Thermal properties of subglacial meltwater plumes at outlet glaciers in Greenland

S. David Comer, Ginny A. Catania, Timothy Bartholomaus, Mason J. Fried

Corresponding author: S. David Comer

Corresponding author e-mail: sdavidcomer@utexas.edu

Meltwater plumes are formed by the upwelling of fresh subglacial discharge at the grounding lines of marine-terminating glaciers. These plumes entrain deep, warm, Atlantic water in glacier fjords, which can melt the terminus face and, in turn, promote terminus undercutting and calving. Despite the importance of meltwater plumes, little is known about their time-varying characteristics (extent and temperature) and few observational studies link them quantitatively to variable subglacial discharge. In this study we use thermal images generated from Landsat 8 thermal bands to determine the temperature of sediment-laden melt plumes occurring at glacier termini in central West Greenland. Calibration is accomplished by assuming that the summer ice sheet is at the melting temperature and through comparison with other sea surface temperature data from ship-based and remote-sensing (e.g. MODIS) observations. We find that meltwater plumes express at the surface as warm temperature anomalies, and hypothesize that they have entrained significant Atlantic water and are thus able to melt glacier termini. Plume temperatures vary seasonally, and are warmest when large discharges of meltwater can be expected. We suggest that remotely sensed meltwater plume temperatures may better constrain the importance of submarine melt relative to other frontal ablation processes acting at Greenlandic tidewater glaciers.


Monitoring the tabular icebergs C28A and C28B calved from the Mertz Ice Tongue using continuous satellite remote sensing data

Tian Li, Xiao Cheng, Yan Liu, Fang Wang, Fengming Hui

Corresponding author: Xiao Cheng

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

Icebergs play an important role in transferring fresh water to, and exchanging heat with the ocean. Their drifting can modify ocean circulation and regional biodiversity. The dynamic mechanisms of the iceberg drifting process have been extensively studied using remote sensing data but their mass loss due to calving and basal melt have not been fully investigated using radar remote sensing. This study analyzes the process evolution of two tabular icebergs (C28A and C28B), calved from iceberg C28 on 2 April 2010. The monitoring of the icebergs started from the calving of C28 from the Mertz Ice Tongue in February 2010 and ended in April 2012 using continuously multisource radar remotely sensed data. The evolution of the iceberg area and perimeter are monitored using ENVISAT ASAR images and the iceberg thickness is derived from CryoSat-2 data. The mass loss due to calving and basal melting of the icebergs is then estimated. We show that the area and thickness of the icebergs continuously decrease while drifting westwards under the effects of oceanic and atmospheric forces. The mass loss due to iceberg calving is about 131.34 Gt during the study period (greater in winter than in summer), with higher loss when the iceberg is floating in open water than when in the stranded state. The basal melting plays a more important role than calving in the iceberg mass loss, which is also found to be greater in winter than in summer. Submarine topography affects the different drifting routes of the icebergs. Our results reveal the complex interaction process between icebergs and their surrounding environment.


Distinguishing the signatures of ice shelf surface roughness, basal roughness, temperature and chemistry in radar sounding data

Dustin Schroeder, Cyril Grima, Mark Haynes, Jamin Greenbaum

Corresponding author: Dustin Schroeder

Corresponding author e-mail: dschroed@gmail.com

Radar sounding observations of ice shelves have long been used to make inferences about their basal properties. For example, stronger bed echoes have been interpreted as a signature of elevated basal melting due to the advection of colder ice and a smoother, more specular, reflecting interface. However, this type of inference can be problematic because bed echo strengths are controlled by a combination of surface and basal roughness as well as englacial temperature and chemistry. Therefore, placing quantitative constraints on either temperature or roughness at a scale relevant to ice or ocean modeling will require the disentangling of these effects. To address this challenge, we present a new analysis approach for radar sounding data that can constrain and correct scattering losses from both two-way propagation and one-way reflection for the ice–air and ice–water interfaces. We demonstrate and validate this approach with a two-layer full-wave raw-data simulator. This enables the estimation of ice-shelf basal roughness at the meter to centimeter scale using a range of airborne radar sounder systems and data sets.


Warm water drives basal melt of the Totten Glacier

Alessandro Silvano, Stephen Rintoul, Beatriz Peña-Molino, Guy Williams

Corresponding author: Alessandro Silvano

Corresponding author e-mail: Alessandro.Silvano@utas.edu.au

Antarctic ice loss has been associated with thinning of ice-shelves forced by an increased oceanic heat supply, especially in West Antarctica. Totten Glacier drains an ice volume equivalent to ~4 m of global sea-level rise and is showing the highest thinning rate in East Antarctica. However the lack of oceanographic data close to the glacier has prevented an assessment of the role of the ocean in the observed thinning. Here we show the results from a recent summer survey on the continental shelf near the Totten Glacier. Warm modified circumpolar deep water (mCDW) is widespread on the shelf below ~500 m depth, with temperatures ∼ 2.5°C above the in-situ freezing point. Fresh (salinity ∼34.3) winter water overlies the mCDW. Moored temperature records confirm that the warm water at depth is present year-round. The mCDW reaches the Totten ice shelf cavity through a deep (>1100 m) channel at the ice front, with maximum temperatures up to 2°C above the in-situ freezing point. A second deep channel provides access for somewhat cooler water to the cavity beneath the Moscow University Ice Shelf. Low oxygen and salinity in the Winter Water at the Totten and Moscow University calving front indicates outflows of a mixture of glacial melt and mCDW, with meltwater concentrations up to 4-5 ml/l. Satellite data show that Totten and Moscow University ice-shelves share the highest basal melt rate in East Antarctica. Our results indicate that this is due to intrusions of warm water into their cavities. The oceanographic conditions at the Totten (fresh Winter Water overlying warm mCDW, absence of high-salinity shelf water and substantial input of meltwater) are more similar to the rapidly thinning glaciers of West Antarctica than to the cold cavity ice-shelves typical of most of East Antarctica.


Melting tidewater glaciers create subsurface acoustic waveguides

Oskar Glowacki, Mateusz Moskalik, Grant B. Deane

Corresponding author: Oskar Glowacki

Corresponding author e-mail: oglowacki@igf.edu.pl

The use of naturally generated underwater sounds is showing promise as an efficient tool for the quantitative analysis of ice–ocean interactions in glacierized bays. However, the correct interpretation of long-term noise recordings requires deep insight into the propagation environment, and its seasonal variations. For that purpose, over 130 CTD casts and several simultaneous acoustic recordings on a vertical hydrophone array were taken in Hornsund Fjord, Spitsbergen between May and November 2015. The application of a ray-based propagation model, together with spectral analysis of the recorded noise, demonstrates the creation of high-gradient, upward-refracting sound–speed profiles during the melt season. Both approaches clearly identify the surface duct as a dominant feature of the overall waveguide, creating a high concentration of acoustic energy in the subsurface layer covered by glacially modified waters. Consequently, inter-seasonal acoustic monitoring of the glacial melt should be performed in the deepest part of the bay, away from the duct, to minimize its effects on the noise signature. Moreover, the distribution of melting ice along the study site has significant impact on the received noise levels, emphasizing the need for directional capability in noise measurements, or other supplementary surveys. Surface ducts in Hornsund effectively trap acoustic energy above 1 kHz, but thicker waveguides expected in other regions most likely affect also much lower frequencies related to, for example, calving events. Therefore, their seasonal evolution must be always taken into account when acoustic research is focused on processes acting at ice–ocean boundaries. The work was founded by the Polish National Science Centre grant no. UMO-2013/11/N/ST10/01729, partially financed from the funds of the Leading National Research Centre (KNOW) received by the Centre for Polar Studies for the period 2014-2018, and supported within statutory activities No 3841/E-41/S/2015 of the Ministry of Science and Higher Education of Poland.


Some aspects of Antarctic and Greenland ice-sheets dynamic and their interactions with the ocean

Alexey N. Markov, Pavel G. Тalalay, Youhong Sun, Danian Huang

Corresponding author: Alexey N. Markov

Corresponding author e-mail: am100@inbox.ru

The generally accepted view of the dynamics of the Antarctic and Greenland ice sheets is based on geodetic measurements of the moving surface. However, on the basis of the results of long-term inclinometry boreholes monitoring at Vostok station (to a depth of 1920 m) and at the Vostok–Vostok-1–Pionerskaya–Mirny 1409 km long traverse (to a depth of about 450 m), and the radar profiling results it was revealed that:
– the ice flow of the Antarctic ice sheet has a layered velocity change and a ‘fan-shaped’ change of flow direction with depth; this change is the result of the cyclical paleoclimatic formation conditions of layers and differentiation of the modern subglacial topography;
– the plastic upper firn layer has individual dynamics parameters and in actual fact ‘slides’ with respect to the underlying more monolithic ‘body’ of the ice sheet;
– the flow rate of lower layers in the ‘body’ of the ice sheet is faster than the rate in the upper layers.
Recently folded structures identical to diaper folds and diapers were registered on the radar images in the lower third part of the Antarctic and Greenland ice sheets. This allows us to conclude that the turbulent flow of ice can take place at the near-bottommost plastic strata of ice sheets. As a result of terrestrial and satellite geodetic observations it was revealed that the ice in the Antarctic ice sheet generally flows outward from the central region to the peripheral but is differentiated into local streams, which are similar to the glaciers. These streams have individual characteristics and flow parameters associated with different ice-catchment basins and various glacier flows. These results suggest that the flow difference between the masses of the ice sheet ‘body’ and the ice sheet surface must be taken into account for reliable forecasting of mass balance and removal of ice masses from coastal areas to the ocean. It was revealed that the modern fan-shaped changing of the direction of the ice flow with depth may be connected with changes in the location of the dominant ice dome and the surface slope in the past. It must be assumed that the dominant ice dome location on the surface of the ice sheet depends on the distribution of moisture in the atmosphere, its circulation over the surface of the ice sheet, andthe geometry (geography) of the moisture sources, e.g. peculiar properties of the open water in adjacent seas.


Influences of platelet and marine ice accumulations on ice–ocean interaction

Natalie Robinson, Craig Stevens

Corresponding author: Natalie Robinson

Corresponding author e-mail: natalie.robinson@niwa.co.nz

Basal melting of ice shelves produces ice crystals that can accumulate in thick layers at the ice-shelf–ocean interface and beneath adjacent sea ice. Observations from the boundary layer beneath ~1.8 m thick platelets in McMurdo Sound demonstrate that velocity in the upper ocean was modified to a depth of ~30 m by the presence of the platelet layer. Drag coefficients were calculated in the range 0.04 < CD < 0.09 – an effect significantly greater than can be explained by the thickness of the platelet layer or its individual crystals. Using these data, we have identified four independent modes by which the platelet layer is able to extract energy from the upper ocean. These modes contribute to an overall ‘effective roughness’ an order of magnitude greater than is currently in use in ice-shelf–ocean models.


Tidal modulation, stagnation and stick–slip in a viscoelastic ice stream model with frictional sliding

Bradley Lipovsky, Eric Dunham

Corresponding author: Bradley Lipovsky

Corresponding author e-mail: brad_lipovsky@fas.harvard.edu

The Antarctic ice streams are the largest ice flux from the Antarctic Ice Sheet. Despite this pivotal role, our understanding of their dynamics remains incomplete. Ice streams experience velocity changes at timescales that range from hours to decades. At the timescale of hours to days ice streams experience velocity changes that are tidally correlated. Two end-member behaviors are observed. In the first, velocity changes are smooth and result in ice surface velocity time series with a sinusoidal appearance. The other end-member behavior, observed at the Whillans Ice Plain (WIP), is characterized by episodic slip events. During a slip event sliding accelerates for 200 s to rates as large as 65 m/d; deceleration then occurs for 1000 s. Over the course of decades, ice streams have long-term velocity variations that may be large enough to result in complete stagnation. In order to explore this broad range of behavior, we conduct numerical simulations of the ice stream motion that arises from a rate- and state-dependent frictional basal sliding law. Our simulations include the effects of tidal forcing, inertia, upstream loading and ice viscoelasticity in a cross-stream, thickness-averaged formulation. Our numerical implementation makes use of higher order accurate, summation-by-parts finite difference operators and weak enforcement of boundary conditions. At the diurnal timescale our model exhibits a spectrum of behavior ranging from smoothly varying motion at high pore pressure to stick–slip type motion at lower pore pressure. Our model recreates many features observed in GPS observations of WIP slip events. The transition to steady sliding occurs when the elastic stress change due to sliding is outpaced by the strength change due to frictional weakening. At the decadal timescale we test two hypotheses: that WIP stagnation is driven by 1) local changes in pore pressure or 2) changes in upstream driving forces. We first calibrate model parameters to match GPS records from the WIP. We then compare the effects of reduced loading rate to that of reduced of pore pressure. We find that a sufficiently reduced loading rate results in steady sliding. Elevated effective pressure, in contrast, does not have this effect. Whether WIP stick–slip continues in the future will therefore be indicative of the mechanism of stagnation.


Modeling granular dynamics of ice melange

Chin-Chang Kuo, Jason M. Amundson, Justin C. Burton, Michael Dennin

Corresponding author: Michael Dennin

Corresponding author e-mail: mdennin@uci.edu

Ice melange, a large-scale, quasi-two-dimensional granular material composed of icebergs and sea ice, has been shown to affect glacier stability and therefore dynamics. One of the dominant features of this highly complex system is the interaction of iceberg–iceberg and iceberg–glacier collisions. We report on results from laboratory-scale modeling of ice melange with plastic particles. Plastic particles of various shapes are placed on the water surface in a meter-long channel. A moving barrier at one end of the channel is used to apply a constant velocity to generate flow of the particles. Our initial studies looked at the impact of particle shape on the flow behavior of the ice melange. We find very distinct behavior, with disk-like particles creating a very smooth flow and rectangular particles having a tendency to jam. To understand the dynamics of the particle raft, the velocity fields and the force required to generate the flow have been measured simultaneously. In addition, we will report on initial comparisons of our experimental results to theoretical simulations and in situ observations of ice melange. Supported by NSF Grant DMR-1506991.


The supercool-o-meter: description & initial results

Matt Walkington, Craig Stevens, Natalie Robinson, Brett Grant, Inga Smith

Corresponding author: Matt Walkington

Corresponding author e-mail: natalie.robinson@niwa.co.nz

The measurability of supercooling is controversial, as ice inside a conductivity sensor may affect the observations, leading to uncertainty. We have modified a SeaBird CTD to include a separate heating chamber that removes self-freezing of the conductivity cell in order to accurately document the presence of supercooled water. McMurdo Sound is a perfect location to look at this problem, not only because of accessibility but also because significant supplies of ice shelf water are consistently delivered there from beneath the nearby Ross Ice Shelf. Along this route, the change in pressure, dictated by the depth of the ice shelf base, is relatively rapid, resulting in supercooled water greater in both degree and volume than has been documented elsewhere. We outline the development to date of the supercool-o-meter and highlight results from two recent deployments.


Implementing a calving parametrization into a flow-line glacier model

Beatriz Recinos Rivas, Ben Marzeion

Corresponding author: Ben Marzeion

Corresponding author e-mail: ben.marzeion@uni-bremen.de

Glacier mass loss is a major cause of sea-level rise. During the 20th century, glacier mass loss and ocean thermal expansion caused about 75% of the observed global mean sea level rise. Modeling glaciers in the global scale remains challenging. Only six global-scale models have been reported in the last 10 years, and only one of these accounts for frontal ablation of marine-terminating glaciers. However, nearly 40% of the global glacier area drains directly to the ocean, and frontal ablation is likely to present a significant part of the mass balance of many tidewater glaciers. Here we present first steps towards implementing a simple calving parametrization into the Open Global Glacier Model (OGGM). OGGM is built upon Marzeion and others (2012) and is currently in intensive development. The model accurately represents glacier geometry and explicitly models ice dynamics. It is designed to reconstruct and project mass-balance, volume and geometry of any glacier contained in the Randolph Glacier Inventory (RGI 5.0). Ice dynamics in OGGM are represented by a depth integrated, multi-branched flow-line model which facilitates the representation of solid ice discharge. Frontal ablation is computed as a function of the height, width and estimated water depth of the calving front. We will show results of the application of the model to well studied tidewater glaciers (e.g. Columbia Glacier) as a starting point for accounting for calving processes in a global scale model.


Understanding interactions between cryosphere elements: a WCRP CliC (Climate and Cryosphere) Initiative

Rob Massom

Corresponding author: Rob Massom

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

Major change is occurring across both polar cryospheres, yet little is known about the inter-relationships between the different component ‘elements’ involved – including ice sheets/glaciers, icebergs and sea ice (connected by oceanic and atmospheric processes). How does change in one element affect others, what processes are involved, and what are the consequences? Recent studies suggest that these linkages may be complex and involve subtle and previously unconsidered feedbacks. Examples include linkages between sea ice and Antarctic ice-shelf basal melt, and landfast sea ice and ice-sheet margin stability around Antarctica. The need for better information on, and understanding of, cross-cryosphere linkages was underlined by the recent SCAR Antarctic Horizon Scan initiative, and represents an exciting challenge. The aim of this CliC Targeted Activity is to raise awareness and understanding of key cross-cryosphere interactions that are currently marginalized/neglected (i.e., in models); foster and facilitate cross-disciplinary and –polar discussion and collaboration; synthesize and integrate existing work and information, to provide new insights and identify crucial gaps; and (in so doing) bridge/connect the polar and climate observational and modeling communities. The ultimate goal is to provide new information towards more realistic parameterization of cross-cryosphere components, linkages and processes in models (Earth System, coupled regional high-resolution and ice sheet/shelf etc.). Initial planned outcomes include a series of community-based review papers e.g., on sea ice-ice shelf interactions.


Monitoring the front dynamics and ice flowing of Dalk Glacier using the Chinese Gaofen-2 satellite and the Polar Hawk-1 unmanned aerial vehicle

Tiancheng Zhao, Sihan Luo, Xiao Cheng, Fengming Hui, Teng Li, Tian Li

Corresponding author: Tiancheng Zhao

Corresponding author e-mail: zhaotc100@gmail.com

Glacier dynamics play an important role in the mass balance of Antarctica. In recent years, scientists have gradually started to use very high resolution (VHR) satellite image data and unmanned aerial vehicle (UAV) technology to investigate glaciers. Gaofen-2 is a Chinese VHR satellite launched on 19 August 2014 that can provide panchromatic images with 1 m resolution and multispectral images with 4 m resolution in four visible–infrared bands. The Polar Hawk-1 is a type of small UAV developed by Beijing Normal University with a 3 m wingspan and a 64 cc gasoline engine. It has a maximum flight of about 180 km in 90 min and a cruising speed of 110 km h–1. In this research, images of Dalk Glacier near Lassemann Hills, East Antarctica, were obtained on two dates, one GF-2 image on 2 November 2014 and UAV images on 2 February 2015. The ice velocity and dynamic changes of the glacier front were derived from these images. The results show that, during 92 days, the front of Dalk Glacier advanced about 35–40 m, and the average ice velocity of this region is about 0.5 m d–1. This study demonstrates that the latest remote sensing observation technology at meter and sub-meter levels can effectively support observation and research on subtle changes in polar glaciers.


Discrete particle simulations of ice melange dynamics

Justin Burton, Jason Amundson, Michael Dennin, Chin Chang Kuo

Corresponding author: Justin Burton

Corresponding author e-mail: justin.c.burton@emory.edu

In tidewater glacial fjords, the open water in front of the glacier terminus is constantly filled with a collection of broken pieces of calved icebergs and sea ice. For glaciers with large calving rates, this ‘melange’ of ice can be jam-packed, so that the flow of the ice is mostly determined by granular interactions, in addition to flow in the underlying fjord waters. As the glacier pushes the ice melange through the fjord, the melange may potentially influence calving rates if the back-stress applied to the glacier terminus is large enough. However, the stress applied by a granular ice melange will depend on its rheology, i.e. iceberg–iceberg contact forces, geometry, friction, etc. Here we report two-dimensional, discrete particle simulations to model the granular mechanics of ice melange. A polydisperse collection of particles is packed into a long channel and pushed at a constant speed. Each individual particle experiences viscoelastic contact forces upon collision with another particle or the walls of the channel, as well as tangential frictional forces. We find that the two most important factors that govern the total force applied to the glacier are the geometry of the channel and the shape of the particles. Using results from the model, the potential magnitude of the back-stress applied to a real glacier terminus will be discussed.


Seasonality of glacier-driven water mass transformation in Greenland’s fjords

Nicholas Beaird, Fiamma Straneo

Corresponding author: Nicholas Beaird

Corresponding author e-mail: nbeaird@whoi.edu

Greenland’s fjords form a link between the open ocean and its reservoir of heat, and the Greenland Ice Sheet and its reservoir of fresh water. The interaction between these two systems has significant consequences for global climate. Glacier-driven water mass transformation in proglacial fjords reflects these important ocean–glacier interactions. We report here on the character and seasonal variation of that water mass transformation from 6 years of mooring and shipboard hydrographic data. Results include a description of the heat, salt and nutrient upwelling induced by glacial buoyancy forcing; seasonal evolution of water mass transformation which extends beyond the melt season; and seasonality in fjord–shelf connectivity as seen from water property convergence and divergence.


Variations in subglacial discharge during tidewater glacier retreat

Jason Amundson

Corresponding author: Jason Amundson

Corresponding author e-mail: jason.amundson@uas.alaska.edu

Subglacial discharge from tidewater glaciers turbulently mixes with warm fjord waters and forms a turbulent upwelling plume that melts glacier termini. Several recent studies suggest that submarine melting can destabilize a glacier terminus and that the melt rate is nonlinearly and strongly dependent on subglacial discharge. For land-terminating glaciers, net annual discharge increases during glacier retreat by about 25% due to climate-driven reductions in glacier storage. The situation is different for tidewater glaciers, which are capable of experiencing rapid changes in glacier dynamics and geometry in the absence of climate change. Using simple theoretical arguments and an ice flow model I show that, in a steady climate, subglacial discharge from tidewater glaciers decreases whenever the rate of retreat exceeds the mean rate of thinning by about a factor of 100 (i.e. when mass losses due to calving are increasing more quickly than the net surface mass balance is decreasing). This reduction in subglacial discharge, which occurs when a glacier retreats into deep water, represents a potentially negative feedback loop within the glacier–ocean system and has implications for fjord ecosystems.


How Ross Sea dynamics are controlled by competing seasonal processes

Stefan Jendersie, Mike Williams, Robin Robertson, Patricia Langhorne, Natalie Robinson

Corresponding author: Stefan Jendersie

Corresponding author e-mail: natalie.robinson@niwa.co.nz

An application of the Regional Ocean Modeling System (ROMS) to the shallow Ross Sea continental shelf has identified a system of three anti-cyclonic and one cyclonic circulation cells that facilitates the water mass transports in the interior, including the ice shelf cavity. Constrained by the banks and depressions, t Sv seasonally – are spatially persistent but experience different individual temporal changes. The main control of their dynamics are the horizontal differences in density that drive three mechanisms: baroclinic pressure gradients, gravity driven bottom flows and barotropic pressure gradients through sea surface height gradients. Circumpolar deep water (CDW), sourced from the Antarctic slope current (ASC), sits next to northward-flowing high salinity shelf water (HSSW) and other dense shelf water. CDW resupply events seem to be triggered by a zonal shift of the ASC in the order of ~10 km that occurs at different times along the shelf break. The second, atmospherically driven mechanism to strengthen density gradients is HSSW production through intense winter sea ice formation in the polynyas of the southwestern Ross Sea. The third mechanism to enhance local horizontal differences in density is ice shelf water (ISW) supplied by melting at the ice shelf base. The timing of ISW occurrence is synchronized not with the atmosphere but with the seasonality of warm water inflow to the cavity, which in turn is controlled by a number of other processes. The model predicts phase lags of up to 1.5 years between heat import events to the cavity and the subsequent ISW pulse leaving the cavity. Thus the seasonality of flow dynamics in the Ross Sea is a superposition of the ASC variability, the atmospheric cycle and the heat import signal to the cavity.


Glaciological settings and recent mass balance of the Blåskimen Ice Rise in Dronning Maud Land, Antarctica

Vikram Goel, Kenichi Matsuoka, Joel Brown, Anja Eichler, Elisabeth Isaksson, Tõnu Martma, Margit Schwikowski, Carmen P. Vega

Corresponding author: Vikram Goel

Corresponding author e-mail: vikram.goel@npolar.no

Ice rises are grounded ice bodies (at least partially) surrounded by the floating ice shelf. They provide buttressing to the neighboring ice shelves and hence play a key role in controlling dynamics and mass balance of the Antarctic Ice Sheet. In turn, evolution of an ice rise is affected by the surrounding coastal environment. The Dronning Maud Land (DML) coast, East Antarctica, is a 2000-km-long interconnected system of ice shelves, outlet glaciers and ice rises, but only a few ice rises have been investigated. Here we report field measurements of the Blåskimen Ice Rise, adjacent to the Fimbul Ice Shelf, fed by Jutulstraumen, one of the largest outlet glaciers in DML. We find that the ice rise is rather dome-shaped, with a pronounced ridge from its summit (404 m a.s.l.) to the southwest. It is grounded on a mostly flat bed below the current sea level (mean elevation: 109 m b.s.l.). Ice flows very slowly (0–2 m a–1) near the summit and accelerates to 8–15 m a–1 within 7–14 km of the local grounding line. Shallow radar-detected isochrones dated by a firn core show that the surface mass balance is higher on the upwind slope than the downwind slope, and this pattern has persisted for at least the past decade. We estimated the mass balance of the ice rise for a plausible range of poorly constrained parameters using these field data. The ensemble of results shows that, over the past decade, the ice rise has been steady-state or has thickened by 0.1 m a–1 or less.


Granular dynamics of iceberg melange

Alexander Robel

Corresponding author: Alexander Robel

Corresponding author e-mail: robel@caltech.edu

Observations from Greenland suggest that the rate of calving-related mass loss from some outlet glaciers is modulated by the presence of sea ice within iceberg melange on seasonal and longer timescales. Despite these and other indications that melange state may play an important role in the long-term evolution of the Greenland Ice Sheet, the dynamics of melange and its interaction with the glacier are still not well understood. In this study, a discrete element model is configured to resemble a region near the terminus of Jakobshavn Isbræ, and is then used to perform some of the first successful numerical simulations of iceberg melange behavior. In these simulations, calving events initiate rapid jamming waves within the melange, and can modulate the propensity for further calving events through back stress at the terminus on timescales of days. Sea-ice-like bonds between icebergs are shown to strengthen melange and more effectively transmit stresses to the glacier terminus. Sea ice can also suppress the post-calving expansion of melange by absorbing kinetic energy in elastic flexure. These simulations reliably reproduce many aspects of both observations and laboratory experiments, and can potentially provide explanations for other vexing aspects of melange behavior.


Large freshwater fluxes from melting ice melange in Greenland glacial fjords

Ellyn Enderlin, Gordon Hamilton, Fiamma Straneo

Corresponding author: Ellyn Enderlin

Corresponding author e-mail: ellyn.enderlin@gmail.com

Iceberg discharge from Greenland’s marine-terminating glaciers has increased over the last two decades and comprises a significant fraction of the ice sheet’s contribution to global sea level. Melting icebergs also contribute cold, buoyant fresh water to glacial fjords, which may influence circulation processes, which in turn will affect the delivery of heat to marine glacier margins. Many of Greenland’s glacial fjords contain an expansive ice melange (a floating matrix of icebergs, bergy bits and sea ice) supplying an unknown amount of fresh water to the fjord systems. Here we use repeat very-high-resolution digital elevation models (DEMs) and Landsat8 images to estimate freshwater fluxes from melange melting in two of Greenland’s largest fjords: Ilulissat Icefjord in West Greenland and Sermilik Fjords Southeast Greenland, where Jakobshavn Isbræ and Helheim Glacier terminate, respectively. Each DEM provides the ice volume above and, by inference using hydrostatic equilibrium, below the water line. Iceberg melt rates are estimated by tracking changes in volume through time for a subset of icebergs in the melange. The submerged area corresponding to the observed ice melange volume is inferred from the iceberg size distribution in the DEM and the melange extent in near-contemporaneous Landsat imagery, then the freshwater flux for the melange is estimated as the product of the iceberg melt rate and submerged melange area. Estimated fluxes range between 549 and 1887 m3 s–1 and between 337 and 707 m3 s–1 for Ilulissat and Sermilik fjords respectively, which we primarily attribute to temporal variations in submarine melt rates for large deep-keeled icebergs. We find that freshwater fluxes from ice melange may exceed surface meltwater runoff from the entire glacier catchment for the majority of the year. Further, we find that the submerged surface area of the ice melange is greater by an order of magnitude than that of glacier terminus in both fjords. These results indicate that freshwater fluxes from iceberg melt in the melange need to be considered in studies of fjord circulation and in estimates of fjord freshwater flux.


Modeling the influence of melt and buttressing on the recent speedup of Smith Glacier

David Lilien, Ian Joughin, Ben Smith

Corresponding author: David Lilien

Corresponding author e-mail: dlilien90@gmail.com

Smith Glacier and the corresponding Crosson and Dotson Ice Shelves constitute the third largest system on the Amundsen Sea Coast. In the past two decades, Crosson Ice Shelf has doubled in speed and this acceleration has propagated more than 50 km upstream into Smith Glacier’s grounded ice. Acceleration has been accompanied by a breakup of Crosson’s shear margins and a >35 km retreat of the grounding line. Although this ice feeds Dotson Ice Shelf as well, slight deceleration has been observed over Dotson. We use a full-Stokes, finite-element model of this system to investigate the effects of differing melt forcing and lateral buttressing on the evolution of the system over the past two decades. Using inverse methods, we initialize a prognostic model mimicking late 1990s conditions and force the model with estimated melt rates and surface mass balance. We compare results to the subsequent velocity measurements to compare the effect of different processes on the acceleration and stability of grounded and floating ice.


Seasonal changes in basal conditions at Helheim and Kangerdlugssuaq Glaciers, southeast Greenland, from 2011–15

Laura Kehrl, Ian Joughin

Corresponding author: Laura Kehrl

Corresponding author e-mail: kehrl@uw.edu

Marine-terminating outlet glaciers are very sensitive to changes at the ice–ocean boundary, which alter the ice-front position and thereby the stress balance. When a glacier retreats into deeper water, more of the driving stress must be balanced by longitudinal stress gradients and/or basal shear stress, so the glacier speeds up. This study uses inverse methods implemented in Elmer/Ice to better understand the seasonal and interannual evolution of the stress balance (basal shear stress, driving stress, and longitudinal/lateral stress gradients) at Kangerdlugssuaq and Helheim glaciers, southeast Greenland, from 2011–15. We use satellite observations to constrain the diagnostic inverse problem and infer basal shear stress roughly four to six times per year. Over this time period, Helheim and Kangerdlugssuaq glaciers exhibited very different seasonal behavior. At Kangerdlugssuaq, the ice-front position varied seasonally by more than 3 km, with glacier advance commencing in December/January and glacier retreat commencing in July/August. The glacier sped up by ~20% during the seasonal retreat and by ~5% during the summer melt season. At Helheim, iceberg calving events, glacier retreat and speedup occurred at all times of the year. Yet, despite these seasonal differences, both glaciers experienced seasonal variations in surface elevation of ~20 m, with glacier thinning starting in late summer and glacier thickening starting in midwinter. We use the inferred evolution in the stress balance to better understand the mechanisms responsible for the observed behavior.


Area changes and the influence factors of Terra Nova Bay Polynya from 2004–15

Yifan Ding, Xiao Cheng, Fengming Hui

Corresponding author: Xiao Cheng

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

The Terra Nova Bay Polynya (TNBP) plays an important role in mass circulation between the ocean and the atmosphere. The distribution of sea ice in TNB directly influences the area of the polynya. According to this, we observed the area changes of TNBP in winter from 2004–15 using AMSR-E and AMSR2 data. Here we extracted the area extent and defined the open water time by using the sea-ice concentration data and by estimating the junction with the north of Drygalski Ice Tongue and the south of Cape Washington. The results show that, in the early years(2004–10), TNBP tends to be more expansive and broken up; however, it gradually becomes relatively stable and the mean area also decreases in later years (2011–15). The area of TNBP has a distinct seasonal nature, which is significantly decreased from June to September, along with the temperature becoming lower. The maximum area is nearly 8800 km2, which usually appears in the primary and end growth of sea ice, i.e. early March and late November. The minimum area is 0 km2, which means TNBP can be completely covered by thick sea ice (sea-ice concentration >70%). Meanwhile, the surrounding glaciers, such as Priestley Glacier and Reeves Glacier, intensify the katabatic winds from the Antarctic interior, which result in frequent fluctuations of area during the whole observation period. The three major influencing factors are the strong katabatic wind from the glaciers, the blocking effect on the sea ice to the south of the Drygalski ice tongue and the new sea ice generated at a lower temperature. We also analyse the relationship between TNBP area changes and the meteorological data from the nearby automatic weather stations, such as temperature and wind speed.


Icebergs in sea ice framework: Ideas towards a numerical model of ice melange

Irena Vankova, David Holland, Alon Stern

Corresponding author: Irena Vankova

Corresponding author e-mail: vankova@cims.nyu.edu

The ice melange is an important factor in glacier ocean interactions; however, it is hard to model numerically and its effects are usually parametrized. We expand on the ideas used in sea ice modeling and present a new approach towards modeling the melange. Sea ice is usually modeled as a continuum characterized by its thickness, compactness and rheology, which allows for large compressive strength but zero tensile strength. In such a model, divergent wind can easily blow sea ice apart while convergent wind slowly piles sea ice up; neither of these are desirable behaviors for an iceberg. Therefore a new iceberg class is introduced in a sea ice model and its properties are modified such that large tensile strength keeps each single iceberg together. Further, an algorithm is introduced to distinguish between individual icebergs, which prevents them from merging after collisions. This is currently being implemented in a one-dimensional cavitating fluid sea ice model proposed by Flato and Hibler and we plan to extend it to two dimensions.


Precursor motion to iceberg calving at Jakobshavn Isbræ, Greenland, observed with terrestrial radar interferometry

Surui Xie, Timothy Dixon, Denis Voytenko, David Holland, Denise Holland, Tiantian Zheng

Corresponding author: Surui Xie

Corresponding author e-mail: suruixie@mail.usf.edu

Time-varying elevations near the calving front of Jakobshavn Isbræ, Greenland, were observed with Terrestrial Radar Interferometry (TRI) in June 2015. An ice block with surface dimensions of 1370 × 290 m2 calved on 10 June. A TRI-generated time series prior to calving shows that ice elevation near the calving front began to increase 65 hours before the event, and can fit a simple block rotation model. We hypothesize that subsurface melting at the base of the floating zone of the floating terminal region of the glacier breaks the gravity–buoyancy equilibrium, leading to slow subsidence and rotation of the block, and its eventual failure.


Historical simulations of ice-shelf/ocean/sea-ice interactions using two different models

Kaitlin Alexander, Ben Galton-Fenzi, Katrin Meissner, Matthew England, Hartmut Hellmer, Ralph Timmermann, Dmitry Sein, Tore Hattermann, Jens Debernard

Corresponding author: Kaitlin Alexander

Corresponding author e-mail: k.alexander@unsw.edu.au

In recent years, ice shelf thermodynamics routines have been successfully implemented by a number of ocean models. However, model intercomparisons of ice-shelf/ocean interactions, simulated on realistic Antarctic domains, have been lacking. In this study, we use two different ocean models, MetROMS and FESOM, to investigate ice-shelf/ocean/sea-ice interactions throughout the Southern Ocean. MetROMS consists of the regional ocean model ROMS and the sea ice model CICE, coupled using MCT, and including ice shelf thermodynamics. FESOM is a global ocean/sea-ice model, including ice shelf thermodynamics, with an unstructured triangular mesh allowing for spatially varying resolution. Using two different configurations of each model, we simulate the period 1992–2005, forced by ERA-Interim atmospheric reanalyses. By analysing these four simulations we investigate the sensitivity of ice-shelf/ocean/sea-ice interactions to the inclusion of tides, model resolution and differences in model physics.


Ice shelf thickness, melt rates and inland response of the Pine Island Glacier, West Antarctica, from a 2008–15 high-resolution DEM record

David Shean, Ian Joughin, Ben Smith, Pierre Dutrieux

Corresponding author: David Shean

Corresponding author e-mail: dshean@uw.edu

The Amundsen Sea sector of the West Antarctic Ice Sheet has experienced significant grounding line retreat, acceleration and thinning in recent decades. These changes are directly linked to ice–ocean interaction beneath ice shelves, but existing observations of the spatial distribution, timing and magnitude of ice shelf basal melt are very limited. We combine all available high-resolution WorldView stereo DEMs (~2 m px–1), SPIRIT DEMs (~40 m px–1), ICESat GLAS, ATM and LVIS altimetry data to produce a 2002–15 DEM time series for the Pine Island Glacier (PIG) shelf and ice stream. We also integrate several available bed datasets, and 2006–15 surface velocity maps from TerraSAR-X/TanDEM-X, ALOS and LS-8. These DEM products are used to compute ice thickness, Eulerian d H/dt and Lagrangian DH/Dt, which document changes in grounding line position, ice shelf thickness and upstream grounded ice dynamics. Ice shelf basal melt rate estimates are derived from Lagrangian DH/Dt measurements and flux gate mass budget analysis. We compare these remote sensing results with in situ borehole measurements and Lagrangian DH/Dt records from 2012–14 GPS data. Observed basal melt rates are ~60–70 Gt a–1 over the main shelf, with localized rates of >150–250 m a–1 within the PIG inner cavity, and significantly lower rates of <10–40 m a–1 for the outer shelf. We document the interannual evolution of melt rates for the 2008–15 period, and compare with oceanographic observations. Our results show a decrease in melt from 2008–10, with limited variability from 2010–15. We also present Eulerian dh/dt records upstream of the PIG grounding line, which show significant thinning (>5–10 m a–1) from 2008–10 following the ~2008–09 ungrounding of the PIG ‘ice plain’. This was followed by ice shelf regrounding on a large transverse seabed ridge and significant ice shelf thickening, likely driven by a combination of increased grounding line discharge and reduced melt rates. We are using a similar approach to document the evolution of adjacent shelves in the Amundsen Sea sector, which will improve our understanding of the relationship between ice–ocean interaction and ice sheet mass loss.


Interactions between ice shelves and ocean in Antarctica: grounding line dynamics and ocean properties

Enea M. Montoli, Florence Colleoni, Dorotea Iovino, Simona Masina

Corresponding author: Enea M. Montoli

Corresponding author e-mail: enea.montoli@gmail.com

Shrinking of ice sheets is expected as a consequence of the rising temperature, leading to an inevitable rise in sea level. The major concern of this projected trend of ice sheets is the Antarctic Ice Sheet (AIS) and in particular its western flank. The reason for concern is that a great part of the West Antarctic Ice Sheet (WAIS) is grounded on a bedrock topography that is below sea level and that deepens inland, defining the WAIS, from a grounding-line dynamics point of view, as an ice sheet in a favourable unstable condition. Therefore, understanding the dynamics of the grounding line and how it may respond to various perturbations, such as basal melting and calving, is critical for forecasting the evolution and reconstructing the past extensions of the AIS. By means of idealized numerical experiments performed with GRISLI, we test the regional and time-period sensitivity of idealized ice shelves to different parameterizations of basal melting and descriptions of calving as well as to atmospheric and oceanic forcing. As the ice-shelf’s dynamics are strongly dependent on the sub-ice-shelf bathymetry and on the initial position of the grounding line, the set of experiments is performed with several idealized ice-shelf bathymetries, based mainly on Gudmundsson et al., 2012 and resembling the various types of ice shelf around Antarctica. In numerical experiments the description of both basal melting and calving has been varied. Since the impact of the ocean on the ice sheet might not be linear but strongly influenced by the mean background climate state, the experiments are forced with different atmospheric and oceanic conditions (i.e. present day, Last Glacial Maximum, Pliocene).


Fully synchronous ice–ocean coupling using MITgcm

James Jordan, Paul Holland, Adrian Jenkins, Robert Arthern, Daniel Goldberg

Corresponding author: James Jordan

Corresponding author e-mail: jamrda26@bas.ac.uk

Very little of the Antarctic ice sheet experiences surface melting; much of its ablation takes place under the floating extensions of the sheet (ice shelves), which are exposed to ocean warming. Ice-shelf thinning due to submarine melting has the potential to cause significant ice loss from portions of the ice sheet, and hence sea-level rise. Oceanographic and remote-sensing observations of this region in recent years have shown a pattern of high ice-shelf melt rates coincident with widespread thinning of grounded ice and therefore sea-level rise. However, the processes by which ocean warming leads to ice-sheet loss are poorly understood. The gap in knowledge is in part due to a lack of coupled numerical ice–ocean models capable of representing the key feedbacks between the ocean and ice shelves. To change this, a fully synchronous, coupled ice–ocean model of the complete ice-stream/ice-shelf/ocean system has been developed using MITgcm (Massachusetts Institute of Technology general circulation model) to investigate these processes. Unlike previous asynchronous coupled models, the position of the ice–ocean interface and grounding line evolves every time step with no need for dump and restart methods. We present preliminary results, demonstrating the capabilities of our new model.


The impact of seasonal variability on ice-shelf structure and melt-water plumes

Christopher MacMackin, Andrew Wells, Ian Hewitt

Corresponding author: Christopher MacMackin

Corresponding author e-mail: christopher.macmackin@physics.ox.ac.uk

Ice shelves can perform an important role in buttressing inland ice sheets and slowing their flow into the ocean. Understanding the evolution and stability of ice shelves is thus necessary if one is to predict the rate of sea-level rise from ice sheets. The stability of ice shelves is known to be effected by the presence of basal channels. In addition to the well established existence of channels running parallel to the shelf’s direction of flow, thickness variations can also be observed at an oblique angle to the flow (for example on Pine Island Glacier’s ice shelf). We evaluate mechanisms that could produce such thickness variations via a coupling to ocean dynamics, using an idealized model of a coupled one-fimensional ice shelf and plume to study time-periodic variations about a background steady state. We consider the potential for seasonal variations in subglacial discharge to modulate ice-shelf thickness, by causing increased entrainment of warm ocean water and hence increased melting near the grounding line at certain times of year. Results of a linear analysis indicate the formation of small ripples, running perpendicular to the shelf’s flow, which are then advected away from the grounding line and towards the calving front. Furthermore, such seasonal variations cause global thinning and thickening of the shelf as a whole. However, these changes are relatively small and cannot, on their own, explain large-amplitude oscillations in ice-shelf thickness. This suggests that other mechanisms may be the main driver of the development of thickness oscillations, such as variations in ocean temperature.


Flow regime and crevasse extent of the McMurdo Shear Zone, Antarctica, using GPR and GPS observations and numerical modeling

Lynn Kaluzienski

Corresponding author: Lynn Kaluzienski

Corresponding author e-mail: lynn.kaluzienski@maine.edu

Sub-ice-shelf circulation plays a fundamental role in ice-shelf mass budget. The shape of the underside of an ice shelf is important, such that the presence of basal crevasses can significantly modulate the transfer of heat at the ice–ocean interface. In situ observations of basal crevasses are challenging to obtain, but surface-based GPR surveys can be used to determine crevasse location and orientation. Here, we use GPR methods to map the internal structures in the McMurdo Shear Zone (SZ) which marks the boundary between the Ross Ice Shelf and the slower-moving McMurdo Ice Shelf. Radar observations reveal the presence of crevasses within a zone of accreted marine ice at a depth of approximately 170 m. A spatial correspondence between surface and basal crevasses suggests that both are formed locally by lateral shearing. We use the Ice Sheet System Model (ISSM) to test this hypothesis. The model estimates the detailed velocity field of the SZ and is constrained by GPS-derived observations of surface motion. The distribution and orientations of surface and basal crevasses are consistent with the gradients in velocity field predicted by the model. This work suggests that high-resolution modeling can be used to predict the locations of basal crevassing which will lead to an improved understanding of ice-shelf mass balance processes.


Firn heterogeneity of Larsen C Ice Shelf from borehole optical-televiewing

David Ashmore, Bryn Hubbard, Adrian Luckman, Bernd Kulessa, Suzanne Bevan, Adam Booth, Peter Kuipers Munneke, Martin O’Leary, Heidi Sevestre

Corresponding author: David Ashmore

Corresponding author e-mail: dwa@aber.ac.uk

There is a growing body of evidence that significant environmental changes are under way on the Larsen C Ice Shelf (LCIS), Antarctic Peninsula. In addition to thinning, acceleration and rapid rift propagation, föhn winds drive intense melt in the northern and western regions of LCIS and, in some years, the formation of melt ponds visible in optical satellite imagery. The firn structure of ice shelves determines the fate of surface melt water, the preconditioning of the ice shelf to hydrofracturing and the interpretation of altimetric data, widely used to reconstruct ice-shelf thickness. The density structure of ice shelves is commonly conceptualized as a depth-integrated ‘firn air content’, the equivalent reduction in ice thickness if it were completely composed of bubble-free glacial ice. Bulk firn air content, as estimated by geophysical methods, necessarily neglects the complex structure of ice-shelf firn and ice. In order to address this knowledge gap we report a suite of ~90 m long optical televiewer (OPTV) borehole logs from the northern and central portions of LCIS collected in austral spring 2014 and 2015. OPTV logs record a LED-illuminated geometrically correct RGB image of the borehole wall. Using a newly calibrated density–luminosity relationship we reveal the density structure of five LCIS sites in detail. The firn/ice column is anomalously dense at all five sites with depth-averaged densities between 3 and 90 m ranging from 867 and 895 kg m3. The minimum estimated densities range between 622 and 816 kg m3 and reveal spatial variation in firn compaction rates, relating to the local effects of accumulation, air temperature and surface melt. Here density profiles are irregular and frequently interrupted by melt layers, which can form continuous ice layers >10 m thick. We identify three distinct ice facies interpreted as: (a) locally accumulated dense firn/ice with refrozen ice layers; (b) continuous massive ice formed by intense melt, commonly associated with surface ponding; and (c) advected, contorted glacial ice. We investigate spatial variability of these facies with complementary ice-penetrating radar data. Radargrams show laterally continuous horizons intersected by discrete reflection-free zones, of typical dimension ~0.02 × 1 km, interpreted to be bodies of massive refrozen ice. Our findings generally support previous estimates of firn air content spatial variability and highlight the complexity of ice shelf structure.


Impact of the Southern Annular Mode on the basal melting under the ice shelves through the Plio/Pliocene transition

Florence Colleoni, Eneal Montoli, Simona Masina, Annalisa Cherchi

Corresponding author: Florence Colleoni

Corresponding author e-mail: florence.colleoni@cmcc.it

Basal melting is an interface process that plays an important part in the reconstruction of past and future ice sheet topography under different mean background climate states. The impact of the main atmospheric and oceanic teleconnections on the basal melting under the ice shelves has been seldom investigated. The Southern Annular Mode (SAM), the main impact of which is to cause a strengthening of the winds around Antarctica and a warming over the Antarctic Peninsula during its positive phase, is also known to impact on the extent of the sea-ice cover and on the sea-surface temperatures. These effects may be of importance for the past and future evolution of basal melting under the ice shelves. By means of stand-alone ice sheet model, we investigate the impact of the SAM on the evolution of the Pliocene and preindustrial Antarctic ice sheet topography. First, we analyse some of the atmospheric and oceanic model outputs obtained in the framework of the PLIOMIP2 project. We analyse the atmospheric and oceanic characteristics of the positive and the negative phases, which we separate with an EOF analysis for both time periods. Second, the climate forcing is split into SAM+ and SAM years to force off-line the SIA/SSA GRISLI ice sheet model (Ritz et al., 2001). In addition, we explore the role of different ice-shelf basal melting parametrizations on the evolution of the topography during through the Plio–Pleistocene transition.


Optical satellite-derived estimates of submarine melt rates along the northern coast of Greenland

Nat Wilson, Fiamma Straneo

Corresponding author: Nat Wilson

Corresponding author e-mail: njwilson23@gmail.com

The modulation of grounded ice by outlet glaciers and ice shelves is a major source of uncertainty in predicting future sea-level rise. In many settings, submarine melt is expected to be an important process, and changes in submarine melt rates have the potential to destabilize marine glacier systems. However, controls on submarine melt rate vary from glacier to glacier and are poorly understood, with subglacial hydrology, fjord geometry and ambient ocean circulation all likely to have varying degrees of importance. Motivated by the need to understand how melt rate variability is forced by oceanic vs glaciological processes, we have calculated melt rate fields for four major ice tongues along the northern coast of Greenland. Using Worldview imagery, we construct surface elevation maps and deformation maps, and use Lagrangian mass-continuity methods to compute melt rates. In addition to rapid melting in the vicinity of glacier grounding zones, we find that shelf-channel geometry influences local melt rates. At Ryder glacier, a single deep ice tongue channel melts rapidly at its lateral margins and more slowly along its keel. We observe submarine melt regimes at Steensby Glacier both before and after that ice tongue experienced a major calving event in 2014 that dramatically changed its geometry. We find evidence for seasonal variability in melt rates. Finally, we compare estimated melt rates to observed ice discharge, ice tongue draft and shape, estimated subglacial catchment size and nearby ocean properties. Attributing observed melt rates to local and regional characteristics is currently under way.


The effect of iceberg calving size distribution on melt patterns and sea-ice formation around Antarctica

Alon Stern, Alistair Adcroft, Olga Sergienko

Corresponding author: Alon Stern

Corresponding author e-mail: sternalon@gmail.com

The freshwater flux from the Antarctic continent into the global ocean occurs through basal melting and the calving of icebergs off the edge of Antarctic ice shelves. After calving, icebergs can drift long distances before melting entirely, depositing their meltwater remotely and affecting sea ice formation away from the Antarctic coastline. In this study, we use a coupled general circulation model with an iceberg component to examine how the initial size of calving icebergs affects iceberg trajectories and melt patterns around Antarctica. Results show that increasing the size of calving icebergs leads to an increase in the westward iceberg freshwater transport around Antarctica. For simulations using larger icebergs, the iceberg freshwater transport out of the Amundsen and Bellingshausen Seas causes sea ice in the region to be reduced by more than 20% (relative to a control simulation where icebergs are melted instantaneously upon entering the ocean). These results suggest that absence of icebergs in some climate models may introduces systematic biases in sea-ice formation and ocean temperatures and salinities around Antarctica.


Simultaneous disintegration of outlet glaciers in the Porpoise Bay region of East Antarctica, driven by sea ice break-up

Bertie Miles, Chris Stokes, Stewart Jamieson

Corresponding author: Bertie Miles

Corresponding author e-mail: a.w.j.miles@durham.ac.uk

Iceberg calving is an important process accounting for around 50% of total mass loss to the ocean in Antarctica. Moreover, dynamic feedbacks associated with retreat in buttressing ice shelves or floating glacier tongues can result in an increased discharge of ice into the ocean. Therefore, improving our understanding of the mechanisms driving glacier calving and how glacier calving cycles have responded to recent changes in the ocean–climate system is important in the context of future sea level predictions. In this study, we analyse the calving rates of several glaciers in the Porpoise Bay region of East Antarctica using Envisat ASAR imagery at approximately monthly intervals between November 2002 and March 2012. We observe a large simultaneous calving event in 2007 where a total of ~2900 km2 of ice calved from multiple glaciers in the region. We link this calving event to a break out of the landfast sea ice that usually occupies the bay. In the absence of regular satellite imagery prior to 2002, we use satellite-based sea ice concentrations between 1972 and 2014 in Porpoise Bay as a proxy for iceberg calving. This infers a potential speed up in the rate of calving for Holmes glacier, the largest in Porpoise Bay.


The effects of the flexural gravity waves on stress regime of the Ross Ice Shelf

Olga Sergienko

Corresponding author: Olga Sergienko

Corresponding author e-mail: osergien@princeton.edu

Recently deployed arrays of passive seismometers record impacts of the long-period ocean waves (e.g. tsunamis, infragravity waves) on the Ross Ice Shelf (RIS). Arrived at the RIS ice front, these long-period waves excite flexural gravity waves that propagate through the ice shelf and its cavity. Numerical results of a three-dimensional model simulating propagation of the flexural gravity waves in the Ross ice-shelf/sub-ice-shelf cavity demonstrate that the geometry of the ice-shelf cavity (its bathymetry and the ice-shelf draft) controls the wave propagation and the amplitude of the ice-shelf flexure. The asymmetry of the RIS cavity – the eastern part is much shallower than the western part – results in larger flexural amplitudes, and hence flexural stresses, of the eastern side of the RIS. Comparison of the model simulations with observations suggests that elastic properties of the Ross Ice Shelf (e.g. Young’s modulus) are spatially heterogeneous.


Recent glacial earthquakes in Greenland

Kira Olsen, Meredith Nettles

Corresponding author: Kira Olsen

Corresponding author e-mail: kolsen@ldeo.columbia.edu

Large calving events at Greenland’s outlet glaciers produce teleseismically detectable glacial earthquakes. These events are observed in the seismic record for the past 22 years, and are valuable remote indicators of glacier dynamics and the timing of mass loss. The published catalog of glacial earthquakes through 2010 numbers only ~300 events. The annual occurrence of these long-period events has increased substantially since the 1990s, which makes recent years especially valuable in expanding the global dataset. We analyze 3 years of glacial earthquake data (2011–13) which increases the number of studied events by 42%. From 2011–13 glaciers on Greenland’s west coast dominated glacial earthquake production. Kong Oscar Glacier, Upernavik Isstrøm, and Jakobshavn Isbræ all produced frequent glacial earthquakes during this time. We also document glacial earthquakes at previously quiescent glaciers on Greenland’s northwest coast. Glacial earthquakes are known to occur when marine-terminating glaciers are close to their grounding lines; events do not occur at glaciers with floating ice tongues. We link patterns in glacial earthquake production and cessation from 2011 through 2013 with the presence or absence of floating ice tongues on glaciers around Greenland. The calving model predicts that the force azimuths produced by glacial earthquakes should be oriented perpendicular to the calving front, and comparisons between seismic data and aerial imagery show this to be true in most instances. At two glaciers we document force azimuths that are not aligned fjord-parallel. These glaciers have retreated past the confines of their fjords and undergone subsequent increases in calving-front area. These results show how patterns in glacial earthquake production correspond to independently observed changes in glacier dynamics.


Velocity variability in Larsen C Ice Shelf

Adrian Luckman, Suzanne Bevan, Laurie Padman, Matt King, Bernd Kulessa, Martin O’Leary, Bryn Hubbard, Helen Fricker

Corresponding author: Adrian Luckman

Corresponding author e-mail: A.Luckman@Swansea.ac.uk

The flow rate of an ice shelf is expected to respond to a number of factors including the onset of instability, changes in thickness, calving losses and structural evolution resulting from surface melting or basal accretion. However, to date there is only limited evidence of speed changes on ice shelve, possibly because data suitable for measuring ice flow has hitherto been sparse, tides can modulate flow, and tidal corrections may be crude. With the launch of Sentinel-1 SAR, however, and ESA’s abundant data acquisition and distribution approach, it is now possible to examine ice-shelf velocity variability and its interaction with tides in more temporal detail. Here we investigate Larsen C Ice Shelf on the Antarctic Peninsula, which is considered vulnerable to the rapidly changing regional climate. We generate velocity maps using feature and speckle tracking from temporally sparse TerraSAR-X data covering small areas at high resolution, and temporally dense Sentinel-1 SAR, covering the whole shelf in a few frames at medium resolution. We use the latest tide model (CATS2008a) to correct for vertical tidal displacement between images, investigate the potential impact of atmospheric pressure, and compare our data to an earlier velocity map. Our high-resolution velocity fields reveal significantly narrower and more intense shear zones than previously reported, with implications for ice-shelf modelling. We find evidence of a slow-down in the northern sector of the ice shelf, and will discuss the magnitude of this signal in comparison with the potential errors in the corrections for tides and atmospheric loading, and of potential speed modulation by tidal forces.


Geoengineering glacial fjords

Michael Wolovick, Olga Sergienko

Corresponding author: Michael Wolovick

Corresponding author e-mail: wolovick@princeton.edu

Climate change threatens the long-term stability of large regions of the Greenland and Antarctic Ice Sheets. The dynamic losses from these ice sheets are governed by the retreat of relatively small outlet glaciers and ice streams forced by warm ocean water at the grounding line. In general, this warm water resides offshore at depth and accesses the grounding line through deep but narrow troughs carved across the continental shelf during previous glacial maxima. In Greenland, warm water access to key outlet glaciers is further constricted through fjords that are only 5–10 km wide. Here, we explore the possibility of blocking warm water transport through these choke points with an artificial sill. We use a simple width-averaged model of ice-stream flow coupled to a buoyant-plume model of ocean-forced melting to investigate whether grounding line retreat can be slowed or reversed if warm water is blocked from accessing the ice front at depth. We investigate what fjord conditions (temperature and salinity) are necessary to produce stable outlet glaciers. We test how these conditions depend on the characteristics of the outlet glacier, such as geometry and basal or lateral drag. We run these tests both for idealized geometries and for flowband profiles representing Jakobshavn, Helheim, Kangerdlugssuaq, and Petermann Glaciers in Greenland, as well as Thwaites and Pine Island Glaciers in Antarctica. We use these tests to define what temperature and salinity profiles would need to be created at the ice front to stabilize the outlet glaciers, and we discuss the feasibility of producing those profiles given the water masses present offshore.


Simulating the evolution of Thwaites Glacier with a coupled ice–ocean model

Helene Seroussi, Dimitris Menemenlis, Eric Larour, Yoshihiro Nakayama, Mathieu Morlighem

Corresponding author: Helene Seroussi

Corresponding author e-mail: helene.seroussi@jpl.nasa.gov

Ice shelves and floating glacier termini play an important role in the stability of ice sheets and interact strongly with the ocean. They account for much of the buttressing against the flow of inland glaciers that drain the Antarctic and Greenland ice sheets. Changes in their geometry due to ice-front retreat, thinning or even collapse profoundly affect the flow of their tributary glaciers, which in turn affects the volume of grounded ice carried by these tributary glaciers into the ocean, and the extent of resulting sea level rise. Recent simulations of glaciers in Antarctica show that the largest climatic impact on ice dynamics is the rate of ice-shelf melting, which rapidly affects glaciers’ speed over several hundreds of kilometers upstream of the grounding line. However, accurate knowledge of these melting rates, as well as their spatial and temporal evolution, remain largely unknown. Coupled ice–ocean models are the best approach to address this question. In this study, we focus on Thwaites glacier in the Amundsen Sea sector, a glacier that has been accelerating, widening and experiencing a complex grounding line retreat pattern over the past three decades. We simulate the coupled ice–ocean system using a new two-way coupled system between the Massachusetts Institute of Technology general circulation model (MITgcm) and the Ice Sheet System Model (ISSM). We investigate the feedbacks between changes in the ice and ocean, and the dynamic response of the glacier to changes in the ocean circulation. Our results show the importance of simulating the coupled ice–ocean system to produce accurate sub-ice-shelf melting rates and suggest that Thwaites Glacier is likely to undergo substantial changes in the coming decades. This work was performed at the California Institute of Technology’s Jet Propulsion Laboratory under a contract with the National Aeronautics and Space Administration, Cryospheric Sciences and Modeling, Analysis and Prediction Programs.


Drift, evolution and break-up of Iceberg B-15

Ted Scambos, Seelye Martin, Sarah Neuhaus, Lei Wang, Doug MacAyeal, David Long, Jennifer Bohlander, Michon Scott

Corresponding author: Ted Scambos

Corresponding author e-mail: teds@nsidc.org

We examine the record of available data on the evolution of the largest iceberg ever observed to calve, B-15, and the subsequent calvings of the berg. The drift patterns of the icebergs show influences from tides, currents and bathymetry at a variety of spatial and temporal scales, revealing strong large-scale controls from bathymetric topography and conservation of vorticity over bathymetric features. Changes in the size of the icebergs during drift shows a pattern of evolution that indicates that sea surface temperature has a strong control on the rate of edge wasting. Ice thickness rates also change rapidly as ocean basal temperatures rise. Surface melt and firn evolution of the large icebergs in the B-15 group is evaluated by the backscatter record from two satellite scatterometers, and shows a consistent trend of increased backscatter after re-freezing with increasing surface melt days until approximately 80 total days of melt have occurred, followed by decreasing backscatter and rapid disintegration. Meter-scale images of iceberg B-15J’s break-up illustrate several processes of iceberg disintegration at the end of this surface evolution process. The set of observed changes in the iceberg and the disintegration events are informative for analysis of ice-shelf disintegration events due to climate warming.


Simulating the failure of Thwaites Glacier using semi-brittle rheology

Liz Logan, Eunseo Choi, Eh Tan, Luc Lavier, Ginny Catania, John Holt

Corresponding author: Liz Logan

Corresponding author e-mail: esl359@gmail.com

We simulate the development of ice geometry resulting from semi-brittle rheology and the formation of basal crevasses at the grounding line of Thwaites Glacier and investigate the role that basal melting has on the spacing of these features. We employ a numerical model, DynEarthSol3D, with the capability of accounting for both viscous and elastic stresses with a Mohr–Coulomb failure criterion to account for the brittle failure of ice. The implementation of a ductile-to-brittle transition based on strain rate in ice is critical in developing the terminus geometry that determines the calving rate at Thwaites Glacier, which has recently retreated back close to its grounding line. This result may help provide constraints to predict whether ice sheets are primed to fail catastrophically or calve in stable ways.


Ocean–ice-shelf interactions in MPAS-Ocean and the Accelerated Climate Model for Energy (ACME)

Mark Petersen, Xylar Asay-Davis

Corresponding author: Mark Petersen

Corresponding author e-mail: mpetersen@lanl.gov

The Accelerated Climate Model for Energy (ACME), a new initiative by the US Department of Energy, includes unstructured-mesh ocean, land-ice and sea-ice components using the Model for Prediction Across Scales (MPAS) framework. The ability to run coupled high-resolution global simulations efficiently on large, high-performance computers is a priority for ACME. We are coupling the MPAS-Ocean the MPAS-Land-Ice models to better understand how changing ocean temperature and currents influence glacial melting and retreat. These simulations take advantage of the horizontal variable-resolution mesh and adaptive vertical coordinate in MPAS-Ocean, in order to place high resolution below ice shelves and near grounding lines. MPAS-Ocean has been tested with compressed vertical coordinates and melt fluxes below ice shelves in idealized test cases and realistic global ocean cases. We are now able to run global simulations using a resolution of 60 km, 10 km or 5 km in the ocean cavities below ice shelves. The land ice draft and sub-ice-shelf bathymetry is initialized using Bedmap2. Current and future work includes coupling to a dynamic MPAS-Land-Ice model within ACME, and the ability to represent moving grounding lines within the ocean/ice system.


On the processes affecting melt rates beneath the ice shelves of the Amundsen Sea sector

Nicolas C. Jourdain, Pierre Mathiot, Gaël Durand, Julien Le Sommer, Paul Spence, Anouk Vlug

Corresponding author: Nicolas C. Jourdain

Corresponding author e-mail: njourdain@lgge.obs.ujf-grenoble.fr

The largest Antarctic contribution to sea level rise over the 21st century will likely be related to the acceleration of a few glaciers in West Antarctica. A large part of the mass loss would occur through Pine Island and Thwaites glaciers, which seem to be engaged in a dynamical instability triggered by coastal ocean warming. Previous studies have emphasized the importance of the amplitude of melt-rates beneath ice-shelves in the timing of glacier retreat. However, modelling melt rates in realistic configurations still remain challenging. In this presentation, we use the regional AMU12.L75 configuration of the NEMO3.6 model to understand the processes that control melt rates beneath seven ice-shelf cavities of the Amundsen Sea. First, the sensitivity of melt rates to exchange coefficients at the ice–ocean interface indicates a linear relationship between melt rates and both the barotropic and overturning circulations within cavities. Our simulations suggest that 35–50% of the barotropic circulation in the Amundsen Sea is explained by melt in ice-shelf cavities. Then, the sensitivity of melt to tides is investigated by prescribing tidal harmonics at the lateral boundaries. We also show that melt rates under the Crosson ice shelf can be significantly impacted by the tabular icebergs that are often grounded/fasten near its front.


Inferring Greenland fjord and coastal bathymetry using icebergs

Jessica Scheick, Gordon Hamilton

Corresponding author: Jessica Scheick

Corresponding author e-mail: jessica.scheick@maine.edu

Warm ocean waters are thought to play an important role in controlling mass loss from Greenland’s marine-terminating outlet glaciers. Heat is transported in relatively warm, salty, dense Atlantic Waters circulating off the coast of Greenland, but the bathymetry of the shelf and fjord environment exerts an important control on the ability of these waters to access the ice-sheet margin. The presence of a shallow bathymetric sill acts as a barrier and can prevent these warm waters from entering the fjord, whereas the lack of a sill means the Atlantic waters have easier access to the ice-sheet margin. The bathymetry of most of Greenland’s glacial fjords is not resolved in sufficient detail to indicate the presence or absence of a sill, which limits our ability to predict dynamic glacier responses to the delivery of oceanic heat. Here we develop a methodology to infer bathymetry using icebergs observed in remotely sensed datasets. First, we identify regions of iceberg stranding. We extract stranded iceberg freeboards from digital elevation models (DEMs) derived from very high-resolution stereo satellite images. The median iceberg freeboard is then used to estimate iceberg draft and hence maximum water depth. Preliminary results in regions where bathymetric measurements are available validate the method and encourage its application in uncharted fjords.


Growth rates of first-year sea ice determined from banding

Kate Turner, Joe Trodahl, Inga Smith

Corresponding author: Inga Smith

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

The growth processes of first-year sea ice are of considerable importance in modelling the energy balance in ice-covered oceans. However, there is a paucity of sea ice growth rate measurements early in the season, when it is unsafe to travel because the ice cover is thin. This study has successfully followed the ice growth in the upper 1 m of first-year sea ice, using the visible banding structure present in an ice core collected at the end of the winter season. Visible banding has been observed for decades in cores and slabs lifted from near-shore first-year sea ice. The banding process is associated with under-ice currents; rapid currents encourage mixing of the concentrated brine that is rejected from the freezing front. The bands appear as variations in turbidity of the ice, which is observed as light and dark bands. The bands commonly show remarkable regularity at sites where the currents are dominated by tidal flows. In this study we exploit that regularity to determine the early-season growth rate along the Ross Island coast of McMurdo Sound, based on the known tidal pattern. The results correlate well with the known date at which the ice reached 1.1 m thickness, and the shallower growth variations show excellent agreement with the air-temperature fluctuations at nearby McMurdo Station and Scott Base. Independent modelling of the ice growth closely mirrored the variation in growth rate with ice thickness which was obtained from the banding. The results offer the potential to monitor early-season sea ice growth with post-growth-season measurements, which would be invaluable to the study of ice growth and behaviour.


Response of the Ross Ice Shelf to tsunami and infragravity wave forcing

Peter Bromirski, Anja Diez, Peter Gerstoft, Ralph Stephen, Olga Sergienko, Douglas Wiens, Rick Aster, Andrew Nyblade

Corresponding author: Peter Bromirski

Corresponding author e-mail: pbromirski@ucsd.edu

The response of the Ross Ice Shelf (RIS) to the 16 September 2015 8.3 Mw Chilean earthquake tsunami (>75 s period) and year-around infragravity (IG) waves (75–300 s period) were recorded by a broadband seismic array spanning the RIS from November 2014 to November 2015. The array included two linear transects, one orthogonal to the shelf front, extending 415 km southward toward the grounding zone, and an east-west transect roughly parallel to, and approximately 100 km landward of, the RIS front. Signals generated by both the tsunami and IG waves were recorded at all stations on floating ice, with little ocean-wave-induced energy reaching three RIS stations sited on grounded ice. Cross-correlation and dispersion curve analyses of the orthogonal array vertical and horizontal component data indicate that signals caused by tsunami and IG waves propagate across the RIS at gravity wave speeds (about 70 m s–1), consistent with flexure of the ice shelf by these waves propagating through the sub-shelf water cavity. IG band signals are continuously observed across the RIS, even during the winter sea ice maximum, indicating IG mechanical straining of the RIS throughout the year. Horizontal displacements are 2–3 times larger than vertical displacements, producing extensional motions that could facilitate expansion of existing fractures. Vertical and horizontal spectral amplitudes of both tsunami and IG-induced flexure displacements decrease progressively with distance from the front. Substantial variability parallel to the front is observed, likely a result of a combination of gravity-wave amplitude variability along the front, bathymetry under the shelf, water-layer geometry, and ice-shelf thickness and properties. In contrast to the very long period gravity waves, ocean swell (<30 s period), which only reaches the ice shelf in the austral summer, excites primarily elastic waves that propagate within the ice shelf at about 2800 m s–1 and produce negligible flexing of the shelf, consistent with relatively little swell energy penetrating the RIS cavity.


NEMO-Elmer/Ice in the context of MISOMIP tests

Nacho Merino, Nicolas C. Jourdain, Gaël Durand, Julien Le Sommer, Fabien Gillet-Chaulet, Olivier Gagliardini

Corresponding author: Nacho Merino

Corresponding author e-mail: Ignacio.Merino@lgge.obs.ujf-grenoble.fr

The West Antarctic Ice Sheet has shown a significant mass loss over the last decades. This is certainly related to an increase in ocean-induced melt near the grounding line. However, the associated processes are not correctly represented in current ice-sheet models, which differ in the magnitude of the melting-induced grounding line retreat and advance. Along the same lines as the inter-comparison experiments MISMIP, and MISMIP3D, a new set of experiments (MISOMIP) with three different sub-tests (MISMIP+, ISOMIP+ and MISOMIP1) has recently been designed in order to improve our understanding of ocean control of ice sheet dynamics. These tests have been performed using the NEMO and Elmer/Ice models and the results are discussed here. In particular, we analyse the impact of SSA equations compared to full-Stokes formulation on the ice dynamics.


Results from the second Ice Shelf–Ocean Model Intercomparison Project (ISOMIP+) and the first Marine Ice Sheet–Ocean Model Intercomparison Project (MISOMIP1)

Xylar Asay-Davis, Daniel Martin

Corresponding author: Xylar Asay-Davis

Corresponding author e-mail: xylarstorm@gmail.com

ISOMIP+ and MISOMIP1 prescribe a set of idealized experiments for ocean models with ice-shelf cavities and coupled ice-sheet–ocean models, respectively. ISOMIP+ and MISOMIP1 were designed together with the third Marine Ice Sheet MIP (MISMIP+) with three main goals: providing 1) a forum for comparing model results during development; 2) a path for testing components in the process of developing coupled ice sheet-ocean models; and 3) a basic setup for later parameter and process studies. The experimental design for the three MIPs is currently under review. We present results from ISOMIP+ and MISOMIP1 experiments using several ocean-only and coupled ice-sheet^ndash;ocean models. Among the ocean models, we show that differences in model behavior are significant enough that similar results can only be achieved by tuning model parameters (notably the turbulent transfer coefficients) for each model. This tuning is constrained by a desired mean melt rate in quasi-steady state under specified forcing conditions, akin to tuning the models to match observed melt rates. We compare the evolution of ocean temperature, kinetic energy, transport (as measured by the barotropic and overturning streamfunctions) and melt rates across models for four ISOMIP+ experiments with prescribed ice-shelf topography, two with static topography and two with dynamic topography. Then we present coupled results from the two MISOMIP1 experiments, comparing melt rates and ice-shelf topography across models. To demonstrate the potential usefulness of the MIPs as a basis for parameter studies, we also present a small number of examples. Using several ocean models, we show that melt rates respond sub-linearly to both changes in the top-drag coefficient (used to control the strength of drag on the ocean from the ice-shelf base) and the turbulent heat-transfer coefficient, and that melting is relatively insensitive to horizontal-mixing coefficients but more sensitive to vertical-mixing coefficients. From the coupled experiments, we demonstrate that varying the resolution of each model as well as the coupling frequency can significantly impact the results.


A synthetic high-resolution bedrock topography dataset for investigating resolution dependence of numerical ice-sheet model simulations

Felicity Graham, Jason Roberts, Duncan Young, Donald Blankenship, Martin Siegert

Corresponding author: Felicity Graham

Corresponding author e-mail: felicity.s.graham@gmail.com

The accuracy of numerical simulations of ice-sheet models, especially near the grounding line, relies to a large degree on the resolution of the underlying bed topography datasets. Current digital elevation models of Antarctic bed topography are heavily smoothed and interpolated onto coarse-resolution grids due to sparsely and unevenly sampled input data. We simulated a 100 m resolution synthetic bed elevation terrain for the whole Antarctic continent. The dataset preserves roughness characteristics of airborne and ground-based ice penetrating radar data from the ICECAP and BEDMAP1 compilations. The simulated bed elevation terrain is subsampled for a series of numerical simulations that investigate the sensitivity of ice flow dynamics near the grounding line to the resolution of the underlying terrain.


Modeling of subaqueous melting in Petermann Fjord, northwestern Greenland, using IceBridge data and a numerical ocean circulation model

Cilan Cai, Eric Rignot, Dimitris Menemenlis, Yoshihiro Nakayama

Corresponding author: CILAN CAI

Corresponding author e-mail: cilanc@uci.edu

Subaqueous melting of the floating tongue of Petermann Glacier in northwestern Greenland is by far the largest process of mass ablation. Melting of the floating tongue is controlled by the buoyancy of the meltwater plume, the pressure dependence of the melting point of sea ice, and the mixing of warm subsurface water with fresh, buoyant subglacial discharge. In prior simulations of this melting process, the role of subglacial discharge has been neglected because in similar configurations (floating ice shelves) in the Antarctic, surface runoff is negligible; this is, however, not true in Greenland (~5.7 Gt in the drainage basin of Petermann Glacier in 2008). Here, we use the Massachusetts Institute of Technology general circulation model (MITgcm) in a two-dimensional configuration and at a high spatial resolution (40 m × 20 m) to simulate the melting process of the ice shelf. The model is constrained by ice-shelf bathymetry and ice thickness from 2011 NASA Operation IceBridge, ocean temperature/salinity data from previous studies, and subglacial discharge at the grounding line deduced from estimated runoff by the Regional Atmospheric Climate Model (RACMO) of Royal Netherlands Meteorological Institute (KNMI). We compare simulated melt rate driven by 2008 seasonal runoff with an estimate obtained from surface mass balance, ice velocity and ice thickness from mass conservation. We also examine the sensitivity of the simulation to thermal forcing from the ocean, to changes of ocean bathymetry, and to the magnitude of subglacial runoff. We conclude on the impact of ocean and runoff from surface melting on the melting regime of the floating ice tongue of Petermann Glacier. The subaqueous melt rate increases ~20% with summer surface runoff. The sill position in the fjord changes the magnitude and pattern of subaqueous melting. This work is performed under a contract with NASA Cryosphere Program.


Sensitivity experiments with a one-dimensional coupled plume–ice flow model

Johanna Beckmann, Mahé Perrette, Andrey Ganopolski

Corresponding author: Johanna Beckmann

Corresponding author e-mail: beckmann@pik-potsdam.de

Over the past two decades net mass loss from the Greenland ice sheet has quadrupled, due to enhanced surface melting and speed-up of the marine-terminating outlet glaciers. This speed-up has been related, among other factors, to enhanced submarine melting, which in turn is caused by warming of the surrounding ocean and by increased subglacial discharge. For the future and recent mass balance changes of the Greenland Ice Sheet, ice–ocean processes potentially play an important role, yet they are not yet properly represented in contemporary Greenland Ice Sheet models. In this work we performed numerical experiments with a one-dimensional plume model coupled to a one-dimensional model of an outlet glacier. We investigated the sensitivity of submarine melt rate to changes in ocean properties (ocean temperature and salinity), to the amount of subglacial discharge and to the glacier’s tongue geometry itself for a number or representative Greenland outlet glaciers. We found that the simulated submarine melt rate is in a reasonable agreement with available empirical data. In a second set of experiments we investigate the response of a coupled ice-flow plume model to possible outcomes of climate change. In particularly, we examine the transient and equilibrium response of the outlet glaciers to changes in ocean temperature and subglacial discharge that affect both glacier geometry and submarine melt rates.


Spatiotemporal patterns of surface elevation change for Totten Glacier ice shelf

Chad A. Greene, Duncan A. Young, Donald D. Blankenship

Corresponding author: Chad A. Greene

Corresponding author e-mail: chad@chadagreene.com

Totten Glacier drains a submarine basin which holds the potential to contribute 5.1 m to global sea level rise. The spatially averaged surface elevation of Totten’s floating ice shelf trended negative during the ICESat epoch, but patterns are poorly constrained in space and time; thus, drivers of change are poorly understood. We assess the spatial distribution and timing of basal melt events between 2003 and 2013 by pairing ICESat satellite and ICECAP airborne laser altimetry. Previous attempts to attain high-precision spatiotemporal signals of melt/freeze from Eulerian analysis have been aliased by surface features which advect through repeat altimetry tracks. We use satellite photoclinometry to characterize the relief of short-wavelength surface features and we artificially translate these features with ice flow to simultaneously account for feature advection and cross-track slopes in repeat-track analysis. The resulting record of change is used to determine potential drivers of sub-ice-shelf melt.


Amundsen Sea ocean, sea ice and thermodynamic ice-shelf simulation with optimized model parameters

Yoshihiro Nakayama, Dimitris Menemenlis, Michael Schodlok, Eric Rignot

Corresponding author: Yoshihiro Nakayama

Corresponding author e-mail: ynakayam@jpl.nasa.gov

The ice shelves and glaciers of the West Antarctic Ice Sheet are thinning rapidly in the Amundsen Sea. The high basal melt rate of these ice shelves is caused by warm circumpolar deep water (CDW), which mainly intrudes via submarine glacial troughs located at the continental shelf break. Above the CDW, winter water (WW), with a potential temperature of ~–1.8°C can be observed at a depth of ~100–400 m. Although there have been many recent studies focusing on Pine Island Glacier (PIG), the objective of explaining observed changes and predicting future evolution remains elusive. One of the difficulties stems from the sensitivity of PIG basal melt to oceanic, sea-ice and atmospheric conditions. For example, a previous study shows that PIG melting is highly sensitive to the depth of thermocline at the Pine Island Ice Shelf front and that during the summer of 2012 basal melting decreased by 50%, which may potentially impact the evolution of PIG. In this study, we use the Massachusetts Institute of Technology general circulation model (MITgcm) in a regional Amundsen and Bellingshausen Seas configuration. The regional MITgcm configuration includes dynamic/thermodynamic sea ice and static thermodynamic ice-shelf representation. Horizontal spacing is about 10 km and there are 50 vertical levels, ranging in thickness from 10 m near the surface to ~450 m near the bottom. We have constrained the model, using a Green’s Functions approach, with available oceanographic observations and satellite-based estimates of ice concentration, ice velocity and ice-shelf melt rate by adjusting key internal model parameters, initial conditions, open boundary conditions and atmospheric forcing. After these adjustments, CDW intrusions into the Amundsen and Bellingshausen Seas are reproduced realistically and melt rates of the ice shelves are mostly consistent with satellite-based estimates. However, compared to the CTD observations, the WW layer tends to become too thin and too saline in most of our simulations. In this presentation, we will present the results of the optimized simulation with emphasis on the parameters that have the largest impact on reducing model–data differences between baseline and optimized solutions. We also discuss the mechanism controlling the thickness of WW, especially focusing on the area close to the Pine Island Ice Shelf front.


Viscosity and elasticity: a model-intercomparison of tidal bending in Antarctic grounding zones

Christian Wild, Oliver Marsh, Wolfgang Rack

Corresponding author: Christian Wild

Corresponding author e-mail: christian.wild@pg.canterbury.ac.nz

The Antarctic ice sheet is separated from its adjacent ice shelves by several-kilometre-wide grounding zones. Processes here determine the mass discharge from the relatively stable grounded ice sheet to the easily disturbed floating ice shelves. Hence, grounding zones are vital to ice-sheet mass balance and its coupling to the global ocean circulation. The response of this transition zone to tidal forcing has been described by both elastic and visco-elastic models, raising questions about our understanding of ice-shelf flexure at tidal frequencies. Here we optimize the values of ice viscosity and the Young’s modulus with available field data of the Southern McMurdo Ice Shelf (–78° 15′ S, 167° 7′ E). Observations of tidal movement were carried out by simultaneous tiltmeter measurements at six locations distributed along a profile across the grounding zone. These observations were complemented by differential GPS measurements over 75 days on the freely floating ice-shelf to record the tidal forcing during the field campaign. Finite-element simulations of temporal changes in tidal bending reveal that variations in Young’s modulus are required to match observations using the elastic approach which cannot be explained by incorporating the time derivative of the tidal wave. A forward model is employed to compare model predictions with tidal flexure patterns determined by SAR interferometry. In addition to the expected bending pattern in the flexure zone, a reversed inland tidal bending is observed, supporting the use of a fulcrum at the grounding line which transmits the oceanic forcing further upstream. These new insights provide important details towards correctly modelling grounding line behaviour and interpreting satellite interferometric data of ice-shelf flexure.


Regime shift in basal melt of the Mertz Glacier after the 2010 calving

Shigeru Aoki, Rio Kobayashi, Stephan Rintoul, Takeshi Tamura

Corresponding author: Shigeru Aoki

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

Mertz Glacier polynya off Adelie Land coast, Antarctica, was formed downstream of the Mertz Glacier Tongue and produced the third highest amount of sea ice among major Antarctic coastal polynyas. The high sea ice production was linked with the formation of dense shelf water, which led to the ventilation of Adelie Land bottom water, one of the varieties of Antarctic bottom water. In February 2010, the calving event of the Mertz Glacier Tongue significantly altered the neighboring icescape and sea ice production has decreased mainly due to the polynya relocation and change in production efficiency. In situ observations revealed salinity decreased downstream of the ‘Glacier Tongue’ region, which is consistent with the decrease in sea ice production. Measurement of the stable oxygen isotope ratio tells us more about what happened after the calving event. Oxygen isotope ratios in 2011 and 2015 were also significantly less than that in 2001, indicating the increased melt of continental ice as well as the decreased production of sea ice. A simple application of three-end member attribution, among meteoric ice, sea ice and modified circumpolar deep water, suggests an increase in meteoric ice fraction of around 20%. The results suggest a potential linkage between sea ice production and continental ice melt triggered by the icescape change. Given the recently observed change in Adelie Land bottom water, regime shift in local icescape on the Antarctic shelf can have potentially global impact.


The effect of ice shelves on grounding line stability

Marianne Haseloff, Olga Sergienko

Corresponding author: Olga Sergienko

Corresponding author e-mail: osergien@princeton.edu

Ice shelves transmit ocean forcing (e.g. variations in basal melt rates) inland by changing the stress balance at the grounding line. However, determining mechanisms and quantifying their effects remains a major challenge for ice sheet models, and the conclusions drawn from these models often depend on the type of model used. Recent simulations with two-dimensional, plan-view, ice-flow models suggest that ice shelves can change the stability properties of grounding lines through buttressing. In contrast, ice shelves can be excluded from the momentum balance in computationally efficient flowline models. To investigate whether parameterized buttressing can bridge the gap between these two classes of models, we extend a recently developed boundary layer approach and calculate how the ice flux at the grounding line changes under different parameterizations. This allows us to determine conditions under which lateral effects can be reproduced by flowline models. We illustrate that some commonly used parameterizations of ice-shelf buttressing do not change the stability properties of marine ice sheets in such models.


Mapping the central Amery Ice Shelf region, East Antarctica: seismic constraints on ice, ocean and seafloor structure

Leo Peters, Richard Coleman, Mark Lackie, Kathleen McMahon, Hugh Tassell

Corresponding author: Leo Peters

Corresponding author e-mail: leoepeters@gmail.com

The need to better understand the role of ocean–ice interactions on the fate of a marine terminating glacier or ice sheet is ever growing, as recent observations have identified warm circumpolar deep water reaching the marine termini of numerous outlet glaciers draining the Antarctic Ice Sheet. However, ice shelves obscure our ability to directly image these ocean–ice interactions along the grounding zone of many outlet glaciers and ice streams, making it difficult to truly determine the degree to which distant oceanographic observations from the continental shelf may ultimately influence grounding zone processes. Detailed observations of the sub-ice-shelf environment are thus necessary to both image and model these ocean–ice interactions and capture their role in ice dynamics. Here we present the results from a seismic campaign conducted across the Amery Ice Shelf between 2002 and 2006. Seismic observations of ice-shelf structure, ocean stratification and seafloor bathymetry were made at 63 locations within a 100 km × 50 km grid extending along the central portion of the ice shelf. The ice shelf varies by >200 m in thickness across the region, thinning in the oceanward direction. A distinct seismic reflector is observed within the ocean that is correlated to local oceanographic observations. Variations in seafloor bathymetry of up to 500 m are also imaged across the region, deepening toward the interior of the Amery Ice Shelf. These improved constraints on the regional structure of the Amery Ice Shelf will assist modellers in better understanding the impact of ocean–ice interactions on the evolution of the Antarctic Ice Sheet.


Dielectric losses in Totten Ice Shelf using multiple reflection from ice-penetrating radar

Laura Lindzey, Dustin Schroeder, Jamin Greenbaum, Duncan Young, Donald Blankenship

Corresponding author: Laura Lindzey

Corresponding author e-mail: lindzey@utexas.edu

Significant work has been done to model or constrain dielectric ice loss since it is one of the main sources of uncertainty when inferring basal conditions from ice penetrating radar, where basal reflection strengths can vary ~20 dB depending on bed roughness and the presence of water. As part of the multinational ICECAP collaboration, UTIG has collected airborne ice-penetrating radar data in East Antarctica for six seasons. This dataset includes more than 50 radar transects over the Totten Ice Shelf. In many of these transects, we see a ‘multiple’ reflection in addition to the main air–ice surface and ice–water reflections. The multiple corresponds to internal reflections within the ice shelf when the radar energy makes two round-trips through the ice. In this work, we use the the returned power in the main ice–water reflection in combination with the multiple to constrain the englacial losses for the ice shelf, and compare this dielectric loss to modeled values. This makes it possible to minimize the sensitivity of these estimates to variations in radar system parameters, geometric spreading loss and transmission/reflection losses. We then interpret these loss profiles in the context of ice-shelf thermal state, including the distribution of basal melt and freeze. This will be used for site selection of in situ studies for phase-coherent radar and validation of sub-ice circulation models.


Semi-automatic mapping of tidal cracks in the fast-ice region near Zhongshan Station in East Antarctica using Landsat-8 OLI imagery

Fengming Hui, Xinqing Li, Tiancheng Zhao, Mohammed Shokr

Corresponding author: Fengming Hui

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

Tidal cracks are linear features that appear parallel to coastlines in fast-ice regions due to the actions of periodic and non-periodic sea-level oscillations. They can influence energy and heat exchange between the ocean, ice and atmosphere, as well as human activity. In this paper, the LINE module of Geomatics 2015 software was used to automatically extract tidal cracks in fast-ice regions near the Chinese Zhongshan Station in East Antarctica from Landsat-8 Operational Land Imager (OLI) data with resolutions of 15 m (panchromatic band 8) and 30 m (multispectral bands 1–7). The detected tidal cracks were determined based on matching between the output from the LINE module and manually interpreted tidal cracks in OLI images. The ratio of the length of detected tidal cracks to the total length of interpreted cracks was used to evaluate the automated detection method. Results show that the vertical direction gradient is a better input to the LINE module than the top-of-atmosphere (TOA) reflectance input for estimating the presence of cracks, regardless of the examined resolution. Data with a resolution of 15 m also gives better results in crack detection than data with a resolution of 30 m. The statistics also show that, in the results from the 15 m resolution data, the ratios in Band 8 performed best with values of the above-mentioned ratio of 50.92 and 31.38% using the vertical gradient and the TOA reflectance methods, respectively. On the other hand, in the results from the 30 m resolution data, the ratios in Band 5 performed best, with ratios of 47.43 and 17.8%, respectively, using the same methods. This implies that Band 8 was better for tidal crack detection than the multispectral fusion data (Bands 1–7), and Band 5 with a resolution of 30 m was best among the multispectral data. The semi-automatic mapping of tidal cracks will improve the safety of vehicle travel in fast-ice regimes.


Sea ice as a record of freshwater fluxes: a comparison of isotopic composition near an Antarctic ice shelf and off the Alaskan coast

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

Corresponding author: Inga J. Smith

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

Sea ice formation is influenced by water masses containing fresh water from ice shelves in Antarctica, and from rivers in the Arctic. Supercooled ice shelf water can have dramatic effects on Antarctic sea ice microstructure through platelet ice formation, yet the ice-shelf-influenced water is only marginally less saline and the supercooling slight (of the order of 0.01 K). In Arctic Alaska, where large ice shelves are absent, land-fast sea ice growth can be influenced by fresher water from rivers and residual summer melt, with much larger inputs of fresh water possible. In this presentation, a method to reconstruct changes in water masses using oxygen isotope measurements from sea ice cores is applied, contrasting the Antarctic and Arctic sea ice. 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, USA, 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. Challenges in applying the method in both regions will be outlined, particularly quantification of changing ocean heat fluxes.


Unlocking controls on frontal ablation rates at Greenlandic tidewater glaciers

Mason Fried, Ginny Catania, Leigh Stearns, Timothy Bartholomaus, David Sutherland, Emily Shroyer, Jonathan Nash

Corresponding author: Mason Fried

Corresponding author e-mail: mason.j.fried@gmail.com

Rates of frontal ablation at tidewater glacier termini represent a significant uncertainty for sea-level rise forecasts. Understanding how frontal ablation rates respond to different forcing mechanisms – both ice dynamics and environmental factors – remains key to resolving this uncertainty. Here, at three tidewater glaciers in central West Greenland, we compare 3-year satellite-derived time series of frontal ablation with coincident high-temporal resolution in situ measurements of ocean and atmosphere conditions and remotely sensed ice dynamics and melange conditions in order to determine controls on frontal ice loss. Moorings deployed in adjacent proglacial fjords enable frontal ablation comparisons to ocean temperature at multiple depths. We find that, independent of glacier size and dynamics, frontal ablation rates are highly variable through time and dominated by isolated calving events rather than more gradual ice removal through submarine melting – although calving enhanced by melt driven undercutting may be an important contributor. Summertime increases in frontal ablation more closely follow atmospheric and near-surface ocean warming and subsequent melange break-up than ice speed or ocean temperature at depth. Finally, we explore the extent to which different forces dominate frontal ablation rates at different timescales in order to form a hierarchy of frontal ablation controls.


Towards the estimation of landfast sea-ice thickness using MODIS IST retrievals

Jonathan Everts, Greg Leonard, Inga Smith

Corresponding author: Jonathan Everts

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

Sea-ice thickness is an important property in the study of thermodynamic interactions between sea ice, ocean and atmosphere. On a local scale it can be directly measured by drilling into the ice, or inferred from upward-looking sonar or electromagnetic induction measurements. On larger spatial scales, sea-ice thickness is routinely estimated from spaceborne and airborne observations. The two most common techniques involve using altimeter (radar or laser) retrievals to estimate sea-ice freeboard, from which sea-ice thickness is derived, or using ice surface temperature (IST) retrievals coupled with a thermodynamic model to solve for ice thickness. Algorithms that derive sea-ice thickness from satellite retrievals allow for the routine monitoring of ice thickness distribution over large areas, but can be flawed due to their underlying physical assumptions. In this study we use in situ thermistor probe data to evaluate one such algorithm that uses MODIS IST retrievals to estimate sea-ice thickness. Over the past 10+ years, the sea-ice group at the University of Otago has routinely installed a thermistor probe into the landfast sea ice in McMurdo Sound, Antarctica during the growth season. The probe measures vertical profiles of sea-ice and near-surface ocean temperature at 10 minute intervals. These data are processed to determine the thickness and growth rate of the sea-ice cover. We compare these data with sea-ice thickness estimates derived from MODIS IST retrievals to evaluate the algorithm’s performance in McMurdo Sound. We find that for both the 2014 and 2015 growth seasons the algorithm underestimates sea-ice thickness, with the best agreement occurring early in the growth season when the sea-ice cover was relatively thin. Drivers for this behaviour are investigated, including thin ice responding more quickly to changes in air temperature, the presence of platelet ice at the site and snow-density changes with time.


Southern Ocean and Antarctic sea ice responses to fresh water from icebergs and ice-shelf basal melting in an Earth System model

Andrew G. Pauling, Cecilia M. Bitz, Inga J. Smith, Patricia J. Langhorne

Corresponding author: Andrew G. Pauling

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

We present results from an investigation into whether recent Antarctic sea-ice expansion resulted from an increase in fresh water reaching the Southern Ocean. Fresh-water fluxes from ice-sheet and ice-shelf mass imbalance is largely missing in models that participated in the Coupled Model Intercomparison Project Phase 5 (CMIP5). However, on average P-E reaching the Southern Ocean has increased in CMIP5 models to a present value that is about 2600 Gt per year greater than pre-industrial times and 5 to 22 times larger than estimates of the mass imbalance of Antarctic ice sheets and shelves (119–544 Gt per year). Two sets of experiments were conducted from 1980 to 2013 in CESM1(CAM5), one of the CMIP5 models, artificially distributing fresh water either at the ocean surface to mimic iceberg melt, or at the ice shelf fronts at depth to mimic basal melting. An anomalous reduction in vertical advection of heat into the surface mixed layer resulted in sea-surface cooling at high southern latitudes, and an associated increase in sea-ice area. Enhancing the freshwater input by an amount within the range of estimates of the Antarctic mass imbalance did not have any significant effect on either sea-ice area magnitude or trend. Fresh-water enhancement of 2000 Gt per year raised the total sea-ice area by 106 km2, yet this and even an enhancement of 3000 Gt per year was insufficient to offset the sea-ice decline due to anthropogenic forcing for any period of 20 years or longer. Further, the sea-ice response was found to be insensitive to the depth of freshwater injection.


Glider measurements: towards directly measuring supercooling beneath ice shelves

Monica J.S. Nelson, Bastien Y. Queste, Inga J. Smith, Gregory H. Leonard, Benjamin G.M. Webber, Karen J. Heywood

Corresponding author: Monica J.S. Nelson

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

Supercooled seawater is seawater cooled below its in situ freezing point temperature (in situ supercooled seawater) or below the surface freezing point temperature (potentially supercooled seawater). Traditionally conductivity–temperature–depth sensors (CTDs), or occasionally conductivity–temperature sensors (CTs) are used to measure supercooling. Due to logistical constraints, measurements of supercooling directly beneath ice shelves are rare. This investigation looked for supercooling in measurements taken from an underwater glider that travelled under the terminal face of the Ross Ice Shelf. The properties logged by the glider used here were: in situ temperature, electrical conductivity, sea pressure, depth, GPS location, and time. At the end of each dive, the glider should have surfaced to communicate with a satellite and accurately log its position using a GPS. However, in the first 100 dives, only 25 ascended to within 1 m of the surface, and in the first 30 dives the glider rose to a reasonably consistent depth, between 80 m and 95 m. This suggests it was under something obstructing its ascent. Comparing the Ross Ice Shelf thickness at the location of the glider’s deployment showed that these surfacing depths matched the ice shelf thickness. Hence, we concluded that the glider was under the ice shelf. For the first 30 dives, when the glider appears to have been stuck under thick ice, water was potentially supercooled to varying levels. No in situ supercooling was found during this period as the depression of the freezing point temperature, due to the pressure effect, was too great.


Multiyear temperature profiles for Amery Ice Shelf and sub-ice-shelf ocean

Roland Warner, Mike Craven, Stefan Vogel, Benjamin Galton-Fenzi, Alan Elcheikh, Adam Christensen, Adam Treverrow, Shavawn Donogue

Corresponding author: Roland Warner

Corresponding author e-mail: Roland.Warner@utas.edu.au

Antarctic ice shelves are coupled to the climate of the Southern Ocean by sub-ice ocean circulation, with interactions ranging from substantial basal melting to the accretion of thick layers of marine ice. They are vulnerable to increased melting from a warming ocean and from changes in ocean currents. Hidden beneath kilometre-thick ice, sub-ice-shelf processes are difficult to study. Optical fibre cables were installed through the Amery Ice Shelf and into the underlying ocean by the Australian AMISOR project during the austral summer of 2009/10 via two hot-water-drilled boreholes. The cables extend beneath the ice shelf to near the ocean floor and permit the measurement of temperature profiles at metre-scale resolution by detection of Raman inelastic back-scattered laser light, using a Sensornet Oryx distributed temperature sensing (DTS) system. The temperature profiles provide detailed records inside the body of the ice shelf, yielding insights into the vertical structure of the ice shelf in both a basal freeze zone (AM05), where the lower part of the ice column consists of accreted marine ice platelets, and a basal melt zone (AM06) with meteoric ice only. Temperature records from the ocean cavity show good agreement with oceanographic instruments fixed on the same moorings, and provide a time series of temperature profiles through several hundred metres of the water column. Knowledge of the temperature regimes within the ice shelf and in the underlying ocean is important for the study of ice–ocean interactions, ice-shelf mass budget and the dynamics of ice-shelf flow. Measurements of ocean temperature profiles throughout the seasonal cycle provide insights into the sub-shelf ocean circulation, heat transport, and basal melting and freezing. The challenges of field conditions and logistics interrupt our records at times. We present an overview of the results gathered over the past 6 years and outline future prospects.


Slow-down of bottom water formation and ventilation after calving of the Mertz Glacier Tongue

Stephen Rintoul, Kate Snow, Bernadette Sloyan, Andrew Hogg

Corresponding author: Stephen Rintoul

Corresponding author e-mail: steve.rintoul@csiro.au

Antarctic bottom water supplies the deep limb of the global overturning circulation, ventilates the abyss and transports heat, nutrients and carbon. Several studies have highlighted surprisingly rapid rates of change in bottom water properties in recent decades. However, interpretation of change observed in the deep ocean basins is hampered by incomplete understanding of the physical processes driving variations in the properties and formation of bottom water. We use the natural experiment initiated by calving of the Mertz Glacier Tongue in 2010 to investigate the sensitivity of bottom water formation to changes in forcing. A 7-year moored time series of temperature and salinity at the sill through which dense water is exported from the Mertz polynya provides a unique perspective on the response to the calving event. Concurrent hydrographic measurements in the polynya, the sill and the deep ocean allow an assessment of the impact of the calving on shelf circulation, ocean–ice-shelf interaction, and bottom water production and ventilation. Calving of the glacier tongue led to a dramatic shift in the regional icescape. Loss of the ice tongue allowed more ice to advect into the region and grounding of the large iceberg B9B caused the polynya region to back-fill with sea ice. Both factors contributed to a sharp reduction in the area and sea ice production of the Mertz polynya, and hence the salinity and density of high-salinity shelf water produced in the polynya. At the sill, the density, thickness and duration of dense water export was sharply reduced. Measurements in the deep ocean downstream of the sill confirm that the changes in export of dense shelf water produced significant changes in Antarctic bottom water. The density, thickness and oxygen content of the bottom water declined after calving, indicating a reduction in bottom water formation and ventilation of the abyssal ocean. Ice shelf water (ISW) produced by glacial melt at depth is observed at the sill after the calving event. Measurements on the shelf show that the ISW is thicker, colder and more widespread after the calving event. These observations suggest that the reduction in polynya activity may have increased the heat available for basal melt after calving, as inferred from modelling studies. These results highlight the sensitivity of bottom water formation, abyssal ventilation and ocean–ice-shelf interaction to changes in the regional icescape.


An image mosaic of Greenland based on Landsat8

Zhuoqi Chen, Xiao Cheng, Fengming Hui

Corresponding author: Xiao Cheng

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

The first Landsat Image Mosaic of Greenland (LIMG) was released by the Greenland Ice Mapping Project (GIMP). This mosaic was created from a combination of Landsat-7 and RADARSAT-1 imagery acquired between 1999 and 2002. This mosaic is a dataset for supporting scientific research in Greenland. However, some scientific researches need at least one another mosaic in this region (e.g. detecting changes of snow cover). A latest image mosaic for Greenland may fit to meet these scientific goals. As the latest satellite of Landsat mission, the Landsat-8 images cover the entire Earth every 16 days. Data collected by the instruments onboard the satellite is available to download at no charge within 24 hours of reception. The standard Landsat-8 products provided by the USGS EROS Center consist of quantiified and calibrated scaled digital. With the support of the USGS portal, a 30 m image mosaic of Greenland is created using 108 multi-band scenes of Landsat-8 OLI data acquired during June–August 2014 and 2015. First, we calculated solar elevation angle for each pixel individually. Second, solar elevation angle is used to calculate planetary reflectance. Planetary reflectance removes the cosine effect caused by different solar zenith angles due to different data acquisition time and compensates for the differences of exoatmospheric solar irradiance across different spectral bands. Finally, all multispectral (Bands 2–4) OLI scenes at 30 m resolution were used to create the mosaic.


Deformation of meteoric and marine ice at the southern McMurdo Ice Shelf, Antarctica

Inka Koch, Sean Fitzsimons, Nicolas Cullen, Adam Treverrow, Jean-Louis Tison

Corresponding author: Inka Koch

Corresponding author e-mail: inkakoch@gmail.com

Ice shelves, which are buttressing glaciers and ice streams, are primarily composed of meteoric ice advected from inland and locally formed marine ice. The latter forms from a mixture of ocean water and fresh water and accretes at the ice shelf base, where it can fill in structurally weak zones, but it remains unknown whether the presence of marine ice influences ice-shelf deformation and stability. In this study, marine ice crystal properties (crystal shape, size and ice fabric) are investigated and compared to meteoric ice crystal properties to help understand whether marine ice deforms differently to meteoric ice in a similar strain setting. Two short ice cores of marine and meteoric ice were extracted from the surface of the southern McMurdo Ice Shelf (SMIS), where local ablation allows for surfacing of basal ice shelf layers. Here SMIS flows toward shore in the order of only a few meters per year. Thus, marine and meteoric ice are exposed to a small horizontally compressive strain rate of around 3.0 × 10–4 a–1 and a shear strain rate that is less than half of that. Both ice types show adjustment to a secondary strain regime, keeping some of their original ice-formation properties. Despite different original accretion processes in meteoric and marine ice leading to isotropic and anisotropic fabrics respectively, crystals from both ice types show a progressive development of girdle fabrics with increased compression down the flowline, whereby this ice fabric is more pronounced in meteoric ice. Adjustment to the predominantly compressive strain regime is corroborated by the development of slightly flattened ice crystals in meteoric ice and the rotation of the originally elongated ice crystals in marine ice together with the presence of smaller granular ice crystals as deformation continues. Results from this study indicate that marine ice does not deform more easily than meteoric ice in a strain setting of primarily horizontal uniaxial compression, since marine ice largely retains original accretion characteristics, such as a weaker anisotropic fabric and elongated ice crystals. However, the SMIS strain setting of primarily lateral compression is only typical for ice-shelf pinning points and margins and further in situ studies in a typical strain setting of horizontal extension are necessary to completely understand the role of marine ice in regulating ice-shelf stability.


Decadal-scale response of the Antarctic Ice Sheet to a warming ocean using the POPSICLES coupled ice sheet-ocean model

Daniel Martin, Xylar Asay-Davis, Stephen Cornford, Stephen Price, Esmond Ng

Corresponding author: Daniel Martin

Corresponding author e-mail: DFMartin@lbl.gov

We present an improved POPSICLES simulation of a 20-year coupled evolution of the full Antarctic Ice Sheet and the Southern Ocean. We use the CORE v. 2 Normal Year forcing data to force the ocean model and tune the initial state of the ice sheet to match present-day observations. Simulations are performed at 0.1° (~5 km) ocean resolution with adaptive ice sheet-resolution as fine as 500 m to adequately resolve the grounding-line dynamics. We discuss the effect of improved ocean mixing and subshelf bathymetry (vs the standard Bedmap2 bathymetry) on the behavior of the coupled system, comparing time-averaged melt rates below a number of major ice shelves with those reported in the literature. We also present seasonal variability and decadal melting trends from several Antarctic regions, along with the response of the ice shelves and the consequent dynamic response of the grounded ice sheet with its implications for incipient marine ice-sheet instability and potential marine-driven collapse. POPSICLES couples the POP2x ocean model, a modified version of the Parallel Ocean Program, and the BISICLES ice-sheet model. POP2x includes sub-ice-shelf circulation using partial top cells and the commonly used three-equation boundary layer physics. Standalone POP2x output compares well with standard ice–ocean test cases (e.g. ISOMIP) and other continental-scale simulations and melt-rate observations. BISICLES makes use of adaptive mesh refinement and a first-order accurate momentum balance similar to the L1L2 model of Schoof and Hindmarsh to accurately model regions of dynamic complexity, such as ice streams, outlet glaciers and grounding lines. Results of BISICLES simulations have compared favorably to similar simulations with a Stokes momentum balance in both idealized tests (MISMIP-3d) and realistic configurations.


An ocean box model to simulate the overturning circulation underneath Antarctic ice shelves.

Ronja Reese, Torsten Albrecht, Ricarda Winkelmann

Corresponding author: Ronja Reese

Corresponding author e-mail: ronja.reese@pik-potsdam.de

Sub-shelf melting is an important component of Antarctica’s mass budget. Although thinning of ice shelves does not directly contribute to sea-level rise, it may have a significant indirect impact through the potential of ice shelves to buttress their adjacent ice sheet. Sub-shelf melt rates depend on temperatures and salinity of the surrounding ocean water. They range from less than 1 m a–1 below the Filchner–Ronne and Ross Ice Shelf to up to ~30 m a–1 below Pine Island Glacier. Accessing the ocean water underneath the ice shelf is challenging, so measurements below Antarctic ice shelves are rare and observational data difficult to obtain. A vertical ocean circulation in a sub-shelf cavity transports ocean water to the grounding line where the ice shelf is melted from below. The so-called ice pump then transports the water along the shelf base towards the calving front where refreezing of the ocean water to the ice shelf can occur. The ocean box model in the Parallel Ice Sheet Model (PISM), a hybrid-type SIA/SSA model for Greenland and Antarctica, aims to simulate this overturning circulation underneath ice shelves. It is implemented with two boxes adjacent to the ice shelf and a reservoir box including the temperatures and salinity of the Southern Ocean. It thus allows Antarctic sub-shelf melt rates to be computed from observations and projections for the Southern Ocean north of the ice shelves. Within the MISOMIP intercomparison project results from the coarse ocean box model are compared to more sophisticated coupled ice–ocean model simulations. Some Antarctic ice shelves are reached by the relatively warm circumpolar deep water, leading to high melt rates, e.g. in the Amundsen region. Changes in atmospheric patterns due to global warming might allow warm water to enter through a sub-shelf trough into the Filchner–Ronne Ice Shelf cavity by the end of the 21st century, leading to an increase in sub-shelf melt rates. With PISM and the ocean box model, we aim at estimating an upper limit of ice loss through sub-shelf melting changes due to changes of circulation patterns of the Southern Ocean.


Temporal variability in melting at the base of the Ronne–Filchner Ice Shelf, Antarctica

Keith Nicholls, Keith Makinson, Svein Østerhus, Lai Bun Lok, Paul Brennan

Corresponding author: Keith Nicholls

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

Ice-shelf basal melt-rates are usually quoted as annual average values, and their calculation requires assumptions that can be difficult to justify. The advent of ground-based phase-sensitive radar has allowed the direct measurement of basal melt rates, removing the need for the assumptions inherent in some other methods, and offering the possibility of obtaining time series of melt rates, to see how ice-shelf basal mass balance responds to its oceanographic forcing. Here we present a year-long time series of basal melt rates from a site on the Ronne Ice Shelf. A phase-sensitive radar (ApRES) was deployed at the site from January 2015 to January 2016, which recorded hourly measurements of the local ice thickness. It also recorded the change in depth of reflecting horizons within the ice column, which allows the effect of vertical strain to be removed from the record, and enables us to calculate the basal melt rate from the thickness record. The basal melt rate exhibits strong variability on timescales from tidal to seasonal. The smoothed melt rate increases from around 0.8 m a–1 to a maximum of nearly 2 m a–1 by March 2015; from June it slowly reduces, to less than 0.5 m a–1 in November. The melt rate does not return to its initial value of 1 m a–1 by the end of the record in January 2016. A strong tidal component of amplitude of up to 1.5 m a–1 is superimposed on the record, resulting in regular periods of basal freezing, especially in the last few months of the record. We also present contemporaneous oceanographic measurements from a mooring deployed beneath the ice shelf at the same site. The combination of the two datasets offers a rare insight into the relationship between the oceanographic forcing and the ice shelf’s basal mass balance.


RTopo-2: A global high-resolution dataset of ice-sheet topography, ice-shelf cavity geometry and ocean bathymetry

Janin Schaffer, Ralph Timmermann

Corresponding author: Janin Schaffer

Corresponding author e-mail: Ralph.Timmermann@awi.de

The RTopo-1 data set of Antarctic ice sheet/shelf geometry and global ocean bathymetry has proved useful not only for modelling studies of ice–ocean interaction in the southern hemisphere. Following the spirit of this data set, we introduce a new product (RTopo-2) that contains consistent maps of global ocean bathymetry, upper and lower ice-surface topographies for Greenland and Antarctica, and global surface height on a spherical grid with now 30 arcsec resolution. We used the General Bathymetric Chart of the Oceans (GEBCO_2014) as the backbone and added the International Bathymetric Chart of the Arctic Ocean version 3 (IBCAOv3) and the International Bathymetric Chart of the Southern Ocean (IBCSO) version 1. To achieve a good representation of the fjord and shelf bathymetries around the Greenland continent, we corrected data from earlier gridded products in the areas of Petermann Glacier, Hagen Bræ and Helheim Glacier assuming that sub-ice and fjord bathymetries roughly follow plausible Last Glacial Maximum ice-flow patterns. For the continental shelf off northeast Greenland and the floating ice tongue of Nioghalvfjerdsfjorden Glacier at about 79°N, we incorporated a high-resolution digital bathymetry model considering original multibeam survey data for the region. Radar data for ice-surface and ice-base topographies of the floating ice tongues of Nioghalvfjerdsfjorden Glacier and Zachariæ Isstrøm have been obtained from the data centers of the Technical University of Denmark (DTU), Operation Icebridge (NASA/NSF) and the Alfred Wegener Institute (AWI). For the Antarctic ice sheet/ice shelves, RTopo-2 largely relies on the Bedmap-2 product but applies corrections for the geometry of Getz, Abbot and Fimbul ice-shelf cavities. The data set is available in full and in regional subsets in NetCDF format from the PANGAEA database.


Ice-flow variation in Larsen B ice shelf interpreted from repeated Landsat images, 2005–15

Gang Qiao, Song Guo

Corresponding author: Gang Qiao

Corresponding author e-mail: qiaogang@tongji.edu.cn

As a basic characteristic of glaciers and ice sheets, ice flow plays an important role in the Antarctic ice-sheet mass balance, which is of important significance for understanding the stability of the Antarctic ice shelves. Recent studies have shown that the northern ice shelves on the Antarctic Peninsula are extraordinarily sensitive to temperature changes, and are thus an important indicator of global warming. During the past 20 years, the Antarctic Peninsula has experienced several large-scale ice-shelf disintegration events, and temperature rise is considered to be one of the main causes. Long-term observations of ice flow in this region are essential for further investigation of the mechanism of disintegration of ice shelves. However, existing studies are mainly focused on the period of 1995–2002, when the ice shelves disintegrated in this region, with very little research since. In this paper, the ice-flow variation in Larsen B ice shelf in the northern Antarctic Peninsula is analyzed based on repeated Landsat optical satellite images from 2005–15. Optical images are widely used for ice-flow interpretation in Antarctic, and a feature-tracking method is commonly employed for ice-flow calculation. In this research, based on an automatic feature-matching technique, a long-term series of surface ice-flow measurements was extracted from Landsat images of Larsen B ice shelf. The Landsat images employed in this research span 2005–15, and the time interval between two adjacent images is about a year. Gap-filling and image enhancement are first conducted to reduce the image noise for fewer image mismatches, followed by filtering of the matching result and accuracy assessment. The resultant ice-flow series is then compared with existing Antarctic ice-flow products. The results show that from 2005–15, the ice-flow velocity in Larsen B ice shelf increased by about 32%, from 520 m a–1 to 690 m a–1. The peak was reached in 2013, and the annual average ice-flow speed decreased by 100 m a–1 in 2014, followed by a relative stabilization in 2015. Meanwhile, Larsen B ice shelf endured continuous collapses from 2005–08, and there has been slow extension of the ice shelf since 2008. Further detailed analysis of the ice-flow changes is being conducted and possible reasons for the variation are being sought.


Towards a regional coupled ice-sheet–ocean model for Antarctica

Ralph Timmermann, Sebastian Goeller

Corresponding author: Ralph Timmermann

Corresponding author e-mail: Ralph.Timmermann@awi.de

To study the interaction between the world ocean and the Antarctic ice sheet, a Regional Antarctic and Global Ocean (RAnGO) model has been developed. The coupled model is based on a global implementation of the Finite Element Sea-ice Ocean Model (FESOM) with a mesh that is substantially refined in the Southern Ocean, particularly in its marginal seas and in the sub-ice-shelf cavities. The Antarctic cryosphere is represented by a regional setup of the ice flow model RIMBAY. As a first step, the RIMBAY domain comprises the Filchner–Ronne Ice Shelf and the grounded ice in its catchment area up to the ice divides. At the base of the RIMBAY ice shelf, melt rates from FESOM’s ice-shelf component are prescribed. RIMBAY returns ice thickness and the position of the grounding line. To adjust the FESOM domain to varying cavity geometries, we use a pre-computed mesh that comprises the present-day ocean plus areas that may potentially become ungrounded. For each coupling step, i.e. once per year, the coupler determines the area covered by ocean and removes grid nodes that are covered by grounded ice. Changes in water-column thickness are easily handled by the terrain-following vertical coordinate system in the sub-ice cavity. Model runs with a 20th-century climate forcing yield a quasi-stable grounding line position close to the presently observed state with only small fluctuations. In a centennial-scale warm-water-inflow scenario, the model suggests a gradual retreat of the grounding line, mainly at Institute, Foundation and Support Force Ice Streams. A more dramatic response is prevented by the steep topography under most of the currently grounded ice.


Iceberg motion and melt from observational and laboratory studies

Anna FitzMaurice, Fiammetta Straneo, Claudia Cenedese, Magdalena Andres, Alessandro Silvano

Corresponding author: Anna FitzMaurice

Corresponding author e-mail: apf@princeton.edu

Icebergs account for approximately half of the mass flux from the Greenland ice sheet, and consequently constitute a leading order component of the ice sheet’s freshwater flux to the North Atlantic. Due to a lack of observational data, the distribution, trajectories and melt of icebergs around Greenland are currently poorly constrained. We analyze acoustic data from Sermilik Fjord in southeastern Greenland to obtain a 1-year record of iceberg speed and draft in the fjord. This is used in conjunction with co-located ocean current and wind velocity data to study the dominant balances controlling iceberg motion. Next, we use the results from the data analysis to inform laboratory experiments of iceberg melting that seek to guide the improvement of parameterizations of melt flux from icebergs in ocean and climate models. In the context of the currently increasing discharge of icebergs from both poles, accurate modelling of this melt flux is an important component of assessing the potential impacts of future climate change.


Transport pathways and consequences for Antarctic ice-shelf basal meltwater

Michael Dinniman, John Klinck, Laurence Padman

Corresponding author: Michael Dinniman

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

Oceanic melting of the base of the floating Antarctic ice shelves is the dominant cause of mass loss for the Antarctic ice sheet. Beyond the direct effects of melting on ice-sheet mass balance, the flux of basal meltwater into the coastal ocean has been proposed to impact important climate processes such as Antarctic bottom water (AABW) formation, Southern Ocean sea ice extent and the delivery of a significant fraction of the limiting micronutrient dissolved iron to the euphotic zone in the extremely productive coastal Antarctic waters. We study generation of, and subsequent pathways for, basal meltwater with a circum-Antarctic ocean/sea-ice/ice-shelf model. The model was initially run with 10 km horizontal grid spacing, but reducing this to 5 km increases the total ice shelf melt by ~20% and improves the simulation of basal melt in several locations, especially ‘warm-water’ shelves in the Amundsen and Bellingshausen seas. Eight independent simulated tracers are used to examine regional differences in the spread of meltwater. Separate tracer release simulations were performed to study seasonal changes in meltwater fluxes. The results confirm the previously reported idea that transport of meltwater from the Amundsen ice shelves to the Ross Sea is significant, especially with respect to possible changes in Ross Ice Shelf basal melt and AABW formation. Meltwater from ice shelves in the Amundsen and Bellingshausen seas is readily transported into the wider Southern Ocean due to the proximity of the Antarctic circumpolar current to the continental shelf. Much of the meltwater reaching the summer coastal polynyas over certain areas of the continental shelf is from non-local sources and there is a strong seasonal variability in the meltwater production. Both of these factors will need to be accounted for in trying to assess the impact of ice shelf melt on phytoplankton nutrient supply.


Sensitivity of ice sheets to ocean forcing over decadal to millennial timescales: the role of isostatic adjustment

Jeremy Bassis, Mac Cathles, Elizabeth Ultee, Yue Ma, Ryan Whitcomb

Corresponding author: Jeremy Bassis

Corresponding author e-mail: jbassis@umich.edu

Observations show that portions of ice sheets in contact with the ocean respond to climate forcing nonlinearly with periods of slow or negligible advance punctuated by abrupt, rapid and often irreversible retreat of the grounding line or calving front. For grounded marine-terminating glaciers, change takes the form of increased discharge and retreat of the calving front. Where ice tongues or ice shelves are present, the response occurs as grounding-line migration and dynamic thinning. In this study we use a range of models to examine the sensitivity of grounded and floating margins to ocean forcing. Over short (decadal) timescales, we find that relatively small increases in ocean forcing can trigger catastrophic retreat for glaciers grounded deep beneath sea level. Under some conditions calving or grounding-line retreat can progress until the glacier stabilizes on a pinning point or constriction in the channel. In the absence of pinning points or constrictions, retreat can lead to a runaway instability, as postulated for portions of the West Antarctic Ice Sheet. However, over longer timescales, isostatic adjustment of the bed becomes significant and adjustment of the bed can even allow the calving front/grounding line to re-advance. We show that the combination of ocean forcing and isostatic adjustment is consistent with the magnitude and timing of Heinrich events, episodic discharges of large volumes of icebergs from the Laurentide Ice Sheet that occurred when surface temperatures were cold. The West Antarctic Ice Sheet, unlike Hudson Bay, is thought to overlie a much weaker, lower-viscosity mantle. We show that the presence of a lower-viscosity upper mantle or asthenosphere significantly increases the stability of Pine Island and Thwaites-like catchment basins over century and shorter timescales. Our results indicate that ocean forcing is an important factor driving ice-sheet retreat, but also suggest that the solid Earth plays a crucial role in regulating ice-sheet stability that should not be ignored by models seeking to assess the stability of ice sheets.


Estimates of ice melange submarine melt rates in Nuup Kangerlua Fjord, southwest Greenland, from satellite imagery

Alexis Moyer, Peter Nienow, Noel Gourmelen, Donald Slater

Corresponding author: Alexis Moyer

Corresponding author e-mail: moyeal03@gmail.com

The accelerating mass loss from the Greenland Ice Sheet is attributed in part to changes in the dynamics of tidewater glaciers. Oceanic forcing of the ice-sheet margin is believed to have promoted recent widespread thinning at tidewater glaciers, leading to increased glacier velocities, calving and retreat. The precise mechanism(s) driving this tidewater glacier instability are poorly understood and, while increasing evidence points to the importance of submarine melt-rates (SMR), estimates of SMR remain uncertain. Estimating submarine melting of tidewater glaciers has been limited by the challenges of collecting in situ measurements in the dynamic and inhospitable glacier–fjord environment. In this study, we present a new approach to estimating SMR, by examining elevation changes in the thick ice melange often present in the fjord adjacent to tidewater glacier termini. At Kangiata Nunaata Sermia (KNS) glacier in Nuup Kangerlua Fjord, southwest Greenland, we derive submarine ice melt from high-resolution (2.5 m) digital elevation models (DEMs) of ice melange made from interferometry applied to TanDEM-X Synthetic Aperture Radar imagery. Along- and across-fjord elevation transects, extending up to 5 km from the calving front, allow us to evaluate spatial and temporal variations in SMR during the winter and spring. Ice melange elevation decreases linearly with distance seaward from the terminus of KNS but with considerable interannual variation in surface gradient (and thus melt-rate). In addition, our estimates of SMR show reasonable agreement with those derived using buoyant plume theory.


High-resolution basal topography and sediment properties beneath Pine Island Ice Shelf from autonomous underwater vehicle surveys

Damon Davies, Robert, G. Bingham, Alastair G.C. Graham, Matteo Spagnolo, Pierre Dutrieux, Adrian Jenkins

Corresponding author: Damon Davies

Corresponding author e-mail: D.Davies@ed.ac.uk

Ice shelves play a crucial role in the sensitivity of West Antarctic ice streams by providing a buffer between the ocean and inland ice. Recent observations in the Amundsen Sea region and, in particular, Pine Island Glacier, have shown rapid thinning of its ice shelf leading to acceleration and grounding-line retreat. Bathymetry beneath Pine Island Ice Shelf controls the transport of warm ocean waters that are currently eroding the ice shelf from beneath and also provides topographic controls on grounding-line retreat. Here we present high-resolution seafloor bathymetry and sub-bottom profiler data obtained by autonomous underwater vehicle surveys in 2009 and 2014. Using these data we explore the geomorphological signature of the unpinning of the grounding line from a submarine ridge, and assess sediment distribution to gain insights into grounding-line retreat processes.


Ice-penetrating radar evidence for Late-Holocene ice-shelf grounding and ice-sheet expansion in the Ronne Ice Shelf, West Antarctica

Jonathan Kingslake, Edward C. King

Corresponding author: Jonathan Kingslake

Corresponding author e-mail: jkingslake@gmail.com

To understand the retreat of the West Antarctic Ice Sheet we must determine how ice flow changes over timescales longer than the decades since the start of the satellite era. Here we provide ice-penetrating radar evidence for Late-Holocene changes in the flow of the Ronne Ice Shelf, West Antarctica. The Henry Ice Rise (HIR) is a 7000 km2 area of 350–940 m thick, slow-moving (<10 m a–1) ice, grounded 250–850 m below sea level and surrounded by the Ronne Ice Shelf. Previous observations of glacio-isostatic adjustment, englacial structure and surface streak lines indicate Late-Holocene ice-flow reorganization in the region to the south of HIR. During a 700 km ground-based ice-penetrating radar survey of HIR, we observed a collection of englacial structures that cannot have formed under present flow conditions, including (1) undulating isochronal layers that intercept the bed in ice too thin to sustain a warm ice-sheet base, (2) steeply dipping basal crevasses in currently slow-moving ice and (3) under-developed Raymond Arches (isochronal layer structures that grow beneath ice divides due to nonlinear ice flow), offset from an ice divide. We argue that these structures can be explained if ice in this location was formerly afloat and later grounded on the seabed to form HIR. A large area of basal crevassing overlies a basal topographic high, where the ice shelf would have first grounded. Parallel to and upstream of HIR’s current grounding line, the boundaries of basally crevassed regions coincide with isochronal-layer troughs and with contrasts in the kilometer-scale ice-surface texture visible on satellite imagery. We suggest that these correspond to past grounding-line configurations. The offset Raymond Arches are consistent with eastward ice-divide migration caused by ongoing outward grounding-line migration. We discuss possible causes of this ice-sheet expansion and consider whether it triggered the flow reorganization indicated by previous observations. These events are generally not included in ice-sheet reconstructions because we have few constraints on their timing. Hence ice-sheet models are being tuned to simulate monotonic Late-Holocene retreat, while the reality was more complex. The impact of this on the reliability of ice-sheet model predictions is unknown. Further work is required to date the grounding of HIR, investigate connections to other flow reorganizations and incorporate these events in ice-sheet reconstructions.


Surprise! Big changes in Seller and Prospect Glaciers in the decades after the collapse of Wordie Ice Shelf

Catherine Walker, Alex Gardner

Corresponding author: Catherine Walker

Corresponding author e-mail: cat.walker@eas.gatech.edu

In recent decades, most Antarctic Peninsula ice shelves have retreated or collapsed completely, leading to associated tributary glacier acceleration in the absence of buttressing forces. One of these major losses was the collapse of the Wordie Ice Shelf (WIS) around 1989. In this study, we use several observational datasets to show that WIS tributaries (Seller and Prospect) have experienced significant acceleration (to ~3 km a–1) beginning around 2008, nearly 20 years after the collapse. During the same period, Icebridge altimetry shows that both glaciers experienced a drawdown at their calving fronts of ~8 km a–1. This represents a near-doubling in rate of elevation change from the 1990s/early-2000s. We show that it is likely that the large loss signal observed in GRACE data in the late 2000s is related to these changes in Wordie Bay. To determine the cause of these changes and their relationship to external forcing, we utilize Palmer Station Long-Term Ecological Research (PAL LTER) ocean CTD-gridded observations, a dataset that represents the longest-lived observations in the Southern Ocean, taken along the coast and continental shelf break along the west Antarctic Peninsula (wAP) beginning in 1993. We also use automatic weather sation data to illustrate changes in atmospheric temperature and increasing positive degree days in the region. Ice-shelf floating-area changes are documented using LandSat and other visible imagery sources for ~50 glacier systems (including WIS and its tributaries) along the wAP from ~1945 to the present. Surface/structural changes in the WIS system, including occurrence of melt ponds, sea/fast-ice presence and crevasse density/orientation, are also documented. In this study, we utilize all these observational constraints to provide a detailed analysis of the big changes observed in the former tributaries of the WIS, specifically focusing on our investigation into the cause and consequences of the significant speed-up and draw-down that began in the late 2000s, which cannot be explained by the loss of ice-shelf buttressing. Our preliminary results suggest that the changes are due to increased upwelling of warm, salty upper circumpolar deep water (UCDW) in the region and frequent incursions of UCDW into Marguerite Bay, former home of WIS, due to the relatively deep Marguerite Trough that connects the bay to the continental shelf break and southernmost boundary of the Antarctic Circumpolar Current.


The transition from order to disorder in Greenland’s glaciers

Catherine Walker, Jeremy Bassis, Britney Schmidt, Joshua Hedgepeth

Corresponding author: Catherine Walker

Corresponding author e-mail: cat.walker@eas.gatech.edu

Iceberg calving is a major process involved in the removal of large volumes of ice almost instantaneously. Because calving is a rapid event, most often what gets captured via remote observation is the before and after rather than the exact moment of failure. Presented here are preliminary results that aim to constrain the relationship between these oft-observed pre- and post-collapse ice states characterizing the transition between order (surface crevasse fractal patterns) and disorder (melange as a rubble pile). This is a technique that in the past has been employed for the most part in civil and weapons engineering. In combination with Icebridge ATM (elevation/change) and MCoRDS (thickness) measurements over our target glaciers (Helheim, Rink, Jakobshavn and Kangerdlugssuaq), we present two modeling approaches: first, a fractal model to describe stress state and distribution within the intact ice; and second, a collapse model (fragmentation analysis) that takes as input the resulting state of a given broken material (in this case, melange) to determine its properties prior to collapse. Characterizing fracture density at different locations within the intact glacier and quantifying a critical factor to describe the transition from highly fractured to collapse/fragmentation region helps understand the distribution of observable surface fracture patterns and underlying differences within the target glaciers. We use a statistical approach to investigate whether or not there exists a quantifiable critical fracture density or ‘critical mass’ of closely spaced crevasses, beyond which dynamic fragmentation of the glacier occurs. By studying the size distribution of fragments in floating proglacial melange, we seek to back out physical properties of the pre-collapsed ice, such as material strength and the energy necessary to create a fragmentation/collapse event. While many studies have focused on the propagation of crevasses in the glacier ice as a means of predicting calving, the main goal of this work is to consider the physical transition between the two phases of collapse. The results of this application yield insight into the lead-up processes involved in collapse and enable us to determine some critical parameters (e.g. critical fracture density) antecedent to collapse events, and to characterize differences between glaciers.


Using radar to determine the mechanical effect of tides on an ice shelf

Keith Makinson, Keith Nicholls, Lai Bun Lok, Paul Brennan

Corresponding author: Keith Makinson

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

Antarctic ice shelves are known to restrain the flow of ice from the continental interior and understanding how they respond to changes in forcing requires models with adequate physics or good parameterizations of the processes that drive their evolution. For reliable predictions, therefore, numerical ice-shelf models must accurately determine the ice flow using suitable rheological parameters. The coupling between ice shelves and ocean creates a unique environment that, unlike the rest of the ice sheet, is exposed to highly dynamic ocean tidal forcing where the ice shelves can rise and fall twice per day by up to 7 m in some grounding line regions. When an ice shelf is tilted, it responds elastically, resulting in the observed diurnal and semidiurnal horizontal flow variations, which are amplified toward the ice-shelf front. Similarly, the ice shelf should also periodically thin and thicken at tidal timescales. Using the recently developed BAS-UCL Autonomous Phase-sensitive Radio Echo Sounder (ApRES), year-round measurements of vertical strain rates are now also possible at tidal timescales. The combined availability at Site 5 on Ronne Ice Shelf of ApRES-derived vertical strain rates and GPS-derived horizontal strain rates enables the key bulk ice-shelf properties, namely Young’s modulus and Poisson’s ratio, to be better constrained as it responds elastically at tidal forcing frequencies. This will enable ice-shelf models to better predict future ice-shelf and ice-sheet evolution.


Mass balance and runoff of glaciers in the Kongsfjord basin, northwestern Svalbard

Ankit Pramanik, Ward van Pelt, Jack Kohler

Corresponding author: Ankit Pramanik

Corresponding author e-mail: ankit.pramanik@npolar.no

Glaciers and ice caps cover 36 000 km2 or 60% of the land area of the Svalbard archipelago. Roughly 60% of the glaciated area drains to the ocean through tidewater glacier fronts. Runoff from tidewater glaciers is posited to have a significant impact on fjord circulation and thus on fjord ecosystems. Ocean circulation modelling under way in the Kongsfjord system requires specification of the freshwater amounts contributed by both tidewater and land-terminating glaciers in its basin. The total basin area of Kongsfjord is ~1400 km2. We use a coupled surface energy-balance and firn model to calculate mass balance and runoff from the Kongsfjord glaciers for the period 1969–2015. Meteorological data from the nearby station at Ny-Ålesund is used for climate forcing in the model domain, with mass-balance data at four glaciers in the Kongsfjord watershed used to calibrate model parameters. Precipitation and temperature lapse rates are adjusted on the study glaciers through repeated model runs at mass balance stake locations to match observed and modelled surface mass balance. Long-term discharge measurement at two sites in this region are used to validate the modelled runoff. Spatial and temporal evolution of melt, refreezing and runoff are analyzed, along with the vertical evolution of subsurface conditions.


Bed topography of Jakobshavn Isbræ, Greenland, from high-resolution gravity data

Lu An, Eric Rignot, Mathieu Morlighen, David Holland

Corresponding author: Lu An

Corresponding author e-mail: lan3@uci.edu

Jakobshavn Isbræ (JKS) is one of the largest marine-terminating outlet glaciers in Greenland, feeding a fjord about 800 m deep on the west coast. JKS has sped up more than twofold since 2002 and contributed nearly 1 mm of global sea-level rise during the period 2000–11. It has been posited that these changes coincided with a change in ocean conditions beneath the former ice tongue, yet little is known about the depth of the glacier at its grounding line and its upstream. The sea-floor depth of the fjord is not well known either. Here, we present a new approach to infer the glacier-bed topography, ice thickness and sea-floor bathymetry near the grounding line of JKS using high-resolution airborne gravity data from AIRGrav. AIRGrav data were collected in August 2012 from a helicopter platform. The data combined with radio-echo sounding data, discrete point soundings in the fjord and the mass conservation approach on land ice. AIRGrav acquired a 500 m spacing grid of free-air gravity data at 50 knots with sub-milligal accuracy, i.e. much higher than NASA Operation IceBridge (OIB)’s 5.2 km resolution at 290 knots. We use a three-dimensional inversion of the gravity data combining our observations and a forward modeling of the surrounding gravity field, and constrained at the boundary by radar echo soundings and points bathymetry measurements. We reconstruct seamless bed topography at the grounding line that matches interior data and the sea-floor bathymetry. The results reveal the true depth at the elbow of the terminal valley and the bed reversal in the proximity of the current grounding line. The analysis provides guidelines for future gravity survey of narrow fjords in terms of spatial resolution and gravity precision. The results also demonstrate the practicality of using high-resolution gravity survey to resolve bed topography near glacier snouts, in places where radar sounding has been significantly challenged in the past. The inversion results are critical to re-interpret the recent evolution of JKS and reduce uncertainties in projecting its future contribution to sea level. This work was conducted at UCI and at Caltech’s Jet Propulsion Laboratory under a contract with the Gordon and Betty More Foundation and with NASA’s Cryospheric Science Program.


Analyzing the effects of sea-ice variability on marine-terminating glacier retreat, central West Greenland

Ashley York, Karen Frey, Sarah Das

Corresponding author: Ashley York

Corresponding author e-mail: ayork@clarku.edu

The response of marine-terminating glaciers to a warming climate is particularly complex, as they are affected by both increasing atmospheric and oceanic temperatures, as well as physical controls such as sea-ice/ice-melange buttressing, underlying bedrock and fjord bathymetry. These environmental and physical factors determine the seasonal and long-term timing and magnitudes of changes observed in the dynamics of glaciers, but have proved difficult to model accurately on more than an individual glacier basis. In this study, we investigated the potential influence of adjacent sea ice conditions on the variability of 18 marine-terminating glaciers in Disko and Uummannaq bays of central West Greenland over the past several decades. One of the simplest ways to assess outlet glacier behavior utilizing satellite data is to monitor termini positions over time, so we constructed a time-series of terminus positions of these 18 glaciers digitized from Landsat imagery between 1985 and 2015. Furthermore, to elucidate potential relationships between sea ice and outlet glacier behavior over both the seasonal and historical record, we combined SMMR-SSM/I (25 km, 1979–2012) and AMSRE/AMSR2 (6.25 km, 2003–11/12–present) passive microwave data to calculate changes in annual sea-ice persistence and the timing of breakup/formation across the broader Davis Strait and Baffin Bay region as well as within Disko and Uummannaq bays. Additionally, glacier-terminus retreat was compared to satellite observations of sea-surface temperature as well as climate reanalysis and automatic weather station data to provide ancillary information on oceanic and atmospheric conditions across the region. Our findings show that the 11 glaciers in Uummannaq Bay retreated an average of –1227.82 ±; 1279.26 m over the entire study period, whereas the retreat of seven glaciers in Disko Bay averaged –1098.94 ± 809.72 m. These results indicate retreat of the majority of glaciers in the region over the study period, yet we observe no latitudinal trend in magnitude of retreat on either a seasonal or long-term scale. This study further finds that sea ice persists almost 1 month longer in Uummannaq Bay compared to Disko Bay; however, the direct effects that sea-ice variability has on marine-terminating glacier retreat throughout this region in a warming climate remain unclear.


Sub-seasonal pressure, geometry and sediment transport changes observed in subglacial channels

Florent Gimbert, Victor Tsai, Timothy Bartholomaus, Jason Amundson, Jake Walter

Corresponding author: Florent Gimbert

Corresponding author e-mail: tsai@caltech.edu

Water from ice melt and precipitation that flows to and pressurizes the base of glaciers contributes to glacier and ice sheet acceleration. Predicting acceleration and its impact on ice mass loss and sea-level rise under global climate warming therefore requires knowledge of subglacial channel evolution and water pressurization, which remains limited by a lack of observations. Here we demonstrate that the detailed analysis of ground motion caused by subglacial channel flow at Mendenhall Glacier (Alaska) allows for the simultaneous measurement of basal water pressure, channel geometry and sediment transport throughout the melt season. We provide observations of the interplay between these physical quantities and discuss the implications for glacier sliding and erosion. By constraining the physics of subglacial hydrology, our framework and its application to outlet glaciers of the Greenland and Antarctic ice sheets may lead to more reliable predictions of ice flow, sea-level rise and subglacial erosion rates.


Bounds on calving rates based on ice thickness and water depths

Yue Ma, Jeremy Bassis

Corresponding author: Yue Ma

Corresponding author e-mail: yuema@umich.edu

Iceberg calving is responsible for nearly half of the mass lost from ice sheets to the oceans. However, a lack of a well parameterized calving model leaves most numerical ice-sheet models incomplete. A variety of models have successfully reproduced patterns of glacier retreat by tuning when water-filled surface crevasses intersect with basal crevasses. But the apparent need for melt water is puzzling because calving events also occur during the winter, when there is little melt water available. Here we seek to reconcile this discrepancy by examining the state of stress associated with crevasse propagation in idealized marine-terminating glacier geometries using a two-dimensional full-Stokes finite element model. Based on this model, we examine when water-free surface crevasses intersect with water-filled basal crevasses using a range of basal boundary conditions that span from a well lubricated (free-slip) to a frozen (no-slip) bed. We find that under free-slip conditions surface and basal crevasses intersect when glaciers thin to near buoyancy, providing a physical reason for the height above-buoyancy calving behavior observed in previous studies. In contrast, when the glacier is very thick, shear stresses exceed the material strength of ice and faulting is likely to occur. Together, the two criteria can be used to map out a stable region in an ice-thickness–water-depth diagram. Comparing our theoretical bounds with observational glacier data shows that available observations lie within our stability envelope. The upper and lower bounds on ice thickness have the potential to be incorporated into numerical ice sheet models in a manner that brackets the magnitude of calving rates possible in marine-terminating glaciers.


A coupled glacier dynamics and fjord circulation model incorporating submarine melting at the glacier front

Eva de Andrés, Jaime Otero, Francisco Navarro, Agnieszka Promińska, Javier Lapazaran, Waldemar Walczowski

Corresponding author: Eva de Andrés

Corresponding author e-mail: eva.deandres@upm.es

Among the processes occurring at the front of marine-terminating glaciers the formation of a buoyant plume is one exerting a critical control on the submarine melting. In this study, a fjord circulation model is coupled with a flowline glacier dynamics model, with subglaciar discharge, allowing the calculation of submarine melting and its influence on calving processes. We focus our study on the Hansbreen Glacier–Hornsund Fjord system, in Southern Spitsbergen, Svalbard, where a large set of data are available for both glacier and fjord. The bathymetry of the entire system has been determined from ground-penetrating radar and sonar data. In the fjord we have got temperature and salinity data from CTDs (May–September, 2010–14) and from a mooring (September–May, 2011/12). For Hansbreen, we use glacier surface topography data from the SPIRIT DEM, surface mass balance from the European Arctic Reanalysis (EAR), centre-line glacier velocities from stakes (May 2005–April 2011), weekly front positions from time-lapse photos (September 2009–September 2011) and ASTER images, and sea-ice concentrations from time-lapse photos and Nimbus-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data. For ocean modelling, we use a general circulation model, MITgcm, to simulate water circulation driven by both fjord conditions and subglacial discharge, and for calculating submarine melt rates at Hansbreen's front. To determine the optimal values of discharge for each season within the year, we perform an analysis of sensitivity of the model to subglaciar discharge variability. Then we establish initial and boundary fjord conditions, which we vary weekly, and calculate the submarine melt rate as a function of depth at the calving front. These data are entered into the glacier flow model Elmer/Ice, to which has been added a crevasse-depth calving model to estimate Hansbreen’s front position at a weekly time resolution.


Surface radiation budget of Helheim and Jakobshavn glaciers, Greenland, as measured with automatic weather stations

Georges Djoumna, David M. Holland

Corresponding author: Georges Djoumna

Corresponding author e-mail: gdjoumna@gmail.com

The representation of radiation budgets in global climate models still shows large discrepancies in surface radiation budget components. Five years of near-surface radiation balance observations from two automatic weather stations (AWS) are presented. The AWS are situated on the northeast coast of Ilulissat Ice Fjord (69°13′19.51″ N, 49°48′57.11″ W), west Greenland and on the south shore of Helheim Ice Fjord (66°19.77′ N, 38°08.10′ W), southeast Greenland. We provide a general description of the meteorological conditions, radiative fluxes and accumulation at Helheim and Jakobshavn glaciers based on observations from five years of operation (2009–14). Cloud cover is a prominent feature of polar regions and glaciated areas, but its effects on radiative fluxes remain problematic, mainly due to lack of cloud observations over ice sheets. The incoming long-wave radiation under clear sky and cloudy sky was examined using a variety of parameterization schemes. These new meteorological observations provide the first long-term in situ data and radiative fluxes at the two glaciers, and complement earlier works on the general description of meteorological conditions and radiative fluxes in Greenland.


Ice-cliff failure via slumping

Knut Christianson, Byron Parizek, Richard Alley, David Holland, Denis Voytenko, Tim Dixon

Corresponding author: Knut Christianson

Corresponding author e-mail: knut@uw.edu

Catastrophic failure is widely observed if terrestrial cliffs reach a sufficient height such that weight from the overlying rock exceeds rock strength, producing landslides and other hazards. Similar failures occur in ice, which is generally weaker than rock. For ice, cliff-failure observations are limited to a few glaciers in Greenland and the Antarctic Peninsula. Parameterized modeling indicates that widespread cliff failure in West Antarctica may be possible past mid-century after large buttressing ice shelves are removed by hydrofracture induced by surface melting, causing collapse of entire ice-sheet drainage basins (>1 m sea-level equivalent) within decades of cliff-failure initiation. Many unknowns remain, including failure mechanism in ice, yield stress of ice, and inland ice-flow response. Recently, we measured ice speed, strain and elevation at Helheim Glacier, Greenland, where ice-cliff failure occurs regularly. Our initial analysis of a single large buoyant-flexure calving event indicates that debris slumps forward and then drops from the cliff face immediately prior to calving. This debris slumping reduces the ice thickness above sea level, allowing buoyancy forces to rapidly float the remaining ice upward, breaking (calving) in the process. Our data-informed process modeling indicates especially high tensile stresses (several hundred kPa) in the surface depression observed ~500 m inland of the ice front and especially high shear stresses (several hundred kPa) just above the water line at the ice front. Given the fractured state of the ice, these stresses may exceed the flexural depth-averaged yield stress of the ice (frequently assumed to be ~1 MPa). Fracture is observed and modeled to initiate at the surface, likely facilitated by the presence of surface meltwater. Crevassing is modeled to follow the maximum stress contour until it reaches the water line, causing the slump and thinning the ice to flotation. Once thinned to flotation, calving via buoyant flexure occurs along most of the ice front. Our results provide a new physical framework for ice-cliff failure and calving that can easily be included in continental-scale ice-sheet models, likely significantly improving the ability of models to simulate sea-level change.


Full-Stokes modeling of grounding-line dynamics and iceberg calving for Thwaites Glacier, West Antarctica

Hongju Yu, Eric Rignot, Mathieu Morlighem, Helene Seroussi

Corresponding author: Hongju Yu

Corresponding author e-mail: hongjuy@uci.edu

Thwaites Glacier (TG) is the second largest and broadest ice stream in the Amundsen Sea Embayment (ASE) sector of West Antarctica. Recent observations have revealed rapid thinning, grounding line retreat and mass loss of TG. The fast retreat have been attributed to high basal melt rate and enhanced iceberg calving as the ocean water is warming and large calving events are observed from satellites. However, the grounding-line migration and iceberg calving involves a complex stress field and the full-Stokes (FS) simulation is required. Here, we present a two-dimensional, FS modeling study of the grounding-line dynamics and iceberg calving of TG. First, we compare the FS model with simplified models to determine the impact of changes in ice-shelf basal melt rate on grounding line dynamics. Second, we combine the FS model with the Linear Elastic Fracture Mechanics (LEFM) theory to simulate crevasse propagation and iceberg calving. In the first experiment, we find that simplified models require an unrealistic basal melt rate for the grounding line to remain stable and are less sensitive to changes in melt rate. In contrast, the FS model requires a melt rate in agreement with remote-sensing observations and the grounding line is more sensitive to changes in basal melt rate. In the second experiment, we find that only the FS model can produce surface and bottom crevasses that match radar observations in size. Additional experiments indicate that iceberg calving is significantly enhanced when prior surface cracks exist near the grounding line, when the ice shelf is shortened or when the ice front is undercut. We conclude that the FS model yields substantial improvements in the description of ice-flow dynamics at the grounding line and also in incorporating crevasse formation and iceberg calving.


Linkages between glaciated coastal mountain watersheds and near-shore waters

David Hill, Jordan Beamer, Anthony Arendt, Seth Danielson, Kate Hedstrom

Corresponding author: David Hill

Corresponding author e-mail: david.hill@oregonstate.edu

The terrestrial–marine interface is an important boundary condition driving oceanographic processes. The delivery of fresh water to the ocean at this interface affects water-column properties, near-shore circulations and mean sea level. The volume, spatial distribution and timing of this freshwater flux all affect the characteristics of the ocean response. The Gulf of Alaska watershed has a complex water balance, due to significant annual precipitation and the complex spatial patterns of that precipitation, temperature and terrain. The watershed has an annual runoff volume of approximately 800 km3 and approximately 60 km3 of this derives from glacier volume loss in the watershed. We apply high spatial resolution hydrologic and oceanographic models to study the coupled dynamics of the Gulf of Alaska and its watershed. The suite of hydrologic models (a) distributes weather reanalysis data to the 1 km model grid, (b) evolves the snowpack (including melt of underlying ice once the seasonal snowpack has disappeared), and (c) transports water across the landscape to the ocean. These models have been calibrated with a broad array of observational data and the regional model predictions demonstrate good agreement with GRACE estimates of water storage. The model results show strong seasonal variation in the runoff and are also able to partition the runoff into contributions from rainfall, snow melt and ice melt. The ROMS oceanographic model uses the 1 km hydrologic model as input to its simulations of the nearshore dynamics. The high spatial resolution is able to resolve the significant freshwater inputs from individual bays and fjords. Simulations comparing spatially distributed vs spatially constant freshwater input have also been carried out.


Ocean circulation and marine-terminating glaciers of the Greenland Ice Sheet

Laura Gillard, Xianmin Hu, Paul Myers

Corresponding author: Laura Gillard

Corresponding author e-mail: gillard2@ualberta.ca

The Greenland Ice Sheet (GrIS) stores the largest amount of fresh water in the Northern Hemisphere, and in recent years has been losing mass at an increasing rate. This lost mass is added to the surrounding oceans as fresh water. Driven with realistic estimates of Greenland meltwater runoff from a glacial mass-balance model, an eddy-permitting coupled ocean and sea ice general circulation model based on NEMO v3.4 is utilized to setup a 0.25° regional (Arctic and Northern Hemisphere Atlantic, ANHA4) simulation to study the fate of Greenland meltwater and the sources of warm water reaching the fjords. Using two approaches to track the meltwater (online passive tracers and an offline Lagrangian package), we examine where this discharge is taken up in the downstream oceans. Here we show that freshwater from western and eastern Greenland has very different fates, at least on a decadal timescale. Fresh water from west Greenland mainly ends up in Baffin Bay, before being exported south down the Labrador shelf. Meanwhile, fresh water entering the interior of the Labrador Sea, where deep convection occurs, comes mainly (~80%) from east Greenland. Our results suggest the need to understand the regional aspects of changes in the GrIS if we wish to understand how its evolution in a warming climate may affect the ocean. It has been shown that relatively warm ocean waters may accelerate the melt production of marine-terminating glaciers. We explore and classify the pathways for the warmer Atlantic waters that reach the fjords along the coasts of Greenland. Preliminary work will be shown of a high-resolution Baffin Bay NEMO model configuration to examine the role of ocean bottom topography in impacting shelf–basin exchange.


Dynamic response of the Ross Ice Shelf to ice-sheet forcing

Adam Campbell, Christina Hulbe, Ted Scambos

Corresponding author: Adam Campbell

Corresponding author e-mail: a.campbell@otago.ac.nz

Variability in the flow and thickness of ice shelves can be driven internally (from the ice sheet) and patterns driven externally (by the ocean or atmosphere). The Ross Ice Shelf (RIS) is known to experience variability due to flux changes in West Antarctic ice streams and transantarctic outlet glaciers, which may complicate identification of climate-forcing signals in observed changes to the present-day system. Changes to ice-dynamic boundary conditions such as glacier influx and traction over ice rises produce both instantaneous changes to the ice-shelf flow field and longer-timescale adjustments in the coupled thickness and flow. How those changes propagate through the system also depends on the morphology of the embayment containing the shelf. For example, flux changes associated with the stagnation and reactivation of Whillans Ice Stream (WIS) are mediated by Crary Ice Rise, a high on the sea floor on the downstream side of the WIS grounding zone. When Crary Ice Rise froze on to the sea-floor rise on which it is grounded, flow upstream of the obstacle slowed and the distribution of mass through the shelf area was changed, becoming thicker upstream and around the ice rise and thinner downstream in its wake. The CIR imprint is then impressed on ice-shelf response to later transient effects, such as the stagnation of WIS, in part by trapping some of the signal upstream of the rise, and by directing mass flux in the shelf and thereby setting its overall thickness pattern. Other features, such as the Steershead ice rise, have similar effects. Here, we use an ice-flow model to identify characteristic patterns associated with perturbations to various ice-dynamic boundary conditions in the expectation that these can be separated from climate-driven change. The results of our experiments are compared with recent changes in ice flow detected via repeat-feature tracking in Landsat and MODIS and changes in ice thickness detected using IceSat. The experiments include ice stream and transantarctic glacier flow variation, grounding of ice rises and calving of tabular icebergs.


A preliminary model of viscoelastic ice-shelf flow of the Ross Ice Shelf

Kelly Brunt, Douglas MacAyeal

Corresponding author: Kelly Brunt

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

Three stations near the calving front of the Ross Ice Shelf, Antarctica, recorded GPS data through a full spring–neap tidal cycle in November 2005. The data revealed a diurnal horizontal motion that varied both along and transverse to the long-term average velocity direction. Based on its periodicity, it was hypothesized that the signal represents a flow response of the Ross Ice Shelf to the diurnal tides of the Ross Sea. To address this hypothesis, a finite-element model that combines viscous and elastic deformation is currently in development. The model treats the ice shelf as a Maxwell viscoelastic system, where elastic-response results can be added to results from a previous model, which was based purely on viscous creep. Here we present preliminary results from the viscoelastic model, which incorporates numerous datasets, including: 1) a domain boundary based on ICESat and InSAR grounding zones; 2) ice-shelf thickness derived from ICESat surface-elevation data, assuming hydrostatic equilibrium; 3) ice-thickness corrections based on IceBridge MCoRDS data; 4) tide corrections and forcing based the Circum-Antarctic Tidal Simulation (CATS) model; and 5) domain velocities, and steady-state model validation, based on MEaSUREs ice-flow velocity data.


Export of supercooled surface water from beneath an ice shelf: past records and future extent

Pat Langhorne, Ken Hughes, Alex Gough, Inga J. Smith, Mike J.M. Williams, Natalie J. Robinson, Craig Stevens, Wolfgang Rack, Greg Leonard

Corresponding author: Pat Langhorne

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

Ice-shelf basal melting freshens and cools the fluid in the ice-shelf–ocean boundary layer, producing ice shelf water (ISW). The potential temperature of ISW is below the surface freezing point. Antarctic sea ice that has been affected by supercooled ISW has a unique crystallographic structure and is called platelet ice. We have synthesized platelet ice observations to construct a continent-wide map of the winter presence of ISW at the ocean surface. The observations demonstrate that, in some regions of coastal Antarctica, supercooled ISW drives an effective negative oceanic heat flux of about –30 W m–2 that persists for several months during winter. This heat flux from the sea ice to the ocean significantly increases the sea-ice thickness close to an ice shelf. In other regions, particularly where the thinning of ice shelves is believed to be greatest, platelet ice is not observed. The most extensive dataset, which includes our new results, dates back to 1902 and extends north of the combined Ross and McMurdo Ice Shelf front in the southern Ross Sea. Here the surface water is held just below its freezing point as it enters McMurdo Sound from beneath the McMurdo Ice Shelf. Since the early 20th century there has been no detectable change in the volume or temperature of this supercooled ISW under sea ice. The inclusion of platelet ice into first-year sea ice is an annual process. Hence it will respond immediately to changes in the sub-ice-shelf circulation pattern and its export of supercooled water. It has recently been demonstrated that the presence of a porous platelet-ice layer beneath sea ice can be detected by electromagnetic induction techniques from the sea-ice surface or from airborne sensors. In this presentation, we outline how this this provides a means to undertake repeated monitoring in suitable, at-risk regions around the coast of Antarctica.


An ice velocity update for the Remnant Larsen-B and Larsen-C ice shelves

Bernd Scheuchl, Jérémie Mouginot, Eric Rignot

Corresponding author: Bernd Scheuchl

Corresponding author e-mail: bscheuch@uci.edu

Synthetic Aperture Radar (SAR) data are a known resource to provide relevant information on ice sheets. Specifically, ice velocity and grounding line and ice-front location can be extracted. Landsat-8, the latest satellite to carry the Landsat mission forward, is the first optical sensor suitable for generating large-scale, high-quality ice-velocity maps. This new wealth of available data sources, together with the commitment by space agencies to collect and make data available, shifts ice-sheet remote sensing from a field exploiting acquisitions of opportunity to an area where information services can be provided on an ongoing basis. There is still an urgent need for a SAR science mission like NISAR, which will shift ice-sheet data availability from regular acquisitions over key regions to ongoing acquisitions of the entire observable area. Here, we provide an update of a time series of the Larsen-B and -C ice shelves and their tributaries. Results show a continued increase in surface velocity on Remnant Larsen-B Ice Shelf, in contrast to the relative stability of the Larsen-C Ice Shelf. We use data from Landsat-8 and a variety of SAR satellite missions that have acquired science data in the region. The time series will be made available as Earth System Data Record (ESDR). Data analysis and ESDR production is conducted at the Department of Earth System Science, University of California Irvine, under a contract with NASA’s MEaSUREs program. Spaceborne SAR data were made available for this project courtesy of the Polar Space Task Group.


Buoyant meltwater flows under gently sloping ice shelves: plumes, gravity currents and hydraulic control

Andrew Wells, Sander Rhebergen

Corresponding author: Andrew Wells

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

The basal ablation of floating ice shelves depends on turbulent heat and salt transfer to the ice–ocean interface, which can be controlled by the buoyant flow of meltwater along the ice face. Quasi-steady meltwater plume models are a useful conceptual tool for characterizing the dynamics of buoyant meltwater flow, but can encounter physical pathologies when applied to ice shelves with very shallow basal slopes. We investigate and remove these pathologies by re-interpreting the meltwater plume dynamics embedded within the framework of a time-dependent model. To illustrate the essential concepts, we consider flow under a locally planar region of a two-dimensional ice shelf. Complementing theoretical arguments with time-dependent simulations, we show that the meltwater flow exhibits distinct regimes depending on a local Froude number for the buoyancy-driven current. For supercritical conditions, plume-like flows are controlled entirely by the source conditions and evolution of the flow downstream. However, subcritical conditions lead to flows similar to a gravity current, where the dynamics depend on a combination of upstream and downstream conditions, so that there is significant downstream influence on the meltwater flow. We also identify hydraulically controlled states where the flow transitions between gravity-current-like behaviour and plume-like behaviour along the length of the ice shelf, as a result of the evolution of buoyancy in the current. Finally, we discuss the implications for using steady-state meltwater plume models under gently sloping ice shelves.


Using an idealized ocean circulation model to assess the role of fjord-glacier geometry on circulation in tidewater glacier fjords

Dustin Carroll, David Sutherland, Jonathan Nash, Emily Shroyer, Laura de Steur, Ginny Catania, Leigh Stearns

Corresponding author: Dustin Carroll

Corresponding author e-mail: dcarroll@uoregon.edu

The acceleration, retreat and thinning of Greenland’s outlet glaciers coincided with a warming of Atlantic waters, suggesting glacier sensitivity to ocean forcing. While Greenland’s fjords often contain ample ocean heat for melting ice, submarine melting is limited by the net ocean heat flux to the terminus through processes that are highly dependent on local fjord circulation. However, we still lack a precise understanding of what factors control the variability of heat transport toward the ice face. Here we use an idealized ocean general circulation model (MITgcm) to systematically evaluate how fjord circulation driven by steady subglacial discharge depends on fjord-glacier geometry, specifically the grounding-line depth, fjord width and sill height. We perform sensitivity experiments to examine how two external forcing mechanisms, the wind stress (both along-fjord and along-shelf) and tides, influence this buoyancy-driven circulation. Our results indicate that, while subglacial plumes in deeply grounded systems can draw deep waters over a sill and toward the glacier, shallowly grounded systems require external forcing to renew basin waters. Passive tracers injected in the plume, fjord basin and shelf waters are used to quantify turnover timescales. Our results underscore the first-order effect that geometry has in controlling fjord circulation and thus ocean heat flux.


Variations at the ice–ocean boundary: insights from terrestrial radar interferometric studies of tidewater glaciers

Ryan Cassotto, Mark Fahnestock, Jason Amundson, Martin Truffer, Shad O’Neel

Corresponding author: Ryan Cassotto

Corresponding author e-mail: ryan.cassotto@wildcats.unh.edu

Tidewater glaciers are sensitive to processes along the ice–ocean boundary. Ocean tides, iceberg calving, variations in ice melange, and precipitation are some of the perturbations that affect tidewater glacier flow. Remote and in situ studies have vastly improved our understanding of these processes and the effect they have on tidewater glacier dynamics; however, observations typically have either high temporal or high spatial resolution, but rarely both. Consequently, many studies lack sufficient resolution to fully characterize the response of a tidewater glacier terminus to perturbations at the boundary layer. Terrestrial radar interferometers (TRI) generate high-resolution (tens of meters) images of surface deformation on very short timescales (~3 min), which can be used to map variations in a glacier’s response to boundary layer processes. Here we present TRI observations of glacier response to processes along two large tidewater glacier systems: Jakobshavn Isbræ, Greenland, and Columbia Glacier, Alaska, USA. At Jakobshavn Isbræ, we show evidence of nonlinear response to calving and the effect speed variations have on ice thickness, and we present observations of melange activity in the proglacial fjord. At Columbia Glacier, we present observations of tidal forcing and a strong response to precipitation; we also show evidence of a change in Columbia’s response to ocean tides that results from a significant precipitation event. These studies demonstrate that understanding tidewater glacier dynamics requires knowledge of short-term variations along glacier termini.


The estimation of glacial ice concentration in sea (icebergs, bergy bits and growlers) using dual-polarized SAR images and multimodal recurrent neural networks (m-RNNs) near the front of the Jakobshavn Glacier, western Greenland

Ester Y. Kashtanov, Adam Y. Kashtanov

Corresponding author: Ester Y. Kashtanov

Corresponding author e-mail: e.kashdan@yahoo.com

The estimation of glacial ice concentration in sea (including icebergs, bergy bits and growlers) is of great interest for offshore operations in the Arctic and subarctic waters. In our research we use multimodal recurrent neural networks (m-RNNs) to estimate concentration of glacial ice in sea from C-band dual-polarized Sentinel-1A satellite images before and after a calving event (14 and 16 August 2015), when Jakobshavn Glacier lost a total area of 12.5 km2, equivalent to a volume of 17.5 km3 (according to NASA). SAR images are used as input and the glacial ice concentration is the direct output from the RNN’s long short-term memory and graphics processing unit. The m-RNNs architecture is demonstrated to produce glacial-ice concentration maps with more detail than image analyses. Sensible glacial ice concentration estimations before and after the calving event are made and areas of low glacial ice concentration are also well captured.


Ice-shelf buttressing and the response of ice-stream flow to vertical ocean tidal motion

Brent Minchew, Bryan Riel, Alex Robel, Mark Simons, Victor Tsai, Pietro Milillo

Corresponding author: Brent Minchew

Corresponding author e-mail: simons@caltech.edu

Ice-shelf buttressing provides an important restraint on inland ice flow. Whether arising from shear in the margins of confined ice shelves or at pinning points, i.e. areas along the base of the ice that are in contact with bathymetric highs, ice-shelf buttressing is susceptible to changes in water-column depth and water temperature at the ice–ocean interface. Here, we study the influence of short (sub-monthly) timescale variations in ice-shelf buttressing on ice-stream flow. We infer time-dependent three-dimensional surface velocity fields using 9 months of continuous synthetic aperture radar observations collected with the COSMO-SkyMed 4 satellite constellation from multiple satellite viewing geometries. The time-varying velocity field components elucidate the spatial characteristics of the response of ice flow on the Rutford Ice Stream to ocean tidal forcing and agree with collocated GPS measurements. The highest amplitude variations in horizontal ice flow are at the period of the Msf tide (14.77 d), which is a beat of the M2 (12.42 h) and S2 (12 h) tides. We show that the response of horizontal ice flow to ocean tidal forcing occurs first and is most pronounced over the ice shelf, subsequently propagating through the grounded ice stream at a mean rate of ~29 km d–1 while decaying quasi-linearly with distance over ~85 km upstream of the grounding zone. Observations of horizontal ice displacement from InSAR and GPS indicate that the relative amplitude at the different tidal frequencies is not simply proportional to the driving tidal forcing as observed in vertical displacement of the ice shelf. Generally, we expect that tides lift the shelf off pinning points, thus decreasing buttressing stress. We demonstrate how the horizontal strain measured downstream from the grounding line, and the associated buttressing contact stress, vary nonlinearly with vertical tidal height. Buttressing stress is more sensitive to changing tides at high tide than at low tide. Consequently, the highest tidal variations resulting from beating of the M2 and S2 tides are amplified by this asymmetric response, leading to a high-amplitude variation in horizontal ice flow at the Msf period. Variations in buttressing stress at the M2 and S2 tidal periods are then attenuated by viscous deformation of the ice shelf, only leaving the observed variations in horizontal velocity at the Msf period.


Coupling of ice sheets and ocean in ACME: progress and challenges

Jeremy Fyke

Corresponding author: Jeremy Fyke

Corresponding author e-mail: fyke@lanl.gov

The Antarctic Ice Sheet (AIS) has the potential to produce multiple meters of sea-level rise on poorly constrained timescales in response to anthropogenic climate forcing. To address this concern, the DOE ACME project aims to couple the Antarctic ice sheet into the global climate system and use the resulting model to project AIS behaviour and consequent sea level rise. Inclusion of the AIS in a coupled model introduces the need to implement unprecedented ice-sheet–ocean coupling in a physically based (non-parameterized) manner. In this poster we document progress and expected challenges on this front, which from a coupling perspective typically involves the need to fundamentally re-engineer coupled model architectures to (for example) provide (1) dynamically migrating boundaries between climate components, (2) multiple stacked climate components in map view, (3) dynamic multi-fraction calculation of inter-component mass/energy fluxes, (4) new canonical climate model boundary condition files (e.g. to potentially allow ocean at the South Pole!), (5) new validation benchmarks and bias detection analyses and (6) new spin-up/initialization techniques


Modelling the evolution of calving-front morphology under submarine melting

Donald Slater, Peter Nienow, Dan Goldberg, Tom Cowton, Andrew Sole

Corresponding author: Donald Slater

Corresponding author e-mail: d.slater@ed.ac.uk

Ocean forcing is thought to be an important driver of tidewater glacier dynamics. Process understanding of the interaction of the ocean with glaciers is, however, limited, severely hampering our ability to make future projections of marine-terminating glacier dynamics. In recent years, progress has been made in modelling rates and variability of submarine melting. However, even with perfect knowledge of submarine melt rates, we would be limited in our understanding of their effect on tidewater glaciers; for example, we understand little of the potential link between submarine melting and calving processes. Here we aim to tackle this knowledge gap by quantifying, for the first time, the effect of submarine melting on tidewater-glacier calving-front morphology. We couple buoyant plume theory to a model of calving-front evolution, enabling us to study the time-evolving morphology of a tidewater-glacier calving front under plume-induced submarine melting. We identify the plume geometry, fjord stratification and magnitude of subglacial discharge as critical controls on calving-front morphology. Undercutting is most prominent for point source plumes in strongly stratified fjords, with the steepest undercutting at the depth where the plume reaches neutral buoyancy. We also note the formation of protruding ‘toes’ near the grounding line, which may promote buoyancy-driven calving. Our results suggest that the feedback of calving-front morphology on plume dynamics and thus local submarine melt rates is weak. This finding implies that the total flux of ice removed at the calving front by submarine melting scales with the magnitude of undercutting, and thus calving front morphology may be an important factor in determining the mass balance of tidewater glacier termini. Our modelling generates morphologies that compare favourably to recent in situ imaging of calving fronts and illuminates links between submarine melting and calving.


Oceans melting Greenland: mission update

Ian Fenty, Josh Willis, Eric Rignot

Corresponding author: Ian Fenty

Corresponding author e-mail: ian.fenty@jpl.nasa.gov

This talk will update the community on the current status of NASA’s Oceans Melting Greenland (OMG) mission. The OMG mission has three observational components: bathymetry, in situ hydrography and ice-sheet elevation. In 2016, OMG conducted its first campaigns to measure glacier elevation near the terminus of tidewater glaciers using radar interferometry and seafloor bathymetry from airborne gravimetry. Summer 2016 plans include multi- and singlebeam sonar surveys in southeast and northwest Greenland and OMG’s first pan-shelf in situ temperature and salinity measurements using Airborne eXpendable Conductivity–Temperature–Depth (AXCTD) probes.


The effect of high-resolution bathymetry from ROSETTA-ICE on model-simulated basal melting of the Ross Ice Shelf

Scott Springer, Laurie Padman, Kirsteen Tinto

Corresponding author: Scott Springer

Corresponding author e-mail: springer@esr.org

Basal melting of the Ross Ice Shelf (RIS), Antarctica, which is currently in approximate steady state, contributes ~50 Gt per year of freshwater to the Ross Sea. Most model-simulated values are a factor of 2–5 times as large. One significant limitation on modeling RIS is the quality of cavity geometry including ice draft and seabed bathymetry. Ice draft is determined from inversions of satellite altimetry with assumed firn density. Current bathymetry models are based on the coarse (~50 km) grid of values obtained from RIGGS seismic surveys in the 1970s. Inversion of gravity data collected during the first year of the ROSETTA-ICE airborne survey mission (late 2015) provides evidence of bathymetric variability on finer scales (~10 km), including locations of seabed ridges that are poorly sampled by the RIGGS seismics grid. We use a high-resolution, thermodynamically coupled ocean–ice-shelf model to compare melt rate maps for simulations with RIGGS-based bathymetry and with updated bathymetry suggested by preliminary inversions of the ROSETTA-ICE gravity data. Differences can be attributed to multiple causes including changes in: mean circulation patterns (access of ocean heat to the ice base, and resulting meltwater plume flow); distribution of tidal currents; and sub-ice-shelf ocean mixing. ROSETTA-ICE will also improve maps of RIS ice draft, which will be incorporated into future model developments. The study provides a framework for investigating how RIS might respond to changes in atmospheric and oceanic states over the Southern Ocean in future climates.


Fast-ice/ice-shelf interactions in the Larsen B embayment

Erin Pettit, Ted Scambos, Martin Truffer, Christina Carr, Hilmar Gudmundsson

Corresponding author: Erin Pettit

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

Following the collapse of the Larsen B ice shelf, Antarctica, in 2002, a remnant shelf located in Scar Inlet began accelerating and thinning. Between 2002 and 2012, the ice-shelf front retreated 25 km, and satellite data show that the ice shelf has progressively weakened, with increased shear along the margins and new internal rifts and crevasses, including basal crevasses near the shelf front. Despite this visible structural weakening, the retreat of the ice front has slowed in recent years and ice-flow speeds have stabilized at approximately 900 m a–1 after increasing from 400 m a–1 in 2002. In the past 2 years, ice-flow speeds near the ice front have slowed slightly, based on in situ GPS and satellite-derived ice velocities. We show that this slowing of frontal retreat and decreasing acceleration is primarily due to persistent fast ice that formed in late 2011. Fast ice presently covers the entire Larsen B embayment and is structurally pinned between the Seal Nunataks, Robertson Island, Jason Peninsula and Cape Disappointment. In 2013 drilling near the fast-ice front indicate that its thickness was ~3 m. During February 2016, we used ground-based radar interferometry, air-deployed GPS poles and time-lapse imagery to study the current fast-ice/ice-shelf interactions. We present these data in the context of 5 years of GPS and imagery from three AMIGOS and GPS stations, satellite imagery and satellite-derived velocities. Currently, the fast ice within 4 km of the ice-shelf front is undergoing compression, while farther from the front it is undergoing extension. Despite compressional strain rates as high as 10% a–1 we do not observe obvious pressure ridging. ‘En echelon’ patterns of leads flanking the coastlines (particularly near Cape Disappointment) demonstrate strong shear strain rates in the fast ice. In aggregate, our data support a hypothesis that buttressing by the fast ice is delaying the collapse of the Larsen B ice-shelf remnant. Our data also suggest seasonality in the buttressing effect, possibly due to strengthening of the fast ice in winter by a combination of melange stiffening in the open leads, decreasing ice temperature, and ice growth. While we still expect the Scar Inlet shelf to collapse in the near future based on trends in local climate forcing, the evolution of this system provides insight into both the process of ice-shelf collapse and notional concepts for the process of ice-shelf regrowth.


Surveying most of Greenland’s marine-terminating glaciers with an airborne radar altimeter

Ala Khazendar, Ian Fenty, Eric Rignot, Josh Willis

Corresponding author: Ala Khazendar

Corresponding author e-mail: ala.khazendar@jpl.nasa.gov

Ala Khazendar, Ian Fenty, Eric Rignot, Josh Willis and the OMG team Oceans Melting Greenland is a NASA suborbital mission to investigate the role of the oceans in ice loss around the margins of the Greenland Ice Sheet. A 5-year airborne and ship-based campaign, the project will directly measure ocean temperatures and glacier changes around most of Greenland. A main component of the airborne campaign is a once-per-year survey of glacier surface elevations of most of the marine-terminating glaciers with GLISTIN-A. This radar is a topographic mapper that measures ice-surface elevations with 50 cm vertical accuracy at 25 m horizontal resolution over a 10–12 km swath, independent of weather. The first survey took place in March 2016 over 10 days and nearly 65 flight hours. We present here a first look at the measurements made during this campaign, how they compare with existing laser and radar altimetry data and a preliminary evaluation of their contribution in estimating ice-thinning rates.


Rapid submarine ice melting underlying glaciological changes in West Antarctica

Ala Khazendar

Corresponding author: Ala Khazendar

Corresponding author e-mail: ala.khazendar@jpl.nasa.gov

The Amundsen Sea Embayment (ASE) remains the region with the highest mass loss in Antarctica. The loss, which rose significantly in the mid-2000s, is strongly influenced by enhanced melting in the cavities beneath the ice shelves. The melting magnitude has hitherto been indirectly estimated at coarse spatial resolutions that mask changes in the critical grounding zones. Here, we directly quantify bottom ice losses along tens of kilometers with airborne sounding radar. We focus on the Dotson and Crosson ice shelves and their tributary glaciers of Smith, Pope and Kohler that exhibit some of the most pronounced glaciological changes in the ASE. Melting rates in the grounding zones of the three glaciers are found to be much higher than steady-state levels, reaching 300–490 m between 2002 and 2009 in the case of Smith Glacier. The intense unbalanced melting supports the hypothesis that a large increase in ocean heat influx into the sub-ice-shelf cavities of the ASE occurred in the mid-2000s. Melting rates between 2009 and 2014 are similar or lower, which also agrees with the slower increases in mass loss by the end of the 2000s. We describe diverse glacier interactions with such changing ocean forcing and identify a relationship among higher melting rates, steepening retrograde beds and farther grounding-line retreats. Our findings illustrate the extensive changes in the ASE as submarine melting is now large enough to detect with radar, which opens new possibilities for monitoring on regional scales the responses of ice shelves to climate variability.


Drivers of glacial fjord circulation: shelf versus local forcing

Rebecca Jackson, Fiamma Straneo

Corresponding author: Rebecca Jackson

Corresponding author e-mail: rebecca.h.jackson@gmail.com

Glacial fjords connect the open ocean to the glaciers of the Greenland Ice Sheet. These fjords are the gateways for importing oceanic heat to melt ice and for exporting meltwater into the ocean. However, is unclear how various drivers of fjord circulation – including fresh water, local winds and shelf variability – affect these heat and meltwater fluxes. Building on observations from Sermilik Fjord at the terminus of Helheim Glacier, we use two models (ROMS and an analytical model) to explore the impact of shelf forcing on glacial fjords and the competition between shelf and local forcing. We investigate the relative importance of shelf versus local forcing across the parameter space of Greenland’s fjords, showing how the fjord response varies with fjord geometry and stratification.


Basal ice dynamics of McMurdo Ice Shelf

Britney Schmidt, Justin Lawrence, Catherine Walker, Luke Winslow, Peter Doran, William Stone, Donald Blankenship, Stacy Kim, Frank Rack

Corresponding author: Britney Schmidt

Corresponding author e-mail: britneys@eas.gatech.edu

Over the 2012, 2014 and 2015 austral summers, the NASA-funded SIMPLE project (Sub-Ice Marine and PLanetary-analog Ecosystems), spent a total of 8 months characterizing ice and ocean exchange processes below the McMurdo Ice Shelf (MIS), Antarctica, a small portion of the larger Ross Ice Shelf easily accessible from USAP’s McMurdo Station. The program developed the Deep-SCINI and ARTEMIS vehicles, and deployed two other vehicles, SCINI and Icefin. We collected imaging and sonar data of the ice, conductivity and temperature profiles of the water column, water samples at the interface for geochemical and biological characterization, and several other complementary in situ measurements. In 2012 we explored at a single location 5 km back from the front of the ice shelf using the ROV SCINI. Here we observed ablation of the ice and a heterogeneous water column, consistent with melting by fast-moving currents; the CTD profiles are consistent with this conclusion. In 2014, we utilized SCINI and the new vehicle Icefin to characterize five hot-water-drilled sites below the MIS. These locations ranged between 10 and 20 km from the shelf front. At each site, we observed very different ice conditions from 2012. Uniformly, large amounts of platelet ice collect in these regions, regardless of ice-shelf thickness. The layer of platelets was between 1 m and several meters thick depending on the site, and ranged between uniform in character and having large platelet spears and columns. At a site near Black Island, we observed what appears to be a solid layer of platelet ice, physically separated from the bottom of the shelf by a thin water lens, at the base of which was newer platelet ice. The water column below the ice was homogeneous, consistent with the formation of platelet ice. We observed a complex community of organisms at the sea floor under this permanent ice cover. In 2015, the team used the ARTEMIS vehicle to perform several ~1 km transects under the shelf from a camp on multi-year ice at the sea-ice–ice-shelf transition. Here, we observed a thin layer of platelets across the shelf, and observed nearly 4 m of platelet ice growth under the multi-year ice from August to November. We collected high-resolution data of the sea-ice–ice-shelf transition. We also collected approximately daily CTD casts to 500 m, which show a homogeneous upper water column and some indication of seasonal/diurnal/tidal changes. We will describe the basal ice conditions and implied ocean interactions.


Fjord sea-surface temperature variability and ice-terminus retreat at the Petermann Glacier, Greenland

Tasha Snow, Ted Scambos, Waleed Abdalati

Corresponding author: Tasha Snow

Corresponding author e-mail: tasn0057@colorado.edu

Enhanced intrusion of warm ocean waters at the terminus of marine-terminating outlet glaciers around the Greenland Ice Sheet has contributed to elevated rates of ice thinning and terminus retreat over the last two decades. Oceanographic studies show that basal melting of glaciers causes buoyant waters to escape at the surface and contribute to fjord outflow circulation. The temperature of these surface waters may hold clues about ice–ocean interactions and small-scale circulation features along the glacier terminus that could contribute to outlet glacier mass loss, but the magnitude and duration of temperature variability remains uncertain. Satellite remote sensing has proved very effective for acquiring sea surface temperature (SST) data from these remote regions on a long-term, consistent basis and shows promise for identifying warm sea-water at the ice front. However, these data sets have not been widely used to date. Here, we extract remotely sensed SSTs from Landsat 7 ETM+ thermal imagery of the Petermann Glacier, which drains 4% of the Greenland Ice Sheet and is experiencing 80% of its mass loss through basal melt. We compare fjord SST variability along the Petermann Glacier terminus to ice-front retreat since 2000. We find that SSTs can vary by more than 1°C along the 16 km ice front. Our initial results provide insights into the physical processes occurring at the ice–ocean boundary and show the utility of the use of SSTs within the fjords of Greenland outlet glaciers. We will use these results to examine the relationship between fjord temperature and ice-front retreat.


Changes in surface roughness of Greenland tidewater glaciers: indicators of glacial acceleration?

Ute Herzfeld, Thomas Trantow, Gavin Medley

Corresponding author: Ute Herzfeld

Corresponding author e-mail: uch5678@gmail.com

Glacial acceleration has been identified as a main source of uncertainty in sea-level rise assessment in the current realm of climatic warming, as established in the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. In this presentation, we investigate the hypothesis that changes in surface roughness may be indicators of glacial acceleration and discriminate causes of acceleration. The hypothesis is motivated by the following analytical background. Using ice-surface-height data collected with the Geoscience Laser Altimeter System (GLAS) during NASA’s ICESat Mission in 2003–09, we derive mathematical parameters that describe spatial surface roughness. Maps of ice-surface roughness are created for the entire Greenland Ice Sheet and several regions of interest in two ways: (1) along-track for the GLAS sub-satellite tracks and (2) interpolated onto DEMs. Investigation of roughness-change maps leads to the interesting observation that in earlier ‘GLAS’ years, roughness-change signals appear to flicker for most glaciers, but over time almost all regions switch to the same roughness-change signal. A similar behavior was observed for acceleration and retreat, with a regional dependence that may suggest a cause associated with warming ocean currents, but exceptions exist. To answer the hypothesis, we further undertake regional studies of roughness change and elevation change, compared to velocity data, satellite image data and ocean temperature data. To highlight the expected potential of NASA’s future ICESat-2 mission (launch 2017) for the use of spatial ice-surface roughness as an indicator variable, spatial surface characteristics of crevassed glaciers in northeastern and western Greenland are derived from data from two ICESat-2 airborne simulator instruments, the Slope Imaging Multi-polarization Photon-counting Lidar (SIMPL) and the Multiple Altimeter Beam Experimental Lidar (MABEL) and contrasted with the much lower-resolution information available from GLAS.


Icefin: a new polar autonomous underwater vehicle/remotely operated undewater vehicle for sub-ice exploration

Britney Schmidt, Matthew Meister, Anthony Spears, Mick West, Catherine Walker, Jacob Buffo

Corresponding author: Britney Schmidt

Corresponding author e-mail: britneys@eas.gatech.edu

The Icefin autonomous underwater vehicle (AUV) was designed to enable long-range oceanographic exploration of physical and biological ocean environments in ice-covered regions. The vehicle is capable of surveying under-ice geometry, ice and ice–ocean interface properties, and water-column conditions beneath the ice interface. It was developed with both cryospheric and planetary-analog exploration in mind. The first Icefin prototype was successfully operated in Antarctica in austral summer 2014. The vehicle was deployed through a borehole in the McMurdo Ice Shelf near Black Island and successfully collected sonar, imaging, video and water-column data down to 450 m. Icefin’s prime driver was to lower the logistical impact of AUV/ROV operation and make such experiments possible in the deep field. The vehicle has a modular design in order to create more field flexibility and portability. Each module is designed to perform specific tasks, dependent on the mission objective. Vehicle control and data systems can be stably developed and power modules added or subtracted for mission flexibility. Multiple sensor bays can be developed in parallel to serve multiple science objectives. This design enables the vehicle to have greater science capability with operational simplicity compared to larger vehicles. Icefin can be deployed through boreholes drilled in the ice, or from a boat, or over the ice edge. Thus, Icefin can easily access sub-ice environments while maintaining a small form factor and the easy deployment necessary for repeated, low-logistical-impact field programs. The current Icefin prototype is 10.5 in (26.7 cm) wide by 10 feet (~3 m) long and weighs 240 pounds (109 kg). It is comprised of two thruster modules with hovering capabilities, a rear thruster module, an oceanographic sensing module, a main control module and a forward-sensing module with sonar for obstacle avoidance and CT sensor for water profiles. The oceanographic sensing module is fitted with a side scan sonar, altimetry profiler and Doppler velocity log with current profiling. Icefin is depth-rated to 1500 m and is equipped with 3.5 km of fiberoptic Kevlar-reinforced cable, which provides point-to-point communications as well as a stable recovery platform between missions. Icefin was designed and built at Georgia Tech, under Schmidt’s startup funds with effort contributed by Georgia Tech Research Institute.


Basal terraces beneath Totten Glacier, East Antarctica

Jamin Greenbaum, Donald Blankenship, Dustin Schroeder, David Gwyther, Duncan Young, Laura Lindzey, Jason Roberts, Roland Warner, Tas Van Ommen

Corresponding author: Jamin Greenbaum

Corresponding author e-mail: jamin@utexas.edu

The Totten Glacier is the primary outlet of the Aurora Subglacial Basin (ASB), draining at least 3.5 m of eustatic sea-level potential into the Sabrina Coast alongside the Moscow University Ice Shelf. Recent work has shown that the ASB has drained and filled many times since large-scale glaciation began, including evidence that it collapsed during the Pliocene. Steady thinning rates near the grounding line of Totten Glacier are the largest in East Antarctica and the nature of the thinning suggests that it is driven by enhanced basal melting due to ocean processes. Warm modified circumpolar deep water (MCDW), which has been linked to glacier retreat in West Antarctica, has been observed in summer and winter on the Sabrina Coast continental shelf by multiple marine expeditions and recent work has shown that depressions in the seafloor likely make the ice shelf vulnerable to MCDW intrusions that could cause enhanced basal melting. Here we use post-processed, focused airborne radar observations of the Totten Glacier Ice Shelf to delineate multi-kilometer-wide basal channels and flat basal terraces that are associated with high basal reflectivity and specularity anomalies and correspondingly large ice surface depressions that indicate active basal melt processes. Sub-ice-shelf ocean circulation modeling and under ice robotic observations of Pine Island Glacier Ice Shelf in West Antarctica and the Petermann Glacier Ice Shelf in Greenland have shown that basal terraces associated with large basal channels are an indication of rapidly melting ice shelves. Therefore, these new results identify an East Antarctic example of rapid basal melting processes and demonstrate that airborne radar can be used to identify basal characteristics relevant to ice-shelf basal processes.


Greenland englacial drainage of water transport through a fractured aquifer

Timothy Creyts, Andrew Fountain, Ute Herzfeld

Corresponding author: Timothy Creyts

Corresponding author e-mail: tcreyts@ldeo.columbia.edu

Recently, the subglacial hydrology of glaciers and ice sheets has garnered intense interest because of its control of ice sliding and potential ice sheet responses as climate warms. Less attention has been given to the englacial water system that connects surface meltwater sources to the basal drainage system. Observations of englacial drainage reveal diametrically opposed behaviors, where the englacial connections either enhance or limit subglacial processes, including sliding. Some observations show cases where water drainage is mainly through an englacial system of fractures so that water flow at the bed is stunted. Other observations show static englacial water systems that play little role in drainage, with primary drainage routes being along the bed. Here, we use a thermomechanical model of englacial water flow to understand the interaction between ice and water along these connections. We assume that water flow is through a series of connected fractures analogous to crevassed Greenland outlet glaciers, and our model is constrained by surface-fracture densities from Jakobshavn Isbræ, Greenland. The fractures are modified by ice flow, and freezing and melting of the water system. Simple mathematical analyses show trade offs between closure rates and melting rates that determine the englacial flowpaths. From numerical experiments, we show that the dominance of englacial flow follows the locations of both bed overdeepenings and areas where the basal water system is compressed dynamically. The preponderance of overdeepenings in Greenland suggests that englacial systems may be favored in critical areas of ice-sheet flow. We conclude by relating the insights from the analytic and numerical results to the broad-scale patterns of change of the Greenland Ice Sheet with particular attention to the major drainage outlets.


Surface speed estimation of Rayner Glacier in the 1960s from ARGON optical imagery

Wenkai Ye, Rongxing Li, Gang Qiao, Xuwen Ma, Song Guo, Zeyang Wang, Xiaohua Tong

Corresponding author: Rongxing Li

Corresponding author e-mail: rli@tongji.edu.cn

Changes in glacier surface motion and elevation, as critical characteristics to evaluate the pattern of ice-sheet dynamics, have been well investigated. Long-term observation records are desired to help unveil the nature of the interaction between the Antarctic Ice Sheet and the global climate. However, few data, including in situ measurements and remote sensing data, are available before the 1970s. Fortunately, three missions of the ARGON program captured scenes covering the continent of Antarctica in the 1960s, which provide the possibility of extending the time period of Antarctic ice sheet observations. A novel method is proposed in this paper to permit simultaneous estimation of glacier surface motion and elevation from stereo ARGON image pairs via a parallax decomposition approach. An experiment involving Rayner Glacier is taken as an example. Specifically, a novel hierarchical image-matching strategy is developed to obtain correspondences on both moving and stable ice surfaces, and the total horizontal parallax of each correspondence is then separated into terrain- and motion-based components with the aid of the ice-flow direction. The image pyramid is established and a repeat process is involved. As an essential element therein, directional information is acquired from ASTER GDEM first, and is updated continually by the newly obtained terrain model. Finally, two point sets are yielded, which represent the terrain and ice flow field of 1963, respectively. Comparison with the ice-velocity map published by Rignot et al. (2011) is made, and the spatial speed pattern appears to be similar, including both the main glacier and tributary ice flows. The terrain point set is compared as well and it turns out an elevation difference of a few meters with ASTER GDEM. Further detailed analysis of the speed and elevation changes is being conducted and possible reasons for the differences and changes are discussed.


Modeling of ocean-induced ice melt rates of five Greenland glaciers over the past two decades

Yun Xu, Dimitris Menemenlis, Michiel van den Broeke, Eric Rignot

Corresponding author: Yun Xu

Corresponding author e-mail: erignot@uci.edu

High-resolution, three-dimensional simulations from the MITgcm ocean model are used to model the subaqueous melt rate of the near-vertical calving faces of Umiamako, Rink, Kangerdlugssuaq, Store and Kangilerngata glaciers, west Greenland, from 1992–2015. Model forcing is provided by monthly reconstructions of ocean state and ice-sheet runoff. Results are analyzed in combination with new observations of bathymetry, bed elevation, ice-front retreat and glacier speed. We find that subaqueous melt is 2–3 times greater in summer than in winter, and has doubled in magnitude since the 1990s due to enhanced runoff production and a 1.5°C warmer ocean temperature. Umiamako and Kangilerngata glaciers retreated rapidly in the 2000s as summer subaqueous melt rates exceeded the glacier calving speeds and the bed geometry favored a rapid retreat. In contrast, Store, Kangerdlugssuaq and Rink glaciers have remained stable because their subaqueous melt rates remain 4–5 times lower than the calving speeds. This study demonstrates the complex interplay between subaqueous melting speeding, calving speed and bed geometry in controlling the evolution of Greenland glaciers and provides a simple framework for modeling ice–ocean interaction at calving margins.


Recent local coastline retreat in some western Antarctic regions

Yixiang Tian, Da Lv, Hexia Weng, Xiaohua Tong, Rongxing Li

Corresponding author: Yixiang Tian

Corresponding author e-mail: tianyixiang@tongji.edu.cn

The Antarctic ice sheet is known as one of the most sensitive parts of the Earth system subjected to the impact of global climate change. It plays an important role in the study of global sea-level change. The dynamics of the ice-sheet margin are reflected by grounding-line changes of ice-shelf areas and coastline changes in non-ice-shelf areas. Non-ice-shelf areas are thought to be stable. For instance, the positions of ice walls on a map such as the coastal change and glaciological map of Antarctica are presented as temporal changing features since the magnitude of change on an annual to decadal scale is generally not discernible on Landsat images. Based on coastline products (Radarsat-1 coastline, MOA2004 coastline and MOA2009 coastline), together with Landsat images captured at summer seasons from 1970s, we investigated two types of coastline change over the last 10 years: (1) retreat of the glacier outlets in the Antarctic Peninsula; and (2) retreat along some of the ice walls in West Antarctica. The ice-front change is mostly attributed to the imbalance between the seaward ice flux and break-back calving, which is related to many factors. A uniform retreat of Antarctic Peninsula glaciers since the 1970s, with a period of small re-advance in the late 1990s, was concluded. Based on the above-mentioned data, it was observed that significant changes (accelerated retreating) had occurred in a few marine-terminating glacier outlets over the last 10 years. To use Wilkinson Glacier outlet as an example, its coastline position did not change obviously between 1987 and 2007. However, it retreated up to 1800 m after 2008. It was reported that, based on satellite altimetry, there has been rapidly accelerated thinning of the glaciers in the southern Antarctic Peninsula since 2009. The recorded coastline changes may have been caused by the combined effect of this thinning and ice-front calving processes. In an ice-wall region in the Bryan Coast, West Antarctica, between Fox Ice Stream and Ferrigno Ice Stream (83°45′–84°57′ W, 73°30′–73°33′ S), along a coastline of about 62 km there has been an average retreat of 2.2 km, with a maximum of about 11 km, between 1988 and 2015. The coastline was stable from 1988–2004; retreat mainly happened 2004–15. The mechanism for such a large local change is being investigated.


Calculation of freshwater melt runoff from the Greenland Ice Sheet from ICESat measurements

H. Jay Zwally, Jun Li, Brooke Medley, John Robbins, Donghui Yi

Corresponding author: H. Jay Zwally

Corresponding author e-mail: jayzwallyice@verizon.net

Time-series of surface elevation, H(t), are derived from ICESat measurements from 2003–09. The seasonal cycle of H(t) in the ablation zone is characterized by (1) a linear upward component evident during winter (fall through spring) that is caused by the net ice flow from above the EL minus the ice discharge at the ice margin plus winter snow accumulation and (2) a superimposed downward component caused by surface melting during the summer. The H(t) are fitted with a multi-variate function with a constant linear component describing the upward component of ice flow (plus a term for winter snow) and a portion of a sine function describing the superimposed summer melting with variable interannual amplitude. By DS, the H(t) for the ablation zones show accelerations of the mass losses in the northern DS1 and the northwest DS8 that are mostly offset by decelerations of the mass losses in the east-central DS3 and the southeast DS4. Over all of the Greenland ablation zone the upward linear component is 1 m a–1, which is adjusted to account for the contribution from winter snow accumulation using net snow estimates from meteorological models, to estimate the net ice inflow into the ablation zone. Preliminary estimates of the melting and runoff for the summers of 2004–09 are respectively 265, 354, 265, 371, 212 and 380 Gt a–1 with an average of 308 Gt a–1. The net mass losses from the ablation zone range from a low of 35 Gt a–1 in 2008 to a high of 203 Gt a–1 in 2009 with an average loss of 131 Gt a–1, which is in good agreement with the average 133 Gt a–1 obtained from the along-track dH/dt solutions. Our analysis provides a unique partitioning of the net dynamic input to the ablation zone (inflow at the EL minus ice discharge at the grounding lines) and the runoff losses from melting.


Seasonal connections between Nares Strait and Petermann Glacier: linking the sea ice, ocean and ice shelf

Emily Shroyer, Laurence Padman, Roger Samelson

Corresponding author: Emily Shroyer

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

Petermann Glacier drains ~6% of the Greenland Ice Sheet through a ~50 km-long ice shelf in Petermann Fjord that opens into Nares Strait to the east of Ellesmere Island. About 80% of the glacier mass loss occurs by basal melting of the ice shelf, with the remainder by iceberg calving, sublimation and surface runoff. Here we describe seasonality in basal melting using a high-resolution coupled ocean/sea-ice model with thermodynamic interaction between the ocean and the ice shelf. The rate of mass loss by basal melting increases by ~25% in summer, with the increased loss primarily restricted to regions of the ice shelf less than ~200 m thick. The increase in basal melt is directly linked to the seasonal shift in ocean circulation within Nares Strait, which is tied to the transition between mobile and landfast sea ice caused by the formation of an ice arch across the Strait’s southern boundary. Under mobile sea ice, warm subsurface waters lie on the eastern side of the Strait at the mouth of Petermann Fjord and instabilities in the flow through Nares Strait effectively flush water through the fjord. Our results suggest that an increase in the length of season for which Nares Strait ice is mobile or absent will lead to more rapid mass loss from Petermann Ice Shelf.


New observations of upward turbulent heat fluxes in the Arctic Ocean

Jennifer MacKinnon, John Mickett, Matthew Alford

Corresponding author: Jennifer MacKinnon

Corresponding author e-mail: jmackinnon@ucsd.edu

Arctic sea ice is retreating at a faster rate than predicted by many models. One hypothesis for enhanced melt rate is that lateral water mass motion, circulation and turbulent mixing within the Arctic ocean allow heat to be brought up from below, towards melting sea ice. In September 2015 the NSF-funded ArcticMix project investigated turbulent mixing and associated heat fluxes associated with a variety of phenomena near remnant multi-year sea ice in the Beaufort Sea. Upward turbulent heat fluxes from sub-surface oceanic heat reservoirs were dramatic – at times larger than solar heat fluxes from above. The strength of observed turbulence was likely set by a complex interplay between direct wind forcing, near-inertial shear and sub-mesoscale interleaving and re-stratification.


Controls on circulation, cross-shelf exchange and dense water formation in an Antarctic polynya

Kate Snow, Bernadette Sloyan, Stephen R. Rintoul, Andy Hogg, Stephanie Downes

Corresponding author: Kate Snow

Corresponding author e-mail: ksnow@ed.ac.uk

The circulation and exchange of water masses across the Antarctic continental shelf break influence the global overturning circulation and the delivery of heat fluxes to the floating ice shelves fringing the Antarctic ice sheets. Processes controlling the shelf circulation and exchange, however, remain poorly understood. Using a unique long-term time series of hydrographic observations on the Terre Adelie continental shelf, East Antarctica, the relationship between surface forcing, circulation, bottom water formation and the exchange of water masses across the shelf break is assessed. Employing inverse box model methods to derive the regional circulation, seasonal variations in surface wind and buoyancy forcing are revealed to induce significantly different winter and summer circulation and cross-shelf exchange. During summer, the southeasterly katabatics drive a northwestward wind-driven coastal current. During winter, buoyancy losses drive the formation and off-shelf flow of dense shelf water. Mass conservation on the shelf leads to a compensating on-shelf flow of modified circumpolar deep water (MCDW). The winter on-shelf flow of MCDW is an order of magnitude larger than that in summer, highlighting the importance of surface buoyancy fluxes in controlling the strength of the cross-shelf exchange. Increased understanding of the role of surface fluxes in controlling shelf circulation aid in developing improved understanding of the drivers of the Antarctic shelf exchange, and ice–ocean interactions.


Asynchronous behavior of outlet glaciers feeding Godsthåbfjord (Nuup Kangerlua) and the retreat of Narsap Sermia in southwest Greenland

Roman Motyka, Ryan Cassotto, Martin Truffer, Kristian Kjeldsen, Dirk van As, Niels Korsgaard, Mark Fahnestock, Ian Howat, Peter Langen

Corresponding author: Roman Motyka

Corresponding author e-mail: rjmotyka@uas.alaska.edu

We assess ice loss and velocity changes of tidewater and land-terminating glaciers that drain into inner Godthåbsfjord (Nuup Kangerlua) in southwest Greenland using digital elevation models, airborne lidar data and satellite imagery. We first focus on the asynchronous behavior of two tidewater glaciers, Kangiata Nunaata Sermia (KNS) and Narsap Sermia (NS). Volume loss in the entire fjord region between 1985 and 2008 was 29.1 km3 with KNS accounting for nearly half of this loss. An additional 14.7 km3 was lost between 2008 and 2014, with 74% of the loss from NS. The losses are equivalent to 0.10 mm eustatic sea-level rise. KNS has retreated 22 km from its Little Ice Age (LIA) maximum since 1761 AD with terminus velocities currently ranging from 5–6 km a–1. KNS appears to have stabilized after retreating onto shallow terrain. In contrast, NS began retreating from its LIA moraine in 2004 (0.6 km), re-stabilized, then retreated 3.3 km between 2010 and 2014 into an overdeepened basin; terminus velocity increased from 1.5 km a–1 to 5.5 km a–1. We assess potential drivers of the observed changes, including glacier thinning, runoff, surface mass balance, ocean conditions, submarine melting, bed topography and ice melange. The 2010–14 NS retreat appears to have been triggered by a combination of all of the preceding factors.


Creep and stick–slip in subglacial granular beds forced by ocean tides

Anders Damsgaard, David L. Egholm, Lucas H. Beem, Slawek Tulaczyk, Nicolaj K. Larsen, Jan A. Piotrowski, Matthew R. Siegfried

Corresponding author: Anders Damsgaard

Corresponding author e-mail: anders.damsgaard@geo.au.dk

Observations show that the flow velocity of several marine-terminating glaciers and ice streams is strongly linked to tidal stage. Deformation of subglacial sediments often accounts for much of the flow, but the mechanical behavior of the sediment remains poorly understood. Measurements and laboratory experiments provide important constraints, but existing models have not been able to explain the internal processes driving transitions from stability to stick–slip within the sediment. In this presentation we use a coupled numerical model of grain and fluid dynamics to show that rapid rearrangements of load-bearing force chains within the granular sediments drive the mechanical transitions. Cyclic variations in driving stresses or pore-water pressure, caused by ocean tides, give rise to strain-rate-dependent creeping motion at stress levels below the point of failure, while disruption of the force-chain network induces fast strain-rate-independent flow above it. This finding contrasts with previous descriptions of subglacial sediment mechanics, which either assume a rate-dependent rheology regardless of mechanical state or unconditional stability before the sediment is stressed to a yield point. Our new micro-mechanical computational approach is capable of reproducing important transitions between these two end-member models, and it can explain multimodal velocity patterns observed in marine-terminating glacier and ice stream systems.


Do icebergs matter to fjord circulation?

David Sutherland, Twila Moon, Jonathan Nash, Emily Shroyer, Leigh Stearns, Ginny Catania

Corresponding author: David Sutherland

Corresponding author e-mail: dsuth@uoregon.edu

Icebergs are ubiquitous in tidewater glacier fjords, representing a freshwater flux to the ocean in addition to submarine melt, subglacial discharge and surface runoff. However, the impact of icebergs on fjord circulation has received little attention. Previous studies have focused on basin-scale iceberg distributions, e.g. the North Atlantic subpolar gyre, or on ice melange, which can more directly affect glacier stability. Inside Greenland’s fjords, however, iceberg residence times can vary from months to years, during whic their melt can alter water column stratification and potentially create a buoyancy-driven circulation. Thus, quantifying the variability in iceberg residence times, estimating the magnitude of melt and predicting where this melt ends up in the water column are all important issues. Here, we focus on icebergs within the Helheim Glacier/Sermilik Fjord system in southeast Greenland, using data that includes seasonal ice discharge, modeled subglacial discharge, residence times from remotely-tracked icebergs, and in situ hydrography. We compare the Sermilik/Helheim system, with its deep, long fjord and relatively warm Atlantic-origin waters, to two west Greenland systems in Uummannaq Bay that have distinctly different characteristics. We find that iceberg meltwater fluxes vary substantially across these systems; in Sermilik Fjord iceberg melt may be as significant as subglacial discharge. The implications are twofold: (1) by changing stratification iceberg melt can change where subglacial-discharge-driven plumes reach their neutral buoyancy, and (2) icebergs make it difficult to isolate the source of meltwater using hydrography alone.


Water properties and circulation in front of tidewater glaciers in northwestern Greenland

Masahiro Minowa, Shin Sugiyama, Yoshihiko Ohashi, Takanobu Sawagaki, Shun Tsutaki, Daiki Sakakibara, Evgeny Podolskiy, Yvo Weidmann, Shigeru Aoki

Corresponding author: Masahiro Minowa

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

Tidewater glaciers in Greenland are rapidly retreating, and their mass loss has contributed to global sea-level rise over recent decades. Previous studies suggested the importance of submarine melting, but physical processes relevant to submarine melting (e.g. heat source, ocean circulation and bathymetry) are not well understood. This is because in situ ocean observations are difficult in front of a calving glacier where icebergs cover the ocean. To better understand the ocean environment and processes near the calving front, we measured temperature, salinity, turbidity and bathymetry in front of Bowdoin and Sun Glaciers, tidewater glaciers in northwestern Greenland, in summer 2014 and 2015. We also carried out high temporal (10 s) time-lapse photography in front of Bowdoin Glacier in July 2015. In the fjord in front of Bowdoin Glacier, Atlantic water (AW) filled the depth below 250–280 m. The mean temperature and salinity within the AW layer was 1.1–1.2°C and 34.4–34.5 g kg–1. The results suggest that warm water flows into the Bowdoin Fjord from the open ocean and acts as the heat source for submarine glacier melting. CIn contrast to these observations in front of Bowdoin, only fresh and cold water was found in front of Sun Glacier. It is likely that the relatively shallow fjord (~100 m) and the existence of a sill (at ~10 m) inhibit warm water from entering from the open ocean. In the vicinity of Sun Glacier (~200 m from the front), water properties were very different from those in the open ocean. Water was highly turbid, fresh and cold, suggesting subglacial discharge of meltwater as the origin of the water. Time-lapse photographs revealed near-surface fjord circulation near the glacier. The observations confirmed that the circulation was driven by a buoyant plume, which was generated by subglacial discharge and/or submarine glacier melting. The buoyant plume was consistently observed in the region from the calving front to approximately 5 km offshore in early July, whereas it was visible only near the front in late July. This change is probably due to the amount of subglacial discharge and the stability of near-surface stratification.


Simulating the upward transport of sediment-derived iron associated with the overturning circulation in ice-shelf cavities of the Amundsen Sea, Antarctica

Pierre St-Laurent, Michael Dinniman, Patricia Yager, Robert Sherrell, Sharon Stammerjohn

Corresponding author: Pierre St-Laurent

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

Recent studies identified ice-shelf meltwater production to be a key parameter explaining differences in the mean chlorophyll concentration between Antarctic coastal polynyas. These biological ‘hot-spots’ are typically replete in macro-nutrients but stressed/limited by iron. The basal melt of ice shelves can enhance the biological production by (1) the release of iron associated with scouring on the bedrock and/or aeolian material accumulated in glacial ice, and (2) the upward transport of deep (sediment-derived) iron inside the ice-shelf cavities. This second mechanism is particularly vigorous in western Antarctica where warm modified circumpolar deep water (mCDW) comes in direct contact with the ice shelves and drives a buoyancy-driven overturning circulation inside the cavities (a warm type of ‘ice pump’). In this poster we focus on this mechanism in the context of the Amundsen Sea polynya (ASP), an area characterized by high biological productivity and fast-melting ice shelves (Dotson, Thwaites and Pine Island ice shelves). We use a realistic three-dimensional sea-ice–ice-shelf–ocean coupled model to simulate the ocean circulation on the continental shelf and within the cavities. Dissolved iron from bottom sediments (dFe) is represented with a passive Eulerian tracer whose concentration at the sea floor is prescribed according to observations. The tracer then evolves following the ocean advection/diffusion, while the biogeochemical processes such as algal uptake are neglected for simplicity. We highlight the effects of the overturning circulation by comparing numerical experiments where this ‘ice pump’ is active/inactive. The pump is effective in importing the deep iron inside the cavities before releasing it back to the continental shelf over a broad range of depths. This mechanism produces in the ASP a general increase in the iron concentration of the intermediate and upper water column, a thickening of the bottom boundary layer, and a large horizontal gradient near the ice shelf front. These features are compared to observed iron distributions from the ASPIRE expedition. The results suggest that the upward transport of deep iron can be partly responsible for the enhanced biological production of certain Antarctic coastal polynyas.


Simulating heat and salt exchange at the ice–ocean interface in semiconvective free convection

Thomas Keitzl, Juan-Pedro Mellado, Dirk Notz

Corresponding author: Thomas Keitzl

Corresponding author e-mail: t@keitzl.com

We use direct numerical simulation of turbulent convection beneath a horizontal, smooth ice–ocean interface to determine the ratio of heat and salt flux in the semiconvective regime. This flux ratio determines the interfacial conditions of an ice–ocean interface. We find that the flux ratio is almost independent of the far-field conditions of the flow and three times as large as previously assessed on the basis of turbulent-flux measurements in the field. As a consequence, melt rates that are given based on the flux ratio have so far been overestimated by up to 40%. Our simulations indicate that direct measurements of the flux ratio based on the turbulent fluxes will be difficult. As an alternative, we present a consistent evaluation of the flux ratio based on the total heat and salt fluxes.


Airborne radar observations of winter snow accumulation on the Larsen C Ice Shelf, Antarctica

Brooke Medley, Daniel McGrath, Peter Kuipers Munneke, Nathan Kurtz

Corresponding author: Brooke Medley

Corresponding author e-mail: brooke.c.medley@nasa.gov

The large ice shelves surrounding the Antarctic continent buttress inland ice, limiting the grounded ice-sheet flow. Determination of their mass balance provides an indicator as to the future of the shelves buttressing capability; however, measurements of surface accumulation are few, limiting the precision of the mass balance estimates. The Larsen C Ice Shelf (LCIS) is the largest ice shelf in the Antarctic Peninsula, located its eastern side. Its location on the climatological leeside of the mountains produces particular patterns in its surface mass balance, especially in surface melt and snow accumulation. Focusing on the latter, we measure winter snow accumulation using the Center for Remote Sensing of Ice Sheets (CReSIS) Ku-band radar data from 2009 and 2010, flown as part of NASA’s Operation IceBridge (OIB). Over a majority of the airborne survey, a single strong reflection is found in the shallow radar data, which likely represents a dielectric contrast due the presence of ice lensing created during the summer melt season. Therefore, the depth of this strong reflection provides a measurement of the amount of snow that has accumulated since the end of melt (i.e. fall/winter/spring accumulation). We confirm this assumption through comparison with in situ sonic ranger snow height changes between the end of melt and collection of the OIB survey at four sites. The survey has excellent spatial coverage because the two flights were specific to the LCIS in order to coordinate with field efforts in 2009. Over the ~3200 km of measurements, the average snow depth is 1.07 m (σ = 0.17 m). Due to the complexities of the local, surrounding topography and its leeside location, the spatial pattern of accumulation, as well as melt, is complex. To evaluate its performance in replicating the observed pattern of accumulation, we compare the radar-derived accumulation rates with those from a high-resolution regional climate model (RACMO2.3).


Variability of warm Southern Ocean water on the East Antarctic shelf during the NBP1503 cruise

David Porter, Frank Nitsche, Raul Guerrero, Guy Williams, Eva Cougnon, Alex Fraser, Ricardo Correia

Corresponding author: David Porter

Corresponding author e-mail: dporter@ldeo.columbia.edu

East Antarctica has long been considered a relatively stable ice sheet but in light of recent studies showing thinning of ice shelves there, it has become the focus of a number of efforts to determine its susceptibility to ocean warming. We present new hydrographic and bathymetric observations from the East Antarctic shelf and margin between the Dibble and Totten ice shelves from the scientific cruise NBP1503 aboard NB Palmer in early 2015. Conductivity–temperature–depth (CTD), multibeam echosounding bathymetry and shipboard acoustic Doppler current profiler (SADCP) data were collected from on and off the continental slope north of the Dibble Glacier, Frost Glacier, Dalton Iceberg Tongue and Totten Glacier. A section of 19 CTD profiles on the continental slope and parallel to the coast shows that modified circumpolar deep water (mCDW) is present near the shelf break over large areas of the margin but absent in others. The depth of the shelf break varies significantly along the margin with values between ~300 and ~500 m. The shallow depths may be obstacles for the flux of heat (via mCDW) onto the shelf and up to the ice–ocean interface. As a result of its larger depth, a ~100 m thick layer of mCDW resides on the ~500 m deep outer shelf north of the Totten Glacier even though no clear trough on the central shelf has yet been identified. The shelf break north of Dibble Ice Shelve is shallower and so mCDW is not observed there. Leveraging coincident SADCP and CTD data, meridional heat flux estimates corroborate the source of water masses on the shelf. The presence of mCDW water and a generally deep shelf break, especially north of Totten Glacier and the Moscow University Ice Shelf, has important implications for the future response of these ice shelves to ocean forcing.


The feasibility of imaging subglacial water systems near the grounding zone using electromagnetic soundings

Kerry Key, Matthew Siegfried

Corresponding author: Kerry Key

Corresponding author e-mail: kkey@ucsd.edu

Subglacial hydrologic systems in Antarctica and Greenland play a fundamental role in ice-sheet dynamics and their susceptibility to changes in response to system-scale perturbations induced by climate change. However, critical aspects of these systems remain poorly understood largely due to a lack of suitable observational constraints. In particular, the mixing of subglacial freshwater and saline ocean water at grounding zones remains a poorly understood detail of ice–ocean interactions in this critical transition region. Ground-based electromagnetic (EM) geophysical methods are well established for mapping groundwater hydrology in many geological environments, but such methods have never been applied to imaging groundwater beneath thick ice sheets nor have they been applied at the grounding zones of major ice streams. Since EM data are directly sensitive to groundwater via its impact on bulk electrical conductivity, EM soundings are highly complementary to other geophysical methods that have been applied for subglacial characterization, such as seismic reflection profiling, radar soundings and gravity mapping. This work exams the feasibility of both passive and active source EM techniques for quantifying the nature of subglacial water in lake systems and at grounding zones, with a particular emphasis on studying the interfaces between ice, ocean and basal meltwater, as well as groundwater systems in the underlying sediments. A suite of systematic model studies exams the sensitivity of EM data as a function of ice thickness, water conductivity and hydrologic system geometry. By combining information on the conductivity structure of the subglacial system from EM soundings with ice thickness and geological structure from established geophysical techniques for glaciology, we are able to provide better physical constraints on the interaction between ice, rock and water at the ocean boundary of ice streams and outlet glaciers.


Viscoelastic response of shear weakening due to periodic drainage of water-filled crevasses

John Cavanagh, Derrick Lampkin, Ryan Walker

Corresponding author: Derrick Lampkin

Corresponding author e-mail: djlumd@gmail.com

The impacts of surface melt water on ice dynamics via supraglacial lake drainage and runoff have been well documented. However, little attention has been focused on the impact of direct melt water injection into the shear margins of fast flowing, marine-terminating outlet glaciers or on the ability of present modeling capabilities to address such processes. Even less attention has been given to seeking to quantify the viscoelastic response of ice sheets to rapid injection of surface melt water to the base. It is possible that injection of surface melt water produces a rapid impulse response in ice flow, before returning to its base flow state, that would likely be undetected by most satellite imagery due to the short timescale. Here, the pulsed response nature of this process is considered by applying an idealized version of drainage perturbations for the summers between May 2007 and September 2015 to a viscoelastic model. We use the model to rapidly increase subglacial water pressurization caused by drainage of seven water-filled crevasses along the Jakobshavn Isbræ shear margin and trace the dynamical response along four 10 km flow lines on the southern margins of the ice stream. This works seeks to quantify the ice dynamic response to surface melt drainage and the anticipated length of time such perturbations can be expected to affect ice flow. This work will also assist in understanding how future ice flow will be affected by prolonged duration of water-filled crevasses due to global warming.


Modeling turbid meltwater plume and associated sediment transport

Yoshimasa Matsumura, Yoshihiko Ohashi, Shigeru Aoki, Shin Sugiyama

Corresponding author: Yoshimasa Matsumura

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

We have performed a numerical experiment of a turbid plume induced by meltwater runoff from a marine-terminating glacier. To explicitly simulate the dynamics of turbidity current, a non-hydrostatic ocean model coupled with a built-in Lagrangian particle tracking system is used. The model domain is an idealized rectangular fjord 10 km × 3 km × 600 m deep, and we set a tunnel-like moulin at the grounding line (500 m depth) of the upright glacier front. The initial stratification is set to idealize the typical summertime Greenland fjord where the cold and less saline polar water layer exists over the warm and saline Atlantic water, while the surface is covered by the warm and fresh surface water. The freshwater runoff is applied into this tunnel-like moulin with various discharge fluxes from 500 m3 s–1 to 2000 m3 s–1, and we also impose sediment particle supply of 1000 kg s–1 inside the tunnel. Since the discharged fresh water is significantly less dense than seawater, strong vertical motion is induced at the exit of the moulin and an upwelling plume is formed along the side of the glacier front. In the relatively greater discharge experiments, the imposed sediment particles are captured inside this upwelling plume and reach the surface; it then forms a similar high-turbidity region to that observed in Greenland fjords. The upwelling plume induces overturning circulation inside the fjord, whereby the upper layer flows offshore and a compensating onshore current exists below. Some portion of imposed sediments are transported offshore by this overturning, mainly along the right-hand wall, due to the Coriolis force. How far the sediment particles are transported offshore from the glacier front depends on the size of the particles; particles <10 μm can be transported to the end of the domain but particles of greater size tend to be deposited near the glacier front. The least-discharge experiment shows a qualitatively different result, where almost all sediment particles cannot reach the surface but tend to stay at the PW–AW layer because buoyancy forcing is not enough in this case to lift sediment to the surface.


Inversions of NASA IceBridge gravity anomalies over the Antarctic Peninsula

David Porter, Kirsty Tinto, James Cochran

Corresponding author: David Porter

Corresponding author e-mail: dporter@ldeo.columbia.edu

We use the full suite of data acquired during NASA Operation IceBridge (OIB) campaigns, including the ice-penetrating radar, the range-finding lidar and the airborne gravimeter, to calculate forward models of gravity anomalies over the Antarctic Peninsula. Glacier characteristics and flight patterns over the Peninsula are similar to OIB fjord-axis flights in Greenland, allowing for quick transference of our expertise to this new region. This also provides an opportunity to enhance and extend the method, resulting in streamlined gravity modeling procedures and increased reproducibility. By extending our experience from generating bathymetry models in Greenland, we have quickly developed a repeatable and robust methodology for gravity inversions over the Antarctic Peninsula. NASA Operation IceBridge (OIB) collected geophysical data from flights along the centerlines of glaciers on the Antarctic Peninsula in 2010/11. The gravity data from these lines along the axes of glaciers, ice streams and fjords provide a means to constrain the bathymetry and bed topography near the ice margin. After surveys are separated into glacier-axis lines, a first-pass inversion using homogeneous densities provides a base model for more targeted, glacier-by-glacier inversions. This newly developed, semi-automated method generates input fields and model horizons for entire regions, allowing for quick prototyping and testing of techniques. After this regional processing step, individual models can be fine-tuned to include additional local constraints, such as corrections for regional long-wavelength gravity anomalies and offshore seafloor depth measurements. Inclusion of glacier-specific constrains are shown to extend and reduce uncertainty in gravity-derived bathymetry models.


A large, rapid subglacial lake drainage beneath Slessor Glacier, East Antarctica, and its potential impact in the Filchner Trough

Matthew R. Siegfried, Dustin M. Schroeder, Ted Scambos, Sasha P. Carter, Helen A. Fricker

Corresponding author: Matthew R. Siegfried

Corresponding author e-mail: mrsiegfried@ucsd.edu

Active subglacial lakes in Antarctica can control the spatiotemporal variability of freshwater flux across the grounding line and into sub-ice-shelf cavities, modifying the local ocean circulation and basal melt rates. Due to the short observational record of subglacial lakes in Antarctica, the impact of subglacial lake drainage events on both the ice sheet and the ocean is unclear, especially for large events with recurrence intervals of a decade or longer. We use new satellite and airborne data to study one of the largest, most rapid subglacial lake drainage events ever recorded, on downstream Slessor Glacier, East Antarctica. As the northernmost major outlet glacier flowing into the Filchner Ice Shelf, the location of Slessor Glacier is oceanographically significant because warm water has been hypothesized to first enter the Filchner–Ronne Ice Shelf cavity through the Filchner Trough. We combine satellite laser and radar altimetry with Operation Ice Bridge (OIB) airborne laser altimetry to generate a nearly 13-year (2003–16) record of surface height changes that document an extended inflation phase from 2003 until mid-2014, when the surface fell by nearly 10 m over ~6 months. During this subsidence event, ~2.5 km3 of ice was displaced, with a peak rate over 350 m3 s–1. Using OIB radar sounding data from before and after the drainage event and a velocity record from optical-image feature-tracking, we investigate the hypothesis that this type of surface subsidence event relates to the drainage of large subglacial lake. We then discuss the potential mechanisms of lake drainage and consider the impact of a large, rapid flood of freshwater into the ocean cavity beneath the Filchner Ice Shelf.


Observations of subglacially driven surface meltwater plumes in a Greenland Fjord

Jonathan Nash, Rebecca Jackson, David Sutherland, Emily Shroyer, Dustin Carroll, Dylan Winters, John Mickett

Corresponding author: Jonathan Nash

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

The terminal depth of meltwater plumes derived from subglacial discharge depends on the freshwater flow rate and fjord geometry/stratification. Under the appropriate conditions, these can reach the ocean surface, be identified visually and be opportunistically sampled. Here we investigate the nearfield structure of a surface-intensified plume in Kangerdlugssuaq Sermerssua, a glacier outlet fjord in central West Greenland. Two field seasons of observations using a towed thermistor chain, near-surface-mounted ADCP, remotely operated surface vessel and CTD profiling find a surface plume that varies over short time (hours) and space (kilometers) scales. The plume creates a cold-water anomaly in the fjord, which numerical modeling suggests scales according to the degree of entrainment in the near-terminus rising plume. Our measurements quantify the near-surface characteristics of the plume, such as its T, S and freshwater flux. This can be used as a consistency check on the estimated subglacial discharge, and to test models of plume dilution, both based on simple theory and more complex numerical models. These findings are important because the surface plume and its three-dimensional structure affect circulation within the fjord and can act as a control to open-ocean heat delivery to the glacier face.


Response of Amundsen Sea ice-shelf thickness to El Niño–Southern Oscillation variability

Fernando Paolo, Helen Fricker, Laurie Padman

Corresponding author: Fernando Paolo

Corresponding author e-mail: fpaolo@ucsd.edu

Atmospheric and sea-ice conditions around Antarctica, particularly in the Amundsen and Bellingshausen seas, respond to climate dynamics in the tropical Pacific Ocean on interannual timescales including the El Niño–Southern Oscillation (ENSO). It has been hypothesized that the mass balance of the Antarctic Ice Sheet, including its floating ice shelves, also responds to this climate signal; however, this has not yet been unambiguously demonstrated. We apply multivariate singular spectrum analysis to an 18-year (1994–2012) time series of ice-shelf height derived from satellite radar altimetry in the Amundsen Sea (AS) region. This advanced spectral method distinguishes between regular deterministic behavior (‘cycles’) at sub-decadal timescale and irregular behavior (‘noise’) at shorter timescales. Although the long-term trends in ice-shelf height change are much larger than the range of interannual variability in the AS region, the short-term rate of change dh/dt can vary about the trend by more than 50%. We extract the principal modes of variability (EOFs) based on common spectral properties from a set of 140 height time series. The mode of interannual variability in the AS ice-shelf height is strongly correlated with the low-frequency mode of ENSO (periodicity of ~4.2 years) as represented by the Oceanic Niño Index. This interannual mode in ice-shelf height, represented by the two leading EOFs, is responsible for about 25% of the variance in the de-trended data set. The ice-shelf height in the AS is expected to respond to changes in precipitation and inflows of warm subsurface circumpolar deep water (CDW) into the ocean cavities under the ice shelves, altering basal melt rates. While we find a correlation between modeled precipitation anomalies and ice-shelf height, we are investigating (a) errors in model precipitation, (b) radar backscatter and firn-density issues, and (c) ocean contribution correlated with atmosphere through wind-stress forcing. We will describe the spatial structure of AS ice-shelf height response to ENSO, and attempt to distinguish the precipitation signal from basal mass balance due to changing CDW inflows.


On the role of buoyant flexure in glacier calving

Till Wagner, Timothy James, Tavi Murray, Dominic Vella

Corresponding author: Till Wagner

Corresponding author e-mail: tjwagner@ucsd.edu

We investigate the role of hydrostatic forces in glacier calving. We develop a mathematical model to account for the elastic deformation of glaciers in response to three effects: (1) marine and lake-terminating glaciers tend to enter water with a non-zero slope, resulting in upward flexure around the grounding line; (2) horizontal pressure imbalances at the terminus are known to cause hydrostatic in-plane stresses and downward acting torque; (3) submerged ice protrusions at the glacier front may induce additional buoyancy forces that can cause calving. Our model provides theoretical estimates of the importance of each effect and suggests geometric and material conditions under which a given glacier will calve from hydrostatic flexure. We find good agreement with observations. This work sheds light on the intricate processes involved in glacier calving and can be hoped to improve our ability to model and predict future changes in the ice–climate system.


ROSETTA-Ice: a new framework for understanding the Ross Ice Shelf

Kirsty Tinto, Christine Siddoway, Laurie Padman, Indrani Das, Helen Fricker, Fabio Caratori Tontini, Robin Bell, Rosetta Team

Corresponding author: Kirsty Tinto

Corresponding author e-mail: tinto@ldeo.columbia.edu

The Ross Ice Shelf (RIS) is the largest ice shelf in Antarctica and overlies the most sparsely mapped sea floor on the planet. Understanding the stability of the RIS is a key part of understanding the history and future of the Antarctic ice sheets. The ROSETTA-Ice project is an interdisciplinary program based around an airborne geophysical campaign using the IcePod instrument suite, flown on LC130 aircraft, to comprehensively map the RIS at 10 km resolution over two field seasons. Instruments include shallow and deep ice radars, lidar, gravity and magnetic measurements that will provide a benchmark dataset of the regional geology, ice shelf height, thickness, draft and internal structure, and sea floor height under the ice shelf. Prior to the ROSETTA-Ice surveys, the most comprehensive mapping of the RIS was from the RIGGS project in the 1970s, where seismic shot points measured ice thickness and sea floor depth at 55 km spacing. ROSETTA-Ice therefore represents an order-of-magnitude improvement in our mapping of this region. The first field season was completed in late 2015, acquiring over 30 000 km of geophysical data. We present preliminary results of the survey, including improved resolution of ice thickness and sea floor bathymetry along survey lines. We will also describe proposed flight lines in Year 2 of the program, which will include higher-resolution gravimetry and airborne deployment of oceanographic sensors. When the survey is complete, the ROSETTA-Ice dataset will provide insight into century to millennial variability of ice flow from both West and East Antarctica, improve our understanding of ice–ocean interactions within the ice shelf cavity; and be used to develop our understanding of the role of tectonic and geomorphological processes that sculpt the region’s bathymetry and topography and, therefore, influence the glaciological response of the Ross embayment.


Reconstructing last interglacial ocean temperature from ice core gas records

Sarah Shackleton

Corresponding author: Sarah Shackleton

Corresponding author e-mail: sshackle@ucsd.edu

With global temperatures ~1–2°C above present, but sea level exceeding current levels by 6–8 m, the last interglacial stage (LIG) may provide valuable insight into Earth system constraints under future warming. Ocean temperature records through the LIG are integral to constraining the relative contributions of thermal expansion and ice-sheet collapse to sea-level rise and understanding the conditions leading up to ice-sheet disintegration. Here we present a novel method to reconstruct globally integrated ocean temperature using atmospheric δKr/N2, δXe/N2 and δXe/Kr. Employing an ocean–atmosphere mass-balance model, the changes in the relative proportions of these gases allow us to estimate oceanic heat uptake during this period. This method will ultimately be applied to a ‘horizontal ice core’ at Taylor Glacier, Antarctica covering the LIG. We anticipate that these findings will advance our understanding of the ocean’s role in ice-sheet collapse in the past and future.


Sources and fate of fresh water in the ocean west of the Antarctic Peninsula

Heather Regan, Paul Holland, Michael Meredith, Jennifer Pike

Corresponding author: Heather Regan

Corresponding author e-mail: hegan41@bas.ac.uk

The Antarctic Peninsula is warming more rapidly than any other location in the Southern Hemisphere, with air temperatures increasing by nearly 3°C since 1950. The consequences of this, such as ocean warming and ice loss, are still not fully understood. Fresh water plays a key role in the dynamics of the Bellingshausen Sea, west of the Antarctic Peninsula, but the various components of the freshwater balance &ndash: glacial ice, sea ice, and precipitation &ndash: are affected by warming in different ways. This problem is compounded by a sparsity of data. To this end, a high-resolution model of the region has been developed, using the MITgcm to represent ocean, sea ice and ice shelves. Passive tracers track the advection, mixing and ultimate fate of fresh water from different sources, and demonstrate the differing spatial and temporal scales of their impacts on ocean structure. They highlight the dominance of sea-ice melt on seasonal timescales, and identify an important site in the south of the region that sees a relative freshening by some sources but a net loss of fresh water overall. Results will be used to investigate the causes and consequences of warming and freshwater change west of the Antarctic Peninsula.


NUMO: a new ice-sheet/ocean interaction model for Greenland fjords using high-order discontinuous Galerkin methods

Michal Kopera, Wieslaw Maslowski, Francis Giraldo

Corresponding author: Michal Kopera

Corresponding author e-mail: makopera@ucsc.edu

One of the key outstanding challenges in modeling of climate change and sea-level rise is the ice-sheet/ocean interaction in narrow, elongated and geometrically complicated fjords around Greenland. To address this challenge we propose a new approach, a separate fjord model using discontinuous Galerkin (DG) methods NUMO (Non-hydrostatic Unified Model of the Ocean). The goal of this project is to build a separate, high-resolution module for use in Earth System Models to realistically represent the fjord bathymetry, coastlines, exchanges with the outside ocean, circulation and fine-scale processes occurring within the fjord and interactions at the ice shelf interface. NUMO is currently at the first stage of development. The DG method provides NUMO with high-order accuracy as well as geometrical flexibility, including the capacity to handle non-conforming adaptive mesh refinement to resolve the processes occurring near the ice-sheet/ocean interface without introducing prohibitive computational costs. Another benefit of this method is its excellent performance on multi- and many-core architectures, which allows modern high performance computing systems to be utilized for high-resolution simulations. The non-hydrostatic model of the incompressible Navier–Stokes equations will account for the stationary ice shelf with sub-shelf ocean interaction, basal melting and subglacial meltwater influx and with boundary conditions at the surface to account for floating sea ice. The boundary conditions will be provided to FDG via a flux coupler to emulate the integration with an ESM. We will present first verification results on simplified test cases. Upon successful completion of the development phase NUMO will be tested for the Sermilik Fjord settings, using real bathymetry, boundary and initial conditions, and evaluated against available observations and other model results for this fjord. The overarching goal of the project is to be able to resolve the ice-sheet/ocean interactions around the entire coast of Greenland and two-way coupling with regional and global climate models such as the Regional Arctic System Model (RASM), Community Earth System Model (CESM) or Advanced Climate Model for Energy (ACME).