composite imagery of the surface, with cloud discrimination provided by the MODIS cloud mask. Typically, several cloud-free views of the surface are available within a 20 day window, leading to blurring of the pack ice in the final composite image and straightforward visual delineation of the sharply defined fast-ice edge. However, particular environmental conditions can present problems for this technique. Such conditions include persistent cloud over the 20 day window, consistent advection of sea ice against the fast-ice edge (i.e. closure of the flaw lead separating pack from fast ice), and inaccuracies in the MODIS cloud mask, particularly during night time. To overcome these difficulties, we have developed a new technique for use with visible or thermal IR data for determining the location of the fast-ice edge. This technique eliminates the need for an accurate cloud mask product, enabling use of the algorithm with any visible or thermal IR sensor with accurate geolocation and frequent polar coverage (e.g. VIIRS). Automation of the new technique is investigated, and resulting fast-ice maps are compared with those from the existing fast-ice product. Automation will be used to extend the MODIS fast-ice product around the entire Antarctic coast, resulting in regular large-scale mapping of Antarctic fast ice for the first time.


A model reconstruction of the Antarctic sea-ice thickness and volume changes over the past decades using data assimilation

François Massonnet, Pierre Mathiot, Thierry Fichefet, Hugues Goosse, Martin Vancoppenolle, Christof König Beatty, Thomas Lavergne

Corresponding author: François Massonnet

Corresponding author e-mail: francois.massonnet@uclouvain.be

Sea-ice variability in the Southern Ocean has a complex spatio–temporal structure. In a global warming context, the Antarctic sea-ice cover has slightly expanded over the recent decades. This increase in sea-ice extent results, however, from the sum of positive and negative regional trends and is influenced by a wide range of modes of climate variability. An additional view on sea-ice thickness and volume changes would improve our understanding. Still, no large-scale multi-decadal well-sampled record of Antarctic sea-ice thickness exists to date. To address this issue, we assimilate real sea-ice concentration data into the ocean–sea-ice model NEMO-LIM2 using an ensemble Kalman filter, and demonstrate the positive impacts on the global sea-ice cover. We find that the global Antarctic sea-ice volume has risen at a significant pace over the period 1980–2008, with an increase in the Ross and Weddell Seas and a decrease in the Amundsen-Bellingshausen Seas. Sea-ice volume anomalies co-vary well with extent anomalies, and exhibit yearly to decadal fluctuations. The results stress the need to analyze sea-ice changes at the regional level first and then at the hemispheric level.


Solar energy budget of first-year sea ice in the central Arctic: autonomous observations from two summer melt seasons

Caixin Wang, Mats A. Granskog, Sebastian Gerland, Stephen R. Hudson, Donald K. Perovich, Marcel Nicolaus, Tor Ivan Karlsen, Kristen Fossan, Marius Bratrein

Corresponding author: Caixin Wang

Corresponding author e-mail: caixin.wang@npolar.no

A spectral radiation buoy (SRB) was developed to autonomously measure the spectral incident, reflected and transmitted solar radiation (350 to 800 nm) above and below sea ice. It was deployed on drifting first-year sea ice near the North Pole prior to melt onset in mid-April 2012 and 2013, co-located with ice mass-balance bouys (IMBs). The buoys drifted southward, reaching Fram Strait in early fall, covering the complete melt season in both summers. In 2012 albedo was large and transmittance was very small until snowmelt onset on 10 June. The decrease in snow depth thereafter was accompanied by a decrease in albedo and an increase in transmittance. The greatest transmission of solar radiation to the ocean occurred in July (average 20 W m–2), following the disappearance of snow and formation of melt ponds, which resulted in low albedo and high transmittance. The solar heat that accumulated in the ocean was enough to cause two-thirds of the observed ice bottom melt; the remaining bottom melt may be explained by radiation entering through open water or melt ponds. Estimates of the energy absorbed in snow and ice corresponded well with observed surface melt, demonstrating that solar heating played an important role in the ice melt in the Arctic basin in 2012. Recent data from 2013 indicate that snowmelt onset was later than in 2012, and it took longer for the snow to melt completely, even though there was less snow in 2013. This may be due in part to lower temperatures in 2013. Thus albedo was higher and transmittance was lower than in summer 2012; however, this may also be due to the difference in the extent of melt ponds that were in the field of view of the SRB. All these factors likely had significant consequences for the energy available for ice melt. In this presentation we will examine in detail the partitioning of solar energy during these two contrasting summers, and showcase the importance of the surface conditions, especially snow cover, in the summer melt of Arctic sea ice.


Improving simulation of drift speed in a coupled ice-ocean model

Einar Olason, Dirk Notz

Corresponding author: Einar Olason

Corresponding author e-mail: einar.olason@zmaw.de

We address the problem of incorrect seasonality and long-term trend in the mean drift speed of modelled Arctic sea ice. This temporal variability is known to be poorly captured by most of the CMIP3 and CMIP5 models. We examine the cause of this inconsistency by considering the parameter space of the Max Planck Institute’s coupled ice–ocean model forced with atmospheric reanalysis. We explore only the parameter space of the ice model, focusing on the relationship between ice strength on the one hand and concentration and thickness on the other, as well as on the effects of altering the drag coefficients. We find that using the default parameter values, variations in ice drift speed in our model depend almost solely on variation in the applied wind, similar to what can be seen in other CMIP3/5 models. We can produce a more realistic seasonal cycle by increasing the ice internal strength P*, indicating that using too low a value for P* results in modelled ice that does not give sufficient resistance to mechanical deformation. When concentration is low, the ice is mostly in free drift and variations in the drift speed are then primarily controlled by variations in wind speed. When concentration is high, however, the ice has much higher internal resistance, which must be correctly parameterized in the model, in order for the seasonal cycle to be correctly reproduced.


Movement of a sea-water front through snow: model and experiment

Felix Matt, Philipp Griewank, Dirk Notz

Corresponding author: Felix Matt

Corresponding author e-mail: felix.matt@zmaw.de

In some parts of the Southern Ocean, snow ice contributes to up to 40% of the total ice mass. However, the formation of snow ice via flooding of snow-covered sea ice with sea water is still poorly understood. To gain more information about the internal processes during a flooding scenario, we investigated the vertical and horizontal movement of sea water through snow. To do so, we developed a 2-D model based on capillary suction in combination with Darcy’s law and conducted laboratory experiments. The model includes the evolution of the temperature field due to advection and diffusion and the evolution of the brine salinity due to advection and phase change. For the experiments, snow of known density and temperature was horizontally flooded with sea water at freezing temperature. Temperature sensors were used to detect the location of the sea-water front inside the snow. Variations of snow height, density and temperature were done. Both model results and experiments show a high dependence of the front velocity on snow density and temperature, and a strong vertical suction above the water table. The comparison of model results and experiment data is followed by further model studies.


Sea-ice movement in Fram Strait

Jennifer King, Grant Bigg, Gunnar Spreen

Corresponding author: Jennifer King

Corresponding author e-mail: arctic_jen@yahoo.co.uk

Fram Strait is the only deep connection between the Arctic Ocean and the other world oceans and as such forms the main gateway through which sea ice leaves the Arctic Ocean: 90% of sea-ice export from the Arctic happens through Fram Strait. When compared with the mean sea-ice circulation in the central Arctic, sea-ice speeds in Fram Strait are high and the ice cover is more fragmented. This makes Fram Strait a challenging region for sea-ice drift tracking, where different approaches to those used in the central Arctic might be suitable. The ITSARI (Iceberg Tracking Using SAR Images) algorithm, originally developed to track icebergs in the Antarctic, has been adapted to track individual sea-ice floes in Arctic conditions. Ice floes are identified and tracked using brightness values and shape characteristics. I will present a case study from late summer 2010 that showcases the use of this algorithm to investigate the movement of individual sea-ice floes. The movement of the individual floes is placed within the context of the movement of the ice pack as a whole as can be inferred from traditional cross-correlation analysis. The tracking of individual distinctive ice floes within the pack allows an improved understanding of the variability in ocean surface circulation and can be related to the meteorology of the area.


Biological and physical controls on DMSP dynamics in ice-shelf-influenced fast ice

Gauthier Carnat, Timothy N. Papakyriakou, Jiayun Zhou, Bruno Delille, Thomas Goossens, Tim Haskell, Véronique Schoemann, Jean-Louis Tison

Corresponding author: Gauthier Carnat

Corresponding author e-mail: gauthier.carnat@gmail.com

Dimethylsulfide (DMS) is a volatile sulphur compound produced by the degradation of dimethylsulphoniopropionate (DMSP), a metabolite synthesized by microalgae as a cryoprotectant and osmoregulator. It is also an important climate-active gas, being the primary source of marine-derived sulphate aerosols, which play an important role in the earth–atmosphere radiation balance. In the last two decades, there has been an increasing interest in the role of the marine cryosphere in the DMSP cycle, motivated by repeated observations of very high DMSP concentrations in sea ice. However, our understanding of the factors driving the spatiotemporal variations of these high concentrations, and hence the fate of the sea-ice DMS pool, remains limited. To date, studies have essentially focused on biotic factors, attributing the high DMSP concentrations to the high biomass of the sympagic communities, and to their strong physiological response to the low temperature and high salinity stresses of the brine habitat. We present here an approach integrating both biotic and abiotic factors, as we investigate the influence of sea-ice growth processes and brine dynamics on the DMSP cycle. We focus on a fast-ice site (Cape Evans, McMurdo Sound, Antarctica) under the influence of ice-shelf waters, and provide measurements covering a full cycle of ice growth. We show a good correspondence between isolated maxima of DMSP in interior ice and the occurrence of platelet crystals in the ice texture. We develop the idea that platelet ice formation in May strongly modifies the production of DMSP by (1) favoring the incorporation of strong DMSP producers and (2) exposing these producers to stronger environmental stresses. We then show the influence of the development and decline of a strong diatom bloom from October to November on bottom ice DMSP concentrations. Finally, we show that the increase in brine volume fraction (permeability) on warming in early December triggers (1) an important release of DMS to the ocean through brine convection and (2) a vertical redistribution of DMSP across the ice.


Modeling frazil ice particles and their impacts on ocean mixed layer

Yoshimasa Matsumura, Kay I. Ohshima

Corresponding author: Yoshimasa Matsumura

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

We have performed a large eddy simulation of wintertime ocean mixed layer with explicitly modeling frazil ice particles by on-line Lagrangian particle-tracking method. The modeled frazil particles generated near surface are widely distributed in the mixed layer by brine-driven convection. Since freezing temperature decreases as pressure increases, frazil particles melt and remove latent heat from sea water after transported deeper. Consequently, the lower level of the mixed layer is selectively cooled and potential supercooling of ~0.03K is realized after 120 hour integration. The upward latent heat flux by frazil advection in the mixed layer is up to 300 W m–2, which is compensated by the downward sensible heat flux as a result of upwelling of cold water formed at the lower level. The upwelling of potential supercooled water results in in situ supercooling at the upper part of the mixed layer (above ~20 m depth). These results are qualitatively and quantitatively consistent with direct observations. As integration proceeds, frazil particles are accumulated at the surface due to buoyancy, and form grease ice. Because the insulating effect of grease ice is less than that of uniform ice cover, the net sea-ice production increase is ~60% as compared with the case where heat loss directly leads to ice cover growth.


The International Association of Cryospheric Sciences and sea-ice research

Ian Allison, Hiroyuku Enomoto, Charles Fierz, Andrew Mackintosh

Corresponding author: Ian Allison

Corresponding author e-mail: ian.allison@utas.edu.au

The International Association of Cryospheric Sciences (IACS) is the youngest association within the International Union of Geodesy and Geophysics (IUGG) of ICSU. IACS was established as a separate association in July 2007, evolving from the International Commission on Snow and Ice that can trace its own legacy back to the Commission International des Glaciers, which was founded in 1894. The objectives of IACS are to promote studies of cryosphere (snow and ice systems) on Earth and other planets of our solar system, and to promote discussion, education and collaboration on cryospheric research. The scientific objectives of the association are managed through five disciplinary divisions, including the Division of Marine and Freshwater Ice, which focuses on studies of sea ice and of freshwater lake and river ice. The Division of Marine and Fresh Water Ice aims to promote collaborative work and discussion in the many scientific and engineering communities with an interest, including oceanography, atmospheric science, remote sensing, coastal/offshore/river engineering, marine/marine and lake ecology, and many others. IACS supports collaborative studies, workshop discussions and symposia (such as this one) as important means of stimulating ideas and research into the nature of snow and ice and their role in the Earth system. It also establishes and supports working groups to explore well-constrained problems or new avenues in cryospheric science. Each working group is devoted to a theme or subject, is composed of experts in the particular field of study and has a lifetime of normally four years. IACS welcomes the involvement of sea-ice researchers in its activities, and invites suggestions of sea-ice research issues that could be advanced through a working group.


Changes in northeast Greenland fast-ice extent 1999–present

Nicholas Hughes, Alexander Fraser

Corresponding author: Nicholas Hughes

Corresponding author e-mail: nick.hughes@met.no

The northeast coast of Greenland, north of 75°N, is fringed by a shallow continental shelf that is host to one of the largest expanses of landfast ice on Earth. This paper examines the annual development of the fast ice during the period 1999 to present and compares the changes observed with earlier studies going back to the 1930s. The period was chosen because it extends on early studies that utilized low-resolution passive microwave data, and can use the large archive of Moderate Resolution Imaging Spectroradiometer (MODIS) data from optical sensors on NASA’s Terra Earth Observing System (EOS AM) and Aqua (EOS PM) satellites. Whilst these provide sufficient detail during clear-sky periods, the region is often cloud-covered and is in darkness during winter. Therefore the study will also analyse the abundant all-weather synthetic aperture radar (SAR) data from the Envisat and RADARSAT-2 satellites for the most recent 6 years. The analysis will evaluate the mapping of fast ice in manually drawn operational sea-ice products during the period and whether changes to the types of satellite images being used have improved accuracy. The area is seeing increased activity from ship traffic including research and tourist cruises, and hydrocarbon exploration. Periodic breakouts of the fast ice in the form of giant ice floes pose a hazard to safe navigation and need monitoring to reduce risk. These floes can also incorporate multi-year ice, ice deformation features, and icebergs that make them difficult for current ice management techniques. The extent and timing of these events is examined through the analysis of the satellite data, as well as the timing of freeze-up that determines the length of the summer period when the fast ice is not present. The loss of fast ice during long periods of the year also has implications for the coastal environment of Greenland that could contribute to the retreat of tidewater glaciers and increased coastal erosion.


Does the precipitation of ikaite (CaCO3·6H2O) in sea ice have any significant implications in other chemical processes within or on sea ice?

Gerhard Stephan Dieckmann, Daiki Nomura, Yubin Hu, Gernot Nehrke

Corresponding author: Gerhard Stephan Dieckmann

Corresponding author e-mail: gerhard.dieckmann@awi.de

Since the discovery and first quantification of ikaite in sea ice, the question of how its precipitation may affect biogeochemical processes within and at the sea–ice interfaces still remains largely unanswered. So too are the processes leading to its precipitation. The fact is that precipitation of ikaite seems to be ubiquitous in Arctic and Antarctic sea ice, however, it seems to be highly variable and heterogeneous since we find it in a variety of sea-ice stages ranging from very young sea ice to multi-year ice. We present both our own field and experimental results and discuss these in the context of recent developments and information regarding the chemistry and possible fate of ikaite in sea ice. In this review we hope to convey the current state of knowledge as well as point out unanswered questions regarding ikaite chemistry in sea ice so as to clarify if the process of ikaite precipitation has any major implications for other chemical processes at sea–ice interfaces.


Present Arctic sea-ice decline from the perspective of a rapid ice loss event (RILE)

Paul Hezel, Cecilia Bitz, Thierry Fichefet, Hugues Goosse, Marika Holland, François Massonnet

Corresponding author: Paul Hezel

Corresponding author e-mail: paul.hezel@uclouvain.be

We explore the possibility that the current recent downward trend in summer Arctic sea-ice extent can be characterized as a rapid ice loss event (RILE), a short period of extreme September sea-ice loss (~5 years). We assess the probability of RILEs in CMIP5 for 1850–2100 (RCP8.5) using 84 ensemble members from 37 models. RILEs occur on average three times in each ensemble member. The probability of a RILE increases dramatically as the sea-ice extent approaches nearly ice-free conditions due to both the increase in variability of the sea-ice extent as the ice thins and the downward trend in sea ice resulting from the increasing radiative forcing. In the models, extrapolation of a RILE trend well overestimates the ice loss within the model realization. Within 10 years of the end of a RILE, the sea-ice extent returns to the background rate of loss, similar to the loss rate over a period of 10–15 years prior to and including the RILE. There seems to be no continued rapid ice loss associated with a RILE event that leads to threshold or tipping-point behavior. Our analysis indicates that it is possible that the sea-ice extent has experienced a RILE this decade, and possibly the the sea-ice extent is on a trajectory back toward the background rate of ice loss. If in fact the sea-ice extent is in a RILE, this may be the cause of some of the discrepancy between CMIP5 modeled and currently observed trends in the Arctic. At any given time, it is impossible to distinguish a RILE from a continuously accelerating rate of loss because the difference depends on the future: a RILE necessarily ends with a period of weaker sea-ice loss, while an accelerating loss has increasingly greater loss each year.


Temporal and spatial variability of fast-ice extent in the southeastern Laptev Sea

Valeria Selyuzhenok, Thomas Krumpen, Ruediger Gerdes

Corresponding author: Valeria Selyuzhenok

Corresponding author e-mail: valeria.selyuzhenok@awi.de

Arctic sea-ice cover has undergone dramatic changes during the last decades. Although fast ice comprises only a small part of the Arctic sea ice, it significantly impacts a number of environmental processes and affects human activity. The Laptev Sea is characterized by the greatest fast-ice extent in the Arctic. However there is a lack of studies concerning fast-ice formation and break-up as well as trends in fast-ice area and extent in this region. This study aims to close this gap by investigating spatial and temporal variability of the southeastern Laptev Sea fast-ice extent. Information on fast-ice extent is obtained from sea-ice charts and remote-sensing data for the period 1996–2012. Here we present a mean climatology of the annual fast-ice cycle for the past 1.5 decades. In addition, the study links seasonal and interannual variability of fast-ice extent to atmospheric processes such as air temperature, wind direction and velocity, as well as global atmospheric circulation patterns. Furthermore, we provide insight into the importance of local bathymetry by linking fast-ice occurrence to bathymetric features.


Coprecipitation of phosphate with ikaite (CaCO3·6H2O) in simulated sea-ice brine

Yubin Hu, Gernot Nehrke, Gerhard Dieckmann, Christoph Völker, Dieter Wolf-Gladrow

Corresponding author: Yubin Hu

Corresponding author e-mail: Yubin.Hu@awi.de

Ikaite (CaCO3·6H2O) in sea ice is only recently discovered, which at the same time is also the first direct proof of CaCO3 precipitation in natural sea ice. Phosphate (PO4) is considered to be crucial for ikaite formation in sea ice. However, the role of ikaite precipitation in relation to PO4 is not understood. PO4 is an important nutrient for the sea-ice biological community. The enrichment or depletion of PO4 in sea ice has an impact on biological activity in sea ice. Generally, the concentration of PO4 in sea ice is often found to be depleted when normalized to sea-water salinity, which is attributed to the biological activity. However, there might be another mechanism explaining the depletion of PO4 in sea ice. In this study, we investigated how the precipitation of ikaite affects PO4 concentrations. Experiments were set up at pH ranging from 8.5 to 10.0, salinities from 0 to 105, temperature from 0 to -4°C and PO4 concentrations from 0 to 50 μmol kg–1 in simulated sea-ice brine, in order to understand how varied conditions affect PO4 removal during ikaite precipitation. Our results show that PO4 is coprecipitated with ikaite under all experimental conditions. The removal capacity of PO4 by ikaite precipitation increases with increasing pH. The changes in salinity (S > 0) as well as temperature do not have a significant impact on PO4 removal and initial PO4 concentrations greatly affect the PO4 coprecipitation with ikaite. These findings may cast some light on the observed variability of PO4 concentrations in sea ice.


Effect of scaling on local sea level and sea-ice freeboard/volume estimation in the Bellingshausen-Amundsen sea-ice zone from NASA’s IceBridge Airborne Topographic Mapping system, 2009–11

David Prado, Stephen Ackley, Hongjie Xie

Corresponding author: David Prado

Corresponding author e-mail: myu357@gmail.com

The Airborne Topographic Mapping (ATM) system flown on NASA’s Operation IceBridge allows for estimation of sea-ice thickness in the Bellingshausen-Amundsen Seas from 2009 to 2011. The estimation of total freeboard is based on the accuracy of local sea-level estimations. The high density of ATM L1B (~1 m footprint) observations allows for local sea-level estimation at varying spatial resolutions (footprints and segment lengths) to assess errors associated with averaging and deviation from the WGS-84 geoid. The ATM dataset will allow for a comparison between IceBridge- and ICESat-derived freeboards with estimates of the errors using the longer segment lengths, which are comparable more directly to ICESat data. Upscaling of L1B and L2 data from three 17.5 km average segments to ~50 km segments (similar to the 50 km × 50 km box used for ICESat) shows very little change when the 17.5 km (control) elevation values are used to estimate total freeboard (-0.049 m to 0.001 m). However, when the same data are processed based on the total 50 km segment length, the total freeboard estimations decrease in agreement from the averaged 17.5 km data (-0.107 m to 0.126 m). This study shows that using a 17.5 km spacing and then converting to a 50 km segment length yields similar freeboard estimations to previous studies (0.05 m) while needing fewer data to estimate local sea levels accurately. Sea-ice distribution is highly variable during each of the 3 years of this study despite 2009 and 2011 having similar sea-ice freeboards. Sea-ice volumes are highly affected by sea-ice extent, causing a 1300 km3 sea-ice volume difference between 2009 and 2011 (total freeboard is 0.58 m for both years).


Landfast sea ice of McMurdo Sound as a source of bio-essential trace metals for primary productivity in the Ross Sea, Antarctica

Veronique Schoemann, Jeroen de Jong, Jean-Louis Tison, Tim Haskell, Hein de baar, Willy Champenois, Jiayun Zhou, Gauthier Carnat, François Fripiat, Thomas Goossens, Sébastien Moreau, Bruno Delille

Corresponding author: Veronique Schoemann

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

Iron (Fe) is an essential micronutrient. Its low abundance limits primary productivity in more than 30% of the oceans, including the Southern Ocean, and has a crucial impact on the biogeochemical cycles of carbon and other elements with ultimate influence on the Earth climate system. Other trace metals, like Mn, Zn, Co and Cu, are also required for microorganism cell metabolism and may be (co-)limiting. Previous data on dissolved and particulate Fe concentration showed that Fe is 10–100 times more concentrated in the sea ice than in underlying sea water and that sea-ice melt can deliver up to 70% of the daily Fe supply to the surface waters. According to budget estimates in East Antarctica and in the Weddell Sea, accumulated Fe would largely derive from the underlying sea water rather than from atmospheric inputs. Most of the available data of trace metals in the sea ice concern pack ice and Fe. Only very scarce data exist on landfast ice and on other trace metal concentrations. In this presentation, the general objective is to assess the role of landfast ice as a source of Fe and other bio-essential trace metals (e.g. Mn, Zn, Cu, Mo, Cd), its impact on primary productivity and on the biological pump. Samples of sea ice, brines and sea water as well as dusts samples have been collected during the land-based sampling program YROSIAE at Cape Evans (Scott Base, McMurdo Sound, Ross Sea, Antarctica) from November 2011 to December 2011 and from August 2012 to December 2012. Dissolved and particulate trace metal concentrations have been measured by a recently developed method, which combines multiple element isotope dilution with preconcentration using the Nobias Chelate PA1 resin and ICP-MS analysis. Concentrations of trace metals in snow collected during the present study are one to up to five orders of magnitude higher than the concentrations previously observed in snow from East Antarctica, showing a much stronger dust input of these metals in McMurdo Sound. When comparing the concentrations obtained in the under-ice sea water with those obtained in the snow at McMurdo Sound, concentrations of Fe, Al, Mn and Co are much lower, whereas concentrations of Cu, Zn and Pb are similar and the concentrations of Ni, Mo and Cd are higher. Inventories of these trace metals in the landfast sea ice give insights on its role as a source of bio-essential trace metal for the fuelling of the seasonal Ross Sea bloom. Other sources of these trace metals will be addressed and compared.


Snow thickness distributions over sea ice in the vicinity of the North Pole, 2005–13

Sebastian Gerland, Christian Haas, Caixin Wang, Mats A. Granskog, Jari Haapala, Alexander Makshtas

Corresponding author: Sebastian Gerland

Corresponding author e-mail: gerland@npolar.no

Knowledge about the snow cover on Arctic sea ice is of central interest for a number of issues. Snow affects the exchange of energy between the atmosphere, the snow-ice system, and the ocean, thus playing an important role in ice growth, melt and upper ocean heating. It also affects the light budget, making it highly important for biota in and below the ice. Furthermore, it is of key relevance for converting any altimetry data, such as from the radar and laser altimeter satellites CryoSat-2 and ICESat, to sea-ice thickness. Recently, the composition of ice types in the Arctic has been changing on large scales. Summer sea-ice extent and ice thickness have been decreasing over the recent decades, along with an increase in ice drift speed. Consequently, sea ice has also become younger, on average. These changes likely also have consequences for the snow cover and related surface processes. We will present data from drift station expeditions as well as so-called ‘last degree’ skiing expeditions between 2005 and 2013 to examine snow thickness distributions in the area between 88 and 90°N. The datasets consist of enough independent data points within one or several limited areas per expedition, to calculate probability density functions for deriving modal and mean snow thicknesses, and to investigate the general nature of the snow thickness distribution for the respective subregions. The results from the different years are compared to quantify the interannual variability of mean and modal snow thicknesses. The new snow thickness data are compared with ice thickness information, where available, and with the commonly used climatology, published by Warren and others in 1999. The impact of the snow cover and its changes on the development of the pan-Arctic sea-ice evolution in a season, such as the snow’s role in affecting freezing rates, surface albedo and melt pond formation, are discussed. Increased use of identical measurement protocols would aid better quantification of the amount of snow on Arctic sea ice, and to detect regional variability and possible trends.


How not to compare models and observations

Dirk Notz

Corresponding author: Dirk Notz

Corresponding author e-mail: dirk.notz@zmaw.de

We discuss limitations of some standard methods that are commonly used to assess the quality of large-scale model simulations against the observational record. We focus in particular on issues that arise from internal variability of the climate system, from uncertainties in the observational record, from different temporal and spatial scales in the observations versus those in the models, and from uncertainties in reanalysis datasets. We show that the identification of model shortcomings can be severely limited by these factors. We show that differences in mean sea-ice concentration fields in the Arctic can be as large between two individual satellite retrieval algorithms as they are between an individual model simulation and a specific satellite dataset. Differences in the observed and modelled retreat of Arctic sea ice can often be explained readily by significant internal variability, while model estimates of sea-ice volume loss crucially depend on the choice of a specific reanalysis dataset for the atmospheric forcing. Differences between observed and modelled annual cycles in sea-ice drift speed are very sensitive to the choice of a specific temporal averaging. In the light of these findings, we suggest that the comparison between a model simulation and a specific observational dataset should not be interpreted as a comparison of model against reality, but instead as a comparison between two quantities with associated individual uncertainty.


The energy budget of landfast sea ice through the melt season

Stephen Hudson, Mats Granskog, Chris Polashenski, Dmitry Divine

Corresponding author: Stephen Hudson

Corresponding author e-mail: hudson@npolar.no

To adequately understand the processes involved in determining the progression of sea-ice melt, we must observe all components of the surface energy budget throughout the transition from a cold, snow-covered surface to a highly ponded, melting surface. Here we examine observations of shortwave radiation fluxes, accounting for the spatial variability of albedo, longwave radiation fluxes, and atmospheric turbulent heat fluxes made between April and June 2012 on fast ice near Barrow, Alaska, to see the progression of the ice energy budget and how the surface melt responded to the changing fluxes. The overall development of the surface is seen clearly from the average broadband albedo measured every 5 m along a 200 m line. In April and early May, the cold snow had an albedo of 0.8; a warm period (up to +2°C) from 17 to 19 May resulted in a wet snowpack, which decreased the albedo to about 0.7, where it remained stable until a rain event on 30 May reduced the albedo to about 0.55. Further melting on the first few days of June reduced the albedo to 0.4, where it remained stable until it plummeted to near 0.2 between 7 and 8 June, as much of the surface opened up into extensive ponds or flooding. Subsequent drainage of the ponds to sea level increased the albedo to about 0.35 by mid-June. Throughout the period, the largest source of energy to the ice was solar radiation, so the albedo is not only a good indicator of the surface state, but also a driving factor in its evolution. The dominance of solar radiation in the overall energy budget did not mean other components were not important. The transition to wet snow occurred as a passing weather system brought southerly winds, increasing the air temperature over the ice. The early June melt was driven by a period of thin and broken clouds, which limited longwave cooling, while allowing significant shortwave heating, while the following period of stability was caused by thicker clouds that reduced the shortwave heating and shifted the spectrum towards shorter wavelengths that are absorbed throughout the ice, rather than at its surface. Finally, the opening of the ponds took place on a clear day, with strong solar heating, at the beginning of a lengthy period of significant atmospheric heating of the surface. Additional lidar and spectral albedo information will be used to further upscale the results and to examine the effect of various weather conditions.


High-speed, high-resolution instrument suite for sea-ice process studies

Leif Riemenschneider, Dirk Notz

Corresponding author: Leif Riemenschneider

Corresponding author e-mail: leif.riemenschneider@zmaw.de

We present a newly developed suite of instruments that allows for multi-channel observations of physical parameters both in sea ice and in the ice-ocean boundary layer with high spatial and temporal resolution. The accessible properties are temperature, conductivity (i.e. salinity), electrical impedance up to 100 kHz, local thermal conductivity and flow velocity. The temperature sensor is distinguished by its ability to measure up to 64 channels quasi-synchronously with an update rate of less than 1 s. This feature allows us to capture highly dynamic processes. The arrangement of the temperature sensors has a minimum spatial resolution of 4 mm. The linear 8-channel conductivity (i.e. salinity) sensor has a spatial resolution of 5 mm and provides an update rate of 3 Hz for the complete acquisition of a profile. Electrical impedance spectroscopy in an adjustable frequency range between 0 and 100 kHz is provided by the according module. The dynamic range of the 1-D thermal mass flow sensor ranges from 10 mm s–1 up to 10 m s–1, with a possible spatial resolution of 10 mm. The low power consumption of the devices and the possibility to store the data on an internal microSD-card allow for future long-term field deployment of our instrument suite.


Development of autonomous measurement systems for observing solar radiation above and below seasonal sea ice: preliminary results and future plans

Stephen Hudson, Caixin Wang, Chris Polashenski, Sebastian Gerland, Mats Granskog, Donald Perovich, Marcel Nicolaus

Corresponding author: Stephen Hudson

Corresponding author e-mail: hudson@npolar.no

Solar radiation and the ice-albedo feedback play an integral role in the melting of sea ice. Observations of the solar albedo and transmittance of sea ice in the Arctic Ocean are rare, and complete seasonal cycles are lacking. Therefore effort has been put into developing autonomous measurement systems that can affordably give this information over a complete season. A spectral radiation buoy (SRB) was developed to autonomously measure the spectral incident, reflected and transmitted solar radiation (350 to 800 nm with 3 nm resolution) above and below sea ice. It was twice deployed in drifting ice in the High Arctic, operating successfully through two full melt seasons and sending spectra in near real time. The results showed the SRB is a suitable system for autonomously measuring solar radiation in the High Arctic. However, the SRB setup is rather expensive, which limits deploying SRBs in large numbers. To have a more affordable measurement system, we are developing systems with simpler broadband radiation sensors, covering the visible – near-infrared band, merged into existing platforms such as the seasonal ice mass-balance buoy, which can then measure both the solar energy budget and the ice mass balance from one platform. These simpler radiation sensors are also being incorporated into a standalone radiation buoy that can be easily deployed where mass-balance observations are being made by a separate system. Autonomous deployments do not come without challenges. The sensor measuring incident radiation is particularly susceptible to being covered with frost or precipitation. Heating can help with the accumulation of precipitation or frost, but it comes with significant costs related to power. The addition of a small camera to photograph the sensors would help to at least identify bad data and could additionally help data interpretation by showing the surface as well. While ensuring data quality is a challenge, the ability to deploy many sensors to record the seasonal cycle of solar energy partitioning at numerous locations at a tiny fraction of the cost of manned stations makes these autonomous buoys an attractive addition to traditional radiation stations on Arctic sea ice.


Multi-year sea-ice export through the Bering Strait during winter 2011/12

David Babb, Ryan Galley, Matthew Asplin, Jennifer Lukovich, David Barber

Corresponding author: David Babb

Corresponding author e-mail: umbabb@cc.umanitoba.ca

Six ice beacons deployed in the Beaufort Sea during August 2011 tracked the anomalous export of multi-year sea ice from the Chukchi Sea through the Bering Strait to the Bering Sea between November 2011 and May 2012. These are the first observations in 34 years of ice beacon export through the Bering Strait. Using 34 years of passive-microwave-derived ice motion fields we find that during 2011/12 southward ice motion in the Chukchi Sea persisted for a record 6 of 7 months and that sea-ice speeds were significantly faster than the long-term mean. The combination of increased ice speeds and reduced likelihood of ice arch development through the strait culminated in the record export of 13.5 × 103 km2 of sea ice through the Bering Strait. Historically, ice export through the Bering Strait was thought to be episodic (1–5 days) in nature with only small southward fluxes of first-year sea ice through the strait. Monthly sea-level pressure fields, dominated by an Aleutian Low and Siberian High, show anomalies in December and January played a role in initiating this event and forced multi-year ice into the southern Chukchi Sea. However these variations were small and typical of this area, yet we find no evidence of a similar export event in the last 34 years even though the forcing was similar to the climatology. This leads us to attribute this event to a change in the responsiveness of the Arctic ice pack to typical forcing mechanisms.


Observed and modelled solar transmittance of very young sea ice in the Arctic

Børge Hamre, Stephen Hudson, Mats Granskog, Sebastian Gerland, Marcel Nicolaus

Corresponding author: Børge Hamre

Corresponding author e-mail: borge.hamre@uib.no

During spring, snow-covered sea ice severely limits the amount of solar energy entering the Arctic Ocean, reducing solar heating and impacting the timing and productivity of biological activity. Leads or other areas of open water can provide temporary windows, allowing one to two orders of magnitude more solar energy into the upper ocean. However, during the first half of the Arctic sunlit season, these windows are often quickly covered by the formation of new ice. The resulting thin ice is a sort of frosted window, allowing much more light into the ocean than the snow-covered ice, but less than the open water. Understanding the transmittance of this thin ice as it grows is therefore important for fully describing the light conditions in the Arctic Ocean in spring (and during autumn freeze-up), but few measurements have been carried out on thin ice because of the difficulty and limitations of working on the weak ice. In April 2010, we measured the albedo and transmittance of new ice growing in Kongsfjorden, near Ny-Ålesund, Svalbard. The measurements were made on fast ice near a pier, on three different days, as the ice, initially 7 cm thick, grew to 11 cm and then to 15 cm. The ice was snow-free during the measurements, and radiation conditions were stable (clear skies). Samples were also collected to measure the absorption by particles and CDOM (colored dissolved organic matter) in the new ice. Though the transmittance of the thin ice dropped quickly in the near infrared, it allowed 70 to 90% of visible light through, showing that such areas can remain viable for photosynthesis for well over a week after they have refrozen. A coupled atmosphere–ice–ocean radiative transfer model (CASIO-DISORT) was used to calculate the albedo and transmittance of the ice, providing a good match to the observations. In the model, observed absorption by impurities had to be significantly reduced, likely illustrating a packaging effect as the impurities were concentrated in brine within the ice. The model was also used to estimate the brine and bubble content of the ice, and to examine the progression of the transmittance as new ice grows in conditions with different amounts of impurities or with different brine and bubble content. The results provide a good dataset for predicting the transmittance of areas of new ice in the Arctic Ocean, and thus to better quantify energy fluxes in this poorly understood and important part of the Arctic sea-ice cover.


Sea-ice–ice-shelf interaction in the eastern Weddell Sea, Antarctica

Mario Hoppmann, Marcel Nicolaus, Stephan Paul, Priska Hunkeler, Thomas Schmidt, Meike Kühnel

Corresponding author: Mario Hoppmann

Corresponding author e-mail: Mario.Hoppmann@awi.de

As of today, many processes influencing Antarctic sea ice and its physical properties still remain unknown. This fact is particularly reflected in the inability of modern climate models to accurately reproduce the observations on the Southern Hemisphere, neither on a regional scale nor on a global scale. One of the processes most likely to contribute to this divergence of model and observation in the Southern Ocean is the poorly known sea-ice–ocean–ice-shelf interaction. Ocean–ice-shelf interaction in the Antarctic is by various means closely linked to sea-ice properties, and is expected to change dramatically in the future. These interactions particularly modify the properties of sea ice directly fastened to ice shelves: basal melting leads to the formation of supercooled water masses, in which small ice crystals may nucleate. These crystals grow further, rise up and accumulate beneath the immobile fast-ice cover to form a dense ice–water mixture. The properties of this platelet layer serve as a valuable source of information about the ocean–ice-shelf interaction and its changes, which is difficult to investigate directly. Moreover, the platelet layer is a unique habitat for different kinds of organisms, modifies the surface energy balance and finally promotes sea-ice growth by being incorporated into the sea-ice fabric as a special ice type: platelet ice. Here we report the results of our investigations on the mass balance of the landfast sea ice in Atka Bay, eastern Weddell Sea, in 2012/13. We focus on the study of ice platelets, which emerge from the cavity below the nearby Ekström Ice Shelf and accumulate beneath the fast ice as early as June. We observed platelet-layer thicknesses of up to 10 m during its seasonal maximum in December. Although the fast ice is covered with thick snow early in the year, thermodynamic growth leads to sea-ice thicknesses of up to 2 m. Texture analysis of sea-ice cores confirms a contribution of platelet ice of more than 50% to total sea-ice mass, which agrees well with a thermodynamic sea-ice growth model.


Dimethyl sulfide and dimethylsulfoniopropionate profiles in sea ice during winter in the Weddell Sea

Christiane Uhlig, Jean-Louis Tison, Janne-Markus Rintala, Gauthier Carnat, Gerhard Dieckmann, Bruno Delille, Ellen Damm

Corresponding author: Christiane Uhlig

Corresponding author e-mail: Christiane.Uhlig@awi.de

This study presents profiles of the organic sulfur components dimethylsulfide (DMS) and dimethylsulfoniopropionate (DMSP) in sea-ice cores collected during the AWECS (Antarctic Winter Ecosytem Climate Study) cruise on RV Polarstern (ANT29-6) in the Weddell Sea. DMS is a semi-volatile sulfur component and under discussion to be climate active, as its oxidation products might act as cloud condensation nuclei – thus cooling the atmosphere. It is produced by enzymatic cleavage of the precursor DMSP, which is synthesized by various types of phytoplankton and serves for example as compatible solute and cryoprotectant. Due to the physico–chemical conditions given, i.e. the high salinity and the icy matrix, sea ice as habitat favors production of high levels of DMSP by the inhabiting microalgae. DMSP and DMS are frequently found in high concentrations in sea ice during spring and summer. The aim of this study was to investigate DMS(P) levels in winter sea ice as data for the winter season is yet scarce, but is of importance for global budgeting. Preliminary results of our study show that DMS(P) production in sea ice in the Weddell Sea is also significant during winter. This stands in contrast to previous measurements in Arctic winter sea ice (CFL-IPY cruise in the Circumpolar Flaw Lead Polynya), where DMS(P) concentrations were very low. Possible explanations for the differences between DMS(P) levels in the Arctic and Antarctic might be the different snow cover and thus insulation, light regimes and also microbial community structure within the ice. DMS(P) levels were generally correlated with chlorophyll a concentrations, although the details are complex and seem to be influenced by species composition and species specific DMSP/Chla ratios. The DMS profiles mirrored the permeability of the sea ice following DMSP in the impermeable areas while showing losses to the ice surface and ice–water interface in the more permeable regions. Winter DMS(P) profiles are furthermore compared with data collected during the following spring cruise of RV Polarstern (ANT29-7) in the Weddell Sea.


Bottom and surface ablation, annual layering, and the mass budget of multi-year sea ice in the Beaufort and Chukchi Seas

Hajo Eicken, Andrew Mahoney, Don Perovich, Stefan Hendricks, Sookmi Moon

Corresponding author: Hajo Eicken

Corresponding author e-mail: hajo.eicken@gi.alaska.edu

Recent thinning and reductions in summer ice extent of Arctic sea ice are particularly pronounced in the Beaufort and Chukchi Seas. Beginning with the then-record minimum summer ice extent of 2007, remote-sensing data show a steady decline in the extent of multi-year ice in the Chukchi and Beaufort sectors of the Arctic Ocean. Starting in 2007, we have conducted airborne electromagnetic-induction thickness surveys and surface-based measurements out of Barrow, Alaska, to track changes in the composition of the ice pack. In 2013, multi-year ice was absent from the study region in the southern Beaufort and Chukchi Seas. Surface observations on level multi-year floes revealed ice thicknesses between 2.2 and 3.4 m, which agree well with modes in the ice thickness distributions obtained from airborne surveys. Identification of annual layers in ice cores allows for dating of level ice floes, and may indicate that during the study period progressively younger ice (from 5–7 years to 3–4 years in age) has been advected into the study region from the High Arctic. Stratigraphic analysis of ice cores extracted from level multi-year ice provides data on net annual ice accretion and highlights the increasing importance of bottom as opposed to surface ablation in this region. This finding is supported by ice mass-balance buoys drifting through the study area. Increased bottom melt is attributed to enhanced solar heating of the surface ocean. Stable isotope data and ice-core stratigraphy show that a substantial fraction of bottom and surface ice meltwater is retained at the bottom of the ice. These meltwater layers, traces of which are found in all multi-year ice cores, form part of the annual ice accretionary layers and contribute to the total mass budget of level ice in the region. At least two of the ice mass-balance buoys drifting through the study region provide a record of summer ice accretion associated with under-ice melt ponds and false bottoms formed through freshwater underplating. In addition to complicating efforts to partition surface and bottom melt of sea ice, underplating also represents a possible source of error in measuring the location of the ice–water interface.


Ocean–ice-shelf interaction at the Ekström Ice Shelf, Antarctica

Mario Hoppmann, Lars Kindermann, Stephan Paul, Priska Hunkeler, Marcel Nicolaus

Corresponding author: Mario Hoppmann

Corresponding author e-mail: Mario.Hoppmann@awi.de

The existence of a several-meter thick layer of disc-shaped ice crystals (ice platelets) under the landfast sea ice of Atka Bay, eastern Weddell Sea, is indirect evidence of extensive ocean–ice-shelf interaction in this region. These ice platelets form in the water column from supercooled water exiting the cavity of the Ekström Ice Shelf. They rise at the base of the ice shelf, accumulate in a dense layer and ultimately some of them are incorporated into the fast-ice fabric. In addition to its role in modifying sea-ice mass and energy budgets, the platelet layer also provides a unique habitat for algae, microorganisms and fish. From the timing and amount of ice-platelet accumulation, conclusions on the ocean–ice-shelf interaction may be drawn. Since December 2005, a CTD probe is permanently installed below the Ekström Ice Shelf as part of the Perennial Acoustic Observatory in the Antarctic Ocean (PALAOA). It records the physical properties of the water body beneath the 100 m thick ice shelf at a distance of 1–2 km from the ice edge in 10 min intervals. Here we present the data from 7 years of measurements, indicating the seasonal and interannual variability of potentially supercooled water masses (ice-shelf water). These data are processed including CTD measurements and time-series observations of platelet-layer thickness from the neighboring landfast sea ice of Atka Bay in 2012. From these comparisons, we expect a better understanding of ocean–ice-shelf interactions and the processes that influence sea-ice growth.


September Arctic sea-ice minimum predicted by spring melt pond fraction

David Schroeder, Danny L. Feltham, Daniela Flocco, Michel Tsamados

Corresponding author: David Schroeder

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

Existing attempts at seasonal forecasts of summer Arctic sea-ice extent and volume are of limited skill. Here we show that the Arctic sea-ice minimum can be forecasted by means of the pond area in spring. Ponds form on the Arctic sea ice from the accumulation of surface melt during spring and affect the heat and mass balances of the ice cover, mainly by decreasing the value of the surface albedo by up to 20%. We have developed a melt pond model suitable for forecasting the presence of melt ponds based on sea-ice conditions. This model has been incorporated into the Los Alamos CICE sea-ice model, the sea-ice component of several IPCC climate models. Simulations for the period 1980 to 2012 are in good agreement with observed ice concentration. We find a strong correlation between pond fraction and September sea ice, which can be physically explained by the positive feedback mechanism: more ponds reduce the albedo; a lower albedo causes more melting; more melting increases the pond fraction. We apply pond fraction in May to forecast September ice extent and ice volume with a positive skill relative to a forecast of climate trend.


Monitoring Atka Bay landfast sea ice: a contribution to the Antarctic Fast Ice Network

Mario Hoppmann, Marcel Nicolaus, Petra Heil

Corresponding author: Mario Hoppmann

Corresponding author e-mail: Mario.Hoppmann@awi.de

Recent observations of a slight increase in sea-ice extent in the Southern Ocean are somewhat contradictive to the expected retreat in a warming world. In addition, it is not yet possible to simulate this behavior with numerical models. This shows that the complex interactions of Antarctic sea ice with atmosphere and ocean remain poorly understood, and we are still lacking sufficient process studies and time series to quantify physical properties of Antarctic sea ice and its snow cover. But gathering these data is most challenging in the drifting pack ice without easy logistical access for repeat and time-series measurements. On the other hand, immobile sea ice fastened to coasts and ice shelves is relatively easy to probe from nearby coastal stations. Taking advantage of this, the Antarctic Fast Ice Network (AFIN) was established as a cooperation of several international partners to provide the scientific community with continuous fast-ice observations. We contribute to AFIN with a suite of regular measurements on the seasonal fast ice of Atka Bay in the eastern Weddell Sea. Through its geographical location near the Ekström Ice Shelf, the fast ice is heavily influenced by ocean–ice-shelf interaction and is generally covered with a thick and highly variable snow cover. Furthermore, it is the breeding ground for a large colony of emperor penguins, and the fast-ice conditions are crucial for the supply of the German scientific station Neumayer III situated on the ice shelf. Here we present the concept and results of our ongoing monitoring program at Atka Bay. Since 2010, we perform repeated drillhole measurements of sea-ice thickness, snow depth and freeboard along transects over the 25 km wide embayment. In addition, we record total thickness distributions by electromagnetic surveys. Since 2011, we also deploy mass-balance buoys, an automatic weather station and a spectral radiation station on the fast ice. These autonomous instruments provide continuous measurements of vertical ice temperature profiles, atmospheric conditions, surface albedo and light transmission through sea ice and snow from mid-winter (June) to early summer (December). At the end of the growth season, we recover sea-ice cores to determine the growth history with a view on contribution from snow ice and platelet ice.


Seasonal melt-freeze transitions over Arctic sea ice retrieved from active microwave instruments

Jonas Mortin, Stephen Howell, Libo Wang, Chris Derksen, Gunilla Svensson, Rune Graversen

Corresponding author: Jonas Mortin

Corresponding author e-mail: mortin@misu.su.se

The duration of Arctic sea ice – inherently linked to the seasonal melt-freeze transitions – moderates the surface energy budget through its high albedo and its strong dampening effect on the ocean–atmosphere heat flux. This has significant implications on the climatic conditions of the region and potentially beyond it. Therefore, in order to understand the Arctic climate system and its variability, it is important to continuously monitor the seasonal melt-freeze transitions over sea ice. Active microwave instruments (scatterometers) are well suited to observe the seasonal transitions. They measure the surface all year and are sensitive to the dielectric properties of the surface, which are strongly related to the amount of liquid water at the surface. The liquid water content, in turn, changes over the year when the surface undergoes seasonal melt-freeze transitions, thus inducing distinct signals in the backscatter measurements. To retrieve the seasonal transitions, we employ an edge detector algorithm that detects the backscatter changes associated with the transitions. An iterative procedure and a constraint using sea-ice concentrations are applied to mitigate erroneous outliers. We obtain the seasonal melt-freeze transitions over all Arctic sea ice using resolution-enhanced data (4.45 km) from three instruments: (1) QuikSCAT, which operated during 1999–2009 at 13.4 GHz; (2) OSCAT, in orbit since 2009 measuring at 13.5 GHz; and (3) ASCAT, from which high-resolution data are available since 2009 at 5.3 GHz, planned for another platform. Because of the different frequencies and orbital parameters of the instruments, we investigate the consistency between the backscatter measurements with an emphasis on the seasonal transitions over Arctic sea ice. Although operating in different frequencies, the scatterometers respond similarly to changes in the liquid water content of the surface. However, differences are evident in some regions, particularly at the early melt. We also compare the retrieved transitions with other, independent datasets: transitions from passive microwave radiometer, skin temperature, surface air temperature, and more. The agreement is generally good and physically consistent between the datasets, which enables a comprehensive multi-dataset trend analysis of the seasonal melt-freeze transitions. Also, the preconditioning and effects of the melt and freeze transitions can be investigated by the surface energy budget using models.


SIPEX 2 primary production field tests: comparison of 14C and 13C methods

Anne-Julie Cavagna, Karen Westwood, Arnout Roukaerts, Delphine Lannuzel, Klaus Meiners, François Fripiat, Frank Dehairs

Corresponding author: Anne-Julie Cavagna

Corresponding author e-mail: acavagna@vub.ac.be

The accurate evaluation of primary production in sea ice is an important challenge since recent studies highlight the importance of sea-ice biogeochemistry in the global carbon cycle. Such a parameter is crucial for flux quantification and further modelling. Various methods are used to estimate primary production ranging from indirect (remote-sensing) to direct (14C/13C tracers) methods. However, intercomparison of methods has rarely been shared in the international community though such action is important to validate quantitative estimation of primary production. During the SIPEX 2 expedition (September to November 2012 – R/V Aurora Australis) we were able to conduct several field tests to compare field methodologies and 14C and 13C tracer methods. Briefly, (1) we tested various slab thicknesses of bottom ice cores (from 2 cm to 25 cm) and (2) we worked either directly on crushed ice (allows working on complete community) or on ice flushed with 0.2 μm filtered underlying sea water allowing differentiation between the liquid phase (algae floating free in brine channels) and the solid ice phase (containing attached algae). Data acquisition is in process. We will discuss results and evaluate the different methodologies applied during SIPEX 2.


A decadal record of Arctic sea-ice thickness change from ICESat, IceBridge and ICESat-2

Sinéad Farrell, Jackie Richter-Menge, Nathan Kurtz, Thomas Newman, Julia Ruth, David McAdoo, H. Jay Zwally

Corresponding author: Sinéad Farrell

Corresponding author e-mail: sineadf@umd.edu

Arctic sea ice has experienced a rapid change in its composition over the last decade, transitioning from a predominantly thick, multi-year ice pack to a thinner, seasonal pack. Altimeters on both airborne and satellite platforms provide measurements of sea-ice freeboard, from which sea-ice thickness and volume may be inferred. Observations from the ICESat and CryoSat-2 missions indicate a significant decline in ice thickness and volume over the last 10 years. The greatest losses have been observed in the oldest multi-year ice areas. We will show the regional trends in ice thickness derived from satellite and airborne altimetry, contrasting observations of the multi-year ice pack with zones now dominated by seasonal ice. We will describe our current efforts to more accurately convert from freeboard to ice thickness, with a modified methodology that corrects for range errors and instrument biases. In particular we will describe an improved treatment of snow depth and ice density, defined with respect to first-year and multi-year ice types. Our analysis spans 2003 to 2013 and utilizes data from the NASA ICESat and IceBridge missions. We discuss our observations in the context of recent results from CryoSat-2, and previous results from Envisat, based on research conducted by Professor Seymour Laxon and Dr Katharine Giles. With the planned launch of ICESat-2 in 2016 we may expect continuity of the sea-ice thickness time series until at least the end of this decade. Data from this mission, together with ongoing observations from CryoSat-2, will allow us to better understand the evolution of Arctic sea-ice thickness over a long-term period. We will briefly describe the current status of the ICESat-2 sea-ice data products, and we will demonstrate the utility of micro-pulse photon-counting laser altimetry over the polar oceans.


Connections between sea-ice changes and recent break-ups of the Petersen Ice Shelf, Ellesmere Island, Nunavut, Canada

Adrienne White, Luke Copland, Derek Mueller

Corresponding author: Adrienne White

Corresponding author e-mail: awhit059@uottawa.ca

In this study we used a combination of field measurements and satellite imagery to complete the first comprehensive assessment of the current state of the Petersen Ice Shelf (thickness, surface mass balance, area), its changes since 1959, and an evaluation of factors contributing to its recent disintegration. Ground-penetrating radar measurements in May 2011 revealed a mean ice thickness of 28.7 m, while ablation stake measurements showed a mean surface mass balance of -1.18 m w.e. a–1 between 2011 and 2012. An analysis of Landsat, ASTER and synthetic aperture radar satellite imagery since 1959 indicate little change to the ice shelf until 2005, when a major break-up resulted in an area loss of 8.07 km2 (16.5%). Subsequent losses in summer 2008, 2011 and 2012 have resulted in a >60% decline in ice-shelf surface area. The satellite imagery reveals that each ice-shelf break-up event was preceded by open water conditions and resulting loss of pack ice pressure across the front of the ice shelf. Prior to 2005 the ice shelf was fringed by long-lived multi-year landfast sea ice (MLSI), which had been in place for >50 years. The removal of this MLSI in 2005 resulted in a switch to first-/second-year sea-ice conditions and seasonal open water in adjacent Yelverton Bay. It appears that MLSI losses and open water conditions in this region have been driven by record temperatures over the past decade (based on NCEP/NCAR climate reanalysis). Given the projected course of climate change, the shift from multi-year to seasonal landfast sea ice and the frequency of recent calving events, the long-term outlook for the Petersen Ice Shelf is poor, making it unlikely that this feature will survive for more than a decade.


Snow cover and short-term synoptic events drive biogeochemical dynamics in winter Weddell Sea pack ice (AWECS cruise – June to August 2013)

Jean-Louis Tison, Bruno Delille, Gerhard Dieckmann, Jeroen de Jong, Julie Janssens, Janne Rintala, Anne Mari Luhtanen, Nikolaus Gussone, Christiane Uhlig, Daïki Nomura, Véronique Schoemann, Jiayun Zhou, Gauthier Carnat François Fripiat

Corresponding author: Jean-Louis Tison

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

This paper presents the preliminary results of an integrated multidisciplinary study of pack ice biogeochemistry in the Weddell Sea during winter 2013. The sea-ice biogeochemistry group was one of the components of the AWECS (Antarctic Winter Ecosystem and Climate Study) cruise (Polarstern ANTXXIX-6). A total of 12 stations were carried out by the sea-ice biogeochemistry group, which collected a suite of variables in the fields of physics, inorganic chemistry, gas content and composition, microbiology, biogeochemistry, trace metals and the carbonate system, in order to give the best possible description of the sea-ice cover and its interactions at interfaces. Samples were collected in the atmosphere above (gas fluxes), in the snow cover, in the bulk ice (ice cores), in the brines (sackholes) and in the sea water below (0 m, 1 m, 30 m). Here we present the results of basic physico–chemical (T°, bulk ice salinity, brine volumes, brine salinity, Rayleigh numbers) and biological (Chla) measurements in order to give an overview of the general status of the Weddell Sea winter pack ice encountered, and discuss how it controls climate-relevant biogeochemical processes. Our results from the first set of nine stations, mainly sampled along the Greenwich meridian and the easternmost part of the Weddell Sea, definitively refute the view of a biogeochemically ‘frozen’ sea ice during the winter. This has already been demonstrated for the spring and summer, but we now see that sea ice sustains considerable biological stocks and activities throughout the winter, despite the reduced amount of available PAR radiation. Accretion of the snow cover appears to play an essential role in driving biogeochemical activity, through warming from insulation, thus favouring brine transport, be it through potential convection, surface brine migration (brine tubes) or flooding. This results in a ‘widening’ of the internal autumn layer (quite frequent in this rafting-dominated sea-ice cover) and increase of the Chla burden with age. Results from the second set of three stations in the western branch of the Weddell Sea gyre confirm that it comprises a mixture of older fast-/second-year ice floes with younger first-year ice floes. The older ice had the highest Chla concentrations of the entire cruise (>200 μg L–1), in an internal community enclosed within desalinized impermeable upper and lower layers. The first-year ice differs from that in the eastern Weddell Sea as it is dominated by


Factors controlling the strength of the sea-ice albedo feedback in the Arctic

Marika Holland, Laura Landrum

Corresponding author: Marika Holland

Corresponding author e-mail: mholland@ucar.edu

Arctic amplification of global warming is a consistent feature of climate model projections and has recently been identified in the observational record. However, the magnitude of the amplification signal varies considerably across different climate models. A primary mechanism driving Arctic amplification is the surface albedo feedback, in which modified snow and ice conditions lead to a reduction in albedo with consequent increases in surface shortwave absorption. There are a number of factors that contribute to changes in the surface albedo in ice-covered waters. These include changes in the surface properties of the sea ice, including the length of snow-free conditions and melt pond conditions, and the loss of the ice cover itself. Here we assess the relative importance of these various factors in transient simulations over the 20th–21st centuries. Comparison of simulations from the Coupled Model Intercomparison Project 5 (CMIP5) provide insight on the factors that control the strength of the sea-ice albedo feedback in models and how variations across the models contribute to uncertainties in future projections of Arctic change.


Observations and simulations of light transmission through springtime high-salinity first-year Arctic sea ice

Bonnie Light, Stephen Hudson, Mats Granskog, Regina Carns, Naomi Goldenson

Corresponding author: Bonnie Light

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

Optical measurements were made in April 2011 on first-year sea ice in Tempelfjorden, Svalbard. The spring ice was snow-covered and still retained significant salt (7 ppt average salinity). This ice is distinguished by the fact that it was not yet permeable and draining, and therefore still contained isolated inclusions, representative of FYI in the Arctic basin prior to melt onset. Average ice thickness was approximately 70 cm. We used four sets of instruments to measure light transmittance independently through the ice at three select sites. The instruments were Trios Ramses and Ocean Optics spectral sensors, Biospherical PUV 7-channel +PAR sensors, and Licor PAR sensors. All measurements at an individual site were made within a 3 hour window, used the same bore hole, and were positioned beneath the same target ice. The most informative measurements were made after the snow was cleared from a 3 m diameter area surrounding the target area. Comparison between the four instrument types shows some disagreement in the absolute transmittance, illustrating the difficulty in using separately calibrated sensors above and below the ice, but we were nevertheless able to use the measurements together with a model to further investigate the properties of the ice. A structural–optical model was used, along with microstructural imagery of thin sections from cores collected in the field and analyzed in the laboratory, to assess the vertical profile of light-scattering inclusions in the ice column. Companion core samples were extracted and segments were melted for analysis of absorbing constituents for the purpose of assigning the vertical variability of absorption coefficient. We observed significant variability in the vertical distribution of both scattering and absorption. These inherent optical properties were used, along with a 4-stream Discrete Ordinates radiative transfer model, to simulate radiative transport through the ice column and compare with observations. The calculations simulate the observed transmittances within the variability of the measurements. The magnitude of the scattering in late spring ice is larger than that observed for melting summertime first-year ice. This difference is likely attributable to the intact microstructure and isolated brine inclusions.


Biochemical and physical properties of multi-year and first-year sea ice in the Lincoln Sea

Benjamin Allen Lange, Christine Michel, Christian Haas, Guillaume Meisterhans, Justin Beckers, Alec Casey, Hauke Flores, Andrea Niemi

Corresponding author: Benjamin Allen Lange

Corresponding author e-mail: benjamin.lange@awi.de

The declining Arctic sea ice extent and replacement of multi-year ice (MYI) by first-year ice (FYI) represents a stunning loss of habitat for sea algae. Sea ice algae account for a large portion of the primary production in ice covered waters and are an important food source for many sympagic organisms. The contribution from ice algae to total primary production in a changing Arctic is uncertain and is partially due to a lack of biochemical and physical observations from the perennial sea ice zone. Here we compare the biochemical and physical properties of MYI and FYI from the Lincoln Sea, a region where the thickest ice in the Arctic is located and one of the last refuges of MYI. During three campaigns in May 2010, 11 & 12, duplicate cores were taken from 11 MYI and 7 FYI sites located in the Lincoln Sea. The biochemical cores were cut into 0.1 m sections and analyzed individually for chl a, phaeopigment, NO3, NO2, PO4 and Si. Texture classification was conducted on thick sections from the second core in addition to salinity, temperature and calculated brine volume. Core lengths ranged between 2.23–3.11 m for MYI and 0.83–1.77 m for FYI. Snow depths ranged between 0.07–0.47 m for FYI and 0.22–0.55 m for MYI sites. Mean integrated chl a for MYI cores was 0.93 mg m–2 (0.17–2.11) and for FYI cores was 0.71 mg m–2 (0.06–2.14). Mean integrated chl a from the bottom 0.20 m of the cores was 0.18 mg m–2 for MYI and 0.52 mg m–2 for FYI, which represented 20% and 46% of the total chl a content, respectively. Using the texture classification and salinity data we were able to determine the previous year's growth layer for 6 MYI cores. The multi-year portions of the cores had a mean length of 2.03 m (1.83–2.41), mean integrated chl a of 0.41 mg m–2 (0.23–0.78) and mean chl a concentration of 0.20 mg m–3 (0.11–0.32); the first-year portions of the cores had a mean length of 0.57 m (0.40–0.70), mean integrated chl a of 0.35 mg m–2 (0.07–1.22) and mean chl a concentration of 0.59 mg m–3 (0.12–1.75). These data represent a unique dataset and provide important information on the biological potential of FYI and MYI which will contribute to more representative projections of primary production in an Arctic covered by seasonal sea ice.


Thickness and optical properties of sea ice in the central Arctic in late summer

Marcel Nicolaus, Stefan Hendricks, Christian Katlein

Corresponding author: Marcel Nicolaus

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

The Arctic climate system is strongly modified through snow and ice cover changes, such as the thinning of sea ice, increasing light transmission and melt pond fractions, prolonging melt seasons, and a more seasonal ice cover. In order to understand and quantify these changes and to reveal the current status of Arctic sea ice, interdisciplinary observation programs are necessary. However, accurate and comprehensive in situ measurements of sea-ice mass balance and energy budgets are also challenging. Technological advances and big logistical efforts are needed to capture the strong spatial variability of sea-ice properties on various scales. Merging such datasets with satellite observations, re-analyses data and numerical models will finally improve Arctic-wide estimates of the sea-ice energy budget and mass balance. Here we present results from the expeditions ARK-XXVI/3 (TransArc, 2011) and ARK-XXVII/3 (IceArc, 2012) with the German research icebreaker Polarstern into the central Arctic Ocean. Both expeditions cover the season from late summer (August) to freeze-up (October). The sea-ice physics program, as part of the highly interdisciplinary research program, focused on the spatial distribution of sea-ice thickness and optical properties during both cruises. Key measurements were airborne sea-ice thickness measurements (EM-Bird), radiation measurements using a remotely operated vehicle (ROV), deployments of different autonomous instruments (buoys), and regular ice observations along the track based on the new ASSIST program.


How do sea-ice concentrations from operational data compare with passive microwave estimates? Implications for improved model evaluation and forecasting

Walter Meier, Florence Fetterer, J. Scott Stewart, Sean Helfrich

Corresponding author: Walter Meier

Corresponding author e-mail: walt.meier@nasa.gov

Passive microwave sensors have produced a 35 year record of sea-ice concentration variability and change. However, coarse spatial resolution and ambiguities in the retrieved emission limit the accuracy in key operational and forecasting regions such as thin ice conditions and near the ice edge. Operational analyses combine a variety of remote-sensing inputs and other sources via manual integration to create high-quality charts of ice conditions in support of navigation and operational forecast models. One such product is the daily Multisensor Analyzed Sea Ice Extent (MASIE) estimate distributed by the National Snow and Ice Data Center based on the Interactive Snow and Ice Mapping System produced by the US National Ice Center. It combines visible/infrared, passive and active microwave, and synthetic aperture radar (SAR) imagery to map the ice areas with >40% coverage at 4 km resolution. The higher spatial resolution along with the higher quality input data and manual analysis provide more precise mapping of the ice edge than passive microwave estimates. However, since MASIE is an extent product, it does not explicitly calculate sea-ice concentration, limiting its usefulness for integration with models and comparisons with passive microwave estimates. In addition, the operational focus is on mapping the ice extent, so regions of low concentration within the ice pack may not be mapped. In the recent low summer ice coverage in the Arctic, such regions are appearing more frequently and are important regions of heat and moisture exchanges potentially important for model processes. Here we derive a concentration product from MASIE by resampling to a 12.5 km grid used for the NASA EOS Advanced Microwave Scanning Radiometer (AMSR-E). Comparisons of the concentrations indicate that MASIE generally shows higher values, particularly near the ice edge. This is primarily due to surface melt during summer as well as small floes and thin ice near the edge that passive microwave sensors tend to underestimate, though open water areas unmapped by MASIE also contribute in some instances. Though less frequent, there are regions where MASIE indicates less ice than AMSR-E. These may be attributable to lags in MASIE analysis due to lack of coverage by input data sources. Such comparisons provide a better understanding of both operational and passive microwave sea-ice estimates that will allow sea-ice models to better incorporate the observations and improve model forecasts.


Observing snow depth on sea ice with a new affordable buoy type

Marcel Nicolaus, Sandra Schwegmann, Clifton Flint, Stefan Hendricks, Fabian Reiser, Andrew Mahoney, Mats A. Granskog, Marius Bratrein, Christian Katlein, Jölund Asseng, Mario Hoppmann, Martin Schiller, Lisa K. Behrens, René Fontes

Corresponding author: Marcel Nicolaus

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

Since sea ice is covered with snow for most of the year, snow depth is one of the crucial parameters for sea-ice research, and more generally for understanding the climate system in polar regions. Snow depth and physical properties determine the mass and energy balance of sea ice to a large degree, and the snow cover strongly influences most remote and autonomous observations and measurements. Most prominent examples are sea-ice thickness, surface albedo and light transmission, as well as surface properties (e.g. melt ponds) of sea ice. But at the same time, snow depth on sea ice is still one of the great unknowns. In order to gather more high quality seasonal data of snow depth from Arctic and Antarctic sea ice, a new buoy was developed. This snow depth buoy measures snow depth around its position by four sonic ranging sensors and transmits the data via iridium satellites. In addition, air temperature and pressure, and GPS position, are recorded and transmitted hourly. The buoy concept and design aim for low unit costs and easy deployment. Through this, we hope to deploy many of these buoys in international efforts to create a comprehensive dataset on snow depth, also from very remote regions. Here we present the buoy concept and design together with first results from different deployments on Arctic and Antarctic sea ice. Up to now, 12 prototypes were deployed during different seasons and in different snow and sea-ice conditions. In addition, the quality, reliability, processing and access of the snow buoy data are discussed. Future datasets are expected to increase our understanding of snow mass balance on sea ice, contributing to comprehensive datasets on snow depth on sea ice. Such data will also be a great contribution to satellite remote sensing and numerical model validation and improvements.


An analysis of acoustic backscatter measurements from 2003–10 made under landfast sea ice adjacent to the McMurdo Ice Shelf

Greg Leonard, Craig Stevens, Pat Langhorne

Corresponding author: Greg Leonard

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

Ice crystals have been observed floating freely in rivers, lakes, seas and oceans in the cryospheric regions of both hemispheres. In Antarctica their formation is commonly attributed to the presence of supercooled water, which is found on regional scales proximate to ice shelves. These freely floating ice crystals are difficult to measure directly in situ due to the nature of their environment, particularly once an ice cover has formed at the ice–water interface. Here we report on the use of volume backscatter strength from acoustic Doppler current profilers (ADCPs) that have been deployed through a sea-ice cover. Variation volume backscatter strength is typically associated with suspended sediment or biology. Here we consider the usefulness of the data as a proxy for measuring ice crystals in the water column. These measurements have been collected between 2003 and 2010 by a Nortek ADCP operating at a frequency of 1000 kHz (2003) and two RDI ADCPs operating at frequencies of 300 (2004, 2005, 2007, 2009 and 2010) and 75 (2010) kHz, respectively. Ancillary data including water salinity and temperature are used to assess the ability of the ADCPs to identify periods of increased number and/or size of ice crystals in the water column. Wavelet analysis is applied to the winter-long data records from 2003 and 2009 to identify whether observed variability in the ADCP volume backscatter signal strength changes over time. These changes are compared with the presence/absence of supercooled seawater, currents and light.


Sea-ice boundary conditions in a global green ocean modelling system NEMO

Martin Vancoppenolle, Olivier Aumont, Laurent Bopp, Gurvan Madec, Christian Ethé

Corresponding author: Martin Vancoppenolle

Corresponding author e-mail: martin.vancoppenolle@locean-ipsl.upmc.fr

Ocean biogeochemistry models typically treat sea ice as biogeochemically inert. However, observations from the last decade suggest a potentially important role of sea ice in the global biogeochemical cycles, promoted by (1) active biological and chemical processes within the sea ice; (2) fluid and gas exchanges at the sea–ice interface through an an often permeable sea-ice cover; and (3) tight physical, biological and chemical interactions between the sea ice, the ocean and the atmosphere. Our global green ocean model NEMO assumes by default that biogeochemical tracer concentrations are identical in the sea ice and in the ocean. In other models, the sea-ice concentration is either prescribed or set to zero. In this study, which is a first step towards understanding how sea ice may affect ocean biogeochemsitry, we investigate how some of the main large-scale simulated biogeochemical fields (chl-a, primary production, CO2 flux, ...) respond to perturbations in the tracer boundary conditions in the sea-ice zone. We found that in both polar regions the boundary condition on dissolved inorganic carbon (DIC) and alkalinity are of significant importance for the distribution and magnitude of the CO2 flux in polar regions. Iron in sea ice is highly important for primary production in the vicinity of the ice edge of the Southern Ocean sea-ice zone, whereas nutrients in sea-ice concentrations only slightly affect marine productivity.


Linking the distribution of under-ice fauna with environmental properties of sea ice: first results from a bi-polar field study

Hauke Flores, Benjamin A. Lange, Carmen David, Fokje Schaafsma, Jan A. van Franeker, Marcel Nicolaus, Ilka Peeken

Corresponding author: Hauke Flores

Corresponding author e-mail: hauke.flores@awi.de

In the polar regions, sea-ice habitats are undergoing rapid environmental change. Because sea ice constitutes an important substrate for numerous species, as well as an important carbon source during critical periods of the year, these changes have a significant impact on ecosystem functioning, biodiversity, species distribution and population sizes. Species dwelling at the ice-water interface (e.g. Antarctic krill Euphausia superba and Arctic cod Boreogadus saida) are assumed to play key roles in these ecosystems. As an important trophic carbon transmitter from the sea ice into pelagic food webs and ultimately to the deep sea benthos, under-ice fauna can contribute significantly to the carbon flux in polar ecosystems. Quantifying under-ice communities was hampered in the past by the inaccessibility of the ice underside to conventional sampling gear. Using a new under-ice trawl, it was demonstrated that Antarctic krill concentrates under sea ice almost year-round, and that krill dwelling under ice are often underestimated by pelagic nets and sonars. An Arctic expedition in 2012 brought evidence of a rich under-ice community even in the biologically poor-considered central Arctic Ocean. The environmental properties of sea-ice habitats which attract under-ice fauna, however, have been difficult to characterize in the past. During the 2012 expedition in the Arctic Ocean, we used a bio-environmental sensor array during under-ice fishing for the first time. The parameters measured included thickness, roughness and spectral light transmission of sea ice, as well as temperature, salinity and chlorophyll a content of the underlying water. The sensors enabled a real-time characterization of sea-ice habitat properties over large (1–5 km) distances. These measurements during fishing were complemented by fine-scale estimation of environmental sea-ice properties during ice stations. Statistical modeling techniques (GLMs, GAMs) were used to model the association of ecological key species with these habitat properties. We will present the progress of our group in linking biological and physical sea-ice data from the Arctic Ocean. This overview will be complemented by preliminary data from a recent winter expedition in the Southern Ocean (August–October 2013), illustrating the benefit of applying an identical multi-disciplinary approach in both hemispheres.


ICESat-2’s capabilities for sea-ice research

Thorsten Markus, Ron Kwok, Tom Neumann, Tony Martino, Sinead Farrell

Corresponding author: Thorsten Markus

Corresponding author e-mail: thorsten.markus@nasa.gov

Understanding the causes and magnitudes of changes in the cryosphere remains a priority for Earth science research. Over the past decade, NASA’s and ESA’s Earth-observing satellites have documented a decrease in both the areal extent and thickness of Arctic sea ice. Understanding the pace and mechanisms of these changes requires long-term observations of sea-ice extent as well as sea-ice thickenss. NASA’s Ice, Cloud and land Elevation Satellite (ICESat) mission, which operated from 2003 to 2009, pioneered the use of laser altimeters in space to study the elevation of the Earth’s surface and its changes. Among other contributions to the cryospheric sciences, ICESat proved adept at making centimeter-level elevation measurements enabling the determination of sea-ice thickness in the Arctic and in the Antarctic, and how that thickness distribution changed over time. Since ICESat stopped collecting data in October 2009, the IceBridge and CryoSat-2 missions continue these important observations. The well-documented and ongoing dramatic and rapid changes in the Earth’s ice cover have strengthened the need for sustained observations beyond what CryoSat-2 and IceBridge are expected to provide. The ICESat-2 mission is now under development for launch in late 2016. One of the primary objectives of the ICESat-2 mission is to enable estimate of sea-ice thickness from freeboard measurements to examine ice–ocean–atmosphere exchanges of energy, mass and moisture. Unlike ICESat, which had a single laser beam measuring 70 m footprints at 150 m along-track intervals, ICESat-2 has three pairs of beams, each pair separated by about 3 km across-track with a pair spacing of 90 m. The spot size is 10 m with an along-track sampling interval of 0.7 m. This measurement concept is a result of the lessons learned from ICESat. Most importantly for sea ice, the dense spatial sampling (eliminating along-track gaps) and the small footprint size are especially useful for sea surface height measurements in the often narrow leads needed for sea-ice freeboard and ice thickness measurements. Additionally the three pairs of beams provide significantly better spatial coverage. The talk with illustrate the concept of ICESat-2 and its potential for sea-ice research.


A first evaluation of the role of wave–ice interactions on the global mass balance

Martin Vancoppenolle, Thierry Fichefet, Steve Ackley, Hayley Shen, Francois Massonnet, Pierre Mathiot, Olivier Lecomte

Corresponding author: Martin Vancoppenolle

Corresponding author e-mail: martin.vancoppenolle@locean-ipsl.upmc.fr

Sea ice frequently forms in wavy waters. Wave motion packs forming ice crystals into small floes, while the ice attenuates the waves as the ice floes increase in diameter and thickness. Swell has been reported up to a few hundred kilometres inside the ice pack. Because of ocean waves, young ice floes take a rounded shape that led hungry early explorers to give them the name of pancake ice. Observations suggest that pancake ice thickness grows up to twice as fast as for ice forming in quiet seas. In this work we try to evaluate whether future large-scale sea-ice models should include wave–ice interactions to properly simulate large-scale distributions of ice concentration and thickness. In the large-scale 3-D ice–ocean modelling system NEMO-LIM, a representation of pancake ice formation is included. First, the ERA-40 ocean wave climatology is extrapolated in the sea-ice zone as if the ocean was ice-free. After diagnosing the simulated ice edge, ocean waves are propagated from the ice edge further inside the ice pack assuming exponential decay of amplitude. Finally, the thickness of newly forming ice is computed as a function of wave amplitude, as given by the equilibrium pancake ice theory. Wavelength is prescribed, which is a strong limitation of the model. In the model, pancake ice formation is found important in regions located in the vicinity of open ocean, namely the Southern Ocean and, in the Arctic, the Bering, Okhostk and Greenland Seas. Pancake ice formation accelerates the ice-edge progression, reduces winter ice concentration and, in turn, enhances ice production and thickness, in particular in the Southern Ocean. In some regions, the ocean responds to changes in ice production and modifies the location of the ice edge, as in East Antarctica. Wave–ice interaction parameters (wave attenuation, equilibrium pancake thickness, …) have a key impact on the simulated response of the model. Given the uncertainty in the model parameters, we conclude that more work is required to couple ocean waves and sea ice in large-scale models.


Ice-shelf-water-influenced ice roughness and ocean turbulence beneath landfast sea ice

C.L. Stevens, N.J. Robinson, M.G. McPhee, P.J. Langhorne, M.J.M. Williams, Inga Smith

Corresponding author: C.L. Stevens

Corresponding author e-mail: c.stevens@niwa.cri.nz

Under suitable conditions sea water flowing out from ice-shelf basal boundary layers is known to be at, or colder than, its in situ freezing temperature. This pressure-affected water promotes the growth of large platelet-like crystals, both freely drifting as well as attached to the underside of sea ice in the vicinity of the shelf. These crystals, often several centimetres or more in characteristic length, oriented at a range of angles and sometimes in layers over a metre thick, dramatically change the surface roughness of the ice-ocean interface. Here we describe observations from a variety of experiments near the combined Ross/McMurdo Ice Shelves. Instruments used include Doppler current profilers (profiles of bulk velocity structure), shear microstructure profilers (profiles of instantaneous variations in velocity at centimetre scale), acoustic Doppler velocimetry (time series of velocity in a single ~cm3 sample volume) and scalar fine structure (high frequency, ~ 4 Hz, time series of temperature and salinity at a single point). The approaches provide evidence of the form of the velocity gradient near the ice–ocean interface as well as the characteristic scales of mixing within that boundary layer. The measurements suggest that the presence of platelet ice on the underside of the sea ice substantially increases the roughness. This enhances drag on the ocean from the landfast ice as well as having a feedback effect on the vertical diffusion of the basal meltwater layer that drives the crystal growth. These roughness estimates are also likely relevant to regions of marine ice production on the underside of ice shelves.


Modelling biogeochemical tracer transport in sea ice due to gravity drainage

Andrew Wells, Joseph Hitchen

Corresponding author: Andrew Wells

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

Sea ice is a porous material formed of an evolving array of solid ice crystals bathed in liquid brine. The liquid-filled pore space provides a habitat for life within the ice and, when the ice is permeable, provides a pathway for exchange of gases and other chemicals between the ice, ocean and atmosphere. This coupling between the physical, chemical and biological evolution of sea ice has poorly constrained implications for biogeochemical processes, such as the impact of sea ice on the carbon cycle. During winter ice growth, so-called gravity drainage drives a convective exchange of brine between the ocean and the porous interior of sea ice. Here, we use two-dimensional mushy-layer simulations of convective flow to provide insight into the resulting transport of passive biogeochemical tracers through the ice. We quantify the chemical concentration in the liquid during periods of quasi-steady growth rate, and determine the total chemical tracer fluxes through the region of convection inside the ice. Chemical concentrations show spatial heterogeneity, and our results predict enhanced chemical concentrations in the ice near to brine channels. These results may provide useful insight for interpreting studies of sea-ice biogeochemistry, and offer a framework to develop models of physical, chemical and biological interactions.


Retrieval of the sub-ice platelet layer thickness with a ground-based multi-frequency electromagnetic device

Priska Hunkeler, Stefan Hendricks, Mario Hoppmann, Stephan Paul, Rüdiger Gerdes

Corresponding author: Priska Hunkeler

Corresponding author e-mail: priska.hunkeler@awi.de

Near Antarctic ice shelves, a dense layer of ice platelets is often observed underneath the immobile fast ice around the coastal margin. These ice platelets form in supercooled water as a result of ice-shelf basal melting. Individual ice platelets float upward and accumulate underneath the solid sea ice. This porous sub-ice platelet layer strongly modifies the total sea-ice mass and energy balance of Antarctic landfast and near-shore sea ice and forms a unique habitat for planktonic organisms. Besides its importance for sea ice, the platelet layer is also an accessible indicator for ice–ocean interaction of Antarctic ice shelves. The properties of this platelet layer have been studied widely on the landfast sea ice of McMurdo Sound, but little is known about its large-scale distribution in the Antarctic. This study presents results of a first field trial to distinguish the platelet layer from solid sea ice with electromagnetic (EM) induction sounding. The method is based on the contrast of electrical conductivity between ocean water, sea ice and the porous platelet layer. We acquired sea-ice and platelet layer thickness data with the multi-frequency electromagnetic device GEM-2 during a field campaign at Atka Bay in the eastern Weddell Sea, Antarctica, from November 2012 to January 2013. Assuming resistive sea-ice (0 S/m), conductive ocean water (2.7 S/m) and a diversity of values for the platelet layer (0–2.7 S/m), the electrical conductivity of this layer can be estimated by comparing the measured EM response to theoretically calculated EM data at locations with known sea-ice and platelet layer thicknesses. Furthermore, we use 1-D inversion techniques to simultaneously estimate the thicknesses and conductivities of these different layers. Data from simultaneous drillings, under-ice cameras and CTD profiles are used to validate the EM data. With this study, we lay the foundation for future large-scale airborne electromagnetic surveys acquiring data of sub-ice platelet layer thickness and extent.


Increased Arctic ice pack deformation in response to 2007 ice loss, related to persistent low sea-ice minima in 2008–12

Jennifer Hutchings, Kim Martini, Ignatius Rigor

Corresponding author: Jennifer Hutchings

Corresponding author e-mail: jenny@iarc.uaf.edu

Over the last decade the Arctic ice pack has decreased in extent and volume, with extreme minima at the end of summer 2007 and 2012. Summer 2007 was the culmination of ice-transport-driven changes in ice age distribution, with thinner young ice melting earlier in summer, enhancing albedo feedback. This has primed the ice pack to be more responsive to surface stresses, increasing ice drift and deformation rates. We have developed a new drifting-buoy ice deformation product from International Arctic Buoy Program (IABP) ice drifter position data. We assess changes in divergence and shear rate for the central Arctic, Beaufort, Chukchi and Lincoln Seas over the last decade. There is a marked change in deformation rate in 2006 and 2007 that coincides with transport-driven reduction in ice age throughout the Arctic. This increase in maximum shear strain rate and increase in variance of total deformation rate persists through 2012. It is not associated with changing wind stress magnitude, and an increase in wind factor suggests that the ice pack has increased its mechanical response to wind stress. We present a case study where enhanced divergence was observed in the eastern Beaufort Sea, which is sufficient to explain enhanced basal ice melt. The new IABP product indicates that low summer ice extents from 2008 through 2012 are an ice pack response to the 2007 minimum. A thermodynamic-dynamic feedback being in effect whereby a younger more mechanically responsive ice pack experiences increased divergence rates in winter, growing increased area of lead ice that melts out early in summer and preconditioning the pack for enhanced albedo feedback and solar absorption into the upper ocean.


Seasonality and spatial distribution of solar radiation under Arctic sea ice

Stefanie Arndt, Marcel Nicolaus

Corresponding author: Stefanie Arndt

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

Arctic sea-ice extent decreased considerably along with the ice cover becoming thinner and more seasonal during the last decades. These observed changes have a strong impact on interactions between atmosphere and ocean and thus play a major role in Earth’s climate system. Until now, it has not been possible to quantify shortwave energy fluxes through sea ice sufficiently well over large regions and during different seasons. In order to obtain Arctic-wide estimates of solar radiation under sea ice, new methods are necessary. Here, an upscaling method combining a newly developed parameterization of light transmittance and remote-sensing and reanalysis data is presented. The main result suggests that 96% of the total annual solar heat input under Arctic sea ice occurs in the time from May to August, hence in the course of only 4 months of the year. Sensitivity studies indicate that once the melt season begins 2 weeks earlier, an increase by 20% of the total annual solar heat input through sea ice is shown. Therefore, the transition period from spring to summer, particularly the timing of the melt season, substantially affects the light availability under ice. Furthermore, a more seasonal ice cover and a higher melt pond coverage lead to a higher fraction of solar radiation being transmitted through the sea ice in summer. This positive correlation between enhanced melting and increasing transmittance can be described as ‘transmittance-melt feedback’. Assuming an ongoing ice thinning, the transmittance-melt feedback results in a further increase in transmitted and absorbed heat fluxes. Changes in timing and amount of light penetrating through Arctic sea ice might also influence melt season, biological and geochemical processes, as well as basal and internal melt and freeze rates. These positive feedbacks affect the mass and energy budget of sea ice and alter crucially the interaction of atmosphere and the upper ocean.


Developing algorithms to detect and quantify sea-ice melt

Stefanie Arndt, Marcel Nicolaus, Wolfgang Dierking, Mario Hoppmann, Sandra Schwegmann

Corresponding author: Stefanie Arndt

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

The mass and energy balance of sea ice are strongly connected through the transfer of solar (shortwave) radiation from the atmosphere through snow and sea ice into the ocean. In order to estimate sea-ice melt and freezing rates, we need (1) to improve our understanding of the radiative transfer into and through Arctic and Antarctic sea ice and its impacts on sea-ice melt, and (2) to improve existing and develop new remote-sensing data products to allow for large-scale estimates of solar radiation fluxes in and under sea ice. Recent studies show that a major uncertainty in sea-ice mass balance is related to the timing and duration of the melt season as well as the very limited knowledge of the snow layer on top. Therefore, we are working on improving the existing data products, resulting from remote-sensing and re-analyses data, for the Arctic, and we are developing a new comparable product for Antarctic sea ice. From this work, we finally expect to elevate our understanding of thermodynamics for both hemispheres. In particular, the reflectivity, absorption and transmittance of sunlight through snow and sea ice and their effects on the sea-ice mass balance are studied. Here we present first results and the concept to develop a melt and freeze onset dataset for Antarctic sea ice. This development is based on passive microwave and (when available) scatterometer satellite data. In addition, we use in situ observations and autonomous measurements (buoy data), e.g. solar radiation over/under sea ice and snow height, from recent Antarctic campaigns during different seasons (Polarstern Weddell Sea cruises, sea-ice monitoring program at the German wintering base Neumayer III) to develop and validate the new algorithms.


Sea-ice thickness variability in Fram Strait between 2003 and 2012 from in situ and airborne observations

Angelika Renner, Sebastian Gerland, Christian Haas, Edmond Hansen, Justin Beckers, Gunnar Spreen, Marcel Nicolaus, Harvey Goodwin

Corresponding author: Angelika Renner

Corresponding author e-mail: angelika.renner@npolar.no

Fram Strait, situated between Greenland and Svalbard, is the main export route for sea ice from the Arctic basin. The sea-ice cover in Fram Strait therefore provides an integrated signal of processes happening upstream in different regions of the Arctic, but is also subject to the highly dynamic regional environment. We present results from sea-ice thickness measurements conducted in 2003–12 as part of the Norwegian Polar Institute’s long-term monitoring program, supplemented with data from process studies and campaigns during the International Polar Year. In all years, fieldwork took place along a transect at 79°N in late summer at the very end of the melt season and at the time of the minimum ice extent. In 2005, 2007 and 2008, additional campaigns were carried out in spring. Total snow and ice thickness was measured by drillings and ground-based electromagnetic sounding (EM); in spring 2005 and 2008, and since 2010, surveys were also conducted using helicopter-borne EM. The late-summer ground measurements show a distinct decline in mean and modal ice thickness of 0.2 and 0.3 m a–1 with long-term averages of 2.5 and 3.0 m, respectively, in line with similar trends derived from moorings. The less extensive spring ice thickness data indicate more variability between years. The helicopter-borne measurements cover a much larger area than the ground EM observations; we therefore use them to assess spatial variability of the drift ice. In spring, we find an east–west gradient in thickness, with thicker ice towards Greenland. However, this gradient is not present in late-summer data. We explore the links to local, regional and remote dynamic and thermodynamic processes that influence the ice thickness distribution to explain the presence or absence of such gradients. In particular, we look at how ice age and ice drift and the observed changes in the Arctic impact the sea ice in Fram Strait.


Light transmittance by a ponded first-year Arctic sea-ice cover

Bonnie Light, Donald Perovich, Melinda Webster

Corresponding author: Bonnie Light

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

Melt ponds are ubiquitous on the Arctic sea-ice cover during the summer melt season. From the perspective of understanding the fate of incident shortwave radiation on the ice and underlying ocean, the impacts of ponds are significant. Ponds backscatter only a fraction of the incident radiation backscattered by bare ice. They also transmit much more light to the ocean below. The representation of melt ponds in climate models is an ongoing challenge. The small physical extent of individual ponds requires them to be described using large-scale averages. Such parameterization needs to accurately describe how ponded ice responds to large-scale forcings, solar radiation in particular. Light propagation through ice is greatly enhanced by individual ponds. And the propagation of light through ponds is influenced by the surrounding ice, making this a complex three-dimensional radiative transport puzzle. In an effort to unravel these mysteries, measurements were made during the 2010/11 ICESCAPE field campaigns on the optical properties of individual melt ponds. At selected sites, through-ice light transmittance measurements were made beneath ponded ice, capturing the transition from pure bare ice to pure pond, and the boundary between. Results from this study illuminate questions about the need for knowledge about melt pond size distributions when parameterizing shortwave radiative transfer for melting sea-ice covers. Typically, ponded ice transmits two to four times more visible light to the ocean, as compared with neighboring bare ice. Some of the questions under investigation include whether the optical properties of interior ponded ice are modified by this higher light transmittance, the extent to which light admitted by a pond propagates into the interior of the neighboring bare ice, and the possibility for quantifying the spatial distribution of beneath-ice light fluxes under a ponded ice cover.


Using bathymetric data to predict anchor ice formation sites in McMurdo Sound, Antarctica

Sarah Mager, Gregory Leonard, Inga Smith, Andrew Pauling

Corresponding author: Sarah Mager

Corresponding author e-mail: sarah.mager@otago.ac.nz

A unique feature of polar seas is the formation of ice clusters attached to the seabed. This ‘anchor ice’, as it is widely known, plays an important role in mobilizing bed sediments, as well serving ecological roles as a habitat and potentially fatal environment. From the few published observations of Antarctic anchor ice, its extent is rarely observed at depths greater than 33 m, with the largest concentrations observed between 5 and 20 m. As such the sublittoral zone associated with the landward margin represents the most likely environment for anchor ice formation, in areas where supercooled ice-shelf water (ISW) has advected from beneath an ice shelf and intersected the sublittoral zone. Here, we present a preliminary spatial model that predicts areas where anchor ice is likely to form in McMurdo Sound using the International Bathymetry Chart for the Southern Ocean (IBCSO) version 1. The spatial model parameters are informed by oceanographic observations and numerical model predictions for McMurdo Sound. From this spatial model, areas conducive to anchor ice formation have been identified, with the largest contiguous area being beneath the McMurdo Ice Shelf (MIS) and extending along Brown Peninsula and Minna Bluff. Anchor ice is also predicted to form along the Hut Peninsula, along pockets of the coast of southern Victoria Land, including Explorers Cove and the Bay of Sails. These predictions are consistent with fossil deposits that have been interpreted as evidence of anchor ice on the MIS, as well as direct observations from diving in the Sound. The limitations of our spatial model include poorly constrained sub-ice-shelf bathymetry and cavity circulation, as well as areas where the terminus and grounding line of floating ice tongues are not well represented in IBCSO (e.g. Koettlitz Glacier). Our spatial model also assumes a northerly flow of ISW discharging from the Ross and McMurdo Ice Shelves, which is seasonally variable, and may explain the episodic observations of anchor ice in McMurdo Sound.


Trophic relationships under sea ice: stable isotope- and DNA-based dietary study in early spring off East Antarctica

Zhongnan Jia, Kerrie M. Swadling, Simon N. Jarman, Patti Virtue, Bettina Meyer, So Kawaguchi, Klaus Meiners

Corresponding author: Zhongnan Jia

Corresponding author e-mail: Zhongnan.Jia@utas.edu.au

Trophic interaction of the sea-ice food web is fundamental knowledge needed for us to understand functions of the Antarctic food web. Furthermore, it helps us to predict how the Antarctic ecosystem might respond to environmental changes. We examined δ13C and δ15N signatures of eight common zooplankton species, together with particulate organic matter (POM) from both the sea ice and the underlying water column in East Antarctica (110–130°E) during two Sea Ice Physics and Ecosystem eXperiments (SIPEX and SIPEX-2012). Interpretations of trophic relationship were helped by comparison with DNA-based dietary analyses on two euphausiid species Euphausia superba and Thysanoessa macrura. The δ13C signatures suggested that the pteropod Limacina helicina relied mainly on sea-ice POM as a dietary source, while the other seven species consumed POM-based food sources from the water column. The δ15N signatures indicated a very clear stepwise enrichment from POM to carnivores. Euphausia superba and T. macrura had very similar δ15N profiles (5.65‰ and 5.53‰, respectively), and were both about 2‰ higher than the other herbivores, highlighting an omnivorous diet for both species. Carnivores (Euchaeta antarctica and Sagitta marri) had more than 5‰ higher δ15N values than the other zooplankton species. These results illustrated a complex food web in the zooplankton community, with pelagic species using both sea-ice and under-ice water column organic matter as food sources. This study highlights the importance of the sea-ice zone to krill and other zooplankton during the winter/early-spring transition.


Forecast of summer sea-ice extent in the Arctic based on winter ice redistribution

Noriaki Kimura, Yuya Nakano, Hajime Yamaguchi, Takaya Uchida, Eiji Hirokawa, Ryosuke Funabashi, Yo Nakahara, Takuya Wada

Corresponding author: Noriaki Kimura

Corresponding author e-mail: kimura@1.k.u-tokyo.ac.jp

Summer sea-ice cover in the Arctic varies largely from year to year due to several factors. We examined one of these factors, the relationship between interannual differences in winter ice motion and ice area in the following summer. Daily ice velocity products on a 37.5 km resolution grid are prepared using the satellite passive microwave sensor Advanced Microwave Scanning Radiometer for EOS (AMSR-E) and AMSR2 data for 2003–11 and 2013, respectively. Derived daily ice motion reveals the dynamic modification of the winter ice cover. The winter ice divergence/convergence is strongly related to the summer ice cover in some regions; the correlation coefficient between the winter ice convergence and summer ice area ranges between 0.5 and 0.9 in areas with high interannual variability. This relation implies that the winter ice motion and resulting redistribution of sea ice is one of the important factors to decide summer ice extent. Assuming the linear relation between the winter ice convergence and summer ice area, we predicted the ice extent during 1 July and 1 November 2013 using the ice velocity data until the end of April. The accuracy of this prediction was examined through comparison between predicted and AMSR2-derived ice concentration. Additionally, the process controlling the summer ice retreat was investigated by considering the contribution of dynamic ice drift and thermodynamic ice melting. The process varies with region and controls the accuracy of the ice prediction.


Effects of different footprint areas on the comparability between measurements of sea-ice freeboard

Sandra Schwegmann, Stefan Hendricks, Robert Ricker, Giulia Castellani, Christian Haas, Andreas Herber

Corresponding author: Sandra Schwegmann

Corresponding author e-mail: Sandra.Schwegmann@awi.de

The significant loss of Arctic sea ice during the last decades shows the sensitivity of the sea-ice system to changes in global climate. To distinguish between natural variability and the impact of global warming, an understanding of processes and feedbacks is necessary, and for that consistent and comprehensive measurements of the most important sea-ice properties are required. While sea-ice concentration is observed routinely year-round since the beginning of the satellite era, strategies to investigate the sea-ice thickness distribution have been developed only recently. These are crucially required for an examination of sea-ice mass changes. Presently, ice thickness observations are mainly based on freeboard measurements by means of satellite laser and radar altimetry. To contribute to the interpretation of these sea-ice thickness products, available airborne thickness and freeboard data were collected within the Sea Ice Downstream Services for Arctic and Antarctic Users and Stakeholders (SIDARUS) EU-Project, and have been analyzed with respect to their usability for the validation of the large-scale satellite products. One major challenge in comparing satellite and airborne measurements is the different footprint area of these methods. Therefore, statistical parameters like the variability of freeboard within the common footprint areas have been analyzed from measurements made during the PAMARCMIP 2011 campaign in order to determine the differences between point measurements and areal averages. It turned out that mean freeboard is less dependent of the freeboard areas than the modal values are. Furthermore, differences in modal and mean values for the range of chosen footprint areas have been related to their dependency on different ice characteristics and length scales of freeboard and thickness profiles. Also for this comparison, mean values are found to be more representative, as the results are mostly independent from the footprint but depend more on the length scales. However, there was no length scale that was representative for all the observed regions. Finally, results are used for the interpretation of the comparison between freeboard and sea-ice thickness data derived from different data sources using laser and radar altimetry.


Properties of snow on landfast sea ice of Atka Bay, Antarctica, and its influence on X-band SAR backscatter

Stephan Paul, Sascha Willmes, Mario Hoppmann, Priska Hunkeler, Günther Heinemann, Marcel Nicolaus

Corresponding author: Stephan Paul

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

High-resolution information about the snow cover on sea ice remains a key problem for satellite-based investigations. Detailed knowledge of snow microphysical properties and especially snow thickness is crucial for the understanding of the mass and energy balance of snow-covered sea ice. Furthermore, high-resolution snow data enhances ice thickness retrievals and/or increases their interpretability. But ground-truth validation data for SAR imagery are rare and the interaction of snow with X-band backscatter is still not well understood. We present results of our field campaign in Atka Bay, Antarctica, between November 2012 and January 2013. The spatial and temporal variability of high-resolution TerraSAR-X backscatter data for Antarctic landfast sea ice in the transition period from austral winter to spring/summer is shown. First analyses suggest that the gross of the backscatter signal variations is influenced by the variability in the signal penetration depth that is due to snow microphysical properties. This yields the possibility of separating different sea-ice surface classes and estimating the snow thickness of cold and dry snow. With the onset of melt these differences equal out due to a higher attenuation of radar waves by meltwater. The small-scale spatial differences in the backscatter signal are correlated to in situ measured snowpack stratigraphic parameters such as snow density, grain size distribution and internal layer structure.


Advances in electromagnetic induction sounding of sea-ice thickness

Stefan Hendricks, Priska Hunkeler, Andreas A. Pfaffhuber, Malte Vöge, Christian Haas

Corresponding author: Stefan Hendricks

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

Electromagnetic induction sounding is a commonly practiced method of sea-ice thickness retrieval. The main advances are the capability to directly measure snow plus ice thickness without the necessity for other significant input parameters, the applicability to all sea-ice types and the feasibility to carry out airborne long-range surveys or ground-based high-resolution studies. Current sensor systems however are optimized for robust detection of level-ice thickness and it has been long known that the simplified representation of sea ice in existing processing schemes leads to an underestimation of the maximum thickness of sea-ice pressure ridges. In similar fashion, internal properties of sea ice, e.g. higher electrical conductivity due to salt water intrusion, are not taken into account. Since airborne electromagnetic (AEM) sea-ice thickness surveys are a vital source of independent validation data for altimetry missions such as CryoSat-2 or sea-ice models, it is important to know what bias present 1-D-1 layer assumptions in data processing impose on the mean AEM sea-ice thickness on the regional scale. Therefore we will present results from numerical studies where we simulated the EM response of complex sea-ice geometries and compared the 1-D result to the preset of the model. These modeling studies stimulate the use and development of more advanced EM sensors that use several frequencies and partly additional receiver configurations to allow 1-D multi-layer or possibly 2-D interpretation of the EM response. We show first results of ground and airborne multi-frequency EM data as well as new processing efforts using inversion techniques. We have tested commercially available sensors for ground-based EM measurements (Geophex, GEM-2) and custom developments for airborne EM (NGI, MAiSIE = Multi-Sensor Airborne Sea Ice Explorer) in order to assess their potential for an advanced and more accurate determination of sea-ice thickness and internal properties.


Comparison of laser altimeter and laser scanner observations from an airborne digital electromagnetic induction device over Arctic sea ice

Justin Beckers, Angelika H.H. Renner, Sebastian Gerland, Christian Haas

Corresponding author: Justin Beckers

Corresponding author e-mail: beckers@ualberta.ca

The thickness and extent of sea ice and the amount of multi-year sea ice present in the Arctic have declined over the past two decades. While extent and ice age are readily monitored from space, ice thickness and surface roughness remain difficult to measure. Furthermore, as the Arctic sea-ice cover becomes increasingly seasonal, new sea-ice surface roughness and thickness data are required in order to provide accurate input for global climate models, for future sea-ice forecasts, and for safe maritime operations in icy waters. In situ roughness and thickness observations are needed for calibration of satellite data for improved pan-Arctic ice thickness estimates. Airborne electromagnetic induction systems (AEM) provide accurate, high-resolution measurements of sea-ice thickness and surface roughness over large areas. Traditionally a laser altimeter is used to determine the height of the sensor above the sea-ice or snow surface. However, the footprint of an AEM system is several times the height of sensor above the conductive water layer, several orders of magnitude larger than the laser altimeter footprint. In 2010, ice thickness surveys were conducted in Fram Strait and north of Svalbard using a new AEM system that included both a point laser altimeter and an across-track laser scanner. A comparison between coincident altimeter and scanner measurements illustrates the importance of accounting for across-track surface roughness information when retrieving ice thickness or estimating spatial properties of sea-ice surface roughness using the AEM system. Using image-processing techniques the dimensions, orientation and spatial distribution of ridges with respect to the flight direction can be retrieved from the scanner. Preliminary results also suggest that the laser scanner may provide new information on the dimensions and distribution of melt ponds. However, specular reflections of the laser beam away from the sensor over some leads and melt ponds caused missing laser altimeter and scanner data. Finally, as the AEM system did not include an inertial navigation unit in 2010, the ability of the laser scanner to estimate the AEM sensor roll angle is presented.


Semi-automated mapping of Antarctic sea ice using Envisat SAR images

Barry Giles, Glenn Hyland, Robert Massom, Petra Heil

Corresponding author: Barry Giles

Corresponding author e-mail: Barry.Giles@utas.edu.au

Feature tracking of sea ice using cross-correlation methods on pairs of satellite synthetic aperture radar (SAR) images has been carried out extensively in the Arctic but not widely in the Antarctic. This is due to the dynamic nature of Antarctic pack ice, less contrast in its microwave signature (compared with Arctic sea ice) and inadequate SAR acquisition sampling across the region. Given these limitations and challenges a semi-automated system, known as IPADS (IMCORR (IMageCORRelation) Processing, Analysis and Display System), has been developed to map both pack ice motion and fast-ice distribution around Antarctica using multiple overlapping pairs of SAR images. The software processing pipeline initially registers images using only information contained in the image headers. Several recent key improvements are presented. To increase computational efficiency a large byte-mask map of Antarctica has been created, based on the NSIDC MOA Antarctic coastline, to exclude non-sea ice from the processing and to specify an offshore zone within which stationary fast ice may exist. The outer edge of this zone is a slightly extended and smoothed version of the 2000 m ocean depth contour. Since precise positioning is important for accurate mapping of coastal fast ice, the geolocation of all images is refined by matching them against the accurately located mask coastline. Pack ice is identified and motion/displacement computed using a cluster-based search method which compares both location and motion information. A hierarchical approach is used to ‘intelligently’ expand the search around already identified features to locate additional correlated structure. Using this method, substantial pack ice motion can be detected in SAR images up to 3 days apart and can also often be detected in the highly dynamic marginal ice zone in images up to a day apart. Fast-ice maps are generated using zero-motion features located within the fast-ice mask zone. All analysis occurs as an off-line batch process, but a viewing tool allows the interactive interrogation of the output NetCDF files. These files form a constantly growing database of Antarctic pack ice motion and fast-ice extent but underline the crucial need for improved satellite SAR coverage around the continent.


A record of Antarctic sea-ice extent in the southern Indian Ocean for the past 300 years and its relationship with global mean temperature

Cunde Xiao, Sharon Sneed, Ian Allison, Runxiang Li, T.F. Dou

Corresponding author: Cunde Xiao

Corresponding author e-mail: cdxiao@lzb.ac.cn

The differing response of ice extent in the Arctic and Antarctic to global average temperature change over approximately the last three decades highlights the importance of reconstructing long-term sea-ice history. Here, using high-resolution ice-core records of methanesulfonate (MS-) from the East Antarctic ice sheet in Princess Elizabeth Land, we reconstruct southern Indian Ocean sea-ice extent (SIE) for the sector 70°E–100°E for the period 1708–2000 AD. Annual MS- concentration positively correlates in this sector with satellite-derived SIE for the period 1973–2000 (p < 0.05). The 293 year MS- record of proxy SIE shows multi-decadal variations, with large decreases occurring in two warm intervals during the Little Ice Age and during the 1940s. However, after the 1980s there is a change in phase between Antarctic SIE and global temperature change, with both increasing. This paradox is probably attributable to the strong anomaly in the Southern Annular Mode (SAM) in the recent three decades.


Winter sea-ice thicknesses in the Weddell Sea and their variability over the past 24 years

Sandra Schwegmann, Priska Hunkeler, Stefan Hendricks, Peter Lemke, Christian Haas

Corresponding author: Sandra Schwegmann

Corresponding author e-mail: Sandra.Schwegmann@awi.de

The sea-ice thickness distribution is one of the most important sea-ice properties, but also one of the less frequently observed ones so far. Satellite retrievals of Antarctic sea-ice thickness are currently limited to laser and radar altimetry observations of snow freeboard with large uncertainties, and electromagnetic measurements have been obtained only sporadically. For the investigation of the variability and changes in the sea-ice thickness distribution over the last decades, data are mainly available from very basic methods such as drilling or ship-based observations following the Antarctic Sea Ice Processes and Climate (ASPeCt) protocol. Thereby it is an advantage that those data also include information on the snow-depth and partly on the sea-ice freeboard distribution, which are as sparse as information on the sea-ice thickness distribution, in particular during winter conditions. The most recent data based on those methods were obtained during austral winter 2013, when various sea-ice parameters were measured in the Weddell Sea as part of the Antarctic Winter Ecosystem Climate Study (AWECS). Here, we present first results of the sea-ice thickness, freeboard and snow-depth distribution obtained by drillhole measurements and ship-based observations from this expedition. The new dataset is compared with results from three previous winter campaigns done in 1989, 1992 and 2006 in the Weddell Sea in order to determine the long-term variability of sea-ice thickness, snow depth and freeboard. A challenge in comparing all those data is that measurement sites are based only on individual floes, which are expected to be representative for an entire region. In addition, sampling rates differ between the considered field experiments. Therefore, drillhole thicknesses are cross-correlated with ground-based EM-measurements in order to identify for the newest dataset, how representative the chosen study areas have been for the respective sea-ice floes and which consequences different measurement spacing has for the comparison of data from different years.


Investigating the relationship between chemical signals and snow accumulation rate from the Mill Island ice-core record and sea-ice extent in East Antarctica

Mana Inoue, Mark A.J. Curran, Andrew D. Moy, Tas D. van Ommen, Helen E. Phillips, Ian D. Goodwin

Corresponding author: Mana Inoue

Corresponding author e-mail: Mana.Inoue@utas.edu.au

High-resolution climate records from ice cores can be used to calibrate instrumental data and to interpret past climatic conditions. In previous study, a significant correlation was observed between the Law Dome ice-core record of methanesulphonic acid (MSA) and sea-ice extent for the area of 80° E to 140° E. In the present study, we investigate the interaction between Mill Island ice core records; including MSA, snow accumulation, and oxygen isotopes; with sea-ice extent changes in Shackleton Ice Shelf region and with sea-ice extent in the wider regional area. Mill Island is 500 km west of Casey station and has a mean snow accumulation rate of 1.01 m a–1 ice equivalent. The record in Mill Island ice core covers the period 1913 to 2009. Since around the 1960s, we observe significant changes in snow accumulation and oxygen isotope data (a proxy for temperature) from the Mill Island ice core. The relationship between this trend and sea-ice changes will be discussed.


Antarctic sea-ice response in ensemble CMIP5 historical and ozone perturbation simulations

Siobhan O’Farrell, Philip Reid, Robert Massom

Corresponding author: Siobhan O’Farrell

Corresponding author e-mail: siobhan.ofarrell@csiro.au

The net Antarctic sea-ice extent exhibits a small statistically significant positive trend since 1979 in contrast to the decay in the Arctic sea-ice cover. This overall increase has been attributed to strengthening of westerly winds (positive SAM) as a result of ozone depletion; trends in ice drift related to wind trends; increasing stratification of the ocean from enhanced sea-ice melt and/or ice shelf melting; and natural variability of the climate system. The CMIP5 models provide key data to investigate several of these theories; in particular the CSIRO Mk3.6 model has a 10-member historical ensemble in which the individual sea-ice concentration members show variation in the net trend in recent ice coverage (1976–2005). Here we aim to better define why the individual differences arise in the individual ensemble members by assessing the atmospheric drivers of the sea-ice change, the sea-ice transport and divergence at monthly/seasonal timescales. We also investigate the changes in ocean temperature, salinity and the ocean stratification in the individual members. As part of CMIP5, several modelling groups undertook either ozone-only or all-forcing historical runs that omitted recent ozone perturbations. CSIRO Mk3.6 undertook the latter, with a 10-member ensemble; these runs are close to the present-day signal so can readily be compared to the existing historical runs and observations. These data have been investigated to ascertain if the ozone signal impacts the ensemble spread of the results, and how individual ensemble members respond compared to the all-forcing case. Preliminary results indicate that this ensemble also had members showing net ice cover increases, whilst others showed decreases, suggesting a similar variation to the original ensemble. Further analysis will focus on whether the sea-ice response in individual Antarctic sectors resembles the recent observed changes using satellite-derived ice concentration (from 1979 to present) and the meteorological reanalysis data, and will build on existing studies of the co-authors. A similar set of experiments has been just been completed for the ACCESS 1.0 and 1.3 model with the ozone distribution at high southern latitudes remaining held at 1960 values until 2020. Although the ensemble size is smaller (three members, which is the same size as the existing historical ACCESS ensembles as the model is more computational expensive) we hope that the ensemble size is sufficient to detect a signal.


Factors driving pCO2 dynamics in sea ice during a large-scale ice tank experiment

Jiayun Zhou, Bruno Delille, Jean-Louis Tison, Riitta Autio, Gerhard Dieckmann, Karl-Ulrich Evers, Linda Jørgenesen, Hermanni Kaartokallio, Gerhard Kattner, Hilary Kennedy, Marie Kotovitch, Harri Kuosa, Anne-Mari Luhtanen, Colin Stedmon, David Neville Thomas

Corresponding author: Jiayun Zhou

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

According to previous studies, pCO2 fluxes measured over Arctic sea ice are higher than those measured over Antarctic sea ice. We hypothesized that this was due to enhanced respiration in Arctic sea ice, as a consequence of higher riverine inputs of dissolved organic carbon (DOC) into Arctic seawater. We tested this hypothesis during the Interice V experiment at the HSVA (Hamburg) environmental test basin facility. We reproduced the growth and decay cycle of sea ice in replicate mesocosms (1 m3) filled with North Sea water (NSW series), and compared these with another series of mesocosms to which humic-rich river water had been added (10%) to increase the DOC concentration (R series). Primary producers were excluded from the experiment. The evolution of the temperature, salinity, DOC, pCO2 and bacterial biomass and production were measured in ice sampled at regular intervals throughout the experiment, as well as in the under-ice water. In addition, ice–air pCO2 fluxes were continuously monitored over both NSW and R mesocosms. pCO2 values in ice were higher in the R ice than in the NSW ice. This is attributed to the DOC content and bacterial respiration, rather than to the ice physical properties (i.e., ice permeability constrained by the ice temperature and salinity). Indeed, R ice had higher DOC content and bacterial production than the NSW ice while both showed similar physical properties. The evolution of the ice–air pCO2 fluxes was consistent with the evolution of pCO2 in ice. The fluxes were, as expected, positive (from sea ice to the atmosphere) during ice growth and negative (from the atmosphere to the ice) during ice melt.


Nutrient uptake rates and primary production in East Antarctic sea-ice (SIPEX 2 results)

Arnout Roukaerts, Anne-Julie Cavagna, Frank Dehairs, François Fripiat, Delphine Lannuzel, Klaus Meiners

Corresponding author: Arnout Roukaerts

Corresponding author e-mail: arnout.roukaerts@vub.ac.be

Sea ice is an important component of Earth’s climate system and a structuring force in Antarctic marine ecosystems. It also plays a crucial role in primary production and biogeochemical cycles of the Southern Ocean. The SIPEX 2 expedition took place in the East Antarctic sector (63–66°S, 115–125°E; R/V Aurora Australis; Sept.–Oct. 2012) and the main objective was to investigate interactions between the physical sea-ice environment and biogeochemistry. In this context, our work focused on improving the understanding of nutrient cycling and primary production in the seasonal drift ice. Various nutrient uptake rates (HCO3, NO, NH4+ and H4SiO4) were measured using a new field methodology. Our approach consists in studying both, bottom and internal communities using enriched stable isotope substrates (13C, 15N, 30Si) and in situ incubations. Results show a clear increase in primary production during the one month course of the expedition, starting in early spring. Between the first and last sampling station, daylight length increased by approximately 3 h. Although we observe an increase in uptake rates with progress of the season this was moderate and typical for pre-bloom conditions. However, we suspect there was light limitation because of extensive snow cover. At one station we observed a snow cover reaching up to 80 cm. Such thick snow cover would essentially screen off incoming light, considering that a 20 cm increase in snow thickness can reduce light transmission by 80%. Overall we observed a clear preference for nitrate uptake compared to ammonium. Especially at the water/ice interface, nitrate is the main substrate, while high rates for ammonium uptake were observed for some internal communities. Results will be further discussed in the light of physical and biogeochemical parameters observed during the SIPEX 2 expedition.


Automatic retrieval of thin sea-ice thickness by remote sensing

Øystein Rudjord, Rune Solberg, Øivind Due Trier, Gunnar Spreen, Sebastian Gerland, Angelika Renner, Nick Hughes

Corresponding author: Øystein Rudjord

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

Monitoring sea ice is important for understanding the dynamics in the climate system as well as practical applications like numerical weather prediction and ship navigation. The thin, seasonal sea ice deserves special attention as it is more vulnerable to summer melt, and it allows for greater heat transfer. Also, as much of the multi-year ice is disappearing in the Arctic due to climate change, the thin, seasonal ice makes up an increasing fraction of the total sea-ice volume. Remote sensing methods for retrieving sea-ice concentration (passive microwave radiometers and radar scatterometers) and ice thickness (altimetry, e.g. CryoSat) exist. However, their accuracy is reduced for low ice concentration and thin ice. Thin ice retrieval from SMOS is limited by the large spatial scales. Algorithms making use of other types of sensors for observing young, seasonal sea ice are therefore valuable. We present here an approach for the automatic retrieval of the extent and thickness of thin sea ice using thermal optical data, which is well suited for analysing time series of data. The method is based on a model in which the heat balance on the surface of the ice is described as a sum of heat fluxes from the atmosphere, solar radiation, and the water. These fluxes are described by various empirical models, of which some depend on ice-thickness information. In particular, the conductive heat flux (describing the heat transfer through the ice) is assumed to be inversely proportional to the ice thickness (with a correction for snow cover). Assuming thermal equilibrium yields an equation that can be solved to find an estimate of the ice thickness in every pixel of a remote sensing image. As input to the algorithm we use the thermal bands from the MODIS sensor on board the Aqua satellite to estimate the surface temperature of the ice. Atmospheric data are acquired from the re-analysed weather data of the ERA project. We also use passive microwave data from the AMSR-E and AMSR2 sensors to separate (thick) first- and multi-year ice from (thin) new ice. The ice-thickness products are validated using in situ ice-thickness measurements from field surveys on Svalbard, as well as operational ice-mapping products. In this presentation we describe the approach and show some preliminary ice thickness retrieval results as well as results from validation work near Svalbard.


On the use of O2/Ar and O2/N2 to estimate the biological carbon uptake in landfast sea ice

Jiayun Zhou, Bruno Delille, Frédéric Brabant, Jean-Louis Tison

Corresponding author: Jiayun Zhou

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

Sea ice is one of the largest biomes on Earth. The net community production (NCP) of the microorganisms living in sea ice impacts the dynamics of pCO2 in sea ice, and therefore the CO2 exchanges at the air–ice–sea interfaces. As oxygen O2 and carbon C are both involved in the photosynthetic and respiration processes, one can theoretically assess NCP (in terms of C uptake) from O2 measurements. However, the concentration of O2 in sea ice depends not only on biological processes (i.e. NCP) but also on physical processes. We present a technique for assessing NCP in sea ice, based on the use of the O2/Ar ratio, which should correct for the physical contribution in O2 variations. We also compare the use of O2/Ar and O2/N2 for deriving NCP, and demonstrate that O2/Ar is more suitable, as it is more sensitive and less affected by gas diffusion and gas bubble formation during sea-ice growth and decay than O2/N2. Using O2/Ar, we then provide conservative estimates of NCP in landfast sea ice, from ice cores collected in Barrow, from January through June 2009. The minimum estimate of the NCP in the whole ice cover reached 229 mg C m–2 d–1 in late spring. This is about 20 times higher than the atmospheric C uptake at that time identified from CO2 fluxes measurements at the ice–air interface, and therefore indicates that the main source of C used in the NCP was from the under-ice water.


Sea-ice biogeochemistry: a guide for modellers

Letizia Tedesco, Marcello Vichi

Corresponding author: Letizia Tedesco

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

Sea ice is a fundamental component of the climate system and plays a key role in polar trophic food webs. Nonetheless sea-ice biogeochemical dynamics at large temporal and spatial scales are still rarely described. Numerical models may potentially contribute integrating among sparse observations, but available models of sea-ice biogeochemistry are still scarce, whether their relevance for properly describing the current and future state of the polar oceans has been recently addressed. We describe a general methodology to develop a sea-ice biogeochemical model directly derived from a generic model of pelagic biogeochemistry. The described methodology is very flexible and allows choice between different levels of ecosystem complexity (1) and vertical representation (2), while adopting a strategy of coupling (3) that ensures mass conservation. We show how to apply this methodology by developing an intermediate complex model used to analyse a typical first-year ice season in the low Arctic. The aim is to: (1) extend our knowledge on the relevant controlling factors limiting sea-ice algae growth; (2) address the importance of sea-ice biogeochemistry; and at the same time (3) encourage ocean modellers of polar regions to add a new component to their modelling framework for a more adequate representation of the sea-ice-covered ocean ecosystem as a whole.


Role of the landfast ice in the Arctic Ocean circulation

Polona Itkin, Martin Losch, Rüdiger Gerdes

Corresponding author: Polona Itkin

Corresponding author e-mail: Polona.Itkin@awi.de

Landfast ice is not represented in the state-of-the-art sea-ice–ocean models and this results in regional overestimation of the sea-ice thickness, regional bias in sea-ice concentration, underestimation of the amount of the brines formed in the polynyas and unrealistic river water distribution on the shelf and in the Arctic Ocean. In our study we use a simple landfast ice parametrization to examine the role of the landfast ice for the river water distribution and cold halocline layer formation. Our results show that the landfast ice prevents freeze-up of the river water into the drift ice and instead channels the overwintering river water plume into the Beaufort Gyre. Furthermore, landfast ice increases the amount of the saline shelf water that reaches the top of the Arctic halocline, where salinity increases by about 0.2 in the 30 years of model simulation. Our approach overestimates the duration and extent of landfast ice and is thus the upper estimate of the landfast ice role, and yet we see no significant influnce on the Fram Strait volume exchange. Based on the landfast ice importance for the regional hydrology we recommend the use of landfast ice parametrization in the studies of the Arctic shelf-deep basin water mass exchange.


First long-term large-scale estimates of primary production in Baltic sea ice

Letizia Tedesco, Elina Miettunen, Byoung Woong An, Hermanni Kaartokallio, Jari Haapala

Corresponding author: Letizia Tedesco

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

In the past century the Baltic Sea has experienced an increase in air temperature larger than the global average and both air temperature and precipitation are projected to continue to increase in this century and, accordingly, also sea-ice thickness and cover are projected to change. Despite several decades of sampling, the role of the sea-ice biota in the ecosystem dynamics of the Baltic Sea still needs to be well understood. As sea ice may shortly disappear from the region, it is now the time to properly describe its quantitative relevance in order to properly project any biological change related to it and to the entire dynamics of the Baltic Sea in the future. We present here the first modelling effort towards describing long-term (1991–2007) and large-scale (the whole Bothnian Bay) biogeochemical properties of Baltic sea ice. We use the BaltiX forcing based on NEMO-LIM3 GCM implementation in the Baltic Sea (e.g. sea-ice thickness, fraction, temperature and salinity) off-line coupled to the Biogeochemical Flux Model in Sea Ice (BFMSI) to simulate the dynamics of sea-ice nutrients (N, P, Si), detritus (POC, DOC), algae, bacteria and fauna. We particularly focus on primary production and we study the correlation with several potential limiting factors, including the ice season length and thickness, PAR, brines volumes, and dependence on winter seawater biomass and nutrients. We analyse the spatial variability within the same year together with the inter-annual and seasonal variability among several years. Our overall results point mainly to a low-productive sea-ice region, but biologically-active throughout the entire ice season, and strongly limited by the low brine volumes that greatly constrain the habitability of the ice. Finally, we present our plans for extending the area of investigation to the entire ice-covered Baltic Sea and we suggest how the same methodology can be successfully applied in any ice-covered ocean.


Oxygen and nitrogen isotopic signatures of nitrate in East Antarctic sea-ice and the underlying water column (SIPEX 2 results)

Arnout Roukaerts, Anne-Julie Cavagna, Frank Dehairs, François Fripiat, Delphine Lannuzel, Klaus Meiners

Corresponding author: Arnout Roukaerts

Corresponding author e-mail: arnout.roukaerts@vub.ac.be

Nitrate is an important source of nitrogen for organisms as it is the largest pool of fixed nitrogen available in the ocean. During biological processing small isotope fractionations occur which affect the isotopic signature of oxygen and nitrogen in nitrate and contain information on past and ongoing processes influencing nitrate. Although isotopic signature analysis for nitrate has become a common tool to study the N-cycling in freshwater and marine systems, measurements in sea ice are still extremely scarce. During the SIPEX 2 expedition (Sep.–Oct. 2012; R/V: Aurora Australis; 63–66°S, 115–125°E) samples were taken in both sea ice and the underlying water column. Nitrate isotopic signals in the water column showed little variation and were close to deep ocean values (δ15N = 4.8‰, δ18O = 2.4‰). Close to the surface of the water column a small elevation was observed for both nitrogen and oxygen suggesting some biological activity. In the sea-ice, slightly higher values of δ15N = 5.5‰ were observed at the water–ice interface and which increased to 14‰ in the center and top layers of the ice cores. Increasing δ15N values coincide with an increase of δ18O and decrease of nitrate concentration. This suggests that nitrate assimilation or denitrification could be one of the main processes explaining the observed isotopic variations. Segregating these two processes with only the nitrate isotopy tool is difficult, however anoxic conditions and denitrification have been observed in sea ice before. At some stations the δ15N and δ18O results deviate from the 1:1 slope (which is generally associated with assimilation and denitrification) indicating that other processes are probably also taking place in the studied sea ice. Further analysis of the data is required to better constrain the processes.


About uncertainties in Antarctic sea-ice thickness retrieval from ICESat

Stefan Kern, Gunnar Spreen

Corresponding author: Stefan Kern

Corresponding author e-mail: stefan.kern@zmaw.de

Sea-ice volume is an important parameter to identify the impact of climate change in high latitudes as it has been shown for the Arctic. We know more about the sea-ice volume of the Arctic than the Antarctic because of submarine upward looking sonar observations and because of a more successful retrieval of the sea-ice thickness using laser or radar altimetry in the Arctic. For the Antarctic, sea-ice thickness validation data sources are extremely sparse. Validation of new approaches for satellite sea-ice thickness retrieval is limited consequently to these few in-situ observations and to inter-comparisons with alternative methods. Therefore careful uncertainty estimations should play an important role which is the motivation of this paper. The basis for sea-ice thickness retrieval from satellite altimetry is the elevation of the ice surface above the sea surface: the freeboard. Thereafter buoyancy or empirical approaches are applied to convert freeboard to ice thickness. Because the sea surface is not known with the spatiotemporal resolution required, it has to be approximated from altimetry data itself. For ICESat in the Arctic it was shown that this is the dominating contribution to the sea-ice thickness uncertainty. The present paper discusses the results of a sensitivity analysis varying several input parameters for a typical freeboard estimation algorithm from ICESat data for Antarctic sea ice. Snow depth is the next important uncertainty source using laser altimetry for sea-ice thickness retrieval. The contributions by sea ice and snow densities are smaller but together can exceed the contribution of snow depth to sea-ice thickness uncertainty. The vertically heterogeneous snow cover on Antarctic sea ice complicates the impact of the snow cover on its thickness retrieval using altimetry. The present paper will review the main aspects related to these three parameters. It will give theoretical uncertainty estimates based on recent snow depth and density observations for Antarctic sea ice. An estimate of the actual sea-ice thickness uncertainty will be given for the Weddell Sea.


Three-dimensional numerical study of electromagnetic induction retrievals of sea-ice thickness

Jesse Samluk, Cathleen Geiger, Chester Weiss, James Kolodzey

Corresponding author: Jesse Samluk

Corresponding author e-mail: sevensam@udel.edu

Since the conductivities of snow and ice are at least a factor of 10 smaller than the underlying seawater, current sea-ice thickness retrievals using electromagnetic (EM) induction methods yield results to 10% uncertainty for level ice and 40–60% uncertainty in proximity of deformed ice features. We explore potential sources for these uncertainties, especially near deformed ice features, using a three-dimensional (3-D) numerical model. Power law relationships of resolution error are the metric we use to compare these 3-D model results to observations in the Beaufort Sea. A more detailed report, intended for Annals of Glaciology, will quantify these results further.


Impacts of different atmospheric forcing on the Arctic sea-ice numerical forecasting cases

Qinghua Yang, Jiping Liu, Martin Losch, Zhanhai Zhang, Cuijuan Sui, Jianyong Xing, Ming Li, Mirong Song

Corresponding author: Qinghua Yang

Corresponding author e-mail: yqh1983@hotmail.com

In this study, a regional Arctic configuration of the Massachusetts Institute of Technology general circulation model (MITgcm) is used as the coupled ice–ocean model for forecasting sea-ice conditions in the Arctic Ocean, and two different datasets as atmospheric forcing are tested: (1) the numerical weather prediction of the National Centers for Environmental Prediction Global Forecast System (NCEP GPS) and (2) the model outputs from Polar WRF simulations. In addition, two different satellite-derived sea-ice products are used as initialization: (1) the Special Sensor Microwave Imager (SSM/I) and (2) the Advanced Microwave Scanning Radiometer for EOS (AMSR-E), respectively. Three synoptic cases which represent typical atmospheric circulations over the Arctic Ocean are selected to carry out the Arctic sea-ice numerical forecasting experiments. Forecast skill assessments of the sea-ice concentration and thickness fields from these three numerical forecasting cases are presented.


Monitoring sea-ice variability in Storfjorden, Svalbard, by combined remote sensing and in situ observations

Gunnar Spreen, Angelika H.H. Renner, Max König, Sebastian Gerland, Olga Pavlova, Anthony P. Doulgeris, Øystein Rudjord, Caixin Wang, Justin Beckers

Corresponding author: Gunnar Spreen

Corresponding author e-mail: gunnar.spreen@npolar.no

Storfjorden, situated between the islands Spitzbergen, Barentsøya, and Edgeøya of the Svalbard archipelago, offers ideal conditions to study the complexity of Arctic sea-ice processes and their atmosphere–ocean interactions in a comparable small and easy accessible location. Storfjorden’s northern part is dominated by a latent heat polynya with associated high rates of local ice production. The polynya is one of the primary sites for brine-enriched shelf water production in the Arctic. The sea-ice composition in Storfjorden is diverse and shows high inter-annual variability. Four main sea-ice classes can be separated: (1) landfast ice growing along the shores, (2) drift ice which was imported from the Barents Sea into the fjord, (3) pack ice grown locally in the aforementioned polynya, and (4) thin ice/open water covering the polynya during northerly wind conditions. The Norwegian Polar Institute conducts a long-term monitoring program of the sea-ice conditions in Storfjorden, primarily of its fast ice. Regular visits of the landfast ice are supplemented with helicopter surveys during recent years. We present a time series of the development of the different sea-ice classes during the two winters 2011/2012 and 2012/2013 obtained from Radarsat-2 Scan SAR wide imagery. Four acquisitions of Radarsat-2 quad-polarized data in 2013 allow a more detailed classification of different sea-ice types, e.g. different surface roughness. Estimates of thin ice thickness in the polynya are obtained from MODIS thermal optical data for selected dates. These results are interpreted in conjunction with measurements obtained during helicopter surveys, snow machine visits of the fast ice, and data from an autonomous sea-ice mass balance buoy and a weather station. These in-situ measurements comprise of sea-ice thickness measurements from electromagnetic induction sounding, optical images, snow depth and snow stratigraphy observations, and temperature and wind measurements. The fast ice in-situ observations are interconnected by 1-D modelling of the ice and snow development. The first remote sensing and modelling studies of Storfjorden started in the mid 1990s. These new, more detailed observations help to establish the monitoring of the long-term development of the Storfjorden sea-ice cover and dense water formation. The obtained results can be applied to observations in other Arctic regions with similar sea-ice types.


Modelling biogeochemistry in the Arctic: limitations of Earth System Model projections

Nadja Steiner, Clara Deal, Martin Vancoppenolle

Corresponding author: Nadja Steiner

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

The 5th Coupled Model Intercomparison Project (CMIP5) includes for the first time a variety of Earth System Models (ESMs, model systems with fully coupled atmosphere, ocean, sea-ice and land components including interactive biogeochemical modules for all components), allowing the study of future projections of the marine carbon cycle and ecosystem behaviour. However, the still fairly coarse horizontal and vertical resolution restricts the ability to resolve biological or chemical processes happening in the euphotic zone as well as small-scale physical processes important for biogeochemistry. Model skill for biogeochemical variables in global ESMs is still low, especially for the Arctic, where observational data are few and seasonally biased, and where shelf areas and narrow passages are common features. Within the SCOR working group 140, Biogeochemical Exchange Processes at the Sea-Ice Interfaces (BEPSII), the task group on modelling is aiming at bridging the gaps between scales by analyzing the applicability and relevance of small-scale processes and model parameterizations in 1-D models for larger scale models as well as identifying model characteristics via model intercomparisons. We will provide an update on this task group and present results from ESM intercomparison studies in the Arctic which show both consistent trends (e.g. acidification) as well as differences (e.g. future primary production, vertical structure, regional distribution) between the models.We will discuss the shortcomings in ESMs and improvements obtained with the application of higher resolution regional models.


CryoSat-2 sea-ice thickness estimates and their uncertainties

Robert Ricker, Stefan Hendricks, Veit Helm, Henriette Skourup, Rüdiger Gerdes, Malcolm Davidson

Corresponding author: Robert Ricker

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

Several studies have shown that there is a notable evidence for thinning of the Arctic sea ice in the last decades. Together with the rapid reduction of ice-covered area this leads to a decline in sea-ice volume. The only remote sensing technique capable of obtaining sea-ice thickness on basin scale is satellite altimetry, applied from platforms such as ICESat and CryoSat-2. The current mission, CryoSat-2, was launched in 2010 and is equipped with the Ku-band SAR radar altimeter SIRAL, which we use to derive sea-ice freeboard, the height of the ice surface above the local sea level. Accurate CryoSat-2 range measurements over open water and the ice surface in the order of centimetres are necessary to achieve the required accuracy of the freeboard to thickness conversion. The local sea-surface height can be determined by automatic detection of leads in the ice cover by the characteristics of the radar signal. The uncertainty of radar freeboard retrievals over ice floes are governed by the quality of automatic lead detection and by temporal and regional dependencies of the Ku-band signal penetrations into the snow layer and the surface roughness. Uncertainties in the freeboard-thickness conversion arise from necessary assumptions of hemispheric and year-around snow depths fields and fixed ice densities for different ice types. To estimate these error sources we compare the CryoSat-2 sea-ice thickness product with validation measurements in the Arctic. The validation and error estimation is based on coincident satellite and airborne altimetry and direct airborne sea-ice thickness measurements. From combined airborne radar and laser altimetry we estimate the uncertainty of CryoSat-2 range and we assess the errors of the freeboard to thickness conversion with direct airborne EM ice thickness measurements. Considering all error contributions we find a decline of 0.32 ± 0.3 m in sea-ice thickness and 2842 ± 1832 km3 in February Arctic sea-ice volume respectively over three winter seasons (2010–13) using a modified snow climatology. While CryoSat-2 freeboard maps show a reasonable regional distribution of sea-ice freeboard, the accuracy of sea-ice thickness and its spatial distribution is influenced by sea-ice type and the inability of the snow depth climatology to map the patterns of actual snow load.


Proxy reconstructions of Arctic-subarctic sea-ice cover variations at millennial time scales : example of the Arctic-subarctic during the Holocene based on dinoflagellate cysts

Anne de Vernal, Maryse Henry, Claude Hillaire-Marcel, Bianca Fréchette, André Rochon, Sandrine Solignac

Corresponding author: Anne de Vernal

Corresponding author e-mail: devernal.anne@uqam.ca

The reconstruction of sea-ice time series beyond instrumental observations is necessary to document the full range of sea-ice variations under natural forcing. Hence, several approaches based on biogenic, geochemical or sedimentological proxies have been developed from marine sediments and coastal records. Among those, the assemblages of dinoflagellate cysts permit quantitative estimates of concentration or seasonal duration of sea-ice cover with an accuracy of ±11%. The reconstructions made in about 30 cores of the northern North Atlantic and Arctic seas show a strong regionalism in the pattern of changes throughout the Holocene. The results suggest high sea-ice concentration (>7/10 per year) with little changes in the Canadian Arctic but large amplitude variations in the Chukchi and Barents seas, where data suggest millennial type oscillations with a pacing almost opposite in Western vs Eastern Arctic. The dinoflagellate cyst data also suggest minimum sea-ice concentration during the mid-Holocene, notably in the eastern Fram Strait, northern Baffin Bay and Labrador Sea. However, the density of dinoflagellate cyst data points is still too low to permit extrapolation at the scale of the Arctic and circum-Arctic. Hence, the use of other proxies such as biomarkers, microfossil assemblages, or ice rafted debris, may help to better constrain occurrence and areal extent of past seasonal sea-ice cover. Efforts to combine several proxies are currently being made by the PAGES sea-ice proxy (SIP) working group and within the framework of the Past4Future project for developing multi-proxy maps of Arctic sea-ice cover.


First ‘in situ’ measurement of bulk gas concentration and gas diffusivity (DO2, DAr and DN2) in natural sea ice

Bruno Delille, Nicalos-Xavier Geilfius, David Thomas, Soren Rysgaard, Jean-Louis Tison, Odile Crabeck

Corresponding author: Odile Crabeck

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

Gas exchange through sea ice is a determining factor in the polar ocean budget of climatically active gases. The pathways and rates of air to sea gas transfer in the presence of a sea-ice cover are not yet well established. We report the bulk gas concentration of permanent naturel gases: O2, N2 and Ar, as well as their diffusivity coefficients, D, in natural landfast sea ice. The survey took place in March 2010 in Kapisillit Kangerluaq fjord (Greenland). The bulk ice gas composition was spanning the interval between the gas composition dissolved in seawater and the atmospheric gas composition. Most past studies suggest that convective transport is the main driver of gas displacement and neglect diffusion processes. According to our data, brines were stratified within the ice, so that no convective transport took place within the brine system. Therefore, diffusive transport was the main driver of gas migration. By analysing the temporal evolution of an internal gas peak within the ice, we deduced the bulk gas diffusivity coefficients, for oxygen (DO2), argon (DAr) and nitrogen (DN2). The values fit to the few previous estimates from experimental work, and are close to diffusivity values in water (10–5 cm2 s–1). We suggest that gas bubbles escaping from the brine medium towards the atmosphere, as the ice gets more permeable, could be responsible for abnormally higher observed diffusivities. These results underline that when there is no convective transport within the sea ice the transport of gas by diffusion through the brines, either in the liquid or gaseous phases, though slower than convective transport, could be an important pathway for ocean–atmosphere exchange.


Linking IceBridge, ICESat, and CryoSat-2 for improved seasonal to decadal-scale estimates of sea-ice thickness and volume

Nathan Kurtz, Natalia Galin, Michael Studinger, Jeremy Harbeck, Thorsten Markus, Vincent Onana, Donghui Yi, Sinead Farrell, Jackie Richter-Menge

Corresponding author: Nathan Kurtz

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

Recent work by Laxon et al., (2013) has connected the observational records of Arctic sea-ice thickness from ICESat and CryoSat-2, and airborne measurements collected by IceBridge and ESA’s CryoVEx campaign. The wealth of observations and public availability of data products from the IceBridge airborne campaign provide a continuing source for evaluation and improvement of the observational sea-ice thickness record. Additionally, the fast availability of the IceBridge quick-look data products are now providing a unique new dataset for improving seasonal forecasts of Arctic sea ice. We present a best estimate of a decadal sea-ice thickness time series from the altimetry records of ICESat, IceBridge, and CryoSat-2. We will present comparisons between IceBridge and CryoSat-2 freeboard and thickness observations through analysis of contemporaneous underflights of CryoSat-2 by the IceBridge mission. We show that improved freeboard and thickness results can be obtained over currently used threshold tracker methods through the usage of a new surface tracker which adapts to varying surface roughness and the associated backscattering variations found over different sea-ice surfaces. Lastly, we will present an evaluation of a merged IceBridge–CryoSat-2 quick look sea-ice product to be made available for seasonal sea-ice forecasts.


Evaluating modeled and observed sea-ice drift using Rotary Taylor Diagrams

Andrew Roberts, Jennifer Hutchings

Corresponding author: Andrew Roberts

Corresponding author e-mail: afrobert@nps.edu

Assessments of simulated sea-ice velocity using traditional Taylor Diagrams is limited to individual comparisons of vector components and magnitude against in situ- and satellite-derived sea-ice velocity. This must be supplemented with comparative graphs of rotary power spectral density and coherence if one is to fully summarize model skill, variability and uncertainty, and observational noise as originally intended for single Taylor Diagrams. The reason for this is partly because sea-ice drift is a vector, not a scalar, and also because a more meaningful analysis is obtained by isolating distinct physical modes in sea-ice dynamics, including inertial oscillations, synoptic variability and seasonal drift. The inability to reduce this information to a single diagrammatic representation poses a problem in assessing sea-ice velocity data from the latest phase of the Coupled Model Inter-Comparison Project (CMIP5), and in validating next-generation coupled models that resolve oceanic eddies with inertial and tidal oscillations in sea-ice zones. We address this deficiency by combining rotary wavelet reconstruction with traditional Taylor diagrammatic methods to create Rotary Taylor Diagrams. By applying a real symmetric wavelet to a Continuous Wavelet Transform of ice velocity, we demonstrate the ability to isolate sea-ice drift to inertial, tidal, synoptic, seasonal, and decadal bands in individual Rotary Taylor Diagrams of clockwise and anticlockwise motion. These indicate correlation and RMS difference between, and variance of, modeled and observed ice drift in the way originally intended for Taylor Diagrams of scalar quantities. We further demonstrate that Normalized Rotary Taylor Diagrams can combine information from multiple bands into a single summary of model ensemble performance and observational uncertainty using measured sea-ice velocity from International Arctic Buoy Program (IABP) drifters, satellite-derived products, and output from the Regional Arctic System Model (RASM). This includes a graphical indication of the signal-to-noise ratio of sea-ice velocity measured using ARGOS and GPS positioning, and passive microwave sensors, verified independently using multiple coherence analysis. Rotary Taylor Diagrams are broadly applicable to other vector quantities, such as surface winds and currents, and a generic tool for creating them is available for download as part of a new Polar System Analysis Toolbox developed at Naval Postgraduate School.


Halogen chemistry in polar environments and the impact of climate change

Paul Shepson

Corresponding author: Paul Shepson

Corresponding author e-mail: pshepson@purdue.edu

It is well known that molecular halogens are released from saline snowpacks, and likely from sea ice and snow cover on the ice, resulting in gas phase halogen atom chemistry in polar regions that can have significant impacts, including total removal of ozone and elemental mercury from the lowest part of the atmosphere. We are beginning to understand some of the complex processes that occur in the condensed phase resulting in production of those halogenated species, and we are at the beginning stages of developing a predictive capability. However, the ongoing assessment of the overall impacts of those processes on the composition and physical characteristics (e.g. aerosols, radiation, and cloud cover) of the atmosphere is complicated by the rapid changes occurring at the surface in the Arctic (and parts of coastal Antarctica) due to climate change. I will review the fundamentals of what we know about the chemistry, on/in what surfaces that chemistry occurs, and how those processes might change as the surfaces change in the context of a warming and developing planet. I will also summarize a little about what we know about why this chemistry might matter.


A case study of high frequency ice–ocean–atmosphere dynamics in the Regional Arctic System Model

Andrew Roberts, John Cassano, Alice DuVivier, Mimi Hughes, Wieslaw Maslowski, Robert Osinski

Corresponding author: Andrew Roberts

Corresponding author e-mail: afrobert@nps.edu

The Regional Arctic System Model (RASM) is a fully coupled limited area pan-Arctic climate model developed as a counterpart to the global Community Earth System Model (CESM). RASM is comprised of the Los Alamos Sea Ice Model (CICE) and Parallel Ocean Program (POP), Variable Infiltration Capacity (VIC) hydrology model and the Weather Research and Forecasting (WRF) model. It uses the same coupling infrastructure as CESM, and is currently configured with 1/12 degree ice–ocean resolution, which is eddy permitting in the Arctic Ocean, and atmosphere–land resolution of 50 km. All components are coupled at 20-minute (super-inertial) intervals. This enables important physical interactions not represented in many IPCC-AR5 models, including increased flux extremes passed between component models, feedbacks resulting from ice–ocean inertial oscillations, and much reduced impulsive wind-stress that can stem from passing discrete, temporally averaged fluxes between models. We demonstrate improvements in simulations of sea-ice dynamics resulting from the RASM coupling configuration through a case study of hindcast ensembles for a single year (1996), comparing three different RASM coupling configurations. First, we demonstrate the ability of the standard RASM configuration to simulate observed ice drift and deformation from 6-hourly to seasonal timescales using spectral analysis and GPS buoys. Next, we demonstrate that by filtering super-inertial ice–ocean wind stress at the coupler interface, ice drift and deformation extremes resulting from storms are considerably damped, with a reduction in maximum ice speeds of inertial oscillations from ~15 cm s–1 to about ~5 cm s–1, and an associated reduction in deformation most noticeable in the marginal ice zone. Finally, we demonstrate that by coupling at ~6 hour intervals, strong damping occurs in the ice–ocean boundary layer in addition to chaotic wander inherent in time-delayed coupled oscillatory systems.


Antarctic sea-ice thickness and volume estimates and uncertainties from ice charts between 1995 and 1998

Rachel Bernstein, Cathleen Geiger, Tracy DeLiberty, Mary Stampone

Corresponding author: Rachel Bernstein

Corresponding author e-mail: erbern@udel.edu

This work evaluates two distinct measures of central tendency for the sea-ice thickness distribution and calculates the impact these measures have on total sea-ice volume in the Southern Ocean between 1995 and 1998. Sea-ice stage of development from operational ice charts serves as a proxy for sea-ice thickness. Based on ice chart features, we estimate the sea-ice thickness distribution and associated uncertainties across five thickness categories. The first measure of central tendency, area-weighted average thickness, is the average thickness of a feature on an ice chart which is then averaged with the thicknesses of all other sea-ice features in the region. The second measure, integrated thickness, uses thickness categories accumulated for an entire region and calculates the central value in a single step. By comparing the two measures of central tendency of the sea-ice distribution, we examine critical problems associated with smoothing smaller scale results for larger scale applications. The propagation of uncertainties around the central tendency is evaluated for the average and integrated thickness procedures. We recognize that accuracy of thickness distribution estimated from proxy information cannot be quantified in an absolute sense because of insufficient ground truth data. However, this dataset is appropriate for evaluating relative uncertainties associated with averaging schemes. Results show that regionally integrated thickness always exceeds area-weighted average thickness due to a skew towards thick and deformed ice, which is prevalent in sea-ice thickness distributions. The resulting sea-ice volume from integrated thickness is up to 60% larger than the volume calculated from average thickness at the ice chart polygon scale. Based on this analysis, we conclude with recommendations for calculating the thickness distribution and volume of sea ice over large-scale regions for remote sensing and modeling applications.


Grease and pancake ice growth in large scale sea-ice ocean models

Lars H. Smedsrud, Torge Martin

Corresponding author: Lars H. Smedsrud

Corresponding author e-mail: larsh@gfi.uib.no

The first stage of sea-ice formation is often grease ice, a mixture of sea water and frazil ice crystals. Within a couple of days grease ice congeals to pancake ice given a minimum of wave activity, finally forming a solid sea-ice cover. While such a solid sea-ice cover reduces the heat flux from the open ocean to the wintery cold atmosphere by two orders of magnitude, grease ice reduces this heat loss only marginally. Hence, the presence of grease ice importantly extends the duration of greatest heat loss also enhancing sea-ice production. Most climate models, however, turn frazil ice instantly into solid sea ice instead of accounting for the slower water grease-pancake transition. Although some sea-ice models include this delay by using a ‘lead closing parameter’, this constant is often used to tune the model and thus does not necessarily reflect the actual relationship of grease ice layer thickness, areal coverage, and congelation time scale to lead width, wind speed, and wave activity. We present a simple approach to include grease and pancake ice in large-scale sea-ice models. Our parameterization captures the basic grease ice properties: a surface temperature at the freezing point, a 1:3 ratio of frazil ice to sea water volume in the grease layer, and no reduction in surface heat fluxes. The open water heat loss governs the grease ice production, and the transition to a solid sea-ice cover follows gradually over the next 24 h when ~50% of the model grease ice area forms pancake ice. Hence, our grease ice parameterization delays the transition from open water to a solid ice cover, and increases the oceanic heat loss during fall and winter. We demonstrate these effects by running sensitivity tests with a 1-D sea-ice model, in which we also compare the effectivity of the grease ice parameterization to different settings of the lead closing parameter. First simulations indicate a 30% increase in mean winter Arctic Ocean heat loss with the new grease ice parameterization. Further, we show the impact of accounting for grease ice in a coupled sea-ice ocean model multi-year simulation of the Arctic Ocean. The heat loss increases with stronger ice divergence and larger open water areas and may therefore gain influence in the future, when the Arctic Ocean is predominantly ice free at the end of summer.


On the sensitivity of sea-ice states to variable parameter space in the Regional Arctic System Model (RASM)

Robert Osinski, Wieslaw Maslowski, Jaclyn Clement Kinney, Andrew Roberts

Corresponding author: Robert Osinski

Corresponding author e-mail: roberto@iopan.gda.pl

The Regional Arctic System Model (RASM) is a regional climate model developed following the framework of the Community Earth System Model (CESM). The complete RASM currently includes the Los Alamos Sea Ice Model (CICE) and Parallel Ocean Program (POP), Weather Research and Forecasting Model (WRF) and Variable Infiltration Capacity (VIC) land hydrology model. These components are coupled using the CESM flux coupler (CPL7). POP and CICE are configured on the same high resolution, rotated spherical 1/12-degree grid, which is eddy-permitting in the Arctic Ocean. Here we use a subset of the full RASM model, where WRF and VIC are replaced with prescribed realistic atmospheric data from the Common Ocean Reference Experiment version 2 (CORE2) 1948–2009 reanalysis. This approach allows investigation of sensitivity of modeled sea-ice states to varying parameters, controlling sea-ice dynamics, thermodynamics and its coupling with the atmosphere as well as direct comparison of model results with observations. The varied parameters include the ice strength, which controls sea-ice deformation, sea ice/snow albedo estimated with the delta-Eddington shortwave radiative transfer scheme, and surface roughness wavelength scale, which determines air-ice coupling. Sea ice results from multiple experiments are compared against each other and against basin-wide observations of sea-ice extent, concentration, thickness, volume and velocity. Our results indicate that many physical parameterizations currently used in sea ice and climate models to represent sub-grid scale processes are scale-dependent and require fine-tuning as model resolution changes. We also show that while results from several tests have realistic sea-ice extent and/or concentration compared to observations, sea-ice thickness distribution and/or drift patterns are acceptable only in a few cases. Hence, we conclude that the use of observed sea-ice extent only to validate the skill of sea-ice models is not a sufficient model constraint.


Mapping of Antarctic coastal polynyas and landfast sea ice from AMSR-E

Sohey Nihashi, Kay I. Ohshima

Corresponding author: Sohey Nihashi

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

The active freezing in Antarctic coastal polynyas leads to dense water formation through large amount of brine rejection. The dense water is a major source of Antarctic Bottom Water (AABW) which is the densest water in the global overturning and is a key player in the climate change as a significant sink for heat and carbon dioxide. The Antarctic coastal polynyas are also the areas of high primary productivity. In some coastal polynyas, it was suggested that landfast sea ice (fast ice), which is a prominent feature of the Antarctic coastal zone, plays an important role in the polynya and AABW formation because it causes divergent ice motion by blocking ice advection. This study presents the first combined mapping of the coastal polynyas (ice production) and fast ice along the entire Antarctic coast using AMSR-E data. The spatial resolution of AMSR-E (about 6.25 km) is two times finer than that of SSM/I which was used in the previous study. This advantage of AMSR-E is critical for the detection of the coastal polynyas and fast ice because their spatial scales are typically less than tens of km. The mapping reveals strong linkage between the coastal polynyas and fast ice: most of the polynyas are formed on the western side of fast ice. The results from the analyses based on the mapping and meteorological dataset are summarized as follows: A wind diverging from the boundary defined by fast ice and land is the primary cause for the polynya formation. A blocking effect of fast ice on sea ice advected by the coastal current is another key factor. A drastic change in fast ice extent causes a dramatic change in the polynya area and ice production, suggesting that the fast ice change can be an important factor of the climate change. The findings of this study suggest that fast ice should be treated in future models for better climate projections. The mapping of this study gives the boundary/validation data of fast ice and ice production for such models.


The role of cyclone activity in the interannual variability of the Beaufort High

Masatake Hori, Jun Inoue

Corresponding author: Masatake Hori

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

The Beaufort High (BH) is a well-known climatological feature of the Arctic with close ties to sea-ice variability. Here, an objective cyclone tracking algorithm is utilized to link the change in cyclone migration pattern and density towards the seasonal mean BH variability. It is found that during winter, the BH variability is influenced by migrating cyclones from the North Atlantic as well as North Pacific cyclones moving northward through the Bering strait where more cyclones entering the BH region corresponds to a weaker BH. Density of cyclone increases along the Eurasian continent coastline and western North Pacific in accordance to a stronger Northern hemisphere annular mode (NAM) and a weakened Aleutian Low. During summer, similar decrease/increase in cyclone number is seen for strong/weak BH activity where region of maximum cyclone density shifts southward toward the Laptev Sea for a strong BH year, and a intensification of ‘Arctic cyclone’ is observed for the weak BH years. These results show that the seasonal mean BH activity and its climatological feature can be explained as a composite of migrating cyclones and its change in density.


Evaluation of pH evolution during the seawater freezing process and sea-ice growth

Alexander Hare, Feiyue Wang, David Barber, Ryan Galley, Nicolas-Xavier Geilfus, Soeren Rysgaard

Corresponding author: Alexander Hare

Corresponding author e-mail: Alexander.Hare@umanitoba.ca

The pH of bulk sea ice and brine in both natural and experimental sea ice was evaluated at near-in situ conditions to study the evolution of pH during the seawater freezing process and the progression of sea-ice development. Sea ice grown outdoors at the Sea-ice Environmental Facility (SERF, Winnipeg, Canada) and in the absence of significant biological activity demonstrated highly alkaline pH values (>9.5) in exterior ice layers and reduced pH in interior layers (~7.0). Frost flowers and young ice had pH well above that of seawater, while sea-ice brine had pH below that of seawater. We interpret these results as indicating CO2-expulsion from exterior ice due to ikaite formation and the rejection of CO2-saturated brine at the ice surface, and CO2 concentration within a semi-closed interior ice system due to reduced salinity and temperatures that decrease brine permeability. Natural sea ice from Godthåbsfjord, Greenland, in April 2013 demonstrated lower surface pH and lower overall variability in pH compared to SERF ice. This may reflect sequential periods of surface melt and the contribution of freshwater glacial melt. Together, these observations provide the framework for a conceptual model of pH evolution in sea-ice growth and melt coupled to the carbonate system.


A method of short-term sea-ice prediction to support ice navigation in the Arctic Ocean

Liyanarachchi Waruna Arampath De Silva, Hajime Yamaguchi, Jun Ono

Corresponding author: Liyanarachchi Waruna Arampath De Silva

Corresponding author e-mail: waruna@fluidlab.sys.t.u-tokyo.ac.jp

Rapid decrease of summer sea ice in the Arctic Ocean has been extending the navigation period in the Arctic Sea Routes (ASR). The passages through the Arctic Ocean are the shortest sea routes from North American and European harbors to Far East Asian harbors. In this regard, precise ice distribution prediction is one of the key issues to realize safe and efficient navigation in the Arctic Ocean. In general, however, most of the available numerical models have shown high uncertainties in the short-term and narrow-area predictions, especially marginal ice zones like ASR. In this study, therefore, we predicted the short-term sea-ice conditions in Arctic sea routes using meso-scale eddy resolving high-resolution ice–ocean coupled model (ice–POM) with explicitly treating the ice floe collision in the marginal ice zones. The ocean part is based on Princeton Ocean Model (POM), while the ice part consists of full thermodynamic and dynamic models. First, numerical issues associated with collision rheology in the ice–POM model were discussed and resolved. Then coarser resolution (about 25 km) whole Arctic Ocean model was developed to investigate the basic performance of the model by discussing the reproducibility of seasonal and interannual sea-ice variability. It was found that coarser resolution model can reproduce seasonal and interannual variations reasonably compared with observations but cannot be used to predict the short-term variation like 1–2 weeks. Therefore secondly, the high-resolution (about 2.5 km) regional models were setup along the Arctic sea routes to investigate the accuracy of short-term sea-ice predictions. High-resolution computation was able to predict the sea-ice extents reasonably in comparison with observations because of the improved expression of the ice-albedo feedback process and ice-eddy interaction process.


In situ measurement of ocean heat flux and its contribution to the thickness of Antarctic coastal sea ice

Pat Langhorne, Alex Gough, Mike Williams, Craig Stevens, Inga Smith, Natalie Robinson, Wolfgang Rack, Daniel Price, Greg Leonard, Ken Hughes, Pat Wongpan, David Dempsey, Christian Haas, Andy Mahoney, Tim Haskell

Corresponding author: Pat Langhorne

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

Climate models fail to predict overall trends in Antarctic sea-ice extent. In particular models disagree about the relative importance of heat from the ocean versus heat from the atmosphere. Here we examine in situ measurements of negative ocean heat flux (i.e. flux to the ocean) in sea ice around coastal Antarctica, using our knowledge of ice–ocean interaction to employ proxy measures to extend coverage. The situation in which the ocean makes a significant contribution to sea-ice growth has the potential to be relatively common in Antarctic waters. Close to an ice shelf, sea ice can form in water that has been supercooled by interaction with the ice shelf at depth. Within a few kilometers of the ice shelf edge the resulting ocean heat flux drives the formation of more than 50% of the thickness of the landfast sea ice: even at a distance of 60 km there is a measurable contribution. This ocean heat flux modifies crystallographic structure, leaving its signature frozen into the sea-ice cover through the formation of platelet ice. The thermal deficit also causes frazil crystals in the water column that accumulate and grow under the sea ice in a very porous layer. The spatial distribution of ocean heat flux is related to the thickness distribution of this subice platelet layer. Some of the earliest Antarctic ice–ocean observational records began a century ago near the combined Ross and McMurdo ice shelves in southern McMurdo Sound. Since then there has been no discernable change in the temperature of the upper ocean. This surface water is held just below its freezing point as it enters the Sound, probably because of the regulating influence of basal melting deep in the ice shelf cavity. We use the ability of sea ice to integrate the effect of ocean heat flux over its annual growth to interpret crystallographic records from a historical time series of sea-ice cores. The distribution of platelet ice (and hence ocean heat flux) appears to be strongly linked to the circulation in McMurdo Sound, although its abundance varies from year to year. The deduced mean, late-winter heat flux to the ocean varies from 0 to 50±15 W m–2 across the region. Finally these multiple sources of data, contextualized within the relatively long time series, are gathered for all ice shelf-influenced regions of Antarctica to provide estimates of ocean/ice shelf contribution to the sea-ice cover.


Simulation of the crystal growth of platelet sea ice with diffusive heat and mass transfer

Pat Wongpan, Pat Langhorne, David Dempsey, Lisa Hahn-Woernle, Zhifa Sun

Corresponding author: Pat Wongpan

Corresponding author e-mail: wonpa625@student.otago.ac.nz

Platelet ice is a sea-ice type found near an ice shelf. Platelet crystals, which originate within a supercooled water column, rise to the surface and deposit under the sea-ice cover in a porous layer, forming a sub-ice platelet layer. The near-surface supercooled water allows the platelet crystals to become frozen into the ice cover as incorporated platelet ice. Several Antarctic field campaigns have collected ice cores, measured crystallographic and physical properties, and simultaneously recorded oceanographic conditions. However, some in situ measurements are difficult to make and theoretical modeling can contribute to solving these difficulties. In particular, the solid fraction of the subice platelet layer is necessary to obtain sea-ice thickness from remote-sensing measurements. Voronoi dynamics is a simple but efficient numerical technique to model grain growth. By integrating this with the analyses of mechanical stability and diffusive heat and mass transfer, platelet deposition, in situ growth, and incorporation into the sea-ice cover are simulated in three dimensions. The results of this model show topological similarity with incorporated platelet ice observed in cores collected from McMurdo Sound, Antarctica. When calibrated, spatial–temporal distributions of porosity, salinity, temperature and crystallographic c-axes can be extracted and compared against observation. For simulations with the growth speed of the platelet tip matched to those observed beneath sea ice in McMurdo Sound (of 10–3 mm s–1), the solid fraction of the sub-ice platelet layer obtained from our simulation due to local growth of platelet crystals within this layer is 0.22 ± 0.01. This is combined with a packing efficiency of 0.06 ± 0.01, previously calculated from the initial deposition of platelet crystals at the interface. Both contributions sum to 0.28 ± 0.01 which is in good agreement with in situ measurement of 0.25 ± 0.06. The major limitations of the model are that it deals with neither density changes nor fluid dynamics. Heat and mass transfer are governed by diffusion only. However, an effective diffusivity is introduced to portray some of the effects of the movement of the fluid. In spite of the simplifications, the model offers qualitative understanding of the role of diffusive processes in this porous layer.


Biological modification of carbonate chemistry in dense water outflows from the Mertz Polynya, East Antarctica

Elizabeth Shadwick, Bronte Tilbrook, Guy Williams, Steve Rintoul

Corresponding author: Elizabeth Shadwick

Corresponding author e-mail: Elizabeth.Shadwick@utas.edu.au

Polynyas, areas of open water with in the sea-ice pack, are often biologically productive, and are sites of enhanced air–sea exchange. The Mertz polynya is formed in East Antarctica by persistent katabatic winds and an ice barrier created by the Mertz Glacier Tongue. The region plays a significant role in the formation of dense shelf water (DSW) primarily due to salt rejection during intense sea-ice formation. Physical and biogeochemical observations indicate that primary production over the shelf in spring/summer preconditions the dense water outflows, supplying both organic material, and water depleted in CO2, to coral and sponge communities on the continental slope. The formation and export of DSW also plays a role in the transfer of anthropogenic CO2 to the deep ocean. In February 2010 a large piece of ice (~70 km long and ~35 km wide) calved from the Mertz Glacier Tongue, dramatically changing the regional fast-ice and pack-ice distributions. The physical and biogeochemical consequences of this event (based on observations pre- and post-calving) included significant surface freshening, a twofold enhancement of biological production, and an increase in the carbonate saturation state.


Satellite remote sensing of Antarctic sea-ice roughness using scatterometer data

Alex Fraser, Takenobu Toyota, Peter Jansen, Noriaki Kimura, Jan Lieser, Guy Williams, Ernesto Trujillo, Katherine Leonard, Ted Maksym, Robert Massom

Corresponding author: Alex Fraser

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

While sea-ice concentration is frequently observed from polar orbiting passive microwave satellite instruments, equivalent large-scale information on sea-ice sub-metre-scale roughness is currently unknown. Roughness on this scale is closely related to sea-ice thickness, and has implications for primary productivity near the ice/ocean interface, and flow-on effects for ecosystems. C-band (~5 GHz) off-nadir microwave backscatter strength from sea ice is sensitive to many physical characteristics of the sea ice. During summertime and early autumn, liquid water-related processes dominate backscatter variability (formation of melt water and superimposed ice, snow/ice interface flooding). However, during freezing conditions, backscatter variability reduces in the inner pack, and backscatter becomes more representative of snow/ice interface or air/snow interface roughness (dry snow on the order of 1 m thick is largely transparent at C-band, though experimental results suggest a significant portion of backscatter occurs at the air/snow interface). We demonstrate that EUMETSAT Advanced Scatterometer (ASCAT) C-band scatterometer isotropic roughness data are sensitive to sea-ice roughness. Validation is provided by helicopter-mounted C-band nadir backscatter radar data and lidar swath data acquired in early spring, 2007, during the Australian-led Sea Ice Physics and Ecosystems eXperiment (SIPEX) campaign, and newly-acquired measurements of rugosity and snow/ice interface roughness from Autonomous Underwater Vehicle (AUV)-derived multibeam sonar acquired during the follow-up marine science voyage, SIPEX II. Using high-resolution sea-ice motion vectors from passive microwave imagery, we also investigate the links between large-scale sea-ice convergence and sub-metre-scale roughness.


Observed and simulated changes in Antarctic sea ice and sea level pressure: anthropogenic or natural variability?

Will Hobbs, Marilyn Raphael, Nathan Bindoff

Corresponding author: Will Hobbs

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

Statistically significant changes in Antarctic sea-ice cover and the overlying atmosphere have been observed over the last 30 years, but there is an open question of whether these changes are due to multi-decadal natural variability or an anthropogenically forced response. A number of recent papers have shown that the slight increase in total sea-ice cover is within the bounds of internal variability exhibited by coupled climate models in the CMIP5 suite. Modelled changes for the same time period generally show a decrease, but again with a magnitude that is within internal variability. However, in contrast to the Arctic, sea ice tends in the Antarctic are spatially highly heterogeneous, and consideration of the total ice cover may mask important regional signals. In this work, a robust ‘fingerprinting’ approach is used to show that the observed spatial pattern of sea-ice trends is in fact outside simulated natural variability in west Antarctic, and furthermore that the CMIP5 models consistently show decreased ice cover in the Ross and Weddell Seas, sectors which in fact have an observed increase in cover. As a first step towards understanding the disagreement between models and observations, modelled sea level pressure trends are analysed using and optimal fingerprinting approach, to identify whether atmospheric deficiencies in the models can explain the model–observation discrepancy.


The increase of Antarctic sea-ice extent during the satellite era

John Turner

Corresponding author: John Turner

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

Since 1979 the extent of Antarctic sea ice has increased throughout the year, with the annual mean extent having increased at a statistically significant rate (p<0.01) of 1.2% per decade. However, this overall increase masks large regional changes, and particularly a couplet of decrease/increase across the Bellingshausen/Ross Seas. But the increase in the Ross Sea has been critical to the overall Southern Ocean positive trend and accounts for ~84% of the net Antarctic sea-ice extent increase. Off West Antarctica sea-ice anomalies have been strongly linked to changes in the near-surface wind field, although other processes, such as freshwater injection from basal ice shelf melting, oceanic change and ice–ocean feedback processes have also been suggested as controlling factors. However, the sea-ice trends in this sector correlate significant with the mean sea level pressure over the Amundsen-Bellingshausen Sea, which is the region of the climatological Amundsen Sea Low. This low has deepened since 1979, increasing the near-surface wind speeds over the Ross Sea, especially during the spring. The depth of the Amundsen Sea Low is significantly correlated with sea surface temperatures across the tropical Pacific Ocean and the phase of the El Nino–Southern Oscillation. High sea surface temperatures result in deep convection that establishes a Rossby Wave train to the Antarctic, giving a weak Amundsen Sea Low. However, since 1979 tropical conditions have shifted more to the La Nina phase, especially in the spring, deepening the Amundsen Sea Low, giving stronger southerly flow over the Ross Sea and increasing the sea-ice extent in this sector. Experiments with an atmosphere-only climate model forced with various sea surface temperature scenarios show the robustness of this teleconnection.


AFIN: An example of East Antarctic fast-ice mass-density

Petra Heil

Corresponding author: Petra Heil

Corresponding author e-mail: petra.heil@utas.edu.au

Mass-density is one of the state variables of sea ice that is not well known, especially for Antarctic sea ice. Knowledge of sea-ice density is important for assessing processes that occur in first-year and multi-year ice, and to correctly describe the sea-ice state within numerical models. It is also needed when deriving sea-ice thickness from freeboard or surface-elevation measurements. The latter aspect is of growing importance with increased access to aerial and satellite-derived freeboard and surface elevation data. Published data on sea-ice density are biased towards the Arctic. Reported densities span the range from 742 to 945 kg m–3, averaging at about 914 kg m–3. However, year-round density data collected off East Antarctica occupy a narrower range, with an average of 903 kg m–3. Here we present sea-ice density data of East Antarctic fast ice off Davis and Mawson stations, obtained from 1996 onward. Data collected since 2003 form part of Australia’s contribution to the Antarctic Fast-Ice Network (AFIN). The Davis fast-ice cores have mostly been obtained at the same location at maximum annual ice thickness, while data from Mawson also includes some austral summer samples. At Davis the maximum ice thickness varies significantly from year to year, however over our data series there appears to be a downward trend in ice mass-density. To explore physical drivers for this apparent change, we make use of our measurements of ice and oceanic surface salinity, as well as atmospheric near-surface parameters.


Physical sea-ice properties in the winter Weddell Sea in 2013

Stefan Hendricks, Sandra Schwegmann, Thomas Krumpen, Priska Hunkeler, Robert Ricker, Mario Hoppmann, Stefanie Arndt, Marcel Nicolaus

Corresponding author: Stefan Hendricks

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

The Antarctic sea-ice cover reached a significant above normal extent during the southern winter in 2013. This behavior stands in a significant contrast to a continuous reducing summer sea-ice cover in the northern hemisphere. Explanations require the understanding of the coupling of ice, ocean and atmosphere in both Polar Regions. However, few observations beside remote sensing products exist especially during the remote and harsh environment of the Antarctic winter season. The assessment of the mass and energy balance of the increasing Antarctic sea-ice cover therefore requires dedicated process studies. Here we present an overview and first results from two expeditions of the German research ice breaker Polarstern (ANT-XXIX/6 and ANT-XXIX/7) in the winter Weddell Sea between June and October 2013. One part of the measurement program was multi-scale sea-ice thickness measurements, ranging from regional airborne survey to local high resolution studies. Together with the study of snow depth and stratification we aim to assess the capability of CryoSat-2 to estimate Antarctic sea-ice thickness by satellite altimetry. Comparisons of sea-ice thickness and ice texture with data from few previous expeditions shall indicate consequences and possibly reasons of the changing Antarctic winter sea-ice cover.


Arctic sea-ice history from isotopic records

Claude Hillaire-Marcel, André Poirier, Anne de Vernal

Corresponding author: Claude Hillaire-Marcel

Corresponding author e-mail: chm@uqam.ca

The history of the Arctic Ocean and its climate remained poorly known until the 2004 IODP coring of Lomonosov Ridge (ACEX) sediments which revealed the existence of an Eocene ‘lake-stage’ prior to marine submergence. As a consequence of high latitude plate tectonics movements, the opening of Fram Strait, possibly related to the Popigai impact dated ca 36 Ma, allowed input of North Atlantic marine water into the Arctic basin and possibly led to inception of an Atlantic Meridional Overturning Circulation. Osmium isotope stratigraphy and Re-Os isochrons concur to assign a Late Eocene age (~36 Ma) to this event, thus an age in accordance with that of the Popigai impact. Surprisingly, the presence of sea-ice seems ascertained as early as the mid-Eocene (~ 45 Ma), during the lacustrine stage, i.e. at about the same time as sea ice/icebergs started to form in the circum-Antarctica. Tracers to reconstruct sea-ice/iceberg activity in the Arctic include sedimentological features (grain size, mineralogy...), microfossils linked to productivity in sea-ice and/or to the duration of the sea-ice cover (diatoms, ostracods, foraminifers, dinocysts...), organic compounds produced by such organisms (sea-ice biomarkers, terrigenous/marine biomarkers), and geochemical proxies (elemental/isotopic tracers of distal/proximal terrigenous supplies, stable isotopes in foraminiferal calcite...). A special issue of Quaternary Science Reviews in press, includes benchmark papers about such proxies and their use to reconstruct past-sea-ice dynamics. We will further illustrate here, specific information provided by radiogenic, U-series and stable isotopes in Arctic sedimentary sequence, notably the ACEX core. We will then examine information about the status of Arctic sea-ice during recent warm intervals, and discuss its potential resilience under a warmer globe.


Ecosystem modelling for sea-ice habitats: from plankton to penguins

Jessica Melbourne-Thomas, Merel Bedford, Elizabeth Fulton, Stuart Corney, Andrew Constable

Corresponding author: Jessica Melbourne-Thomas

Corresponding author e-mail: Jess.Melbourne-Thomas@aad.gov.au

Antarctic sea ice provides critical habitats for diverse biota at a range of spatial and temporal scales, from microscopic phytoplankton through to Antarctic top predators. Modelling approaches are critical to understanding and predicting the responses of these ecosystem components to climate change impacts. A series of recent, significant advances in modelling approaches for microbial communities associated with sea ice provide increased predictive power for these lower trophic level components of sea-ice ecosystems. However, models that represent sea-ice habitat characteristics that are important to species such as krill, penguins and seals are still lacking. We present a ‘toy’ model of an Antarctic ecosystem that incorporates both pelagic and sea-ice habitats and that represents the association of phytoplankton, zooplankton, krill, fish, penguins, seals and whales with these different habitats. By simulating a set of scenarios for change in seasonal sea-ice habitats we explore and evaluate appropriate within-model representations to realistically capture biological responses to habitat change. These scenarios also provide the first model predictions for how diverse sea-ice ecosystems might respond to climate-driven change in Antarctic sea-ice environments.


Statistical modelling of ice algal distribution in Antarctic sea ice

Natalie Radojcic, Jessica Melbourne-Thomas, Stuart Corney, Klaus Meiners

Corresponding author: Natalie Radojcic

Corresponding author e-mail: natalie.radojcic@gmail.com

Algae incorporated in sea ice plays an important role in primary production and therefore sea-ice algal concentration impacts polar ecosystems. In Antarctica sea ice algae are a vital food source for zooplankton during the winter months, with juvenile krill survival reliant on the availability of algae. As it is the bottom layer of algae that is freely available for pelagic feeding, it is the biomass of algae in this region that is of particular interest from an ecosystem perspective. We present a series of statistical models to determine the vertical distribution and integrated biomass of sea-ice algae in the Antarctic circumpolar region. Numerous variables impact the distribution of sea-ice algae. Current research suggests that sea-ice thickness, snow cover, the ratio of ice and snow thickness and exposure to photosynthetically active radiation (PAR) are particularly important variables. We use sea-ice chlorophyll a data from the Antarctic Sea Ice Processes and Climate – Biology (ASPeCt – Bio) circumpolar dataset to extend previous work and model sea-ice algal distribution using ice thickness, snow cover and exposure to PAR as predictors. Here, latitude is used as a proxy for exposure to PAR, based on the assumption that older ice (with greater cumulative exposure to PAR over the course of a season) will occur at more southerly latitudes. Given the potential for climate change to alter the physical determinants of ice algal distribution, modelling approaches such as ours are an important first step towards understanding the potential implications of climate change for Antarctic sea-ice ecosystems.


Controls on the causes and patterns of Arctic ice shelf calving events

Luke Copland, Colleen Mortimer, Adrienne White, Miriam Richer McCallum, Derek Mueller

Corresponding author: Luke Copland

Corresponding author e-mail: luke.copland@uottawa.ca

There have been rapid changes to the ice shelves of northern Ellesmere Island over the past decade, with a reduction in their area by ~50% since 2005, leaving a total of ~500 km2 today. Total losses over the past century have reached almost 9000 km2, but these losses have been uneven over space and time. To investigate the controls on these losses, a comprehensive assessment of the factors controlling ice shelf stability and mass balance along northern Ellesmere Island has been undertaken from an analysis of historical air photos, sea-ice charts, satellite imagery, field observations and previously published literature. This assessment suggests that the ice shelves have been weakening over at least the past century due to a long-term negative surface mass balance, reductions in input from surrounding glaciers, and reductions in the thickness and presence of adjacent sea ice, such as the loss of extensive regions of multi-year landfast sea ice. Many of these changes can be related to a rapidly warming climate in this region, in which air temperature has warmed by an average of ~0.48°C decade–1 since 1948, with warming particularly focused in the fall (0.70°C decade–1) and winter (0.68°C decade–1). Once pre-weakened, major ice shelf break ups have occurred during periods in the summer with offshore winds and little or no adjacent sea ice. These changes illustrate the cascading effects that sea-ice losses have had on adjacent cryospheric features in the Arctic Ocean, and the crucial role that sea ice plays in Arctic ice shelf stability. If the current climate regime continues and Arctic sea-ice losses continue at their present rate it is unlikely that any Arctic ice shelves will last beyond the middle of the 21st Century.


The influence of the atmospheric circulation on Antarctic sea ice during ice advance and retreat seasons

Marilyn Raphael, Will Hobbs

Corresponding author: Marilyn Raphael

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

Using observed sea-ice concentration data this research isolates distinct regions of sea-ice variability around Antarctica and determines advance and retreat seasons for each of them. The latter are then statistically linked with observed geopotential height data to determine the atmospheric circulation mechanism most closely associated with the sea-ice seasons. The Amundsen Sea Low, the Southern Annular Mode, Zonal Wave Three and ENSO, are shown to influence different regions of sea ice around the continent. The timing of their influences vary and these influences maybe of similar or opposing signs. The results clarify which atmospheric circulation mechanism is of primary importance to sea-ice variability in the different regions around Antarctica. As these circulation mechanisms respond to a changing climate, sea-ice variability around Antarctica will also change.


Modeling sea-ice extent and basal melting of Antarctic ice shelves at the Last Glacial Maximum

Kazuya Kusahara, Tatsuru Sato, Hiroyasu Hasumi, Ralf Greve

Corresponding author: Kazuya Kusahara

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

We estimate the sea-ice extent and basal melting of Antarctic ice shelves at the Last Glacial Maximum (LGM) by an ice-shelf/sea-ice/ocean coupled model. The cryosphere/ocean coupled model can roughly reproduce the sea-ice distribution and ice shelf melting amount under the present-day climate. The shape of Antarctic ice shelves at the LGM (location and draft) is derived from output of an ice-sheet/ice-shelf dynamics model, SICOPOLIS. A steady ice-sheet/ice-shelf configuration is derived from a simulation from the Eemian to the LGM. We have performed several numerical experiments in which the surface air temperature is modified. In this study, present-day surface boundary conditions are used as background forcing to drive the coupled model. Negative temperature anomalies are added on the background surface air temperature for producing the LGM surface conditions. Sea-ice extent in the experiments with 6–8°C cooling is consistent with that estimated from ice core data. In this model we estimate about 5000 Gt a–1 for the basal melting of the Antarctic ice shelves at the LGM. The basal melting amount at the LGM is about five times larger than that at present. The model results show that warm oceanic waters originated from the Circumpolar Deep Water in the Southern Ocean inflow directly into the ice shelf cavities and melt actively the ice shelf bases. In particular, there are very active melting areas at rates up to 30 m a–1 in the ice shelves in the Bellingshausen and Amundsen Seas. Change of the basal melting of Antarctic ice shelves is about 1000 Gt a–1 among the experiments of 2–10°C cooling compared to present-day conditions. We found that basal melting of Antarctic ice shelves is strongly determined by the ice shelf configuration, more than by surface air temperature change.


Reconstruction of circum-Arctic sea-ice cover during the Holocene

Marit-Solveig Seidenkrantz, Anne de Vernal, Sandrine Solignac, Christof Pearce, Longbin Sha, Simon Belt, Beth Caissie, Thomas M. Cronin, Marc Macias-Fauria

Corresponding author: Marit-Solveig Seidenkrantz

Corresponding author e-mail: mss@geo.au.dk

Is the extensive reduction of Arctic sea-ice cover of recent decades unprecedented in Earth history or are such variations in sea ice within normal natural variations for the Arctic region? Large amplitude variations of Arctic sea ice have been recorded in detail for the instrumental period, with a strikingly fast decrease in recent years, which might result in the disappearance of Arctic summer sea ice in the near future. The recent sea-ice decline has been partly attributed to the impact of anthropogenic climate changes. However, little is known about the natural variability of the sea-ice cover at multi-decadal to multi-millennial timescales. This is particularly critical since the few available historical time series and reconstructions of Arctic sea ice tend to demonstrate that the recent observational records of Arctic sea-ice cover is far from centennial to millennial averages. Hence, there is a need to establish longer sea-ice time series to document the full range of sea-ice variations under natural forcings, thus establishing a baseline for natural Arctic sea-ice variability. Here, we re-evaluated >80 proxy sea-ice records to test if sea-ice cover has remained stable during the Holocene or if at any point there may have been a severe reduction or full disappearance of summer sea ice. The reconstructions are based on biological (diatoms, dinoflagellate cysts, foraminifera, ostracods), sedimentological (IRD), and geochemical (IP25, PIP25) proxies as well as driftwood. Sea-ice reconstructions from all sites were grouped into classes: perennial (>80%), dense (50–80%), common (i.e. winter occurrence; 20–50%), occasional (<20%), rare (0% but with a few occurrence peaks) and absent (never). Reconstructions were made for selected time slices: 0–2000 kyr BP, 2000–4000 kyr BP, 4000–6000 kyr BP, 6000±500 kyr BP, 6000–8000 kyr BP and 8000–10,000 kyr BP. Our study shows that winter sea ice was present during the entire Holocene, but summer sea ice may have been reduced in some areas, albeit probably not absent, during the earliest Holocene. The Atlantic sector of the Arctic has undergone a subtle but consistent increase in sea-ice cover since the early Holocene. In the Nordic Seas and North Atlantic minimum sea-ice conditions are seen from ~10,000–6000 years ago, whereas in the eastern Labrador Sea minimum sea-ice occurred ~6000–4000 years ago. Since then, sea-ice cover has increased in nearly the entire Atlantic sector and the eastern Arctic region.


Evolution of melt pond geometry

Kenneth Golden

Corresponding author: Kenneth Golden

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

During the Arctic melt season, the sea-ice surface undergoes a remarkable transformation from vast expanses of snow covered ice to complex mosaics of ice and melt ponds. Sea-ice albedo, a key parameter in climate modeling, is largely determined by the complex evolution of melt pond configurations. In fact, ice-albedo feedback has played a major role in the recent declines of the summer Arctic sea-ice pack. However, understanding melt pond evolution remains a significant challenge to improving climate projections. Here we will discuss recent findings on the evolution of melt pond geometry. In particular, we have found that as the ponds grow and coalesce, their fractal dimension undergoes a transition from 1 to about 2, around a critical length scale of 100 square meters in area. As the ponds evolve they take complex, self-similar shapes with boundaries resembling space-filling curves. We will also discuss how mathematical models of composite materials and statistical physics, such as percolation and Ising models, are being developed to describe this evolution.


Electromagnetic monitoring of sea-ice processes

Kenneth Golden

Corresponding author: Kenneth Golden

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

Fluid flow through sea ice mediates a broad range of geophysical and biological processes in the polar marine environment. For example, the evolution of melt ponds and sea-ice albedo, which is critical to climate modeling, is constrained by drainage through the porous brine microstructure. Fluid flow also governs snow-ice formation, the evolution of the salt budget, and biomass build-up sustained by nutrient fluxes. However, for brine volume fractions below about 5%, columnar sea ice is effectively impermeable to fluid flow. In two different experiments conducted in the Arctic and Antarctic, we have found that this critical transition in fluid flow exhibits a strong electrical signature, with sea-ice resistivity sharply rising over three orders of magnitude near the brine connectivity threshold. The data are accurately explained by percolation theory for the effective resistivity of sea ice, treated as a composite of ice and brine, with the same universal critical exponent of 2 which captures the behavior of the effective fluid permeability. The data were obtained using cross-borehole tomography and a novel direct measurement technique on extracted sea-ice cores. Our results demonstrate that classical lattice models of phase transitions in statistical physics and theories of homogenization for composites can help unravel the complexity of transport in this multiscale random medium. The theory enables electrical classification of sea-ice layers in terms of their fluid flow properties, thus connecting specific electrical signatures to important transport processes such as melt pond drainage, CO2 pumping, and nutrient fluxes. These findings also have implications for measuring ice thickness, an important gauge of the impact of global warming.


The percolation threshold for fluid flow in Antarctic granular sea ice

Christian Sampson, Kenneth Golden

Corresponding author: Kenneth Golden

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

The fluid permeability of sea ice governs a broad range of physical and biological processes in the polar marine environment. For example, in the Arctic melt pond drainage is largely controlled by the fluid permeability of the ice, which in turn has a significant effect on ice albedo, a critical parameter in climate models. Algae depend on nutrients from the ocean transported through the porous microstructure of sea ice when it is permeable. However, columnar sea ice is effectively impermeable for brine volume fractions below about 5%, while above this threshold fluid can flow through the ice. In the Antarctic, granular ice with a different crystallographic structure makes up a significant portion of the ice pack. Data gathered during SIPEX II in 2012, as well as mathematical models, suggest that the percolation threshold for the fluid permeability of granular sea ice is around 10%, which is significantly higher than for columnar ice. These findings are significant, as both ecological models involving nutrient transport and physical process models must take this into account, such as in modeling snow-ice formation, an important component of ice production in the Southern Ocean.


On a positive sea-ice–ocean dynamic feedback in the Arctic Ocean

Wieslaw Maslowski, Joanne Haynes, Jaclyn Clement Kinney, Robert Osinski, Andrew Roberts

Corresponding author: Wieslaw Maslowski

Corresponding author e-mail: maslowsk@nps.edu

The reductions in the Arctic sea-ice cover due to warming climate have been most pronounced in the western Arctic, in particular over the Chukchi, East Siberian and Beaufort seas. The outflow of warm summer water from the Chukchi and East Siberian shelves, including coastal runoff, combined with local absorption of solar radiation and the transport due to the boundary current along the slope and eddies in the basin control the heat content in the upper ocean and contribute to the overall state of sea-ice cover. The oceanic heat accumulation below the surface mixed layer and above the halocline has significantly increased in the region since the late 1990s, according to observations and some models. Thinning and shrinking of the Arctic sea-ice cover can promote a positive, dynamically driven sea-ice–ocean feedback loop, in addition to the thermodynamically based snow/ice-albedo and water vapor–surface temperature feedback. We propose that on top of changes in sea-ice kinematics, a thinner/reduced ice cover promotes a positive feedback in the upper ocean, where excess heat has accumulated and the new ice regime, including increased sea-ice drift and resulting larger ice–ocean stress act to enhance turbulent mixing and upward heat entrainment making it available to further reduce ice growth in winter and melt ice in summer. Such processes and interactions currently are not realistically represented in Global Climate and Earth System models (GC/ESMs). We hypothesize that this is one of the primary reasons why their predictions of decline of the Arctic sea-ice cover under warming climate are too conservative. To investigate the operation of the proposed ice–ocean positive feedback loop we use the Regional Arctic System Model (RASM), which is a high resolution and limited-area analogue of a GC/ESM. Our model results, including realistic representation of sea-ice and ocean dynamics at process scales and feedback processes between the upper ocean, sea ice and atmosphere, corroborated with observational data appear to support both the existence and increasing importance of such a positive ice–ocean dynamic feedback in the Arctic.


Investigating iron and organic matter incorporation in growing sea ice

Julie Janssens, Jeroen De Jong, Jean-Louis Tison, Bruno Delille, Klaus Meiners, Gerhard Dieckmann, Delphine Lannuzel

Corresponding author: Julie Janssens

Corresponding author e-mail: Julie.Janssens@utas.edu.au

High concentration of exopolysaccharides (EPS) and iron have been found in sea ice surrounding the Antarctic continent. However, the mechanisms leading to the enrichment remain unclear. Scavenging of iron by organic matter in seawater and entrainment during sea-ice formation are thought to be responsible for the accumulation of iron in sea ice. Exopolysaccharides could also play a role in the iron passive chelative scavenging process in sea ice and in the increase of iron bioavailability. Our study investigates the processes responsible for the accumulation of iron (dissolved, particulate and total dissolvable fractions), EPS, dissolved and particulate organic matter, macro-nutrients (silicic acid, nitrate and nitrite, phosphoric acid and ammonium) and chlorophyll a in young sea ice during an Australian-lead spring voyage off East Antarctica (SIPEX II September–November 2012) and a German-lead winter voyage in the Weddell Sea (AWECS June–August 2013). We used a combination of field- (‘in situ’) and laboratory-based sea-ice growth time-series experiments. In addition, different types of newly formed sea ice (pancake ice, grey ice, frost flowers and slush) were collected during both voyages as a means to compare and validate the experimental data. To our knowledge, this is the first report on the biogeochemical properties of newly formed Antarctic pack ice samples in winter. Ice temperature, salinity and textures are also presented to support the biogeochemical observations at the onset of sea-ice formation.


Ice core proxy of Antarctic sea-ice extent

Mark Curran, Tessa Vance, Andrew Moy, Jason Roberts, Tasman van Ommen, Jan Lieser

Corresponding author: Mark Curran

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

Prior to the satellite era, ice core records have been used as a proxy for past sea-ice extent in Antarctica. This information is important for investigating long-term trends in sea-ice coverage. One of the first studies to correlate ice core data with modern satellite sea-ice extent records was conducted by Curran et al. They used data from the Law Dome ice core and found a significant correlation between the annual concentration of methanesulphonic acid (MSA) in the ice core and the maximum extent of sea ice (for Aug–Sept–Oct) in the region 80°–140°E. Curran et al. concluded there was a 20% decline in the sea-ice extent in this region between the 1950s and 1990s, and observed there was a high degree of decadal variability in the record. Support for this approach was provided by a study in Wilhelm II Land and an array of ice cores in West Antarctica. Recently a comprehensive review of sea-ice proxy information from polar ice cores outlined the importance of studies of this type. Here, we present an updated record for the MSA data from the Law Dome site (to 2013) and investigate the recent changes in Antarctic sea-ice extent. We also include a reanalysis of old Nimbus I satellite information from Sept 1964, which agrees with the sea-ice extent reconstruction.


Polar tropospheric aerosol formation – discovering a significant source to the Antarctic and Southern Ocean region

Ruhi Humphries, Andrew Klekociuk, Robyn Schofield, Andrew Robinson, Neil Harris, Melita Keywood, Jason Ward, Stephen Wilson

Corresponding author: Ruhi Humphries

Corresponding author e-mail: rsh615@uowmail.edu.au

The contribution of Antarctic aerosol formation to the Southern Hemisphere aerosol load, and consequently their impact on the Earth’s albedo, climate and chemistry, is currently unknown. In the Antarctic region, aerosol measurements have been limited primarily to boundary layer air-masses at spatially sparse coastal and continental research stations. Focussed aerosol studies in this region have been limited primarily to continental and coastal locations where permanent stations exist. Measurements in the sea ice have been limited to ice shelf measurements and those made as campaigns pass through the pack ice to and from the continent. Presented here are the results of the first focussed aerosol study in the Antarctic pack ice during polar spring off the coast of East Antarctica as part of SIPEXII. Boundary layer concentrations of ultrafine aerosols (3–10 nm diameter) were found to increase sharply as we moved south across the Polar Front, with median concentrations in the Polar cell of 530 cm–3 compared to 45 cm–3 in the Ferrel cell to the north. Polar cell aerosols showed no signs of growth, suggesting long-range transport. Using back-trajectory modelling and satellite data, we propose a model for new particle formation in the Antarctic free troposphere from precursors uplifted in the sea-ice region. After reaching the sea-ice surface, these aerosols formed in the free troposphere are found to escape to lower latitudes where they could have significant impact on the radiative balance of the whole Southern Ocean region.


Assessing the robustness of sea-ice classifications from polarimetric radar images

Mari-Ann Moen, Torbjørn Eltoft, Stian N. Anfinsen, Anthony P. Doulgeris, Angelika H.H. Renner, Sebastian Gerland

Corresponding author: Mari-Ann Moen

Corresponding author e-mail: Mari-Ann.Moen@uit.no

This paper investigates automatic segmentation and classification of polarimetric synthetic aperture radar (SAR) satellite images of Arctic sea ice. The objective is to assess the robustness of the results obtained under varying environmental conditions and sensor geometries. We obtain unsupervised segmentations of images from three consecutive days and consider two strategies for class labeling: The first is based on in-situ data and expert knowledge; the other uses the previous day labeling as prior knowledge in a supervised classifier. The results are compared and discussed. Sea ice maps are valuable for environmental monitoring and climate change surveys. SAR is considered one of the most important remote sensors for acquiring sea ice information, especially in Arctic areas where in-situ data is limited. During a field cruise in April 2011 measurements on first-year sea ice north of Svalbard were collected. Two categories of data were obtained: In-situ data (ice station measurements of surface properties, drift measurements, ice thickness measurements from an EM-bird, shipboard observations and photographs from ship and helicopter) and quadrature–polarimetric SAR data from the C-band satellite Radarsat-2. This investigation is based on three quad-pol images from consecutive days (11–13 April) with incidence angles varying from 24 to 41 degrees. Initially, the three SAR images were clustered into unlabeled segments using an unsupervised mixture of Gaussian clustering algorithm utilizing six features extracted from the polarimetric data and contextually smoothed with a built-in Markov random field. Two different labeling processes were conducted and the results were compared: (1) The 12 April image coincides with various ground truth sources, hence the segmentation result is labeled utilizing relevant in-situ measurements and polarimetric features derived from the SAR data. The other two images are subsequently labeled based on a visual inspection of the labeled 12 April image, matching of the multidate segments in feature space, and Pauli images of the polarimetric SAR data. (2) We used the labeled 12 April image as training data to do a supervised classification of the remaining two unlabeled segmentations. Each segment in the images from the 11th and the 13th were assigned to the statistically nearest class. The consistencies between the results are compared and the effects of changing incidence angles and meteorological conditions are discussed.


Temporal and spatial changes in biogeochemical properties of Antarctic sea-ice during spring and early summer in the western Weddell Sea: the sulfur case

Jacqueline Stefels, Mattias Steffens, Stathys Papadimitriou, David N. Thomas

Corresponding author: Jacqueline Stefels

Corresponding author e-mail: j.stefels@rug.nl

In polar regions, major climate-gas exchange events occur during the period of sea-ice melt and associated growth of phytoplankton. An important compound that is released is dimethylsulfide (DMS), a potentially climate-cooling gas. The non-volatile compound dimethylsulfoniopropionate (DMSP) is the precursor of DMS and is solely produced by algae. Algae produce DMSP to keep osmotic balance, especially under extremely low temperatures and high salinities found in sea ice. Another important sulfur compound is dimethyl sulfoxide (DMSO), the (photo-)oxidation product of DMS and potentially a sink for reduced sulfur. DMSP, DMSO and DMS are all produced in sea ice at concentrations that are three orders of magnitude higher than in surface waters. A general rule for open ocean systems is that only a few percent of the originally produced DMSP is emitted to the atmosphere as DMS, since the majority of DMSP is converted by bacteria or photo-oxidized to DMSO. What happens in sea ice is of interest given the extremely high concentrations and close connection with the atmosphere. Both the temporal variation of concentrations and conversion processes as the spatial heterogeneity of the S-compounds is of importance in order to realistically model sea-ice associated biogeochemical processes and their potential impact on climate. Here we present on overview of biogeochemical parameters followed during a temporal study on an ice floe in the Weddell Sea with RV Polarstern (ANTXXII-2: ISPOL) and a spatial study through first- and multi-year sea ice of the Western Weddell Sea (ANTXXIII-7: WWOS). During ISPOL, physicochemical parameters and biological components were followed within a plot of 12 m x 12 m over the course of 1 month and analyzed every 5th day in high-resolution sea-ice profiles. Photochemical oxidation and grazing activity appeared important processes in the conversion of S-compounds. The spatial and temporal heterogeneity of both the concentrations and conversion pathways of S-compounds where investigated with the use of stable isotope additions. The data provide a showcase for the importance of sea ice as a source for DMS(O/P) in the marginal ice zone.


Important role for ocean warming and increased ice-shelf melt in Antarctic sea-ice expansion

Caroline Katsman, Bert Wouters, Sybren Drijfhout, Richard Bintanja, Geert Jan van Oldenborgh

Corresponding author: Richard Bintanja

Corresponding author e-mail: bintanja@gmail.com

Changes in sea ice significantly modulate climate change because of its high reflective and strong insulating nature. In contrast to Arctic sea ice, which is retreating dramatically, sea ice surrounding Antarctica has expanded over the past decades. Several explanations have been put forward for this Southern Hemispheric sea ice expansion, the majority of which invoked dynamical atmospheric changes that induce atmospheric cooling. In this presentation it will be shown that warming deep ocean water and accelerated basal melting of Antarctic ice shelves have likely contributed significantly to sea-ice expansion. Specifically, observations indicate that melt water from Antarctica’s ice shelves accumulates in a cool and fresh surface layer that ‘shields’ the surface ocean from the warmer deeper waters that are melting the ice shelves. Simulating these processes in a coupled climate model (EC-Earth) it is found that, indeed, cool and fresh surface water from ice-shelf melt leads to expanding sea ice in austral autumn and winter through reduced vertical ocean mixing. Although changes in atmospheric dynamics contribute to sea ice increase regionally, the analyses indicate that the overall upward sea-ice trend is dominated by increased ice-shelf melt. This negative feedback is sufficiently strong to counteract Southern Hemispheric atmospheric warming, and is shown to continue to do so in the (near) future when ice shelf melt from Antarctica becomes even more widespread.


Diurnal observations of coincident C, X, Ku-band microwave backscatter and snow geophysical properties over smooth first-year sea ice: implications for synthetic aperture radar remote sensing

Jagvijay P.S. Gill, John Yackel, Torsten Geldsetzer

Corresponding author: Jagvijay P.S. Gill

Corresponding author e-mail: jpsgill@ucalgary.ca

The cloud penetrating capability of microwaves at C, X and Ku frequencies has generated increased importance and use of Synthetic Aperture Radar (SAR) data for snow and sea ice studies in the Arctic. Both ascending and descending orbital modes are frequently used and at times considered interchangeable or equivalent in backscatter response for the same day acquisitions. In the current study we present the observations of the diurnal coincident geophysical and dielectric properties of snow covered smooth first-year sea (SFYI) ice and the polarimetric backscatter response at C, X and Ku-band frequencies. The objective is to analyze how snow properties and the associated multi-frequency backscatter changes over a period of twenty-four hours using measurements acquired every two to three hour. The multi-frequency microwave data is acquired in-situ in Resolute Bay; high Arctic over SFYI using ground based scatterometers. Snow geophysical and dielectric properties are measured within 10–15 m of the scan area. The analysis is performed at two temperature regimes: (1) late winter period with diurnal air temperatures ranging between -11°C and -17.5°C, and (2) early melt period with diurnal air temperatures ranging between -5°C and -11°C. Our preliminary analysis demonstrates that polarimetric backscatter at radar incidence angle from 20° to 60° is highly correlated with diurnal variations in the air temperature and subsequent snow properties for the late winter period. This difference in backscatter at C-band for the highest and the lowest temperature of the day is to the order of 3–5 dB. For the early melt period, we do not observe any diurnal variations in the polarimetric backscatter with changes in the diurnal air temperatures and snow properties. This suggests that SAR ascending and descending orbital data must be used cautiously for snow and sea ice studies at certain temperature conditions.


Uncertainties in thickness estimates of floating ice when applying buoyancy assumption

Jan Lieser, Benjamin Galton-Fenzi, Jason Roberts, Robert Massom

Corresponding author: Jan Lieser

Corresponding author e-mail: jan.lieser@utas.edu.au

The largest impediment to accurately measuring changes in ice volume, both of land origin and sea ice, is with uncertainties in ice thickness estimates. Since the satellite era, the extent and seasonality of sea ice and the location and size of ice shelves and icebergs is quite well known; but those satellites provide only a two-dimensional representation of a three-dimensional entity. Surface elevation measurements by air- or space-borne altimeters provide an estimate of the ice or snow–air interface above a reference surface, the freeboard. In the case of floating ice the reference surface is usually the open water surface. Computing the thickness and subsequently volume of floating ice from altimetry data relies critically on the validity of the parameters used when converting surface elevation measurements into ice thickness. The underlying assumption is that ice and ocean are in hydrostatic equilibrium derived from the buoyancy principle first described by Archimedes in ‘On floating bodies’ proposition 5 discovered in 212 B.C. In this study we present a numerical analysis of the most widely used formula to compute ice thickness from freeboard measurements. We are investigating the errors associated with reasonable uncertainty estimates of the parameters of the buoyancy conversion when applied to surface elevation estimates in the Antarctic marine cryosphere.


A simple parameterization of sub-ice-shelf melting constrained by results from the ocean model ROMS, and its implementation in the ice sheet model SICOPOLIS

Tatsuru Sato, Ben Galton-Fenzi, Ralf Greve

Corresponding author: Tatsuru Sato

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

Ice-shelf basal melting plays a crucial role for the past and future evolution of the Antarctic ice sheet and its contribution to sea level change. To capture the general structure of the Antarctic basal mass balance, output of the ocean model ROMS is analyzed: mass balance at the ice base, temperature, sea-ice conditions etc. We find strong basal melting not only near the grounding zone, but also around the calving fronts, in agreement with observations. This demonstrates the importance of the ocean surface conditions for the mass balance of the ice shelves. We describe a simple, physically-based parameterisation that calculates the basal melting of ice shelves. The parameters are constrained based on the analysis of the ROMS output, which can be done differently for various Antarctic sectors conditions in order to achieve reasonable agreement with the modern spatial distribution of ice shelf basal melting. The parameterisation is then implemented in the Antarctica module of the dynamic/thermodynamic, large-scale ice sheet model SICOPOLIS. We discuss the response and long-term evolution of the ice sheet to various oceanic and climatic forcings. The model will help provide new insight into estimating both past and future contributions of the ice sheet to sea level rise and the heat budgets of the ocean.


Comparing methods of measuring sea ice density in the East Antarctic

Jennifer Hutchings, Petra Heil, Jan Leiser, Roger Stevens, Oliver Lecomte

Corresponding author: Jennifer Hutchings

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

Sea ice density is a key parameter in estimation of sea ice thickness using remote sensing methods. Density is typically between 840 and 940 kg m–3, dependent on where the ice is in relation to the waterline and the ice type. There are relatively few measurements taken of sea ice density around Antarctica, and the mean ice density often used (around 910 kg m–3) is estimated from primarily Arctic data. During the Sea Ice Physics and Ecosystem eXperiment in September to November 2012 (SIPEX-2), sea ice density was measured with several techniques. Error estimates for weight/volume and water volume/equation of state of salt water methods are presented. With 10 cm core samples, the error on individual density estimates is 10 kg m–3. Errors increase to 30 kg m–3 if a salt water volume method is used, and there appears to be a high bias with this method. Larger errors are apparent for smaller block sizes, increasing to 70 kg m–3 for machined blocks from 5cm core sections. The cruise sampled ice pack in the East Antarctic that was highly deformed first year ice, 1.2m thick on average, consisting of 6–7/10 columnar ice and 3/10 granular ice in layers consistent with multiple rafting. Ice density was found to be consistently lower than 910 g m–3, with columnar ice mean density of 870 g m–3, and lower densities for granular and snow ice fractions. At two different core sites we estimate the mean density of the ice to be 870 kg m–3 and 800 kg m–3. The lower density at the second core site reflecting the higher percentage of granular and suposition ice at that site.


Reversibility of Antarctic sea ice decline

Jeff Ridley, Helene Hewitt

Corresponding author: Jeff Ridley

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

Heat transport to the Antarctic coast, in the HadGEM2-ES climate model, is modulated by large transient barotropic eddies spawned from the Brazil–Falklands confluence. The ocean eddies travelling eastwards in the Antarctic Circumpolar Current, arrive at the Antarctic coast off Enderby Land (~60°E) and the transported heat is picked up by the easterly shelf-break current to dissipate in the Weddell Sea. The associated heat release influences sea ice cover, and is responsible for some of the internal variability in the winter ice extent. A climate simulations in which the atmospheric CO2 is ramped up to four times pre-industrial levels then ramped back down, indicates that the above heat transport in non-reversible, resulting in prolonged ice loss from the Weddell Sea.


Simulation of the energy and freshwater budget of the Arctic by a CMIP5 model

Alex West, Matt Collins, Helene Hewitt, Ann Keen, Jeff Ridley

Corresponding author: Alex West

Corresponding author e-mail: alex.west@metoffice.gov.uk

Coupled climate models are among the most useful tools in investigating future Arctic climate change. Many are now able to capture large scale aspects of past and present Arctic climate, but in recent years a mismatch between trends in sea ice extent in models and observations has been observed. Here, an attempt is made to better understand the strengths and weaknesses in Arctic climate simulation by examining terms in the energy and freshwater budget of the Arctic in HadGEM2-ES, one of the models submitted to the CMIP5 ensemble. The Arctic is split into the components of atmosphere, ice, ocean and land; fluxes passing between these components, and into the region from midlatitudes, are calculated on annual and monthly timescales, and are compared to observations in regions where these exist. A number of clear discrepancies between model and observations are identified, for example a too-high conductive flux through sea ice, and greater than observed upwards shortwave in the early summer. Attempts are made to understand the causes of these discrepancies, and to gauge their likely impact on the accuracy of future Arctic climate projections.


Drivers of Arctic sea ice decline in the HadGEM1 coupled model

Ann Keen, Helene Hewitt, Jeff Ridley

Corresponding author: Ann Keen

Corresponding author e-mail: ann.keen@metoffice.gov.uk

The Arctic sea ice cover has been declining for several decades, with an associated reduction in the ice thickness and volume, and climate model projections to the end of the century show a continuing decline in both extent and volume. Using projections from the HadGEM1 global coupled model, we analyse changes in the heat budget of the sea ice and overlying snow in order to understand the factors driving the decline of the Arctic sea ice in this model. Most of the long-term decline in ice volume in the HadGEM1 model is due to an increase in ice loss during the summer melting period. As the atmosphere warms there is more surface melt in June, and extra melting at the sides and base of the ice due to extra heat from warmer ocean surface in August. In June the surface albedo of the ice reduces as the surface temperature warms, allowing extra surface melting. As the ice retreats, in situ warming of the ocean surface results in an increased ocean to ice heat flux, which peaks during August when the ice extent is approaching its seasonal minimum. In both caes an ice albedo feedabck plays an important role in determining which months have extra melting, illustrating the importance of having a physically realistic representation of albedo in models used to make projections of future climate change.


A method to automatically determine sea level for referencing snow freeboards and computing sea ice thicknesses from NASA IceBridge airborne LIDAR

Xianwei Wang

Corresponding author: Xianwei Wang

Corresponding author e-mail: xianweiw@vip.qq.com

The NASA Operation Ice Bridge flights have obtained critical observations for Earth’s polar ice since ICESat stopped collecting data in 2009. This study develops a fully automatic method in processing IceBridge Airborne Topographic Mapper (ATM) altimeter L1B data (one elevation per 3–4 m horizontally) to derive a local sea level height for referencing snow freeboards and then computing sea ice thicknesses. Four 30 km L1B profiles (A, B, C and D) flown on October 30, 2010 and October 21, 2009 over the Bellingshausen Sea in Antarctica are selected. The local sea level reference is first obtained by visual examination of ATM L1B heights over leads or thin ice identified on images simultaneously acquired from the Digital Mapping System camera (called manual selection). This sea level reference is then used as ground truth to validate sea level heights derived by automatic calculations using five thresholds of 2%, 1%, 0.5%, 0.2% and 0.1% of the lowest L1B data over a DMS image area (called auto1 selection). The L1B_0.1% method gives a similar sea level height as from the L1B manual selection, by mean (absolute) difference of 0.04 (0.09) m. Then a fully automatic method averaging the lowest 0.1% L1B data within a ± 5 km section is also compared and validated with the manual selection (called auto2 selection) by a mean absolute difference of 0.12 m to those from the L1B manual selection on October 30, 2010 and 0.08 m on October 21, 2009. The sea level heights demonstrate a near linear gradient of 0.02 m km–1 to 0.03 m km–1 within each ~30 km L1B profile along the flight track from section A to D. The resulting mean snow freeboards are 0.95 m, 1.07 m, 0.73 m, and 1.04 m on sections A, B, C and D, respectively. The mean ice thicknesses estimated by the buoyancy equation from derived snow freeboard (with zero ice freeboard assumption) are 3.01 m for section A, 3.37 m for B, 2.30 m for C, and 3.28 m for D.


Iron contribution from melting icebergs estimated from airborne remote sensing data

David Prado, Steve Ackley, Hongjie Xie

Corresponding author: David Prado

Corresponding author e-mail: myu357@gmail.com

Melting icebergs have been considered a potential source of iron to fuel phytoplankton blooms in the Southern Ocean, a region generally considered iron poor due to the lack of terrestrial iron sources nearby. Regional estimates of iceberg iron contributions however have not been available since simultaneous data on iceberg size distributions, both the low freeboard (elevation) (~5 m) and small area (~100 m2 area) ‘bergy bits’ that can melt in the photic zone are needed. Two sensors, airborne Lidar for elevation and Digital Mapping System (DMS) photography for surface areas, flown on the NASA IceBridge mission were used to provide this information at the high resolution needed to estimate bergy bit volumes. Data were taken from the IceBridge missions in the Bellinghausen-Amundsen Seas in 2010/11. Bergy bits account for 0.68 km2 (2010) and 0.17 km2 (2011) of the flight lines providing potentially numerous, small areas of nutrient rich water for algal blooms. Based on Digital Mapping System imagery and Airborne Topographic Mapper LIDAR elevations the total volume can be estimated, further accounting for iceberg shape. Previous studies have shown a wide range of Fe concentrations within icebergs (0.5 nMol L–1–120 nMol L–1). Bergy bits were therefore estimated to deliver iron from 2.33 × 109 to 5.58 × 1011 nMol (2010) and 3.88 × 108-9.32 × 1010 nMol (2011) based on volume estimates derived from airborne data, distributed over the total areas measured by DMS photography under the flight lines. Since the bergy bits are not evenly distributed over the flight lines, the concentration of bergy bits is used to delineate potential ‘oases’ and ‘deserts’ of biological activity in the Bellingshausen-Amundsen Seas.


A viscous–elastic–brittle rheology for sea ice modeling

Véronique Dansereau, Jérôme Weiss, Pierre Saramito

Corresponding author: Véronique Dansereau

Corresponding author e-mail: vero_d@mit.edu

In recent years, statistical analysis of available ice buoy drift and RGPS data have revealed the strong heterogeneity and intermittency of Arctic ice pack deformation and thereby demonstrated that the viscous–plastic (VP) rheology widely used in climate and operational models does not simulate adequately the mechanical behaviour of sea ice. A new rheological framework named ‘elasto–brittle’ (EB) has therefore been developed as an alternative to the VP model, which combines the linear elasticity of a continuum solid, a Mohr–Coulomb criterion for brittle failure and a progressive damage mechanism for the elastic modulus that allows for long-range elastic interactions inside the pack. Recent implementation of this rheology into 3-days stand-alone realistic simulations of the Arctic ice pack without advection reproduced the strong localization of damage and agreed well with the deformation fields estimated from RGPS data. In the context of longer-term simulations of ice conditions and coupling to an ocean component, a suitable rheological framework should however distinguish between the permanent and recoverable (elastic) deformations in order to estimate the adequate ice drift velocities from the computed displacements, i.e. allow the passage from small to large deformations. To achieve this, a viscous relaxation term is added in the elastic constitutive relationship of the EB model together with an ‘apparent’ viscosity that evolves according to the local thickness, concentration and damage of the ice, much like the elastic modulus. The coupling between the level of damaging and both mechanical parameters is such that within an undamaged ice plate the viscosity is infinitely large and deformations are strictly elastic, while along highly damaged zones such as leads the elastic modulus vanishes and most of the constrain is dissipated through permanent deformations. In this augmented EB model the irreversible and recoverable deformations are solved for simultaneously, hence ice drift velocities are defined naturally. This new rheological framework is presented along with preliminary results of simplified numerical experiments.


Influence of oceanic eddies on sea ice variability: a winter study of numerical model results and observations in the Western Arctic Ocean

Joanne Haynes, Wieslaw Maslowski, Robert Osinski, Jaclyn Clement Kinney

Corresponding author: Joanne Haynes

Corresponding author e-mail: joanne.haynes@defence.gov.au

The rapid decline in Arctic sea ice over the past few decades has prompted scientists to better understand the factors driving sea ice variability. Analyses and syntheses of numerical model results and available observational data are presented in order to advance the understanding of critical processes and feedbacks affecting the oceanic forcing of sea ice in the western Arctic Ocean, where melt has been particularly pronounced. Observational data from ice-tethered profilers (ITPs) and co-located ice mass-balance buoys (IMBs) are analyzed over the Beaufort Sea, with a focus on the entrainment of heat into the mixed layer. Results indicate that mixed layer warming and freshening driven by mesoscale eddies may locally affect the thermodynamic mass balance of sea ice, especially in winter. To obtain a qualitative measure and spatial distribution of mesoscale eddy activities in the western Arctic, results from two numerical ice–ocean models, an eddy permitting (~9 km) and an eddy resolving (~2.3 km), are analyzed. Gains in the eddy-resolving model due to its explicit representation of the local Rossby radius of deformation (or order 10 km or less) are also evaluated.


An assessment of the reliability of sea-ice motion and deformation from SAR data

Stefanie Linow, Thomas Hollands, Wolfgang Dierking

Corresponding author: Stefanie Linow

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

To understand the interaction between ocean and atmosphere in the polar regions, it is necessary to consider sea ice, as it influences heat exchange and salinity transport. The motion of sea ice is therefore a key factor to improve our understanding of climate processes. Sea-ice motion can be observed from space by SAR sensors and quantified by drift detection algorithms. Due to the scarcity of field observations, it remains a challenging task to validate the resulting motion fields. In our presentation, we will show a detailed analysis of the quality of sea-ice motion fields derived from SAR data. To this end, we analyzed a number of motion fields from the region of the Ronne polynia in the Weddel sea and from Fram Strait. This dataset gives us a large variety of sea ice conditions for evaluation purposes. We derived a quality indicator for sea ice motion fields which is independent of field data and evaluated it with reference data from different sources. Together with the motion field, sea-ice deformation can in principle be retreived from SAR images. Similar to ice motion, it is very difficult to obtain field data to validate of the results. We analyzed ice deformation fields derived from satellite data and examined the possibilities to assess the reliability of the resulting information.


Modelling sea ice in a coupled atmosphere–ocean chemistry–climate model

Olaf Morgenstern, Guang Zeng, Andrew Klekociuk, Mark Curran, Luke Abraham, Manoj Joshi, Jonathan Gregory, Annette Osprey

Corresponding author: Olaf Morgenstern

Corresponding author e-mail: olaf.morgenstern@niwa.co.nz

Sea ice is amongst the most challenging elements of the climate system to adequately represent in climate models. Models used for the 5th Assessment Report of the Intergovernmental Panel for Climate Change (IPCC) produce negative or insignificant trends for total sea ice extent in the Southern Ocean, whereas observations show a positive trend in Antarctic sea ice since the onset of space-based observations in 1978. Hence it is presently not understood why Antarctic sea ice is behaving differently from Arctic sea ice, which is in steep decline. Here we present simulation results using a coupled atmosphere–ocean chemistry–climate model. The model comprises a deep ocean, a state-of-the-art physical formulation of sea ice, and interactive whole-atmosphere chemistry. Simulations presented here include a small ensemble using all anthropogenic forcings, and sensitivity simulations in which either long-lived greenhouse gases or ozone-depleting substances are held constant. The simulation without ozone depletion shows a near-continuous, long-term decline of Antarctic sea ice extent, as expected in a warming climate. By contrast, the full-forcing simulations, and the simulation in which long-lived greenhouse gases are held constant, exhibit a steep decline of Antarctic sea ice extent before 1980 followed by four decades or more of high variability and on average an increasing trend in Antarctic sea ice. During the 2020s, Southern Ocean sea-ice extent starts to shrink. We will assess the realism of these developments also using sea ice proxy data derived from ice cores. The model results suggest that the growth in Antarctic sea ice characterizing the last few decades is driven by Antarctic ozone depletion, the ‘ozone hole’. This finding is generally consistent with other modelling studies which implicate ozone depletion in recent climate change across the Southern Hemisphere. Further work is needed to characterize in more detail the physical links in the atmosphere and ocean which connect sea ice extent to ozone depletion, and also to understand remaining pertinent model problems e.g. regarding total sea ice extent and also regional patterns of trends.


Time series space-borne Ku-band scatterometer and MODIS albedo observations over landfast first-year sea ice for snow thickness discrimination

John Yackel, Torsten Geldsetzer, Jagvijay P. S. Gill, Girish Bhardwaj

Corresponding author: John Yackel

Corresponding author e-mail: yackel@ucalgary.ca

Daily observations of space-borne Ku-band (QuikScat) scatterometer and MODIS-derived albedo are used to investigate geophysical changes to snow thickness distributions over first-year sea ice. Data are presented for in-situ measured snow thickness distributions in Franklin Bay, Canadian Arctic, during the late-winter to spring transition in 2008. Observations were made during the International Polar Year (IPY) Canadian Arctic Shelf Exchange Study (CASES). Previous research suggests that a thin snow cover on first-year sea ice will exhibit an early, rapid and homogeneous transition to extensive melt ponds or a fully flooded ice surface; in contrast, a thicker snow cover will exhibit a slower transition to a highly variable melt pond and snow patch icescape. In-situ measured snow thickness endmembers (thin and thick) are used to test the temporal co-evolution of the microwave and optical data for differences in timing and magnitude. Microwave and optical data thresholds are used to identify significant geophysical changes. Local meteorological data and snow-melt modeling provide supporting evidence. Results show that snow melt onset is readily observed in the microwave data, in contrast to minor changes in the optical data. During the transition to melt ponds, microwave and optical data exhibit opposite temporal trends. Pond formation is evident in the optical data, whereas ambiguity may exist in the microwave data. The timings between the microwave and optical thresholds are related to the snow thickness distributions. This suggests that bulk snow thickness on first-year sea ice can be differentiated by observing the co-evolution of microwave and optical data signatures.


Spring snow depth on arctic sea ice using the icebridge snow depth product

Melinda Webster, Ignatius Rigor

Corresponding author: Melinda Webster

Corresponding author e-mail: melindaw@uw.edu

Snow has dual roles in the growth and decay of Arctic sea ice due to its high albedo and insulation effects. Knowing snow thickness and distribution are essential for understanding and modeling sea ice thermodynamics and the surface heat budget. Therefore, an accurate assessment of the current snow cover is needed for identifying its impacts in the changing Arctic. This study assesses springtime snow conditions in the Arctic using airborne snow thickness measurements from Operation IceBridge (2009–12). The 2012 data were validated with coordinated in situ measurements taken in March 2012 during the Bromine Ozone and Mercury EXperiment (BROMEX) field campaign, yielding a statistically significant correlation coefficient of 0.59 and RMS error of 5.8 cm. The comparison between the IceBridge snow thickness product and the 1954–1991 Soviet drifting ice station data suggests the snow cover has thinned by 29% in the western Arctic, and 54% in the Beaufort and Chukchi seas. The relationship between the 2009–12 snow depth distribution and sea ice freeze-up dates was statistically significant, with a correlation coefficient of 0.59. These results may help us better understand the surface energy budget in the changing Arctic, and may improve our ability to predict the future state of the sea ice cover.


Very thick and heavily deformed Antarctic sea ice captured in 3-D by autonomous underwater vehicle

Guy Williams, Jeremy Wilkinson, Ted Maksym, Clay Kunz, Hanumant Singh

Corresponding author: Guy Williams

Corresponding author e-mail: guy.darvall.williams@gmail.com

Sea-ice thickness is a fundamental component of the polar climate system and there is an urgent need to advance our capability to monitor it from space and to model its response and feedback to climate change. Whereas previous in situ observations in support of these efforts have been restricted to point measurements, a new generation of autonomous underwater vehicles (AUV) are delivering unique 3-D floe-scale maps of sea ice draft. Here we present sea ice draft observations from ten floes (up to 400 m2) during two recent AUV expeditions to the near-coastal regions of Weddell/Bellingshausen and Wilkes Land sectors in early spring. These data provide the first complete statistical characterisation of sea ice draft morphology, providing new insights into ecosystem habitats and the role of deformation processes in controlling total sea ice volume. We find mean drafts ranging from 1.4 m to 5.5 m, with maximum drafts up to 17 m and drafts >5 m accounting for 10–30% of the mean, these are the thickest observations of Antarctic sea ice to date. Similarly ‘thick’ ice is being reported from new remote sensing products in areas outside of these near-coastal regions, prompting the question: ‘Are we underestimating Antarctic sea ice thickness?’.


Studying Antarctic minke whales in sea ice regions in East Antarctica

David Peel, Natalie Kelly

Corresponding author: Natalie Kelly

Corresponding author e-mail: Natalie.Kelly@csiro.au

Antarctic minke whales (Balaenoptera bonaerensis) are the most common and one of the smallest baleen whales in the Southern Hemisphere. According to vessel-based assessments of Antarctic minke whales in the Southern Ocean, there has been an apparent decline in their total (i.e. circumpolar) abundance of around 30% in the decade between the late 1980s and late 1990s. One compelling hypothesis to explain, this decline – at least to some degree – is that intra- or inter-annual changes in the position of the sea-ice edge, in concert with changes in the relative proportion of minke whales in the ice, may have decreased the number of animals available to be counted in later years by research vessels, which operated outside of ice-covered regions each summer season (and consequently leading to lower abundance estimates). In order to begin to estimate how many minke whales might be inside sea-ice areas in the present day, including how this varies spatially – which could provide an indication of past dynamics in minke whale distribution – and to understand their habitat preferences, the Australian Government undertook two aerial surveys in East Antarctica during the 2008/09 and 2009/10 austral summers. These surveys covered a range of ice habitats, between the continent out to the open water, and were the first fixed-wing aerial surveys for cetaceans in the Antarctic. Results from these aerial surveys indicate that whilst the sea ice areas in East Antarctica probably do not support large numbers of Antarctic minke whales during the summer months (relative to big, productive embayments like the Ross and Weddell Seas), intra- and inter-annual variations in their observed densities may go at least some way towards explaining differences in estimated abundances of this species in East Antarctica between the late 1980s and late 1990s (albeit these being small compared to changes at the circumpolar level).


Modeling sea ice as a granular material

Keguang Wang

Corresponding author: Keguang Wang

Corresponding author e-mail: keguang.wang@met.no

We present the basic framework for the modeling of sea ice as a two-dimensional granular material. The sea ice is assumed to fail when its principal internal ice stresses are located on the curved diamond yield curve. The deformation of the sea ice under failure is assumed to follow the co-axial flow rule. Sea-ice advection is modeled using the particle-in-cell technique. The model is used to simulate the motion and deformation of the Arctic sea ice with a grid resolution of 10 km. The initial ice concentration is obtained from the AMSR2 observations, and the initial ice thickness is obtained from SMOS ice thickness data together with the estimate from OSISAF ice type data. We show that the model is of high capacity in modeling the highly localized sea-ice deformation field and the discontinuous sea-ice velocity field. The resulting ice concentration field after 6 days captures the major fracture features in the sea ice when compared with the MODIS sea-ice observations.


Sea-ice production and dense shelf water formation around Prydz Bay

Guy Williams, Fabian Roquet, Laura Herraiz-Borreguero, Takeshi Tamura, Mark Hindell

Corresponding author: Guy Williams

Corresponding author e-mail: guy.darvall.williams@gmail.com

Antarctic Bottom Water (AABW) production, from the downslope mixing of Dense Shelf Water (DSW), supplies the abyssal layer of the world ocean and is vital to both physical and biogeochemical climate cycles. Recently, the first observations of overflows of newly formed AABW on the continental slope near 68šE have refocused attention on the enhanced sea ice production in the polynyas of Cape Darnley and Prydz Bay. Here we examine the DSW sources for this ‘Cape Darnley’ Bottom Water (CDBW), using unique data from instrumented southern elephant seals (Mirounga leonina). We present the general circulation and distribution of inflowing modified circumpolar deep water from offshore and ice shelf water from the Amery Ice Shelf. Time series of salinity at 300 m depth, during the wintertime occupation of Barrier, Davis, Mackenzie Bay and Cape Darnely polynyas by the seals, show the first observational evidence that Prydz Bay provides a distinct, albeit lower-salinity, contribution of DSW to Cape Darnley bottom water. These time series are converted into rare wintertime estimates of sea-ice production, whose temporal variability can be used to evaluate both satellite remote sensing products and numerical model output.


Ice–ocean drag in the Arctic in late Summer

Achim Randelhoff, Arild Sundfjord, Angelika H.H. Renner

Corresponding author: Achim Randelhoff

Corresponding author e-mail: achim@npolar.no

Ice–ocean fluxes of momentum and buoyancy interact with the under-ice turbulent boundary layer and, together with the atmospheric boundary layer, control the drift, melt and freeze of sea ice. When sea ice melts, buoyant freshwater is mixed into the upper ocean, which inhibits turbulent mixing. Previous extensive drift experiments have usually been conducted in thick multiyear ice, but the ongoing climate change in the Arctic will likely lead to a thinner, highly seasonal ice cover. The longer melt season will therefore change the underlying hydrography and thereby also the coupling of turbulent stresses across the boundary layer. The area north of Svalbard is dominated by first-year ice and is thus an ideal environment to investigate the new sea ice regime. We analyze measurements of currents, oceanic and atmospheric stresses and oceanic buoyancy fluxes from an ice drift station at 82.25° N, 21° E from 26 July to 3 August 2012, when stable stratification in the upper ocean extended all the way to the surface. We find a very small quadratic drag coefficient CD = 9.5 × 10–4 and large turning angles of 32–39° between ocean surface velocity and stress. Buoyancy fluxes at the ice–ocean interface and in the upper ocean are compared with a Rossby similarity scaling that includes the effects of stable stratification. The drag law does not follow Rossby similarity, but the deviations from previously measured drag coefficients and turning angles can instead be explained by the shallow pycnocline forcing the Ekman transport into a thin layer of about 12 m depth.


A study of formation processes of supercooled water and frazil ice in a coastal polynya

Masato Ito, Kay. I. Ohshima, Yasushi Fukamachi, Daisuke Simizu, Katsushi Iwamoto, Yoshimasa Matsumura, Andrew R. Mahoney, Hajo Eicken

Corresponding author: Masato Ito

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

We have conducted mooring measurements off Barrow in the Chukchi Sea since 2009 to examine the process of sea-ice formation in a coastal polynya. Remote sensing imagery, including Synthetic Aperture Radar (SAR), show that coastal polynyas, characterized by streaks of new ice in SAR scenes, occur episodically in this region and further along the coast to the southwest. Each mooring at the nearshore (depth of 45 m) and offshore (depth of 57 m) sites consists of a conductivity-temperature recorder, an acoustic Doppler current profiler and an ice-profiling sonar which can detect frazil ice in the water column. Previous laboratory experiments showed that the underwater frazil ice formation occurs in association with supercooled water resulting from large surface heat loss under the turbulent conditions. When this process is applied to a coastal polynya in the real ocean, ice production rates can be quite high, because of the lack of an insulating ice cover. However, in-situ observations of ice production processes in such polynya settings are greatly limited due to logistic challenges. One of the few places where frazil ice formation in supercooled water has been documented in such a setting is the St. Lawrence Island polynya in Alaska. Our mooring data collected 30–40 m below the surface reveal several potential supercooling events, i.e. with the potential temperature of the subsurface water below the surface freezing point, and two in-situ supercooling events. The potential supercooling persisted for several days under conditions of a shoreparallel northeastward current. According to ice thickness data derived from the Advanced Microwave Scanning Radiometer for EOS (AMSR-E), a series of coastal polynyas had formed along the coast southwest of the mooring sites prior to this event. Thus, water masses exhibiting potential supercooling were likely advected from these polynyas. On the other hand, in-situ supercooling occurred at times when a coastal polynya had formed at the mooring sites with large heat loss to the atmosphere of ~450 W m–2. When the supercooling occurs with strong wind condition, frazil ice was episodically detected by the ice-profiling sonar down to 2–5 m water depth. These findings indicate that frazil ice formation associated with in-situ supercooling occurs through strong direct surface cooling and turbulence.


Distribution and interannual variability of sea-ice thickness in the pack-ice zone off Lützow-Holm Bay, Antarctica

Fuko Sugimoto, Haruhito Shimoda, Daisuke Simizu, Shotaro Uto, Kazutaka Tateyama, Seita Hoshino, Toshihiro Ozeki, Takeshi Tamura, Yasushi Fukamachi, Shuki Ushio, Kay I. Ohshima

Corresponding author: Fuko Sugimoto

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

Under the Japanese Antarctic Research Expedition (JARE), sea-ice thickness has been measured with the electro–magnetic inductive sounder (EM) on board the icebreaker Shirase since the 2000/01 summer season. The results of the measurement in the landfast ice zone have been reported. However, there are very few reports on sea-ice thickness in the pack-ice zone off Lützow-Holm Bay, East Antarctica. Since sea ice is advected westward by the Antarctic Coastal Current, the thickness data off Lützow-Holm Bay are likely represent those in broader area of the Indian Ocean sector. This study examines the distribution and interannual variability of sea-ice thickness in the pack-ice zone off Lützow-Holm Bay based on the EM measurements from 2000/01 to 2011/12 and discusses the cause of the interannual variability. The measurement was carried out between mid-December and early January each summer. For analyses of the interannual variability, we focus on the area of 68°–68.8°S, 38.1°–38.9°E defined as the repeated area, where the measurement has been carried out for 9 seasons except 2004/05. We also utilize visual observational data by a simplified ASPeCt protocol. Both from the EM and visual observations, the mean seaice thickness is about 1.5 m with the mean snow depth of 0.5 m in most summer, while the mean sea-ice thickness reaches 3–6 m in some years, suggesting the significant interannual variability. In the 2011/12 season the ice thickness was particularly large with active ridging. The extremely large ice thickness with active ridging in the 2011/12 season is suggested to be the result of the strong convergent condition. Using the meteorological reanalysis data (ERA-Interim), we calculate convergence (-divergence) of sea-ice drift in the repeated area, assuming that sea-ice drift is directed 30° to the left of the wind and is zero at the southern boundary with landfast ice. The mean sea-ice thickness correlates well with the convergence of ice drift suggesting that severe ridging in 2011/12 is due to the strong convergent condition.


Ice–ocean heat and salt fluxes for Arctic sea ice in advanced stages of melt

Achim Randelhoff, Arild Sundfjord, Angelika H.H. Renner

Corresponding author: Achim Randelhoff

Corresponding author e-mail: achim@npolar.no

The melt of sea ice is controlled both by atmospheric and oceanic heat. Open water and thin, ponded ice allows heating of the mixed layer, contributing to accumulation of fresh meltwater and further trapping of solar heat under the stratification. The ongoing climate change in the Arctic will likely lead to a thinner, highly seasonal ice cover, which entails a longer melt season and more solar heat accumulating in the upper ocean. This makes the ice–ocean heat and salt fluxes an important player for both melt process and summer upper ocean stratification. The area north of Svalbard is dominated by first-year ice and is thus an ideal environment to investigate the new sea ice regime. We analyze measurements of boundary layer hydrography, ice–ocean stress and heat fluxes from an ice drift station at 82.25° N, 21° E from 26 July to 3 August 2012, when the ice was in advanced stages of melt with large fractions of open water and meltwater ponds. By comparing under-ice turbulence measurements with wind stress data, local heat flux measurements can be rescaled to yield values representative of the floe scale. A new parameterization of the heat flux FH = ρw cw (CH u* ΔT + δq) with friction velocity u*, local temperature elevation above freezing ΔT, effective heat transfer coefficient CH, heat flux offset δq and density times heat capacity of seawater ρw cw, is derived analytically from a set of equations expressing the ocean–ice exchange of heat and salt as proportional to friction velocity and gradient between mixed layer and interface. We find a transfer coefficient CH = 0.0056 and an average heat flux reduction ρw cw δq = -7.1 W m–2 which is likely attributable to additional salt or heat fluxes, e.g. from melt water in leads. The bulk Stanton number parameterization St = FH/( ρw cw u*ΔT) = 0.005–0.006 is recovered as a special case.


Surface flooding of summer Antarctic sea ice and related growth of sea ice algae

Stephen Ackley, Zach Brown, Donald Perovich, Kevin Arrigo, Blake Weissling

Corresponding author: Stephen Ackley

Corresponding author e-mail: stevejacackley@gmail.com

The surface flooding of Antarctic sea ice in summer has been found extensively in the two major summer ice packs, in the western Weddell Sea and the Bellingshausen-Amundsen Seas. Sea ice physical–biological investigations were made during and after the Oden 2010/11 summer cruise in the pack ice of the Amundsen Sea. Ice core, surface slush, and sea water sampling were made at fourteen stations on the pack ice during Jan 2011. At two of the sampling stations, CRREL ice mass balance buoys were installed, providing time series of related physical–biological parameters throughout the summer period after the sampling took place. The ice, slush and under-ice station sampling showed higher levels of chl a in the ice and slush than in the under-ice water. The highest vertically-integrated chl a levels (20–72 mg m–2) were found in the cores, in their upper portions, with deeper snow covers. Temperature, ice thickness and snow depth time series records were obtained at both the ice buoys. At one buoy, an uplooking seven channel under-ice spectroradiometer recorded changes in biologically sensitive light levels in the ice during the summer period. Buoy temperature records from thermistors embedded in the snow and ice vertically showed increases in the depth of the flooded layer by 30–40 cm during Jan–Feb. While the snow depth was relatively unchanged, ice thickness decreased by up to a meter from bottom melting during this period. The radiometric proxy for vertically integrated sea-ice biology from the under-ice spectroradiometer showed relatively constant values for the summer period, indicating sea-ice production continued despite the loss of nearly half the ice thickness at the same site. The mechanism for increasing the depth of flooding was therefore caused by the decrease in ice thickness, which progressively raised the sea level into the snow cover to maintain isostatic balance in the floe. High biological production could be maintained in these flooded layers throughout summer since nutrients were recharged as the flooded layer increased in thickness as ice bottom melting occurred. Estimates of the nutrient input during flooding were determined. Continued seawater flooding at the snow–ice interface impacts continued growth of sea ice biological communities in Antarctic summer sea ice.


Surface flooding of Antarctic summer sea ice

Stephen Ackley, Donald Perovich, Ted Maksym, Blake Weissling

Corresponding author: Stephen Ackley

Corresponding author e-mail: stevejacackley@gmail.com

The surface flooding of Antarctic sea ice in summer covers 50% or more of the sea ice area in the two major summer ice packs, the western Weddell and the Bellingshausen-Amundsen Seas. Sea water flooding at the snow–ice interface affects the remote sensing characteristics of Antarctic sea ice, and the composition of the ice cover as snow ice forms during freezing of the flooded layer. Mechanisms previously discussed for surface flooding have concentrated on sudden winter increases in snow depth causing the ice surface to be pushed below sea level and sea water to intrude. We investigated instead the development of surface flooding in summer due to ice bottom melting. Four ice mass balance buoys, two CRREL buoys and two SAMS IMBs, were deployed on the Amundsen Sea pack ice at distances of ~100 m from floe edges in late December 2010 from the icebreaker Oden, lasting from six weeks up until four months, bridging the summer period (Jan–Feb 2011).Temperature records from thermistors embedded in the snow and ice vertically showed progressive increases in the depth of the flooded layer on the ice cover during January and February by 30–40 cm at some of the sites. Snow-depth and ice-thickness changes were also recorded during this time. While the snow depth was relatively unchanged, ice thickness decreased by up to a meter from bottom melting during this period. The mechanism for increasing the depth of flooding was therefore caused by the decrease in ice thickness which progressively raised the sea level into the snow cover to maintain isostatic balance in the floe. The process to increase the depth of the flooded layer was vertical in nature, and did not require additional snowfall. Contemporaneous with the high bottom melting, under-ice water temperatures of up to 1°C above the freezing point were found. Two possible sources were investigated for the high ocean heat flux. Upwelling of warm deep water into the mixed layer can bring heat from below, while solar heating of the upper mixed layer can occur when ice concentration in the local area falls and lower albedo ocean water is exposed to radiative heating. Both relatively high snowfall in winter and the summer dynamics, where bottom melting results in less snow freeboard above sea level, can therefore increase the depth of flooded snow. The high proportion of snow ice found for Amundsen Sea pack ice in previous survey cruises therefore results from both winter snowfall and summer ice bottom melt.


Ice shelf basal melt control by small-scale coastal polynya

Benjamin Keith Galton-Fenzi, Simon James Marsland, Alexander Donald Fraser, Noriaki Kimura, Mike Craven, Ian Allison

Corresponding author: Benjamin Keith Galton-Fenzi

Corresponding author e-mail: Ben.Galton-Fenzi@aad.gov.au

Ice shelves are an important component of the interaction between the Antarctic ice sheet and the global ocean. How ice shelves evolve in a changing climate is crucial to informing our understanding of future global sea level rise. Many ice shelves are observed to be thinning, likely due to changing oceanic conditions. However, ocean drivers of ice shelf basal melting are only poorly understood, particularly in East Antarctica. We present a new mechanism applicable to ice shelves in the vicinity of coastal latent heat polynyas (intense sea ice production regions). Using observations of the Amery Ice Shelf (AIS)/Prydz Bay (PB) ocean system in East Antarctica, we show that the timing of the arrival of warm water from the continental shelf to the ocean beneath the AIS is controlled by intense sea ice production in a local polynya. Dense shelf water formation due to sea-ice production in the polynya causes the shutdown of the PB gyre, resulting in flooding of the sub-ice-shelf cavity with cooler waters. Sensitivity studies from an ocean model show, contrary to previous studies, suppression of dense water production leads to warmer waters entering into the sub-ice-shelf cavity, resulting in increased basal melt. This implies that continued suppression of dense shelf water formation, induced through atmospheric warming, will further increase ice shelf basal melting, exposure of the Antarctic ice sheet, and ultimately sea level.


Importance of sea-ice derived iron for primary production in the Ross Sea polynya: results from the PRISM project

Peter Sedwick, Dennis McGillicuddy, Michael Dinniman, Thomas Bibby, Blair Greenan, Eileen Hofmann, John Klinck, Stefanie Mack, Christopher Marsay, Walker Smith, Bettina Sohst

Corresponding author: Peter Sedwick

Corresponding author e-mail: psedwick@odu.edu

The Ross Sea polynya is among the most productive regions in the Southern Ocean, sustaining an average annual net primary production of ~20 Tg C. Field observations indicate that phytoplankton growth in this region during summer is limited by the availability of dissolved iron, which can be drawn down to growth-limiting concentrations (~0.1 nM) in surface waters by as early as mid-November. Despite this, the polynya supports substantial production during the December–February period, implying that there are sustained inputs of dissolved iron to surface waters over the summer months. The PRISM project aims to constrain these inputs by combining field observations and numerical modeling, focusing on the 201/12 austral summer growing season. Our results indicate that the major sources of dissolved iron to primary producers are vertical resupply via winter convective mixing (~41%), and sea ice meltwater (~41%), with lesser contributions from circumpolar deep water (~15%), and glacial meltwater (~3%). On a regional basis, sea ice is most important as an iron source in the western Ross Sea, where we estimate that it accounts for around 50% of the seasonal input, assuming a dissolved iron concentration of 10 nM in the sea ice end member. However, the limited data that are available indicate order-of-magnitude scale variability in the dissolved iron concentration of sea ice meltwaters, highlighting the critical need for additional measurements of iron in Antarctic sea ice.


Circulation-driven response of Antarctic sea ice to external anthropogenic forcings: why do models and observations disagree?

Alexander Haumann, Dirk Notz, Hauke Schmidt

Corresponding author: Alexander Haumann

Corresponding author e-mail: alexander.haumann@gmail.com

It is still unclear how Antarctic sea ice responds to the combined effect of anthropogenic perturbations of the climate system, such as the depletion and recovery of stratospheric ozone and the increase of greenhouse gases. Satellite observations over the past three decades suggest a zonally asymmetric response of Antarctic sea ice, with a significantly increasing ice cover in the Ross Sea and a slightly weaker decreasing ice cover in the region of the West Antarctic Peninsula. This asymmetry is due to an asymmetric lowering of the surface pressure, which is strongest for the Amundsen-Bellinghausen Seas Low, causing stronger southerly winds and an enhanced advection of cold, continental air over the Ross Sea and thus the expansion of sea ice there. As a net effect, the sea-ice cover in the Antarctic shows over the past few decades a small but significant overall increase. In contrast, global climate models consistently show a rather uniform decrease of the ice cover in large parts of the Southern Ocean. We here compare the observed changes with a suite of model experiments in order to explain these discrepancies and to assess the effects of external anthropogenic forcings. The model simulations do reproduce the significant lowering of the surface pressure in coastal regions that is clearly induced by both ozone depletion and greenhouse-gas increase. However, the decreasing surface pressure is much more zonally symmetric than the observed one. As a consequence, the asymmetric response of the sea-ice cover is visible but not as pronounced as in the observations. Additionally, the simulated zonal wind increase leads to an enhanced overturning circulation. This advects warmer water masses from below and reduces the ice cover. We conclude that anthropogenically induced meridional circulation changes act to produce an asymmetric sea-ice response and an overall increase of the Antarctic sea-ice cover, but a zonal circulation intensification might lead to an overall decline. Whereas the observed change is asymmetric, models tend to have a too zonal response which acts to rather decrease the ice cover.


The role of the sea-ice carbon pump for the marine carbon cycle

Rosina Grimm, Dirk Notz, Søren Rysgaard, Ronnie Glud

Corresponding author: Dirk Notz

Corresponding author e-mail: dirk.notz@zmaw.de

The role of sea-ice growth and decay for the global carbon cycle is still largely an open question. Several studies propose a net oceanic CO2 uptake in response to the release of excess bulk sea ice alkalinity (TA) to the surface ocean during ice melt, the so called sea-ice carbon pump. During the current global climate change, both the enhanced seasonal cycle of sea-ice volume due to intensified summer sea-ice melt as well as the increased atmospheric CO2 concentration modify the efficiency of the sea-ice carbon pump. We here aim to quantify the role of sea ice for the marine carbon cycle using a global ocean–sea-ice–biogeochemical model (MPIOM/HAMOCC). In the model, we assign to the sea ice constant ratios of TA:DIC covering the range of observations. To separate the climatic influence on the sea-ice carbon pump, we carry out simulations for three idealized climate scenarios: First, we simulate a steady state control climate by forcing the model with the present-day climate (OMIP atmospheric fields) and atmospheric pCO2 of 337 ppm. In the second scenario, we simulate a reduction of the summer sea-ice extent but keep the pCO2 at 337 ppm. In the third scenario, we increase the atmospheric pCO2 by 10 ppm per year but the radiation scheme is not influenced by this rise so that the climate remains at present-day level. The results show that for each climate scenario the higher the ratio and the concentration of TA:DIC prescribed in sea ice, the more efficient is the additional oceanic uptake of CO2 by the sea-ice carbon pump. Locally, the air–sea CO2 flux patterns are modified dependent on both the prescribed TA:DIC concentrations and the climate scenario. In both the steady state and CO2 climate scenarios, the oceanic CO2 uptake by the sea-ice carbon pump is enhanced in sea-ice import areas (e.g. Gin Sea, import of TA), and reduced in sea-ice production areas (e.g. Arctic Ocean, export of TA). In the climate scenario simulating the enhanced seasonal cycle of sea ice, this trend is reversed due to the reduced sea-ice (TA) export during summer. After 100 years of simulation, the global oceanic uptake of CO2 due to the sea-ice carbon pump ranges between 3 and 30 Tg a–1, which corresponds to a 0–10% change of the global uptake, and is highest in the steady state and lowest in the CO2 scenario. On the long term (Millennial scale), the enhanced oceanic CO2 uptake by the sea-ice carbon pump will balance with an enhanced CO2 outgasing.


Antarctic and Arctic landfast sea ice: deriving parent water mass properties changes from sea ice cores

Inga J. Smith, Alexander J. Gough, Patricia J. Langhorne, Hajo Eicken, Andrew R. Mahoney, Gregory H. Leonard, Robert J. Van Hale, Yasushi Fukamachi, Stefan Jendersie, Joshua M. Jones, Timothy G. Haskell

Corresponding author: Inga J. Smith

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

In Antarctica, where approximately half the coastline is fringed by ice shelves, landfast sea-ice formation can be influenced by supercooled sea water originating from ice shelf basal melt. Although the ice shelf influenced water is only marginally less saline and the supercooling slight (of the order of 0.01 K), the effects on sea ice structure, through platelet ice formation, are dramatic. In the Arctic, where large ice shelves are absent, landfast sea ice growth can be influenced by fresher water from rivers and residual summer melt. In this presentation, a method to reconstruct changes in water masses using oxygen isotope measurements from sea-ice cores is examined. Direct measurements of sea-ice growth rates are used to validate the output of a sea ice thermodynamic model driven with Modern-Era Retrospective Analysis for Research and Applications (MERRA) reanalysis data, along with observations of snow depth and freeze up dates. The output of that model is used along with sea ice oxygen isotope measurements and fractionation equations to determine changes in sea water isotope composition over the course of the ice growth period. It is shown that for Barrow, Alaska in 2011/12, large changes in the isotope composition of the ocean waters were captured by the sea ice, indicating episodic advection of less saline water during the growth season. For McMurdo Sound, Antarctica in 2009, the method is at the limits of resolution for determining the presence of ice shelf influenced surface waters, and improvements to the methods are suggested for that region.


The outcome of different sea-ice melting methods to brackish sea-ice biology measurements

Janne-Markus Rintala, Markus Majaneva, Jonna Piiparinen, Susann Müller, Jari Uusikivi, Jaanika Blomster, Riitta Autio

Corresponding author: Janne-Markus Rintala

Corresponding author e-mail: janne.rintala@helsinki.fi

Brackish water of the Baltic Sea forms structurally similar sea ice to Polar areas with brine channels and pockets that are colonized by algae and protists. These sympagic communities cannot be attained in biological studies without destroying their natural habitat by melting the ice prior to any analysis. Practically there are three different melting procedures: brine drainage, direct melting or buffered melting. Brine drainage is not sufficient enough because it is estimated to represents only 5% of the whole sea ice volume. Direct melting is probably most widely used, commonly accepted and also the recommended method, for example in salinity measures. In practice the ice is left to melt in just above freezing temperatures until it is thawed. This method is seldom acceptable for sea ice samples intended for biological determinations because it has been associated with up to 70% cell losses. The lysis of the cells is explained to be caused by osmotic shock, the change in salinity, which is avoided if samples are melted into osmotic buffer. Addition of filtered sea water, obtained from the same site with the sea ice samples is commonly used buffer in the handling of the sea ice samples. Melting the samples in the sea water of the same salinity is easily accepted as it resembles the actual melting event that would occur in nature. Yet, if the sea ice is not in the marine environment, such as a brackish water estuary, the water underneath the ice can also be fresh water originating from the rivers, which would improve the lysis of the cells. Therefore melting the ice samples in increased salinities similar to those found inside the brine salinities would be better. Yet there are no available results if it is better to melt the samples in raised salinity solution that is prepared from nutrient poor MQ water or for example the sea water obtained underneath the ice. This may become significant if the sea ice in question is floating on top of eutrofied water mass, such as the Baltic Sea. Melting the sea-ice samples in the eutrophied Baltic Sea water would resemble an incubation in nutrient enrichment, which is likely to enhance the growth of sympagic organisms. Therefore an orthogonal hierarchical factorial experiment with 5 replicates was designed to investigate the effects of different melting methods to the brackish water sea-ice biogeochemistry: nutrients, pigments, cell abundances and their photosynthetic activities both in situ and laboratory.


The role of wind-blown snow in Antarctic sea-ice formation

Rob Massom, Katie Leonard, Ted Maksym, Takenobu Toyota, Lars Smedrud, Petra Heil, Steve Warren

Corresponding author: Rob Massom

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

Past studies have shown how wind-blown redistribution of snow is an important process governing not only snow thickness distribution on Antarctic sea ice but also the input of snow into leads. In this work, which follows on from previous work in the Arctic, we use observations from various cruises to investigate the possibility that, under freezing and windy conditions, snow is not simply ‘lost’ to leads (to melt and represent freshwater input) but actively contributes to the formation of sea ice, with individual blowing-snow particles at a low temperature (equivalent to that of the air) bombarding the surface waters to act as foreign nuclei for frazil crystal formation and the collective generation of a mixed frazil-snow slurry – which we term ‘snow frazil’. Given the widespread occurrence of leads and strong winds prevalent over the Antarctic sea-ice zone, we propose that ‘snow seeding’ may be an important process in both sea ice formation and the infilling of narrow leads or back-filling of wider leads. As the snow is significantly depleted in 18O, we suggest that this process may contribute to the high proportion of snow-ice observed in Antarctic sea ice cores, possibly explaining the significant fraction of snow-ice identified based on only a slightly negative δ18O signature. It also has potentially-important oceanic and biological implications.


Sensitivity of sea ice, water masses and ice shelf basal melt in the Ross Sea to changes in the winds and atmospheric temperatures

Michael Dinniman, John Klinck, Walker Smith, Jr

Corresponding author: Michael Dinniman

Corresponding author e-mail: msd@ccpo.odu.edu

Strengthening of the cold southerly winds over the Ross Sea is thought to be one of the causes for the observed increases in sea-ice extent in this area and may have significant effects on other aspects of the ocean circulation. A regional ocean/sea-ice/ice shelf model of the Ross Sea is used to examine the effects of changes in the winds on sea ice, water masses and basal melt of the Ross Ice Shelf (RIS). Simple increases in the wind speed with no other atmospheric changes actually reduced the summer sea-ice area and increased the duration of the summer polynya, the opposite of what has been observed over the past three decades. Increases in the winds combined with spatially uniform decreases in the air temperatures led to realistic increases in sea-ice concentrations. The increased winds worked against the cooler air temperatures in changing the basal melt rate of the RIS and the slight change (a 4% increase) in the basal melt makes it difficult to distinguish the dominant forcing factor. AR4 future scenario simulations typically show atmospheric warming and changes in wind speed (increases and decreases) and direction over the Ross Sea. Results from simulations forced with winds and air temperatures from the SRES A1B scenario simulations from the MPI ECHAM5 model show a decrease in the summer sea-ice area over the continental shelf of 56%/78% for 2046–50/2096–2100 compared to 1996–2000 with a mean 12/33 day increase in the ice free summer period (winter extents were unchanged). The basal melt rate of the RIS increased slightly by 2046–50 (6% increase) and 2096–2100 (9% increase). There has also been an observed freshening of the Ross Sea over the last 50 years and it has been proposed that this is a signature of increased meltwater advected from the Amundsen Sea. A simplistic freshening of the water advected into the model domain for the 2046–50 and 2096–2100 simulations did not have a significant effect on the modeled sea-ice extent. However, the freshening reduces the vertical mixing over the shelf, increases the basal melt rate of the RIS for 2046–50 (10% increase compared to the 2046–50 simulation with no freshening, 17% increase compared to the end of the 20th century) and 2096–2100 (12%/22% increase) and leads to a major reduction in the export of cold/salty shelf water.


NPZD-iron lower level ecosystem model

Elodie Salmon, Michael Dinniman, Eileen Hofmann

Corresponding author: Michael Dinniman

Corresponding author e-mail: msd@ccpo.odu.edu

The Ross Sea continental shelf is the single most productive area in the Southern Ocean, and may comprise a significant but unaccounted for oceanic CO2 sink, largely driven by phytoplankton production. However, the processes that control the magnitude of primary production in this region are not well understood. During summer, an observed abundance of macronutrients and scarcity of dissolved iron are consistent with iron limitation of phytoplankton growth in the Ross Sea polynya, as is further suggested by shipboard bioassay experiments. Field observations and model simulations indicate four potential sources of dissolved iron to surface waters of the Ross Sea: (1) circumpolar deep water intruding from the shelf edge; (2) sediments on shallow banks and nearshore areas; (3) melting sea ice around the perimeter of the polynya; and (4) glacial meltwater from the Ross Ice Shelf. These potential iron sources are isolated, either laterally or vertically, from the surface waters of the Ross Sea for much of the growing season. We hypothesize that hydrodynamic transport via mesoscale currents, fronts, and eddies facilitate the supply of dissolved iron from these four sources to the surface waters of the Ross Sea polynya. We propose to test this hypothesis through a combination of in situ observations and numerical modeling, complemented by satellite remote sensing. We have embedded a simple nutrient model into an idealized Antarctic coastal circulation model. This formulation includes explicit dynamics for iron, and three groups of phytoplankton (Phaeocystis antarctica solitary cells and colonies and diatoms), two groups of zooplankton and detritus. The parameterizations used for primary production are revised to reflect physiological responses of Ross Sea phytoplankton communities. Preliminary simulation results suggest that phytoplankton growth is mainly controlled by light related to sea ice coverage and dissolved iron availability. A smaller value of nitrate uptake rate is needed to reproduce fast growth of phytoplankton during the Antarctic spring season than in Artic region. We also found that the two major sources of iron come from the circumpolar deep water and mixing of sediments on shallow banks.


Evaluation of multiple proxies for sea-ice variability in an ice core from Derwael Ice Rise, Dronning Maud Land, Antarctica

Morgane Philippe, Jean-Louis Tison, Frank Dehairs, Bryn Hubbard, Frank Pattyn

Corresponding author: Morgane Philippe

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

We drilled an ice core at the top of the coastal Derwael Ice Rise, Dronning Maud Land, Antarctica. This 120 m long ice core should cover about 200–400 years of past accumulation, according to our preliminary results of annual layers identification from optical televiewer-based borehole luminosity. Major ions and water stable isotopes measurements are conducted at high resolution to reveal seasonal variations and improve annual layers identification. At the same time, the potential of sea-ice extent proxies such as sea salt sodium (ssNa), non-sea-salt sulfate (nssSO4) and Methyl Sulfonic Acid (MSA), is investigated. Several studies have shown various positive and negative relationships between these ions and sea ice extent, but very few concerned the region of Dronning Maud Land. Additionnally, NH4, as an indicator of marine biogenic activity, and deuterium excess, as a proxy for moisture source region fluctuations (and potentially sea ice extent) will also be discussed.


Prognostic salinity in sea-ice models and sea-ice biogeochemistry

Adrian Turner, Elizabeth Hunke

Corresponding author: Adrian Turner

Corresponding author e-mail: akt@lanl.gov

In order to properly model sea-ice biogeochemistry in climate simulations the processes that transport brine (and so nutrients) around sea ice need to be modelled. We present a new thermodynamic component of the Los Alamos sea-ice model, CICE, that has both prognostic temperature and bulk salinity. Processes key to representing the transport of brine within the sea ice, such as gravity drainage, melt-water flushing and snow-ice formation, are parameterized by the model. The thermodynamic component is numerically efficient and suitable for global climate simulations. We compare both idealized laboratory and fieldwork experiments and global simulations to observations to validate the salinity model.


On brine rejection induced ocean mixing in CESM climate model

Meibing Jin

Corresponding author: Meibing Jin

Corresponding author e-mail: mjin@alaska.edu

Brine excluded from sea-ice formation has a much higher salinity than ambient ocean and thus falls into ocean in the forms of long filaments without appreciable diffusion. The ocean mixing caused by brine is a major driver for ocean mixing and the seasonal variations of halocline in polar oceans. The horizontal scale of the mixing is much smaller than climate model grid and model tends to have a bias of much deeper ocean mixing than observed. Here, a two-column ocean grid (TCOG) scheme is implemented in a global coupled sea ice–ocean model to explicitly solve the different vertical mixing in the two water columns with and without brine rejection. Compared with the control simulation using a regular ocean grid, the TCOG simulations showed consistent improvements in mixed layer depth (MLD) vertical profiles of salinity (S) and relatively smaller improvements in temperature (T). Improvements in vertical S profile were consistent in the upper 150 m, including the upper mixed layer and the cold halocline below it. The fraction of grid with brine rejection was tested to be the fraction of open water in a grid or a small constant fraction. A small constant fraction was found to produce best model comparison with observations in the Arctic, and the concept was qualitatively supported by observations of brine formation in pack ice and brine spread in ocean.


The role of ice age distribution in the trends and variability of summer sea ice extent in the Beaufort and Chukchi seas

Walter Meier, Mark Tschudi

Corresponding author: Walter Meier

Corresponding author e-mail: walt.meier@nasa.gov

The extent of Arctic summer sea ice is declining rapidly. However, there is strong regional and interannual variability in the location of the end of summer ice cover. An important factor in this variability is the change in distribution of ice age types within the Arctic basin. In particular, multi-year ice was one predominant in the Beaufort and Chukchi seas and the region was factory for older ice as the sea ice circulated within the Beaufort Gyre though multiple melt seasons. However, less old ice is making the transit around the gyre. One contributing factor appears to be greater advection of ice across the Transpolar Drift Stream, but there is also significant summer melt of multiyear ice within the Beaufort and Chukchi seas causing an in situ loss of the old ice types. Here we investigate the trends of extent and age in this important region using satellite-derived extent and age estimates and Lagrangian tracking of ice parcels. We also discuss the causes of the loss and changing distribution, and the potential implications of the observed changes. In addition to examining the long-term trends we present a case-study contrasting the 2012 and 2013 minimum extent and the role played by the ice age distribution versus other factors.


CO2 and CH4 profiles in Arctic fjord sea ice

Odile Crabeck, Bruno Delille, David Thomas, Nicolas-Xavier Geilfus, Soren Rysgaard, Jean-Louis Tison

Corresponding author: Odile Crabeck

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

We present bulk ice [CH4] and bulk ice pCO2 from sub-Arctic, landfast sea ice in the Kapisillit fjord, located in NE of Nuuk (Godthabsfjord, SW Greenland). The bulk ice CH4 concentration ranged from 1.8 to 12.1 nmol L–1. This corresponds to a partial pressure range of 3–28 ppmv in sea ice, and is markedly higher than the average atmospheric methane content of 1.85 ppmv. Most of the trapped methane within the sea ice seems be contained inside bubbles and only a minor part are dissolved in the brine network. While the CH4 from the seawater is accumulated within the sea-ice cover, the sea cover provides an interface in which the methane could be stored and transformed over time by biogeochemical processes. Further studies based on longer times series and carbon isotope signatures will provide us the opportunity to study the potential methane oxidation rate within the sea ice cover. In the case of the CH4 being oxidised over time within the sea ice cover, the sea ice could provide an interface where CH4 is degraded and, hence, act as sink for oceanic CH4. The bulk ice pCO2 ranged between 60 and 330 ppmv showing that ice with a temperature above -4°C is under-saturated relative to the pCO2 in the atmosphere (390 ppmv). Finally, the sea ice investigated was influenced by river input.


The mass balance of Arctic sea ice

Don Perovich, Jacqueline Richter-Menge, Chris Polashenski, Bruce Elder, Kathy Jones, Todd Arbetter

Corresponding author: Don Perovich

Corresponding author e-mail: jonesperovich@myfairpoint.net

In recent years, the Arctic sea ice cover has undergone significant changes, with a reduction in summer ice extent, a thinning of the ice, and a shift from multiyear to first year ice. The sea-ice mass balance provides insights into understanding these changes by attributing them to variations in ice growth, melt season length, surface melt, and bottom ablation. With this data, the relative impact of atmospheric and ocean forcing can be determined. As part of an Arctic Observing Network, we have routinely deployed ice mass-balance buoys (IMB) at multiple locations to monitor autonomously the sea ice cover since 2000. Mass-balance results show considerable spatial variability across the Arctic in a given year, with the largest amount of melting in the Beaufort Sea and the smallest in observed in ice north of the Canadian Archipelago. The thinning of ice in the Beaufort Sea is primarily due to increases in bottom melting. At the end of summer, peak ocean heat fluxes to the ice in this region have exceeded 100 Watts per square meter. Bottom melting is closely related to the solar heat inputted to the ocean through leads. Results from IMBs deployed at the North Pole Environmental Observatory (NPEO) from 2000 to 2012 exhibit no definitive trends, but there are a number of interesting findings. There is large interannual variability, with surface melting ranging from 0.02 m to 0.50 m and bottom melting from 0.10 m to 0.57 m. The years with the largest amount of surface melting at NPEO were also years of record minimum ice extent (2005, 2007, and 2012). The largest amounts of bottom melting have occurred in the past few years. However, for all years the ice was at least 1.2 m thick at the end of melt season.


Effect of sea-ice distribution on spatial and temporal variation of spring bloom in the Sea of Okhotsk

Hajime Yamaguchi, Noriaki Kimura, Yuya Nakano, Makoto Kayano

Corresponding author: Makoto Kayano

Corresponding author e-mail: marcador.kp@1.k.u-tokyo.ac.jp

The Sea of Okhotsk belongs to the belt of subarctic basins in the Northern North Pacific, which are among the most important fishery grounds in the world. It is the southernmost sea with a significant seasonal sea ice area and is characterized by the rich marine ecosystem. The spring bloom that is a strong increase in phytoplankton occurs and contributes to high primary production in the Sea of Okhotsk. It is said that sea ice plays an important role in maintaining the richness. Recent sea ice extent in the Sea of Okhotsk has declined associated with global warming. We need to understand the mechanism of this rich ecosystem and to examine the response of marine ecosystem to the ice decline. The object of this study is to clarify relationship between sea ice and phytoplankton. To understand spatial and temporal distribution in phytoplankton, we analyze satellite derived chlorophyll-a (chla) concentration. This study uses the data from Moderate Resolution Imaging Spectroradiometer (MODIS) boarded on Terra since its error is smaller compared with Aqua/MODIS or Sea-viewing Wide Field-of-view Sensor (SeaWiFS); it is well correlated with the observed data near Hokkaido by Hokkaido National Fisheries Research Institute (HNFRI) in the Sea of Okhotsk (R = 0.91). Spring bloom is noticeable near the coast, especially the Hokkaido coast, the Terpenia Bay and off the western coast of the Kamchatka Peninsula First, we examine the regional difference of each spring bloom, by considering the interannual change in the volume of ice melting, timing of ice retreat and weather conditions. We investigate where the sea ice remains before spring bloom occurs and how much the sea-ice melts. Finally, we discuss the relationship between the rich ecosystem and sea ice in the Sea of Okhotsk.


Reduced sea ice in the northwestern Weddell coincides with onset of Larsen Ice Shelf structural weakening

Ted Scambos

Corresponding author: Ted Scambos

Corresponding author e-mail: teds@nsidc.org

Ice shelf/ice tongue disintegrations and break-ups have a major effect on glacier mass balance, and nowhere has this been more evident than in the northern sections of the Larsen Ice Shelf in the Antarctic Peninsula. Ice flux in this region surged 2- to 6-fold after the 1995 and 2002 ice shelf disintegration events, driven by a group of processes based on the presence of extensive surface melt lakes. However, precursor changes in the ice shelves beginning more than a decade before the events have been identified in satellite imagery, beginning around 1990. Prior to this time, available satellite data suggests steady-state evolution of the ice shelf areas. The onset of non-steady structural changes appears to coincide with reduced sea ice extent in the northwestern Weddell Sea. Sea ice in the Weddell trended slightly lower in the 1990s than for earlier or later periods. An examination of satellite imagery of the Larsen A and B ice shelves spanning 1963 to the present shows that shear margins on the Larsen B and Scar Inlet were essentially unchanged through 1986. Following that time, ice shelf shear zones show significant evolution, including increased and expanded areas of rifting, concentration of shear, and ice flow speed increases. These early changes, occurring prior to shelf area loss, suggest either increased ocean-driven basal melt or effects of increased meltwater are the cause of early shelf weakening that led to disintegration. The reduced sea-ice extent in the 1990s in the region suggests that wind traction on the ocean surface may have influenced sub-shelf ocean circulation.


East Antarctic sea ice: time of change?

Petra Heil, Adam Steer, Robert Massom

Corresponding author: Petra Heil

Corresponding author e-mail: petra.heil@utas.edu.au

Ship-based observations from seven recent cruises off East Antarctica (90–110°E) are used to describe regional change in the thickness distribution and characteristics of sea ice and snow cover thickness during austral spring. These observations were collected between 2003 and 2012 and form part of the Scientific Committee on Antarctic Research Antarctic Sea Ice Processes and Climate (ASPeCt) dataset and are compared to previous observations (1992–2007) in the region. The total ice thickness is derived from observations of level ice thickness and surface topography using a simple ridge model. The long-term mean (standard deviation) total ice thickness, which includes deformed ice for the region during austral spring was 0.72 (±0.80) m, and that of level ice only was 0.53 (±0.43) m. For the region, the long-term mean ice concentration was 71% with about 12% of the ice area covered by surface ridges. The regional long-term mean snow thickness for level ice was 0.15 m. By comparison for the same region during the most recent decade the mean total ice thickness was 0.87 (±0.92) m, with a mean of 0.45 (± 0.38) m for level ice only. At the same time the snow cover of level ice increased to 0.20 m. To investigate these changes we review the recent data in context of regionally increased ice concentration, pol-ward retreat in equatorial ice extent, and changes in the atmospheric surface circulation. Our preliminary finding suggests that a southward shift of autumnal and early winter atmospheric storm trajectories and a change in their frequency are associated with increased sea-ice deformation, which consequently increases the aerial coverage of ridged ice as well as the total ice thickness. Increased snow thickness over level ice might be associated with the changes in the atmospheric circulation.


Arctic sea ice prediction dependence on sea-ice thickness anomalies

Cecilia Bitz, Edward Blanchard-Wrigglesworth

Corresponding author: Cecilia Bitz

Corresponding author e-mail: bitz@uw.edu

Skillful Arctic sea-ice forecasts may be possible for lead times of months or even years owing to the persistence of thickness anomalies. We characterize sea-ice thickness in terms of its variability in fully-coupled global climate models (GCMs) and sea ice–ocean only models (IOMs) that are forced with an estimate of observations derived from atmospheric reanalysis and satellite observations. Overall, the variance in sea-ice thickness is greatest along Arctic Ocean coastlines. In the GCMs, sea-ice thickness anomalies have a typical timescale of up to 20 months, which can extend to 30 months when accounting for their transport, and a typical lengthscale of about 500–1000 km. The range of these scales across GCMs implies that an estimate of the number of thickness monitoring locations to characterize the full Arctic basin sea-ice thickness variability field is model dependent, and would vary between 3 and 10. Models with a thinner mean ice state tend to have ice-thickness anomalies that are generally shorter lived and smaller in amplitude, but have a larger spatial scale. Additionally, sea-ice thickness variability in IOMs is damped relative to GCMs due to strong negative coupling between the dynamic and thermodynamic processes of sea ice. The implications of these findings on the design of prediction systems will be discussed.


Year round survey of ocean–sea ice–air exchanges – the YROSIAE survey

Bruno Delille, Tim Haskell, Willy Champenois, Bernard Heinesch, Jiayun Zhou, Véronique Schoemann, Gauthier Carnat, François Fripiat, Thomas Goossens, Sébastien Moreau, Martin Vancoppenolle, Frédéric Vivier, Antonio Lourenço, Jean-Louis Tison

Corresponding author: Bruno Delille

Corresponding author e-mail: Bruno.Delille@ulg.ac.be

YROSIAE survey aimed to carry out a year-round survey of landfast sea ice focusing on the study of sea ice physics and biogeochemistry in order to (a) better understand and budget exchanges of energy and matter across the ocean–sea ice–atmosphere interfaces during sea ice growth and decay and (b) quantify their potential impact on fluxes of climate gases (CO2, DMS, CH4, N2O) to the atmosphere and on carbon and macro-nutrients and micro-nutrients export to the ocean. Ice cores, sea water, brines and exported material were collected at regular intervals about 1 km off Cape Evans from November 2011 to December 2011 and from September 2012 to December 2012 in trace-metal clean conditions. Samples are processed to characterize both the vertical distribution and temporal changes of climate gases (CO2, DMS, CH4, N2O), CO2-related parameters (dissolved inorganic carbon, total alkalinity and CaCO3 amount), physical parameters (salinity, temperature, texture, 18O), biogeochemical parameters (macro-nutrients, particulate and dissolved organic carbon, δ13C, δ30Si and δ15N, micro-nutrients – including iron) and biological parameters ( chlorophyll a, primary production within sea ice derived from O2:Ar and O2:N ratios, autotrophic species determination, bacterial cell counts a.s.o.). In addition, we deployed a micro-meterological tower and automatic chambers to measure air–ice CO2 fluxes. Continuous measurements of ice temperature and ice accretion or melting, both at the ice–ocean and the ice–atmosphere interfaces were provided by an ‘Ice-T’ ice mass balance buoy. Sediment traps collected particles below the ice between 10 and 70 m, while dust collectors provided a record of a full suite of trace metal and dust at different levels above the ground. We will present the aims, overall approach and sampling strategy of the YROSIAE survey. In addition we will also discuss CO2 dynamics within the ice and present temporal air–ice CO2 fluxes over the year. We will provide a first budget of air–ice CO2 fluxes during ice growth for Antarctica sea ice and discuss the impact of the snow cover on air-ice CO2 fluxes.


An Australian-led integrated process study of the Antarctic Marginal Ice Zone in spring 2016

Guy Williams, Delphine Lannuzel, Robert Massom, Klaus Meiners

Corresponding author: Guy Williams

Corresponding author e-mail: guy.darvall.williams@gmail.com

The seasonal retreat and melting of Antarctic sea ice is one of the largest physical changes on Earth, fundamentally impacting key mechanisms of the global climate system and marine ecosystem. It is also a key factor in the planning and logistics necessary to run icebreakers in support of National Antarctic programs. Extending poleward from the sea ice edge from October through December, the Marginal Ice Zone (MIZ) is a major conduit through which this dramatic seasonal event unfolds. A site of intense ocean–atmosphere interactions, the MIZ is a highly variable circumpolar zone that forms a dynamic interface between consolidated pack ice zone to the south and open ocean to the north. More than zone, the MIZ is a structural phase in process space that ultimately passes over the entire seasonal ice zone area. The MIZ also is critical to the ecology of Antarctic organisms as it supports major phytoplankton blooms in spring and summer and it is an important habitat for krill and dependant predators. The blooms are stimulated by the release of the key micronutrient iron from the melting sea ice. In spite of this crucial importance, our knowledge of the key physical processes occurring in the MIZ and their interaction with the biology and biogeochemistry is severely limited. Complex inter and intra-disciplinary connections present inherent sensitivities to external forcing and important feedbacks to the overall system. While traditionally ‘fly-through country’ for Australian Antarctic sea ice missions and logistic operations, a new multinational integrated process study of the MIZ is planned for spring 2016. This will capitalise on emergent techniques to resolve spatial and temporal variability and change, with the goal of advancing our mechanistic understanding of the Antarctic MIZ so that we can effectively predict and monitor its response to future climate change. Here we present the proposed fieldwork for feedback and potential contributions from the greater sea-ice community.


Is fast drifting sea ice resulting in unrealistic Antarctic sea ice in climate models?

Petteri Uotila, Paul Holland, Timo Vihma, Simon Marsland

Corresponding author: Simon Marsland

Corresponding author e-mail: simon.marsland@csiro.au

For the first time, we compute the sea-ice concentration budget of a fully coupled climate model, the Australian ACCESS model, in order to assess its realism in simulating the autumn–winter evolution of Antarctic sea-ice. The sea-ice concentration budget consists of the local change, advection and divergence, and the residual component which represents the net effect of thermodynamics and ridging. Although the model simulates the evolution of sea-ice area reasonably well, its sea-ice concentration budget significantly deviates from the observed one. The modelled sea-ice budget components deviate from observed close to the Antarctic coast, where the modelled ice motion is more convergent, and near the ice edge, where the modelled ice is advected faster than observed due to high ice drift speeds. In the central ice pack the agreement between the model and observations is better. Based on this, we propose that efforts to simulate the observed Antarctic sea-ice trends should focus to improve the realism of modelled ice drift speed.


The effects of UV radiation on brackish fast ice algal communities in northwest coast of the Gulf of Finland

Sara Enberg, Markus Majaneva, Jonna Piiparinen, Anssi Vähätalo, Riitta Autio, Janne-Markus Rintala

Corresponding author: Sara Enberg

Corresponding author e-mail: sara.enberg@helsinki.fi

The effects of ultraviolet radiation (UV radiation) on species diversity, cells numbers and photosynthetic effiencies in fast ice algal communities were studied in situ in the northwestern part of the Gulf of Finland during February and March 2011. The study site consisted of nine squares each one of which was 1 m2 and there were three different treatments: treatment with snow cover (UNT), treatment that was exposed to solar radiation by removing the snow cover (PAR+UV) and treatment without snow cover but covered with a foil that filtered UV radiation (PAR). The experiment was carried out in the vicinity of Tvärminne Zoological Station between 28 February and 14 March. The amount of biomass increased due to the increased amount of solar radiation during the experiment. The biggest biomass growth was observed in the bottom part of the ice in all three treatments. The biomass increase was the smallest in the PAR+UV treatment, which indicates that exposure to UV radiation had a negative effect on the biomass of algae and protists. The photosynthetic parameters differed between ice layers at the beginning and between ice layers and treatments at the end of the experiment. Photosynthetic capacity was highest in the surface layer of the PAR-treated ice, but in the UNT treatment the highest photosynthetic capacity values were found in the deeper layers of the ice. These results show that UV radiation has a negative effect on algal growth in the topmost 10 cm of the ice. No such effect was found in the deeper layers. This study shows that the effect of UV radiation on sea ice algal communities is biggest in the top part of the ice and some species may be more sensitive than others. The pennate diatoms were highly dominant species throughout the whole experiment. The pennate diatom Nitzschia frigida Grunow in Cleve & Grunow 1880 increased in the top 5 cm in the PAR treatmen,t indicating that UV radiation can be more harmful to some species than to others.


Variability mechanism and ice production budget of the Ross Ice Shelf polynya, Antarctica

Kazuki Nakata, Kay I. Ohshima, Sohey Nihashi, Noriaki Kimura, Takeshi Tamura

Corresponding author: Kazuki Nakata

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

Antarctic coastal polynyas, where dense water is formed due to high sea-ice production, play an important role in the formation of Antarctic Bottom Water (AABW). Pease (1987) provided a simplified polynya model, by considering a balance between the offshore sea ice drift and ice production within the polynya. This model, which incorporates both the dynamic and thermodynamic processes of the polynya in the simplest way, has become the base of subsequent polynya models. However, few studies have made direct comparison of the model with the observation incorporating the temporal variation, because of lack of appropriate data. In this study, we have examined to what degree the model can explain a real coastal polynya and how the model should be modified, using the polynya area and sea-ice production data by the thin ice thickness algorithm and sea-ice drift data from AMSR-E. We also estimated the collection depth of frazil ice (thickness when the frazil ice accumulates at the polynya edge), which is a key parameter in the model. Our target area is the Ross Ice Shelf Polynya (RISP), which has the highest sea ice production in the Antarctica and is a major AABW formation region. This study would also contribute to better understanding of the variability mechanism of the polynya and ice production related to the AABW formation. It is found that the Pease model represents the polynya variability very well by considering the lag time (1.5 days in the case of RISP) in which the frazil ice is accumulated at the polynya edge. The most suitable collection depth obtained by a least square fitting is about 14 cm, which is consistent with the values assumed in the past modeling studies. The extension of the polynya is made by the ice divergence of offshore drift and the retreat of the polynya is made mostly by sea ice production. We divided the sea-ice drift into the wind component and the remaining component (mainly regarded as the ocean current component), and evaluated each contribution to the polynya extension. It is found that the both components contribute comparably and that the ocean current component is mainly responsible for the offshore drift at the western part of RISP. This may partly explain larger extension of the polynya in the western part.


Antarctic coastal exposure index and its relationship to trends in sea ice

Phillip Reid, Rob Massom

Corresponding author: Phillip Reid

Corresponding author e-mail: p.reid@bom.gov.au

Ice shelves are considered an important and vulnerable component of the Antarctic cryosphere under and enhanced greenhouse environment. They act as a restrainer, slowing the pace of glacial (ice sheet) movement and hence sea level rise. In turn, sea ice itself acts as a protective barrier, reducing the impact of ocean processes (waves, swell and storms) on ice shelves. In this study we introduce and examine a simple, large-scale index called the Coastal Exposure Index. It effectively measures the amount of Antarctic coastline that is not protected by a sea ice barrier to its north on any given day. Hence, an increase in coastal exposure implies a potential removal of the protective sea ice barrier. Initial analysis suggests that, from 1979 through to the present day, there is a statistically significant trend for a decrease in coastal exposure during the late retreat through early advance sea ice season and a small increase in coastal exposure during the mid-sea-ice advance season. Here we put these trends into perspective and relate them to regional and seasonal processes. We find that the trend towards an increase in coastal exposure during the mid-sea-ice advance season is dominated by the reduction of sea ice around the Western Antarctic Peninsula. The trend towards a decrease in coastal exposure during the summer period is predominantly associated with an increase in sea ice in the Weddell and Ross Seas. We conclude that monitoring the Coastal Exposure Index, and coastal exposure itself, is worthwhile. It allows for detection of trends and emerging potential problem regions.


Thermodynamics of melting sea ice versus melting freshwater ice

Mareike Wiese, Philipp Griewank, Dirk Notz

Corresponding author: Mareike Wiese

Corresponding author e-mail: mareike.wiese@slf.ch

We investigate in detail the thermodynamics of melting sea ice in a combined lab experiment and 1-D modelling study. We contrast the melting of sea ice with that of freshwater ice to investigate the impact of salt within the ice. Both in our controlled laboratory experiments and in the model simulations, we find a distinct evolution of melt rates for constant boundary conditions. After a sudden increase in air temperature to above 0°C, freshwater ice quickly reaches a constant melt rate for constant boundary conditions. Sea ice, in contrast, initially melts slower because of its high heat capacity. But because of its lower solid fraction, sea-ice melt rates then quickly increase to exceed those of freshwater ice. Towards the final stages of melt the remaining sea-ice has been thoroughly desalinated and melt rates decrease again. An analysis of outgoing longwave radiation with an infrared camera reveals that during melt the mean surface temperature for both ice types usually exceeds 0°C, with higher surface temperatures found for sea ice compared to freshwater ice. This high surface temperature is caused by a thin liquid film that covers much of the surface of the ice during melting, which allows for an increase in outgoing longwave radiation compared to bare ice.


On the significance of blowing snow above sea ice as a source of polar sea salt aerosol

Markus Frey, Phil Anderson, Ian Brooks, Kouichi Nishimura, Anna Jones, Eric Wolff

Corresponding author: Markus Frey

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

The sublimation of saline blowing snow above sea ice has recently been suggested to generate more sea salt aerosol than is produced from a similar area of open ocean. To date no measurements exist to test this hypothesis, however understanding the production of polar sea salt aerosol is important for several reasons: sea salt aerosols contribute directly to the radiative balance and can act as cloud condensation nuclei. They can also significantly impact the lifetime of methane, ozone or mercury through the photochemical release of reactive halogens. Furthermore, knowing the origin of sea salt measured in ice cores and how it relates to sea ice surfaces will allow to develop a quantitative proxy of past sea ice extent, which is an important climate variable. A winter cruise of the German research vessel Polarstern through the Weddell Sea in June–August 2013 provided unique access to a potential source region of sea salt aerosol in the Antarctic sea ice zone during a time of year when it is thought to be most active. Reported are first measurements of snow particle as well as aerosol concentrations and size distributions during blowing snow events above the sea ice. Snow particle spectra are similar to those observed on the continent. Even though the salinity of surface and blowing snow was very low (<0.1 psu) a significant increase of aerosol in the size range of sea salt particles is observed during and after blowing snow events. The dependance of observed aerosol production rates on environmental conditions is discussed as well as their significance compared to the open ocean source. Active blow experiments show that frost flowers, which are highly saline (>60 psu), do break up and get airborne when exposed to strong airflow and therefore cannot be disregarded as an additional sea salt aerosol source as suggested recently.


Multi-scale granular modelling of sea-ice dynamics

Matthias Rabatel, Stéphane Labbé, Jérôme Weiss

Corresponding author: Matthias Rabatel

Corresponding author e-mail: matthias.rabatel@imag.fr

At intermediate ice concentrations, sea ice consists of separate ice floes of various sizes and shapes interacting with each other through inelastic collisions, in a way similar to a granular medium. We developed a multi-rigid-body frictional contact model in order to describe the dynamics of a population of floes of arbitrary shapes and sizes submitted to wind forcing and ocean drag. This granular model is based on simplified momentum equations for ice floe motion between collisions and on the resolution of linear complementarity problems to deal with ice floe collisions. Between collisions, the motion of an individual floe satisfies the linear and angular momentum conservation equations, with classical formulations to account for atmospheric and oceanic skin drag. In order to describe the rotational motion of a floe, especially when the floe is not a disk, each floe is meshed with finite-elements. Regarding ice floe collisions and interactions, unlike what is done in classical discrete elements models of granular media, we do not introduce fictitious interpenetrations to estimate interactions forces. Instead, we first use disks surrounding ice floes in order to accelerate the collision detection and to precise the description of the collision, i.e. of the geometry of the contact surfaces, of the normal and tangential forces between floes and of the resulting effects on floes dynamics. Secondly, to deal with the collisions before they lead to an interpenetration we write a linear complementarity problem with the Signorini-Fischera conditions and Coulomb’s law. The nature of the contact, elastic or inelastic, is described by a coefficient of restitution used in Signorini-Fischera conditions. In the present version of our model, this coefficient is fixed. At regional scales, this kind of model is suited for a dynamical description of an ice cover with a relatively low concentration, such as in the marginal ice zone. Improving the understanding of interactions between ice floes and offshore structures is important for the design of these structures and, as a consequence, is important to improve safety for on board personnel and environmental integrity. The next step will be to incorporate ice deformation and crack initiation and propagation as a mechanism of collisional energy dissipation. This will allow to follow the evolution of the floe population in terms of sizes and shapes as well.


Weddell Sea ice thickness from ICESat using an empirical approach based on in-situ drillings

Burçu Ozsoy-Cicek, Stefan Kern

Corresponding author: Burçu Ozsoy-Cicek

Corresponding author e-mail: burcu.ozsoycicek@gmail.com

The Antarctic sea-ice thickness distribution is still widely unknown despite substantial progress in satellite technology. The current method of choice to obtain an estimate of Antarctic sea ice thickness distribution is to take data of the freeboard from satellite altimetry and compute the sea ice thickness using the buoyancy approach. Using radar altimetry requires using the sea-ice freeboard. However, the complex structure of snow on Antarctic sea ice seems to limit applicability of radar altimetry for Antarctic sea ice. In contrast to radar altimetry, using laser altimetry involves the total (sea ice plus snow) freeboard. For both types of sensors additional parameters are needed: snow depth and densities of water, snow and sea ice. Densities cannot be measured contemporarily and therefore have to be taken from in-situ measurements which might not represent actual conditions. Snow depth can be measured by satellite microwave radiometry in parallel to satellite altimetry. However, such snow-depth data are known to underestimate actual snow depth for rough sea ice and during wet snow conditions. In order to circumvent the unknown snow depth bias over rough sea ice and/or flooded sea ice and in order to mitigate uncertainties due to density values that do not match the actual conditions, we apply a recently suggested empirical approach. This approach is based on in-situ drillings carried out on Antarctic sea ice and allows calculating sea ice thickness as a linear function of freeboard. This approach has been applied in the Bellingshausen-Amundsen Sea with encouraging results. In the present paper we calculate the sea ice thickness distribution in the Weddell Sea using the new empirical approach and using the classical buoyancy approach. Results will be inter-compared and discussed in the context of other studies and sea ice thickness datasets.


Snow depth on Antarctic sea ice from ICESat

Stefan Kern, Burcu Ozsoy-Cicek, Marcel Nicolaus, Georg Heygster, Torben Frost

Corresponding author: Stefan Kern

Corresponding author e-mail: stefan.kern@zmaw.de

Snow on sea ice is known as reducing the ocean–sea ice–atmosphere heat flux efficiently during winter because of its heat conduction being an order of magnitude smaller than that of sea ice. Snow on Antarctic sea ice can contribute substantially to the sea-ice volume formed every season via snow-ice formation. Snow on sea ice poses a significant impediment for an accurate sea ice thickness retrieval using satellite altimetry in the Arctic but more so in the Antarctic. The reason for this is that currently used sea-ice thickness retrieval approaches require snow depth. Snow depth derived from satellite microwave radiometry is the only long-term dataset currently available for Antarctic sea ice. The approach used, however, is sensitive to sea-ice type, surface roughness, and snow parameters such as wetness and grain size. This is caused by several factors, for instance, the average thin sea-ice cover promoting ice–snow interface flooding, the environment dominated by frequent cyclone passages promoting sea ice deformation caused by sea-ice retreat, and enhanced snow metamorphism. As a consequence, the dataset for Antarctic snow depth on sea ice has not well quantified uncertainty and likely underestimates actual snow depth over deformed sea ice and flooded sea ice. We apply an alternative method to derive snow depth on Antarctic sea ice. It uses empirical relationships which were developed from in-situ drillings on sea ice around Antarctica. These relationships suggest that snow depth can be derived from measurements of the elevation of the snow surface (the freeboard) above the sea surface. Such measurements were already provided by ICESat (Ice, Cloud, and land Elevation Satellite). Snow depth is calculated using the empirical relationships in combination with ICESat freeboard data in the Weddell Sea. Results are compared to in-situ drillings carried out during ship expeditions into the Weddell Sea, to ASPeCt (Antarctic Sea Ice Processes and Climate) protocol ship-based visual estimates of the snow depth, and to snow depth derived from contemporary passive microwave radiometry.


Is sea ice running the Atlantic water inflow?

Polona Itkin, Michael Karcher, Rüdiger Gerdes

Corresponding author: Polona Itkin

Corresponding author e-mail: Polona.Itkin@awi.de

Maximal compressive sea ice strength (P*) is an empirical parameter that is used in the rheology of sea ice models and controls the response of the sea ice to stresses. Its size has a considerable effect on the sea ice motion, its thickness and the location of open water in the simulation. Its value is usually fitted to achieve the best results in the sea ice drift and it has a wide range of acceptable values. In our coarse resolution sea ice–ocean model we compare simulations with two commonly used P* values: 27500 N m–2 in control run (CTRL) and 15000 N m–2 in weak sea ice run (WEAK). The differences in the sea ice simulations are the largest in the central Arctic sea ice motion and in the sea ice concentration in the Barents Sea and in the Nordic Seas. The latter two areas coincide with Atlantic water (AW) sinking areas. Our analysis show that the low sea ice concentration in the Barents Sea contributes to low ocean surface heat fluxes and to formation of cool and dense Barents Sea branch water (BSBW). The increased sea-ice mobility in the central Arctic in WEAK comparing to CTRL results in faster and deeper anticyclonic Beaufort Gyre which results in reduction of the cyclonic AW circulation beyond the Lomonosov Ridge and a stronger loop of the AW in the Eurasian Basin. As a results of both mechanism, the stronger BSBW formation and stronger Baufort Gyre, WEAK has about 0.2 K cooler mid-depth temperatures and about 0.3 Sv higher Fram Strait outflow than CTRL. Thus, sea-ice cover extent and sea-ice mobility have an indirect impact not only on the AW characteristic and circulation, but also on the volume transport balance between the Barents Sea Opening, Fram Strait and Davis Strait and finally also on the water mass characteristics in the deep convection areas of the Nordic Seas and Labrador Sea.


Multi-year/multi-decade ice about Mertz Glacier, East Antarctica, and its temporal evolution

Neal Young, Jan Lieser, Rob Massom, Alex Fraser, Dana Floricioiu

Corresponding author: Neal Young

Corresponding author e-mail: neal.young@utas.edu.au

Large areas of multi-year and multi-decade sea ice occur to the east of the Mertz Glacier Tongue (MGT) in East Antarctica. Measurements of the free-board height of this ice obtained by satellite-borne laser and radar altimeters indicates that the ice is many metres thick and in places tens of metres thick. Visual observations from ships of the above water layers of thick floes from the region suggest that the free-board is composed of snow. Snow accumulation rate for the region is estimated at around 1 m a–1 from firn core data retrieved from the large iceberg B09B when it was grounded to the east of MGT. So the primary process contributing to the large thickness of the floes appears to be accumulation of snow at the surface and flooding to create snow-ice. The thick ice in the region has been described as landfast, however long fractures do form through the ice-field and allow the ice to diverge somewhat and move. But, the ice floes typically remain in place. They converge again, and export of ice appears to occur rarely. Re-location of iceberg B09B and calving of MGT in early 2010 brought about major changes. Large sections broke-up and drifted out to the west through what was the location of the Mertz Glacier polynya. The breakup and melting of the snow cover introduced a fresh-water pulse into the upper layers of the ocean in the region. We derive the temporal evolution of the spatial distribution of this thick ice and its growth through analysis of time series of SAR images and altimeter data from ICESat and CryoSat-2.


Robotic lasers, sea ice science and hacky sacks

Adam Steer, Petra Heil, Rob Massom, Jan Lieser, Guy Williams

Corresponding author: Adam Steer

Corresponding author e-mail: adam.steer@utas.edu.au

During the Sea Ice Physics and Ecosystems eXperiment (SIPEX) II in austral spring 2012, the deployment of a suite of state-of-the-art surveying equipment enabled a new approach to sea-ice measurements. At each ice station, a local coordinate system was established using a Total Robotic Station (TRS) and a pair of geodetic GPS stations. For the first time in the AAD sea-ice science program, every activity on the ice floe could be given a spatial context relative to other measurements taken at the site. Further, sea-ice thickness observations could be taken anywhere on an ice floe with its location referenced precisely. This capacity was explored firstly as a way to choose an optimum geometry for ice-thickness and draft measurement sites and create ‘tie points’ for under- and over-ice surveying craft. It also allowed the creation of an ‘ad hoc’ transect line for one ice station, and led to a semi-random sampling technique in which a weighted marker – a ‘hacky sack’ – was thrown to determine sampling sites. For each of these activities, the positions of sampling sites are precisely recorded with respect to other floe activities, and given a geographic location by the GPS sites which formed the ‘datum line’ at each site. While the ability of TRS-guided semi-random and ad-hoc sampling methods to characterise sea-ice thickness variability needs to be further verified, use of the TRS frees direct sea-ice sampling from the constraints of one or more transects across the ice floe; and allows integration with coincident under- and above-ice remote sensing platforms.


The role of Arctic sea ice on near-surface temperature trends

Eveline van der Linden, Richard Bintanja, Wilco Hazeleger, Caroline Katsman

Corresponding author: Eveline van der Linden

Corresponding author e-mail: linden@knmi.nl

Long-term global near-surface temperature trends in response to rising greenhouse gas concentrations in climate models vary by almost a factor two, with greatest intermodel spread in the Arctic region, where sea ice is a key climate component. Three factors govern the intermodel spread: (1) model formulation of physics and dynamics, (2) control climate (CTRL) state, and (3) internal climate variability, but how much each factor contributes is as yet unknown. Our aim is to obtain a better understanding of the contribution of the CTRL sea ice state on near-surface warming, by expanding upon previous studies with a multi-scale analysis: varying from a classification based on local sea ice conditions to the entire Arctic. The relation between CTRL sea-ice conditions and near-surface warming is examined through analysis of idealized 1% per year CO2 increase simulations of a multimodel ensemble of 33 state-of-the-art global climate models participating in the fifth phase of the Coupled Model Intercomparison Project (CMIP5). On the Arctic mean scale, the intermodel spread in near-surface temperature trends is only weakly related to CTRL ice volume or area, and is not dominated by internal variability. This suggests that other processes, such as ocean heat transport and meteorological conditions, play a more important role in the spread of long-term warming than CTRL sea ice conditions. However, on a local scale, sea-ice-warming relations show that in regions with more sea ice, models generally simulate more warming in winter and less warming in summer. The local winter warming is clearly related to CTRL sea ice and universal among models, whereas summer sea-ice-warming relations are more diverse, and are probably dominated by differences in model formulation. To obtain a more realistic representation of Arctic warming it is recommended to simulate CTRL sea ice conditions in climate models so that the spatial pattern is correct.


Changes in Fram Strait sea-ice volume export between 1992 and 2012 observed in combined ULS and satellite data

Gunnar Spreen, Edmond Hansen, Ron Kwok, Jennifer King, Sebastian Gerland

Corresponding author: Gunnar Spreen

Corresponding author e-mail: gunnar.spreen@npolar.no

Fram Strait between Svalbard and Greenland is the main gateway for sea-ice export out of the Arctic Basin. Changes of the sea-ice mass balance can either have thermodynamic (melting) or dynamic (export) causes, or a combination of both. Therefore, to better understand the recent decrease in Arctic sea ice volume it is of special importance to monitor changes in the sea ice export. Additionally, inter-annual perturbations in the sea ice transport through Fram Strait can modify the major water mass formation processes in the Greenland Sea and further downstream with consequences for the deep water formation and global ocean circulation. To estimate the sea-ice volume export through Fram Strait the three variables sea-ice drift, area, and thickness have to be known and combined. The long-term Fram Strait sea-ice area export obtained from passive microwave satellite observations since 1982 shows no significant trend. However, for the last decade 2001–09 a positive trend in sea-ice area export can be observed. The sea-ice thickness in Fram Strait at 79°N is monitored by Upward Looking Sonars (ULS) since 1990. A newly reprocessed and extended ULS time series for 1990–2011 shows negative ice thickness trends of about 3 and 5 cm per year for the mean and modal ice thickness, respectively, with an increasing negative trend since about 2003. A preliminary sea-ice volume export time series based on monthly data for winter months 1992 to 2011 shows a negative trend in sea-ice volume export through Fram Strait. This reduction in ice volume export is mainly due to the reduced ice thickness as the ice area export does not change significantly during this time period. We will present an updated Fram Strait sea ice volume export time series for the period 1992 to 2012, combining in-situ ULS ice thickness data with satellite passive microwave observations of sea ice drift and area. Long-term and decadal changes of the ice export are discussed in context with the recent sea ice area and volume decrease in the Arctic Basin.


Snow thickness retrieval from SMOS satellite data over Arctic sea ice: sensitivity to surface temperature and ice and snow conditions

Nina Maaß, Lars Kaleschke, Xiangshan Tian-Kunze, Rasmus T. Tonboe

Corresponding author: Nina Maaß

Corresponding author e-mail: nina.maass@zmaw.de

The Soil Moisture and Ocean Salinity (SMOS) mission measures brightness temperatures at a low microwave frequency of 1.4 GHz in the L-band. SMOS provides a daily coverage of the polar regions and its measurements have been used to retrieve ice thickness over thin sea ice. In addition, there has been an attempt to extract information on the snow thickness over thick Arctic sea ice. Information on snow thickness is required for the freeboard-based estimation of sea ice thickness from lidar and radar altimetry, for example. For the L-band retrieval of ice thickness over thin ice and snow thickness over thick ice, we have used a radiation model that describes the brightness temperature over sea ice mainly as a function of surface temperature, ice salinity, snow density, and of the ice and snow thicknesses. In this presentation, we focus on the potential to retrieve snow thickness from SMOS measurements. Considering the expected uncertainties for the remaining ice parameters and their impact on the brightness temperature, we investigate for which ice conditions the brightness temperature’s sensitivity to snow thickness is higher than the sensitivities to the other ice parameters, which is the basis for the feasibility of SMOS data for the retrieval of snow thickness. We compare the brightness temperature’s sensitivities to the ice conditions, as estimated from our radiation model, with the sensitivities obtained from brightness temperature simulations that are based on a more complex emission model, namely a sea ice version of the Microwave Emission Model for Layered Snowpacks (MEMLS). Subsequently, we estimate the uncertainty of the snow thickness retrieval with regard to the other ice parameters for different ice conditions. Additionally, we compare our theoretically derived retrieval uncertainties with deviations between snow thicknesses as retrieved from SMOS and as measured during the IceBridge flight campaigns in the past winter seasons. Compared to ice salinity and snow density, surface temperature in general has the largest impact on the brightness temperatures. Thus, we here, more specifically, use the IceBridge validation data to compare the SMOS retrieval’s performance in case our radiation model is driven with surface temperature information from reanalysis data, from space-borne thermal imagery as provided by Moderate Resolution Imaging Spectroradiometer (MODIS) measurements, or from IceBridge’s infrared measurements, respectively.


A sensitivity study of the viscoelastic wave–ice interaction model

Jingkai Li, Hayley Shen

Corresponding author: Hayley Shen

Corresponding author e-mail: hhshen@clarkson.edu

A rheological model devised to capture both the elastic and viscous properties of various ice covers was proposed recently. Under proper limiting conditions, this model was shown to converge to previously established continuum models: the thin elastic, the viscous layer, and the mass loading. There are two mechanical parameters in this model: the elasticity and the viscosity. Depending on the ice cover thickness, the wave frequency, and the water depth, the sensitivity of the model outcome on these parameters, namely the wave speed and its attenuation, may be varied. We present the results of a sensitivity study for this model. This study may help to determine the relative importance of these parameters in different ranges of the physical conditions.


Spectral analysis of Amundsen Sea pack ice roughness and estimates of air–ice drag coefficient

Blake Weissling, Stephen Ackley

Corresponding author: Blake Weissling

Corresponding author e-mail: blake.weissling@utsa.edu

High spatial resolution surface topography from terrestrial laser scanning (LiDAR) and sea-ice thickness and snow depth were obtained from in situ electromagnetic induction (EMI) sounding surveys at three Amundsen Sea ice stations on the transit of the Oden Icebreaker during the Southern Ocean 2010–11 expedition. These data have been investigated in order to assess snow surface, ice surface, and ice bottom roughness, associated wave number dependency, and air–ice drag coefficients through spectral power analysis. The roughness of an ice sheet (top ice surface and bottom) has previously been demonstrated to reflect its formation and deformation history while the snow surface roughness may be frictionally coupled with the atmosphere, parameterized as the neutral stability drag coefficient (CDN10 – height referenced to 10 m). This study represents the first attempt at parameterizing sea ice roughness through spectral techniques for the Amundsen Sea, with the Winter Weddell Gyre Study of 1989 providing the only Antarctic dataset for comparison. The Weddell project, based on over 4700 holes drilled on 47 sets of 100 m transects on the pack ice, yielded wavenumber (k) dependencies of approximately k-1 for snow, top and bottom-side ice surfaces – in stark contrast to studies of Arctic sea ice with spectral roll-offs of >k-2. Our spectral analysis has indicated snow and ice surface spectra for the Amundsen pack ice roll off at rates matching Arctic ice, for wavenumbers between 0.1 and 3 rad m–1, or wavelengths between 2 m and 60 m – suggesting more Arctic-like ice formation and deformation conditions for the Amundsen. Air–ice momentum coupling through the estimation of the drag coefficient is computed from a roughness scale ζ for wavelengths between 2 m and 13 m. Our CDN10 fall within the range of those seen in the Weddell pack ice and are consistent with 10 m tower-based eddy covariance assessments. Terrestrial LiDAR scanning coupled with EMI sounding for spectral roughness studies provides for a viable alternative to the traditional ice drilling and transit/stadia surveying approach, with at least an order of magnitude higher spatial resolution and a fraction of the time commitment. Sampling considerations and other lessons learned in terrestrial LiDAR scanning in a pack ice environment, in the context of spectral analysis studies, are discussed.


Spatial variability of surface roughness and snow resistance of snow covered sea ice

Nander Wever, Katherine Leonard, Michael Lehning

Corresponding author: Nander Wever

Corresponding author e-mail: wever@slf.ch

Exchange processes of heat, moisture, particles and trace gases between sea ice and the atmosphere are heavily influenced by the surface roughness. Remote sensing methods also require surface roughness estimates for retrieval algorithms. The most important direct influence of sea ice surface roughness is on the large-scale movement of the ice. Surface roughness of sea ice exhibits large spatial and temporal variability, as it varies with snow cover thickness, ice type and the presence of ridges. Direct in-situ measurements and in-situ information of the temporal evolution of surface roughness elements are scarce and mostly limited to summer time. This study presents wind speed profile measurements using a 2 m high measurement mast with wind speed sensors at 3 heights, terrestrial laserscanning (TLS) data of the snow surface and Snow Micro Penetrometer (SMP) measurements of the internal micro-structure of the snow pack down to 36 cm depth. The data are used to analyze the variability of surface roughness for snow covered sea ice at typical length scales of 10–100 m. Data have been collected during the Antarctic Winter Ecosystem Climate Study (AWECS) winter expedition in the Weddell Sea (June–August 2013), at several ice stations at the east–west transect followed by the expedition. Ice stations were located on both newly formed ice from the beginning of the winter season with a shallow snow cover to multi-year ice with a thick snow cover. Surface roughness estimates from the TLS data are compared to roughness estimates from the measured wind speed profile and the turbulence signal of the measured wind speed. It is shown how these roughness estimates varies with wind direction for a specific floe and between different floes with varying topography. Furthermore, high density SMP measurement transects are used to explore potential links between surface features and snow microstructure properties. This improves the understanding of the persistence of snow surface features under the influence of snow fall and snow drift.


Ocean heat flux under Antarctic sea ice in the Bellingshausen and Amundsen Seas

Stephen Ackley, Elizabeth Murphy, Hongjie Xie

Corresponding author: Hongjie Xie

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

The role of Ocean Heat Flux (OHF) in sea ice growth and melt is examined using data from autonomous ice mass-balance buoys deployed on drifting ice in the Bellingshausen Sea and on fast ice in the Amundsen Sea during the spring–summer and summer–fall transitions, respectively. The buoys collected half-hourly measurements of ice thickness, temperature gradients through the ice, and the below-ice ocean salinity and temperature. We derive OHF using three methods that examine (a) changes in the sub-ice ocean properties (OHF-0), (b) changes in the ice thickness and temperature gradient (OHF-1), and (c) a hybrid of both (OHF-H). Good agreement is found between the time-averaged estimates of OHF-0 and OHF-1, using the same heat exchange coefficients in both cases. Average OHF measured was 8±2 W m-2 under the drifting pack ice in the Bellingshausen Sea from Oct – Dec 2007 and 18±2 W m–2 under the landfast ice in the Amundsen Sea from Feb – Mar 2009. Some short-term values of OHF-0, from ocean measurements in both the Bellingshausen and Amundsen Seas exceeded 55 W m–2. Semi-diurnal fluctuations of OHF in the Bellingshausen Sea coincided with semi-diurnal variations of ice velocity. Temperature excursions in the mixed layer were irregular, with several intermittent spikes (separated by days) to 0.13°C and ~0.35°C above the in situ freezing point in the Bellingshausen Sea and Amundsen Sea, respectively. The spring OHF variations in the Bellingshausen Sea were periodic and controlled by semi-diurnal ice velocity fluctuations. Larger temperature fluctuations, originating from incursions of warm deep water masses, drove OHF variability in the summer Amundsen Sea contributing to the twice higher OHF as compared to the Bellingshausen Sea.


Summer surface albedo of Pacific Arctic sea ice: a case study from ice stations during the 2010 cruise

Wentao Xia, Hongjie Xie

Corresponding author: Hongjie Xie

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

Assessing Arctic snow/ice surface albedo change is essential to understand local energy budget and ice/albedo feedback under global warming scenario. In-situ measured albedo plays an essential role by providing ground truth reference for remote sensing monitoring of albedo change in large scale. Such measurements with three portable spectroradiometers were conducted in one 12-day and seven short-term (3–4 h) ice stations over the Pacific Arctic sector during the 2010 Chinese Arctic Expedition. Results show the albedo of decomposed granular ice was 0.65~0.69 in the low latitude and marginal ice zone, while snow albedo reduced from 0.86 to 0.67 in one week during the high latitude long-term ice station. The albedo of melt ponds varied due to different surface cover over the ponds such as water, refrozen ice, and snow. Snow covered melt pond was only seen in the long-term ice station, with albedo reduced from 0.75 to 0.55 in one week. The ice covered melt pond had albedo from 0.38 to 0.61, varied greatly in different latitudes. The water covered melt pond had albedo of 0.27, higher than the sea water albedo of 0.18. Such rapid decreasing of sea ice albedo and evident difference in polar region and marginal ice zone contribute much to the surface ice/albedo feedback and rapid shrinking summer sea ice extent in Arctic Ocean.


The record 2013 Southern Hemisphere sea-ice extent maximum

Phil Reid, Sharon Stammerjohn, Rob Massom, Ted Scambos, Jan Lieser

Corresponding author: Phil Reid

Corresponding author e-mail: p.reid@bom.gov.au

September 2013 saw a new Southern Hemisphere (SH) satellite-era maximum sea ice extent record set, just one year after September 2012’s record. NSIDC reported a 18 September 2013 extent of 19.471 × 106 km2 (7.52 × 106 miles2), higher (but within the range of measurement error) than 26 September 2012’s previous maximum value of 19.470 × 106 km2 (7.52 × 106 miles2). The two winters were markedly different in terms of climate and regional pattern, however. We examine the large-scale processes that led to the 2013 event, contrast the climate patterns in 2013 and 2012, and put the record into perspective with recent regional and seasonal sea ice trends. From January through at least September 2013, SH total sea-ice extent and sea-ice area were well above average, with many days setting new record high extents for those dates (for the period 1978–present). In September the regions showing the largest positive extent anomalies were the northern Ross Sea, northern Amundsen, and northwestern Weddell – a pattern inconsistent with the regional extent trends from the last 30+ years of sea-ice records. During the sea-ice advance period (April–early August) the Southern Annular Mode (SAM) index was mostly positive, with a neutral ENSO index. A wind-driven sea ice edge expansion augmenting normal thermally-driven ice growth occurred during this period. In mid-August, wind-driven sea ice expansion decreased as westerly winds subsided, replaced by a thermally-driven sea-ice advance through August and September. The SAM index value is strongly negative in mid-August. A high air pressure anomaly covered continental areas, and lower-than normal pressures were present just outside the sea ice edge, reducing westerly wind strength. Slightly cooler-than-average temperatures prevailed near the ice edge as well. Additionally, from May through September, persistent cyclonic circulation in the Ross Sea provided conditions suitable for upwelling along the sea ice edge, which in turn induced a cold pool of water to the north of the ice edge in that region. This provided ideal conditions for combined thermal and wind-driven expansion in this region – the main region of positive sea-ice extent anomaly. We find that the 2013 anomalies are the result of wind-driven ice expansion during the early growth period, followed by thermally driven ice growth in the latter phase of winter 2013. We also discuss the role of climate change in these recent record sea ice extents.


The changing Antarctic sea ice environment

Sharon Stammerjohn, Rob Massom, Phil Reid, Ted Maksym, Marilyn Raphael

Corresponding author: Rob Massom

Corresponding author e-mail: Rob.Massom@aad.gov.au

The stark contrasts between widespread decreases in Arctic sea ice versus the small but significant increases in Antarctic sea ice poses many questions about the nature of climate change in Antarctica. Interpreting these changes is further complicated by the fact that most models simulate Antarctic sea-ice decreases, not increases as observed. The poleward intensification of westerly winds over this time period is thought to be one factor contributing to the positive Antarctic sea-ice trend, but models that successfully capture this change show it leads to sea-ice decreases, not increases. So what are we missing here and how can we improve this situation? We discuss what we know (and don’t know) about Antarctic sea-ice change, the regional/seasonal differences in wind-driven versus thermal-driven sea-ice changes (and feedbacks), the regional/seasonal connections to large-scale climate variability, the role of snow and freshwater changes in affecting air–ice–ocean interactions and ice mass balance. If we wish to improve our understanding of Antarctic sea ice and how it is changing, we need integrated datasets e.g., in situ multi-platform time series observations that capture air–ice–ocean interactions from before the autumn freeze/advance to after the spring melt/retreat. But, until recently, such coordinated observations have been difficult, if not impossible, to obtain. Fortunately, new technologies and techniques are fast emerging, and recent process studies are proving their success.


Seasonal cycle and variability of Arctic sea ice and relationships with land/coastal cryosphere

Hiroyuki Enomoto, Yasuhiro Tanaka, Nuerasimuguli Alimasi, Kazutaka Tateyama

Corresponding author: Hiroyuki Enomoto

Corresponding author e-mail: enomoto.hiroyuki@nipr.ac.jp

Seasonal cycle is a very principal variation of sea-ice change and has been studied for basic understanding of sea-ice study. In recent dynamic changing of sea ice in the Arctic, we summarize recent understanding and discuss the fundamental seasonal cycle. This study describes seasonal sea-ice variation in the Arctic, then discusses sea-ice variation in the other ice covered regions. Summer Arctic sea decreasing and summer minimum has the largest interannual variability. Satellite data focuses summer ice reduction and the minimum area is strong concern. Although Arctic summer sea-ice minimum is strong concern, other important timing or inflections are in the other season has been not well evaluated. Winter sea ice shows also large variability, however this is due to sea ice variability in the sub Arctic sea. Sea-ice fluctuation in the Sea of Okhotsk influences winter sea-ice variation in the northern hemisphere, and produces the decreasing trend in winter maximum sea-ice area. There are two timing of small deviations, appearing as tie points. These timings appear in May and November. Although winter and summer standard deviations are large, the standard deviations in these two timings are small. This indicates strong control of sea ice extent in these seasons. The extent of Arctic Ocean is essential condition. As the ice edge reached the most coastal regions, the interannual fluctuations are masked in the following period, by the present Arctic climate conditions. From the areal information, summer large deviation seems to start after May. It is well expected that, even in same ice area, the ice types are different. This can be indicated by thickness or age of ice. This preconditioning will affect sea ice retreat after this period. Some examinations on sea-ice forecast shows ice thickness at the minimum ice cove in the end of summer influences greatly to the next years ice conditions. There are interesting approaches on the summer ice retreat based on the winter ice dynamic behaviors. Summer melting shows trends of longer duration. This is due to extending of melt season. The peak time is also shifted to later. The coastal freezing and snow cover conditions are prceeded to the sea ice formation in autumn and also same thing happens for melting. This study also discusses land and ocean cryospheric relationships.


New Japanese Arctic climate change research project and cryosphere/sea-ice research activity

Hiroyuki Enomoto

Corresponding author: Hiroyuki Enomoto

Corresponding author e-mail: enomoto.hiroyuki@nipr.ac.jp

GRENE-Arctic project is a new initiative of Arctic study by more than 30 Japanese universities and institutes as the flame work of GRENE (Green Network of Excellence) of MEXT (Ministry of Education, Culture, Sports, Science and Technology, Japan). The new Arctic Climate Change Research Project ‘Rapid Change of the Arctic Climate System and its Global Influences’ has started in 2011 with strategic research targets: – Understanding the mechanism of warming amplification in the Arctic – Understanding the Arctic system for global climate and future change – Evaluation of the effects of Arctic change on weather in Japan, marine ecosystems and fisheries – Prediction of sea-ice distribution and Arctic sea routes. This project aims to realize the strategic research targets by executing following studies. The project is aiming to integrate interdisciplinary research work and collaboration of observation and modeling work. This Arctic project is establishing Arctic data archive System (ADS) for effective collaboration among project member and also for public use.


Potential links between sea-ice change and recent Antarctic ice-shelf disintegration

Rob Massom, Ted Scambos, Vernon Squire, Tim Williams, Sharon Stammerjohn, Phil Reid, Mike Pook, Ian Simmonds

Corresponding author: Rob Massom

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

The Antarctic coastal zone is an area where two keys elements of the global cryosphere, namely ice sheet and sea ice, are closely associated and strongly interact for much or all of the year (depending upon the region). Sea ice has an indirect effect on floating ice-sheet margins by greatly modifying the properties of the ocean, thereby affecting ice shelf basal melt and thinning. In the case of fast ice, the sea ice is also locked onto the ice sheet margin and is mechanically bonded to it – perennially in certain areas and as a consolidated plate that can attain considerable thicknesses. Sea ice also dampens the potentially destructive energy of ocean waves encroaching upon the Antarctic coast, while itself being modified by the waves. Given these factors and the potentially important (though poorly understood) role of sea ice in ice sheet margin processes, it follows that any significant change in the regional coastal sea ice environment may affect adjacent ice-sheet margins which, when floating, are particularly vulnerable to climate change. In this work, we introduce the possibility that well-documented regional changes in the duration of seasonal sea-ice coverage adjacent to the Larsen A and B and Wilkins ice shelves over the past three decades, and particularly since the early 1990s, may have played some role in their extraordinarily abrupt and rapid disintegration (since 1995). Specifically, we implicate change in regional sea ice seasonality in the nature and timing of these disintegration events, which are iconic examples of recent environmental change, and also in the subsequent dispersal of the resultant plethora of icebergs. We hypothesize that the increasingly prolonged reduction in sea-ice coverage may have exposed weakened outer ice-shelf margins to enhanced wave attack over prolonged periods, to contribute to the loosening and subsequent release of the outer margins in the form of characteristic elongated bergs that preceded each disintegration event. In the case of the Wilkins Ice Shelf, we also examine the possibility that change in fast ice coverage was also involved.


A seven-year analysis of Arctic sea ice using Envisat/RA-2 and IceBridge measurements over ice in the Canada Basin

Laurence Connor, Sinead Farrell, Andrew Ridout, Thomas Newman, David McAdoo, William Krabill

Corresponding author: Laurence Connor

Corresponding author e-mail: laurence.connor@noaa.gov

On 27 March 2006, the Arctic Aircraft Altimeter campaign was carried out jointly by NASA and NOAA using airborne laser altimetry and photo imagery to validate sea ice elevation measurements derived from the Envisat/RA-2 microwave altimeter. A NASA P-3 aircraft was flown along an Envisat ground track northwest of the Canadian Archipelago and just north of the Beaufort Sea, marking the beginning of a seven-year sequence of repeated airborne measurements along this track. Operation Ice Bridge (OIB) took up this flight line again in its early spring Arctic campaign in 2009 and continued to repeat measurements along the line in the spring campaigns of 2010, 2011 and 2012. The OIB flights also saw the addition of ku-band and snow radars and high-resolution digital photography. This 2006–12 time span encompasses a period of particularly large, interannual loss of Arctic sea ice. The satellite and airborne data collected over the seven-year span, from 2006 to 2012, offers an unprecedented measurement timeline over a large region of the Arctic which holds one of the last areas of extensive, thick, winter sea ice. Altimetry and snow radar data from these flights are used to examine springtime evolution of sea ice freeboard and thickness between 2006 and 2012. Additionally, high-resolution digital photography is used to supplement altimetry measurements, enabling an analysis of lead distribution and sea ice drift dynamics in this region of the Arctic. Results from this analysis are presented in the context of a rapidly changing Arctic environment that can only be understood through large-scale, diverse and frequently repeated observations.


Ice Watch and ASSIST (Arctic Shipborne Sea Ice Standardization Tool)

Alice Orlich, Scott MacFarlane, Tiffany Green, Jennifer Hutchings

Corresponding author: Jennifer Hutchings

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

Ship bridge based visual observations of sea ice have been recorded in a standardized way from research vessels in the Antarctic since 1996. The Antarctic Sea Ice Processes and Climate (ASPeCt) observation methods and tools have been increasingly utilized in the Arctic, though these protocols do not account for the unique nature of the Arctic pack ice were melt ponds in summer are prevalent. Through consultation and collaboration with sea ice field researchers we have developed extensions to ASPeCt protocols to account for typical surface melt conditions in the Arctic. We have also developed a new web browser based tool to record observations and disseminate this data to the Ice Watch data network hosted by the Geographic Network of Alaska (GINA) and the International Arctic Research Center. Data is freely available (http://www.iarc.uaf.edu/en/icewatch), and we welcome submission of historical data to this new database of northern hemisphere ship based visual sea ice observations.


A comparison of sea-ice surface elevation and local sea-level estimates from Cryosat-2 and Operation IceBridge under-flights in Arctic Ocean

Wentao Xia, Hongjie Xie, Stephen Ackley

Corresponding author: Hongjie Xie

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

Abstract: The European Space Agency Cryosat-2 (CS-2) mission has been providing high-spatial resolution cryosphere observations with space-borne synthetic aperture interferometric radar altimeter (SIRAL) since 2010. Arctic sea-ice surface elevations from the CS-2 SAR products are compared with the NASA Operation IceBridge (OIB) under-flight observations, to validate the CS lead detection and local sea level estimation. With the comparison of the OIB under-flights during March 2012, the CS-2 altimetry and lead detection algorithm are analyzed with reference to the OIB’s Airborne Topographic Mapper (ATM) and the Digital Mapping Camera System (DMS). Results show that: (1) There is a consistent bias (1–2 m less) in the measured surface elevation (with respect to WGS 84 ellipsoid) of the CS-2 L2A products as compared with the ATM L2 data, partly due to snow penetrating ability of the Ku-band radar onboard CS-2; (2) Referenced with the DMS, the CS-2 lead detection algorithm is sensitive to lead as small as 5–10m in width or diameter; (3) However, the CS-2 elevation for those footprints with leads seems primarily respond to signal from the snow-covered ice, since the lead is too small in March.


Polar research initiatives and opportunities of the US Navy’s Office of Naval Research (ONR) and ONR Global

Marvin McBride

Corresponding author: Marvin McBride

Corresponding author e-mail: marvin.b.mcbride.mil@mail.mil

The Office of Naval Research Global (ONR Global) works with scientists around the world to improve scientific understanding through international collaboration. This presentation will present an overview of the funding programs ONR Global uses to foster collaboration and new research, which includes funding Non-US researchers to travel, support conferences/workshops and finding cost sharing opportunities to fund research proposals. The aim of these tools is to develop international collaborations in order to find and support the best science while. The Arctic and Global Prediction research initiatives of ONR are the Marginal Ice Zone (MIZ) initiative and the Sea State research initiative. The MIZ experiment is a 5-year Department Research Initiative (DRI), scheduled from FY12-FY16. The main field experiment will be held in spring–summer–early autumn 2014 in the Beaufort Sea and will consist of deployed acoustic navigation and communication arrays, under-ice gliders and floats, ice-tethered profilers, ice mass balance buoys, and wave buoys on ice and in open water. The Sea State experiment is a 5-year, $9M DRI held from FY13-FY17. The main field experiment will be conducted in spring–summer–early autumn 2015 in the Beaufort and Chukchi seas. ONR anticipates the need for international cooperation for vessel(s) to support deployment of moorings, and wave buoys in open water and in marginal ice zone.


Bromine explosion events driving mercury deposition over the East Antarctic during the SIPEX II cruise in polar spring 2012

Robyn Schofield, Ruhi Humphries, Caitlin Gionfriddo, Karin Kreher, Paul Johnston, Sarah Connors, Neil Harris, Xin Yang, Michael Tate, David Krabbenhoft, John Moreau, Suzie Molloy, Ian Galbally Andrew Klekociuk Andrew Bowie

Corresponding author: Robyn Schofield

Corresponding author e-mail: robyn.schofield@unimelb.edu.au

Bromine explosion events occurring over first year sea ice in the polar spring cause mercury deposition and dramatically alter the oxidative power of the polar boundary layer. Atmospheric chemical measurements made during the Sea Ice Physics and Ecosystems eXperiment (SIPEX II), in the polar spring of 2012 are presented here. UV-Vis Multi-AXis Differential Optical Absorption Spectrometer (MAX-DOAS) measurements of BrO and HCHO profile concentrations are presented along with gaseous elemental mercury (GEM) and in-situ ozone observations. In addition, snow–sea–ice and ocean samples made at several ‘ice stations’ between 62– 65°S and 115–121°E that were analyzed for total mercury and methylmercury are presented. Several GEM depletion events were observed. This suite of observations provides the most complete suite of observations to date to improve our understanding of the implications of BrO explosion events in the Antarctic. We present a comparison of these atmospheric chemistry observations with the UK Chemistry and Aerosol (UKCA a module within the UK’s Unified Model) global chemical transport model simulations of BrO, O3 and a newly integrated mercury scheme.


Before satellites: reconstructing past sea ice changes using proxy records

Nerilie Abram

Corresponding author: Nerilie Abram

Corresponding author e-mail: nerilie.abram@anu.edu.au

Despite its importance in Earth’s climate system, very little information exists to constrain sea ice processes prior to the satellite era. The enigmatic nature of sea ice means that it doesn’t not impart a direct signal on the geological record, however there are many indirect indicators of sea ice presence. This presentation will give an overview of the various proxies that exist for reconstructing aspects of sea ice change prior to direct observations. There will be a particular focus on the advances that have been made over the last decade in understanding the history of Antarctic sea ice from ice cores. Some examples of the challenges in linking proxy reconstructions of sea ice with physical climate processes will also be discussed.


To the two ends of the Earth: Arctic and Antarctic sea-ice variability and environmental change

Ian Simmonds

Corresponding author: Ian Simmonds

Corresponding author e-mail: simmonds@unimelb.edu.au

This presentation will highlight the nature of sea-ice variability in both the Arctic and Antarctic regions over recent decades. The trends in these two polar domains have been very different, a circumstance which represents a challenge to our scientific understanding of the responsible mechanisms. The broad morphological environments of the Arctic and Antarctic zones are very different, a fact which goes part of the way to explaining why their sea-ice behaviour differs. Focus will be placed on how atmospheric and oceanic circulations are influencing regional sea ice distributions. Among these drivers are the atmospheric (northern and southern) annular modes, cyclonic activity, baroclinicity, teleconnections and changes in water mass distributions. Attention will also be devoted to exploring how changes on sea ice may be impacting on both local and midlatitude weather and climate.


Local effects of ice floes in the marginal ice zone and of ice-sheet margins in fjords from airborne platforms

Christopher J. Zappa, Scott Brown, David F. Porter, Robin E. Bell, William Emery, John Adler, Gary A. Wick, Michael Steele, Scott Palo, Gregory W. Walker, Jim Maslanik

Corresponding author: Christopher J Zappa

Corresponding author e-mail: zappa@ldeo.columbia.edu

Airborne remote sensing, in particular InfraRed (IR), offers a unique opportunity to observe physical processes at sea-ice margins including marginal ice zones (MIZ) and fjords. MIZ, or areas where the ‘ice-albedo feedback’ driven by solar warming is highest and ice melt is extensive, may provide insights into the extent of the extreme changes in sea ice in a changing climate. The calving fronts in Greenlandic fjords are an important boundary condition for understanding circulation and melt processes within the fjord, and crucial to the understanding of ocean–ice interactions at the margins of the Greenland ice sheet. The inaccessibility of much of these polar systems leaves many regions poorly surveyed, with important data gaps remaining throughout the Arctic. Airborne surveys provide a valuable platform for the study of sea ice–ocean dynamics by offering wide coverage and the ability to survey otherwise inaccessible regions. Airborne based IR imagery permits characterizing mesoscale to sub-mesoscale, and small-scale temperature variability as well as investigating the physical processes at play. It permits monitoring the ice extent and coverage, as well as the ice and ocean temperature variability. It can also be used for derivation of surface flow field allowing investigation of turbulence and mixing at the ice–ocean interface. During test flights of IcePod in the Summer of 2013, multiple passes were made with both visible and infrared cameras of the inner, iceberg-filled region of Godthåbsfjord in Greenland. IcePod is a multi-instrument pod flown on an LC130 operated by the New York Air National Guard. In this study, we investigate the movement of sea ice and melt as well as the circulation, transport, and mixing in Godthåbsfjord at the calving front of Kangiata Nunata Sermia. Also in July–August of 2013, visible and IR imagery was taken during the Marginal Ice Zone Ocean and Ice Observations and Processes EXperiment (MIZOPEX) in the Beaufort Sea north of Oliktok Point AK. The instruments were flown in the unmanned airborne vehicle (UAV) ScanEagle. The IR imagery show distinct cooling of the skin sea surface temperature (SST) surrounding ice floes as well as an intricate circulation and mixing pattern that depends on the surface current, wind speed, and near-surface vertical temperature/salinity structure. We investigate the structure of circulation and mixing in the presence of ice floes and the effect of the ice melt on the skin SST.


On impacts of ocean waves in marginal ice zones and their repercussions for Arctic ice/ocean models

Vernon Squire

Corresponding author: Vernon Squire

Corresponding author e-mail: vernon.squire@otago.ac.nz

Associated with a gradual metamorphosis of summer Arctic sea ice – from a quasi-continuous ice sheet punctuated by pressure ridges and leads to a melange of ice floes resembling a MIZ, is an augmented presence of sizeable ocean waves that may have propagated into the pack ice from distant storms or have arisen within the MIZ itself due to the larger fetches that are now more common. If sufficiently forceful as they pass through the ice field, these waves can break up the ice floes to create a new floe size distribution (FSD), change local concentration by moving floes around, and supplement the melting that is occurring because of ice albedo feedback. In turn, the ocean waves themselves attenuate due to conservative scattering from the randomly sized, spatially disordered floes and cakes making up the MIZ that diffuse the waves and return energy to neighboring open water, and lose energy through several prospective dissipative processes. Consequently, the omission of ocean waves from ice/ocean models is unwise, as they can potentially alter atmosphere–ice–ocean coupling appreciably by affecting MIZ morphology so radically. In a series of three research projects, involving scientists from Norway, Canada, Australia and NZ, we have systematically investigated how ocean wave interactions with sea ice can be embedded in an ice/ocean model; first at high resolution in Fram Strait and later in other MIZ around the Arctic Basin. In each case it has been possible to track how the MIZ forms and, on the basis of its FSD or an abrupt change of concentration, how wide it becomes as a result of an inbound wave field provided by a spectral model such as WAM. Initially unidirectional seas were considered but more sophisticated 2-D scattering paradigms are now being developed that allow directionally defined seas to be modeled. Based upon the recognition that a MIZ can be delineated into a number of contiguous bands of ice floes, each with possibly different concentrations and randomized floes present in some FSD, the manner in which a long crested sea with its intrinsic directional spread advances through a conglomeration of dispersed multiply-scattering floes can be tracked, with the purpose of finding how the waves diminish in amplitude and whether the sea ice will be broken up. The details of how this is done are the subject of an AGU paper.


Sea ice and the biogeochemistry of ice-covered oceans

Martin Vancoppenolle

Corresponding author: Martin Vancoppenolle

Corresponding author e-mail: martin.vancoppenolle@locean-ipsl.upmc.fr

Will the future polar oceans be more biologically productive than today? Will the pumping of carbon through the dissolution of CO2 increase or decrease in the future near the Poles? How the ocean contributed to past climate variations? Scientists usually address this question by simplifying the role of sea ice as much as they can. Is this right to do so? Maybe not. A clear observational fact – which has emerged over the last two decades – is the existence in the sea-ice zone of specific biological and chemical processes at work, of previously unknown pathways for atmosphere–ocean gas exchanges, and of potential biophysico-chemical interactions. Unravelling and upscaling those observed processes is the focus of recent research, including field, lab, remote sensing and modeling studies. As the first process parameterizations could be part of the next generation of Earth System Models, we should soon be able to further elaborate the future of the polar oceans.


On thin ice: sea ice – ecosystem linkages in a changing ocean

Hauke Flores

Corresponding author: Hauke Flores

Corresponding author e-mail: hauke.flores@awi.de

In both hemispheres, polar sea ice environments are changing at dramatic speed. Because both the Arctic and the Antarctic Oceans host a large variety of sea ice-associated life forms, these drastic changes of the physical environment have already significantly affected Polar ecosystems, and are predicted to cause more severe ramifications in the future. A key aim of this talk is to raise awareness of the need to better understand and protect the endangered beauty of sea ice-associated biodiversity. I will investigate several examples of critical relationships between the physical structure of sea-ice habitats and sea ice-associated organisms and communities in both polar regions. The crucial importance of inter-disciplinary research for understanding these relationships and predicting future developments will be highlighted. Concepts of current and future inter-disciplinary sea ice research will be discussed.


Sea ice in the cryosphere: trends, observations, and challenges

Ron Kwok

Corresponding author: Ron Kwok

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

I will summarize the trends in Arctic and Antarctic sea ice cover reported in the Fifth Assessment Report (AR5) of the IPCC. Along with the persistent negative trends in Arctic ice extent since AR4, the current assessement also highlights for the first time our confidence in the decline in Arctic ice thickness over the last few decades based on observations from a variety of sources: moorings, submarine, EM probes, and satellite altimetry. For Antarctic sea ice, although attention is typically directed at the trends in total ice extent, it is the large regional trends that are remarkable although longterm/large-scale observations of crucial parameters for understanding these trends are still quite limited. For both sea-ice covers, I list observations that are currently available and what we might expect in the future; emphasis will be on snow depth, ice thickness, and ice kinematics. There are recent advances in the large-scale observations of these parameters but challenges remain especially in the derivation/validation of large-scale fields of snow thickness that are important not only for understanding surface heat exchanges but also for the calculation of snow loading in the retrieval of sea ice thickness. Arctic and Antarctic sea-ice trends as obtained from IPCC models do usually not agree with observed trends. Hence, the models must be wrong. This reasoning is surprisingly widespread both in the scientific literature and in the general discussion of the quality of IPCC model simulations. In this presentation, I will outline why this reasoning is wrong, and why we can often learn surprisingly little about model quality through a direct comparison of model simulations and observations. The presentation will also discuss what kind of observations would be most useful for the identification of shortcomings in sea-ice models, and why in many areas of sea-ice research a closer cooperation of modelers and observationalists is highly desirable.


The interdependency of sea ice and ice shelves

Patricia J. Langhorne

Corresponding author: Pat Langhorne

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

Global climate models are not able to reproduce both the trend and the mean of Antarctic sea ice extent. Further there are large variations between model predictions of sea ice coverage, some of which arise because of differences in representation of the relative importance of heat from the atmosphere versus heat from the ocean. It is very well known that Antarctic coastal waters are conditioned by interaction with ice shelves, but the pervasiveness of the influence of this water on sea-ice formation is poorly documented. Not only does proximity to an ice shelf alter sea-ice growth, there is mechanical coupling between an ice shelf and the sea ice at its edge that probably buffers the shelf from the ravages of the ocean. Here we focus on observations of the physical processes involved in the interaction between ice shelf-conditioned waters and sea ice formation, with the ultimate aim of guiding model development. Close to an ice shelf, sea ice often grows in water that has been supercooled by interacting with the ice shelf at depth. An important consequence is that the sea ice loses heat to the ocean as well as to the atmosphere. This ‘negative’ ocean heat flux causes the sea ice to grow thicker than it would without the ice shelf. The thermal deficit also means there are tiny frazil crystals in the water column. While they sometimes accumulate and grow on any object suspended in the nearsurface ocean, they also accumulate under the sea ice where they form a porous layer of crystals in an evolving state of consolidation. This consolidation results in a modified crystallographic structure that leaves a detectable signature frozen into the sea ice cover through the formation of incorporated platelet ice. Here we use new results to extend some of the longest available Antarctic ice ocean observational records that were initiated a century ago near the combined Ross and McMurdo ice shelves in southern McMurdo Sound. Over that time there has been no discernable change in the temperature of the upper ocean. This surface water is held just below its freezing point as it enters the Sound, probably because of the regulating influence of basal melting deep in the ice shelf cavity. We use the ability of sea ice to integrate the effect of ocean heat flux over its annual growth to interpret crystallographic records from a historical time series of sea ice cores. This information is supplemented by airborne and satellite remote sensing of sea ice thickness and modeling at a range of scales. The distribution of platelet ice (and hence negative ocean heat flux) appears to be strongly linked to the circulation in McMurdo Sound although its abundance varies from year to year. These multiple sources of data, contextualized within the relatively long time series, are extended around coastal Antarctica to provide estimates of the influence of ice shelf-conditioned surface water on sea ice. These data may also enable indices of the ice shelf ‘health’ to be constructed.


Seasonal evolution of sea ice motion in the Beaufort Sea and icepack response to atmospheric forcing

D.G. Babb, J.V. Lukovich , G. McCullough, K. Hochheim, R. Scharien, D.G. Barber

Corresponding author: D.G. Babb

Corresponding author e-mail: umbabb@cc.umanitoba.ca

The arctic ice pack within the Beaufort Sea undergoes a transition from a consolidated ice state in late winter (March/April) to a less extensive, weaker ice pack in summer. Using an array of autonomous instruments deployed in early April 2012 along the periphery of the multiyear ice pack in the Beaufort Sea, we highlight the seasonal increase in ice drift speeds. Hourly in situ observations of ice drift, surface winds, air temperature and ice mass balance were collected between April and July 2012. Monthly mean ice drift speeds increased during all four months from 6.56 cm s–1 in April to 19.36 cm s–1 in July. It is shown that increased ice drift speeds are not the result of increased wind speeds, as monthly mean wind speeds remained around 4 m s–1 throughout the study with equal wind speed distributions during April and July, and May and June. Given that winds did not increase during the study we attribute the change in ice drift speeds to the weakening and gradual breakup of the Beaufort ice pack, which allowed ice to enter a state of free drift and become more responsive to atmospheric forcing. We show that both the turning angle and the scaling factor between ice and wind speeds increased during each month of our study. A monthly mean turning angle of –0.8° in April indicates a compact ice pack with great internal stresses which limit ice drift, comparatively a turning angle of 31.4° in July indicates an icepack in free drift that responds freely to the Coriolis force. As the ice pack deteriorated so did the coherent motion between ice floes; in April ice drift was correlated across scales of 100 km, while in July motion showed weak and even negative correlations across much shorter length scales. Using monthly meander coefficients we show the seasonal decrease in linear ice drift as inertial loops and random walks became present within ice floe trajectories during June and July. Using in situ ice mass balance observations, we highlight the declining strength of the ice pack through spring as the thick multiyear ice became isothermal around July 12. Radarsat-2 ScanSAR imagery is used to quantify the seasonal evolution of open water coverage and floe size distributions within a 100 km box around the drifting study sites, which are then compared to ice drift speeds through the transition from the winter to the summertime ice pack regime.