Radio-echo sounding of active subglacial lakes in Institute Ice Stream, West Antarctica

Martin Siegert, Neil Ross

Corresponding author: Martin Siegert

Corresponding author e-mail: m.siegert@imperial.ac.uk

Institute Ice Stream (IIS) is known to be at a physical threshold of major change, as its grounding line is perched at the top of a reverse sloping bed, similar in gradient to that beneath Thwaites Glacier. Moreover, numerical ice-sheet modelling has shown it to likely be more sensitive to external forcing than any other ice stream in Antarctica at present. Knowledge of ice-flow processes is critical to IIS stability, as accelerated grounding line retreat could be triggered by ice-flow reduction and the resulting net mass loss at the grounding line, and through thinning-induced flotation if flow is increased. Basal hydrology is clearly important to the flow of ice, as a series of ‘active’ subglacial lakes have been detected by ICESat laser altimetry through measurements of discrete ice surface elevation changes (both uplifts, where lakes are gaining water and depression as lakes lose water). Although radio-echo sounding (RES) is well suited to detecting subglacial lakes in slow-flowing regions of the ice-sheet interior, measurements of ‘active’ lakes have been more difficult to resolve. Indeed RES from many sites in East Antarctica where such lakes are thought to exist reveal no evidence for stored basal water. Here we provide evidence from RES of bed conditions over and around all of the ‘active’ subglacial lakes in IIS. We reveal that very little evidence of pooled basal water is apparent in the data. We discuss why this might be through investigating (1) the ability of RES to detect subglacial lakes under certain conditions, (2) the dynamic translation of basal hydrology to the ice surface and (3) the potential and likelihood of processes other than subglacial hydrology. We show that hydrological processes beneath IIS, and the role played by ‘active’ subglacial lakes, are far from understood well, despite being critical to sustaining the flow and stability of this highly sensitive ice stream.


Large-ensemble modeling of last deglacial and future variations of the Antarctic ice sheet

David Pollard, Robert DeConto, Won Chang, Patrick Applegate, Murali Haran

Corresponding author: David Pollard

Corresponding author e-mail: pollard@essc.psu.edu

Recent observations of thinning and retreat of Pine Island and Thwaites Glaciers identify the Amundsen Sea Embayment (ASE) sector of West Antarctica as particularly vulnerable to future climate change. To date, most future modeling of these glaciers has been calibrated using recent and modern observations. As an alternate approach, we apply a hybrid 3-D ice-sheet–shelf model to the last deglacial retreat of Antarctica, making use of geologic data from ~20 000 years BP to present, focusing on the ASE but including other sectors of Antarctica. Following several recent ice-sheet studies, we use large-ensemble statistical techniques, performing sets of ~600 runs with varying model parameters. The model is run for the last 40 ka, both on continental domains and on nested domains over West Antarctica. Various types of objective RMS scores for each run are calculated using reconstructed past grounding lines, relative sea-level records, measured uplift rates, cosmogenic elevation-age data, and modern ice distribution. Runs are extended into the future few millennia using simple warming scenarios. The goal is to produce calibrated probabilistic ranges of model parameter values and quantified envelopes of future ice retreat. Two types of results are described and compared, using (1) statistical techniques with emulation, likelihood functions and MCMC, and (2) a much simpler technique of taking ensemble-mean averages weighted by aggregate RMS scores. One robust conclusion is that for future warming scenarios, most reasonable parameter combinations produce retreat deep into the West Antarctic interior.


Circulation and hydrography in the Filchner Depression, Weddell Sea

Elin Darelius

Corresponding author: Elin Darelius

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

Cold and relatively dense ice-shelf water formed below the Filchner–Ronne Ice shelf flows northward along the eastern flank of the Filchner Depression to eventually descend the continental slope and become bottom water. Meanwhile, there is an adjacent seasonal inflow of warm water towards the ice-shelf cavity that is projected to increase drastically in a warmer future. We present results from four 1 year long time series from oceanographic moorings – three deployed across the eastern flank of the depression at 77° S and one at the ice-shelf front – and describe the observed mean hydrography and circulation and the seasonal variability thereof. The highest temperature observed at the front at depths below the ice-shelf draft is –1.5°C in April 2013.


Investigating the flow dynamics at ice-shelf calving fronts

Martin Wearing, Richard Hindmarsh, Grae Worster

Corresponding author: Martin Wearing

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

Ice-shelf calving rates and the buttressing ice shelves provide to grounded ice are both difficult to model and quantify. An increased understanding of the mechanics of this process is imperative in determining the dynamics of marine ice sheets and consequently predicting their future extent, thickness and discharge. Alley and others (2008) proposed an empirically derived calving law, relating the calving rate to the strain rate at the calving front. However, Hindmarsh (2012) showed that a similar relationship could be deduced by considering the viscous flow of the ice shelf. We investigate the relationship between the ice-shelf flow field and the strain rate field in the area close to the calving front. Analysis is undertaken of ice surface velocity data for a range of Antarctic ice shelves (data from Rignot and others, 2011) and an inferred strain rate field produced from that data. These geophysical results are compared with a simple mathematical model for laterally confined ice-shelf flow. Correlations are calculated between the same variables as Alley and others but using a new and larger data compilation, which gives a greater degree of scatter. Good agreement is observed between the expected theoretical scaling and geophysical data for the flow of ice near the calving front in the case of laterally confined ice shelves. This lateral confinement ensures flow is aligned in the along-shelf direction and resistance to flow is provided by near-stationary ice in the grounded margins. In other cases, the velocity is greater than predicted, which we attribute on a case-by-case basis to marginal weakening or the presence of ice tongues. We develop statistical methodologies for identifying these outliers.


Sea-ice formation on continental shelves: a comparison of water mass signals 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, Joshua Jones, Timothy G. Haskell

Corresponding author: Inga J. Smith

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

Sea-ice formation on the continental shelves of Antarctica and Arctic Alaska is influenced by water masses containing fresh water from ice shelves and rivers, respectively. The effects of supercooled ice-shelf water on Antarctic sea-ice microstructure, through platelet ice formation, can be dramatic, 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, landfast sea-ice growth can be influenced by fresher water from rivers and residual summer melt, with much larger input fresh water possible. In this presentation, application of a method to reconstruct changes in water masses using oxygen isotope measurements from sea-ice cores is contrasted for the Antarctic and Arctic continental shelves. Direct measurements of sea-ice growth rates are used to validate the output of a sea-ice thermodynamic model driven with Modern Era Retrospective-analysis for Research and Applications (MERRA) reanalysis data, along with observations of snow depth and freeze-up dates. The output of that model is used along with sea-ice oxygen isotope measurements and fractionation equations to determine changes in sea-water isotope composition over the course of the ice growth period. It is shown that for Barrow, Alaska, in 2011/2012, 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.


Characterizing the attenuation and temperature structure of Thwaites Glacier, West Antarctica

Dustin Schroeder

Corresponding author: Dustin Schroeder

Corresponding author e-mail: dustin.m.schroeder@jpl.nasa.gov

The temperature structure of ice sheets and glaciers is a fundamental control on ice flow, rheology and stability. However, it is difficult to observationally constrain temperature structures at the catchment to ice-sheet scale. The englacial attenuation of radar sounding data is strongly dependent on the temperature structure of ice sheets. Therefore, echo strength profiles from airborne radar sounding observation do contain temperature information. However, direct interpretation of englacial attenuation rates from radar sounding profiles is often difficult or impossible due to the ambiguous contribution the geometric and material properties of the bed to echo strength variations. To overcome this challenge, this work presents techniques that treat radar sounding echo strength and ice thickness profiles as continuous signals, taking advantage of along-profile ice thickness, propagation distance, and echo strength variations to constrain the spatial pattern of attenuation. These techniques are then applied to an airborne radar sounding survey of the Thwaites Glacier catchment in West Antarctica and the resulting patterns of englacial attenuation and temperature are interpreted in the context of local ice-sheet velocities, force balance and bed conditions.


Parameterization of basal hydrology near grounding lines: parameter sensitivity and transient results in one- and three-dimensional ice-sheet models

Gunter Leguy, Xylar Asay-Davis, William Lipscomb

Corresponding author: Gunter Leguy

Corresponding author e-mail: gunter@nmt.edu

Ice sheets and ice shelves are linked by the transition zone, the region where the grounded ice lifts off the bedrock and begins to float. Adequate resolution of the transition zone is necessary for numerically accurate ice-sheet–ice-shelf simulations. In previous work we have shown that by using a simple parameterization of the basal hydrology, the knowledge of basal lubrication across the grounding line improves the numerical accuracy of a one-dimensional vertically integrated fixed-grid model. We used a set of experiments based on the Marine Ice-Sheet Model Intercomparison Project (MISMIP) to show that reliable grounding-line dynamics at resolutions of ~1 km is achievable. In this presentation we show how our one-dimensional model and parameterization respond to basal variations when run to steady state. We further show transient results with time-varying basal stress and ice-shelf melting. Finally, we use the Community Ice Sheet Model (CISM) to demonstrates how our hydrology parameterization impacts three-dimensional models using the MISMIP-3D experiment. To this end we will compare three different numerical models: the shallow-shelf approximation (SSA), a depth-integrated higher-order approximation, and the Blatter–Pattyn model. The results from our one-dimensional model carry over to the 3-D models and a resolution of ~1 km remains sufficient to accurately predict grounding-line dynamics.


Disintegration mechanism of Mertz Ice Tongue revealed by sea-floor topography

Xianwei Wang, David Holland, Xiao Cheng, Peng Gong

Corresponding author: Xianwei Wang

Corresponding author e-mail: xw21@nyu.edu

Evolution of the Mertz Ice Tongue (MIT) was investigated using remote-sensing data from 1989 to 2013. A slight area increasing trend 36.9 km2 a–1 of the MIT was found after disintegration in February 2010 (35.3 km2 a–1 before that event). From boundary changes of the MIT, two interesting phenomena relating to evolution of rifts in the middle and front of the MIT are revealed. One is an opposite area-changing trends of both rifts in the middle of the MIT after the end of 2002, and the other is the changing direction of the ice front by ~42° eastward, or clockwise. The sea-floor topographic high (Mertz Bank) was verified to be the underlying cause of these phenomena. Comparison of ice bottom elevation and sea-floor topographic highs indicates a slightly grounded ice front started at least from the end of 2002. The high elevation of the sea-floor topography of the Mertz Bank forced the MIT to bend eastward, or clockwise, fed by upstream ice flowing towards the bank. Even without the collision of a large iceberg, the MIT would eventually disintegrate due to the presence of the topographic feature in its path. The disintegration cycle of the MIT is presumably ~70 years. Sea-floor-controlled disintegration of an ice tongue has implications for not only the Mertz break-up, but other such ice tongues in Antarctica.


Insights into ice-stream dynamics through modelling their response to tidal forcing

Sebastian Rosier, Hilmar Gudmundsson, Mattias Green

Corresponding author: Sebastian Rosier

Corresponding author e-mail: sebsie46@bas.ac.uk

The tidal forcing of ice streams at their ocean boundary can serve as a natural experiment to gain an insight into their dynamics and constrain the basal sliding law. Observations show that the flow of Rutford Ice Stream (RIS) is strongly modulated by the ocean tides, with the strongest tidal response at the 14.77 day tidal period (Msf). This is striking because this period is absent in the tidal forcing. A number of mechanisms have been proposed to account for this effect, yet previous modelling studies have struggled to match the observed large amplitude and decay length scale. We use a nonlinear 3-D viscoelastic full-Stokes model of ice-stream flow to investigate this open issue. As a first step we attempt to match observations qualitatively and find that a nonlinear sliding law can produce the Msf response. Matching the data qualitatively is more difficult and we find that we cannot match the large amplitudes of the long period modulation in flow without including a coupling between basal sliding and tidal subglacial water-pressure variations. Furthermore, the subglacial water system must be highly conductive and at low effective pressure, and the relationship between sliding velocity and effective pressure highly nonlinear in order for the model results to match GPS measurements. Coupled model results show the presence of tides result in a ~12% increase in mean surface velocity. Observations of tidally induced variations in flow of ice streams provide stronger constraints on basal sliding processes than provided by any other set of measurements.


Simulating marine ice sheets with the Community Ice Sheet Model

William Lipscomb, Gunter Leguy, Xylar Asay-Davis

Corresponding author: William Lipscomb

Corresponding author e-mail: lipscomb@lanl.gov

Version 2 of the open-source Community Ice Sheet Model (CISM, a successor of the Glimmer model) includes a higher-order velocity solver. CISM uses finite-element methods on a structured grid to solve several sets of flow equations, including the shallow-shelf approximation (SSA), a depth-integrated higher-order approximation, and the Blatter–Pattyn approximation. Here we show results for Marine Ice-Sheet Model Intercomparison Project (MISMIP and MISMIP3d) test problems. Using a newly implemented subgrid grounding line parameterization (GLP), CISM simulations at moderate grid resolution (~1 km) give excellent agreement with published benchmarks. For the original MISMIP experiment on a downward-sloping bed, CISM’s SSA results are in good agreement with the theoretical boundary-layer solution, and higher-order results are similar to those from a full-Stokes model. For MISMIP3d experiments with basal-sliding perturbations, CISM successfully simulates reversible migration of curved grounding lines. Without a GLP, much higher resolution is needed for comparable accuracy. These results suggest that accurate simulation of whole marine ice sheets at ~1 km resolution may be feasible. We discuss the work remaining to achieve this goal.


Modelled glacier response to centennial temperature and precipitation trends on the Antarctic Peninsula

Bethan Davies, Nicholas Golledge, Neil Glasser, Jonathan Carrivick, Stefan Ligtenberg, Nicholas Barrand, Michael Van den Broeke, Michael Hambrey, John Smellie

Corresponding author: Bethan Davies

Corresponding author e-mail: bethan.davies@rhul.ac.uk

The northern Antarctic Peninsula is currently undergoing rapid atmospheric warming. Increased glacier surface melt during the 20th century has contributed to ice-shelf collapse and the widespread acceleration, thinning and recession of glaciers. Therefore, glaciers peripheral to the Antarctic ice sheet currently make a large contribution to eustatic sea-level rise, but future melting may be offset by increased precipitation. Here we assess glacier–climate relationships both during the past and into the future, using ice-core and geological data and glacier and climate numerical model simulations. Focusing on Glacier IJR45 on James Ross Island, northeast Antarctic Peninsula, our modelling experiments show that this representative glacier is most sensitive to temperature change, not precipitation change. We determine that its most recent expansion occurred during the late Holocene ‘Little Ice Age’ and not during the warmer mid-Holocene, as previously proposed. Simulations using a range of future Intergovernmental Panel on Climate Change climate scenarios indicate that future increases in precipitation are unlikely to offset atmospheric-warming-induced melt of peripheral Antarctic Peninsula glaciers.


Design and sample results from the ISOMIP+ (ocean-only) and MISOMIP (coupled ice-sheet–ocean) intercomparison projects

Xylar Asay-Davis, Stephen Cornford, Daniel Martin, David Holland, Denise Holland

Corresponding author: Xylar Asay-Davis

Corresponding author e-mail: xylar.asay-davis@pik-potsdam.de

The MISMIP and MISMIP3D marine ice-sheet model intercomparison exercises have become popular benchmarks, and several modeling groups have used them to show how their models compare to both analytical results and other models. Similarly, the ISOMIP (Ice Shelf–Ocean Model Intercomparison Project) experiments have acted as a proving ground for ocean models with sub-ice-shelf cavities. As coupled ice-sheet–ocean models become available, an updated set of benchmark experiments is needed. A third ice-sheet-only intercomparison, MISMIP+, is already under development. We propose a sequel ocean-only exercise, ISOMIP+, with an end goal of coupling the two in a third intercomparison exercise, MISOMIP (the Marine Ice Sheet–Ocean Model Intercomparison Project). The four ISOMIP+ experiments have been designed to make use of bedrock topography and ice-shelf geometries from MISMIP+ produced by the BISICLES ice-sheet model in a shallow-shelf approximation (SSA) configuration. The first two experiments use static ice-shelf geometries to explore the effects of changes in far-field forcing on melt rates. In one experiment, the initial state is cold but the far-field forcing is relatively warm, wherease the initial state is cold and the forcing warm for the second experiment. The third and fourth experiments prescribes 100 years each of dynamic ice-shelf geometry, one with ice retreat and the other with ice advance. The MISOMIP experiment couples MISMIP+ and ISOMIP+. Changes in far-field ocean forcing lead to a rapid (over ~5 years) increase in sub-ice-shelf melting, which drives 100 years of ice-shelf retreat. Then, the far-field forcing switches abruptly to a cold state, leading to a rapid decrease in melting and a subsequent advance over the next 100 years. To illustrate, we present results from both intercomparison exercises using the Parallel Ocean Program 2 extended (POP2x) and the POPSICLES (POP2x-BISICLES) coupled model.


Evidence for increase basal ice-shelf melting in the Weddell Sea from oceanic noble-gas observations, 1990–2013

Oliver Huhn, Monika Rhein, Michael Schröder

Corresponding author: Oliver Huhn

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

We use oceanic noble-gas observations from the Weddell Sea from the period 1990 to 2013 to infer basal ice-shelf melting and the spatial distribution and temporal variability of the meltwater input into the ocean. This helium and neon data were used to compute glacial meltwater fractions and their contribution to Antarctic Bottom Water formation. Oceanic measurement of low-solubility and stable noble-gases helium and neon provide a useful tool to quantify and trace basal glacial meltwater. Atmospheric air with a constant composition of these noble gases is trapped in the ice matrix during formation of the meteoric ice. Due to the enhanced hydrostatic pressure at the base of the floating ice, these gases are completely dissolved, when the ice is melting from below. This leads to a substantial excess of helium and neon in pure glacial meltwater. From that 23 year long time series of noble-gas observations in the Weddell Sea we find an increasing trend in helium, neon and, hence, in the glacial meltwater content in Weddell Sea Deep and Bottom Water. Meltwater fractions along a repeated section in the northwestern Weddell Sea are almost doubling from 1990 to 2013, indicating substantial increase of glacial melting in the Weddell Sea.


Discriminating between steady-state and transient controls on englacial structures

Nicholas Holschuh, Byron Parizek, Sridhar Anandakrishnan, Richard Alley

Corresponding author: Nicholas Holschuh

Corresponding author e-mail: ndh147@psu.edu

Direct sampling of the subglacial environment is costly, and will therefore never supply the spatial coverage required for large-scale ice-sheet modeling. Studies of the West Antarctic ice sheet (WAIS) show that the frictional and rheologic properties of the bed ultimately control the evolution of the system, so developing geophysical methods to help us better constrain the basal characteristics of the WAIS will reduce uncertainty in predictions of the timing and magnitude of future sea-level rise. Radar-imaged structures within the ice are an attractive dataset for this pursuit, as they must reflect the flow dynamics that transform the horizontally deposited layers to their modern configuration; however, they can be challenging to interpret, given their non-unique nature. Research has shown that for some cases, the internal structures are spatially locked, and reflect the steady-state flow field of the ice; however, at other locations, complex internal structures were likely advected into place, and represent more transient phenomena. Using the ice-flow model of Parizek and others (2013), we attempt to determine the reasonable range of steady-state structures within the ice sheet to prevent the interpretation of transient features as a product of temporally stable bed characteristics. We also investigate whether it is possible to discriminate between structures generated by a plastic, high-friction bed as opposed to a linear-viscous, low-friction bed, to aid in future inversions of basal properties. Finally, we work toward an automated method of interpreting internal structures without the need for sophisticated layer tracing algorithms.


Imaging basal crevasses at the grounding line of Whillans Ice Stream, West Antarctica

Robert Jacobel, Knut Christianson, Adam Wood

Corresponding author: Robert Jacobel

Corresponding author e-mail: jacobel@stolaf.edu

In a recent Annals of Glaciology paper we presented a preliminary interpretation of basal crevasses in the grounding zone of Whillans Ice Stream as imaged by ground-based ice-penetrating radar. There we described elongated basal crevasses that typically extend for more than a kilometer and are imaged primarily in along-flow profiles. Often the phase of the crevasse echo is reversed relative to the bed echo, and arrives later, appearing as a return from ‘below’ the basal interface. We argued that this location was only apparent; the echoes result from off-nadir returns enhanced by the antenna radiation pattern of resistively loaded dipoles. The additional phase reversal is due to a second reflection from the ice–seawater interface. We also found a second class of basal crevasse that consists of peculiar ‘mirror-like’ echoes that appear both above and below the basal reflection and are highly symmetric. One cluster of these crevasses is associated with a portion of the grounding line with the dip of the crevasse(s) oriented downflow. We have since investigated additional occurrences of this type of crevasse echo and posit that they too result from the radar wave reflecting from a bright basal interface in such a way that the crevasse is illuminated from two directions: directly from above, and as a result of the reflection off the basal interface. In the latter case, energy is returned to the surface as a later arrival after a second bottom reflection, mirroring the geometry of the direct return and having the same phase. Here we investigate the details of this reflection mechanism. We also explore what the distribution of these crevasses can reveal regarding grounding-line location and basal stress distribution within the grounding zone.


Measuring changes in the vicinity of the Seal Nunataks ice shelf remnant from imagery and altimetry

Christopher Shuman, Etienne Berthier, Ted Scambos

Corresponding author: Christopher Shuman

Corresponding author e-mail: cshuman@umbc.edu

Acquisition and analysis of a combination of repeated imagery and ICESat laser altimetry has enabled the ongoing losses from the northern Larsen B ice shelf remnant to be assessed. The remnant, the Seal Nunataks ice shelf (SNIS), has four ICESat tracks that cross it, as well as adjacent tracks that cross Robertson Island (RI) and the margin of the Antarctic Peninsula. The available altimetry data from ICESat (2003–2009) show that elevation losses increase from west to east across the SNIS. Ice elevation losses suggest mean ice-shelf thinning rates of up to 1.6 m a–1 and reveal processes impacting the remaining shelf ice. Asymmetric elevation changes across RI suggest the magnitude of regional climate impacts although its sloping terrain shows distinct local variations in ice loss depending on slope aspect. Imagery analysis using Landsat 7 and ASTER images acquired during 2001–2013 shows that ice area losses continued on the shelf remnant following the Larsen A break-up in early 1995 as well as the Larsen B break-up in early 2002. The largest losses (~350 km2) occurred on the north side of the remnant in late 2004 into 2005, with smaller losses over the years along the remaining margins. Despite a slight regional cooling in recent years and more persistent sea ice in the area, the SNIS is still calving and appears to be retreating past its pinning nunataks. In contrast to SNIS, RI has experienced minor ice area losses that suggest most of its ice is grounded and thus less directly impacted by ocean interactions. The combination of these remote-sensing datasets provides additional insights about ongoing ice loss processes.


Passive underwater acoustics gives insight into glacier–ocean interactions

Oskar Glowacki, Jaroslaw Tegowski, Grant B. Deane, Mateusz Moskalik, Philippe Blondel

Corresponding author: Oskar Glowacki

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

Dynamic phenomena taking place at the ice–ocean boundaries are one of the main processes that clearly reflect progressive climate shifts. Melting tidewater glaciers lose mass as a result of calving and submarine melting, thereby providing large amounts of fresh water to glacial bays and fjords. In this way, they cause sea-level rise and modify local water circulation patterns. Therefore, these issues require considerable attention and there is also a growing need for new state-of-the-art measurement methods and tools. Recently, application of passive underwater acoustics made it possible to gain insight into complex glacier–ocean interactions. Ice detachments from grounded marine-terminating glaciers and submarine melting were investigated during several field experiments carried out in Hornsund fjord, Spitsbergen, in 2013 and 2014. Three different types of calving event were unambiguously detected and classified, both acoustically and photographically: typical subaerial, sliding subaerial and submarine. The energy of the block-water impact was found to be strongly correlated with acoustic emission at 200 Hz for analyzed subaerial events. Moreover, measurements of ambient noise directionality showed that low-frequency sounds are relatively stable and originate mostly from the glacier terminus. The directional characteristic of high-frequency signals, in turn, is strongly dependent on the distribution of floating chunks of glacial ice around the fjord. These findings illustrate possible applications and effectiveness of passive underwater acoustic methods in the monitoring of melting glaciers. This work has been supported by the Polish National Science Center, grants Nos. 2011/03/B/ST10/04275 and 2013/11/N/ST10/01729, Office of Naval Research, Ocean Acoustics Division, grant No. N00014-1410213, and the statutory activity of the Institute of Geophysics Polish Academy of Sciences.


Pan-Arctic controls on the rapid retreat of marine-terminating Arctic outlet glaciers

Rachel Carr, Chris Stokes, Andreas Vieli

Corresponding author: Rachel Carr

Corresponding author e-mail: rachel.carr@newcastle.ac.uk

Arctic ice masses have lost mass rapidly during the past two decades and accelerated discharge through marine-terminating outlet glaciers has been a primary component of these deficits. However, previous studies have focused on individual glaciers or regions, meaning that little is known about broadscale trends in outlet glacier retreat rates, their spatial variation across the Arctic and/or their differing response to forcing. Here we present the first pan-Atlantic-Arctic assessment of outlet glacier retreat rates and their relationship to forcing factors during the past two decades (1992–2010). Outlet glacier retreat was almost ubiquitous and increased fivefold between 1992–2000 (30.5 m a–1) and 2000–2010 (105.8 m a–1). During the latter period, 97% of all study glaciers (264 glaciers) retreated. In contrast, only 74% underwent net retreat between 1992 and 2000, 18% advanced and 8% showed no discernible change. For the period 2000–2010, recession was greatest in northern Greenland (>600 m a–1), followed by central-west (168.0 m a–1), southeast (135.9 m a–1) and northwest (116.6 m a–1) Greenland. Retreat rates also increased dramatically on the Barents Sea coast of Novaya Zemlya and exceeded those in east Greenland between 2000 and 2010. Oceanic warming appears to be a widespread control on glacier retreat and the influence of sea ice is significant in several regions. Despite regional trends, we document large variations in retreat rates between individual glaciers, with the areas exhibiting the highest retreat rates also showing the greatest variability. We show a widespread statistical relationship between fjord width variability and outlet glacier retreat rate, particularly in areas where glaciers are constrained by mountainous topography.


Multi-decadal retreat of Novaya Zemlya outlet glaciers in response to climatic forcing

Rachel Carr, Heather Bell

Corresponding author: Rachel Carr

Corresponding author e-mail: rachel.carr@newcastle.ac.uk

Arctic ice masses have rapidly lost ice from the mid-1990s, through a combination of negative surface mass balance and accelerated ice discharge from marine-terminating outlet glaciers. In the past decade, substantial mass deficits have been identified on Novaya Zemlya (NVZ), Russian High Arctic, and its outlet glaciers have retreated dramatically, likely due to declining sea-ice concentrations. However, little is known about longer-term glacier behaviour on NVZ, and its potential impact on overall mass balance. Here we greatly extend the available record of retreat and assess multi-decadal glacier response to forcing between 1976 and 2014 using remotely sensed data. Following at least 25 years of gradual recession, retreat rates accelerated substantially from circa 2000 and again from 2011 onwards. The rate and temporal pattern of retreat were strongly dependant on terminus type: marine-terminating outlets receded an order of magnitude faster than land- or lagoon-terminating glaciers and land-based termini showed limited change in retreat rate over time. Furthermore, retreat was markedly higher on the Barents Sea coast than the Kara Sea. Comparison with forcing data shows that accelerated retreat from 2011 onwards coincided with exceptionally low sea-ice concentration and duration between 2011 and 2013. This suggests that sea ice is an important controlling factor, which agrees with shorter-term studies on NVZ. Although air temperature information is more limited on NVZ, data from both meteorological stations and reanalysis highlight 2011 and 2012 as two of the warmest years during the period 1950–2014. The limited response of land-terminating outlets suggests that air temperatures do not cause retreat directly, via melting or enhanced basal lubrication, but may have an indirect influence through melting of sea ice or hydrofracture.


Subglacial bathymetry and sediment distribution beneath the Pine Island Glacier ice shelf modeled using aerogravity and in situ geophysical data

Atsuhiro Muto, Leo Peters, Karsten Gohl, Ingo Sasgen, Richard Alley, Sridhar Anandakrishnan, Kiya Riverman

Corresponding author: Atsuhiro Muto

Corresponding author e-mail: aum34@psu.edu

The Amundsen Sea sector of the West Antarctic ice sheet is losing mass at a rate that has more than doubled in the past four decades, and continues to increase. Pine Island Glacier (PIG), the second largest drainage basins in this sector, experienced the fastest grounding-line retreat and its ice-mass loss increased more rapidly than any others in the last two decades. The large mass imbalance of PIG is attributed to the increased sub-ice-shelf melting by the incursion of relatively warm Circumpolar Deep Water (CDW) beneath the PIG ice shelf (PIGIS), although the lack of precise bathymetry data has restricted thorough understanding of the ice–ocean interactions. Here we present updated bathymetry and sediment distribution beneath PIGIS, modeled by inversion of aerogravity data with constraints from active-source seismic and autonomous underwater vehicle data, and the regional gravity anomaly derived from satellite gravity observations. Modeled bathymetry shows that the submarine ridge beneath the middle of PIGIS appears to continue across the width of the ice shelf, with no major deep troughs crossing it, consistent with previously predicted sub-ice-shelf ocean circulation. However, the relatively low resolution of the aerogravity data and limitations in our inversion method leave a slight possibility that there is an undetected, few-kilometer-scale narrow trough that may alter this predicted sub-ice-shelf ocean circulation. Modeled sediment distribution indicates that the submarine ridge marks the transition from a thick sedimentary basin (soft, smooth bed for ice flow) around the current grounding zone of the main PIG trunk to a region of thin-to-no sediment with some exposed crystalline basement (rough, resistant bed for ice flow) that extends seaward into Pine Island Bay. We hypothesize that this transition in basal conditions caused the post-Last Glacial Maximum retreat of PIG to stabilize near this geological boundary.


Highly concentrated melting and channel formation at the grounding line of the southern Ross Ice Shelf

Oliver Marsh, Helen Fricker, Matt Siegfried, Keith Nicholls, Hugh Corr, Ginny Catania

Corresponding author: Oliver Marsh

Corresponding author e-mail: oliver.marsh@canterbury.ac.nz

Ice shelf channels are observed in radar and satellite imagery around Antarctica and their presence has been shown to affect surface velocity patterns, the internal strength of ice shelves and the background rate of oceanic basal melting. Using a combination of GPS, satellite data and phase sensitive radar we conduct a detailed analysis of channel formation at the southern end of the Siple Coast, around the Whillans Ice Stream/Mercer Ice Stream suture zone. ICESat data show consistent surface thinning and channel expansion over the period 2003–2009. In addition, we observe basal melt rates greater than 20 m a–1 in December–January 2014–15 at a number of phase-sensitve radar sites, decreasing rapidly in ice shelf flow direction. This melting is concentrated where the channel initiates at a site immediately downstream of the grounding line of Mercer Ice Stream. Localized, intense melting in this region produces a deep basal incision in the ice shelf and forms a channel with a clearly visible surface impression. MODIS imagery from 2004, 2009 and 2014 shows that the surface depression advects downstream with ice flow and, while it is created at the upper end, it does not disappear at the lower end. The highly focussed nature of the melting contrasted against very low background values (<1 m a–1) and low horizontal strain rates suggest the channel is formed by a buoyancy driven meltwater plume directly linked to the active subglacial drainage system upstream of the grounding line. Similar discrete surface features 30–40 km downstream of the grounding line to the north indicate a possible major switch in subglacial drainage pathways between 85 and 120 years ago. The intensity of the observed basal melting in the channel in 2014–15 and delayed ice shelf surface response may also provide further evidence for episodic drainage of subglacial lakes upstream.


Marine ice-sheet profiles and stability under Coulomb basal conditions

Victor Tsai, Andrew Stewart, Andrew Thompson

Corresponding author: Victor Tsai

Corresponding author e-mail: tsai@caltech.edu

The behavior of marine-terminating ice sheets, such as the West Antarctic ice sheet, is of interest due to the possibility of rapid grounding-line retreat and consequent catastrophic loss of ice. Critical to modeling this behavior is a choice of basal rheology, where the most popular approach is to relate the ice-sheet velocity to a power-law function of basal stress. Recent experiments, however, suggest that near-grounding-line tills exhibit Coulomb friction behavior. Here we address how Coulomb conditions modify ice-sheet profiles and stability criteria. The basal rheology necessarily transitions to Coulomb friction near the grounding line, due to low effective stresses, leading to changes in ice-sheet properties within a narrow boundary layer. Ice-sheet profiles ‘taper off’ towards a flatter upper surface, compared with the power-law case, and basal stresses vanish at the grounding line, consistent with observations. In the Coulomb case, the grounding-line ice flux also depends more strongly on flotation ice thickness, which implies that ice sheets are more sensitive to climate perturbations. Furthermore, with Coulomb friction, the ice sheet grounds stably in shallower water than with a power-law rheology. This implies that smaller perturbations are required to push the grounding line into regions of negative bed slope, where it would become unstable. These results have important implications for ice-sheet stability in a warming climate.


Fjord geometry and marine outlet glacier stability on centennial timescales

Henning Åkesson, Kerim H. Nisancioglu, Faezeh M. Nick

Corresponding author: Henning Åkesson

Corresponding author e-mail: henning.akesson@uib.no

Many of Greenland’s marine-terminating glaciers have thinned, accelerated and retreated during the last decade, broadly consistent with warmer atmospheric and oceanic conditions. However, these patterns involve considerable spatial and temporal variability, with a large diversity in glacier behaviour within the same regions. Similarly, reconstructions of marine-terminating glaciers indicate highly asynchronous retreat histories. This highlights the danger in extrapolating glacier trends in space and time, and point towards topographic controls of marine outlet glacier behaviour. Here we test the hypothesis that marine outlet glacier stability is largely controlled by fjord geometry, and to a lesser extent by regional climate and ocean forcing. We employ a dynamic flowline model on idealized glacier geometries (representative of different outlet glaciers) to investigate geometric controls on decadal to millennial timescales. The model accounts for driving and resistive stresses of glacier flow as well as along-flow stress transfer. It also has a physical treatment of iceberg calving and a time-adaptive grid allowing for continuous tracking of grounding-line migration. We assess the relative importance of basal and lateral pinning points, and whether shallow and wide fjords are more likely to host rapid retreat than deep and narrow ones. The relevance of these results to past and future marine-terminating glacier stability is also discussed.


Parameterization of melting along the calving face of Jacobshavn Glacier, Greenland

Pierre Mathiot, Adrian Jenkins, David Holland

Corresponding author: Pierre Mathiot

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

Today ocean models at climate resolution (O(1–10 km)) are not able to compute the melt rate driven by subglacial discharge along the calving face of a tidewater glacier such as Jacobshavn Glacier. The reasons are simple: the processes along the calving face need a very high horizontal resolution O(1–10 m) to be explicitly simulated, and subglacial discharge is neglected in most ocean models. However, this deep input of fresh water has two major consequences. The first one is that the waters confined within the overdeepened basin of a fjord can be renewed in summer by a vigorous overturning circulation driven by the subglacial discharge and the associated, enhanced terminus melting. The second one is that the elevated melt rates may contribute to glacier recession in Greenland, which is focussed on its marine outlet glaciers. In order to incorporate these processes in a relatively coarse-resolution ocean model (~5 km), we implement a parameterization based on the theory of buoyant plumes in unstratified environments. Although Greenlandic fjords are far from unstratified, their structure is more accurately approximated using two or three well-mixed layers separated by sharp interfaces than by a constant linear stratification. In an unstratified environment there are two length scales that govern plume behaviour: one based on the ratio of the initial input of buoyancy to the buoyancy input by melting; the other based on the change in the freezing point with depth. The first is appropriate for times of high subglacial discharge, while the latter applies when the discharge is low or non-existant. Suitable nondimensionalization of the plume equations using these length scales allows the melt rate to be approximated reasonably well by simple polynomial functions. The melt rates computed by these parameterizations and the subglacial runoff are included in an ocean model using techniques recently developed to include the melting of floating ice shelves around Antarctica. As a proof of concept, this method is implemented in an idealized fjord configuration at coupled climate model resolution (~5 km) and compared with observations and results from high-resolution ocean models (~10 m at the calving face).


Strong effects of thermodynamic interactions of the Antarctic ice shelves with the ocean circulation on the Southern Ocean and sea-ice formation in a global coupled ocean circulation model

Olga Sergienko, Mathew Harrison

Corresponding author: Olga Sergienko

Corresponding author e-mail: osergien@princeton.edu

Melting/refreezing of ice shelves have strong impacts both on ice shelves (through modification of their shape) and on the ocean circulation (through modification of their water masses). Representation of ice-shelf/ocean interaction in the global ocean circulation models continues to be challenging. Using a high-resolution (1/8 deg) global isopycnal ocean model, MOM6, and a sea-ice model, SIS, we investigate the effects of thermodynamic coupling of the Antarctic ice shelves on the various aspects of ocean circulation. Such high (3–8 km) horizontal spatial resolution allows for detailed resolution of the sub-ice-shelf cavity circulations. The computed ice-shelf melt rates are in very good agreement with observationally derived melt rate estimates. The spatial distributions of simulated melting/freezing rates indicate enhanced melting in the vicinity of the grounding line and very strong melting at the ice-shelf front. Results of our simulations show strong effects of sub-ice-shelf meltwater on circulation of the Southern Ocean. We also find that simulations accounting for the thermodynamic coupling of the Antarctic ice shelves produce consistently thicker sea ice compared with the uncoupled simulations.


A simple parameterization of ice-shelf basal melting for long-term, large-scale simulations of the Antarctic ice sheet

Ralf Greve, Ben Galton-Fenzi, Rupert Gladstone

Corresponding author: Ralf Greve

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

Ice shelf basal melting, and its change with time, plays a crucial role for the past and future evolution of the Antarctic ice sheet and its contribution to global sea level. Since integrating coupled ocean and ice sheet models for long time periods is hard to achieve within a reasonable amount of time using present-day computers, long-term simulations of the Antarctic ice sheet require that ice-shelf basal melting is parameterized. Here we describe a simple, physically based parameterization that calculates the basal melting of ice shelves as a function of both the depth of ice below mean sea level and far-field ocean temperatures. The parameterization is tuned differently for eight Antarctic sectors in order to achieve reasonable agreement with the modern spatial distribution of ice-shelf basal melting. In addition, we consider a dependence on water column thickness in order to reduce melting to zero very close to the grounding line. The parameterization is implemented in the Antarctica module of the dynamic/thermodynamic, large-scale ice-sheet model SICOPOLIS (www.sicopolis.net), and we discuss several future climate runs over the next centuries in order to explore sensitivity to various forcings.


Iceberg interactions in a coupled GCM

Alon Stern, Olga Sergienko, Alistair Adcroft, Robert Hallberg

Corresponding author: Alon Stern

Corresponding author e-mail: sternalon@gmail.com

Iceberg calving accounts for ~50% of the ice mass loss from the Greenland and Antarctic ice sheets. Once calved, icebergs drift into the open ocean where they melt, redistributing fresh water to the ocean and altering water mass properties. This study focuses on the representation of icebergs in a coupled global climate model (CGCM). This representation is developed by allowing icebergs to occupy physical space, and to interact with sea ice, topography and other icebergs. Model results from simulations using interacting icebergs are compared with simulations where icebergs are represented as non-interacting point particles. This comparison shows that including the effects of iceberg interactions in a CGCM has a large effect on the the spatial location of the fresh-water fluxes from icebergs into the ocean. A representation of large tabular icebergs is also developed. Tabular icebergs are represented as a collection of smaller circular icebergs ‘bonded’ together by stiff springs. This representation allows for tabular icebergs to break/calve in the open ocean, which is an essential part of the iceberg decay process.


Changes in ice dynamics of tributary glaciers of former Larsen A and Prince Gustav Channel Ice Shelf (Antarctic Peninsula)

Thorsten Seehaus, Matthias Braun, Sebastián Marinsek, Pedro Skvarca, Veit Helm

Corresponding author: Thorsten Seehaus

Corresponding author e-mail: thorsten.seehaus@fau.de

Within the last decades various ice shelves in the Antarctic Peninsula disintegrated or retreated considerably. Tributary glaciers started to accelerate due to the missing buttressing forces and a significant ice mass loss was reported for the Antarctic Peninsula. Estimates of the mass changes converge for different studies, but they show still considerable error bars. In this study we provide a detailed analysis of former tributary glaciers of the Larsen A and Prince Gustav ice shelves. The goal is to investigate their response to the ice shelf disintegration, which started at the end of the 1980s, and to better quantify temporal variations in the ice mass loss. Surface velocity changes are investigated by applying intensity offset tracking on time series of different SAR satellites (ERS-1/2, Envisat, RADARSAT-1, ALOS PALSAR, TerraSAR-X) of the last 23 years. Ice front position changes are mapped in conjunction with the velocity data. Surface elevation lowering is determined from high-resolution bi-static TanDEM-X satellite data, ASTER and SPOT stereo images. Mapping of former trimline altitude extents the elevation change analysis to pre-collapse conditions. Airborne laser scanning, ground-penetrating radar (NASA Operation IceBridge) and differential GNS data from field campaigns support the mass loss analysis. At the Dinsmoor–Bombardier–Edgeworth glacier system, former tributary to the Larsen A ice shelf, the surface velocity increased from 0.65 m d–1 up to 3.9 m d–1 between 1993 and 2000 at the glacier front. Subsequently, surface velocities decreased to 1.5 m d–1 in 2014. The glacier system lost an area of 58 km2 of previously grounded ice. An average surface lowering of 140 m since 1995 was measured on the lower parts of the glacier system, with a significantly decreased surface lowering rate during the last years. The former Prince Gustav Channel tributaries Sjögren and Boydell Glacier showed a similar behavior during the last decades. The total ice mass loss is estimated by combining all datasets. Our results will support the imbalance calculation in the research area and the interpretation of how ice shelf disintegration affects the tributary glaciers.


A virtual laboratory for understanding shear margin dynamics

Jenny Suckale, Cooper Elsworth, Thibaut Perol, John Platt, Jim Rice

Corresponding author: Jenny Suckale

Corresponding author e-mail: jsuckale@stanford.edu

Ice stream dynamics remains an important unknown in ice-sheet models. The Siple Coast ice streams exhibit spatial variability on the timescale of decades to hundreds of years, suggesting that they may respond dynamically to future climatic changes. The mechanisms controlling the positions of the lateral ice-stream margins, where ice transitions from fast to slow, remain poorly understood. Three possible mechanisms include the transition from thawed to frozen bed, the formation of subglacial drainage channels, or the existence of sticky spots. This study aims to evaluate these mechanisms with comparisons to ice-stream observables. The distinct behavior of ice streams depends upon the strong coupling between till rheology, ice rheology and shear margin position. For this reason, we propose a time-dependent 2-D thermomechanical model that resolves the margin location self-consistently through force balance. We assume a Mohr–Coulomb till, with basal slip occurring where the shear stress equals the local yield strength of the till. Yield strength is strongly dependent on pore pressure resulting in variations of basal stress across the width of the stream with the presence of subglacial water. The model can therefore provide constraints on basal strength distributions based on surface observations of margin position and stream velocity. Furthermore, melt production in the ice column is linked to pore pressure to determine possible feedback between glacial melt and basal sliding. Realistic ice rheologies are implemented to provide insight on rheological effects on model results. We present consistent basal strength distributions for velocity observations spanning the Siple Coast ice streams, with particular emphasis on margin migration. We discuss the implications of basal strength distributions on pore pressure variation and compare the different mechanisms thought to control margin stability. We determine that velocity distributions along the length of a stream require different basal strength profiles to match data, suggesting the development of different hydrological regimes. Time-dependent simulations provide insight into the evolution of subglacial conditions and effects on margin migration and stability. Our model was developed for the Siple Coast, but could provide a virtual laboratory for modeling shear margin behavior more generally.


Transport pathways of Circumpolar Deep Water on the Amundsen continental shelf

Karen Assmann, Anna Wåhlin, Ben Webber, Karen Heywood, SangHoon Lee

Corresponding author: Karen Assmann

Corresponding author e-mail: karen.assmann@gu.se

Oceanic heat transport has been identified as a critical control on the mass balance of the West Antarctic ice sheet. Its outlet glaciers in the Amundsen Sea show the highest mass loss in Antarctica and this has been traced to the thinning of the ice shelves. Circumpolar Deep Water (CDW) with temperatures several degrees above the freezing point is commonly observed under these ice shelves and linked to their high melt rates. Transport of CDW onto and on the shelf occurs through a series of troughs from the shelf break to deep basins in front of the ice shelves. It has generally been assumed that each of the troughs separately links the shelf break to the ice shelves. There is, however, evidence that the warm inflow into the Getz–Dotson trough already contains a meltwater signature that has likely been acquired further east on the shelf. This would constitute an additional transport pathway for glacial meltwater in addition to westward surface transport in the coastal current. It implies that northward transport of modified CDW at the western flank of the troughs does not actually leave the shelf, but follows the bathymetry of the shallower banks separating them and preconditions the CDW inflow in troughs located further west. We will use data from a shelf break mooring close to the eastern Getz–Dotson trough to examine this hypothesis. Comparing water masses here to observations across the shelf break and from the central trough will allow us to identify the origin of the water flowing onto the shelf at this location. Analysing the co-variability between the shelf break mooring and a mooring site located in the inflow mid-shelf will allow us to identify the fate of the shelf break flow and possible additional transport pathways located further south on the shelf.


Identifying meltwater pathways in the Amundsen Sea

Louise C. Biddle, Karen J. Heywood, Jan Kaiser, Adrian Jenkins, Brice Loose

Corresponding author: Louise C. Biddle

Corresponding author e-mail: louise.biddle@uea.ac.uk

The Amundsen Sea is the location of some of the fastest-melting ice shelves, linked to warm ocean waters causing basal melt. This increased meltwater production may be the root of the freshening in the Ross Sea over the past 30 years. Tracing the meltwater pathways is important for identifying the regions most affected by the increased input of this water type. We use water mass characteristics derived from 105 CTD casts during the iSTAR/Ocean2ice cruise between January and March 2014 to calculate meltwater fractions. Fractions immediately in front of Pine Island Glacier (PIG) are as high as 2.4%. Away from the ice front, the meltwater maximum follows the 27.7 kg m–3 potential density level sandwiched between Winter Water (WW) above and Circumpolar Deep Water (CDW) below. The ice-shelf meltwater maximum is coincident with increased turbidity close to the front of the ice shelf. Traditional methods using temperature, salinity and dissolved oxygen concentrations to calculate meltwater fractions become less reliable with distance from the ice-shelf front. Processes such as atmospheric interaction and biological activity affect these properties. We analysed the effect of these processes on the reliability of meltwater fraction results across the region using a bulk mixed-layer model based on the one-dimensional Price–Weller–Pinkel model (1986). The model includes sea ice, dissolved oxygen concentrations, air–sea gas exchange and a simple biological model. Independent estimates of meltwater fractions are derived from noble gas (mainly helium and tritium) concentration measurements from water samples collected during the same cruise. Oxygen isotope measurements of water are used to identify the contributions of glacial and sea ice melt to the meltwater budget.


Determining the basal properties and englacial temperature of the Greenland ice sheet from radio-echo sounding

Thomas Jordan, Jonathan Bamber, Christopher Williams, John Paden, Martin Siegert

Corresponding author: Thomas Jordan

Corresponding author e-mail: tom.jordan@bris.ac.uk

Mapping the basal properties of ice sheets (primarily the presence of basal melting and the surface roughness) provides valuable boundary conditions for, and tests of, 3-dimensional thermomechanical models of ice sheets. In addition, constraining the spatial variation of englacial temperature can provide a test of the validity of the steady-state solution of such models. Whilst direct measurements of the basal properties and englacial temperature can only be made at a limited number of borehole sights, radio-echo sounding provides the means to infer the spatial distribution of both throughout the extent of entire ice sheets. The radar inference of the basal melting is possible due to regions of subglacial water having a ~10–15 decibel greater power reflection coefficient than bedrock, and the radar inference of depth-averaged temperature is possible due to there being an exponential relationship with the radar attenuation rate. In this study we map regions of basal melting and the spatial distribution of depth-averaged temperature for ~10 years of radio-echo sounding data from the Greenland ice sheet. Determining the spatial distribution of the radar attenuation rate is a necessary precursor to our investigation, and we infer this from the relationship between basal returned power and ice thickness. This approach requires that the glacier bed has a statistically uniform reflection coefficient in a local region, and we select appropriate sites using the radar attenuation loss/basal reflection coefficients calculated from a thermomechanical ice-sheet model. The radar-inferred attenuation rate is firstly used to obtain radar-inferred power reflection coefficients (which in turn are used to identify areas of basal melting), and secondly used to obtain the horizontal distribution of depth-averaged temperature (via inverting the exponential relationship at each location). Our study aims to be the first to provide estimates of the spatial distribution of radar-inferred temperature for Greenland over the entire depth of the ice column, and to provide spatially extensive mapping of regions of basal melting in Greenland.


Bed properties beneath the tributaries of Pine Island Glacier from seismic investigations

Alex Brisbourne, Andrew Smith, Edward King, David Vaughan, Damon Davies, Robert Bingham

Corresponding author: Alex Brisbourne

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

Pine Island Glacier, in the Amundsen Sea sector of West Antarctica, is one of the most rapidly changing glaciers on the continent and currently one of the most significant contributors to sea-level rise. The ongoing changes are almost certainly driven by increased oceanic melting at the base of the floating ice shelf, resulting in reduced buttressing of the glacier. The glacier’s tributaries are however not responding to this forcing in a uniform manner and it is not fully understood what controls this variable response. iSTAR is a UK NERC funded programme studying the ocean, ice shelf, glacier and drainage basin system of Pine Island Glacier. One component of this multi-disciplinary project is aimed at determining the basal conditions beneath the glacier and the role these conditions play in controlling changes in the ice flow. Radar grids acquired on the first iSTAR traverse, which determine basal topography and character, were used to define the location of seismic lines, which can be used to discriminate bed materials. Consequently, during the second iSTAR traverse, a series of 7 km seismic lines were acquired on the main glacier trunk and a number of its tributaries. A range of features such as subglacial hills, lineations, flat planes and complex topography was targeted. These data have been analysed to determine the nature of the bed material, to help characterize the basal conditions, their variability, and their potential influence on the ice flow. Results from a number of these seismic lines will be presented.


Sensitivity of the recent increase in Antarctic sea ice in ocean models

Joakim Kjellsson, Paul Holland, Gareth Marshall, Pierre Mathiot, Yevgeny Aksenov, Andrew Coward, Sheldon Bacon, Alex Megann, Jeff Ridley

Corresponding author: Joakim Kjellsson

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

We study the recent increase in Antarctic sea ice using a coupled ocean–sea-ice model forced by atmospheric reanalysis. We investigate the impact on sea ice from both model parameters, e.g. vertical mixing and eddy parameterization, as well as external forcing, e.g. precipitation and meltwater from the Antarctic continent. We use the NEMO ocean model coupled to the CICE sea-ice model at 1 degree horizontal resolution forced with ERA-Interim reanalysis. The results will have impacts for our understanding of the Southern Ocean, its sea ice and their representation in future coupled climate model studies, e.g. CMIP6. Since the dawn of the satellite era there has been a slow increase in Antarctic sea ice with pronounced spatial structure. The reason for this increase is not yet fully understood and very few climate model simulations reproduce the observed mean state and/or increase. By varying model parameters and external forcing, we determine that obtaining a realistic sea-ice cover requires a complex balance of horizontal and vertical mixing as well as freshwater input. The surface freshwater balance impacts the vertical salinity gradient and thus vertical fluxes of heat and salt. Underestimation of precipitation or meltwater results in deep convection in the open ocean and the opening of large polynyas in the Weddell and Ross Seas. The presence of polynyas reduces the sea-ice extent. The depth of the mixed layer has a large impact on the sea-ice seasonal cycle. The summer mixed layer must be sufficiently deep to prevent SST from becoming too high but not so deep as to mix up heat and salt from below. In winter, a deep mixed layer lets brine rejected from sea ice mix down to depths below that of the summer mixed layer thus maintaining a necessary stratification.


Water properties at the Antarctic shelf edge in high-resolution climate models

Jeff Ridley

Corresponding author: Jeff Ridley

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

The next generation of climate models will start to incorporate ice-shelf cavities and interactive ice sheets. However, the processes of entraining circumpolar intermediate waters onto the continental shelf, where it may interact with the ice shelves, may not be well represented. This study investigates the sensitivity of the intrusion of warm waters onto the Antarctic shelf as a function of coupled climate model atmosphere and ocean resolution. The global HadGEM3 model will be investigated at 25 and 8 km ocean and 60 and 25 km atmospheric resolution. Increasing the ocean model resolution improves the representation of latent heat polynyas and the Southern Ocean overturning circulation. However, this study will focus on the shelf edge in the Amundsen and Weddell Seas, and will provide a baseline analysis of the temporal variability of shelf-edge water properties against which future model improvements may be assessed. It is expected that the development of ocean hybrid vertical coordinate systems and the representation of tides will improve the mixing across the continental shelf edge.


Investigating elevation change in the Bellingshausen sector of the West Antarctic ice sheet

Allen Pope, Benjamin Smith, Paul Holland, Heather Regan

Corresponding author: Allen Pope

Corresponding author e-mail: allen.pope@nsidc.org

The Amundsen–Bellingshausen sector of West Antarctica is one of the fastest-changing sections of the entire icy continent, contributing to both current and future sea-level rise. While ice loss in Pine Island and Thwaites catchments has been well-documented, changes in the Bellingshausen region (including the Abbott, Allison, Venable, Fox, Ferrigno, and Stange) has been less studied. In this study, we attempt to link observations of elevation change with drivers of elevation change of grounded ice in the Bellingshausen sector of Antarctica. Elevation changes are measured with ICESat and IceBridge data. Oceanic forcing (i.e. ocean heat content interacting with ice) is interrogated by integrating high-resolution regional ocean models. Atmospheric forcing (i.e. surface mass balance and/or precipitation minus sublimation) is investigated using regional and global reanalyses. Atmospheric reanalyses are also used to model the presence of possible non-surface retracking targets for radar altimeters. The elevation change signal for the Bellinghausen sector is complex and somewhat sparse, and there are many potential components contributing to elevation change in the region. This study attempts to bridge ice, ocean, and atmospheric effects to understand the elevation changes in an important sector of the West Antarctic ice sheet.


High-resolution modelling of ocean circulation and melt rates beneath the Filchner–Ronne Ice Shelf

Ruth Mugford, Paul Holland, Keith Nicholls

Corresponding author: Ruth Mugford

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

The Filchner–Ronne Ice Shelf plays an important role in the modification of the temperature and salinity of shelf water to produce dense bottom waters, which are significant for the global ocean circulation (Nicholls and others, 2009; Foldvik and others, 2004). The MITgcm is run at high resolution (0.1°) in order to resolve the effect of tidal currents and eddies on ocean circulation and melt rates in the sub-ice-shelf cavity. Tidal forcing has been found to have an important influence on the water flux and basal melt rate in the cavity (Makinson and others, 2011). The model output is validated against the latest oceanographic observations from moorings deployed through boreholes on the ice shelf. The ice melt rates from the model are compared with glaciological estimates. The model is used to investigate temporal variability in the ocean circulation and melt rates beneath the ice shelf.


Dating ice-rise formation in the Ronne Ice Shelf region, West Antarctica, using ice-penetrating radars

Jonathan Kingslake, Richard Hindmarsh, Edward King, Hugh Corr

Corresponding author: Jonathan Kingslake

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

The history of the West Antarctic ice sheet in the region currently occupied by the Ronne Ice Shelf is poorly known. This reflects a lack of accessible recently deglaciated surfaces, which prohibits conventional paleoglaciological techniques that can provide evidence of past ice-sheet extent and retreat, for example ocean coring or exposure-dating of geological material. We use a glaciological technique, Raymond Effect Dating, to constrain the retreat of the ice sheet through the Ronne Ice Shelf region. During two Antarctic field seasons, we used a pulse-echo ice-penetrating radar to image the base and internal stratigraphy of four ice rises – areas of grounded ice containing ice divides. Towing the radar with skidoos, we conducted over 2000 km of surveys on the Skytrain, Korff, Henry and Fowler Ice Rises and the ice shelf between them. We also used a step-frequency radar called pRES and dual-band GPS to measure ice flow in the vicinity of each ice divide. Isochronal ice layers imaged during the surveys deform in a predictable way with ice flow, so their shape contains information about past ice flow. Directly beneath ice divides the downward motion of the ice is impeded by an ice-dynamical phenomenon called the Raymond Effect. This causes layers beneath the divides to form ‘Raymond Arches’ that grow over time. We will present the data and simulate the growth of Raymond Arches using ice velocities measured with pRES and GPS to date the onset of ice-divide flow at each ice rise by comparing simulated arches to arches imaged during pulse-echo surveys. We consider the main sources of uncertainty associated with these ice-rise formation dates and discuss what they can tell us about the retreat of the West Antarctic ice sheet through this region during the last few thousand years.


The pattern of flow convergence across Antarctic ice-stream networks

Felix Ng

Corresponding author: Felix Ng

Corresponding author e-mail: f.ng@sheffield.ac.uk

The tributary organization of ice streams is an unsolved complexity that hinders reliable prediction of ice-sheet dynamics. Thermomechanical models can mimic branched and evolving ice-stream flow, and, with suitably tuned basal parameterizations, match observed ice-sheet-wide flow velocities. However, how networks of interacting ice-stream tributaries form, and what controls their dendritic (sometimes anastomosing) pattern, remain poorly understood. In this study, I present the first map of planimetric ice-flow convergence across the Antarctic ice sheet, calculated from satellite-measured surface velocities, and use it to explore the spectrum of interconnected flow structures that make up this complexity. Besides elucidating how tributaries draw ice from the ice-sheet interior, the map reveals curvilinear convergence features along lateral shear margins of streaming, and abundant convergence ‘ripples’ associated with nonlinear ice rheology and changes in bed topography and friction. Remarkably, flow convergence on ice-stream tributaries and their feeding zones is highly uneven and interspersed with divergence at length scales of kilometres. For individual drainage basins as well as the whole ice sheet, the range of convergence and divergence decreases systematically with flow speed, implying that fast flow cannot converge or diverge as much as slow flow. I consider several mechanisms that may cause this hitherto unknown, universal property of networked ice flow. These findings (the convergence map and the property) provide new targets and constraints for numerical ice-flow simulations, motivate more research into tributarization, and invite a new generation of theories to address the rich spatial structure of ice-sheet flow.


Dissolution by turbulent compositional convection of ice-shelf fronts and the sides of tabular icebergs

Ross Kerr, Craig McConnochie

Corresponding author: Ross Kerr

Corresponding author e-mail: ross.kerr@anu.edu.au

We present laboratory experiments and theoretical analysis that quantify the dissolution by turbulent compositional convection of a vertical ice face in homogeneous salt water. Our experiments use a purpose-built laboratory apparatus in which we grow bubble-free ice that is 1.2 m high. In our experiments, we varied the temperature of the salt water between 0 and 6°C. We observed laminar flow in the lower 10–20 cm of the ice wall, and turbulent convective flow up the remainder of the wall. In the region of turbulent flow, both the dissolution rate of the ice and the temperature of the ice–water interface were found to be independent of height. These results are explained using a theoretical model of the heat and salt transfer by turbulent compositional convection at a vertical boundary. We find that the dissolution velocity depends on the 4/3 power of the difference between the ocean temperature and its freezing point, and that it is reasonably consistent with observations of the ablation velocities of the sides of icebergs. Our analysis also predicts that a transition from dissolution to melting occurs at about 6°C.


The turbulent convective plume at ice-shelf fronts and the sides of tabular icebergs

Craig McConnochie, Ross Kerr

Corresponding author: Craig McConnochie

Corresponding author e-mail: craig.mcconnochie@anu.edu.au

We present laboratory experiments and theoretical analysis that quantify the turbulent buoyant plume formed by the dissolution of a vertical ice face in homogeneous salt water. In our experiments, we vary the temperature and salinity of the salt water and measure the dissolution rate of the ice, the temperature of the ice–water interface, the maximum vertical velocity of the buoyant plume, and the rate at which the laboratory tank becomes stratified with buoyant fluid. Using this experimental information, we then construct a theoretical model of the turbulent buoyant plume as a function of height. The plume has a top-hat entrainment coefficient of 0.048 ± 0.006, and is found to have substantial drag. The plume model is used to calculate a plume width, velocity, buoyancy and Reynolds number for typical dissolving icebergs and ice-shelf fronts. Our plume model is also adapted by adding peeling detrainment from a linear velocity profile and a quadratic density profile, which gives an accurate description of the stratification that forms in the tank as a function of time. Our laboratory experiments also examine the effect of a linear salinity gradient on the dissolution of a vertical ice face. As the stratification is increased, the dissolution rate, the interface temperature and the maximum vertical plume velocity all decrease, and their dependence on height changes. We also outline a method of scaling the effects of stratification from our laboratory experiments to the much larger vertical scales of ice shelves and icebergs.


Using idealized models to explore uncertainties in ice-shelf–ocean interaction

David Gwyther, Eva Cougnon, Benjamin Galton-Fenzi, Michael Dinniman, Jason Roberts, John Hunter

Corresponding author: David Gwyther

Corresponding author e-mail: david.gwyther@gmail.com

Ice shelves are an important control on Antarctic mass loss and global sea-level rise. The ability of ice shelves to provide back-stress on tributary glaciers is controlled by the rate at which mass is lost into the ocean. This is principally through basal melting and iceberg calving (calving has recently been suggested to be enhanced by strong basal melting). However, processes of ice-shelf–ocean interaction that lead to and influence basal melting are poorly understood. We present results from idealized ice-shelf–ocean numerical simulations, designed to investigate several uncertainties. The experiments are based on the Ice Shelf–Ocean Model Intercomparison Project framework, and run with a version of the Regional Ocean Modeling System modified for ice-shelf–ocean interaction, including a modification to the parameterization of basal melting at low circulation. These experiments are as follows: (1) The effect of varying basal drag on melting/freezing is presented. Melt rate varies as a function of drag; higher drag increases melting. However, very high drag decelerates the boundary layer flow, and increased turbulence is offset by reduced flow, leading to a plateau and slight reduction in melting. Refreezing zones display more complex behaviour. (2) Simulations with a spatially varying drag coefficient (high drag for refreeze zones and low drag for melt zones) that evolves with the spatial pattern of melting/freezing are presented. A high drag coefficient in refreeze zones leads to modified melting across the entire ice shelf. This has important implications for ice-shelf–ocean models which produce refreezing, but apply a uniform low drag coefficient that is more suitable for melting. (3) Simulations investigating both cold and hot cavity environments and, in particular, the transition between the two environments are shown. For progressively warmer cavity environments, the distribution of melting shifts from a maximum on the western (outflow) side to a maximum on the eastern (inflow) side. In the cold ice-shelf cavity environment, melting is controlled by circulation, as opposed to the hot cavity, where melt distribution is driven by the availability of heat.


Modelling the impact of a major calving event in a region of high Dense Shelf Water formation

Eva Cougnon, Ben Galton-Fenzi, Guy Williams, Steve Rintoul, John Hunter, Benoit Legresy

Corresponding author: Eva Cougnon

Corresponding author e-mail: Eva.Cougnon@utas.edu.au

Global climate is partly governed by the thermohaline circulation. The densest water mass in the ocean, Antarctic Bottom Water (AABW), in the Australian-Antarctic basin has freshened and decreased in volumes by 50% over the last few decades. The reasons for these changes are poorly understood but one mechanism is by freshening of Dense Shelf Water (DSW) formed on the continental shelf that contributes to the formation of AABW. The main source of DSW in the Antarctic-Australian basin is formed in the region of the Adelie basin, near the Mertz Glacier Ice Tongue (MGT). However, this region has subsequently changed with the calving of the MGT in February 2010. This major calving event impacted the DSW formation rates on the continental shelf, mainly due to changes in the surface forcing due primarily to sea-ice production and cover. However we still do not know the combined effect of calving and of changes of the surface forcing in the longer term. Here we present results from a numerical modelling study showing the impact of the MGT calving event on the DSW export. We compare results from simulations with the pre-calving and post-calving ice conditions, in a way to predict the DSW formation in the Mertz region, once the region (iceberg and fast-ice positions) has stabilized. We also perform simulations with alternate surface forcing in order to simulate the natural variability of the sea-ice production. Small changes in surface forcing associated with the calving event can significantly impact the DSW formation rates.


Seasonal resonance in the diurnal frequency band on the continental slope in the southern Weddell Sea: do shelf waves enhance tidal eddy kinetic energy?

Stefanie Semper, Elin Darelius

Corresponding author: Stefanie Semper

Corresponding author e-mail: stefanie.semper@student.uib.no

At the shelf break and upper continental slope in the southern Weddell Sea, diurnal tidal currents are exceptionally strong and enhance the cross-shelf exchange of water masses. Shelf break processes affect the interaction of cold, dense shelf waters with warmer off-shelf water masses contributing to deep water formation, and are important for heat transport onto the continental shelf. All available current meter moorings from the southern Weddell Sea (29 moorings), spanning the period from 1968 to 2014, have been used to calculate eddy kinetic energy (EKE) in the diurnal tidal frequency band. Tidal EKE is anomalously high at the shelf break and decays rapidly with increasing distance. The tidal EKE at the shelf break shows a pronounced maximum in austral summer and a second maximum in austral winter, while tidal models give two symmetric maxima. The asymmetry between observed and modelled tidal EKE may be explained by seasonally present continental shelf waves (CSWs). We use a numerical code describing shelf waves to investigate the possibility of seasonally resonant diurnal CSWs. The dispersion relation shows a maximum (i.e. zero group velocity and trapped energy) that, depending on the seasonally varying stratification properties of the slope current, move in and out of the diurnal tidal frequency band. A seasonal cycle of hydrography on the slope is constructed using mooring data and CTD profiles collected by ship and CTD.


Controls on the front positions of tidewater glaciers

Jaime Otero, Francisco Navarro, Javier Lapazaran, Roman Finkelnburg, Darek Puczko, Ethan Welty

Corresponding author: Jaime Otero

Corresponding author e-mail: jaime.otero@upm.es

Hansbreen is a grounded tidewater glacier in southern Spitsbergen, Svalbard, with a rich history of field and remote-sensing observations. These include: glacier surface topography from the SPIRIT DEM; subglacial topography from ground-penetrating radar (GPR); hourly 2 km resolution surface mass balance from the European Arctic Reanalysis (EAR); center-line glacier velocities measured from a set of stakes (May 2005–April 2011); weekly front positions from time-lapse photographs (September 2009–Sept. 2011) and ASTER images; and sea-ice concentrations from time-lapse photographs and Nimbus-7 SMMR and DMSP SSM/I-SSMIS passive microwave data (http://nsidc.org/data/nsidc-0051.html). The available data make this glacier a good candidate for evaluating and comparing various mechanisms and controls of calving, some of which are tested in this contribution. In particular, we use a full-Stokes thermomechanical flow model (Elmer/Ice), incorporating a crevasse-depth calving model, to estimate Hansbreen’s front position at a weekly time resolution. The basal sliding coefficient is calibrated after each month by solving an inverse model. We investigate the effect of some possible controls of calving, such as back-pressure at the front (in turn related to sea-ice concentration) or basal water pressure, by examining the model’s ability to reproduce the observed seasonal cycles of frontal advance and retreat.


Antarctic ice-rise formation, evolution and stability

Lionel Favier, Frank Pattyn

Corresponding author: Lionel Favier

Corresponding author e-mail: lionel.favier@ulb.ac.be

Antarctic ice rises originate from the contact between ice shelves and one of the numerous topographic highs emerging from the edge of the continental shelf. While investigations of the Raymond effect indicate their millennium-scale stability, little is known about their formation and their role in ice-shelf stability. Here we present for the first time the simulation of an ice rise using the BISICLES model. The numerical results exhibit several field observable features, such as the substantial thinning downstream of the ice rise, and the successive formation of a promontory and ice rise with stable local flow, showing that ice rises are formed during the ice-sheet deglaciation. We quantify the ice-rise buttressing effect, found to be mostly transient, delaying grounding line retreat significantly but resulting in comparable steady-state positions. We demonstrate that such topographic features are key in controlling simulations of Antarctic deglaciation.


Modelling on-shelf ocean heat transport along the West Antarctic Peninsula

Jennifer Graham, Michael Dinniman, John Klinck

Corresponding author: Jennifer Graham

Corresponding author e-mail: j.graham@uea.ac.uk

The flux of warm deep water onto the Antarctic continental shelf plays a vital role in determining water mass properties on the shelf, and availability of heat to melt ice shelves. Two regional models, with differing grid resolution, have been used to simulate ocean processes along the West Antarctic Peninsula (WAP) for 2006–2012. At both 4 km and 1.5 km resolution, the model is able to reproduce the locations of cross-shelf transfer. However, the 1.5 km simulation shows a greater on-shore heat transport, leading to improved representation of heat content on the shelf, consistent with observations from the region. The increased heat transport is attributed to increased eddy activity both at the shelf break and off-shore. The internal Rossby radius on the WAP shelf is approximately 5 km. Therefore, a grid resolution of 4 km is unable to resolve such small-scale eddies. The mean along-shore transport shows a similar structure for both simulations, with no significant change in position or strength of ocean fronts. However, increased eddy activity leads to a greater cross-front heat transport, and therefore increased on-shelf heat transfer above the continental slope. Cross-shelf troughs are know to be key locations of on-shelf heat transfer. Comparison of two troughs, Belgica Trough and Marguerite Trough, shows differing responses to increased resolution. At higher resolution, there is an increased on-shore transport at Belgica Trough, but not at Marguerite Trough. This is related to the differing structure of the shelf-break jet between these two locations. Eddy activity also plays a significant role in on-shelf heat transport away from these troughs. While the two simulations have differences in the magnitude of heat transport, they do show similar patterns of interannual variability. Since both models are forced with ocean climatology at the boundary, this variability must result from changes in local atmospheric forcing over the domain. These results have significant implications for understanding the cross-shelf transport of water masses around Antarctica, as well as variability in ice-shelf melt rates.


Ice flow pattern and dynamical evolution of the Greenland ice sheet during the last deglaciation

Lisbeth T. Nielsen, Christine S. Hvidberg

Corresponding author: Lisbeth T. Nielsen

Corresponding author e-mail: lisbetht@nbi.ku.dk

The basal boundary conditions of the Greenland ice sheet have an important control on the ice flow pattern and dynamical evolution of the ice sheet. The flow in ice streams is dominated by basal sliding thought to depend strongly on basal hydrology and material, and both ice rheology and the basal sliding processes are important factors for the ice-sheet dynamics and contribute to the pattern and timing of ice-sheet response to climate changes. However, parameterizations of these processes in large-scale ice-sheet models remain uncertain due to lack of direct observations of the conditions at the base of the ice sheet, contributing to uncertainties in the projections of the ice sheet’s past and future behavior and stability. Large climatic shifts experienced by the ice sheet in the past offer the possibility to investigate the long-term effects of the evolution of the ice flow in response to these changes on the ice-sheet evolution. In this study, we investigate the evolution on millennial timescales of the large-scale dynamics of the Greenland ice sheet during the transition from the Last Glacial Maximum to the present warm period, using the Parallel Ice Sheet Model (PISM). We focus on the sensitivity of ice-sheet response to uncertainties in ill-constrained model parameters describing basal conditions and ice deformation properties. These parameters include the enhancement factor, the exponent in the basal sliding relation employed in the model and the parameterization of the strength of the basal material. In order to investigate the sensitivity of the ice flow pattern and dynamical evolution of the ice sheet to parameters controlling ice flow dynamics, we run the model through a glacial–interglacial cycle for an ensemble of model parameter sets. In all model runs, we use a reconstruction of surface boundary climate forcing based on ice-core-derived temperature anomalies. All experiments show a considerable increase in ice velocities over the entire ice sheet at the transition to warmer temperatures, and additionally show that the largest changes occurs along the margins, with a similar distinct, regional pattern of the response found in all experiments.


Susceptibility of the Antarctic ice sheet to changes in ice-shelf buttressing

Johannes Fürst, Gaël Durand, Fabien Chillet-Gaulet, Laure Tavard, Olivier Gagliardini

Corresponding author: Johannes Fürst

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

Higher surface air temperatures over the Antarctic Peninsula are hypothesized to have caused melt-pond formation, destabilization and sudden disintegration of the Larsen B ice shelf in 2002. The almost total removal of the shelf resulted in an acceleration of the extant glacier fronts, upstream thinning and unabated ice loss up to this day. Similar thinning is observed for Thwaites and Pine Island Glaciers in the Amundsen Sea sector, but here ocean warming is suspected for enhancing the shelf melting. In both cases, shelf geometries were altered in a way that upstream buttressing was reduced, an explanation for the observed accelerations. Since more than half of all Antarctic glaciers extend into floating shelves and since most of them showed no significant accelerations in the recent past, it remains unclear how susceptible the upstream ice sheet is to geometric changes of the corresponding shelves under further warming in the future. In this context, we aim at quantifying the dynamic susceptibility using ice geometry and surface velocities inferred from observations. To obtain the stress distribution near the grounding line, the shelf viscosity field is determined using a variational inverse method that optimizes the mismatch between observed and modelled surface velocities. This allows us to compute a buttressing factor along the grounding line. Using this factor as one criterion, we succeed to a priori discern the segments of the grounding line in the Amundsen Sea sector that, in fact, retreated by now. An abrupt drop-off in buttressing across the main trunk of Thwaites Glacier can explain its asymmetric retreat pattern. Moreover, other regions in this sector are recognized as susceptible to further loss of shelf buttressing, where, for now, perturbations remain too weak for a distinct migration of the grounding line. With the chosen criteria, we are able to localize the regions that are prone to changes in the downstream shelves. This identification enables us to classify each Antarctic glacier and ice stream according to the retreat susceptibility.


A new 3-D full-Stokes model as a tool for basal inversions

Teresa Kyrke-Smith, G. Hilmar Gudmundsson, Patrick Farrell

Corresponding author: Teresa Kyrke-Smith

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

High-resolution models of ice-sheet dynamics are required to make accurate predictions of the future mass balance of ice sheets. These require knowledge of flow conditions at the bed of the ice; however, the inaccessibility of the bed means there exist few observational constraints. Inverse methods are therfeore commonly used to obtain information about the nature of basal control using given surface observations. We present a new 3-D Stokes solver written using FEniCS with the potential to carry out second-order inversions for basal slipperiness. We show solutions and calulcations of the first and second order adjoint and prelimiary results from basal inversions over patches of Pine Island Glacier. Pine Island Glacier is one of the fastest-flowing and most rapidly changing ice streams in Antarctica, and is currently contributing to sea-level rise at an increasing rate. Recent field seasons as part of the iSTAR project have acquired high-resolution in situ geophysical measurements; results from our model will be compared with these to try and increase understanding about the conditions at the bed of Pine Island Glacier.


Landsat- and InSAR-derived grounding-line dynamics in the Bellingshausen Sea sector of West Antarctica

Frazer Christie, Robert G. Bingham, Noel Gourmelen

Corresponding author: Frazer Christie

Corresponding author e-mail: F.Christie@ed.ac.uk

Recent studies in West Antarctica have aimed to elucidate the drivers of ‘dynamic thinning’, a process by which oceanic and/or atmospheric forcing instigates accelerated thinning of coastal ice. These studies have been especially concentrated in the Amundsen Sea sector, where the largest losses of ice have occurred over the past decade, and from where there are now considerable observational records of glacial, ice-shelf and ocean change to constrain ice–ocean–atmosphere interactions. There are, however, other parts of West Antarctica, notably the sector draining into the Bellinghausen Sea, where dynamic thinning is also occurring. In this sector, the pace and timing of dynamic thinning, and whether it is responding to the same or different atmospheric or oceanic stimuli to forcing in the Amundsen Sea sector, have remained poorly resolved. Using grounding-line migration as a proxy for dynamic thinning, we present changes to grounding-line position along the Bellingshausen Sea facing coastline of West Antarctica as observed over multiple epochs from 1985 with optical and radar satellite imagery. Using Landsat MSS/TM/ETM+ imagery, we derive advance and retreat rates along the coastline for 5 year periods from 1985 to 2010. For parts of the coastline, we validate the trends inferred from the optical imagery with ERS-1/ERS-2 double-differenced interferometric synthetic aperture radar (InSAR)-derived grounding-line positions for 1992, 1994, 1996 and 2011. Analysis of grounding-line positions reveals: (1) significant grounding-line retreat (~330 ± 0.06 m a–1) at Ferrigno Ice Stream, which unlike glaciers and ice streams throughout the remainder of the coastline is not buttressed by substantial downstream ice shelves; and (2) additional noteworthy retreat at the glaciers and ice streams flowing into the southern Stange Ice Shelf (~100 ± 0.06 m a–1). These findings correspond with observations of recent rapid glacial thinning across both regions. However, we find: (3) insignificant grounding-line retreat along the margin of Venable Ice Shelf in the last 25 years, despite observations of significant surface thinning over the last decade.


Ocean mixing beneath the Pine Island Glacier ice shelf

Satoshi Kimura, Pierre Dutrieux, Adrian Jenkins

Corresponding author: Satoshi Kimura

Corresponding author e-mail: satmur65@bas.ac.uk

Pine Island Glacier (PIG) has shown a dramatic acceleration and thinning over the past few decades. One of the leading causes of this thinning is the increased basal melting due to warming of Circumpolar Deep Water (CDW) on the adjacent continental shelf. The basal melting is controlled by the turbulent transport of momentum, heat and salt beneath ice shelves; however, direct measurements of turbulence beneath ice shelves are rare, mostly due to the difficulty in gaining access through the ice shelf. Here we present turbulence measurements collected from the NERC autonomous underwater vehicle (AUV), AUTOSUB 3, missions beneath the PIG ice shelf. The AUV completed two missions totalling 92 hours covering 460 km of track beneath the PIG ice shelf. Typically a level of turbulence, ε, is estimated from a free-falling instrument to minimize the vibrations of instruments, whereas our measurements are influenced by the vibrations of the AUV. We will discuss a technique to estimate ε from the AUV. The estimates give a reasonable agreement with the measurements from the free-falling instrument near the AUV track. The grounding line of PIG and the rest of the cavity are separated by a ridge. The water column above the crest of the ridge limits the replenishment of CDW into the grounding line. Mixing above the ridge fluctuates with a tidal cycle. On the grounding line side of the ridge, the inflow of CDW and outflow of meltwater sustain a baroclinic shear leading to elevated mixing regardless of the tidal cycle.


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. Further, there is evidence that ocean forcing in the Bellingshausen Sea, west of the Antarctic Peninsula, is contributing to the loss of glacial ice. Salinity is the dominant control on density near the freezing point of sea water, hence fresh water plays a key role in the dynamics of the Bellingshausen Sea. However, the different components of the freshwater balance – glacial ice, sea ice and precipitation – are affected by warming in complex ways. Oxygen isotope data exist from which information on the freshwater balance can be obtained, both in a contemporary context (from research cruises and time series sites) and on longer timescales from sediment records. However, interpretation is made difficult by the general sparsity of such data. To better understand the freshwater balance of the Bellingshausen Sea, a high-resolution model of the region has been developed, using MITgcm to represent ocean, sea ice and ice shelves. Experiments have been conducted that track the advection, mixing and fate of fresh water from different sources, and demonstrate the differing spatial and temporal scales of their impacts on ocean structure. Results will be used to investigate the causes and consequences of warming and freshwater change west of the Antarctic Peninsula.


Coupling ice–ocean models 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 coupled ice–ocean model is being developed using MITgcm (Massachusetts Institute of Technology general circulation model) to investigate these processes. The ocean model is a z-coordinate model which presently represents circulation under static shelves. A wetting/drying scheme has been developed to allow for synchronous ice-shelf thickness changes, which in turn are provided by the Land Ice component of MITgcm. We present preliminary experiments with our coupled model.


Impact of highly resolved atmospheric forcing on the Southern Ocean circulation using the Finite Element Sea-ice Ocean Model (FESOM)

Marta Kasper, Ralph Timmermann, Willem Jan van de Berg

Corresponding author: Marta Kasper

Corresponding author e-mail: marta.kasper@awi.de

The acceleration of fast-flowing outlet glaciers has been identified as the main reason for increased mass loss of the Antarctic ice sheet (AIS) in recent decades. This was attributed to enhanced ice-shelf basal melt rates, which depend on the circulation in the ice-shelf cavities. In this study we aim to simulate the ocean circulation around Antarctica as well as in the cavities using two high-resolution models. We use the global Finite Element Sea-ice Ocean Model (FESOM), which proved to be a useful tool for such investigations. In this configuration the model includes ice-shelf–ocean interaction using a three-equation system for computation of temperature and salinity at the interface. Together with bathymetry data from RTopo-1 resolving the cavities, FESOM is able to simulate the circulation below the ice shelves with a resolution up to 4 km. For the first time the ocean model was forced with high-resolution (27 km) atmospheric data for the Southern Ocean, using the output from the Regional Climate Model RACMO2. The latter is able to resolve katabatic winds, which locally have a strong impact on sea-ice formation and thus ocean circulation. Here we present first results from experiments where we replaced ERA-Interim forcing by RACMO2 in some regions of the Southern Ocean. The focus of this study is on the impact of the resolution of atmospheric forcing on ocean circulation and the distribution of sea-ice properties in the Antarctic marginal seas.


Controls on heat transport, submarine melting and ice melange stability at Kangerdlugssuaq Fjord (East Greenland) from numerical modelling experiments

Tom Cowton, Andrew Sole, Donald Slater, Peter Nienow

Corresponding author: Tom Cowton

Corresponding author e-mail: tom.cowton@ed.ac.uk

Fjords provide a key link between the Greenland ice sheet and the ocean, yet their role in modifying ice–ocean interaction is poorly understood. In particular, the circulation of waters within the fjords remains uncertain. As a result, it is not clear how the temperature of water adjacent to the glaciers responds to processes occurring on the continental shelf, whether the up-fjord transport of oceanic heat could change in a warming climate, or how the movement of water adjacent to the termini influences submarine melt rate, calving and glacier stability. We explore these questions by using a modified version of the MIT general circulation model (MITgcm) to simulate the circulation of water in the Kangerdlugssuaq Fjord system (East Greenland). We find that while coastal wind events are responsible for the most energetic periods of exchange between the fjord and continental shelf, these events have only limited effect on water properties adjacent to the glaciers at the head of the fjord. Instead we find that estuarine circulation resulting from the input of glacial runoff is the mechanism most effective at transporting warm subsurface waters from the shelf to the head of the fjord. Through this mechanism, variability in ocean temperature is transmitted to the glaciers at sub-seasonal timescales. In addition to its effect on fjord-scale circulation, the subglacial input of runoff further influences the glacier by modifying the circulation close to the calving front. In keeping with previous studies, we find that the turbulent upwelling of buoyant runoff generates a significant increase in the rate of submarine melting at the calving front. However, we also find that these plumes are responsible for transporting large volumes of warm subsurface waters into the cooler upper part of the water column. This process could enhance glacier calving by increasing the thermal and mechanical weakening of the ice melange buttressing the glacier termini. We therefore propose that increased runoff from the ice sheet during warmer summers will have a three-fold effect on Greenland’s marine-terminating glaciers: drawing more warm water along the fjord, increasing the plume-driven melting of the calving front and transporting a greater volume of warm water up into the ice melange. As such, we expect the greatest destabilizing effect when periods of high ocean temperatures coincide with warm summers and high runoff.


The thermal structure of the Recovery Glacier drainage basin, Antarctica, derived from three-dimensional numerical flow modelling

Thomas Kleiner, Daniel Steinhage, Angelika Humbert

Corresponding author: Thomas Kleiner

Corresponding author e-mail: Thomas.Kleiner@awi.de

The basal conditions of the ice sheet are of crucial importance for the dynamics of the ice, as they determine the sliding at the ice base. Since the base is hardly accessible for in situ observations, indirect measurements of the basal properties are of great interest. The bed reflectivity derived from the radar-measured bed returned power (BRP) has been widely used to distinguish between wet and dry basal areas of a glacier. As the BRP depends on the loss due to englacial dielectric attenuation, which itself is primarily a function of ice temperature, the vertical thermal structure within the ice must be known along the radar profiles. In January 2014 an airborne radio-echo sounding campaign of the Alfred Wegener Institute was carried out in the Recovery Glacier drainage area. The newly derived ice thickness measurements have been incorporated into the Bedmap2 datasets for ice thickness and bedrock elevation replacing data in the so far unmapped area. In addition to the original Bedmap2 geometry this new geometry is used for numerical ice flow simulations using the Parallel Ice Sheet Model (PISM). Simulations are carried out across a broad range of datasets for present-day boundary conditions for the Antarctic ice sheet. Here we present our analysis of the thermal structure of the Antarctic ice sheet based on the model results with a strong focus on the Recovery Glacier area. We compare our results with simplified 1-D vertical heat transport assumptions and discuss the implications for radar bed returned power estimates.


Flow speed within the Antarctic ice sheet and its controls inferred from satellite observations

Robert Arthern, Richard Hindmarsh, C. Rosie Williams

Corresponding author: Robert Arthern

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

There are now a great number of satellite and airborne observations of the large ice sheet that covers Antarctica. These include maps of the surface elevation, the ice thickness, the snowfall rate, the surface flow speed of glaciers, and maps of how the surface elevation is changing over time. Uncertainty in the possible rate of future sea-level rise motivates using all of these observations and computational models of the flowing ice to make projections of how the ice sheet as a whole might behave in future, but this is still a challenge. The ice sheet can be several kilometres thick, but most of the observations identify quantities at the upper surface of the ice sheet, not within its bulk. To make useful predictions using glaciological models of the ice sheet we need accurate information from beneath the surface. We need to find out where cold-based parts of the ice sheets are frozen immobile onto rock, and where the ice rests on slippery water-lubricated sediment that allows it to flow rapidly towards the ocean. We also need to assess which parts of floating ice shelves around the edge of Antarctica are damaged and unable to provide much resistance to the ice that flows from the main part of the ice sheet into the ocean. We need information about how quickly the ice is flowing deep within the ice sheet, and which physical processes beneath the ice can affect this flow. Finally, we need to understand how the future behaviour of the ice sheet relates to what we observe today and how changes happening now at the edge of the ice sheet could in due course affect ice that is presently hundreds of kilometres inland. We describe how a combination of ice-sheet modelling and measurements from satellites and aircraft has allowed us to explore these questions and visualize what is happening in parts of Antarctica that we cannot visit.


Melt pond formation on the Larsen C ice shelf, Antarctica

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

Corresponding author: Suzanne Bevan

Corresponding author e-mail: s.l.bevan@swansea.ac.uk

The Larsen C ice shelf (LCIS) is the largest ice shelf on the Antarctic Peninsula and is gradually being exposed to conditions that preceded the collapse of its northerly neighbour, the Larsen B ice shelf. Collapse preconditions include increased surface melt and firn densification, melt ponding, thinning, rift propagation and ice-flow acceleration. In Cabinet Inlet on the northern part of the LCIS, föhn winds descending from the Antarctic Peninsula mountains can lead to positive surface air temperatures and increased summer surface melt. When the meltwater generated exceeds the storage capacity of the firn layer, surface melt ponds are able to form. Upon refreezing, the subsequent modification of density and temperature will potentially impact the rheology and dynamics of the ice shelf and hence its stability. Melt ponds are clearly visible in optical satellite images, and an ongoing analysis of MODIS data from 2000 to 2015 for Cabinet Inlet reveals that ponding does not occur every year. It is likely that the formation of visible surface ponds depends on the balance between snow accumulation and snowmelt over a number of years leading up to the summer melt season. Optical televiewing of a borehole drilled in Cabinet Inlet in November 2014, as part of the MIDAS project, shows an unusually high density ice body commencing within 3 m the surface extending to a depth of 45 m. Once the summer melt exceeds the storage capacity of the firn down to the top of this ice body, melt ponds can form. We will present a time–depth analysis of snow density based on RACMO2.3 output for Cabinet Inlet from 1990 to 2015, which shows a firn layer that varies in depth between 0 and 8 m overlying a dense ice layer many metres thick. The firn layer accumulates over successive low melt/accumulation years only to be rapidly depleted in high melt summers. For example, after a high-melt season in 2008/09 and until 2015, low ratios of melt to accumulation allowed a 3.1 m deep layer of firn to build, during this period we see no evidence of surface water in Cabinet Inlet within the record of MODIS data.


Instability and sensitivity of the Amundsen Sea ice streams

Isabel Nias, Stephen Cornford, Tony Payne

Corresponding author: Isabel Nias

Corresponding author e-mail: isabel.nias@bristol.ac.uk

Present-day ice loss is centred on the Amundsen Sea Embayment (ASE), in West Antarctica. The stability of this area is a key control on global sea level. Within the ASE, ice loss is primarily associated with ice streams draining the area, including Pine Island (PIG), Thwaites (TG) and Smith (SG) glaciers. A perturbed parameter ensemble was performed in order to understand the differences in sensitivity between these three ice streams. This was achieved using BISICLES, a vertically integrated higher-order flow model with adaptive mesh refinement, which provides fine resolution in the region of the grounding line and shear margins, while applying a coarse mesh elsewhere. Latin hypercube sampling was used to generate 64 parameter sets in which three physically based parameters (associated with basal traction, ice rheology and sub-shelf melt rate) were altered. For each parameter set BISICLES was run for 50 years using two bed geometries: one based on BedMap2 and another based on an inverse method, which applies mass conservation to velocity and surface elevation data. Initial results show that the ice streams respond differently to changes in the various parameters. Dynamic retreat of the ice streams is particularly sensitive to changes in the viscosity of the ice. The sub-ice-shelf melt rate plays a role in the stability of PIG and SG, as they both have constrained ice shelves, which acts as a buttress to the grounded ice. The results are also very sensitive to bedrock geometry. The mean rate of sea-level rise across the whole domain over the 50 year period is 0.44 mm a–1 for the BedMap2 simulations, compared with 0.34 mm a–1 for the simulations using the geometry modified according to mass conservation.


Föhn winds on Larsen C ice shelf generate an unusually massive ice layer

Adrian Luckman, Bryn Hubbard, Dave Ashmore, Bernd Kulessa, Suzanne Bevan, Peter Kuipers Munneke, Ian Rutt, Martin O’Leary, Michiel van den Broeke, Paul Holland, John King, Daniela Jansen

Corresponding author: Adrian Luckman

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

Surface melt and ponding has been observed on many Antarctic ice shelves and is implicated in ice-shelf collapse through firn compaction and hydrofracture (van den Broeke, 2005). Ice-shelf surface meltwater can percolate into the firn, transferring heat to deeper layers by refreezing (Vaughan, 2008). This can lead to warming and densification to the point where firn air content approaches zero (Holland and others, 2011), potentially impacting ice dynamics and fracture toughness. Surface processes also contribute to the recent thinning observed on Larsen C ice shelf (LCIS; Pritchard and others, 2012; Holland and others, 2015). In northern LCIS, föhn winds provide extra sensible heat to drive surface melt (Luckman and others, 2014). The NERC MIDAS Project (Impact of Melt on Ice Dynamics and Stability: 2014–2017) aims to investigate the mechanisms of melt and ponding, and test their impact on the stability of the LCIS. In November 2014 we visited Cabinet Inlet in northern LCIS to install an ARGOS-enabled automatic weather station, drill a 100 m borehole, and determine ice layering and temperature using an optical televiewer (OPTV) and logged thermistor string. The OPTV log shows a ~40 m layer of refrozen meltwater perched above glacier ice presumably advected from beyond the grounding line 40 km away. The thermistor data show the ice to be many degrees warmer than mean atmospheric temperatures would lead us to expect. We infer the source of this unusual ice configuration and temperature profile as föhn-driven surface melt. In April 2015, well after normal mean surface temperature had fallen below freezing, several föhn wind events lasting a few days raised the temperature well above 0°C and, in MODIS satellite data, melt ponds were seen to form. We continue to investigate using numerical modelling the potential impact on ice-shelf stability of melt, warming and massive ice layers.


New high-resolution views of the bed of Pine Island Glacier, West Antarctica

Robert G. Bingham, Damon Davies, Edward C. King, David G. Vaughan, Stephen L. Cornford, Alex M. Brisbourne, Alastair G.C. Graham, Jan de Rydt, Andrew M. Smith, Matteo Spagnolo

Corresponding author: Robert G. Bingham

Corresponding author e-mail: r.bingham@ed.ac.uk

Pine Island Glacier (PIG) in West Antarctica is currently losing ice at a rate equivalent to ~7% of current sea-level rise, and predicting its future is therefore an important scientific goal. Though the glacier has now been the focus of several modelling studies, the different models disagree on the likely future pace of loss and its spread inland. Significantly, all models depend critically on the form of the subglacial conditions used, and though the general form of the bed has been mapped from surveys over the last decade, the resolution of bed required for modelling to be improved, i.e. at the sub-km scale, has hitherto been unavailable. Addressing this dearth of detailed bed information was therefore a key objective for the 2013/14 UK iSTAR (ice Sheet sTAbility Research programme) traverse across PIG. We deployed the British Antarctic Survey’s DEep-LOoking Radio Echo Sounder (DELORES) to sound 10 × 15 km patches of the bed in six locations across PIG. Each patch was surveyed in 22 parallel transects lying 500 m apart and which were each 15 km long. Along each radar transect, the bed was sounded approximately every 5 m. The patches sample the main trunk of the ice stream, the beds of four of the main tributaries, and, as a control site, an inter-tributary ridge. Here we present an image of the ice-sheet bed at each of the six sites. We show that the nature of the bed varies significantly between sites, and make some interpretations of the subglacial forms relative to those observed in palaeo-ice stream settings.


Exploring the use of transformation group priors and the method of maximum relative entropy for Bayesian glaciological inversions

Robert Arthern

Corresponding author: Robert Arthern

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

Ice-sheet models can be used to forecast ice losses from Antarctica and Greenland, but to fully quantify the risks associated with sea-level rise probabilistic forecasts are needed. These require estimates of the probability density function (PDF) for various model parameters, such as the basal drag coefficient and ice viscosity. To infer such parameters from satellite observations it is common to use inverse methods. Two related approaches are in use: (1) minimization of a cost function that describes the misfit to the observations, often accompanied by explicit or implicit regularization, or (2) use of Bayes’ theorem to update prior assumptions about the probability of parameters. Both approaches have much in common and questions of regularization often map onto implicit choices of prior probabilities that are made explicit in the Bayesian framework. In both approaches questions can arise that seem to demand subjective input. One way to specify prior PDFs more objectively is by deriving transformation group priors that are invariant to symmetries of the problem, and then maximizing relative entropy subject to any additional constraints. Here we investigate the application of these methods to the derivation of priors for a Bayesian approach to an idealized glaciological inverse problem.


Year-long monitoring and imaging of the Pine Island Glacier ice shelf using phase sensitive radar

Lai Bun Lok, Paul Brennan, Matt Ash, Keith Nicholls

Corresponding author: Lai Bun Lok

Corresponding author e-mail: l.lok@ucl.ac.uk

In January 2014, three ground-based ApRES radars were deployed on the ice shelf of Pine Island Glacier. The radars, including antennas, were custom-designed to monitor and image ice shelves in monostatic and multiple input multiple output (MIMO) modes, respectively. The systems operated throughout the year, each powered by a single lead-acid battery. The data from all three systems were successfully recovered in December 2014. This presentation summarizes the results of our initial analyses of these ‘year-long’ RES datasets, the first to be obtained from the ice shelf of Pine Island Glacier. At the location of deployment, the monostatic data reveal an observed 4 m reduction in the range to the ice-shelf base during the year. The data, recorded at two-hour intervals, highlight the sub-millimetre precision monitoring capability of the ApRES system. We also describe a method of processing the MIMO data to generate two-dimensional images through the ice shelf. At the time of writing, all three radar systems are still operating on the ice shelf and collecting RES data. It is planned to recover the new datasets in the 2015/16 field season.


Surging and calving activity in Svalbard from an unusually dense series of velocity maps

Adrian Luckman, Doug Benn, Heidi Sevestre

Corresponding author: Adrian Luckman

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

Measurement of glacier dynamics is important in understanding processes of glacier flow as well as monitoring the glacial response to climate change. Glacier velocity can be measured by in situ techniques such as GPS, or by remote sensing using feature-tracking or satellite radar interferometry. In situ methods provide very high temporal resolution but for limited duration and over a very restricted area. Remote sensing, on the other hand, allows measurement of extensive velocity fields but is usually quite restricted in temporal resolution by standard satellite acquisition strategies and/or cloud cover and polar night. The CRIOS Project (Calving Rates and Impact on Sea level) has provided the funding for a unique set of observations to bridge the gap between these two observational extremes. During 2013 and 2014 we acquired images from TerraSAR-X almost every satellite cycle for a number of image frames over Svalbard and Greenland. The repeat image reliability of SAR, and the high spatial resolution (2 m) and short revisit time (11 days) of TerraSAR-X has led to an unprecedented series of velocity maps covering dozens of glaciers. This has led to new insights into seasonal variability in glacier dynamics in Svalbard and Greenland, and particularly into the processes of calving at tidewater margins and surging in tidewater and land-terminating glaciers. In this presentation we will consider controls on calving speed in Svalbard glaciers and highlight the wide variety of styles of glacier surge observed within our dataset.


Modelling the dynamic response of Jakobshavn Isbræ, West Greenland, to calving rate perturbations

Johannes H. Bondzio, Hélène Seroussi, Mathieu Morlighem, Thomas Kleiner, Martin Rückamp, Angelika Humbert, Eric Larour

Corresponding author: Johannes H. Bondzio

Corresponding author e-mail: johannes.bondzio@awi.de

Jakobshavn Isbræ, one of Greenland’s major outlet glaciers, displayed rapid changes since the mid-1990s. Its floating ice tongue broke up around 1997, followed by a rapid calving front retreat over tens of kilometres, possibly linked to a prior warming of the ocean waters adjacent to the fjord. Parallel to this major retreat, a quasi-simultaneous process of acceleration and thinning of the glacier has been observed, currently making it a major contributor to eustatic sea-level rise. However, the causal interplay between the various factors involved has not yet been fully understood. Numerical studies of Jakobshavn Isbræ so far are either 2-D plan view ice flow models with a fixed calving front or 2-D flowline models, which have to parameterize lateral stresses. Hence the interaction between changes in calving front position and ice dynamics could not be studied consistently. To overcome this limitation, we implemented an implicit boundary tracking method in the Ice Sheet System Model (ISSM). This tool allows us to freely evolve the calving front by prescribing a calving rate, the ice front velocity being therefore the sum of ice velocity and calving rate. A suite of sensitivity experiments perturbing an initial steady-state calving rate has been performed to study its impact on the dynamics of the glacier. Our numerical results suggest a high sensitivity of the glacier dynamics to the applied calving rate. Changes in calving rate quickly affect upstream areas of the ice stream through a combination of changes in calving front position, ice velocity, thickness, grounding and ungrounding, and surface gradient change. Consequently, acceleration triggered at the calving front quickly affects the entire drainage basin. Moreover, the model results suggest that the ice stream does not recover from a short duration (about a year) calving rate perturbation over timescales on the order of a century. We present selected results of the sensitivity experiments to support the discussion clarifying the causes of the current changes occurring at this dynamic ice stream.


Ice fabric characteristics from shear-wave anisotropy using passive icequakes in Rutford Ice Stream, West Antarctica

Emma C. Smith, J. Michael Kendall, Alan F. Baird, Alex M. Brisbourne, Andrew M. Smith

Corresponding author: Emma C. Smith

Corresponding author e-mail: emit1@bas.ac.uk

Ice fabrics are formed by the deformation of ice under stress and will, in turn, affect the future response of that ice to the stresses induced by ice flow. Fast-flowing ice streams are the dominant method of mass discharge for the West Antarctic ice sheet (WAIS) and therefore understanding the ice fabric in these areas is key both to understanding their deformation history and our ability to model and predict their future evolution. We investigate the ice fabric of Rutford Ice Stream, West Antarctica, by measuring shear-wave splitting in naturally occurring basal icequake signals. 3698 good quality splitting measurements are made on data recorded over a 34 day period using ten three-component geophones 40 km upstream of the grounding line. A shear wave is split as it travels from the base of the ice stream through a seismically anisotropic fabric, travelling faster in one orientation than another and being recorded at the surface as two separate arrivals. Measurements of the polarization of the fast shear-wave direction and the time difference between the two arrivals are made. The magnitude of the splitting ranges from 1 ms to 90 ms and decreases with increasing angle of incidence at the receiver. The orientation of the fastest shear-wave varies across the ice stream, indicating a spatially changing fabric pattern. These results are inverted for a variety of fabric models that can be used to infer the types of crystal preferred orientation (CPO) fabrics within Rutford Ice Stream. This informs both deformation history and current rheological anisotropy, which can be used for improved ice-flow modelling.


The ocean’s influence on glacier retreat in Patagonia

Carlos Moffat

Corresponding author: Carlos Moffat

Corresponding author e-mail: carlos.moffat@ucsc.edu

The observed retreat of the vast majority of glaciers in the Patagonia Icefields has been well documented, but the role that the large network on fjords and the adjacent South Pacific Ocean is playing in this process has not. Here the only existing multi-year oceanographic observations from a proglacial fjord in Patagonia are discussed. The site, adjacent to Jorge Montt Glacier (the northernmost glacier in the Southern Patagonia Icefields) has been retreating for most of the 20th century, with increased rates nearing 1 km a–1 in recent years. Seasonal oceanographic cruises from 2010 to 2015 are used to characterize the evolution of the freshwater plume generated by the glacier melt in tidal to interannual timescales. High-resolution velocity fields were measured using long-term moorings and a novel, shipboard dual acoustic Doppler current profiler system designed to resolve, simultaneously, the surface fresh layer and the deep, warm ocean inflow. Concurrent microstructure profiles were collected during the surveys to characterize the intensity and spatial structure of the mixing in the fjord. The results show that Jorge Montt (and, is suggested, many other Patagonian glaciers) are particularly sensitive to the variability of surface not deep water off the coast of Patagonia, which can reach up to 11°C. This makes the Patagonia Icefields subject to one of if not the warmest ocean forcing in the world. The supply processes for this warm oceanic water to the ice involve a combination of (1) shelf and fjord processes away from the proglacial fjord, (2) the estuarine-like circulation driven by the glacier itself and (3) a katabatic wind-forced supply mechanism acting in the proglacial fjord, which is heavily modulated by the glacier itself. The resulting freshwater discharge from the glacier varies significantly in synoptic to interannual scales, and appears to be leaving the fjord system largely at depth due to the strong surface stratification caused by other sources of fresh water in the region. The results are discussed in the context of the ocean influence of similar systems in Alaska, Greenland and Antarctica, as well as of the potential future changes in the Patagonia Icefields driven by the ocean.


Developing observationally constrained, regime-specific projections of basal melting

Christopher Little, Nathan Urban

Corresponding author: Christopher Little

Corresponding author e-mail: clittle@aer.com

Process studies suggest that ocean model horizontal resolutions of approximately 1 km are required to capture key oceanographic exchanges across the Antarctic continental shelf break, under ice-shelf fronts, and through the ocean–ice-shelf boundary layer. Atmosphere–ocean general circulation models (AOGCMs) do not resolve these small-scale processes and thus exhibit considerable biases in relevant Antarctic Continental Shelf Bottom Water (ASBW) properties. Furthermore, ice shelves (and some continental shelves) are absent from current AOGCMs. For the purposes of assessing ocean-driven changes in Antarctica’s mass balance, AOGCMs may thus be better used to inform the expected range of large-scale changes in ocean temperature; models that account for oceanic, atmospheric and glaciological processes may be used to link far-field boundary conditions to ice dynamics. This strategy (analogous to the use of regional climate models for surface mass-balance analyses) requires choosing appropriate ‘source’ regions in AOGCMs, assessing uncertainty in ocean heat content in those regions, and translating that to the ice sheets. In Antarctica, this is complicated by limited observations and differing oceanographic regimes on continental shelves. Here we present analyses comprising part of a more comprehensive strategy for developing (1) probabilistic projections of ice-sheet mass changes and (2) standardized forcing scenarios for ice-sheet model intercomparisons. First, we develop large-scale 21st century projections of ASBW thermal anomalies using water-mass-based definitions, continuous sampling from AOGCM ensembles and observational constraints. We then test regime-specific relationships between large-scale thermal anomalies and ice-shelf basal melt rates using: (1) reanalyses and observations of ASBW and (2) oceanographic and glaciological measurements. We conclude by comparing basal melt rates to other recent approaches and suggesting how information from ongoing regional observational and modeling efforts might be used to better constrain the ASBW/melt rate relationship.


Adding antiplane shear to Röthlisberger channels

Colin R. Meyer, Matheus Fernandes, James R. Rice

Corresponding author: Colin R. Meyer

Corresponding author e-mail: colinrmeyer@gmail.com

We examine superimposed antiplane shear on the in-plane creep closure of a Röthlisberger channel, a meltwater conduit that is maintained due to a balance between turbulent heat dissipation melting the channel walls and in-plane viscous creep closure of the ice. Although the shearing does not factor explicitly in the in-plane creep closure, it enters through the effective viscosity when assuming that ice deforms as a power-law shear-thinning fluid with a rheological power given by Glen’s law. We first develop a closed-form small perturbation solution, for which the amount of superimposed antiplane shear strain rate is much smaller than the in-plane strain rate. At linear order, the antiplane shear perturbation has no effect on the creep closure of the Röthlisberger channel because the effective viscosity is set by the in-plane creep closure. We also examine the limit where the antiplane strain rates dominate the in-plane strain rates, and find a simple scaling for the average creep closure velocity. Then, using numerics, we compute the viscous creep closure and determine the size of the Röthlisberger channel as a function of the superimposed antiplane shear. The results are interpreted for Röthlisberger channels that flow out of the grounding line.


Variability of seasonal Greenland glacier velocities and implications for ice-sheet sensitivity to ocean and surface meltwater changes

Twila Moon, Mark Fahnestock, Ted Scambos, Marin Klinger, Terry Haran

Corresponding author: Twila Moon

Corresponding author e-mail: twila.moon@nsidc.org

Greenland ice sheet mass loss is a primary contributor to sea-level rise. Understanding current and future mass loss, however, requires knowledge of the ice-sheet response to other components of the climate system, including the atmosphere and ocean. Earlier research suggests that ice-sheet motion, a key parameter for calculating mass loss, can indicate outlet glacier sensitivity to seasonal subglacial hydrologic evolution and terminus advance and retreat. These variances are expressed as distinct seasonal velocity patterns across the ice sheet. Differences in seasonal velocity behavior include local and regional divisions, as well as changes from year to year. Via these distinct velocity signatures, it is possible to improve understanding of the ice-sheet response to increases in surface meltwater production via atmospheric warming and changes in ocean conditions. To expand on earlier research, we use a new velocity dataset derived from Landsat 8, which has unprecedented temporal coverage (weeks to months) across the full Greenland ice sheet. We investigate the spatial variability of glacier surface velocities from the terminus towards the ice-sheet interior and across neighboring glaciers, including those with different seasonal velocity patterns. Results suggest that glaciers across the ice sheet may respond differently to the same climate forcing, providing valuable insight for understanding future mass loss from the Greenland ice sheet.


Evidence of multiple distinct subglacial meltwater plumes observed using the REMUS-100 autonomous underwater vehicle

Laura Stevens, Fiamma Straneo, Sarah Das, Albert Plueddemann, Michiel van den Broeke, Mathieu Morlighem

Corresponding author: Laura Stevens

Corresponding author e-mail: stevensl@mit.edu

Buoyant plumes fed by submarine melting and subglacial discharge drive mass loss along marine-terminating margins of the Greenland ice sheet. Near-ice (<200 m) hydrographic measurements are lacking due to the difficulty in working at the margin of calving glaciers. The consequent extrapolation of far-field hydrographic measurements to the ice/ocean boundary in observation-based interpretations and model parameterizations results in highly uncertain estimates of submarine melt rates and in an inadequate understanding of how marine-terminating glaciers respond to oceanic and atmospheric forcing. Here we present detailed hydrographic and bathymetric measurements collected as close as 150 m from the ice/ocean interface of the Sarqardliup Sermia/Sarqardleq Fjord system, West Greenland, with the REMUS-100 autonomous underwater vehicle. We find evidence of two distinct subglacial meltwater plumes discharging along the tidewater margin as indicated by the spatial distribution of glacially modified waters observed in the REMUS-100 observations of temperature, salinity and turbidity. These observations are consistent with the existence of two subglacial subcatchments beneath Sarqardliup Sermia routing subglacial meltwater out of two primary, distinct channels along the grounded margin, and resulting in two distinct glacially modified water masses. Discrepancies between these data and idealized buoyant plume models suggest that significant mixing takes place over the 150 m between our observations and the vertical ice/ocean interface.


Simulations testing the thermal regime of an ice sheet on a regional domain

Mark Pittard, Jason Roberts, Ben Galton-Fenzi, Christopher Watson

Corresponding author: Mark Pittard

Corresponding author e-mail: mark.pittard@utas.edu.au

The temperature of the Antarctic ice sheet has an important influence on the ice flow. For example, ice temperature modifies where sliding will occur and directly influences the rate of deformation through the rate factor in the flow law. Numerical models often do not use temperature as a diagnostic variable due to the paucity of available observational constraints. One of the key observational-derived inputs that controls ice temperature is the geothermal heat flux. This is poorly constrained and at low resolution due to difficulties in observing under thick ice sheets. Using the enthalpy scheme based Parallel Ice Sheet Model (PISM) we assess the impact of varying the geothermal heat flux on ice-sheet evolution on a regional ice-sheet model of the Amery–Lambert basin. We compare two different geothermal heat flux datasets, and test the effect of three different variants of each. The variants tested include a spatially constant mean model, and two scaled variants at 50% and 150% to test a plausible range of uncertainty. We compare the regional model to borehole temperature data at six locations on the Amery Ice Shelf and discuss the thermal regime of the Amery–Lambert basin. We find that the geothermal heat flux has a non-trivial effect in regions of low ice velocity with differences of up to a few hundred metres in thickness, which, while seemingly small, has a large effect downstream on the ice-sheet evolution. The temperature of the ice in the model compares poorly to the observed borehole temperatures, which was not unexpected given the simplicity of the basal melting within the model.


Development of a numerical ice-sheet/ice-shelf model IcIES: numerical exercises in ice-sheet simulation

Fuyuki Saito, Kunio Takahashi, Ayako Abe-Ouchi

Corresponding author: Fuyuki Saito

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

Ice sheet model for Integrated Earth system Studies (IcIES) has been developed to simulate past and future ice-sheet evolution. The model adopts the shallow ice approximation for the grounded part, the shallow shelf approximation for the floated part, and optionally adopts the Schoof (2007) parameterization for computing grounding-line flux. Because of strong non-linearity in the governing equation, large computing resources are demanded to solve the equation with high accuracy. Practically, truncation at some stage has to be decided on the basis of the balance between available computer resources and numerical accuracies, especially for long-term application. This study presents impact on the ice-sheet/shelf simulation of various timescales by different technical details such as a convergence criteria in the matrix solver, under ideal and realistic configuration including the Antarctic ice sheet.


Uncertainty in the contribution of the Greenland ice sheet to future sea-level rise assessed with an integrated approach

Reinhard Calov, Andrey Ganopolski, Alexander Robinson, Mahe Perrette, Johanna Beckmann, David Alexander

Corresponding author: Reinhard Calov

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

We present simulations of the contribution of the Greenland ice sheet to sea-level rise under short- and long-term future scenarios. Here, we utilize the ice-sheet model SICOPOLIS coupled with the regional climate system model of intermediate complexity REMBO. The loss of ice into the fjords via outlet glaciers is resembled in a parameterization. In particular, the impact of ocean temperature on mass loss is included in our parameterization. This work is a preparation of our planned more comprehensive approach, which will resolve all important processes in a model of intermediate complexity of the Greenland glacial system. We perform large ensembles of simulations under future scenarios to investigate uncertainties in atmosphere and ocean temperature, precipitation, geothermal heat and basal sliding. A suite of initial configurations of the ice sheet is generated via palaeo simulations over two glacial cycles by introducing present-day and Eemian constraints, such as a range in mass-balance partition, the shape of the ice sheet, the Eemian elevation drop at an upstream position of the NEEM and further local constraints. Our work adds to previous approaches. In particular, we fathom the range of possible future contribution of the Greenland ice sheet to sea-level rise and we assess the role of different processes governing the dynamics of the Greenland glacial system.


Marine ice-sheet model resolution depends on basal processes

Rupert Gladstone, Ben Galton-Fenzi, Roland Warner, Ralf Greve

Corresponding author: Rupert Gladstone

Corresponding author e-mail: rupertgladstone1972@gmail.com

Computer models are necessary for understanding and predicting marine ice-sheet behaviour. However, there is uncertainty over implementation of physical processes at the ice base, both for grounded and floating glacial ice. Here we implement several sliding relations in a Stokes-flow marine ice-sheet flowline model, and demonstrate that model resolution requirements are strongly dependent on both the choice of basal sliding relation and the spatial distribution of ice-shelf basal melting. Sliding relations that reduce the magnitude of the step change in basal drag from grounded ice to floating ice (where basal drag is set to zero) provide improved convergence with resolution compared to a commonly used relation, in which basal drag is purely a power-law function of basal ice velocity. Sliding relations in which basal drag goes smoothly to zero as the grounding line is approached from inland (due to a physically motivated incorporation of effective pressure at the bed) provide further improvements to convergence with resolution. A similar issue is found with the imposition of basal melt under the floating part of the ice shelf: melt parameterizations that reduce the change in basal melting from grounded ice (where basal melt is set to zero) to floating ice provide improved convergence with resolution compared to parameterizations in which high melt occurs adjacent to the grounding line. Thus physical processes, such as subglacial outflow (which could cause high melt near the grounding line), would impact on capability to simulate marine ice sheets. For any given marine ice sheet the basal physics, both grounded and floating, governs the feasibility of simulating the system. The combination of a physical dependency of basal drag on effective pressure and low ice-shelf basal melt rates near the grounding line mean that some marine ice-sheet systems can be reliably simulated at a coarser resolution than currently thought necessary.


Marine- and land-terminating glacier retreat in Disko and Uummannaq Bays, West Greenland, 1985–2014

Ashley York, Karen Frey, Sarah Das

Corresponding author: Ashley York

Corresponding author e-mail: ayork@clarku.edu

The variability in glacier termini positions is an important indicator of overall glacier balance and the net effects of ice–ocean–atmosphere interactions. Glacier margins fluctuate on both seasonal and interannual timescales, and owing to logistical difficulties associated with field observations, satellite imagery provides a critical spatially and temporally extensive resource for monitoring glacier behavior. In general, glacier termini have been retreating globally over recent decades, but the magnitude of seasonal variation and overall retreat has proven unique to each glacier. The glaciers in central West Greenland are generally experiencing the same regional atmospheric forcing, yet previous studies have shown varying magnitudes of retreat among both marine- and land-terminating glaciers over the last 40 years. Some studies have found marine-terminating glacier retreat to be three to five times greater than that observed at land-terminating glaciers. Furthermore, land-terminating glaciers (unlike marine-terminating glaciers) have more of a direct response to atmospheric forcing and it is the interaction with the ocean that differentiates variability in fluctuations between the two types of outlet glaciers. In this study, we utilize Landsat imagery between the years 1985 and 2014 to digitize a seasonal and historical time series of glacier front positions of 19 marine-terminating glaciers in the Disko and Uummannaq Bay regions of central West Greenland. Additionally, we use a combination of different spectral indices to calculate changes in ice-covered area on Disko Island and Nuussuaq Peninsula. Our retreat results of both glacier types are inter-compared to better understand the importance of the interaction at the ice–ocean interface in dictating fluctuations of marine-terminating outlet glaciers.


Glacial fjord circulation: how is heat imported and meltwater exported?

Rebecca Jackson, Fiamma Straneo

Corresponding author: Rebecca Jackson

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

Fjords form a key link in the climate system by connecting glaciers of the Greenland ice sheet to the ocean. Submarine melting of glaciers in these fjords has been implicated as a driver of dynamic glacier changes in the past decades. However, there are no direct measurements of this melting, and little is known about the fjord processes that might affect melt rates. Here we explore the drivers of fjord circulation and ocean heat transport in Sermilik Fjord, near the terminus of Helheim Glacier, from moored records of velocity and water properties. We investigate the competing roles of buoyancy forcing from meltwater and external forcing from the shelf. Building on estuarine studies of salt fluxes, we assess the flux of heat and fresh water through the fjord and develop a new framework for inferring submarine melt rates from glacial fjord budgets.


Noble gases quantify meltwater distribution in a West Greenland fjord

Nicholas Beaird, Fiamma Straneo, William Jenkins

Corresponding author: Nicholas Beaird

Corresponding author e-mail: nbeaird@whoi.edu

The five stable noble gases do not interact with the biology or chemistry of the ocean, and are modified only by interactions with the atmosphere and with ice, making them excellent tracers of ice–ocean interaction. Noble gases have been used as effective tracers of glacial meltwater in the waters around Antarctica, but not in Greenland where their value is enhanced by the ability to differentiate submarine melt from subglacial discharge. We present the first systematic tracer study using noble gases in a glacier/fjord system in a mid-sized glacial fjord in West Greenland. The noble gas observations allow us to unambiguously quantify the fraction and distribution of submarine melt and subglacial discharge in the waters near the glacier. They also reveal the overturning circulation driven by glacier buoyancy forcing showing that the deep waters, near the glacier, contain water modified by submarine melt (melt-driven convection) only and the region near the surface contains both melt and subglacial discharge (convection-driven melting).


Antarctic Eemian ice-sheet–shelf dynamics controlled by sustained ocean warming

Johannes Sutter, Gerrit Lohmann, Malte Thoma, Klaus Grosfeld

Corresponding author: Johannes Sutter

Corresponding author e-mail: johannes.sutter@awi.de

The dynamic evolution of the West Antarctic ice sheet (WAIS) under a warming climate is of considerable interest since a destabilization of the marine-based grounded ice could lead to sea-level rise of several meters and distinct shifts in ocean circulation. We investigate the behaviour of the Antarctic ice sheet in warm climate scenarios with a focus on the last interglacial. Combining ocean temperature sensitivity experiments with a variety of basal shelf melt parameterizations, in a state-of-the-art 3-D ice-sheet model (ISM), we asses potential thresholds of WAIS collapse triggered by ocean warming and their consequences for LIG eustatic sea level. Forcing the ice sheet with LIG climates derived from a coupled GCM, with different initial ice-sheet configurations (Greenland and Antarctica), we cover a range of potential atmospheric and ocean states and their impact on the Antarctic ice sheet. Furthermore we assess the sensitivity of the ISM to different geometries (BEDMAP1 and 2) and resolutions. The transient model runs are forced by time slice experiments of a fully coupled atmosphere–ocean global circulation model, as well as different sets of sea level and bedrock reconstructions. First results show strong evidence for a WAIS collapse at peak interglacial ocean temperature anomalies exceeding 2°C. Ice-sheet–shelf dynamics feature strong sensitivities to the applied shelf melt parameterization, stressing the importance of a realistic representation of present-day melt rate distribution in ISMs. GCM LIG ocean temperature anomalies do not model sufficient warming to destabilize WAIS ice shelves and their tributary glaciers, due to coarse resolution and missing ice ocean feedbacks.


Observational constraints for a generalized constitutive relation for ice creep with damage

Chris Borstad, Ala Khazendar, Mathieu Morlighem, Bernd Scheuchl, Eric Larour, Eric Rignot

Corresponding author: Chris Borstad

Corresponding author e-mail: chrisb@unis.no

The remnant Larsen B ice shelf in Scar Inlet has progressively weakened and accelerated since the collapse of the northern part of the ice shelf in 2002. At the same time, the tributary glaciers feeding the shelf have thinned and accelerated, illustrating the importance of reductions in ice-shelf buttressing to the flux of grounded ice into the ocean. Here we use an ice-sheet model to assimilate velocity observations for the ice shelf in the years 2000, 2006 and 2010 to determine the changes in rheology, buttressing and damage for the shelf. We explore the constitutive parameter space occupied by the shelf through time, and uncover a characteristic pattern of ice softening once a threshold stress is reached. By defining a simple two-parameter softening curve beyond this threshold stress, we generalize the constitutive framework for ice creep to include evolving damage. Both of these new parameters – a threshold stress and a ductility parameter – have clear physical interpretations and are determined from the assimilated observations. We derive an analytical relation that associates the level of damage to the present level of strain rate, which reduces the problem of modeling damage ‘evolution’ to simply a problem of modeling the evolution of the parameters and processes that influence the stress balance in the ice. As in modeling damage growth in fully elastic modes of loading, damage in this new framework is an explicit part of the new constitutive framework, in contrast to all previous approaches to damage evolution that add or subtract increments of damage using ad hoc flux terms independent of the stress balance equations. We discuss the advantages of this new constitutive framework and ongoing applications for modeling the structural evolution of ice shelves and their buttressing influence on the ice sheet under various perturbation scenarios.


Heat flow pathways at 79North

Nat Wilson, Fiamma Straneo

Corresponding author: Nat Wilson

Corresponding author e-mail: nwilson@whoi.edu

Constraining melt rates beneath ice tongues and ice shelves is important in settings such as northern Greenland and Antarctica where submarine melt is a major or dominating source of mass loss. Furthermore, variations in melt rates driven by the ocean are thought to have upstream influences by modulating the capacity for the ice shelf to provide buttressing, restricting the ice discharge of marine glaciers. For floating ice tongues in northern Greenland, submarine melting is thought to be largely controlled by the ocean temperatures at the ice interface and the geometry of the ice tongue and the cavity, including sill depth, cavity width and fjord length. Here we present measurements of water properties collected from the 79North glacier cavity, an outlet glacier of the Northeast Greenland Ice Stream in northeastern Greenland. We identify pathways as well as possible modes of water exchange between the cavity and the continental shelf. We use a simple formula based on heat balance to infer recirculation rates in the cavity. Rapid flushing rates mean that the cavity properties may respond rapidly to offshore variability. We then compare the 79North system with a number of other ice tongue systems along the northern coast of Greenland in terms of different regimes of cavity renewal and submarine melting.


Ice stream stick–slip: dynamic rupture simulations

Brad Lipovsky, Eric Dunham

Corresponding author: Brad Lipovsky

Corresponding author e-mail: lipovsky@stanford.edu

Whillans Ice Stream (WIS), West Antarctica, flows to the sea by way of stick–slip sliding. Episodic stick–slip events represent a natural prod to the ice stream system that can be exploited, via judicious mechanical analysis, to understand both streaming ice dynamics and the properties of the bed. Stick–slip motion requires a rate-weakening rheology over some portion of the bed. While such a constitutive law is a foundational component of our understanding of tectonic earthquakes (e.g. Scholz, 2002), it is not commonly invoked to represent glacier sliding (e.g. Cuffey and Paterson, 2010). We present 2-D and 3-D numerical simulations of the stick–slip motion of WIS that capture the full inertial dynamics, seismic waves, and the evolution of sliding with rate- and state-dependent basal friction. We seek possible explanations for the observed slow average rupture velocities (~200 m s–1), maximum basal sliding velocity amplitudes (~50 m d–1) and duration of acceleration (~200 s) observed at on-ice GPS stations during the onset of sliding. One possibility is the existence of a very long nucleation length at the ice–bed interface, such that WIS events never quite escape the nucleation stage of the seismic cycle. Long nucleation lengths could arise from low basal effective pressure and/or rate-strengthening rheology over much of the bed. This initial acceleration of the ice stream is, from Newton’s second law, A ~ T / (rho H) for average shear stress drop T, ice thickness H, and ice density rho. The observed A leads to an estimated 3 Pa stress drop that must be reconciled with the final stress drop of 300 Pa inferred from the total slip and fault dimensions (Bindschadler and others, 2003). This apparent contradiction, as well as very long duration decay time, ~30 min, of the elevated sliding velocities during a stick–slip event, remain enigmatic features of the WIS stick–slip events. One idea we are exploring is that deceleration of the ice in the latter portion of a slip event is associated with a substantial decrease in basal traction from extensive rate-strengthening regions.


Submarine melting of icebergs in Sermilik Fjord, southeast Greenland, based on satellite remote sensing and hydrographic observations

Ellyn Enderlin, Gordon Hamilton, Fiamma Straneo

Corresponding author: Ellyn Enderlin

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

Changes in the magnitude and spatial distribution of freshwater input within glacial fjords can influence their properties and, in turn, their circulation and the heat delivered to marine-terminating outlet glaciers. Given the recent increase in iceberg discharge from many of Greenland’s marine-terminating glaciers and the simultaneous increase in the subsurface ocean temperatures on Greenland’s continental shelves, it is likely that the freshwater input into the glacial fjords due to melting of icebergs has increased, yet spatial and temporal variations in iceberg melting are unknown. Here we present iceberg submarine melt rates estimated from very high resolution stereo satellite image-derived digital elevation models using a volumetric-differencing technique. We analyze spatial and temporal variations in iceberg melt rates and water temperature from Sermilik Fjord, southeast Greenland, to elucidate the controls of iceberg melting and freshwater fluxes in Greenland’s glacial fjords.


Monitoring rift activity on the Filchner–Ronne, Amery and Ross ice shelves and their role in ice-shelf–ocean interaction

Catherine Walker, Britney Schmidt

Corresponding author: Catherine Walker

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

Iceberg calving from ice shelves accounts for nearly half of the mass loss from the Antarctic ice sheet, yet our understanding of this process remains fairly limited. The precursor to iceberg calving is large through-cutting fractures, or rifts, which can propagate for decades after they have initiated until they become iceberg detachment boundaries. We previously monitored the lengths of 78 rifts in 13 Antarctic ice shelves (Walker and others, 2013) using satellite imagery from the Multiangle Imaging Spectroradiometer between 2002 and 2012. We build on that work here by focusing on the rifts identified as ‘active’ by that study, particularly those on the Filchner–Ronne, Amery and Ross ice shelves, and compare them with those identified as stagnant. We will also focus on rifts in which we observed (abnormal) wintertime propagation. We also use NASA ICESat and Icebridge altimetry and elevation datasets to infer the vertical dimension of the rifts and their setting within the ice shelf. We will show profiles over the rifts, in addition to using the data to estimate the melange thickness within the rift to possibly shed light on the mechanical role of melange in stabilizing a rift (or not). We also employ Landsat and ASTER visible imagery to investigate the surface changes in the rift. The high resolution of each of these datasets allows for the detection of ‘small-scale’ features of ice shelves, like these rifts. We additionally focus on rifts that propagate near or into suture zones. Marine ice likely underlies suture zones and leads to mechanical heterogeneity associated with different ice types within the ice shelf through which the rift propagates. Suture zones have previously been noted to impose a stabilizing effect on rift propagation (Hulbe and others, 2010; McGrath and others, 2012); however, recent results including the studies by Walker and others (2013) and Jansen and others (2015) have shown that rifts can propagate across these features. Many of the rifts that were classified as active in the Walker and others’ (2013) study were front-initiated, and open to the ocean (thus, have only one propagating tip). These front-initiated rifts have been found to be sensitive to the arrival of tsunamis. Here we present additional work showing the injection of a pressure wave associated with a tsunami or similarly large wave anomaly into front-initiated rifts and associated widening and/or propagation, and discuss the role of these rifts on front stability.


Sustained ocean cooling observed in front of Pine Island Glacier in 2011–13

Benjamin Webber, Karen Heywood, David Stevens, Pierre Dutrieux, Adrian Jenkins, Povl Abrahamsen, Stan Jacobs

Corresponding author: Benjamin Webber

Corresponding author e-mail: k.heywood@uea.ac.uk

During the iSTAR research cruise to Pine Island Bay in the Amundsen Sea in early 2014, two moorings were recovered from in front of Pine Island Glacier (PIG), one of which was deployed by the NB Palmer cruise 09-01. Together, these provided an unprecedented 5 year time series of temperature, salinity and current velocity. These data reveal considerable seasonal and interannual variability in deep ocean temperatures, of sufficient magnitude to make a substantial impact on the melt rate of PIG. In particular, the period August 2011 to August 2013 was anomalously cold; comparison with ship-based summertime observations suggest the heat content at the glacier front in December 2012 was the coldest in the observational record. Similar cold anomalies are observed at other moorings within Pine Island Bay, and current observations suggest a reduction in the inflow of warm Circumpolar Deep Water to the glacier, as well as a reduction in the outflow of meltwater and in the concentration of ice cavity water. The broad spectrum of variability observed by the mooring network suggests that a combination of remote and local processes is at work. At seasonal timescales, local weakening in wind forcing is correlated with cooling, consistent with a weakening of the circulation bringing warm waters to Pine Island Glacier. At interannual timescales, a combination of weakened inflow at the continental shelf edge and regional processes could be at play, but the record is too short to draw statistically significant conclusions.


Melting ice shelves in the Amundsen Sea Embayment: role of poleward shifting winds and changing cavity geometries

Nicolas Jourdain, Pierre Mathiot, Gaël Durand, Julien Le Sommer, Paul Spence

Corresponding author: Nicolas Jourdain

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

The mass loss of West Antarctic glaciers has accelerated over the last 15 years, most likely in response to ocean warming in Antarctic coastal waters. This oceanic warming in Antarctic coastal waters has recently been suggested to be caused by the positive trend of the Southern Annular Mode. But the mechanisms controlling the changes in melt rates underneath outlet glaciers are still poorly understood. For instance, despite recent developments in glacier modeling, melt rates are usually prescribed in glacier models. This strongly limits the ability of glacier models to predict the future evolution of West Antarctic glaciers. Several ocean models are now able to simulate ocean circulation beneath ice shelves, therefore allowing a direct study of the mechanisms controlling the changes in melting rates underneath outlet glaciers. Building upon these developments, we here investigate the relative influence of poleward shifting winds and changes in ice-shelf cavern geometries on melting rates underneath West Antarctic glaciers. To this purpose, we use a regional ocean/sea-ice model configuration based on NEMO, centered on the Amundsen Sea, which explicitly represents flows in ice shelf cavities. A series of sensitivity experiments is conducted with different cavity geometries and under different atmospheric forcing scenarios in order to identify the leading mechanism controlling the changes in melt rates underneath West Antarctic glaciers over the 21st century. Our results provide a first assessment of the importance of coupling glacier models to ocean models for predicting the future evolution of outlet glaciers.


Supplementing ice-core time series at a small-scale Alpine glacier with a 3-D full-Stokes ice flow model using Elmer/Ice

Carlo Licciulli, Pascal Bohleber, Dietmar Wagenbach, Olaf Eisen, Olivier Gagliardini, Martin Hoelzle

Corresponding author: Carlo Licciulli

Corresponding author e-mail: Olaf.Eisen@awi.de

The cold glacier saddle Colle Gnifetti (CG) is the unique drilling site in the European Alps offering ice-core records substantially exceeding the instrumental period. In spite of an ice thickness not much exceeding 100 m, CG provides long-term ice-core records due to its low net accumulation and rapid layer thinning. However, net accumulation at CG is characterized by strong spatio-temporal variability causing depositional noise and, combined with a complex flow regime, upstream effects. These intricate glaciological settings hamper the full exploitation of the unique potential for long-term ice-core records of this site. Here we present first results from developing a new model attempt, i.e. 3-D full-Stokes with consideration of firn rheology, fully thermomechanically coupled, utilizing the finite-element software Elmer/Ice. A major objective is to map source trajectories of existing ice-core sites in order to evaluate potential upstream effects. An additional focus is to assist in finding a reliable age scale, especially targeting depths where annual layers can no more be counted. This includes the calculation of isochronous surfaces for intercomparison of different drilling sites within the CG multi core array. A considerable amount of empirical data has been collected at CG over the last few decades. Model input quantities comprise density and temperature profiles measured at the ice-core sites, surface topography and GPR-based bedrock topography. The model accuracy is limited especially by the latter, due to an uncertainty of typically 15%. Additional limitations arise from other model parameters that are not directly constrained by measurement, for example the mechanical stress on the glacier boundaries. To achieve better constraints, the model input quantities are iteratively adjusted to provide the best fit between model-derived and directly measured quantities. Here we present first results regarding the model validation based on comparison with empirical data, using for this purpose the measured surface velocities and net snow accumulation. Finally we discuss the next steps in building our model approach, which include comparing model results with ice-core-derived depth-dependent information like e.g. the observed layer thinning or the measured vertical age distribution as well as to use a flow law taking into account ice anisotropy, as empirical evidence suggests.


Rapid subglacial water system evolution triggered by a subglacial flood in West Antarctica

Matthew R. Siegfried, Helen A. Fricker, Sasha P. Carter, Slawek Tulaczyk

Corresponding author: Matthew R. Siegfried

Corresponding author e-mail: mrsiegfried@ucsd.edu

Water at the ice–bed interface exerts a primary control on the motion of glaciers and ice streams, yet a calibrated, quantitative link between subglacial hydrology and ice flow in Antarctica remains elusive, in part due to a lack of observations. One component of the Antarctic subglacial hydrologic system that is readily observable is the interconnected system of active subglacial lakes, which fill and drain on sub-decadal timescales. Often located beneath regions of fast ice flow, active lakes have been shown to account for a large fraction of the subglacial water within an individual hydrologic basin, thereby controlling the extent and timing of lubrication of the ice–bed interface downstream. Here we used 5 years of continuous GPS data on Whillans and Mercer ice streams, West Antarctica, to observe a cascading subglacial lake flood event and its impact on ice dynamics. This region moves via a tidally paced stick–slip cycle resulting from a delicate basal force-balance, and any perturbation to this force-balance should alter the tidal pacing of slip events. We identified a 10 month interruption to the typical stick–slip cycle of ice motion, which began as the lake reached peak volume and ended halfway through the flood event. We also identified an unambiguous, though modest, ice acceleration (relative to background regional ice deceleration) sustained for 2 years, with velocity fluctuations that correlate to lake drainage evolution. These observations suggest that changes in the Antarctic subglacial water system can fundamentally alter the basal force-balance of the ice stream and that the ice dynamic response is immediate, complex and evolves rapidly. Long-term, continuous observations are therefore critical to understanding the spatiotemporal scales on which subglacial hydrologic processes influence the larger ice-sheet system.


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, William Collins

Corresponding author: Daniel Martin

Corresponding author e-mail: DFMartin@lbl.gov

We present POPSICLES simulation results covering the full Antarctic ice sheet and the Southern Ocean spanning the period from 1990 to 2010. We use the CORE v. 2 interannual forcing data to force the ocean model. Simulations are performed at 0.1o (~5 km) ocean resolution with adaptive ice-sheet resolution as fine as 500 m. We compare 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. 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 . Stand-alone 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 comparable simulations with a Stokes momentum balance in both idealized tests (MISMIP-3d) and realistic configurations.


Patterns of variability in the Helheim Glacier/Sermilik Fjord system, southeast Greenland: a 5 year synthesis

Fiamma Straneo, Leigh Stearns, Dave Sutherland, Gordon Hamilton

Corresponding author: Fiamma Straneo

Corresponding author e-mail: fstraneo@whoi.edu

Increased submarine melting at the margins of Greenland’s glaciers is indicated as a potential trigger for the speed-up and retreat that began in the late 1990s. Progress on this hypothesis, however, has been hindered by a limited knowledge of the variability of the fjord waters, in the proximity of the glacier, and a limited understanding of how such variability may affect submarine melt rates. Here we present a synthesis of 5 years (2008–2013) of oceanic, atmospheric and glaciological data collected at or near Helheim Glacier, southeast Greenland, including in the proglacial fjord and nearby continental shelf, from moored instrumentation and summer surveys. These data show that the seasonal to interannual variability in ocean properties at the edge of the glacier can be attributed to changes in the adjacent subpolar gyre of the North Atlantic Ocean and to the variations in subglacial discharge from the glacier’s catchment basin. We discuss how these records can be used to estimate changes in submarine melt rates and how, in turn, they are related to changes in glacier velocity and terminus retreat.


Autonomous ocean observations beneath the Pine Island Glacier ice shelf, West Antarctica

Pierre Dutrieux, Adrian Jenkins, Satoshi Kimura, Stanley S. Jacobs, Karen Heywood

Corresponding author: Pierre Dutrieux

Corresponding author e-mail: dutrieux@uw.edu

Warm Circumpolar Deep Water reaching 3.5°C above the in situ freezing point pervasively fills a network of glacially carved troughs in the Amundsen Sea, West Antarctica, and melts and thins neighbouring ice shelves, including the Pine Island Glacier ice shelf (PIIS). Hydrographic, current and microstructure observations obtained in austral summer 2009 and 2014 by an autonomous underwater vehicle beneath the PIIS are used here to detail the complex ice–ocean interaction and resulting ocean circulation. The theoretical schematic of deeply incoming warm and saline water melting the grounding line and generating a buoyant plume upwelling along the ice draft is consistent with observations. The cavity beneath PIIS is clearly divided in two by a sea-bed ridge, constraining the oceanic circulation and water mass distribution. On the seaward side of the ridge, a thick warm deep water layer circulates cyclonically and is overlaid by a thin meltwater plume. Only intermediate depth waters are allowed to overflow from the ridge top into the inner cavity, where a much thinner warm water layer is now overlaid by a thicker meltwater layer. At the ice/ocean interface, melt-induced freshening is forcing an upwelling which in turn injects cyclonic vorticity and participates in creating a vigorous cyclonic recirculation in the inner cavity. The top of the ridge, where warm waters overflow in the inner cavity, is a dynamical boundary characterized by northward along-ridge currents up to 0.2 m s–1, and enhanced mixing.


1000 year adaptive mesh simulations of Antarctic ice dynamics

Stephen Cornford, Daniel Martin, Antony Payne, Esmond Ng

Corresponding author: Stephen Cornford

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

Numerical modelling of Antarctic ice dynamics becomes more demanding as the simulation time increases, partly because drainage basins evolve and even merge over long timescales, and to a great extent because fine-scale features – such as the grounding line – can migrate over continental length scales. Century long calculations – for example, the simulations of Pine Island Glacier described by Joughin (2010), Favier (2014) and Seroussi (2014) – need only consider single ice streams, and can take advantage of the relatively little grounding line migration likely to occur to limit fine resolution to a region close to the present-day grounding line. As integration times grow the grounding line tends to sweep out a larger area – meaning that the region of fine resolution must either cover that growing area or evolve with it. At the same time, neighbouring ice streams may merge, so that they can no longer be treated separately. Ultimately, it becomes necessary to carry out simulations of the whole of Antarctica, potentially applying fine resolution everywhere. We present 1000 year simulations of the whole Antarctic response to simplified ocean forcing using the BISICLES ice-sheet model. Some of the simulations feature the complete collapse of the West Antarctic ice sheet, and we are able to use time-evolving adaptive mesh refinement to track the grounding line and reduce the computational complexity by orders of magnitude. We compare the size of pure numerical errors, caused by spatial and temporal under-resolution, with the differences due to approximation made in the model physics, and estimate an upper bound on the speed of West Antarctic collapse.


Validation of Antarctic coastline products

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

Corresponding author: Yixiang Tian

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

Sequential mapping of ice-sheet margins provides a simple and direct method for measuring the area and volume if ice-sheets advance or retreat in the Antarctic coasts. However, the coastal area is a significantly active area of Antarctica, since ice fronts, icebergs, glacier tongues and others are the most dynamic and changeable features. Therefore, constructing accurate coastlines is an important step for future change detection studies in order to understand the response of the Antarctic ice sheet dynamics to climate change. There are two commonly used methods to extract ice-sheet coastline from images. One is the digitizing method based on visual interpretation and the other is based on automated image processing and boundary extraction. The resulting coastline positions may not always be accurate because of image resolution, lighting, clouds, fast ice, terrain complexity and other factors. It is not easy to decide whether the difference between the coastlines of two periods is related to mass change or seasonal/annual change. We validated the quality of two Antarctic coastline products (MODIS coastline and RADARSAT coastline) and found that these products must be cross-inspected by using multitemporal remote-sensing data before they can be used for change detection and mass-balance estimation. Based on previous work, we separated non-ice-shelf coastlines from both coastline products. Then the distance between two coastlines for each region is calculated. Big change regions can be indicated by distance value and further analysis is done for these regions. Additional multitemporal Landsat images over the austral summer season from 1990 to 2006 are used to analysis the change between 1997 and 2004. Both bedrock and surface elevation data are used to reveal the 3-D ice sheet and terrain characteristics, which are very useful to distinguishing the actual changes and misinterpretations. After the analysis of the non-ice-shelf region, the error rates are calculated. The error rate of MODIS coastline is 3.88%. Correspondingly, the error rate of RADARSAT coastline is 2.56%. We inspect the image mosaics from which they were extracted along with the coastlines. It was found that most of them were extracted appropriately. Fast ice is very difficult to be distinguished from both MODIS and RADARSAT images. The backscatter of rock and water is similar in MODIS images, which is another main factor affecting the position extraction of coastlines.


Internal wave-driven mixing on the Amundsen Sea Shelf, West Antarctica

Georges Djoumna, David M. Holland, SangHoon Lee, Tae Wan Kim

Corresponding author: Georges Djoumna

Corresponding author e-mail: gdjoumna@gmail.com

Recent results on turbulent dissipation and mixing in the Southern Ocean have revealed that topographically rough regions are the prominent location for the generation and breaking of internal waves with associated diapycnal diffusivities K_{rho} of O(10–4–10–3) m2 s–1. However along the Amundsen coast where the observed thinning and acceleration of glaciers produces the majority of Antarctica’s contribution to sea-level rise little is known about the level of mixing. Mesoscale eddies play a key role for exchanges across the Antarctic shelf break, their interaction with small-scale topography is believed to generate internal lee waves. Enhanced generation of internal waves of tidal frequency is expected on the Amundsen Shelf due to its geographical location relative to the critical latitude (74°28’ S) for lunar semi-diurnal M_2 tide. We show that the critical latitude coincides with near-critical topography on the shelf and this condition favors the generation of M_2 internal waves. In this work, we describe and estimate the intensity and spatial distribution of the rates of turbulent kinetic energy dissipation $\epsilon$ and K_{rho} on the Amundsen Sea Shelf and study the relationship between the observed turbulence and the internal wave field in this region. Estimates of K_{rho}$ were obtained using fine-scale parameterizations. Higher mixing rates up to 5.4 × 10–3 m2 s–1 were seen in the depth between 200 and 600 m. The highest mixing rate exceeding 10–2 m2 s–1 was also seen in the southern end of Pine Island Glacier, attributed to observed shear. A latitudinal variability in K_{rho} near the bottom is reported, with K_{rho} increases near the critical latitude for M_2 tides. We show that the critical latitude coincides with near-critical topography on the Amundsen Shelf and this condition favors the generation of internal waves of M_2 frequency. Analysis of current time series from the moored instruments reveals a thickening of the frictional bottom boundary layer near the critical latitude. Although a weak semi-diurnal tidal dynamics was observed at the continental shelf, its combination with the critical latitude effects lead to enhanced mixing that potentially affects the heat budget and the circulation of the Circumpolar Deep Water below the ice shelves.


Testing a new physically based parameterization of ocean-induced melting

Hilmar Gudmundsson, Adrian Jenkins

Corresponding author: Hilmar Gudmundsson

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

We compare predictions of ocean-induced melt on ice shelves based on the commonly used plume model against melt rates estimates derived from flux divergence considerations. The plume model, in its most usually employed form, is found to predict unrealistic melt-rate distributions across both the Pine Island and the Thwaites Ice Shelves. The possible reasons for this discrepancy are discussed and modifications to the traditional plume model are suggested that allow for more realistic estimates of melt rates. We couple the improved melt-rate model to an ice-sheet flow model and discuss the implications for Pine Island and Thwaites Glaciers.


Impact of observed changes in Antarctic ice sheet mass balance on southern ocean properties and sea ice

Nacho Merino, Julien Le Sommer, Gaël Durand, Nicolas Jourdain, Pierre Mathiot, Gurvan Madec

Corresponding author: Nacho Merino

Corresponding author e-mail: ignacio.merino.cue@gmail.com

Ongoing changes of the Antarctic ice sheet (AIS) are suspected to be involved in the widespread changes of physical properties in the Southern Ocean. Changes in freshwater release from the AIS are likely participating in the observed freshening of the Antarctic Bottom Waters (AABW) around Antarctica. The observed increase in freshwater release from the AIS may also affect ocean surface stratification and convection regimes thereby contributing to the observed positive trend of sea-ice cover in the Southern Ocean. Quantifying how different AIS evolution scenarios may affect oceanic properties is therefore a key question for physical oceanographers and climate scientists alike. However ocean model studies usually use crude representations of the distribution and evolution of freshwater release from Antarctica. Here, we reconstruct two scenarios of Antarctic freshwater release (present-day and 90s) and study their impact on the Southern Ocean in a comprehensive eddy permitting ocean–sea-ice model with interactive icebergs. The scenario that will be presented explicitly accounts for the changes in ice-shelf volumes over the period. We will show the large-scale and regional patterns of AIS-driven changes in sea-ice cover and AABW properties. The impact of future changes in specific ice shelves around Antarctica will also be discussed.


Seasonal variations in the thermal structures of proglacial lakes in the Southern Patagonia Icefield

Masahiro Minowa, Shin Sugiyama, Daiki Sakakibara, Pedro Skvarca, Yoshihiko Ohashi, Takanobu Sawagaki, Nozomu Naito, Kazuhisa Chikita

Corresponding author: Masahiro Minowa

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

Patagonia Icefields are characterized by a number of outlet glaciers calving into lakes and the ocean. Some of the calving glaciers are rapidly retreating, playing a key role in the recent ice mass loss in Patagonia. Recent studies in Greenland and Alaska suggest the importance of submarine melting of the calving front in the recent mass loss of tidewater glaciers, but little is known about frontal melting of freshwater calving glaciers. To better understand frontal melting of freshwater calving glaciers, we measured temperature and turbidity in proglacial lakes in the Southern Patagonia Icefield. Filed measurements were carried out at Glaciar Upsala (840 km2) and Perito Moreno (259 km2). Upsala is one of the most rapidly retreating glaciers in the region, which flows into a ~600 m deep water at a rate of ~2 km a–1. Perito Moreno flows into a shallower lake (~170 m deep), having shown no significant retreat over the past century. Temperature and turbidity were measured in December 2013 and October 2014 at 8 (Upsala) and 21 (Perito Moreno) locations, by lowering a conductivity–temperature–depth profiler from a boat. In October 2014, water temperature and turbidity were uniformly distributed in the two lakes. Contrasting to this observation in spring, temperature and turbidity showed steep vertical gradients in summer. Mean water temperature in front of Upsala was similar in summer (2.4°C) and spring (2.6°C), which implies a larger amount of meltwater discharge from the glacier. In the summer measurement, turbid and cold water (~0.8 g L–1, <1°C) was found at the deepest part of the lake (>500 m deep), which is a strong indication of subglacial discharge. Contrasting to the observations in Upsala, cold deep water was not observed in the lake of Perito Moreno. Summer water temperature (6°C) was ~3°C warmer than in spring, and the warmest water (~8°C) was found near the lake surface. These data indicate seasonally and regionally different thermal structures in the two proglacial lakes. Water temperature was affected by subglacial meltwater discharge as well as the lake bathymetry. For example, relatively warm water was found in front of Perito Moreno probably because meltwater discharge is relatively small and a shallow lake is more influenced by energy input from the surface. The observed thermal structures and seasonal variations should play crucial roles in the melting of freshwater calving glaciers in Patagonia.


The Weichselian deglaciation of central Scandinavia

Trevor Faulkner

Corresponding author: Trevor Faulkner

Corresponding author e-mail: trevor@marblecaves.org.uk

‘Central Scandinavia’ extends from the Norwegian coast to the Caledonide thrust front in Sweden and from Grong northwards to Ranafjord. It is assumed that the Weichselian deglaciation of this area was driven by two summer heat fluxes: warming by the sea as it encroached up fjords and caused the ice-sheet margin to retreat eastwards; and warming by direct solar insolation that caused the ice sheet to ablate from its upper surface. An earlier empirical parabolic time relationship of H = 0.75t2 for melting height has been reconstructed as H = 1700+5(YD isobase–220)–0.000075(13500-t)2 metres, where t = 14C years BP, using information about the isostatic uplift that increased inland. Deglaciation was modelled by drawing deglacial maps at c. 300 year intervals, after the ice began to melt vigorously in the Bølling interstadial below the present 1700 m altitude (at the 220 m YD isobase at Børgefjell) that was already ice-free. The maps show the advance of the sea, the recession of the ice-sheet margin, the synchronous thinning of the ice sheet, and the evolution of ice-dammed lakes (IDLs). After c. 11k yr BP, the melting height lowered at roughly 0.5 m a–1 at all isobases. The formula agrees with many marine dates to within 300 years in the central and northern part of the area, but downwasting by calving was much faster along the coast, as the sea submerged large areas of the isostatically depressed strandflat. Across the Swedish border, there is evidence that large IDLs and glaciers remained longer than predicted. Most of the area was probably systematically flooded by glacial meltwater for periods of 800–1200 14C years during deglaciation as local IDLs lowered and coalesced (commonly at jökulhlaups). Six types of deglacial IDL are described, with hydrological flows initially via spillway cols, lateral meltwater channels and englacial conduits, and later into subglacial waterways that ran to the sea. The ice margin lost its sharp, morainal definition as it retreated under the increasing influence of topography. The IDLs reduced in size as the ice sheet shrank into separate valley glaciers, which, in Norway, eventually became tidewater glaciers at the deglaciation marine limit.


Sea-ice indices of ice-shelf ‘health’

Pat Langhorne, Craig Stevens, Wolfgang Rack, Mike Williams, Inga Smith, Christian Haas, Cecilia Bitz, Kenneth Hughes, Greg Leonard, Stefan Jendersie, Natalie Robinson, Andrew Pauling, Alex Gough, Pat Wongpan, Daniel Price, Tim Haskell

Corresponding author: Pat Langhorne

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

Antarctic coastal waters are conditioned by interaction with ice shelves. In turn, the formation and decay of sea ice must be influenced, although the areal extent of this influence is not well documented. Yet this is an important measure because sea-ice growth and decay are annual processes and likely to respond very rapidly to ice-shelf-induced changes in the near-surface ocean. Close to an ice shelf, sea ice often grows in water that has been supercooled by interacting with the ice shelf at depth. An important consequence is that the sea ice loses heat to the ocean as well as to the atmosphere. This ‘negative’ ocean heat flux causes the sea ice to grow thicker than it would without the ice shelf. The thermal deficit also means there are tiny frazil crystals in the water column. While they sometimes accumulate and grow on any object suspended in the near-surface ocean, they also accumulate under the sea ice where they form a porous layer of crystals in an evolving state of consolidation. This porous layer harbours some of the highest concentrations of sea-ice algae on Earth. The crystallographic structure is modified, leaving a detectable signature frozen into the sea-ice cover. Using in situ sea-ice measurements, we have derived an oceanic heat flux index that indicates the influence of ice-shelf-conditioned surface water on sea ice. This includes the longest ice–ocean record for Antarctica, which dates back to 1902 near the McMurdo Ice Shelf. These historical data indicate that, during the 20th century, any change in the volume of very cold surface outflow is less than the uncertainties in the measurements. In situ measurements are restricted in scale so this information is supplemented at the regional scale by the detection of the porous layer beneath the sea ice from airborne electromagnetic induction sounding, and computed sea-ice thickness anomalies derived from satellite altimeter measurements. These data are supported by modeling of ice–ocean interaction, at scales ranging from the growth of individual crystals, through regional models of ice-shelf–ocean interaction, to the injection of fresh water close to the surface in global climate models. It is critical that the thickness and extent of the porous layer beneath sea ice is monitored because of its importance to Antarctic ecosystems, and because it is likely to be highly sensitive to changes in the inaccessible space deep within an ice-shelf cavity.


Marine ice-sheet dynamics investigated using laboratory experiments

Roiy Sayag

Corresponding author: Roiy Sayag

Corresponding author e-mail: roiy@bgu.ac.il

The grounding zone of marine ice sheets separates between two distinguished flow regimes. One in which the flow is dominated by shear and the other in which the flow is dominated by extension. The ice response can differ substantially between these two flow regimes, resulting in viscous shear-thinning and brittle failure, respectively. Such nonlinear response, primarily due to the ice complex rheology and the transition across the grounding zone, can have important implications to grounding-line and ice-shelf stability, and calving. We study the role of rheology in such dynamical transitions using laboratory-scale experiments combined with theoretical analysis. We performed experiments with fluids that are much less viscous than ice, but with deformation law that follows Glen’s law (power-law fluid) though with a larger exponent. The fluid properties were measured by combining experiments of axisymmetric gravity currents (similar to a grounded ice sheet), together with theoretical modeling and measurements using a rheometer. We used the same fluids to investigate the flow across a grounding line in a marine ice-sheet apparatus, in which the fluids were discharged axisymmetrically into a tank filled with a denser salt solution representing an ocean. As the viscous fluid intruded the level of the salt solution was continuously maintained by an optical control. Consequently, two flow regimes were formed corresponding to a grounded sheet and a floating shelf, and separated by a grounding line. The grounding line remained axisymmetric throughout the experiments. However, unlike the case for Newtonian fluids, the flow axisymmetry broke down across the grounding line, and an array of tongue-like shelves was formed. In one set of experiments the grounding line was steady and the number of tongues did not change in time. In a second set of experiments the grounding line was allowed to evolve, resulting in the number of tongues increasing in time via splitting of pre-existing tongues. Using image-processing techniques we follow the evolution of the grounding line and of the arclength of each tongue, identifying characteristic dimensions at which tongue splitting initiate. Preliminary modeling confirms the critical role of rheology in the patterns we measure.


Analysing fracture and calving on the Totten Ice Shelf

Sue Cook, Jan Åström, Richard Coleman, Ben Galton-Fenzi

Corresponding author: Sue Cook

Corresponding author e-mail: sue.cook@utas.edu.au

Calving is a significant mass loss component for the Antarctic ice sheet, occurring largely at the edges of floating ice shelves. Due to the complicated nature of fracture and calving, current implementations in ice-sheets models do not allow us to predict how calving rates will be affected by changing climate in the future. However, calving rates are likely to be significantly affected as ice shelves become exposed to increased basal melting and possible hydrofracture from surface melt. The Totten Ice Shelf in East Antarctica is an example of a region where mass loss has already been observed, with thinning thought to be driven by high melting rates underneath the ice shelf. The ice shelf has also shown increases in velocity, and given that the calving front has remained stable an increase in calving rate can be inferred. We apply a discrete particle model to the Totten Ice Shelf and use it to analyse the pattern of fractures on the ice shelf. Unlike continuum models, the discrete particle method allows fractures to be included explicitly in the model, and is ideally suited to an analysis of the sources of fracture in the ice shelf. The model also allows us to examine how fracture patterns and calving rates might change as the acceleration and thinning of the ice shelf continue.


Ice melt rates in a subglacial outflow plume

Kenneth Mankoff, Fiametta Straneo

Corresponding author: Kenneth Mankoff

Corresponding author e-mail: kmankoff@whoi.edu

We present oceanographic observations collected in and immediately surrounding a buoyant freshwater sediment-laden subglacial outflow plume rising up the marine-terminating front of Saqqarliup Glacier, Greenland (68.9° N, 50.4° W). Subglacial outflow plumes entrain the relatively warm fjord waters and bring them in contact with the ice front, and are correlated with enhanced submarine melt and increased calving. Few in situ observations exist due to their location at the calving front of glaciers. Our data were collected from a small boat with winched sensors, with XCTDs dropped from a helicopter, and from the JetYak, a remote-controlled jet-ski-powered kayak. The plume delivers 56 m3 s–1 to a fjord with an almost uniform 1°C temperature. The freshwater flux is diluted to ~10% by the time it reaches the surface. Meltwater concentration in the plume is low (~0.1%), but ice melt rates >10 m d–1 in the vertical plume core are far higher than plume modeling studies predict for this low thermal forcing. Although there is disagreement between published models and our observations of meltwater concentration and melt rate, simple analytical and numerical models do reproduce plume flux and velocity with errors less than ~10%.


Detail mapping of a new subglacial lake at the ice divide between Institute Ice Stream and Minnesota Glacier in West Antarctica

Andrés Rivera, José Uribe, Rodrigo Zamora, Jonathan Oberreuter

Corresponding author: Andrés Rivera

Corresponding author e-mail: arivera@cecs.cl

The discovery of a near 20 km2 subglacial lake at the ice divide between Institute and Minnesota Glaciers (79°15’ S, 87°34’ W) in West Antarctica is reported. This water body was found in January 2014 during an oversnow traverse conducted by Centro de Estudios Científicos (CECs), Valdivia, Chile. The expedition included a VHF radar coherent system with high output power (200 W) using a pulse-compression method working at a central frequency of 155 MHz and a bandwidth of 20 MHz, among several other instruments. The expedition began at Union Glacier (79.8° S, 83.4° W) and was conducted to the high Antarctic plateau following a previously designed track aiming to map several subglacial valleys that extend in parallel to the Ellsworth Mountains. A second traverse in December 2014 confirmed the existence of the lake and provided a much more detailed mapping of its characteristics, thanks to the use of a regular and dense grid of measurements. The lake is located underneath a mean of 2653 m of ice at an altitude of 626 m below sea level. The radar and GPS data analysis determined that the lake is comprised of fresh water at a temperature of –1.8°C. The lake is located in a deep trough and is surrounded by steep flanks with peaks more than 100 m above the lake/ice interface. In general, the subglacial topography in the vicinity of the lake is very rough topography. Considering a surface mass balance between 18 and 25 cm w.e., a snow temperature at the ice-sheet surface between –25°C and –30°C and a geothermal heat flux between 110 and 240 Wm2, we have estimated that the melt rate at the lake is 7–20 mm a–1. Available IceSat data from the ice surface above the lake do not show any evidence of vertical movements, indicating that the lake must be stable at least in the order of tens of years. This lake possesses properties that together make it quite unique and comparatively advantageous for a deep exploration program.


Annual supraglacial lake drainage linked to plume formation at Helheim Glacier, southeast Greenland

Alistair Everett, Nick Selmes, Tim James, Tavi Murray, Ian Rutt, Dominic Reeve, Violetta Moloney, Harshinie Karunarathna

Corresponding author: Alistair Everett

Corresponding author e-mail: a.everett.743498@swansea.ac.uk

Supraglacial lake drainage events are common on the Greenland ice sheet. The majority occur on the west of the ice sheet, with only a small number in the southeast, where they are thought to have little impact on ice dynamics. We present data linking the drainage of a near-terminus supraglacial lake with the appearance of a plume of subglacial discharge at the front of Helheim Glacier, southeast Greenland. From a combination of Landsat and MODIS imagery we monitor the size of the lake and identify an annual pattern of growth during June followed by drainage in early July. A clearing in the ice melange, indicative of a point-source plume of subglacial discharge, is identified using Landsat and camera imagery at the ice front. The plume can typically be identified a few days after the lake drainage and can persist for up to 2 weeks. We also note the occurrence of a number of major calving events during this period, though a direct relationship is less clear. Air temperature data indicate that in most years the plume cannot be caused by runoff alone, though in 2012 we see evidence that the early onset of an unusually warm melt season may cause a plume that is independent of the lake drainage. Modelling of plume behaviour is compared to lidar data of the water surface elevation over the plume in order to estimate the discharge and configuration of the subglacial system required to maintain the opening in the melange. The linkage of a supraglacial lake drainage with the presence of a plume gives an important insight into the subglacial hydrological system of a tidewater glacier, allowing us to identify when we can be sure a channelized system is present and roughly estimate its discharge. Information such as this is vital for improving estimates of submarine melt rates, fjord circulation and the dynamics of tidewater glaciers.


Glacier dynamics near the calving front of Bowdoin Glacier, northwestern Greenland

Shin Sugiyama, Shun Tsutaki, Daiki Sakakibara, Jun Saito, Mihiro Maruyama, Naoki Katayama, Takanobu Sawagaki, Martin Funk, Andreaas Bauder

Corresponding author: Shin Sugiyama

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

The Greenland ice sheet is rapidly losing mass under the influence of retreat, thinning and acceleration of marine-terminating outlet glaciers. Rapid glacier changes cannot be attributed to climate change alone. Ice dynamics should play a key role, but lack of in situ data near the calving front hampers our understanding of the mechanism of the glacier changes. To better understand recent rapid recession of marine-terminating glaciers in Greenland, we performed satellite and field observations near the calving front of Bowdoin Glacier, a 3 km wide outlet glacier in northwestern Greenland. Studying glaciers in this region is particularly important because the rate of mass loss is increasing in northwestern Greenland. We analyzed satellite images to measure the changes in the glacier front position and ice speed since 1987. In summer 2013 and 2014, glacier and ocean bed geometry was surveyed in the vicinity of the calving front, using a ground-based ice radar and a sonic depth sounder mounted on a boat. Short-term ice speed variations were measured by running GPS on the glacier. We also performed hot-water drilling at approximately 2 km from the calving front. Satellite data showed a clear transition to a rapidly retreating phase in 2008 from a relatively stable glacier condition that lasted for more than 20 years. Ice near the calving front was grounded but very close to the floating condition, suggesting ice flotation as the driver of the recent retreat. GPS data showed complex short-term ice speed variations, which were correlated with air temperature, precipitation and ocean tides. When the drilling reached the bed of the 260 m thick glacier, borehole water level dropped to about 30 m above sea level. This observation confirmed the existence of a highly pressurized subglacial drainage system. Based on these in situ data collected near the glacier front, the mechanism of the recent rapid retreat of Bowdoin Glacier will be discussed.


Pathways of warm water to the 79°N Glacier

Janin Schaffer, Paul Dodd, Wilken-Jon von Appen, Torsten Kanzow, Christoph Mayer, Ursula Schauer

Corresponding author: Janin Schaffer

Corresponding author e-mail: janin.schaffer@awi.de

The ocean plays an important role in modulating the mass balance of the Greenland ice sheet by delivering heat to the marine-terminating outlet glaciers around Greenland. The warming and accumulation of Atlantic Water in the subpolar North Atlantic has been suggested to be the driver of the glaciers’ retreat around the coast of Greenland. The shelf regions thus play a critical role for the transport of Atlantic Water towards the glaciers. A key region for the mass balance of the Greenland ice sheet is the Northeast Greenland Ice Stream. This large ice stream drains the second-largest basin of the Greenland ice sheet and feeds three outlet glaciers. The largest one is Nioghalvfjerdsfjorden (79°N Glacier) featuring an 80 km long floating ice tongue. In order to study the relevant processes of glacier–ocean interaction we analyse historic and recent hydrographic observations in the Northeast Greenland shelf region. The complex bathymetry steers the flux of warm water of Atlantic origin from the open ocean onto the continental shelf and into the subglacial cavity of the 79°N Glacier. We show that the warmest water observed in the cavity of the 79°N Glacier origins from the southeastern entry via Norske Trough, where modified Atlantic Water recirculated in Fram Strait enters the shelf area. We found that these Atlantic Waters, both on the shelf and in the cavity, have become warmer by about 0.5°C during the last two decades. We propose that an increase in Atlantic Water temperatures in Fram Strait likely propagates onto the shelf and underneath the 79°N Glacier, where it may cause increased basal melt.


Bed radar reflectivity measurements at Institute Ice Stream and the ice divide between Minnesota, Pine Island and Rutford Glaciers in West Antarctica

Rodrigo Zamora, José Uribe, Andrés Rivera, Jonathan Oberreuter

Corresponding author: Rodrigo Zamora

Corresponding author e-mail: rzamora@cecs.cl

In January and December 2014, approximately 2200 km of radio-echo sounding data were collected over the Antarctic Plateau as part of a field campaign conducted by Centro de Estudios Científicos (CECs). The oversnow traverse covered areas of Union Glacier, Institute Ice Stream and the ice divide between Minnesota, Pine Island and Rutford Glaciers. Data were collected using a pulse-compression coherent radar sounder, working at a 155 MHz, yielding a maximum of 3300 m of ice thickness. Here we present a map of the spatial distribution of reflected power returned from the boundary between the ice and bedrock. Bed reflection power (BRP) analyses allowed for the characterization of bed roughness, basal conditions, subglacial drainage and the fast-moving and slow-moving portions of the ice sheet. Strong internal reflectors were also found in some areas, distinguished by internal reflection power (IRP) values, interpreted to be isochrones. Compared with Pine Island Glacier and Institute Ice Stream, the radar profiles collected reveal lower reflection powers for Rutford Glacier. It is therefore hypothesized that Rudford Glacier may have different basal conditions compared especially with Pine Island, which shows a high reflectivity at the bed. Overall, the data shown represent an important first-order measurement for understanding the ice dynamics close to the ice divides of this region.


Radar survey at the triple divide Institute Ice Stream, Pine Island and Minnesota Glaciers, West Antarctica

José Uribe, Rodrigo Zamora, Andrés Rivera, Jonathan Oberreuter

Corresponding author: José Uribe

Corresponding author e-mail: juribeparada@cecs.cl

Two oversnow radar surveys were performed at the Antarctic Plateau, West Antarctica, by Centro de Estudios Científicos (CECs) in January 2014 and December 2014. We surveyed a combined total distance of approximately 2200 km. The results of radar measurements at the multiple ice divide Institute Ice Stream, Pine Island and Minnesota Glaciers are presented here. The radar was composed of two parts: (1) a coherent 155 MHz pulse-compression radar, with 20 MHz of bandwidth and 200 W output power, for ice thickness detection; and (2) a fine-resolution frequency-modulated continuous wave (FMCW) radar, with a frequency range from 203 MHz to 1019 MHz, used for accumulation layer detection. The ice thickness at the divide is ~900 m, and the maximum ice thickness around the divide is ~2200 m. The fine-resolution radar detected approximately 150 m of shallow layers at the divide, where the stratigraphy is clearly visible. Both radar outputs were combined to produce one radargram for each profile, with a processing to enhance isochrone visualization. A shallow firn core at the divide and 12 stakes were used to determine the accumulation rate at this site. Additionally, accumulation rate and firn compaction derived from the fine-resolution radar was performed, using the data difference between the two oversnow traverses. The preliminary result indicates that the accumulation rate is ~0.23 m w.e. a–1 at the divide.


Seasonal water mass properties in the Amundsen Sea, Antarctica, using seal-borne tags

Helen Mallett, Karen Heywood, David Stevens, Lars Boehme, Mike Fedak

Corresponding author: Helen Mallett

Corresponding author e-mail: h.mallett@uea.ac.uk

Global attention is focused on the melting of the West Antarctic ice sheet (WAIS); however, the link between the ice-sheet mass loss and the hydrodynamics of the surrounding sea is poorly understood due to the difficulty in collecting data in this area, especially in the winter season. Pine Island Glacier (PIG) is the largest outlet for the WAIS, leading into the Amundsen Sea. There have been suggestions that the warm Circumpolar Deep Water (CDW) is increasingly crossing the continental shelf, contributing to increased ice mass loss. In order to better understand this link the iSTAR Ocean2Ice project equipped 14 southern elephant and Weddel seals in the Amundsen Sea with tags designed to collect salinity, temperature and depth information along with location and diving behaviour (CTD-SDRLs, SMRU, St Andrews). Between February and October 2014 more than 11 000 hydrographic profiles were collected, many times more than the previous sum total of profiles in the area. Previous data through the winter season are limited, coming from a few moorings of fixed location. The seal-tag profiles are spread throughout the Amundsen Sea basin, from the shelf break to the coast, and under the sea ice, offering a unique glimpse into the dynamics of the CDW. Using monthly maps of maximum subsurface potential temperature, the dataset reveals that the CDW enters the shelf sea primarily at the eastern trough. There is some indication of seasonal variation in location of maximum CDW temperature, and change in the CDW properties as it crosses the shelf sea. We also present maps of the depth of the subsurface potential temperature maximum, and similar plots for salinity. The outstanding dataset presented here provides a great tool for testing current theories as well as developing new ideas.


Seasonal and interannual evolution of meltwater production, storage and discharge from a tidewater glacier: insights from in situ and remotely sensed data

Eleanor Darlington, Richard Hodgkins, Adrian Jenkins

Corresponding author: Eleanor Darlington

Corresponding author e-mail: e.f.darlington@lboro.ac.uk

Tidewater glaciers are a significant drainage catchment of glacierized areas, directly transporting meltwater from the terrestrial to the marine environment. Meltwater delivery not only freshens fjord waters, but also plays a key role in glacimarine sedimentary processes. This has implications for sea-level rise, the marine ecosystem and glacier stability. Despite this, the temporal evolution of meltwater production, storage and release from tidewater glacier systems remains elusive. Sampling such systems for prolonged periods is expensive and comes with inherent safety risks. This study has utilized in situ sampling of total suspended sediment (TSS) and spectral reflectance to calibrate MODIS satellite images, addressing the seasonal and interannual variability of meltwater discharge. This has been achieved by quantifying the TSS and sediment plume size emerging from the tidewater glacier Kronebreen from 2002 to 2013. Freshwater entering at the head of Kongsfjorden from terrestrial sources is an order of magnitude greater than meltwater produced by direct submarine melting of Kronebreen’s ice face. This has been determined by a water mass mixing model which uses temperature and salinity from CTD sections. Sediment plume size is a suitable proxy to quantify meltwater delivery to Kongsfjorden and, as such, the seasonal characteristics of meltwater delivery have been identified. The relationship between plume formation and meteorological parameters has revealed the process of freshwater production and storage, both seasonally and interannually. Plume formation lags behind atmospheric temperature by over a week during June and July. This is reduced to 1–2 days in August and September, indicating that melt is no longer stored, and the drainage efficiency increases. Winter temperatures exert control over freshwater discharge in the following summer, with increased discharge after colder winters, highlighting the significance of winter storage. Sediment plume studies not only quantify fresh water emerging from tidewater glaciers, but also provide valuable insights to the supraglacial and englacial processes that control the timing and magnitude of discharge.


Coupled ice-sheet–sea-level modelling: on the stability of marine ice sheets and the interaction with relative sea-level change

Bas de Boer, Paolo Stocchi, Roderik S. W. van de Wal

Corresponding author: Bas de Boer

Corresponding author e-mail: b.deboer@uu.nl

Recent advances in numerical modelling of ice sheets with sea-level models have shed new light on the precise interaction of marine ice sheets with the change in near-field sea level and the related stability of the grounding line position (Gomez and others, 2013; de Boer and others, 2014). These studies with fully coupled ice-sheet–sea-level models have shown that including the full gravitational self-consistent sea-level change will act to slow down the retreat and advance of marine ice-sheet grounding lines. Here we present results for the past glacial cycle to the Holocene from a set of 3-D ice-sheet–shelf models coupled to a global sea-level model. The sea-level model incorporates all the glacial isostatic adjustment feedbacks for a Maxwell viscoelastic and rotating Earth model with variable coastlines. Ice volume is computed with four 3-D ice-sheet–shelf models for North America, Eurasia, Greenland and Antarctica (de Boer and others, 2014). To determine model uncertainties, we perform model simulations while perturbing model parameters that determine the deformation of the Earth, surface melting of the ice sheets and ice flow. We will present a first-order comparison of the ensemble of model realisations with relative sea-level reconstructions derived from paleo sea-level indicators over the globe and asses the stability of marine ice sheets related to different warming and cooling phases during the last glacial cycle.


Circulation of modified Circumpolar Deep Water and basal melt beneath the Amery Ice Shelf, East Antarctica

Laura Herraiz-Borreguero, Richard Coleman, Ian Allison, Stephen R. Rintoul, Mike Craven, Guy D. Williams

Corresponding author: Laura Herraiz-Borreguero

Corresponding author e-mail: lherraiz@nbi.ku.dk

Antarctic ice sheet mass loss has been linked to an increase in oceanic heat supply, which enhances basal melt and thinning of ice shelves. Here we detail the interaction of modified Circumpolar Deep Water (mCDW) with the Amery Ice Shelf, the largest ice shelf in East Antarctica, and provide the first estimates of basal melting due to mCDW. We use sub–ice-shelf ocean observations from a borehole site (AM02) situated ~70 km inshore of the ice-shelf front, together with open ocean observations in Prydz Bay. We find that mCDW transport into the cavity is about 0.22 ± 0.06 Sv (1Sv = 106 m3 s–1). The inflow of mCDW drives a net basal melt rate of up to 2 ± 0.5 m a–1 during 2001 (23.9 ± 6.52 Gt a–1 from under about 12 800 km2 of the northeastern flank of the ice shelf). The heat content flux by mCDW at AM02 shows high intra-annual variability (up to 40%). Our results suggest two main modes of sub–ice-shelf circulation and basal melt regimes: (1) the ‘ice pump’/high salinity shelf water circulation, on the western flank; and (2) the mCDW meltwater–driven circulation in conjunction with the ‘ice pump’, on the eastern flank. These results highlight the sensitivity of the Amery’s basal melting to changes in mCDW inflow. Improved understanding of such ice-shelf–ocean interaction is crucial to refining projections of mass loss and associated sea-level rise.


Pine Island Glacier retreat through the asynchronous coupling of an ice and ocean model

Jan De Rydt

Corresponding author: Jan De Rydt

Corresponding author e-mail: janryd69@bas.ac.uk

Recent observations and modelling work have shown a complex mechanical coupling between Antarctica’s floating ice shelves and the adjacent grounded ice sheet. Changes to atmospheric and ocean conditions that affect the ice shelf can therefore lead to feedback mechanisms that affect the mass balance of the ice sheet at sub-decadal timescales. For example, Pine Island Glacier currently has a strongly negative mass balance caused by warming ocean conditions that have led to enhanced under-ice melting, weakening of the ice shelf, and subsequent speed-up of the glacier. In order to study ice–ocean feedback mechanisms through physical models, a coupled approach is needed, which relates changes in ocean properties to changes in ice-shelf geometry and ice-sheet dynamics. The full coupling of an ice dynamics model with an ocean model is a difficult problem, though a possible way forward is provided by an asynchronous relationship, where both model components are run forward in time with different time steps. We present preliminary results on the asynchronous coupling between a 2-D shallow-ice model (Ua) with grounding line resolving capabilities, and a 3-D ocean gcm (MITgcm) with a static implementation of the ice shelf. A series of idealized experiments are performed to investigate the retreat of Pine Island Glacier from a subglacial ridge, and to better understand the sensitivity of this retreat to observed changes in the ambient ocean.


Resolving flow and deformation of Store Glacier, West Greenland, using phase-sensitive FMCW radar

Tun Jan Young, Poul Christoffersen, Keith Nicholls, Lai Bun Lok, Paul Brennan, Bryn Hubbard, Alun Hubbard, Samuel Doyle, Marion Bougamont, Craig Stewart, Joe Todd

Corresponding author: Tun Jan Young

Corresponding author e-mail: tjy22@cam.ac.uk

Studies have revealed that the dynamics of glacier flow are influenced by basal lubrication and motion at the ice–bedrock interface, and by internal deformation within the ice column. In Greenland, the majority of glacier flow can be characterized by enhanced basal motion, and the associated ice-sheet dynamics are responsible for up to half of its contribution towards sea-level rise. The processes and mechanisms involved in basal motion are, however, still poorly understood. Furthermore, there is a paucity of data to describe how the penetration of surface meltwater influences basal motion on diurnal and seasonal timescales. To study this process, we installed a phase-sensitive frequency-modulated continuous-wave (FMCW) radar array on Store Glacier in West Greenland to obtain a continuous time series of changes through the ice column at temporal resolution of 1 hour. We present results that show dynamic strain throughout the vertical profile over a 3 month period that includes the 2014 ablation season. When strain is resolved, we obtain a record of basal motion, which shows rapid changes occurring around the ice–bed interface. The results from this project significantly improve our understanding of subglacial and englacial environments and processes, and provide first results of radar-measured strain in a fast-flowing glacier.


Modelling the response of the Larsen B glaciers after the ice-shelf collapse

Jan De Rydt, Hilmar Gudmundsson, Helmut Rott, Jonathan Bamber

Corresponding author: Jan De Rydt

Corresponding author e-mail: janryd69@bas.ac.uk

Following the rapid disintegration of the Larsen B ice shelf, Antarctic Peninsula, in 2002, regular surveillance of its ~20 tributary glaciers has revealed a response that is varied and complex in both space and time. The major outlets have accelerated and thinned, smaller glaciers have shown little or no change, and glaciers flowing into the remnant SCAR Inlet Ice Shelf have responded with delay. We present the first area-wide numerical analysis of glacier dynamics before and after the collapse of the ice shelf, combining new datasets and a state-of-the art numerical ice flow model. We revisit the role of the compressive arch for the stability of the ice shelf, and simulate the loss of buttressing at the grounding line as the ice shelf progressively disintegrates. A good agreement is found between modelled changes in glacier flow and observations in response to the loss of buttressing. Through this study, we seek to improve confidence in our numerical models, and infer important information about their ability to capture the complex mechanical coupling between floating ice shelves and grounded ice.


External forcing modulates Pine Island Glacier flow

Knut Christianson, David Holland, Pierre Dutrieux, Ian Joughin, Mitch Bushuk, Byron Parizek, Sridhar Anandakrishnan, Richard Alley, Adrian Jenkins, Karen Heywood, Tim Stanton, Atsuhiro Muto

Corresponding author: Knut Christianson

Corresponding author e-mail: knut@uw.edu

Nearly 50 years ago, Mercer first suggested the Eemian sea-level highstand was a result of a collapse of the marine portions of the West Antarctic ice sheet. Since this time, special attention has been paid to West Antarctica’s Amundsen Sea Embayment due to its steeply sloping retrograde beds that are well below sea level, and observations of rapid grounding-line retreat, high ice-shelf basal melt rates, and basin-wide glacier thinning and acceleration here. Despite this focus, accurate assessments of the past and future behavior of this embayment remain elusive, due to a lack of understanding of glacier response to ocean forcing. Here we present a continuous 2 year (2012–2014) time series of oceanographic, borehole, glaciological and seismological observations of Pine Island Glacier, its sub-ice ocean cavity, and the adjacent Amundsen Sea. With these data, we captured the glacier’s response to the largest fluctuation in temperature of water entering the sub-ice ocean cavity in the last 35 years. Initially, as the water cools by ~0.5°C, the glacier slows by 5% with a 3 month lag to the ocean temperature; then, as the ocean warms by 1°C, the glacier nearly recovers its earlier speed by the end of 2014. The smooth glacier flow response to ocean forcing is punctuated by rapid (2–3 week), high-amplitude (~2.5% of the glacier’s speed) speedups and slowdowns that do not correspond with changes in ocean temperature. Satellite and seismological observations indicate that speedups are caused by reduction of lateral drag along the ice stream’s shear margins as a large iceberg calves and that slowdowns are most likely due to ephemeral regrounding on bed highs. Our observations indicate that a glacier can rapidly respond to external forcings in ways that may mimic the response expected from the marine ice-sheet instability. Thus, long-term, multifaceted investigations are necessary to determine whether a glacier is responding to external forcing or undergoing a terminal collapse.


Observable consequences of calving laws

Martin O’Leary

Corresponding author: Martin O’Leary

Corresponding author e-mail: m.e.oleary@swansea.ac.uk

When modelling tidewater glaciers, one of the least well constrained aspects is the choice of frontal boundary condition, or ‘calving law’. Ideally we would like to relate this choice to observations, such as the many available time series of glacier front positions, typically over timescales of a few years to decades. Unfortunately it is often unclear how to relate these observations to the choice of calving law, and so the choice is often made for reasons of convenience, rather than accuracy. Here I show how we can relate the locations of stable calving front positions to properties of the calving law. By simulating ensembles of randomized tidewater glaciers, I demonstrate that calving laws divide into two classes, based on their stability properties. I also show how this distinction can be mapped to the real-world difference between melt-driven and internally paced calving. I also show that the existence of stable calving front positions is in itself a strong constraint on the possible calving laws. I demonstrate how many reasonable-seeming calving laws are incapable of producing stable calving fronts on long timescales.


Modelling the steady-state crystal size of deforming ice

Felix Ng, T.H. Jacka

Corresponding author: Felix Ng

Corresponding author e-mail: f.ng@sheffield.ac.uk

Knowledge of how the grain size of deforming polycrystalline ice evolves can shed light on the recrystallization processes that influence ice rheology. The distribution of crystal size across the ice sheets must also reflect their flow histories, although data are presently limited to the few places where ice cores have been drilled. In a set of laboratory experiments by Jacka and Li (1994), the mean grain size D of ice deforming under uniaxial compression to the tertiary creep stage was found to reach steady values (D_eq) that depend on applied stress (tau) in a power law: D_eq^2 proportional to tau^-m, where m ≈ 1.5. Here we study this phenomenon by building a differential-equation model that describes the coupled evolution of D and mean dislocation density rho. Processes represented in the model include grain flattening under compression, grain-size reduction due to rotation recrystallization (polygonization), dislocation generation by the deformation and dislocation removal by grain-boundary motion and polygonization, grain growth driven by grain-boundary energy, and migration recrystallization induced by stored strain energy. Our description of the last process remains tentative, but we consider plausible formulations for it in this homogeneous model where the population distribution of grain sizes is ignored. The model predicts D and rho to evolve toward equilibrium values at large time, with m = 4/3 in the D_eq vs stress relationship, close to the laboratory value. Furthermore, the corresponding predicted steady rheology (strain rate proportional to tau^n) has n in the range from 8/3 to 4, depending on how significant are the strain-induced effects. These findings may aid efforts to seek a generalized model of ice rheology incorporating texture and fabric evolution.


Feedbacks of glacial water and primary production on the carbonate system and ocean acidification state in the Djimphna Sound, northeast Greenland

Agneta Fransson, Melissa Chierici, Paul Dodd, Mats Granskog, Colin Stedmon, Edmond Hansen

Corresponding author: Agneta Fransson

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

We investigated the processes affecting the carbonate system and the ocean acidification (OA) state (i.e. pH and calcium carbonate saturation,) in summer 2012 in the Dijmphna Sound and Nioghalvfjerdsfjorden (79°N). This fjord system is located in one of the main outlets of the large Northeast Greenland Ice Stream, which feeds meltwater from the Greenland ice sheet (GrIS) to the fjord and adjacent east Greenland shelf. The tracer relationships between salinity, total alkalinity (AT) and oxygen isotopic ratios (δ18O) indicate three layers in the fjord system and the adjacent shelf; a fresh surface water, cold waters originating from the Arctic (Polar water) and warm and salty water from modified Atlantic water. We found that biological CO2 uptake and freshwater addition were major drivers for the variability of the carbonate system and ocean acidification state in the surface waters. In the upper 10 m, freshening due to glacial water contributed with decreased Ar between 0.2 and 0.4, whereas CO2 uptake due to phytoplankton production resulted in an increase of about 0.35 extending from the glacier front to the central part of the fjord system. The increase in  due to biological CO2 uptake thus alleviates/mitigates some of the  decrease due to increased freshwater supply, which has implications for the effect of the observed freshening of the Arctic Ocean and changes in the processes affecting Arctic biological primary production.


An overview of the ISMIP6 effort

Sophie Nowicki, Tony Payne, Eric Larour, Ayako Abe-Ouchi, Heiko Goelzer, Jonathan Gregory, William Lipscomb, Andrew Shepherd, Helene Seroussi

Corresponding author: Sophie Nowicki

Corresponding author e-mail: sophie.nowicki@nasa.gov

We present the framework for a new effort, ISMIP6, the Ice Sheet Model Intercomparison Project for CMIP6. The primary goal of ISMIP6 is to improve projections of sea-level rise by focusing on the evolution of the Greenland and Antarctic ice sheets under a changing climate, along with a quantification of associated uncertainties (including uncertainty in both climate forcing and ice-sheet response). This goal requires an evaluation of AOGCM climate over and surrounding the ice sheets; analysis of simulated ice-sheet response from standalone models forced ‘offline’ with CMIP AOGCM outputs and, where possible, with coupled ice-sheet–AOGCM models; and experiments with standalone ice-sheet models targeted at exploring the uncertainty associated with ice-sheet physics, dynamics and numerical implementation. A secondary goal is to investigate the role of feedbacks between ice sheets and climate in order to gain insight into the impact of increased mass loss from the ice sheets on regional and global sea level, and of the implied ocean freshening on the coupled ocean–atmosphere circulation. These goals map into both Cryosphere and Sea-Level Rise Grand Challenges relevant to Climate and Cryosphere (CliC) and the World Climate Research Program (WCRP).


Seismic evidence of widespread firn compaction on the Larsen C ice shelf, Antarctic Peninsula

Bernd Kulessa, Alex Brisbourne, Peter Kuipers Munneke, Ed King, Suzanne Bevan, Adrian Luckman, Bryn Hubbard, David Ashmore, Paul Holland

Corresponding author: Bernd Kulessa

Corresponding author e-mail: b.kulessa@swansea.ac.uk

Rising mean annual surface temperatures have been causing firn layers on Antarctic Peninsula ice shelves to compact, a process that is strongly implicated in ice-shelf disintegration. Firn compaction warms the ice column and, given sufficiently wet and compacted firn layers, allows meltwater to penetrate into surface crevasses and enhance the potential for hydrofracture. Indeed, the firn air content of the Larsen B ice shelf was strongly reduced when a critical compressive arch in its stress regime was breached, and meltwater-driven fracture expedited the ensuing rapid ice-shelf disintegration. Larsen B’s large southern neighbour, Larsen C, is now a considerable cause for concern because a compacting firn layer has been inferred from airborne radar and satellite data, with strongly reduced air contents in Larsen C’s north and northwest. Here we present direct evidence for widespread firn compaction on the Larsen C ice shelf using 15 seismic refraction datasets recorded in four austral summers and covering most ice-shelf sectors. All but the extreme northwesterly refraction travel-time dataset, recorded in a surface depression in Cabinet Inlet, were converted to density–depth profiles using a well-publicized method that assumes a monotonic increase of snow and firn density with depth. The Cabinet Inlet depression dataset violated this assumption, and is instead characterized by a ~2.7 m thick surface layer with a mean compressional wave velocity of ~1545 m s–1, overlying a medium with a mean velocity of ~3685 m s–1. The surface layer is interpreted as snow and firn, and the seismic velocity of the underlying medium is consistent with refrozen meltwater, as imaged by a concurrently acquired borehole optical televiewer (OPTV) borehole log. Taken together the 15 seismic refraction datasets substantiate the presence of a northwesterly to southeasterly gradient in firn compaction, characterized by minimum firn air contents in Cabinet Inlet and relatively uncompacted firn in the southeast. We compare these seismically derived firn-density profiles with previous estimates of firn-air content and new results from a one-dimensional firn model for the Larsen C ice shelf that includes melt percolation and refreezing. In the future the NERC MIDAS project will acquire an additional seismic refraction transect along a central flowline on the Larsen C ice shelf, and constrain the firn model with density profiles from seismic, borehole OPTV and snow-pit measurements.


External forcing and geometric controls on surface elevation changes of central-west Greenland outlet glaciers

Ginny Catania, Timothy Bartholomaus, Dustin Carroll, David Sutherland, Emily Shroyer, Niels Korsgaard, Kurt Kjær, Michiel van den Broeke, Leigh Stearns, Jonathan Nash, Ryan Walker, Denis Felikson

Corresponding author: Ginny Catania

Corresponding author e-mail: denis.felikson@utexas.edu

Along the central-west coast of the Greenland ice sheet (GIS), regional ice loss has been accelerating since around 2007 (Khan and others, 2010). However, the changes in glacier velocity are not uniform across the region and we do not understand the physical processes controlling the loss here. To add to the accounting of ice loss, we examine ~30 year elevation differences of outlet glaciers using digital elevation models (DEMs) from present-day WorldView-1 and -2 stereo imagery and a 1985 aerial photo survey (Kjær and others, 2012). Large spatial variability in both surface elevation and terminus position changes is observed. The observed variability cannot be explained by changes in surface mass balance from the Regional Atmospheric Climate Model (RACMO2.3) alone. We examine two additional sources of external forcing on the ocean-terminating glaciers and their effects on glacier behavior: effective pressure at the glacier bed and terminus melt. Effective pressure is calculated from the DEMs and a combination of a Greenland bed dataset and available bathymetry. Terminus melt caused by turbulent flow of buoyant subglacial discharge at the glacier terminus face is inferred from a plume model driven by runoff from RACMO2.3. Additionally, we examine the role that geometric controls play in glacier response. Glacier thickness and surface slopes govern whether surface perturbations will diffuse upstream or advect downstream and the relative importance of advection to diffusion can be quantified by the Peclet number. We observe strong thinning where the Peclet number is small, indicating that thinning may be diffusing upstream from the terminus.


Temporal changes in basal conditions of Pine Island Glacier, West Antarctica, from repeat radar surveys

Damon Davies, Robert G. Bingham, Edward C. King, David G. Vaughan, Andy M. Smith, Alex M. Brisbourne

Corresponding author: Damon Davies

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

Remote-sensing observations over the last two decades have revealed rapid acceleration, thinning and grounding line retreat of ice streams draining the Amundsen Sea Sector of the West Antarctic ice sheet (WAIS). This has renewed concern over the long-term future of the ice sheet and its potential contribution to global sea-level rise. Pine Island Glacier (PIG) is currently exhibiting the greatest mass loss of any ice stream and recent numerical models have suggested it may be in a state of irreversible retreat. However such models do not account for the influence of local, temporal changes in basal conditions on ice dynamics. Recent studies in West Antarctica, including on Pine Island Glacier itself, have revealed rapid erosion rates on the order of 1 m a–1, indicating mobilization and removal of sediment from a soft bed. During the 2013/14 iSTAR traverse of Pine Island Glacier we undertook dedicated high-resolution ‘repeat radar surveys’, specifically acquiring radar data along traverses that were first radar-profiled in 2007/08 (or, for one traverse, 2010/11). Using these data we aim to (1) assess changes in morphology of the glacier bed over a 6 year interval and (2) make quantitative assessments of basal erosion rates. These data offer a unique opportunity to image the temporal evolution of the ice/bed interface beneath a fast-flowing Antarctic ice stream with the ability to test pre-existing hypotheses of subglacial processes.


Controls on the Sabrina Coast grounding line, East Antarctica

Jamin Greenbaum, Donald Blankenship, Duncan Young, Jason Roberts, Tas van Ommen, Martin Siegert

Corresponding author: Jamin Greenbaum

Corresponding author e-mail: jamin@utexas.edu

Totten Glacier is the primary outlet of the Aurora Subglacial Basin (ASB), draining 3.5 m of eustatic sea-level potential into the Sabrina coast alongside the Moscow University Ice Shelf that fringes the coastline. 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 in the 400–500 m depth range. Entrances to the cavity deeper than this range of thermocline depths indicate that the TGIS is vulnerable to intrusions of MCDW. We delineate a region of marine ice-sheet instability landward of the modern grounding line where MCDW may be available. This zone is bounded to the south by ice grounded near the modern grounding line depth, beyond which the rest of the ASB descends to more than 1.7 km below sea level. We use recent aerogeophysical data to delineate areas of the grounding line along the Sabrina coast where satellite-derived datasets disagree and present the result in the context of the zone of local marine ice-sheet instability, providing the latest subglacial and submarine controls on the most rapidly changing coastline of East Antarctica.


Observed propagation of a large rift in the Larsen C ice shelf: rift development and possible consequences for the stability of the ice shelf

Daniela Jansen, Adrian J. Luckman, Alison Cook, Suzanne Bevan, Bernd Kulessa, Bryn Hubbard, Martin O’Leary, Paul R. Holland

Corresponding author: Daniela Jansen

Corresponding author e-mail: daniela.jansen@awi.de

The Larsen C ice shelf is the most northerly of the remaining major Antarctic Peninsula ice shelves and is vulnerable to changes in both ocean and atmospheric forcing. It is the largest ice shelf in the region and its loss would lead to a significant drawdown of ice from the Antarctic Peninsula ice sheet. There have been observations of widespread thinning, melt ponding in the northern inlets, and a speed-up in ice flow, processes that have all been linked to former ice-shelf collapses. Previous studies have also highlighted the vulnerability of the Larsen C ice shelf to specific potential changes in its geometry, including a retreat from the Bawden and Gipps Ice Rise. Rift tips in the vicinity of Gipps Ice Rise have been observed to align at a suture zone between two flow units within the shelf. Several studies have provided evidence for marine ice in these suture zones, which has been found to act as a weak coupling between flow units with different flow velocities. It has been concluded that this ice inhibits the propagation of rifts because it can accommodate strain in the ice without fracturing further. In a change from the usual pattern, a northwards-propagating rift from Gipps Ice Rise has recently penetrated through the suture zone and is now more than halfway towards calving a large section of the ice shelf. The rate of propagation of this rift accelerated during 2014. When the next major calving event occurs, the Larsen C ice shelf is likely to lose around 10% of its area to reach a new minimum area for the ice shelf. We followed the rift propagation on MODIS and Landsat imagery and used a numerical model to investigate the influence of the future calving event on ice-shelf stability. We find that the ice front is at risk of becoming unstable when the anticipated calving event occurs.


On calving rates and the determination of critical stress states at the ice front vicinity of Antarctic ice shelves

Julia Christmann, Ralf Müller, Angelika Humbert

Corresponding author: Julia Christmann

Corresponding author e-mail: jchristm@rhrk.uni-kl.de

In this study the rate and position of small-scale calving are investigated for Antarctic ice shelves using viscous and different viscoelastic material models. The aim of this study is to discuss the applicability of different material models to get a physically based calving criterion. If the position of calving and the time during calving events are known, a computation of the calving rate, including the knowledge of the flow velocity, is possible. The concept of reaching a critical principal stress as a calving criterion is discussed and the influence of different model parameters is analyzed. The ice shelf is modeled as a two-dimensional body loaded by gravity and water pressure. There are two responses of ice to loading: on long timescales ice reacts as a viscous fluid, and on short timescales as an elastic solid. Therefore, the stress response of different material models is compared for the same geometric setup and the application to ice-shelf calving is verified. The analysis of the influence of the most crucial parameters on the maximum principal stress and the position of this stress shows the importance of the freeboard, the stress-free part at the ice front. Thereby, the thickness of the freeboard is highly dependent on the ice-shelf thickness, the salt water density, and the ice density. In contrast, the time between two calving events is only influenced by the material parameters and therewith the characteristic time of the material model. Further investigations consider the kinematic relations of the computation, as small-scale calving happens on rather continuous timescales of several days to a few years. During these time intervals, the ice shelves could spread out more than hundreds of meters in flow direction inducing strains, which cannot be assumed to be small and do not fulfil the kinematics for small strains. Therefore, a comparison of the stress states in the vicinity of the ice front using small or large deformation is performed and the influence on the maximum stress value and its position is discussed.


Flow variation of Bindschadler and MacAyeal Ice Streams at decadal timescales

Christina Hulbe, Ted Scambos, Marin Klinger, Mark Fahnestock

Corresponding author: Christina Hulbe

Corresponding author e-mail: christina.hulbe@otago.ac.nz

Ice streams on the Ross Sea side of the West Antarctic ice sheet are known to experience flow variability on hourly timescales (tide influence near the grounding line), on annual timescales (basal water system variability) and on multi-century timescales (stagnation and reactivation linked to internal thermomechanical feedbacks). We report here on observations that fill in the missing timescale, flow variability at the decadal scale, on the Bindschadler and MacAyeal Ice Streams (formerly D and E, respectively). The analysis makes use of new and archived ice velocity from Landsat imagery. Both streams speed up and slow down at rates of a few meters per year, which are close to the detection limit following the errors in the various observations. Patterns of flow variation are guided by local geographic features and changes to surface crevasses suggest that the influence of sticky spots at the downstream ends of the ice streams is relatively recent.


Observations and modelling of centimetre-scale folding in the NEEM ice core

Daniela Jansen, Maria-Gema LLorens, Julien Westhoff, Florian Steinbach, Paul D. Bons, Albert Griera, Ilka Weikusat, Sepp Kipfstuhl

Corresponding author: Daniela Jansen

Corresponding author e-mail: daniela.jansen@awi.de

Disturbances on the centimeter scale in the layering of the NEEM ice core (North Greenland) can be mapped by means of visual stratigraphy as long as the ice has a visual layering, such as, for example, cloudy bands. Different focal depths of the visual stratigraphy method allow, to a certain extent, a three-dimensional view of the structures. In this study we present a structural analysis of the visible folds, discuss characteristics and frequency and present examples of typical fold structures. The structures evolve from gentle waves at about 1500 m to overturned z-folds with increasing depth. Occasionally, the folding causes significant thickening of layers. Their shape indicates that they are passive features and are probably not initiated by rheology differences between layers. Layering is heavily disturbed and tracing of single layers is no longer possible below a depth of 2160 m. Lattice orientation distributions for the corresponding core sections were analyzed where available in addition to visual stratigraphy. The data show axial-plane parallel strings of grains with c-axis orientations that deviate from that of the matrix, which shows a well developed single-maximum fabric at the depth where the folding occurs. We conclude from these data that folding is a consequence of deformation along localized shear planes and kink bands. The findings are compared with results from other deep ice cores. The observations presented are supported by microstructural modeling. We are using a crystal plasticity code that reproduces deformation, applying a fast Fourier transform method (FFT), coupled with the microstructural platform ELLE to include dynamic recrystallization processes. The model results reproduce the development of bands of grains with a tilted orientation relative to the single maximum fabric of the matrix and also the associated local deformation.


A model study of the mass balance and firn structure of ice shelves in the Antarctic Peninsula

Melchior van Wessem, Willem Jan van de Berg, Michiel van den Broeke, Elizabeth Morris, John Turner, Carleen Reijmer

Corresponding author: Melchior van Wessem

Corresponding author e-mail: jmvanwessem@gmail.com

Following the subsequent disintegration of Antarctic Peninsula (AP) ice shelves (e.g. Larsen A, Larsen B and (parts of) Wilkins ice shelf), various studies have investigated the role of atmospheric forcing in these collapses. Meltwater ponding and subsequent hydrofracturing of surface crevasses is believed to be an important process for the disintegration of AP ice shelves. The process of meltwater ponding starts when no pore space remains in the firn, which occurs when, over successive years, enough meltwater has percolated into the firn, refreezing, heating and saturating it. Snowfall counteracts this process by adding pore space to the firn pack. Because of the scarcity of observations, modeling of these processes is important to gain an improved understanding of the past, current and future state of AP ice shelves. Here we present the first attempt to relate the temporal and spatial evolution of several modelled key variables, such as air temperature, surface mass balance (SMB) components snowmelt, snowfall, meltwater percolation and runoff, with ice-shelf viability in the AP region. We use a high-resolution (5.5 km) Regional Climate Model (RACMO2.3) coupled to a Firn Densification Model (FDM) to simulate the SMB, with a focus on meltwater (runoff) fluxes, over the AP for the last 35 years (1979–2014). We evaluate model results with available observational data, such as measurements of SMB, melt, firn air content and snow/air temperature. Results show that the model correctly represents observed trends in air temperature. The firn model suggests that melt strongly influences AP firn-pack conditions over large regions. From these results it becomes clear that predictions of ice-shelf viability, which are usually based on surface and air temperature, should in fact be based on the actual firn-pack conditions.


Multichannel acoustic backscattering of frazil ice: a case study in McMurdo Sound, Antarctica

Akos Kungl, Daniel Schumayer, Patricia Langhorne, Gregory Leonard

Corresponding author: Akos Kungl

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

Frazil ice crystals have been observed floating freely in rivers, lakes, seas and oceans in the cryospheric regions of both hemispheres. In Antarctica their formation is commonly attributed to the presence of ice-shelf water (ISW), which is produced when dense shelf water is modified at depth through interaction with an ice shelf. The resulting water mass is buoyant due to an injection of meltwater and potentially supercooled as it has lost heat in the melting process. The ISW then ascends the basal slope of an ice shelf where it becomes supercooled in situ as its pressure-dependent freezing point increases, and hence is capable of sustaining a population of frazil ice crystals. When these plumes of ISW reach the ice-shelf edge they can deliver frazil crystals beneath the sea ice, which grows thicker than it would otherwise. Although there are existing casual observations and proxy measurements in McMurdo Sound that confirm the presence of frazil crystals beneath a sea-ice cover, quantification of the crystals’ size distribution and number density proves to be a difficult task due to the challenging nature of the environment. Quantitative observations of frazil crystals generally use acoustic backscattering techniques that represent crystals as equivalent spheres. Here we provide a more realistic theoretical model that treats the crystals as oblate spheroids and present a novel method for evaluating acoustic backscatter data. Within this oblate spheroidal model we show that the equivalent sphere approach may consistently overestimate the scattering cross section of the frazil crystals, potentially leading to biased estimates of crystal size. The method is assessed by applying it to data collected from a four-frequency, active sonar instrument termed an Acoustic Water Column Profiler (AWCP), which was deployed in McMurdo Sound in 2012. The AWCP is manufactured by ASL Environmental Sciences and the particular instrument used in this study operates at frequencies of 125, 200, 455 and 774 kHz. This frequency range covers the Rayleigh regime for the expected size distribution of frazil crystals and is predicted to be capable of estimating the number density and size distribution of the crystals. We believe that the oblate spheroidal model, via the rescaled backscattering cross section, can provide an alternative data analysis procedure to the equivalent sphere approach as the model simultaneously captures the Rayleigh and geometric scattering regimes.


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

Ralph Timmermann, Janin Schaffer

Corresponding author: Ralph Timmermann

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

The RTopo-1 dataset of Antarctic ice-sheet/shelf geometry and global ocean bathymetry has proven useful not only for modelling studies of ice–ocean interaction in the Southern Hemisphere. Following the spirit of this dataset, we introduce a new product (RTopo-2) which provides consistent maps of not only Antarctic but also Greenland ice-sheet/shelf topography along with global ocean bathymetry/land surface topography on a spherical grid with now 0.5 min resolution. The backbone datasets of RTopo-2 were derived from the GEBCO, IBCAO, IBCSO and Bedmap-2 products. We also considered multibeam survey data for the bathymetry on the continental shelf off northeast Greenland as well as high-resolution gridded data for upper and lower ice surface topography and cavity geometry of Nioghalvfjerdsfjorden Glacier which go back to seismic surveys and to data from the IceBridge campaign. We show the improvements achieved in RTopo-2 for the Greenland fjord systems in comparison to bathymetry products commonly used for numerical simulations so far. However we also demonstrate that using Bedmap-2 sub-ice cavity bathymetry does not universally yield an improvement over previous datasets. We aim at restoring these bits of information before the dataset is released.


Modelling ice streams in the northern Antarctic Peninsula

Sam Royston, Hilmar Gudmundsson

Corresponding author: Sam Royston

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

Glaciers on the Antarctic Peninsula (AP) have shown a varied response to recent climatic change. Most commonly, AP tidewater glaciers have retreated at the calving front and their flux across the grounding line has accelerated, increasing the contribution to sea-level rise. We utilize new datasets and tools that exploit remote-sensing data from before distinct periods of ice-shelf collapse in the northern AP to set-up the model domain and parameterize a higher-order vertically integrated ice-stream model. The model is run diagnostically to quantify the flux of ice across the grounding line. We investigate the ice flux and dynamics in the short satellite-era period before ice-shelf collapse, noting the ice shelves in this region have been thinning and retreating for a number of decades. We then aim to quantify how well the model replicates the changes in ice dynamics immediately following an ice-shelf collapse, by comparing the diagnostic results from a model with the ice shelf removed against observed ice surface velocities.


Pre-disintegration precursors to the Larsen ice shelf disintegrations: the climate–ocean conspiracy

Ted Scambos, Mattias Cape, Bruce Huber, Marin Klinger, Terry Haran

Corresponding author: Ted Scambos

Corresponding author e-mail: teds@nsidc.org

Precursor changes in the Larsen A and B ice shelves beginning nearly a decade before the first disintegration event in 1995 have been identified in satellite imagery, and coincide with a trend towards reduced sea-ice cover in the far northwestern Weddell Sea and increased föhn winds on the eastern Antarctic Peninsula. This is linked to climate changes based on increased frequency of positive Southern Annular Mode conditions. Prior to this time, available satellite data suggest steady-state evolution of the ice-shelf areas. Examination of satellite imagery of the Larsen A and B ice shelves spanning 1963 to the present shows that shear margins and suture zone features were essentially unchanged through 1986. Following that time, but well prior to the break-up events, ice-shelf shear zones show increased and expanded areas of rifting, concentration of shear, and ice flow speed increases. These early changes suggest either increased ocean-driven basal melt or effects of increased surface meltwater on grounded glacier outflow are the cause of early shelf weakening that led to disintegration. The reduced sea-ice extent in the 1990s–early 2000s in the region suggests that wind traction on the sea ice and ocean surface may have influenced both surface layer movement (outward from the shelf cavity) and increased sub-shelf ocean circulation. Ocean temperature and salinity profiles show that modified Weddell Deep Water, a highly modified version of Circumpolar Deep Water, is present near the ice-shelf fronts in many but not all profiles during 1995–2012. This mid-ocean layer, mWDW, has potential temperatures 0.1–0.2°C above freezing in this region. This may have been sufficient to initiate basal thinning of the shelves, and especially the ice-shelf margins, leading to reduced internal compressive stress and an increased susceptibility to hydrofracture. The same wind events that produce the shift in surface and subsurface circulation also act to increase surface melt on the ice shelves, and so create a nefarious conspiracy.


Committed near-future retreat of Smith, Pope and Kohler Glaciers inferred from transient model calibration

Daniel Goldberg, Patrick Heimbach

Corresponding author: Daniel Goldberg

Corresponding author e-mail: dngoldberg@gmail.com

A glacial flow model is used to investigate near-future thinning and and grounding line retreat of Pope, Smith, and Kohler Glaciers, West Antarctica. The model is calibrated against observations to infer unknown parameters. We investigate two methods of calibration: the more commonly used ‘snapshot calibration’, which does not consider time dependence and assumes all observations are contemporaneous; and ‘transient calibration’, which accounts for the transient nature of observations. The transiently calibrated model achieves good agreement with time-dependent observations of surface elevation and velocity from 2001 to 2011, while snapshot calibration is unable to reproduce transient observed behaviour – although the poor fit of the snapshot-calibrated model is not apparent when examining an areally integrated metric such as total 2001–2011 sea-level contribution from the region. The models are then run from 2011 to 2041 with no additional forcings. The transiently calibrated model predicts near-steady grounded ice loss of 22.5 km3 a–1 over this period, while the snapshot-calibrated figure, while still large, is nearly 50% less. Moreover the transiently calibrated model reproduces past grounding line retreat and predicts further retreat, while the snapshot-calibrated model does neither. This demonstrates the need for ice models to be able to reproduce time-dependent observational histories in order to make more accurate predictions of near-future ice-stream behavior. Still, these results – along with additional sensitivity studies – suggest that this region will continue to have significant sea-level contributions over the next several decades, regardless of external forcings or uncertainties in unknown parameters.


Heat transfer regimes for the buoyant flow of meltwater next to ice shelves and submerged glacier termini

Andrew Wells, Grae Worster

Corresponding author: Andrew Wells

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

The melting of floating ice shelves and glacier termini submerged in a warm, saline ocean can contribute significantly to glacial mass balance and impact marine ice-sheet dynamics. When background ocean currents are relatively weak, the buoyancy-driven flow of fresh meltwater rising along the ice face provides a key control on melting rates. Model predictions of melting rates typically rely on a description of heat and salt transfer to the ice–ocean interface across a turbulent ocean boundary layer. Some commonly used parameterizations are based on theory and experiments for shear-driven boundary layers. However, whilst the parameterizations are applied on geophysical scales, they do not capture the behaviour observed in certain laboratory experiments of buoyancy-driven flows generated at melting ice faces. Building on recent experimental observations, we present results from a theoretical analysis of the turbulent boundary layer dynamics for buoyancy-driven flow of fresh meltwater released by ablation of a locally planar ice face in a saline ocean. Our model extends previous work on thermal transport, by accounting for the impact of salt transport and freshwater fluxes at the ice–ocean interface. The analysis reveals that the viscous sublayer close to the ice–ocean interface plays a key role in moderating heat and salt transfer from the ocean, and different dynamical regimes are observed depending on whether this viscous sublayer is driven by buoyancy, or driven by shear stresses. We identify different parameterizations for the heat and salt fluxes in each regime, which produce different scalings for the ice ablation rates. The former buoyancy-driven regime recovers scalings consistent with the laboratory-scale experiments, whilst the latter shear-driven regime is consistent with parameterizations that use heat and salt transfer coefficients dependent on the velocity in the core of the turbulent boundary layer. We determine a criterion for the transition between these regimes, and estimate the range of oceanographic conditions and ice-face geometries under which each regime occurs. Finally, we discuss the consequences for modelling the melting of ice shelves and glacier termini.


Measurements of snow density, accumulation and compaction along the iSTAR traverse, Pine Island Glacier, Antarctica

Elizabeth Morris, Robert Gurney, Andrew Smith

Corresponding author: Elizabeth Morris

Corresponding author e-mail: elizabeth.morris@reading.ac.uk

At 22 sites along a traverse across the Pine Island Glacier catchment in the 2013/14 austral summer snow density profiles were measured to 13 m depth, using a neutron scattering technique. At six of these sites, nine profiles were collected on a 1 km2 nested grid, to allow the effect of spatial variability on scales of 1 m, 10 m 100 m and 1 km to be assessed. In the 2014/15 austral summer existing sites were remeasured and new profiles were collected close by. From these data we determine the accumulation rate over the period between traverses, estimate mean annual accumulation over the last decade and calculate vertical strain rates at each site. These provide ground truth for verification of the meteorological and snow densification models used to interpret satellite measurements of elevation change in the Pine Island Glacier catchment.


Interannual climate response of large Antarctic ice shelves

Laurie Padman, Fernando Paolo, Helen Fricker, Susan Howard

Corresponding author: Laurie Padman

Corresponding author e-mail: padman@esr.org

The 18 year time series of Antarctic ice shelf height variability h(t) reported by Paolo and others (2015; Science) provides an opportunity to assess the response of Antarctica’s large ice shelves (Ross, Filchner–Ronne and Amery) to interannual climate variability. Although these ‘cold-water’ ice shelves have low mean basal melt rates (<1 m a–1) and small 18 year average rates of change (dh/dt) compared with the rapidly thinning Amundsen and Bellingshausen Sea ice shelves, their contribution to net change in Antarctic ice-shelf volume is comparable to other regions because melting is integrated over large areas (>65% of the Antarctic ice-shelf total). On interannual timescales, time series of dh/dt for the large ice shelves show significant fluctuations, including sign reversals. We expect that the dominant causes of this variability are changes in precipitation and net basal melt rate. Here we compare measured dh/dt(t) with atmospheric variables (vector wind stress, air temperature and precipitation) from the ERA-Interim reanalysis model, and with sea-ice concentration from passive microwave. Change in precipitation over the ice shelves has a direct and immediate impact on ice-shelf volume. Wind stress over the continental shelf directly influences ventilation of the sub-ice-shelf cavities. Advection of sea ice and the surface ocean layer across the continental shelf influences circulation under the ice shelves, especially fluxes of warm water to the base of the shallower ice near the ice front.


Proglacial discharge plumes at tidewater glaciers: scalings for ice–ocean interaction from buoyant plume theory and time-lapse imagery

Donald Slater, Pete Nienow, Dan Goldberg, Andrew Sole, Tom Cowton, Doug Mair

Corresponding author: Donald Slater

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

Rapid dynamic changes at the margins of the Greenland ice sheet, synchronous with ocean warming, have raised concern that tidewater glaciers can respond sensitively to ocean forcing. Our understanding of the processes encompassing ocean forcing nevertheless remains embryonic. Here we investigate submarine melting using buoyant plume theory with application to Kangiata Nunata Sermia (KNS), a tidewater glacier in southwest Greenland. We use buoyant plume theory to model plumes of fresh subglacial discharge emerging from the grounding lines of tidewater glaciers. Avoiding recourse to complex numerical simulations, buoyant plume theory can describe plume dynamics and associated submarine melt in idealized scenarios. We examine fundamental scalings for submarine melt m with a particular focus on variations in subglacial discharge Q (i.e. we investigate the value of the exponent γ in the relationship m ~ Q^γ). A value γ = 1/3 is prevalent in the literature but values as high as γ = 0.88 have been reported in numerical modelling. We suggest the appropriate value of γ depends on a combination of the subglacial discharge and fjord stratification, and that variation in these factors can explain the differing values of γ found in the literature. We further propose that higher values of γ are appropriate for large tidewater glaciers in Greenland, and therefore submarine melt volumes at these glaciers might be more sensitive to variation in subglacial discharge than previously thought. A second set of scalings are obtained for characteristic plume heights, facilitating prediction of (1) the depth of plume outflow in a fjord and (2) whether the plume should be visible on the fjord surface. In combination with time-lapse imagery of KNS, we use these scalings to probe near-terminus subglacial hydrology, a recently identified but important control on submarine melt volume and distribution.


Grounding zone sediment accumulation and tidewater and ice-shelf stability

Ross Powell, Timothy Hodson

Corresponding author: Ross Powell

Corresponding author e-mail: rpowell@niu.edu

In general, fluxes of subglacial sediment to grounding zones are poorly constrained. From studies of modern systems, grounding zone accumulation rates vary over orders of magnitude ranging from millimeters to tens of meters per year, dependent primarily on glacial regime and flux magnitudes by meltwater discharges. Temperate glacial regimes have dynamic subglacial water systems fed by abundant surface runoff and are capable of transferring millions of cubic meters of sediment to build morainal banks at a grounding zone. The latter are unconstrained in height due to their termini being tidewater cliffs, and can build vertically and in volume to decrease water depth at a terminus if it remains quasi-stable for a short period of time. The morainal bank structure can thus decrease exposure to marine melting and provide back-stress, which combined, also decrease fracture propagation in surface and basal crevasses. Thus if ablation forces change at a terminus, it may remain stable for a period of time, delaying retreat. As glacial regime cools subglacial stream fluxes decrease, which due to them being the fastest and most voluminous process of sediment transfer, results in progressive lowering of sediment accumulation rates at colder grounding zones. For the polar ice sheet end member, sediment flux is primarily by shearing of subglacial till – yielding sediment accumulation rates on the order of mm to cm per year. When polar glaciers terminate in ice shelves, sedimentation is also constrained by the height of the sub-ice-shelf cavity, and its primary accumulation site is under grounded ice rather than in the ocean cavity. As a consequence, both the sediment flux and its effect at changing grounding zone stresses are less effective than for morainal banks. These grounding zone wedge deposits could be sufficient to counteract rates of sea-level rise, but not ice thinning driven by high rates of ocean melt.


Modeling today’s sea-level contribution of glacial Antarctica

Torsten Albrecht, Ricarda Winkelmann, Anders Levermann

Corresponding author: Torsten Albrecht

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

The present dynamic state of the Antarctic ice sheet (and the bed it rests on) is mainly a product of past climate evolution, namely the history of advance and retreat during the last glacial cycles. Bed uplift and grounding line migration can occur delayed in response to past oceanic and atmospheric forcing. To this end, we need to account for this internal memory in order to better understand present changes and to project future contributions to sea-level rise, particularly with regard to anthropogenic climate change. This requires a fully dynamic model including ice-shelf dynamics as well as a continental-scale treatment of the transition zone and a proper coupling to oceanic and atmospheric forcing data, all of this is included in the Parallel Ice Sheet Model (PISM). Instead of aiming at a best-guess simulation, we provide an ensemble of model simulations for 15 km resolution that incorporates uncertainties from climate boundary conditions, internal process-modeling and ice parameter choices. With this approach we produce a broad ensemble of model representations of the present-day Antarctic ice sheet, which is at the same time well constrained by paleo-climatic data (e.g. LGM configuration) and present-day observations.


Greenland ice sheet–ocean–atmosphere interactions in a fully coupled GCM (EC-Earth–PISM), evaluated using high-resolution regional climate modelling and observations

Ruth Mottram, Shuting Yang, Christian Rodehacke, Marianne Sloth Madsen, Synne Høyer Svendsen, Jens Hesselbjerg Christensen

Corresponding author: Ruth Mottram

Corresponding author e-mail: rum@dmi.dk

The cryosphere in the Earth system is critical for modulating global and regional sea level, and has an important role in controlling regional climate at the high latitudes and influencing seasonal weather patterns in the mid-latitudes. Recently observed rapid changes in key elements of the cryosphere including sea ice, snow cover and glacier surface mass balance show that the Arctic region is currently experiencing a rapid transition. However, projections of both present-day and future cryospheric changes made by the CMIP5 models, for example in sea-ice extent, have a widespread and, in historical simulations, often a relatively poor fit to observations. Equally, only a few of the CMIP5 models have a realistic representation of land ice or include, for example, an interactive ice sheet. Here we present results from the EC-Earth GCM, which has been fully two-way coupled to the PISM for Greenland, in order to study the feedback processes between the climate system and the Greenland ice sheet. Results from experimental simulations show that fully coupling an ice-sheet model can alter ocean circulation, sea-ice extent and regional climate when compared with uncoupled experiments. For example, coupled experiments in EC-Earth–PISM show a dampening of the expected increase in Arctic temperatures under high RCP scenarios, possibly related to differences in freshwater forcing and sea-ice concentration. To assess the quality of the climate forcing from the GCM to the ice-sheet model we compare the energy-balance and surface mass-balance (SMB) output with that from very high resolution (0.05 degrees) simulations by the regional climate model (RCM) HIRHAM5. The RCM has been thoroughly evaluated over a wide range of climate parameters for Greenland, which allows us to be confident it gives a representative climate forcing for the Greenland ice sheet. In order to make a fair comparison, we use downscaled simulations forced with both ERA-Interim reanalysis and the EC-Earth RCP8.5 scenario to run an offline energy-balance model (EBM) and compare with output from the same EBM forced directly with EC-Earth output. The SMB forcing from the EC-Earth model is comparable to that provided by HIRHAM5 but important differences in the distribution of, for example, melt area and extent also emerge. A cold bias in the Arctic during the historical period within EC-Earth is similarly propagated through the HIRHAM downscaling.


The ice–ocean interface; a dynamic boundary (the effects of treating ice as a porous medium)

Jacob Buffo, Britney Schmidt

Corresponding author: Jacob Buffo

Corresponding author e-mail: jacob.buffo@eas.gatech.edu

Here I describe the procedure for how we are producing an initial one-dimensional model of ice-shelf dynamics, treating the bottom layer as a porous media. There are a number of crucial parameters that need to be taken into account when designing such a model. Additionally there is a need for methodologies by which to truth test its validity. The architecture of the model will consist of a set partial differential equations that represent conservation equations necessitated by the laws of thermodynamics and the theory of fluid transport in porous media. Boundary conditions for the model will be implemented using a combination of theoretical and experimental results found in the literature. The most basic model of this type will be capable of producing temporally and (one-dimensional) spatially dependent solutions of temperature, salinity, brine velocity and accretion/melt rates for the ice–water interface in question. This model can then be tested against in situ data obtained by the Antarctic submersibles Icefin and ARTEMIS, part of the SIMPLE project, over the 2014 and 2015 field seasons. This initial model will provide a benchmark study that expresses how cycling between the ice and ocean can affect the large-scale properties and dynamics beneath ice shelves, including basal melt and accretion rates, and investigate gas exchange through ice. Such models are necessary to go beyond equations of state and describe the systems-level behavior of these environments. This plays directly into an additional goal of the model: the utilization of small-scale, high-fidelity results from this model to benefit large-scale Earth systems models that currently struggle to incorporate the dynamics of the world’s ice shelves, and their effects on the Earth’s global cycles and climate. I will discuss preliminary results of an implemented numerical model, and outline future research goals that include moving towards two- and three-dimensional models capable of simulating the basal dynamics of entire ice shelves, incorporating more complex dynamics such as currents and frazil formation, and the applicability of the model to astrobiology.


Greenland outlet glacier dynamics from a complete remote-sensing elevation change record

Beáta Csathó, Anton Schenk, Cornelis van der Veen, Michiel van den Broeke

Corresponding author: Beáta Csathó

Corresponding author e-mail: bcsatho@buffalo.edu

Elevation changes are sensitive indicators of ice-sheet mass balance. When partitioned into changes caused by ice dynamics, surface mass-balance (SMB) anomalies and firn-compaction, they allow investigation of ice dynamic processes directly. We updated our 1993–2012 elevation-change record of the Greenland ice sheet (GrIS) with adding 2012–2014 airborne laser altimetry data and using SMB estimates from RACMO2.3. Most outlet glaciers exhibited steady or intermittent dynamic thinning during the whole period. However, a significant variability of elevation changes was detected both within single drainage basins and between neighboring glaciers, indicating that the response of individual outlet glaciers to external forcings is highly modulated by local conditions. To extend the coverage from the narrow swaths of laser altimetry to the entire fast-flowing region of the ice sheet, we incorporated digital elevation models into our Surface Elevation Reconstruction And Change detection (SERAC) method. Using the laser altimetry-derived time series as control information, we reconstructed elevation changes with sub-meter accuracy, even in areas where no stable control surfaces exist. The new record, partitioned into elevation changes due to ice dynamics and surface processes, allows us to address two major questions regarding the dynamic evolution of the GrIS over the next 100–200 years: (1) How representative are the four well-studied glaciers (Jakobshavn Isbræ, Helheim, Kangerlussuaq and Petermann glaciers) for other outlet glaciers? (2) What physical processes are the most likely cause for observed glacier changes in Greenland? The presentation shows detailed reconstructions of dynamic thickness changes for different types of outlet glacier behaviors including: steady and intermittent thinning; rapid thinning that abruptly terminates, quickly followed by thickening; steady thickening; changes due to proglacial lake drainage and surging. We explore these changes at different spatial scales and resolutions, from entire drainage basins to the crevasse patterns that are changing throughout the seasons. Special attention is given to the periods when ice dynamics transitions between opposing behaviors, for example from thinning to thickening. Finally, time series of height-above-floatation and effective pressure at the bed will be constructed to aid the interpretation of dynamic changes.


A new model for polythermal ice

Ian Hewitt, Christian Schoof

Corresponding author: Ian Hewitt

Corresponding author e-mail: hewitt@maths.ox.ac.uk

The dynamics of ice sheets and glaciers depend sensitively on their thermal structure. Many ice masses are polythermal, containing both cold ice, with temperature below the melting point, and temperate ice, with temperature at the melting point. The temperate ‘ice’ is really an ice–water mixture, with water produced at grain boundaries by dissipative heating. Although the water content is typically small, it can have an important effect on ice dynamics; water content controls ice viscosity, and internal meltwater percolation affects hydrology. Locations where this may be important are in the enhanced shear layer at the base of fast-flowing outlet glaciers, and in the shear margins of ice streams. In this study, we examine a simplified model to determine the heat and water content of polythermal ice, with particular attention to the water transport driven by gravity in the temperate ice. We study the behaviour of the model at boundaries between cold and temperate ice, and compare this behaviour with other methods in the literature such as the commonly used enthalpy method. In some circumstances, the model reduces to these methods, but in some cases it does not. Based on the results of this analysis, we suggest a modified enthalpy method that allows for drainage under gravity but that can be relatively easily implemented in ice-sheet models.


Seismic observations of ocean stratification beneath the Pine Island Glacier ice shelf, West Antarctica

Leo Peters, Sridhar Anandakrishnan, Tim Stanton, William Shaw, Kiya Riverman, David Holland, Robert Bindschadler

Corresponding author: Leo Peters

Corresponding author e-mail: leoepeters@gmail.com

Pine Island Glacier ice shelf (PIGIS) may control the flow speed of Pine Island Glacier from interior West Antarctica to the open ocean. Rapid (and increasing) basal melting and thinning of the ice shelf has been observed in recent decades, highlighting the potential for the continued inland migration of the grounding line of PIG and further acceleration of Pine Island Glacier. The likely cause of these rapid changes in the structure and dynamics of the PIGIS is an increase in warm Circumpolar Deep Water penetrating beneath the ice shelf. We present the results of a seismic reflection experiment conducted in the center of the PIGIS, where we image a stratified ocean column in the along-flow direction beneath the ice shelf. Two distinct reflective zones are identified in the seismic data, both of which align with in situ oceanographic observations of temperature and salinity structure through the water column. Furthermore, even though these seismic data were collected over the course of 3 weeks in the field, the observed seismic reflectors are spatially continuous, suggesting that these scattering layers are stable over timescales of at least a month. These results highlight the effectiveness of employing active seismic methods to capture ocean water-mass boundaries beneath ice shelves, which could prove useful in better understanding the interior reaches of the large ice shelves of Antarctica where the use of unmanned submersibles is implausible.


Rapid growth and persistence of efficient subglacial drainage under kilometre thick Greenland ice

David Chandler, Peter Nienow, Jemma Wadham, Sam Doyle, Andrew Tedstone, Alun Hubbard

Corresponding author: David Chandler

Corresponding author e-mail: peter.nienow@ed.ac.uk

The relationship between surface melting and ice motion will affect how the Greenland ice sheet responds to climate and the structure of the subglacial drainage system may be crucial in controlling how changing melt rates impact ice motion. However, the extent to which subglacial channels can develop tens of kilometres inland from the ice margin under thick (>1 km) ice remains equivocal. In particular, several numerical modelling studies suggest that efficient subglacial channels cannot evolve on seasonal timescales, even under extreme inputs of surface meltwater. Here we present a suite of hydrological and ice-motion data collected in summer 2012 in the vicinity of a moulin located ~40 km from the western margin of the Greenland ice sheet where ice is ~1 km thick. Supraglacial discharge into the moulin was monitored from the onset of surface drainage and the tracer sulphur hexafluoride (SF6) was injected into the moulin at repeat intervals following the initiation of surface drainage and its emergence monitored at the ice-sheet margin. The tracer results indicate a rapid evolution from a slow, inefficient drainage system to a fast, hydraulically efficient system within ~3 weeks from the onset of surface drainage. Once the efficient en/subglacial pathway was established, it remained open (as evidenced by the fast tracer return times) even during periods of low surface melt (~0.01 m d–1) when discharge into the moulin was <4 m3 s–1 and ceased overnight. Ice motion in the vicinity of the moulin slowed following the establishment of the channelized drainage pathway with a clear diurnal cyclicity driven by variations in supraglacial discharge. Our results confirm that hydraulically efficient subglacial drainage can exist tens of kilometres from the ice-sheet margin where ice is ~1 km thick, that the drainage configuration can form rapidly in a matter of weeks and that it persists even during cool periods when local surface melt rates are low. In addition, given that the maximum measured input to the moulin was ~11 m3 s–1, the impact of the drainage of order of magnitude larger volumes of meltwater from supraglacial lakes on the subglacial drainage system needs further investigation.


Evolution of Antarctic ice dynamics from ICESat

Greg Babonis, Beáta Csathó, Toni Schenk, Michiel van den Broeke

Corresponding author: Greg Babonis

Corresponding author e-mail: gbabonis@buffalo.edu

Altimetry measurements, combined with climate model outputs, can directly reveal the spatial and temporal changes of ice dynamics. Here we present the first synoptic reconstruction of ice dynamic elevation changes of the Antarctic ice sheet (AIS), providing details of the temporal evolution from 2004 to 2009. Using the Surface Elevation Reconstruction and Change Detection (SERAC) method, combined with output from the Regional Atmospheric Climate Model 2 (RACMO 2), we are able to partition surface elevation change records into the corresponding contributions from climate and ice dynamics; resulting in maps of the annual evolution of dynamic thickening and thinning for the AIS. In addition to producing annual change grids, we also developed a method that directly compares elevation changes as a function of time; a robust solution to explore dynamic elevation change patterns. 120 000 individual records of surface elevation change are generated across the AIS and converted into dynamic thickness change by correcting for isostatic processes and elevation changes driven by surface mass balance, and gridded annually. Those records with identifiable elevation change trends are clustered by comparing dynamic changes as a function of time. Clusters are grown to distinguish between localized, large-amplitude, nonlinear dynamic changes, likely caused by changes in subglacial hydrology, and regional-scale linear dynamic thickening/thinning events in response to processes related to larger ice-sheet stability. The resulting interpretation of the dynamic changes coupled with the elevation changes due to surface processes allows for the characterization of the complex spatial and temporal patterns of dynamic mass loss across Antarctica 2004–2009, especially along individual outlet glaciers. We show the 5 year evolution of rapid thinning at Totten Glacier, detect regional thickening changes near the Foundation ice stream analogous to the Kamb Ice Stream stagnation on the Siple Coast, and show a complex pattern of changes to the thinning of Pine Island Glacier. Additionally, these data allow us to quantify the annual contribution of Antarctic ice loss to sea-level rise.


Observations of basal melt channels on Antarctic ice shelves

Karen Alley, Ted Scambos

Corresponding author: Karen Alley

Corresponding author e-mail: karen.alley@colorado.edu

The formation of basal melt channels has the potential to affect the stability of floating ice shelves. We use remotely sensed data to classify and survey the locations of basal melt channels on ice shelves around the Antarctic continent. Channels and associated features are identified using visible satellite imagery and, where data are available, their presence is confirmed using IceBridge MCoRDS ice-penetrating radar. Identified channels are ~1–5 km wide and ~50–250 m deep. Though constituting just 11% of Antarctic ice shelf area, the ice shelves of the Amundsen and Bellingshausen Sea have ~41% of all melt channels. The high concentration of melt channels in the region, combined with (1) a regional correlation between grounding line depth and melt channel density, and (2) a continent-wide correlation between ice-shelf melt rates and melt channel density, suggests an association between the density of melt channels and warm Circumpolar Deep Water presence. The presence of persistent polynyas at the termini of 22 melt channels in the Amundsen/Bellingshausen Sea also demonstrates significant warm water forcing associated with these channels. In at least one location, channels are exhibiting rapid change. At the head of a channel that has grown ~20 km towards the grounding line over the last ~30 years on the Getz Ice Shelf, ICESat data reveal a basal deepening rate on the order of 10–20 m a–1. Melt channels at a shear margin of the Getz Ice Shelf are also associated with a region of fracture that has developed over the past decade, suggesting that basal melt channels may be associated with the weakening of ice shelves.


Basal ice conditions under McMurdo Ice Shelf

Britney Schmidt

Corresponding author: Britney Schmidt

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

The McMurdo Ice Shelf is an interesting test bed for observing ice–ocean exchange. In the 2012, 2014 and 2015 austral summer Antarctic field seasons, NASA’s SIMPLE project (Sub-Ice Marine and PLanetary-analog Ecosystems), has been tasked with characterizing ice and ocean processes below and within the McMurdo Ice Shelf (MIS), a small portion of the larger Ross Ice Shelf easily accessible from USAP’s McMurdo Station. Using sub-ice vehicles, ice-penetrating radar and other measurements of this unexplored region, the SIMPLE team is building a comprehensive picture of processes at the ice–ocean interface and within the brine-infiltrated ice shelf. Between the two field excursions, the SIMPLE team has had the opportunity to produce initial characterization of the ice–ocean interface and processes. We have collected imaging data of the ice and inhabitants, as well as conductivity and temperature profiles of the water column. In 2012 the team explored at a single location 5 km back from the front of the ice shelf using the small ROV SCINI (Moss Landing). Here we observed ablation of the ice and a heterogeneous water column, consistent with melting by fast-moving impingent currents. Both imaging data and the CTD profiles are consistent with this conclusion. In 2014, SIMPLE utilized SCINI and a new AUV/ROV vehicle, Icefin (Georgia Tech), to characterize five hot-water-drilled sites below the MIS. These locations ranged between 10 and 20 km from the shelf front. Here we observed very different ice conditions from that in 2012. Uniformly, large amounts of platelet ice were observed, regardless of ice-shelf thickness. The layer of platelets was between 1 and several meters thick depending on the site, and ranged between uniform in character to regions with large platelet spears and columns. At one site, we observed the possible formation of a compressed layer of platelet ice, physically separated from the bottom of the shelf by a thin water lens, at the base of which was newer forming platelet ice. We observed homogeneous water column below the ice, consistent with the formation of platelet ice. We also observed a complex community of organisms at the sea floor near Black Island under this permanent ice cover. The 2015 season with ARTEMIS (SAS) will allow extended exploration range and allow us to observe gradients in processes. I will describe the project and results to date, including imaging and preliminary mapping of the conditions below the MIS.


High-resolution velocity observations and model results reveal a strong bed at stable, tidewater Rink Isbræ, West Greenland

Timothy Bartholomaus, Ginny Catania, Ryan Walker, Leigh Stearns, Denis Felikson, Mark Fahnestock, Ryan Cassotto, Jonathan Nash, Emily Shroyer, David Sutherland

Corresponding author: Timothy Bartholomaus

Corresponding author e-mail: tbartholomaus@ig.utexas.edu

At tidewater Rink Isbræ, on the central west coast of Greenland, satellite observations reveal that glacier velocities and terminus positions have remained stable, while the lowest 25 km have thinned 30 m since 1985. Over this same time period, other tidewater glaciers in central west Greenland have retreated, thinned and accelerated. Here we present field observations and model results to show that the flow of Rink Isbræ is resisted by unusually high basal shear stresses. Terrestrial radar interferometry (TRI) observations over 9 days demonstrate weak velocity response to 4 km wide, full thickness calving events. Velocities at the terminus change by ±10% in response to rising and falling tides within a partial-width, 2.5 km long floating ice tongue; however, these tidal perturbations damp out within 2 km of the grounding line. Inversions for basal shear stress and force balance analyses together show that basal shear stresses in excess of 300 kPa support the majority of the driving stress at thick, steep Rink Isbræ. These observational and modeling results tell a consistent story in which a strong bed may limit the unstable tidewater glacier retreats observed elsewhere. An erosion-resistant quartzite bed, with low fracture density, may play a major role in this bed strength.


Reducing uncertainties in projections of Antarctic ice mass loss

Gaël Durand, Frank Pattyn

Corresponding author: Gaël Durand

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

Climate model projections are often aggregated into multi-model averages of all models participating in an intercomparison project, such as the Coupled Model Intercomparison Project (CMIP). A first initiative of the ice-sheet modeling community, SeaRISE, to provide multi-model average projections of polar ice sheets’ contribution to sea-level rise recently emerged. SeaRISE Antarctic numerical experiments aggregate results from all models willing to participate without any selection of the models regarding the processes implemented in. Here using the experimental set-up proposed in SeaRISE we confirm that the representation of grounding line dynamics is essential to infer future Antarctic mass change. We further illustrate the significant impact on the ensemble mean and deviation of adding one model with a known biais in its ability of modeling grounding line dynamics. We show that this biased model can hardly be discriminated from the ensemble only based on its estimation of volume change. However, tools are available to test parts of the response of marine ice-sheet models to perturbations of climatic and/or oceanic origin (MISMIP, MISMIP3d). Based on recent projections of Pine Island Glacier mass loss, we further show that excluding ice-sheet models that do not pass the MISMIP benchmarks decreases by an order of magnitude the mean contribution and standard deviation of the multi-model ensemble projection for that particular drainage basin.


Sensitivity of Antarctic Bottom Water formation to fresh water from sea ice and basal ice-shelf meltwater using noble gases

Laura Herraiz-Borreguero, Peter Schlosser, Stanley S. Jacobs, Beatriz Peña-Molino, Stephen R. Rintoul, Bob Newton

Corresponding author: Laura Herraiz-Borreguero

Corresponding author e-mail: lherraiz@nbi.ku.dk

Antarctic Bottom Water (AABW) plays a key role in the global exchange of heat, fresh water, nutrients and oxygen between the surface and the deep ocean. However, the sensitivity of AABW formation to fresh water from sea ice and continental ice/ocean interaction are poorly understood. Helium and neon noble gas measurements provide a means to trace sea ice and glacial meltwater influences via their unique dissolved gas saturation pattern (Schlosser, 1986; Schlosser and others, 1990). Here we use measurements of helium, neon and oxygen stable isotopes obtained near the Mertz Glacier Tongue (MGT), East Antarctica, to study the impact of the MGT calving on AABW formation and local ocean circulation. The tracer data are integrated with hydrographic surveys. We will show that the source of the observed helium in ice-shelf water is associated with subglacial fresh water that enters the ocean by crossing the MGT grounding line. Preliminary results from the helium and oxygen isotopes also indicate that the source water masses involved in the formation of AABW in Adélie Land post-calving are Circumpolar Deep Water and Antarctic Shelf waters (ice-shelf water and dense shelf water) in an almost 50:50 contribution. Results from this analysis will further our understanding of the physical processes driving the recent freshening of AABW observed in the Indian sector of the Southern Ocean.


Ensemble predictions of future Antarctic mass loss with the FETISH model

Frank Pattyn, Brice Van Liefferinge

Corresponding author: Frank Pattyn

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

Ice-sheet models become more and more components of global climate system modelling instead of stand-alone features to study cryospheric processes. Full coupling of ice-sheet models to atmospheric and ocean models requires a standard for ice-sheet models, and more precisely for marine ice-sheet models, where complex feedbacks between ice and ocean, such as marine ice-sheet instability, and the atmosphere, such as the elevation mass-balance feedback, operate at different timescales. Recent model intercomparisons (e.g. SeaRISE, MISMIP) have shown that basic requirements for marine ice-sheet models are still lacking and that the complexity of many ice-sheet models is focused on processes that are either not well captured numerically (spatial resolution issue) or are of secondary importance compared to the essential features of marine ice-sheet dynamics. Here we propose a new and fast-computing ice-sheet model, devoid of most complexity, but capturing the essential feedbacks when coupled to ocean or atmospheric models. Its computational efficiency guarantees to easily test its advantages as well as limits through ensemble modelling. FETISH (Fast Elementary Thermomechanical (marine) Ice SHeet model) is a vertically integrated hybrid (SSA/SIA) ice-sheet model extended with a Weertman sliding law. Although vertically integrated, thermomechanical coupling is ensured through a simplified representation of ice-sheet thermodynamics based on an analytical solution of the vertical temperature profile, enhanced with strain heating. The marine boundary is represented by a parameterized flux condition similar to Pollard and Deconto (2012), based on Schoof (2007). A simplified ice shelf is added to account for buttressing of ice shelves in this parameterization. The ice-sheet model is solved on a finite-difference grid and special care is taken to its numerical efficiency and stability. We perform an ensemble experiment with FETISH by forcing the model with different ice-shelf melt rates and basal sliding perturbations and compare results to recent model intercomparisons of the Antarctic ice sheet (e.g. SeaRISE; Favier and others, 2013).


The evolution of Antarctic ice-shelf channels in observations and models

Reinhard Drews, Sophie Berger, Lionel Favier, Brice van Liefferinge, Kenichi Matsuoka, Frank Pattyn

Corresponding author: Reinhard Drews

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

Ice-shelf channels are prevalent in many Antarctic ice shelves. They distinctly appear in satellite imagery as elongated lineations extending from the grounding line to the ice-shelf front. The thinner ice inside the channels mediates (and originates from) channelized melting, with basal melt rates which can be significantly higher inside than outside the channels. Because satellite sensors are often too coarsely resolved and because dedicated field studies are sparse, it is difficult to quantify the basal mass balance inside channels. This is a major drawback for understanding the ice–ocean interactions inside the channels, and ultimately for assessing the role of the channels in defining the ice-shelf stability. Here we present a geophysical dataset collected over the last 3 years imaging a number of channels on the Roi Baudouin Ice Shelf, Dronning Maud Land, Antarctica. Static GPS markers give yearly averages of ice-flow velocities in the channel’s surrounding and at a nearby pinning-point; two single-frequency GPS monitor daily and seasonal variations of ice velocity; repeat measurements with a stationary phase-sensitive radar (pRES) quantify vertical velocities across a channel (and hence melt rates); and stationary radar wide-angle data map spatial variations in radar wave propagation speed (and hence density). Ground-based radar profiling (10 MHz, 400 MHz) visualizes increasingly deformed internal layers along the entire ice column in 11 transects across channels. We investigate the channel evolution in synthetic geometries using a full-Stokes ice-flow model (Elmer/Ice) and find that melting channels imprint the surface velocities characteristically. This effect is also visible in the observational data and it can be used as a proxy to detect channelized melting from space. The deformed radar isochrones inside the channels are likely witness to both a decreased basal mass balance and an enhanced surface mass balance. Using the ice-flow model, we aim to separate the two effects and to solve for basal melt rates using an inversion scheme. The combination of field data and ice-flow modeling presented here is a step towards incorporating channelized melting in ice–ocean modeling.


SAR observation and inverse modelling of ice-shelf pinning point dynamics and channel formation

Sophie Berger, Reinhard Drews, Lionel Favier, Veit Helm, Wolfgang Rack, Frank Pattyn

Corresponding author: Sophie Berger

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

Ice-shelf channels (elongated lineations in which ice is thinner) and pinning points (locally grounded features embedded within the otherwise freely floating ice shelves) are ubiquitous in Antarctic ice shelves. Although these features are readily visible in satellite imagery, ice-thickness and ice-velocity variations in their surrounding are typically heavily undersampled. This lack of data is the main bottelneck in assessing the role of pinning points and ice-shelf channels in defining the ice-shelf stability. Pinning points are often only a few kilometers wide, but the increased friction nevertheless causes an ice-shelf-wide slowdown in velocities. Ice-shelf channels, on the other hand, mediate channelized melting and significantly alter the basal mass balance on short horizontal scales. We present highly resolved surface velocities of the Roi Baudouin Ice Shelf, Dronning Maud Land, derived from interferometric synthetic aperture and speckle tracking of ERS 1/2 and PalSAR data (50 and 100 m postings, respectively). The velocities are validated using a set of ground-based GPS measurements. The high spatial resolution allows us to investigate flow anomalies near channels, which presumably originate from enhanced basal melting. We combine the flow velocities with surface elevation derived from TanDEM-X and CryoSAT-2. Using an inversion scheme based on an ice-flow model (BISICLES), we show the importance of the pinning point on the ice-shelf dynamics. Without observational data that adequately resolves the pinning point, the inversion results in an erroneous ice-shelf rheology which is problematic for prognostic simulations. Using ground-based GPS surface elevation, we demonstrate that TanDEM-X is an ideal sensor to map the channel morphology at the ice-shelf surface. The corresponding elevation models allow us to derive the basal melt rates inside the channels using the continuity equation in a Lagrangian framework circumnavigating steady-state assumptions. The combined approach of satellite remote-sensing and ice-flow modeling presented here clearly demonstrates the need for high-resolution data in order to better understand the effects of ice-shelf buttressing and channelized melting.


Observationally constrained projections of Antarctic ice sheet instability

Catherine Ritz, Tamsin Edwards, Gaël Durand, Tony Payne, Vincent Peyaud, Richard Hindmarsh

Corresponding author: Catherine Ritz

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

Large parts of the Antarctic ice sheet lie on bedrock below sea level and may be vulnerable to a positive feedback known as marine ice-sheet instability (MISI), a self-sustaining retreat of the grounding line triggered by oceanic or atmospheric changes. There is growing evidence that MISI may be underway throughout the Amundsen Sea Embayment (ASE) of West Antarctica, induced by circulation of warm Circumpolar Deep Water. If this retreat is sustained the region could contribute up to 1–2 m to global mean sea level, and if triggered in other areas the potential contribution to sea level on centennial to millennial timescales could be two to three times greater. However, physically plausible projections of Antarctic MISI are challenging: numerical ice-sheet models are too low in spatial resolution to resolve grounding line processes or else too computationally expensive to assess modelling uncertainties, and no dynamical models exist of the ocean–atmosphere–ice-sheet system. Furthermore, previous numerical ice-sheet model projections for Antarctica have not been calibrated with observations, which can reduce uncertainties. Here we estimate the probability of dynamic mass loss in the event of MISI under a medium climate scenario, assessing 16 modelling uncertainties and calibrating the projections with observed mass losses in the ASE from 1992 to 2011. We project losses of up to 30 cm sea level equivalent (SLE) by 2100 and 72 cm SLE by 2200 (95% credibility interval: CI). Our results are substantially lower than previous estimates. The ASE sustains substantial losses, 83% of the continental total by 2100 and 67% by 2200 (95% CI), but in other regions losses are limited by ice dynamical theory, observations, or a lack of projected triggers.


Icefin: a new small modular AUV for polar under-ice exploration

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

Corresponding author: Britney Schmidt

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

Icefin was designed and built at Georgia Tech and fielded in the austral summer 2014 in McMurdo. The vehicle is a modular, field-portable hybrid autonomous underwater/remote-operated vehicle designed as a long-range and deep water under-ice robotic oceanographer that can survey cavity geometry, ice properties and ocean conditions beneath floating ice that are not resolvable in remotely sensed observations or using localized mooring data. Icefin’s unique design provides greater capability than small vehicles and improved operational simplicity relative to large vehicles. The current Icefin vehicle is 12 inches in diameter, 10 feet long and weighs 220 pounds. Icefin consists of five modules including a nose cone, two vertical/horizontal thruster modules, a sensor bay, an electronics module and a rear propulsion module. Because of its small size and modularity, the vehicle can be broken down by module, transported in Pelican boxes and deployed easily. As opposed to the larger vehicles which require much greater logistics with much larger costs, the relatively small modular Icefin can be deployed through small holes drilled in the ice. This modular design is robust to failures, allows easy switching out of science modules that increases the number of sensors available for science without increasing vehicle size, reduces the assembly time of the vehicle, provides a simpler mode for disassembling while in the deep field, and allows the vehicle to be customized for each mission, with imagers, sonar and sensors pointing up or down depending on whether the ice or the silicate interface is under study. Here vehicle control and data systems can be stably developed and power modules added or subtracted for mission flexibility, while multiple sensor bays can be developed to serve multiple science objectives. Icefin is currently fitted with sensors for scientific analysis of the ice–ocean system, including a sensor bay with Side Scan Sonar (SSS), Doppler Velocity Log (DVL) with current profiler, altimeter, and imaging sonar. This sensor bay may be pointed in the down position for ocean bottom mapping or in the upward position for ice–ocean interface mapping. The forward module includes a forward-looking blazed array sonar, a CTD sensor and obstacle avoidance camera. Initial work in Antarctica and local work in coastal Georgia has the vehicle operational, well-characterized and ready to go for future collaborations in any ice–water environment.


Tracking the propagation of crevasses in Helheim Glacier in Greenland

Joshua Hedgepeth, Catherine Walker, Britney Schmidt

Corresponding author: Joshua Hedgepeth

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

Related to the high level of crevassing observed in Greenland’s glaciers is the fact that calving events generally occur in the form of many small pieces collapsing from the ice front, rather than the large, tabular style often observed in Antarctic ice shelves. Satellite-enabled investigations and field studies have focused on the retreat of Greenland’s glaciers and properties of the proglacial ice melange (Sohn and others, 1998; Reeh and others, 2001; Joughin and others, 2004; Howat and others, 2005, 2008, 2010; Amundson and others, 2008; Moon and Joughin, 2008; Luthi and others, 2009; Walter and others, 2012). However, the actual transition mechanism from intact crevassed glacier – possessing a large amount of gravitational potential and fracture energy – to what is essentially a rubble pile in the proglacial fjords below remains something of a mystery. To address this, we are attempting to develop a fracture density criterion to relate observable surface appearance to likelihood of collapse in ice using high-resolution satellite imagery. Following our digitized mapping of glacier crevasses, we will seek to determine the fractal dimension of the features. With evidence of fractality, formation of crevasses on glaciers can be analyzed as a scale-invariant process. To begin our study, we focused on Helheim Glacier in Greenland. What is the relationship between the crevasse patterns, glacier flow and calving rate? The purpose of this study is to better understand these questions. Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) images were employed, beginning with images from May 2001, July 2003 and June 2005. The crevasses in the ice were traced electronically for further modeling. This was continued using image sources available on USGS Earth Explorer for 2007, 2010 and 2014. There are clear flow patterns within the ice that were obvious when viewing on a large scale, but upon closer look the crevasses become more sporadic. The margins tended to exhibit a larger amount of noise; the fractures did not always appear to follow any specific direction. What is clear, along the margins, is a type of sandwiching effect on the fractures. As they approach the calving front of the glacier, the crevasses become more evenly divided across the glacier. Further demonstrated is the relationship between crevasses and the concave shape formed at the calving front by calving.


Extraterrestrial glaciology: the role of subsurface lakes and basal fractures in ice-shell–ocean interactions on Jupiter’s moon Europa

Catherine Walker, Britney Schmidt

Corresponding author: Catherine Walker

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

We use terrestrial glaciology to better understand dynamics of icy water worlds of the solar system. For subsurface water to erupt onto the surface of Jupiter’s moon Europa, a surface- or bottom-initiated fracture must vertically propagate and penetrate the entire shell thickness (~30 km) to create a channel from the liquid reservoir below to the surface. We consider two processes in the formation of terrains observed on Europa: (1) the propagation of fracture systems and (2) their coalescence. Initiation and penetration of a surface crevasse is driven by stress at/near the surface. At depth, the compressive overburden pressure from the weight of the ice opposes these forces. Tensile stresses at the surface and within the brittle surface layer must be great enough to overcome the compressive stress at the base to allow for full fracture penetration of the shell. Aside from the whole problem of basal fracture propagation as opposed to surface fractures, another factor that is often ignored in icy moon fracture studies is the role of highly fractured materials in fracture propagation. The stress-shadow effect and fracture interaction, phenomena observed both in Earth’s ice shelves and in other media, such as wave-breaker walls, airplane propellers, Arctic permafrost, mud flats in Death Valley, and others, make a significant difference in fracture penetration depth of both surface and basal cracks (e.g. Walker and Bassis, 2013). Additionally, even on Earth it is unlikely that surface fractures alone could lead to fully ice-penetrating features, and we find the same result here (e.g. Bassis and Walker, 2011). To address the problem of crack penetration to a subsurface reservoir, we model the propagation of cracks that initiate at the bottom of the ice shell, and illustrate the parameters that might affect their propagation height towards the surface. We model water pressure within the crack dependent upon reservoir size and depth below the surface. The work of Schmidt and others (2011) and Walker and Schmidt (2015) showed that it is likely that Europa’s chaos terrains formed after the surface above a perched water pocket flexed and allowed cracks to initiate at the base of the ice lid. Additionally, we present a characteristic timescale of this process based on local energy balance. We will discuss the implications of this process timescale on the likelihood that the observed water plumes represent an interior water body that interacts with the outer ice.


Impact of basal melt rate on ice flow through the Totten ice shelf

Sainan Sun, Stephen Cornford, John Moore

Corresponding author: Sainan Sun

Corresponding author e-mail: sainansun1985@sina.com

Increased basal melting of ice shelves is likely, and may have a dramatic effect influence on ice flow speed and Antarctic mass loss. Totten Glacier is the primary outlet of Aurora Basin and discharges the largest volume of ice in East Antarctica. Rapid thinning of the Totten ice shelf is observed, which is suggested to be the result of rising melt rates caused by advection of warm deep ocean water near the grounding line. Here we test the sensitivity of the BISICLES adaptive mesh ice-sheet model to different melt rates beneath the ice shelves in the region of Aurora Basin. The initial melt rate comes from a 3-D primitive equation finite-different ocean model ROMS (Reginal Oceanic Modeling System). We take the ROMS model simulation result during 1997–2012, parameterize the melt rate and make it fluctuate to represent melt rates under different ocean forcing scenarios. We calculate the change of ice-stream speeds, grounding line evolution and mass loss of the whole basin with different ice-shelf melt rates to evaluate the influence of ice-shelf basal melt on Aurora Basin evolution.


Oceanographic observations at the Dotson Ice Shelf front, West Antarctica, and calculations of basal melting

Deb Shoosmith, Adrian Jenkins, Pierre Dutrieux, Stanley Jacobs, Tae Wan Kim, Sang Hoon Lee, Ho Kyung Ha, Sharon Stammerjohn

Corresponding author: Deb Shoosmith

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

It is well known that the ocean plays a key role in the process of mass loss from ice sheets through iceberg calving and basal melting. The Amundsen Sea, in the eastern Pacific sector of the Southern Ocean, is a region where the ice shelves are rapidly thinning. The widespread, coherent nature of the thinning suggests oceanic forcing, which has now been well documented for Pine Island Glacier. Studies using satellite data have indicated that Dotson Ice Shelf is melting at a rate of 8 m a–1 and thinning by about 3 m a–1. This study works on the problem from an oceanographic perspective. Observations spanning nearly a decade and a half (2000–2014) have been obtained at the Dotson Ice Shelf front. A total of seven hydrographic sections reveal the oceanographic environment in front of the ice shelf as well as changes in water properties and meltwater content over time. We investigate the variability in circulation and meltwater production beneath the ice shelf and produce estimates of the basal melt rate for this 14 year period.


Assessing the impact of iceberg calving using an ice-sheet model for present-day

Victoria Lee, Stephen Cornford, Antony Payne, Andrew Taylor

Corresponding author: Victoria Lee

Corresponding author e-mail: v.lee@bristol.ac.uk

We investigate the dynamic response of the Greenland ice sheet (GrIS) to different calving rates using an ice-sheet model. GrIS has been losing mass at an increasing rate over the last two decades and a significant proportion is due to dynamic thinning of narrow outlet glaciers. Dynamics are strongly linked to changes in the stress balance at the glaciers’ front caused by calving. We evolve the ice sheet from our present-day initial state with a calving model that determines the position of the calving front as the point where surface and basal crevasses penetrate the full thickness of the ice. The ice-sheet model uses adaptive mesh refinement which allows coarse resolution in the slow-moving interior of the GrIS and finer resolution in fast-flowing ice to capture the behaviour of the outlet glaciers. We investigate the effects of a sudden change in the calving rate using 100 year runs with different values of the crevasse water depth parameter and compare changes in the modelled ice velocity of the main outlet glaciers with observations from the 1990s and early 2000s.


Grounding line change analysis of the Antarctic ice shelves using existing grounding line products, satellite altimetry and topographic data

Huan Xie, Yang Xu, Saisai Lu, Shuang Liu, Lei Chen, Xiaohua Tong, Rongxing Li

Corresponding author: Huan Xie

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

The aim of this research is to investigate grounding line changes in the Antarctic ice shelves by comparing four existing grounding line products covering the period 1992–2009, which are available at National Snow and Ice Data Center (Haran and others, 2005; Brunt and others, 2010; Bindschadler and others, 2011; Rignot and others, 2011). These products used satellite remote-sensing data such as MODIS images, Landsat images, ERS‐1 and 2 data, RADARSAT‐1 and -2 imagery, ALOS PALSAR imagery, and ICESat/GLAS laser altimetry data. As the MODIS grounding line (MOA) and Landsat grounding line (ASAID) are represented by polylines, and the InSAR and ICESat grounding lines are represented by feature points that describe the grounding zone, three measures are used to inspect the grounding line differences, including: (1) passing through rate of grounding zone; (2) distance between different grounding line products; and (3) buffer between different grounding line products. The objectives of this research were to analyze uncertainties, derive grounding line changes, and discuss potentials and limitations of using grounding line products at local and continental scales. Great efforts were made to detect grounding line changes beyond the uncertainties of grounding line products. Satellite altimetry (including ICESat and Cyrosat-2), DEMs and bedrock maps were used to validate the results. Finally, a grounding line change map of the Antarctic ice shelves was generated.


Comparison of Antarctic surface elevation model products in representative regions

Tiantian Feng, Haifeng Xiao, Xiaohua Tong, Rongxing Li

Corresponding author: Tiantian Feng

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

Investigating the surface elevation changes of the Antarctic ice sheet is essential to the understanding of the associated mass change and global climate change. Great efforts have been made to produce the DEM products of Antarctica by using a variety of observations, such as laser altimetry data, radar altimetry data, high-resolution remote-sensing images and cartographic data. There are a variety of DEM products of Antarctica covering different time periods published by researchers from different institutions. These DEM products have been validated and applied in many science applications. However, a comparative study of the DEM products should be performed if they are used jointly to detect changes. The difficulty comes first due to the lack of detailed ground truth. As a result, the elevation differences between different DEM products may be the mixture of both elevation changes during corresponding time periods and the uncertainties in the DEM products. In this paper, we focus on the quality assessment of four DEM products. They include: (1) RAMP 2 DEM that was produced from ERS-1 data, ADD data and airborne RES data, etc; (2) ERS DEM produced by using ERS-1 altimetry data from 1993 to 1995; (3) ICESat DEM generated from ICESat altimetry data from 2003 to 2005; and (4) ERS-ICESat DEM produced by combining both ERS-1 data from 1993 to 1995 and ICESat data from 2003 to 2008. Considering the difficulty of obtaining the ground truth of surface elevation in multiple time periods, cross-validation strategy between these four DEM products is adopted. Three kinds of representative regions (lake, blue ice and bare rock) are chosen to carry out the comparison of the DEM products, because the surface elevation and changes in these regions are more understandable. We summarized the data sources and data-processing methods, including wave retracking methods, slope correction methods and elevation interpolation methods. Then, we calculated the elevation differences among the DEM products. The computed changes are analyzed against a number of factors such as radar signal penetration, surface slopes, surface processes and dynamics and others. The levels of uncertainty of the DEM products are given when applied in the chosen types of region. It is expected that the results of this study will provide some more useful information about the products and promote objective applications of the models.


Ice flow velocity estimation from ZY-3 and Landsat satellite imagery: an experimental test in Fisher Glacier, East Antarctica

Shijie Liu, Gang Qiao, Fansi Kong, Wenkai Ye, Xiaohua Tong, Weian Wang, Rongxing Li

Corresponding author: Shijie Liu

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

Ice flow velocity is a fundamental characteristic of glaciers and ice sheets that measures the rate at which ice is transported from the interior regions toward the ocean and how ice mass evolves with time (Rignot and others, 2011). Traditional ground-based measurements are spatially sparse and limited in coverage. Satellite remote-sensing provides an alternative method to map glacier flow with large and continuous coverage. In our study, Landsat-7 and ZY-3 satellite images were used to map the ice flow in Fisher Glacier, which is one of the three major glaciers discharging to the Amery Ice Shelf in East Antarctica. The Landsat-7 imagery was acquired on 14 February 2013, with a 15 m ground resolution and a swath width of 185 km. ZY-3 is China’s first civilian high-resolution stereo mapping satellite, and the imagery used in the study was collected on 3 March 2014, with a 2.1 m ground resolution and a swath width of 50 km. The workflow for image processing and ice flow velocity estimation is as follows: (1) ZY-3 image orthorectification: as the Landsat-7 imagery was an orthorectified product, the ZY-3 images need to be orthorectified. (2) Image resolution adjustment: to facilitate image matching, the ZY-3 imagery was downsampled to the same resolution of the Landsat-7 imagery. (3) Image co-registration: the ZY-3 and Landsat-7 images were co-registered by using corresponding points which are stable on the ground such as points on bare rocks. (4) Dense image matching: the ZY-3 and Landsat-7 images were dense-matched by using a coarse to fine hierarchical image-matching method. (5) Ice flow velocity estimation: ice flow velocities were estimated from the disparities of the matched points and the time difference of the two images, and they were further interpolated to produce the ice flow map. The derived ice flow velocity result from ZY-3 and Landsat-7 was compared with the Antarctic ice flow product, which was generated by Rignot and others (2011) from multiple satellite interferometric synthetic aperture radar data acquired from 2007 to 2009. It showed a general consistency in both speed magnitude and spatial distribution between the two maps. An analysis of the technical performance and a comparison of the speed changes in the two maps are given. This experimental test is one of the few early studies on the capability of ZY-3 imagery in ice flow mapping in Antarctic ice sheet mass-balance research.


Feature-based image matching for Antarctic ice flow measurement using DISP images from the 1960s

Gang Qiao, Fansi Kong, Wenkai Ye, Song Guo, Xuwen Ma, Zeyang Wang, Guanjie Tang, Xiaohua Tong, Rongxing Li

Corresponding author: Gang Qiao

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

Current ice flow products from remote-sensing techniques are mostly produced by satellite SAR and optical images which date back to the 1970s by Landsat images. Historical images such as aerial photographs or the first generation satellite images like declassified intelligence satellite photographs (DISP) images provide a valuable view of the Antarctic ice sheet’s configuration in the 1960s, extending the coverage of Antarctic ice flow observations. This research presents a framework and the initial result of our efforts in surface ice flow measurement in the Antarctic ice sheet by using DISP images from 1963. The framework mainly consists of two parts: geometric modelling of the DISP images, and feature-based image-matching strategy to derive the surface ice velocity. In the geometric modelling part, the camera calibration parameters, including focal length and lens distortions, were collected from the camera calibration reports. The tie points and control points were selected by considering topography, shadowing and with help of existing image mosaics and DEMs. The initial estimates of the exterior parameters were derived from the orbital ephemeris of specific missions, and they were then refined in a bundle adjustment process. The accuracy of the bundle adjustment was assessed and discussed. In the image-matching part, a feature-based hierarchical matching technique was proposed and implemented. After preprocessing, extracted feature points were matched hierarchically in each pyramid layer under different geometric constraints. A special process was devoted to handle the point displacement caused by the ice flow. The results show the reconstructed terrain and ice flow speed in 1963 in the experimental area.


Circulation and processes beneath the Ronne Ice Shelf, Antarctica

Svein Østerhus, Keith Nicholls, Keith Makinson

Corresponding author: Svein Østerhus

Corresponding author e-mail: svein.osterhus@uni.no

Ice–ocean interactions and circulation beneath the ice shelves fringing the Antarctic continent are of great climatic interest since they determine the basal melt rates, which influence (1) the ice-shelf mass balance, and thus its capacity to buttress the ice sheet feeding into them and (2) the production of cold, dense ice-shelf water, which is a source of Antarctic Bottom Water. The Filchner–Ronne Ice Shelf in the southern Weddell Sea, Antarctica, is the largest ice shelf by volume and a key site that contributes to bottom water formation. While basal melt rates are currently low, studies have suggested that they have the potential to increase dramatically in a future, warmer world, although the underlying processes are subject to large uncertainty. We operate a number of monitoring stations across the southern Weddell Sea. In 1999 we established site 5 on the Ronne Ice Shelf where access to the 402 m water column was gained through the overlying 763 m thick ice shelf using a hot-water drill. Results from the multiyear time series show the sensitivity of the sub-ice-shelf circulation to changes in conditions over the continental shelf and highlight the importance of monitoring the ice-shelf cavity. In the 2014/15 austral summer, site 5 was reoccupied and three instrumented moorings for long-term monitoring of the circulation beneath the Ronne Ice Shelf were deployed. In addition, three phase-sensitive radars (ApRES) were deployed at the snow surface to monitor the melting/freezing rate at the ice-shelf base. Here we present observations from site 5 for the period 1999–2015 and demonstrate the benefits of combining in situ oceanographic observations with the ApRES radar observations to resolve the temporal details in melting and freezing at the ice-shelf base


Long-term observing system for the oceanic regime of the Filchner–Ronne Ice Shelf, Antarctica

Svein Østerhus, Michael Schröder, Hartmut Hellmer, Keith Nicholls, Keith Makinson, Elin Darelius

Corresponding author: Svein Østerhus

Corresponding author e-mail: svein.osterhus@uni.no

Long-term observations of the flow of dense waters from their area of formation to the abyss of the world ocean, and the return flow of warm waters, are central to climate research. For the Weddell Sea an important component of such a system entails monitoring the formation of high-salinity shelf water (HSSW) on the continental shelf north of Ronne Ice Front, the transformation to ice-shelf water (ISW) beneath the floating Filchner–Ronne ice shelf, and the flux of ISW overflowing the shelf break to the deep Weddell Sea. Equally important is the return flow of warm water toward the Filchner–Ronne Ice Shelf system. We operate a number of monitoring stations in the southern Weddell Sea. The systems build upon techniques and methods developed over several decades and have a proven record of high data return. Here we present plans for extending, integrating and operating the existing long-term observatories to increase our knowledge of the natural variability of the ocean–ice-shelf system, and to allow early identification of possible changes of regional or global importance. The S2 observatory at the Filchner sill was established in 1977 and continues to deliver the longest existing marine time series from Antarctica. As a key site for monitoring the ISW overflow. The existing S2 observatory consists of subsurface mooring carrying sensors for current velocity, temperature, salinity and dissolved oxygen measurements. Site 5 at the Ronne Ice Shelf was first established in 1999 and in the 2014/15 austral summer the site was reoccupied and three instrumented moorings for long-term monitoring of the circulation beneath the Ronne Ice Shelf were deployed. In addition, three phase-sensitive radars (ApRES) were deployed at the snow surface to monitor the melting/freezing rate at the ice-shelf base. Some of the systems transmit in real time and are designed to operate for more than 10 years. In 2015/16 we will extend the observing network by deploying observatories on the Filchner Ice Shelf. The Filchner–Ronne Ice Shelf and S2 observatories will provide the first ever concurrent observations from the ice-shelf cavity where ISW is formed, and the sill where it starts its descent towards the deep Weddell Sea, and will provide a unique dataset allowing us to link processes and variability within the cavity directly to overflow properties and deep water formation.


Is the Larsen C ice shelf ungrounding from Bawden Ice Rise?

Helen Amanda Fricker, Fernando Paolo, Matthew Siegfried, Ted Scambos, Paul Holland, Adrian Luckman, Laurie Padman

Corresponding author: Helen Amanda Fricker

Corresponding author e-mail: hafricker@ucsd.edu

With the recent dramatic collapses of Antarctic Peninsula ice shelves, including Larsen A (1995) and B (2002), and the common understanding that these collapses are related to the southward expansion of a strong warming signal along the Antarctic Peninsula, much attention has been placed on the stability of the next ice shelf to the south (Larsen C; LCIS). Here we use a new continuous high-resolution record of surface-height change constructed from satellite radar altimeter (RA) data for 18 years from 1994 to 2012 to show that the onset of thinning on the LCIS has occurred progressively later from north to south, consistent with atmosphere-driven climate forcing. The highest observed thinning rates on Larsen C (local maximum of 16.6 ± 8.1 m decade–1) are found near Bawden Ice Rise (BIR). Various modeling studies have pointed out the stabilizing role that the ice rises (Bawden and Gipps) have on the LCIS and noted that, when contact with these is lost, the ice shelf will become less stable. An IceBridge radar profile across BIR shows that the ice is only ~40 m above floatation. If the surface lowering continues at the rate observed for 1994–2012, then the ice-shelf base will lift off BIR by ~2065, perhaps triggering rapid loss of ice-shelf volume at that time. We examine additional data up to 2015 (RA height data from CryoSat-2, velocity estimates from satellite imagery (Landsat and SAR), and long-term ice-front and rift changes from visible images) to assess recent dynamic changes near BIR.


Glaciological mass-balance measurements along oversnow traverses in West Antarctica

Francisca Bown, Andrés Rivera, Rodrigo Zamora, José Andrés Uribe

Corresponding author: Francisca Bown

Corresponding author e-mail: fbown@cecs.cl

The West Antarctic ice sheet is considered potentially unstable because its bedrock is well below sea level and the total disintegration could contribute significantly to global sea-level rise. The above, when taking into account ongoing global changes, especially oceanic warming in areas of the Southern Ocean, is leading to grounding line migration as clearly observed, for instance, in glaciers of the Amundsen Sea. In turn, dynamic responses of ice streams to coastal processes may have a large and rapid impact on ice acceleration and thinning, negative mass balance and ice divide migration. This latter is of the utmost importance for ice-sheet variability in the long term, where direct measurements could improve our understanding of ongoing and future changes. Several oversnow tractor traverses have been performed in West Antarctica by Centro de Estudios Científicos (CECs) during the last 10 years in order to perform glaciological and geophysical surveying over long distances. In austral summer of 2007/08, we carried out a terrestrial traverse from Patriot Hills towards the South Pole, where a stake network deployed every 20 km in a previous campaign in 2004 allowed us to survey the mass balance along 1100 km. Most of the stakes from 80° to 84°S were missing, probably due to wind drift and high ice flow rates, but in the remaining traverse, several stakes were gradually appearing accounting for decreasing accumulation with latitude and a mean value of 6.4 ± 3.9 cm w.e. More recently (summer of 2013 and 2014), oversnow surveys in West Antarctica have been performed from Union Glacier towards the Institute Ice Stream and Pine Island ice divide along total distances larger than 2000 km. This region has been described as stable in the last thousands of years, but significant changes taking place at Pine Island are thought to be spreading upstream with uncertain effects on the ice divides. As expected, accumulation rate at this ice divide yielded higher than at the South Pole traverse, accounting for 23 ± 5 cm w.e. These results seem consistent when compared with the Quantarctica database model.


Interannual variation of cyclonic eddy in the Amundsen Sea Polynya, Antarctica

Tae Wan Kim, Anna K. Wåhlin, Chang Sin Kim, Kyoung Ho Cho, Ho Kyung Ha, SangHoon Lee, Jae Hak Lee

Corresponding author: Tae Wan Kim

Corresponding author e-mail: twkim@kopri.re.kr

The Amundsen Sea has recently attracted particular interest because it is the most rapidly warming in the Western Antarctic (WA). Many studies reported that the relatively warm Circumpolar Deep Water (CDW) and its seasonal variation of intrusion are associated with the regional warming. However, few studies have examined why the intrusion varies seasonally. The strong seasonal cyclonic eddy in the Amundsen Sea Polynya was found from shipborne measurement LADCP data during the 2010/2011 and 2011/2012 IBRV Araon expeditions. Also this ocean surface circulation was reconfirmed from the OSCAR (Ocean Surface Current Analysis) what calculated from quasi-linear and steady flow momentum equations using sea surface height, surface vector wind and sea surface temperature obtained from satellites. A polynya larger than 200 km in diameter formed effect of wind and ocean currents in front of the DIS (Dotson Ice Shelf) during the austral summer season. The strong upwelling induced by latitudinal varying of westward wind velocities and sea-ice concentration at the northern boundary of the polynya. Such upwelling generates the cyclonic eddy that maximum current speed is larger than 10 cm s–1. However, during austral winter, the polynya was almost closed, so upwelling is reduced compared to the summer season and disappear the cyclonic eddy in front of the DIS. This seasonal variation in the strong ocean surface circulation may be influence on oceanic heat transport to ice selves.


Projections of ice-shelf basal melting and sub-ice-shelf circulation changes in a warming climate

Kaitlin Alexander, Katrin Meissner, Ben Galton-Fenzi, Matthew England

Corresponding author: Kaitlin Alexander

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

Ice shelves are essential to the stability of the Antarctic ice sheet, but a warming ocean puts them at risk. Further research into the possible response of ice-shelf–ocean interactions to climate change is therefore vital to better anticipate future changes in the Antarctic ice sheet and subsequent sea-level rise. Previous modelling of ice-shelf/ocean interactions mainly simulated present-day conditions or highly idealized warming scenarios. Here we instead derive atmospheric and oceanic boundary conditions from existing CMIP5 simulations of RCP (representative concentration pathway) scenarios. We apply these boundary conditions to a circumpolar regional ocean model including ice-shelf/ocean interactions (ROMS-Ice). The resulting changes in basal melting and circulation are examined for major ice shelves along the Antarctic coastline.


Effect of glacial drainage water on the CO2 system and ocean acidification state in an Arctic tidewater-glacier fjord during two contrasting years

Agneta Fransson, Melissa Chierici, Daiki Nomura, Mats Granskog, Svein Kristiansen, Tonu Martma, Gernot Nehrke

Corresponding author: Agneta Fransson

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

In order to investigate the effect of glacial water on the CO2 system in the fjord, we studied the variability of the total alkalinity (AT), total dissolved inorganic carbon (CT), dissolved inorganic nutrients, oxygen isotopic ratio (δ18O), and freshwater fractions from the glacier front to the outer Tempelfjorden on Spitsbergen in winter 2012 (January, March and April) and 2013 (April) and summer/fall 2013 (September). The two contrasting years clearly showed that the influence of fresh water, mixing and haline convection affected the chemical and physical characteristics of the fjord. The seasonal variability showed the lowest calcium carbonate saturation state (Ω) and pH values in March 2012 coinciding with the highest freshwater fractions. The highest Ω and pH were found in September 2013, mostly due to CO2 uptake during primary production. Overall, we found that increased freshwater supply decreased Ω, pH and AT. On the other hand, we observed higher AT relative to salinity in the freshwater end-member in the mild and rainy winter of 2012 (1142 μmol kg–1) compared to AT in 2013 (526 μmol kg–1). Observations of calcite and dolomite crystals in the glacial ice suggested supply of carbonate-rich glacial drainage water to the fjord. This implies that winters with a large amount of glacial drainage water partly provide a lessening of further ocean acidification, which will also affect the air–sea CO2 exchange.


Interaction of the Weddell Sea continental shelf with the Antarctic coastal current and the Antarctic slope front – an idealized model study

Kjersti Daae, Elin Darelius, Ilker Fer, Tore Hattermann

Corresponding author: Kjersti Daae

Corresponding author e-mail: kjersti.daae@uib.no

Future climate scenarios suggest an increased flow of warm water towards the Filchner–Ronne Ice Shelf cavity as the coastal current is redirected onto the continental shelf where the shelf opens up and widens at about 10W. We use an idealized, eddy resolving numerical model to study the dynamics of the Antarctic coastal current and the Antarctic slope front as the flow encounters the widening continental shelf representative of the southern Weddell Sea. A 1700 m deep narrow shelf channel is connected to a 400 km wide and 400 m deep shelf with a continental slope of 0.026. The channel is forced with idealized wind stress and zonally periodic boundary conditions. Mean and eddy driven cross-shelf transport will be analyzed in experiments with alternate wind forcing and different water mass on the continental shelf, in order to study the processes controlling the access of Warm Deep Water at the Filchner Trough.


Seasonal variability of the Antarctic slope front and implications for onshore eddy heat transport in the southeastern Weddell Sea

Tore Hattermann, Peygham Gaffari Nooran, Qin Zhou, Pål Erik Isachsen, Elin Darelius

Corresponding author: Tore Hattermann

Corresponding author e-mail: tore.hattermann@awi.de

The Weddell Sea Antarctic slope front efficiently separates warmer Southern Ocean Warm Deep Water from the coast. Average ocean temperatures on the continental shelf are close to the surface freezing point and ice-shelf basal melting is generally rather low in this sector of Antarctica. But observations also show an intermediate access of warmer water onto the continental shelf and models suggest that the onshore heat transport is controlled by a subtle frontal balance at the shelf break that might undergo dramatic changes under future global warming. In order to better understand those dynamics, we analyze data from a set of oceanographic moorings in the southeastern Weddell Sea. We find a distinct seasonal variability of the shelf break current that relates to hydrographic changes caused by sea-ice melting and freezing. During summer, the wind-driven accumulation of buoyant Antarctic Surface Water on the continental shelf leads to a steepening of isopycnals near the coast, which enhances the vertical shear in the upper part of the water column and affects the stability properties of the coastal current. Linear growth rates and associated scales are obtained using multiple-layer shallow-water formulations, in order to study the cross-slope eddy heat transport in the seasonally varying frontal structure and assess the dynamic response to possible future climate change.


Four years at Coulman High: what do a long term mooring and a model tell us about the oceanography and ice shelf melt near the front of the Ross Ice Shelf?

Mike Williams, Stefan Jendersie, Craig Stewart, Robin Robertson, Pat Langhorne

Corresponding author: Mike Williams

Corresponding author e-mail: mike.williams@niwa.co.nz

An oceanographic mooring has been deployed near the front of the Ross Ice Shelf at Coulman High (30 km east of Ross Island and 7 km back from the ice front) since December 2010. In addition to oceanographic instrumentation the mooring was fitted with an upward looking sonar to directly measure ice shelf melt. During summer, melt rates average 2.7 m a–1, but the relatively high background melt rate of 1 m a–1 throughout winter causes approximately half of the observed ablation of the ice shelf base. The ocean current data shows the mean northwest flow out of the cavity, but the currents are highly variable. About half of the variance is due to the diurnal tides. At longer timescales the velocities show a distinct seasonal component with southward flow in late summer and autumn and northwest flow, consistent with the annual mean, over the remainder of the year. We also find seasonal variation in the ocean circulation on the Ross Sea continental shelf in a Regional Ocean Model System (ROMS) climatological simulation of the Ross Sea. Here we will utilise the ROMS model output to understand what drives the seasonal variability observed at Coulman High, the implications this has for understanding melt patterns along the front of the Ross Ice Shelf, and potential pathways for warm water intrusions, with the objective of determining if we have monitoring in place to observe these intrusions.


The distribution of freshwater from meteoric sources and sea ice melt across the western and northern fjords of Svalbard

Lewis Drysdale

Corresponding author: Lewis Drysdale

Corresponding author e-mail: antysd43@bas.ac.uk

Accelerated retreat of glaciers in regions of the Arctic, such as Svalbard, causes the release of excess freshwater to coastal estuaries. This freshwater has a stratifying effect on the water column, thus influencing processes such as shelf exchange, heat delivery to the surface, and biological productivity. The freshwater budgets of high latitude fjord systems, however, are poorly understood, with multiple sources (including also sea ice melt and direct precipitation) contributing. The use of stable isotopes of oxygen as a water mass tracer, when measured alongside salinity, is a powerful technique for quantitatively decomposing the freshwater budget. We present a new record of glacial melt and sea-ice melt distributions from three different fjords around Svalbard using full depth profiles of oxygen isotopes. We find that in autumn 2013, glacial meltwater is widespread in both Isfjorden and Billefjorden where there has been a strong influence of Atlantic water, while in 2014 sea ice melt was the dominant source of freshwater in to the north of the archipelago in Rijpfjorden. These results will improve the representation of freshwater in simple models of the region.


Simulating the dynamics of Helheim Glacier with a crevasse-depth calving criterion and 2D ice-flow model

Andrew Taylor, Stephen Cornford, Anthony Payne

Corresponding author: Anthony Payne

Corresponding author e-mail: a.j.payne@bristol.ac.uk

Between 2000 and 2005, Helheim Glacier, one of Greenland’s largest outlet glaciers, retreated around 7 km inland, with the period of most rapid change occurring between 2004 and 2005. Subsequently, the glacier re-advanced 3 km to its present-day location over the course of the next few years. This change in front position was accompanied by an increase in surface velocity at the terminus from 8 km a–1 in 2000 to 11 km a–1 in 2005 and a subsequent decrease to 8.5 km a–1 in 2006. These changes have been successfully reproduced in a one-dimensional flowband model using a front-stress perturbation, but have not been repeated using a calving model such as a crevasse-depth criterion or with higher-dimensional models. With this in mind, the two-dimensional, finite volume ice sheet model, BISICLES, and an extension of the crevasse-depth calving model to two dimensions is used to model calving events on Helheim Glacier. To create a two-dimensional glacier domain, ice thickness, surface elevation and bedrock topography taken from the 500 m DEM product produced by CReSIS and surface velocity from 2005 to 2006 are used with BISICLES to generate basal traction and ice rheology fields. In a further inverse problem, individual flightlines of ice thickness are used to additionally solve for mass conservation. By keeping surface elevation constant, basal topography is adjusted and a second DEM is created. The results of the two inverse problems are used to generate two initial conditions located at the observed 2001 front position, following which different formulations of the crevasse-depth calving model are implemented to determine their sensitivity to model parameters and ability to reproduce observed results. Results from both initial conditions show that observed rates of retreat, front position progression and changes in surface velocity can be reproduced. Topography influences the locations at which the Helheim is more stable, and the modelled calving fronts match observations of retreating glacier. Rates of retreat and calving front shapes are dependent on calving model structure and model parameters.