FRISP and WAIS Workshop 2008

Abstracts

---

A model of ocean processes in the Bellingshausen Sea

Paul R Holland¹, Adrian Jenkins¹ and David M Holland²

A regional sea ice-ocean-ice shelf model is applied to the Bellingshausen Sea, a region that oceanographically is coarsely observed. The model is based on MICOM and CICE and forced by NCEP/NCAR reanalyses for the period 1979-2007. The model reproduces both sparse 'snapshot' ocean observations and a simulation-length time series of monthly sea ice concentrations, so it may be used to investigate ocean variability and its impact on the nearby ice shelves. The model suggests that a steady mid-depth eastward flow along the continental shelf is opposed by westward flow at depth offshore of the shelf break. A strong westward surface coastal current appears in summer and autumn, but weakens in winter and spring as the wind direction changes and the presence of sea-ice redistributes the surface forcing. Barotropic flow through George VI Sound oscillates in direction but is northward on average. The model suggests a relatively-constant presence of warm Circumpolar Deep Water on-shelf, leading to simulated George VI Ice Shelf melt rates of a magnitude that, in agreement with observation, suggest that the ice shelf is out of balance and losing mass. Other ice shelves in the region are poorly resolved.

---

Ice-sheet variations as depicted in seismic records of the Amundsen Sea Embayment

Estella Weigelt¹, Karsten Gohl¹, Gabriele Uenzelmann- Neben¹ and Rob Larter²

New multichannel seismic reflection data of high-resolution offer a record of the glacial development and processes in the western Amundsen Sea Embayment. We present our interpretation of five seismic profiles collected in this area as part of the RV Polarstern cruise ANT-XXIII/4 (2006).
The sea-floor morphology of the observed area is highly variable. Close to the present coast, a rough topography includes troughs of several 100 m depth which were probably formed by palaeo-icestreams cut through the shelf. Here, the seismic lines reveal hardly any internal sequences. Only some thin sediment pockets (< 80 ms TWT ~ 60 m) in between these steep and rugged structures can be identified.
On the middle shelf and outside the troughs, the topography is generally smooth and only shows small surface undulations suggesting a quieter deposition realm which was not directly affected by ice streams. Hard surface matter as revealed in parasound sections indicates the flush away of soft sediments by melt-water streams. Such streams might have also formed the shallow and about 10 km wide channel-like structures observed at some locations.
The most interesting features are pronounced northwest-dipping reflector series of more than 1 s TWT thickness (> 800 m), indicating well layered sedimentary units. The strata reveal a striking alternation of reflection-poor, almost opaque, units and sequences of closely spaced, continuous reflectors. This alternation probably represents changes in depositional conditions due to sea level variations or periods of ice-sheet extension. Reflection-poor units probably consist of a homogeneous sedimentary matrix with uniform physical properties. An extended ice-shelf cover may have led to the deposition of diamictites and till. In contrast, high-amplitude reflectors result from marked changes of sedimentary densities and P-wave velocities which are the parameters controlling the seismic reflectivity. During these times, open water conditions might have prevailed which allowed a stronger diversification of deposits. Unfortunately, no age control by drill sites is available for this region. However, borehole records of the ANDRILL project in the western Ross Sea show a similar sedimentary pattern, that is why we tentatively suggest at least 10 cycles of ice-shelf expansions and retreats since the middle Miocene in the western Amundsen Sea Embayment.

---

A Mid-Pleistocene depositional anomaly in Antarctic continental margin sediments: Evidence for a collapse of the West Antarctic Ice Sheet?

Claus-Dieter Hillenbrand¹, Gerhard Kuhn² and Thomas Frederichs³

Modern global warming may trigger a collapse of the West Antarctic Ice Sheet (WAIS), which would raise global sea level by up to 5-6 meters. Despite the importance of the WAIS for sea-level changes, its response to the Pleistocene climatic cycles is poorly constrained, and the geological evidence for a WAIS disintegration within the last 1 Myr is ambiguous. Here we compare two marine sediment cores recovered from the West Antarctic continental margin with two cores from the East Antarctic margin. The age models of the cores are based on magnetostratigraphy, lithostratigraphy and/or foraminiferal oxygen isotope stratigraphy and show that the recovered sequences span >800 kyr. Within the sediment cores proxies for biological productivity and the supply of lithogenic detritus from the Antarctic hinterland exhibit cyclic fluctuations in accordance with the glacial-interglacial cycles of the Pleistocene. Discrepancies in these fluctuations between the two study areas and cores from the same area are minor and attributed to regional and/or local environmental impacts on sediment deposition. A prominent depositional anomaly observed at all four sites spans Marine Isotope Stage (MIS) 15 to MIS 13 (621 ka to 478 ka before present). In particular on the West Antarctic margin proxies for productivity and lithogenic sediment supply indicate that the interval MIS 15-13 has the characteristics of one, prolonged interglacial period. In the cores from the West Antarctic margin no proxy suggests environmental conditions much different from today, but if the WAIS collapsed during the last 800 kyr, then MIS 15-13 seems to be the most likely time period. A comparison with various marine and terrestrial climate archives from around the world reveals that unusual environmental conditions prevailed during MIS 15-13. We speculate that some of these anomalies are a direct consequence of major ice-sheet drawdown in Antarctica.

---

Initial Surface Observations and New Inferences of the Pine Island Glacier Ice Shelf

Robert Bindschadler¹, David Holland², David Vaughan³ and Patricia Vornberger¹

A new field program has begun on the ice shelf fed by the Pine Island Glacier, West Antarctica, a glacier experiencing dramatic thinning, acceleration and retreat. The spatial pattern of change suggests the cause is related to interaction between the ice shelf underside and warm water from the Southern Ocean. A quick summary of the reconnaissance work conducted in 2007-08 will be followed by a report of new analyses of recent airborne geophysical data, satellite altimetry and optical imagery. A rather sudden and dramatic shift in the sub-shelf circulation pattern seems to have occurred in 1999 as evidenced by the persistence of three distinct polynyas at the ice shelf front. In addition, the study of a dramatic surface wave pattern supports the inference that there is a strong seasonal modulation to sub-shelf melting rates.

---

Improved bathymetry from iceberg groundings in the western Weddell Sea

Adrian Luckman¹, Laurie Padman² and Daniela Jansen¹

Measurements of bathymetry around Antarctica are important for understanding past glacial extent and for modelling ocean circulation and its influence on the marine cryosphere (ice shelves and sea ice). In the western Weddell sea, bathymetric soundings are sparse because of persistent sea ice cover and there is considerable uncertainty in the possible pathways for warm water from the adjacent deep ocean to the front of the Larsen C Ice Shelf (Nicholls et al., 2004). Consequently basal melt rates under the shelf, which appear to be determined by tidal currents, remain poorly constrained.
In this study, we use multiple Envisat ASAR observations to detect regions of persistent iceberg groundings in the western Weddell Sea, which we then interpret as submarine ridges (Viehoff and Li, 1995). While single SAR images cannot easily discriminate between grounded bergs, drifting bergs and sea ice, mean backscatter from the multiple Envisat Global Monitoring Mode images allows stationary features to be readily detected. Two large (50 x 50 km) regions of persistently grounded bergs are detected in this way and Envisat Wide-Swath Mode data is used to measure the iceberg planimetric dimensions. ICESat laser altimetry data are used to estimate the maximum draft of a sample of grounded icebergs and infer the depth in this region.
Results confirm submarine ridges of less than 300 m water depth in a region where the best available bathymetry grids report depths of between 400 and 500 m. Observations of similar persistent groundings elsewhere off Antarctica may yield further improved bathymetric information.

REFERENCES

Nicholls K. W., C. J. Pudsey, P. Morris (2004), Summertime water masses off the northern Larsen C Ice Shelf, Antarctica, Geophys. Res. Lett., 31, L09309, doi:10.1029/2004GL019924.

Viehoff, T., Li, A., (1995), Iceberg observations and estimation of submarine ridges in the western Weddell Sea, International Journal of Remote Sensing 16(17): 3391-3408.

---

Regional sea level rise from a collapse of the West Antarctic Ice Sheet

Jonathan Bamber¹, Riccardo Riva² and Bery Vermeersen²

Theory has suggested that the West Antarctic Ice Sheet may be inherently unstable. Recent observations lend weight to this hypothesis. We reassess the potential contribution to eustatic and regional sea level from a collapse of the ice sheet and find that previous assessments have substantially overestimated its likely contribution by 80-100%. We obtain an upper estimate for the eustatic sea level contribution of 3.1 m, with important regional variations. The maximum impact is concentrated along the Pacific and Atlantic seaboard of the US. In this region the rise is about 20% larger than the global mean, even for the case of a partial collapse.

---

First (1957-58) Geophysical-Glaciological Investigation of the Filchner-Ronne Ice Shelf System Made During the International Geophysical Year (IGY)

John C Behrendt^1,2

The only major field project of the U.S. International Geophysical Year (IGY) Antarctic program was a series of oversnow traverses (Behrendt, 1998; 2003) mostly in West Antarctica, starting in 1957, making seismic reflection ice soundings (and other geophysical measurements) and glaciological measurements. The Filchner Ice Shelf Traverse of 1957-58 made the initial glaciological and geophysical observations of the Filchner-Ronne Ice Shelf on a 1900-km oversnow traverse based at Ellsworth Station on the Filchner Ice Shelf front that acquired (among various data) surface elevation and seismic reflection determinations of ice shelf thickness (a maximum 1300 m) and depth to bedrock and made a geological reconnaissance of the Dufek Massif. Results included the preliminary mapping of the Filchner-Ronne Ice Shelf and its vast area, a maximum thickness of 1300 m, the discovery of Berkner Island, Henry Ice Rise and Korff Island, and the definition of the Thiel trough beneath the Filchner Ice Shelf. Observed thinning of the Filchner Ice Shelf from ~700 m in the area of the former Grand Chasm to 250 m at the ice front in about 130 km was interpreted as due to a melt rate of ~9 m per year in this area.
In 1965-66 the U.S. Geological Survey made seismic reflection measurements of ice shelf thickness and depth to bedrock southwest of the Filchner Ice Shelf Traverse data in the area south of Berkner Island-Henry Ice Rise (Behrendt et al., 1966). However, these sparse surveys remained the only coverage until the 1994-95 seismic traverse of Johnson and Smith (1997). I compare this data set with the earlier data and infer measurable thinning of the ice shelf during the 1957-1995 interval.
Ideally I would compare seismic reflection soundings at the same locations, but the seismic data are not located in the same positions in the 1957-58 and 1994-95 surveys. The 1957-58 seismic measurements of ice shelf thickness were used to control calculations of ice shelf thickness (Behrendt, 1962)from barometric elevations assuming hydrostatic equilibrium for floating ice over a broad area of the Filchner-Ronne ice shelf.
The 1957-58 data show a maximum ice shelf thickness of 1300 m south of Henry Ice Rise, the result of ice entering the Filchner-Ronne Ice Shelf System from the East Antarctic Ice Sheet via the Foundation Ice Stream. Direct comparison of the 1957-58 (Behrendt, 1962) and 1994-95 (Johnson and Smith, 1997) contour maps of ice shelf thickness show the greatest thickness of ~1100 m in the same area in the 1994-95 data. The apparent thinning indicated is >100 m in the intervening 37 years. I assume that the errors are solely in the 1957-58 contour map because of the errors in barometric elevation used to calculate the floating ice thickness. However the apparent >100-m thinning must be greater than any possible error I can identify. I conservatively estimate a minimum thinning of 50 m from 1957-58 to 1994-95.

---

Ice shelf morphology and the efficiency of basal melting

Christopher Little¹, Anand Gnanadesikan² and Michael Oppenheimer^1,3

Ice shelf basal melting is strongly influenced by turbulent mixing through meltwater-freshened oceanic layers in the uppermost part of the water column. Buoyancy-driven, boundary-trapped currents in these locations are dynamically similar to downslope flows, in which mixing has been shown to be controlled by local bathymetry. However, basal melting is a function of the entrainment of warmer water from below and heat flux through a viscous sub-layer. Here, idealized numerical simulations are analyzed to determine how ice shelf morphology drives spatial decorrelation of these heat fluxes, and thus the distribution and efficiency of basal melting.
Near grounding lines, the slope of ice shelves top and basal surfaces may be an order of magnitude larger than tens of kilometers down-glacier. Idealized model configurations representing these large-scale morphologic constraints reveal three slope-driven, spatially distinct, dynamic regimes in the sub-ice shelf mixed layer. Steep basal slopes drive entrainment of heat in initiation regions. Heat is advected upslope and used to melt ice in the maintenance region; however, flow convergence in the outflow region limits heat flux to flatter portions of the ice shelf. The ice shelf thickness gradient thus controls the area-integrated efficiency (the fraction of entrained heat used to melt ice). In these simulations, efficiency is a function of ice shelf shape, turbulence parameterizations, and the ocean temperature. One robust result is a loss of efficiency as the ocean temperature is increased, providing insight into the response of basal melting to temperature described in previous modeling efforts.

---

Structure and Variability of the Filchner Overflow Plume

Elin Darelius^1,2, Lars H. Smedsrud², Svein Østerhus², Arne Foldvik¹ and Tor Gammelsrød¹

Properties of the dense Ice Shelf Water plume emerging from the Filchner Depression in the southwestern Weddell Sea are described, using available current meter records and CTD-stations. A mean hydrography, based on more than 300 CTD stations gathered over 25 years is used. This data points to a cold, relatively thin and vertically well defined plume east of two ridges crosscutting the continental slope about 60 km from the Filchner sill. The dense bottom layer is warmer, more stratified, and much thicker west of these ridges. The data partly confirm the three major pathways suggested earlier and agrees with recent theories on topographic steering by submarine ridges. A surprisingly high mesoscale variability in the overflow region is documented and discussed. The variability is to a large extent due to three distinct oscillations (with periods of about 35 hours, 3 days and 6 days) seen in both temperature and velocity records on the slope. The oscillations are episodic, barotropic and have a horizontal scale of ~20-40 km across the slope. They are partly geographically separated, with the longer period being stronger deeper down. Energy levels are lower west of the ridges, and in the Filchner Depression. The observations are discussed in relation to existing theories on eddies, commonly generated in plumes, and in continental shelf waves.

---

Acceleration of Jakobshavn Isbrae Triggered by Warm, Subsurface Irminger Waters

David M. Holland¹, Robert H. Thomas², Brad deYoung³, Mads H. Ribergaard⁴ and Bjarne Lyberth⁵

Observations over the past decade show a rapid acceleration of several outlet glaciers in Greenland and Antarctica. Among the largest changes seen, Jakobshavn Isbrae (JI), an outlet glacier feeding into a deep ocean fjord on the west coast of Greenland, recently, and suddenly, switched its behavior from slow thickening prior to 1997 to subsequent rapid thinning and a doubling in glacier velocity. Suggested reasons for the JI speedup range from increased lubrication of the ice-bedrock interface as more meltwater drains to the bed during recently warmer summers, to weakening and breakup of the floating ice tongue. Here, we present evidence that the changes of the JI were triggered by an increase in subsurface ocean temperature, based on hydrographic data showing a sudden jump during 1997 along the entire west coast of Greenland. This arrival of upstream Irminger Sea warm water was driven by changes in regional atmospheric circulation.

---

Summary and review of 'Circulation in Lake Vostok: A laboratory analogue study' by M.G. Wells and J.S. Wettlaufer, GRL (2008)

Malte Thoma

Laboratory studies can complement field work and modelling. I present some interesting results of the recent publication of M.G. Wells and J.S. Wettlaufer in Geophysical Research Letters (2008). This presentation also briefly referes with a wink to GRL-policies.

---

Subglacial Lakes in Antarctica: Numerical modelling of temperature distribution and accreted basal ice

Malte Thoma^1,2, Klaus Grosfeld¹ and Christoph Mayer²

Subglacial lakes are a widespread phenomena across the Antarctic Ice Sheet. These lakes are isolated from direct exchange with the atmosphere by several kilometers of ice since millions of years and provide unique environments for potential life forms. The hydrological role of the lakes is still not clear, but there are indications for an active hydrological system underneath the ice sheet. Their inaccessibility increases the importance of numerical models to investigate the physical conditions in these environments. Hydro-dynamical considerations reveal that the high pressure leads to convective flow in the lakes, as long as the overlying ice thickness exceeds about 1500 m. Using a three dimensional numerical model and a compilation of the best available geometry, we analyse the flow regime and the basal mass balance within two Antarctic subglacial lakes, Lake Vostok and Lake Concordia in East Antarctica. Both lakes are targets for water sampling in the near future which demands a good knowledge of potential in situ conditions. Results demonstrate that both lakes show a permanent mass gain from melting ice, indicating a permanent flux of water across the lake boundaries. In combination with observed ice velocities we are able to calculate the thickness and distribution of accreted ice at the ice-lake interface which also is a crucial parameter for sampling strategies.

---

Use of adaptive mesh refinement in ice sheet models

Rupert Gladstone¹, Victoria Lee¹, Tony Payne¹ and Andreas Vieli²

Grounding line migration is a key process affecting the stability of marine ice sheets such as the West Antarctic ice sheet (WAIS). Vieli and Payne [2005] test existing models and conclude that no reliable model of grounding line migration is currently available. In particular, they find that models which employ a fixed spatial grid can not be used to solve this problem in a robust manner but that moving-grid models (which explicitly track grounding line migration) show a great deal of promise. While the latter type of model is relatively easy to implement in one dimension, it would be very difficult to employ in full, three-dimensional ice sheet models. One possible solution is the use of AMR (adaptive mesh refinement) technique, which is based on a nested set of fixed-grid models. Here we use this methodology in a 1D model to address this issue by providing greatly increased resolution in the vicinity of the grounding line. We have developed a 1D ice sheet model in which sub regions of the whole model domain (and sub regions thereof, etc) are run at higher resolutions, forced at their boundaries by data from the encompassing lower resolution region, allowing very high resolution where it is needed without the high computational costs of running the whole domain at high resolution. The higher resolution regions evolve during the simulation according to predefined criteria (e.g. crossing a given threshold in estimated truncation error or proximity to grounding line). This 1D AMR ice sheet model shows close agreement with the analytical solution for a simple ice sheet simulation. The model will be used to attempt to simulate grounding line migration in ice sheet/ice shelf simulations.

Reference:
Vieli, A and Payne, A.J., 2005, "Assessing the ability of numerical ice sheet models to simulate grounding line migration", J. Geophys. Res 110 F01003

---

Generation of a dense coastal current by an Antarctic polynya

Alexander Wilchinsky¹ and Daniel Feltham^1,2

Descent and spreading of high salinity water generated by salt rejection during sea ice formation in an Antarctic coastal polynya is studied using a hydrostatic, primitive equation three-dimensional ocean model, POLCOMS. The shape of the polynya is assumed to be a rectangle 100 km long and 30 km wide, and the salinity flux into the polynya at its surface is constant. The model has been run at high horizontal spatial resolution (500 m) and numerical simulations reveal a buoyancy-driven coastal current. The coastal current is a robust feature and appears in a range of simulations designed to investigate the influence of a sloping bottom, variable bottom drag, variable vertical turbulent diffusivities, higher salinity flux, and an off-shore position of the polynya. It is shown that bottom drag is the main factor determining the current width. This coastal current has not been produced with other numerical models of polynyas, which may be because these models were run at coarser resolutions. The coastal current becomes unstable upstream of its front when the polynya is adjacent to the coast. When the polynya is situated off-shore an unstable current is produced from its outset due to the capture of cyclonic eddies.

---

Subglacial Lakes: Modelling the interaction between ice- and water flow

Malte Thoma^1,2, Klaus Grosfeld¹, Christoph Mayer² and Frank Pattyn³

Subglacial lakes are a widespread phenomena across the Antarctic Ice Sheet. We shortly introduce the challenges Antarctic subglacial lakes hold for the scientific community, followed by a brief history of subglacial water flow modelling and ice flow modelling. In order to join these modelling efforts we finally present first results of a coupled ice - lake system which will help to understand the complex interactions between the Antarctic ice sheet and subglacial lakes.

---

Large Eddy Simulation of Gravity Currents

Tamay M. Ozgokmen

In light of the importance of stratified mixing processes for both coastal and large-scale ocean circulation, the feasibility of LES is explored as a small-scale process tool.
A 3D lock-exchange problem is put forth as a useful canonical problem to test the accuracy of mixing using LES, because it contains shear-driven mixing, internal waves, interactions with boundaries and convective motions, while it is very simple and unambiguous to set up regarding the domain, initial and boundary conditions, and forcing.
A DNS is used as the benchmark solution, and under-resolved DNS is performed on three coarse meshes, for which the CPU time is 2,292 to 14,864 times shorter. We experiment with two general classes of SGS models, namely eddy-viscosity (EV) models and an approximate deconvolution model, based on the rational approximation of a Gaussian filter.
The main measure of mixing is taken as the background potential energy.
Based on the solutions from 56 experiments involving different SGS models and parameter combinations, it is noted that the dynamic EV and rational models have different strengths in that the former is good in providing appropriate dissipation while the latter in preserving the detail of coherent structures on coarse resolution meshes. The hybrid model, namely the superposition of both models, shows a good agreement with the time evolution of the mixing curve from DNS, while being computationally 1,212-fold faster.

---

Modeling the Contribution of Tides to Basal Melt Rate for Larsen-C Ice Shelf

Rachael Mueller^1,2, Laurie Padman² and Helen A. Fricker³

The Larsen-C is showing pre-collapse signs of retreat and thinning, but our understanding of the processes affecting its evolution is still too poor to predict whether it will ultimately collapse like Larsen-A and Larsen-B. Among these poorly understood processes is basal melting at the ice-ocean interface. We hypothesize that basal melt under Larsen-C is sensitive to the strength of tidal currents, as a source for turbulent mixing of warmer ocean water up to the ice shelf base. This hypothesis is explored through comparison of basal melt rates obtained with several runs of the Regional Ocean Modeling System (ROMS 2.2), which includes explicit thermodynamic coupling between the ice-ocean interface. We use simplified ice shelf geometry and initially homogeneous ocean stratification. Tidal forcing is simulated by an idealized coastal-trapped Kelvin wave at the M2 tidal frequency. By varying the amplitude of the Kelvin wave, we explore both the change in basal melt rate at different stages of a hypothetical spring/neap tidal cycle as well as the sensitivity of melt rate to uncertainties in water column thickness (w.c.t.) under the shelf. Our results show that basal melt rate is strongly affected by tidal strength. Increasing the Kelvin wave amplitude from 0.1 to 5.0 times the value of the base case, which correspond to the real M2 tides near Larsen-C, increases the shelf-averaged basal melt rate from ~0.5 m/yr to ~5 m/yr. The base case scenario (with tidal elevation of 0.8 m and velocity of ~0.07 m/s) corresponds to averaged melt of 2 m/yr, comparable to the thinning observed over much of the northern portion of the shelf with ERS radar altimetry. These results indicate the need to include tides in coupled ocean/ice-shelf models and demonstrate the importance of mapping w.c.t. under ice shelves due to its influence on tidal current speeds, hence, mixing and basal melt.

---

Antarctic Ice Shelf Melting in the 21st Century, "A Model Study"

Hartmut H. Hellmer, Frank Kauker and Ralph Timmermann

It has been proposed that huge ice shelves might face less basal melting in a warmer climate due to less sea ice formation on the continental shelf and thus a reduced density gradient between ice shelf front and the cavern interior. We present the results of a 100-year simulation with BRIOS-2.2 forced with the ECHAM5-MPIOM output for the IPCC-A1B scenario. The results show that basal melting enhances for all ice shelves, but the enhancement is minor for huge ice shelves like Filchner-Ronne and Ross while the smaller ones like Fimbulisen and Getz are threatened by an up to 100% increase. Further analysis reveals that a decrease in salinity (due to a reduced sea ice cover) in parallel to slightly higher temperatures on the broad continental shelves are responsible for a mitigated response of the huge ice shelves to climate warming. In contrast, minor decreases in salinity in combination with increased near-bottom temperatures on the narrow continental shelves cause smaller ice shelves to be highly vulnerable to a warmer climate.

---

Physical conditions in Subglacial Lake Ellsworth

A.M. Smith¹, J. Woodward², N. Ross³, M.J. Siegert³, H.F.J. Corr¹, R.C.A. Hindmarsh¹, E.C. King¹, D.G. Vaughan¹ and M.A. King⁴

Subglacial lakes are regions of ice sheet overlying substantial bodies of water. They are of particular interest for any life they contain, for their record of ice sheet history and for their potential impact on ice dynamics. So far, none has been accessed directly, but Subglacial Lake Ellsworth (SLE) has been identified as the most suitable target in West Antarctica where access should proceed. We have completed a geophysical reconnaissance of Subglacial Lake Ellsworth and present the first results.
Seismic, radar and GPS surveys were conducted over SLE between November 2007 and February 2008. The radar surveys mapped the lake outline and the ice thickness in the region, and showed the internal structure within the ice sheet. GPS receivers measured the ice flow over the lake and the surrounding area. The seismic data gave information on the water depth, the ice-water interface and the bed of the lake itself, and we concentrate on these findings. The aims of the seismic surveys were to determine the lake bathymetry and to give some indication of both the physical conditions within the lake, and the nature and structure of any sediments at the lake floor.
The lake surface covers an area approximately 22.5 km2. Maximum water depth is more than 150 m, showing that SLE is a substantial body of water. The ice-water interface shows no evidence for freeze-on of basal ice, indicating that the ice is melting at its base over the whole of the lake. Hydrostatic analysis indicates that bridging stresses influence the floatation level over virtually all the lake. In places the ice is supported above its hydrostatic equilibrium level; elsewhere it is depressed below this level. The lake bed is composed of high-porosity, low-density sediments with acoustic properties very similar to material found on the deep ocean floor. Seismic reflections indicate a substantial thickness of this soft sedimentary material, accumulated at the lake bed in a low-energy environment. These results have implications for subglacial conditions and ice sheet history, as well as significant practical implications for lake access operations, including access location, preparations for break-through and expectations for the bed sedimentary sequence.

---

Glacial history of the Ellsworth Mountains, West Antarctica: Implications for the source of Meltwater Pulse 1A

Michael Bentley^1,2, Christopher Fogwill³, Anne Le Brocq¹, Alun Hubbard⁴ and David Sugden⁵

The source of meltwater pulse 1A has been debated for several years, with studies variously ascribing it to a dominant Laurentide source, a dominant Antarctic source, or some combination of both. Part of the difficulty in defining the source region has been the relative lack of glacial geologic observations from Antarctica that can constrain former ice sheet volume during this interval. We report here new geomorphologic observations and cosmogenic surface exposure dates for erratic boulders that together constrain the glacial history of the West Antarctic Ice Sheet where it flows across the Ellsworth Mountains in the Weddell Sea embayment. This is a critical sector of West Antarctica as it arguably represents the largest unknown component of the post-LGM Antarctic ice volume (sea level contribution) budget. Our data show that the ice sheet surface reached its Last Glacial Maximum elevation ~400-500 m above present-day ice. This contrasts to some previous estimates of > 800 m. The limit was occupied at 15 ka BP. Thinning from this limit was progressive throughout the Holocene with no evidence of rapid thinning through the meltwater pulse 1A interval, within the resolution of our data.
We have applied an ice sheet model to quantify the excess volume of ice in the Weddell Sea embayment, constrained by recently acquired field evidence from the Ellsworth Mountains, Shackleton Range and Berkner Ice Core. This model suggests that the Weddell Sea embayment contributed only ~1-3 m of sea level since the Last Glacial Maximum, and so implies that the region could not have provided the dominant contribution to meltwater pulse 1A.

---

Data Integration and Shifting Glacier Flow in the WAIS Amundsen Sea Embayment

Frank O. Nitsche and Suzanne M. Carbotte

Studies of the West Antarctic Ice Sheet (WAIS) involve such disciplines as glaciology, remote sensing, geology, oceanography and geophysics, each producing different types of data. Integrating and visualizing the different data sets can provide new insights that might not be identified by single disciplines. Various databases and interfaces have been developed to access different data sets, maximize their use, and avoid unnecessary redundancy of data collection. Here we use the NSF sponsored GeoMapApp interface and the underlying Antarctic Multibeam Bathymetry and Geophysical database, to integrate and compare bathymetry and ice flow information in the Amundsen Sea and adjacent sections of the WAIS. A new bathymetric map of the Amundsen Sea continental shelf, based on the database and other sources, defines the shelf break and two major trough systems that dominate the continental shelf. One trough emerges from the Pine Island, Thwaites and Smith Glaciers, and the other from the eastern Getz and Dotson ice shelves. Both are similar in size and depth, and show clear indications of formation or at least modification by grounded ice streams. The trough morphologies suggest similar ice burdens and movements. However, the greater Pine Island Bay Trough now leads to large ice drainage basins that together accommodate about a third of the WAIS outflow, while the Getz-Dotson-Trough is only linked to much smaller glaciers. If ice streams of similar strength/magnitude were necessary to form these troughs, their similar features would indicate a different drainage pattern in the past, with larger ice streams then flowing through the Getz-Dotsen Trough. Such ice streams would have drained a much larger portion of the WAIS than today, denoting shifts in drainage patterns during or between glaciations. Analyzing the characteristics of troughs on the continental shelf and their connections to present and modeled past ice flow could be an important tool for verifying and calibrating reconstructions of long-term ice sheet behavior.

---

Sensititity study of the influence of islands on the ice cavity circulation using unstructured meshes

Adriana Huerta-Casas and Hartmut Hellmer

Most of the current ocean models that include ice cavities employ structured grids and are well established. This work presents the first attempt to introduce ice cavities in an unstructured grid model, the Finite Element Ocean Model (FEOM). As a first step, we present an idealized setup used to study the influence of islands in the circulation underneath ice cavities. The setup resembles the Filcher-Ronne ice cavity. The extent to which the location and number of islands influence the circulation and freezing and melting patterns is discussed. It is also reviewed how the presence of depressions modifies the water mass exchange with the continental shelf circulation and the melting/freezing system.

---

How stable is the Brunt-Stancomb-Wills Ice Shelf? The roles of marine ice, rheology and fracture

Ala Khazendar¹, Eric Rignot^1,2 and Eric Larour¹

Marine ice, sometimes as part of an ice mélange, significantly affects ice shelf flow and ice fracture. The highly heterogeneous structure of the Brunt-Stancomb-Wills ice shelf system in the east Weddell Sea offers a rare setting for uncovering the difference in rheology between meteoric and marine ice. Here, we use data assimilation to infer the rheology of the Brunt-Stancomb-Wills Ice Shelf by an inverse control method that combines InSAR measurements with numerical modeling. We then apply the inferred rheology to support the hypothesis attributing the observed 1970s ice shelf flow acceleration to a change in the stiffness of the ice mélange area connecting Brunt proper with Stancomb-Wills, and to examine the consequences of frontal rift propagation. We conclude that while the Brunt-Stancomb-Wills system is currently not susceptible to severe fragmentation similar to that of the Larsen B Ice Shelf in 2002, our inverse and forward modeling results emphasize its vulnerability to destabilization by relatively rapid changes in the ice mélange properties, resulting from the interaction of its marine ice component with ocean water, or by the further propagation of a frontal rift.
We also present preliminary results from the application of the same methods to the Larsen C Ice shelf.
This work was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration, Cryospheric Sciences Program.

---

How old are subglacial meltwater channels on the Antarctic continental shelf?

James A. Smith¹, Claus Dieter Hillenbrand¹, Robert D. Larter¹, Alastair G.C. Graham¹ and Gerhard Kuhn²

Several recent studies have demonstrated the movement of considerable volumes of water beneath today's Antarctic ice sheets, potentially in well-organised channel systems. In addition, the flow of water is also thought to be sporadic and inherently unstable. Since basal meltwater facilitates ice motion through increased sliding or sediment deformation a clear understanding of the magnitude and rates of subglacial meltwater movement is important for our understanding of ice stream dynamics ice sheet stability. Here we present new swath bathymetric data from the western Amundsen Sea Embayment (ASE), West Antarctica, showing a network of anastomosing channels eroded into granitic bedrock. The features are similar in scale to those described elsewhere in Antarctica (e.g., Labyrinth, Dry Valleys) with some channels measuring up to 25 km long, 4.5 km wide and 400 m deep. Their morphological characteristics (reverse gradient, undulating thalweg) and size are consistent with incision by subglacial meltwater. In order to obtain a better understanding of how and when these channels formed we have performed detailed sedimentological analyses on three cores recovered from within the channels. Diamictons deposited beneath or proximal to an expanded grounded West Antarctic Ice Sheet (WAIS) are present in two of the channels and these are overlain at both sites by glaciomarine sediments deposited when the ice sheet retreated from the continental shelf. The sediment core from the third channel recovered a turbidite sequence that was probably deposited after the last deglaciation. The presence of deformation till at one core site suggests that channel incision pre-dates overriding by fast flowing grounded ice during the last glacial period and thus provides a minimum age for when the channel was last active. Moreover, the presence of the deformation till within one of the channels is consistent with findings from the Dry Valleys which indicates that the channels are preserved beneath a wet-based eroding ice sheet and hence, it is possible that they survived numerous advances and retreats of grounded ice. Given the overall scale of the channels and their incision into bedrock it is likely that the channels probably formed over multiple glaciations, possibly since the Miocene, and have been reoccupied on several occasions. In addition our sedimentological data differs from one previous study in the eastern ASE (Pine Island Bay) which reported graded gravels and sands interpreted as meltwater deposits associated with channel incision. We explore several possibilities for this regional variability.

---

The role of the shore lead at the Filchner-Ronne ice front

Keith W. Nicholls

Over the past few some decades we have become used to the idea that coastal polynyas play a vital role in the production of sea ice and the accompanying transformation of shelf water masses. This has become so much the norm that the ocean to atmosphere heat flux through neighbouring sea ice has often been considered negligible, regardless of the fact that the area of wintertime sea ice cover over the continental shelf is an order of magnitude greater than that of the open water. In this presentation I review what we know about the fluxes of water that flush the southern Weddell Sea continental shelf, and estimate the heat fluxes required to convert the inflowing water to High Salinity Shelf Water (HSSW), the key precursor of the Antarctic Bottom Water that has its origins in the Weddell Sea. I then show that heat fluxes from the wintertime shorelead meet only a fraction of this requirement. A simple thermodynamic/kinematic model illustrates that locally formed sea ice, exported by wind action from the continental shelf, allows sufficient heat to be lost to explain the water mass transformations that take place. Thus the vast majority of the water mass transformation takes place over the broader continental shelf, rather than in a narrow strip along the Filchner-Ronne ice front. It seems that in the southern Weddell Sea the narrow band of intense ice production associated with coastal polynyas might play an important local role, but it is unlikely to be significant at regional and greater scales.

---

Extent and flow dynamics of the Last Glacial Maximum ice sheet offshore of Alexander Island, and features of its last retreat

Alastair G.C. Graham, James A. Smith, Robert D. Larter, Claus-Dieter Hillenbrand and Joanne S. Johnson

Although the Last Glacial Maximum (LGM) ice sheet is commonly assumed to have advanced to the shelf break, its true extent offshore of Alexander Island remains unclear at present. Previous workers have reconstructed a small and localised ice cap on Alexander Island at the LGM (Clapperton and Sugden, 1982, Quat. Res.) while marine geological studies have indicated overriding of the island by a more extensive Antarctic Peninsula Ice Sheet (APIS; Kennedy and Anderson, 1989, Quat. Res.). Moreover, in both scenarios the flow dynamics and retreat pattern of the LGM ice sheet in this region is virtually unknown. To the west and east of the area, major ice streams drained significant drainage basins of the West Antarctic Ice Sheet (WAIS) and APIS through Belgica Trough and Marguerite Trough during the last glacial period. However, the broad ‘inter-stream’ area between these two outlets has gone largely unsurveyed to-date. If a local ice cap existed on Alexander Island, then the seabed to the northwest may represent one of the few areas of the West Antarctic shelf, which was not covered with grounded ice at the LGM. Thus, the open shelf area offshore of Alexander Island may have been crucial to the survival of marine benthos through the last glacial cycle and there would be a greater potential for low-lying terrestrial refugia on Alexander Island itself. (Maslen and Convey, 2006, Soil Bio. And Biochem.). If proven incorrect however, then it completes a picture of almost entire coverage of the West Antarctic (Pacific margin) shelf with grounded ice at the LGM, and challenges current refugia hypotheses for this part of Antarctica.
Here we present a compilation of newly-surveyed and existing marine geophysical and geological data from the area offshore of Alexander Island, outlining first results on the extent, flow dynamics and retreat of grounded ice in this sector, at and subsequent to the LGM. Our analyses, though preliminary, lend support to an extensive and widespread glaciation of the continental shelf during the last glacial period, followed by a stepwise, episodic deglaciation. The palaeoglaciological features identified from our new datasets should be incorporated into, and reproduced within, future numerical simulations of the APIS and WAIS at the LGM.

---

Seals measure bathymetry and ocean hydrography around Wilkins Ice Shelf, Antarctica

Laurie Padman¹, Dan Costa², Mike Dinniman³, Helen Fricker⁴, Ted Scambos⁵, Michael Goebel⁶ and Birgitte McDonald²

Seals instrumented with Conductivity-Temperature-Depth Satellite Relay Data Loggers (CTD-SRDLs) have provided detailed information on bathymetry and ocean hydrography over the continental shelf around Wilkins Ice Shelf (WIS). Maximum dive depth exceeded 900 m. We use the dive depth data to identify previously unmapped deep troughs that could act as conduits for warm water across the continental shelf and into the sub-ice-shelf cavity. One such trough leads between Charcot and Latady islands towards the location of the recent collapse. The dive depth data allow us to construct a bathymetry map (at ~5 km resolution) that significantly improves upon the previous maps for the southern Bellingshausen Sea. The CTD data reveal that warm Upper Circumpolar Deep Water, with maximum subsurface temperature >1 degree C, is present over most of the shelf to the west of WIS, and in the passages between the islands that buttress the ice shelf. Thus, the WIS setting shares many of the characteristics with the nearby George VI and the Amundsen embayment ice shelves, and basal melt rates may be similar to the 2-5 m/yr estimated for George VI by Jenkins and Jacobs (2008: JGR). Satellite-derived trends in surface elevation for the WIS, if interpreted as the hydrostatic response to basal melt, are generally consistent with these melt rates.

---

Tides, Rifting and Calving of the Mertz Glacier

Benoit Legrésy¹, Lydie Lescarmontier², Richard Coleman^2,3, Pascal Lacroix¹, Laurent Testut¹, Neal Young³ and Frederique Rémy¹

The CRAC-ICE project aims at monitoring the calving of the Mertz Glacier tongue in East Antarctica, an ice tongue which extends 140km from its grounding line. Legresy et al. (2004) observed dramatic ice flow changes at daily scales linked to tide currents using limited GPS observations and SAR interferometry. In November 2007, during the IPEV R0 Astrolabe voyage, we deployed a network of year-round GPS beacons along a flow line of the glacier, including upstream of the grounding line, the upper part of the ice tongue each side of the main rift and near the glacier front. Two GPS were also installed on rock sites each side of the glacier. Two months of GPS data were collected at the end of the field season from 2 stations (GPS4 & GPS5) around the main rift. We have analysed these GPS data and interpret the results together with historical analysis of remote sensing data in relation with tides, tidal currents, and glaciological stresses. The ice tongue is freely floating and has horizontal motions of about 3m/day. It is clear that the ice flow is affected at daily scales by the tides. A kind of stick-slip effect can also be considered to occur at daily scales. We see a modulation of the flow at fortnightly scales, however, we also observe that the maximum speed occurs a few days after the higher tides. The ice tongue is moving in an E-W direction following the force exerted by tide currents at all scales. The 2 GPS sites are situated each side of the main rift about 3km apart. The rift is opening quickly at some 12 cm/day at an angle of 35° from the main flow direction. When we remove the mean flow, we observe a residual rotation of the rift opening with a radius of 15 km. The rotation center is situated in the eastern part of the rift, which appears active at the daily scale. The GPS-derived vertical velocity at GPS4 (closer to the grounding line) is 60cm/yr larger than at GPS5 (1.8m/yr instead of 1.2m/yr). Using satellite images and an estimated ice thickness profile, we propose that the part of the ice tongue (~8x8km) where this GPS site is situated, is presently subject to a very large extension, leading to further rifting.

---

Stratiegies for identification and prediction of unstable marine ice sheet dynamics

Robert Arthern

Predicting the sea level contribution from the marine part of the West Antarctic ice sheet has proved difficult, partly because it has been uncertain whether the margins of the ice sheet are stable to perturbation, or whether an small retreat could propagates quickly into the interior. A recent theory of marine margins suggests that rapid, unstable retreat should be considered a possibility, but it is not obvious that the available observations from West Antarctica support this theory. This talk will consider strategies, based upon the approach of data assimilation, that could be helpful in diagnosing whether the changes presently underway in West Antarctica are likely to evolve in an unstable manner.

---

On water mass formation and melting in ice shelf caverns

Dirk Olbers and Hartmut Hellmer

A simple box model of the circulation into and inside the ocean cavity beneath an ice shelf is used to estimate the melting rates of Antarctic glaciers and ice shelves. The model incorporates a coarse parameterization of the overturning circulation, the melting/freezing physics at the ice shelf/ocean interface, and vertical mixing. It is driven by the thermohaline inflow conditions and coupling to the heat and freshwater exchanges at the sea surface in front of the cavity. The model is tuned for the Pine Island Glacier and then applied to six other major caverns. The dependence of the melting rate to thermohaline conditions at the ice shelf front is investigated for this set of caverns, including sensitivity studies and warming scenarios.

---

Frazil ice beneath ice shelves in a three-dimensional ocean model

Ben Galton-Fenzi^1,2,3, John Hunter^2,3 and Richard Coleman^1,2,3

Preliminary results are presented of numerical simulations of frazil ice formation and motion beneath ice shelves. The study used the Regional Ocean Modelling System (ROMS) specifically adapted to simulate the ocean circulation, basal melt/freeze and frazil crystal dynamics and precipitation beneath ice shelves. ROMS is a stretched sigma-coordinate, three-dimensional, primitive equation hydrodynamic ocean model. A brief overview of the ocean model and basal boundary conditions are given before discussing the implementation of frazil ice. In the model, frazil ice that is suspended in the water column is transported, like other conservative tracers (e.g. temperature and salinity), by solving the advection-diffusion equation with two additional terms. These terms are for (1) vertical settling, and (2) sources or sinks due to melting and freezing, and secondary nucleation and precipitation. Results from simulations using simplified cavities show the formation of frazil ice due to supercooled water created in the layer adjacent to the ice shelf base. The supercooling is initiated when buoyant water that is created through basal ice-ocean interaction begins to rise. Frazil ice laden water is initially formed near the ice shelf base but can become detached and move deeper where the frazil ice then melts. Simulations that include an open ocean show that large concentrations of frazil ice can form in the surface layer. The water properties beneath ice shelves and basal melt/freeze are shown to be strongly modified by the presence of frazil ice when compared to models without frazil ice.

---

Reasonable grounding line behavior in large-scale ice sheet models: In search of a parameterization

Stephen Price¹, Bill Lipscomb¹, Todd Ringler¹ and Tony Payne²

Recent work demonstrates that grounding line behavior in ice sheet models is strongly dependent on the model numerics (e.g. Vieli and Payne (2005)). While the recent two-dimensional boundary layer theory of Schoof (2007) provides a potential "benchmark" for verification of grounding line behavior in different models, it is impractical for implementation in large-scale, three-dimensional predictive models, many of which solve a continuous equation set on a structured grid at limited spatial resolution (order 1-10 km). With respect to assessing the contribution of ice sheets to future sea level rise, these same models provide the most likely path forward over the next few years. This motivates efforts to define parameterizations of the basal boundary condition near the grounding line (e.g. Pattyn et al. (2006)), which result in "reasonable" grounding line behavior for reasonable grid resolution. Here, we report on two-dimensional, grounding-line-motion experiments that use several such parameterizations over a range of grid resolutions.

---

Force Balance of Narrow Ice Shelves

Richard C.A. Hindmarsh

Narrow ice shelves are defined as those with a low width-to-length ratio. Scale estimates of the force-balance of such ice-shelves indicate that the total back-force transmitted across the grounding line is independent of the length of the shelf, but depends upon the difference between the thickness at the grounding line and calving front. Implications for the coupling between ice-stream dynamics and sub-ice-shelf seas are discussed.
A numerical model is used to investigate the validity of the scale analysis by varying the width -length ratio in idealised ice-shelf models. The question of whether the Pine Island Glacier shelf is narrow is addressed.

---

Rift treatment in ice-shelf modelling: A comparative study of the Brunt Ice Shelf - Stancomb-Wills Ice Tongue System

Thomas Kleiner¹, Angelika Humbert¹, Ralf Greve² and Manfred A. Lange^1,3

The Brunt Ice Shelf - Stancomb Wills Ice Tongue System (BIS-SWIT) located off Caird Coast, Oates Land, Antarctica is a prominent example of an ice shelf with major rifts. Rifts are arising in shear mode along the margin of the fast moving Stancomb-Wills Ice Tongue and in the tensile mode in area between SWIT and the slower moving BIS.
Two diagnostic, dynamic ice-shelf models, the Darmstadt (DA) and the Münster (MS) model, are applied to the BIS-SWIT. The DA model uses finite element technique to solve diagnostically the elliptic boundary-value problem for the horizontal velocity in the shallow-shelf approximation (SSA), while the MS model uses finite differences. We perform numerical simulations of the present flow regime of the ice shelf that results from the ice-thickness distribution derived from ICESat (Ice, Cloud, and Land Elevation Satellite) GLAS (Geoscience Laser Altimeter System) surface elevation data and the inflow at the grounding line, for which remote sensing datasets from the British Antarctic Survey (H. Pritchard; in Humbert and Pritchard, 2006) and the Canada Center of Remote Sensing (L. Gray; in Gray, 2001) are used.
Focus of this study is the incorporation of two dominant rift systems in the different numerical schemes of the models. This is done by two different approaches, numerical decoupling in the MS model and physical softening in the DA model. Our main findings are that the incorporation of the rift system improves the modelling results in both model approaches, as it leads to increased shear between SWIT and the adjacent slow moving part. However, the extreme velocity contrast across this shear mode rift, as seen in observational data, cannot fully be simulated. In case of the tensile mode rift, the model results are representing the jump of speed across the rift very well. A major result of this study is that both numerical implementations of the rifts are resulting in a similar effects on the flow field.

---

Influence of tides on the lateral movement of Ronne Ice Shelf, Antarctica

Keith Makinson¹, Keith Nicholls¹ and Matt King²

Five GPS stations, deployed during the Austral summer of 2007-08 on Ronne Ice Shelf, Antarctica, recorded ice shelf motion over observational periods of 2 to 10 weeks. Two sites were located close to the ice front while the remainder were approximately 200 km further inland. The data show a strongly elliptical semidiurnal horizontal motion, with velocities of up to ±3400 m a-1 superimposed on a mean annual ice velocity of 500-1300 m a-1. Analysis shows that the horizontal motion occurs at diurnal and semidiurnal tidal frequencies indicating that tides, which are strongly semidiurnal in the southern Weddell Sea, are responsible for this ice shelf motion. Here we present an initial examination of the GPS data with a view to understanding the coupling between the ocean tidal motion and the motion of Ronne Ice Shelf.

---

Side drag and the generation of buttressing: Comparing flow line and two-dimensional approaches

Todd K. Dupont and Daniel S. Lindsey

The buttressing provided by an ice shelf is a function of the drag along its sides and over locally grounded areas. In flow line models of shelves (and streams) the side drag has to be parameterized. In an effort to evaluate the utility of one such parameterization we compare results from a 1-d flow line model with those from a 2-d (plan view) model with similar input data. Both models simulate depth-integrated stream/shelf flow using the MacAyeal/Morland equations, in one or two dimensions. The side drag and buttressing profiles generated by the models for various scenarios are compared, as is the relationship between the center-line speed and side drag. This latter exercise provides an assessment of the quality of the side drag parameterization adopted, wherein side drag is proportional to the nth root of the center-line speed, n being the flow-law exponent for Glen's Law. Preliminary results suggest that the parameterization is at least of heuristic value.

---

The Role of Ice Shelf Ocean Interaction in the ECCO2 Data Synthesis

Michael Schodlok and Dimitris Menemenlis

Dense water masses form on the continental shelves around Antarctica and spread as a major contributor to the overturning circulation into the global abyss. So far, as many other global ocean circulation models the ECCO2 Project (Estimating the Circulation and Climate of the Ocean Phase II) did not adequately resolve high latitude processes in particular ice shelf ocean processes.
The ECCO2 solution is obtained by a high-resolution global-ocean and sea-ice data synthesis. Its solutions are obtained by fitting a high resolution (18 km horizontal grid spacing, 50 vertical layers, z-coordinates) global-ocean and sea-ice configuration of the MITgcm to all available ocean and sea-ice in-situ and/or satellite data.
The ice shelf package included in these studies is based on the ISOMIP parameterisations for heat and freshwater exchanges at the ice-ocean interface. IceSAT data provided the ice shelf thickness and BEDMAP the water column thickness in the cavities. The ice shelves were treated as floating slabs on water without flexural deformation.
The results presented here focus on a domain taken from the global solution reaching from the Amundsen Bellingshausen Sea to 30º E. The simulations cover the period 1992 to 2006 and show, e.g., an increase in sea ice thickness in agreement with previous studies. A variety of sensitivity simulations of this configuration are used to optimise model parameters which, in turn, contribute to further improve the global model solution.

---

New Instrumentation, scientific and technological challenges for sub ice research - opportunities for international & interdisciplinary collaboration

Stefan W. Vogel and Ross D. Powell

Processes in subglacial environments beneath floating and grounded ice are key to solving remaining questions about ice sheet stability and ocean circulation processes, provide a habitat for life in extreme environment and make an unknown contribution to the global geochemical cycling of matter. Despite this recognized importance, subglacial environments are virtually unexplored territory, providing many scientific, technological and logistical challenges. Due to the complexity of this system and the technological challenges and logistical investment need to access the subglacial environment, future investigations of this environment will likely to be of interdisciplinary nature.
Here we present i) results from a study of chemical properties of a West Antarctic subglacial hydrological system illustrating the complexity of subglacial system and global linkages, ii) an update report on the development of an ice borehole deployable Remote Operated underwater Vehicle (ROV) and an oceanographic and geochemical instrumentation package and iii) discuss other technological challenges opportunities and opportunities for international collaboration in the study of sub ice environments.

---

Present and Future Stability of the Larsen C Ice Shelf (SOLIS) - Fracture Mechanics

Daniela Jansen¹, Bernd Kulessa¹, Adrian Luckman¹, Edward King² and Peter Sammonds³

The SOLIS project, funded by the UK Natural Environment Research Council, aims to identify the role of ice composition, balance of stress intensity and fracture toughness, as well as fracture mechanics in regulating the present stability of the Larsen C ice shelf. In particular, we will investigate ice shelf mechanical heterogeneity, manifesting e.g. in the South East, where the growth of rifts appears to be restricted along what appears to be the boundary between neighbouring meteoric ice units. We plan to achieve our goals by a combination of geophysical field measurements on the Southern Larsen C ice shelf (active and passive seismics, ground-penetrating radar, and GPS) and remote sensing techniques, which will enable upscaling of the field data to the entire ice shelf. The acquired data will be used as forcing for a fracture mechanics model (Rist et al. 1999, 2002), which has been already applied successfully to the Filchner-Ronne and Larsen B Ice Shelves. Based on the results concerning the present conditions of the Larsen C Ice shelf, the next step will be the simulation of likely future scenarios.
At the workshop, we will give an introduction to the SOLIS project with emphasis on the fracture mechanics model. The current status of the adapted model will be presented along with first results for both crevasse distribution and penetration depth from test simulations for the Larsen B Ice Shelf.

---

Geophysical survey of Thwaites Glacier, West Antarctica

K. Christianson, H. Horgan, L. Peters, D. Voigt, L. Zoet and S. Anandakrishnan

We present preliminary results from our survey of Thwaites Glacier. The seismic and radar work were conducted at a location as close to the grounding line as possible (about 60km upstream), where a grid of radar lines and intersecting seismic lines were acquired. In addition, GPS receivers were installed along the length of the glacier to compare velocities to previously measured values. We find that Thwaites is relatively stable, except near the grounding line. The region of the grounding line is complex with flow nearly perpendicular to local surface slope and considerable topography and structure throughout.
We will present a short description of our future work in the session on field work information.

---

Ross Sea Area Freshening, Ice Shelf Melt Increasing?

Stan Jacobs¹, Claudia Giulivi¹, Richard Cullather¹ and Eric Rignot²

A lengthening record of shelf water salinity in the southwest Ross Sea shows a persistent decline for several decades, accompanied by slight warming. The study area lies in the westward coastal current, where surface water salinity has decreased more rapidly, consistent with inferences that the primary freshwater sources are upstream in the Amundsen-Bellingshausen sector. Under similar winter conditions, fresher shelf water then forms in the large Ross Sea Polynya, replacing saltier predecessors being converted to ice shelf and bottom waters. The salinity trend is best developed during a period when the annual SAM index became more positive. Related changes in atmospheric forcing have been postulated to drive a stronger upwelling near the continental shelf break, increasing on-shelf ocean heat transport. Identification of the actual freshwater sources requires volumetric determinations of seawater properties over time, and consideration of possible changes in the rates of melting continental ice, precipitation and sea ice production. A potential ice shelf impact of the freshening trend may also be noted.

---

Aerogeophysical surveys of the Evans, Carlson, and Rutford systems, West Antarctica

Robert G. Bingham¹, Hugh F.J. Corr¹, Richard C.A. Hindmarsh¹, Fausto Ferraccioli¹ and Ian Joughin²

During the austral summer of 2006-07, an aerogeophysical team from the British Antarctic Survey (BAS) undertook ~22 000 km of new ice-penetrating radar surveys over the West Antarctic catchments drained by the Evans, Carlson, and Rutford ice stream systems (traversing an approximate area of 250 000 km2). Operating from a Twin Otter aircraft, equipped with dual-frequency carrier-phase GPS enabling <1 m navigation accuracies, the recently-developed BAS ice-sounding radar system PASIN (first used in the 2004-05 Amundsen Sea sector surveys) was deployed to image both the bed and internal reflecting horizons to depths > 3 km. The surveys had three objectives: (i) To detect and record internal reflecting horizons (and the bed) directly along ice stream flowlines inferred from remotely-sensed surface velocities both along the active Rutford and Evans Ice Streams (and tributaries) and the stagnant Carlson system. Collecting such flow-parallel englacial data represents an attempt to improve the coherence of internal layers, which in turn will improve the information provided to ice flow models that are calibrated with internal layering geometry. Surveys were additionally flown transverse to ice flowlines to obtain flow-perpendicular layering, and along the ice stream edges to obtain marginal ice-thicknesses. (ii) To acquire internal layering data over a number of ice divides across the region, ranging from inter- ice-stream ridges, major catchment divides, and selected ice rises. These data will be examined for trends in layer separation due to accumulation variation or flow-dynamical effects, and for the occurrence, or otherwise, of ‘Raymond bumps.’ (iii) To measure ice thickness along the grounding line from Evans to Rutford Ice Stream to determine ice discharge across this dynamic sector. This forms part of the wider International Polar Year (IPY) objective of calculating ice discharge from the entire Antarctic Ice Sheet. In this paper we present the survey geometry and provide some early insights gained from preliminary analyses of the data.

---

In situ glacial geophysical investigations on Pine Island Glacier, 2006-08

Robert G. Bingham, Julian B.T. Scott, Andrew M. Smith, G. Hilmar Gudmundsson, David G. Vaughan, Richard C.A. Hindmarsh and Hamish D. Pritchard

Satellite measurements made since the 1990s have exposed dramatic rates of ice thinning, acceleration, and grounding line retreat across the region of the West Antarctic Ice Sheet that drains into the Amundsen Sea. One of the largest catchments, Pine Island Glacier (PIG), drains an area of 175 000 km2, and consists of a complex system of tributaries feeding a central ice stream. Over the last two Antarctic field seasons (2006-07 & 2007-08), we have conducted the first major ground-based studies of PIG. The fieldwork involved: (i) in situ GPS measurements of ice motion; (ii) extensive radar surveys of basal conditions and internal layers; (iii) shallow ice coring up to 21 m depths; and (iv) active seismic surveys of basal, and sub-basal, conditions. In this talk we present an overview of the fieldwork methodologies employed and a first glimpse of some of the findings. In particular, we focus on the findings from the GPS measurements. At all measurement sites, current rates of acceleration exceed estimated rates over the last two decades. Mean annual increases in velocities correlate closely with changes in surface slope and driving stress. Superimposed onto this general acceleration trend are irregular changes in velocities over periods of weeks and months with no measurable coincident changes geometry. There is an increased thinning rate along the centreline that is greatest downstream. This is causing the slope changes and suggests that the PIG is not stabilising.

---

Effect of Circumpolar Deep Water on ice shelf asal melt under Ross and George VI Ice Shelves: A regional model comparison

Michael S. Dinniman and John M. Klinck

Relatively warm Circumpolar Deep Water (CDW) can be found near the continental shelf break around most of Antarctica. The advection of this warm water across the continental shelf to the base of floating ice shelves around the continent is thought to be critical in determining why ice shelves of similar thickness can have vastly different basal melt rates. For example, George VI Ice Shelf (GVI) has a similar range (100-600 m) of ice shelf thickness as the Ross Ice Shelf (RIS), yet the estimated basal melt rate of GVI (2-5 m/yr) is approximately 20 times greater than that of RIS (12-22 cm/yr). The RIS is much larger than GVI, but much of the difference in the basal melt rate has been attributed to the warm (> 1.8 C) CDW along the west Antarctic Peninsula shelf break that travels to the GVI cavity while still retaining much of its heat (> 1.0 C). Meanwhile, the cooler (~ 1.0 C) CDW along the Ross Sea shelf break is more strongly modified as it advects towards the RIS cavity entrance where only a small amount (relative to the cross sectional area) of modified CDW (< -1.0 C) is found.
In this study, high resolution (4-5 km) regional models of the Ross Sea (including the cavity beneath the Ross Ice Shelf) and the west Antarctic Peninsula coastal ocean (including cavities beneath several of the ice shelves in the area) are used to compare how the differences in CDW affect the ice shelf basal melt in each area. Modeled basal melt rates are within the observed range for the RIS (15 cm/yr) and above it for GVI (7 m/yr). Examining the fluxes of not only heat, but a simulated "dye" representing oceanic CDW, show that there is a relatively greater flux of CDW into the GVI cavity as expected. The CDW moving towards the base of GVI not only starts out warmer and travels a shorter cross-shelf distance than that being advected towards the base of RIS, but it is also subjected to less vertical mixing and exchange with surface waters.

---

Extent and dynamics of the West Antaractic Ice Sheet on the outer continental shelf off Pine Island Bay during the Last Glacial Maximum

Robert D. Larter¹, Alastair G.C. Graham¹, Claus-Dieter Hillenbrand¹, James A. Smith¹, Tara J. Deen¹, Karsten Gohl² and Gerhard Kuhn²

Previous studies have shown that during the Last Glacial Maximum (LGM) the grounding line off Pine Island Bay (PIB) advanced onto the outer continental shelf, and probably reached the shelf edge. Two troughs on the outer shelf containing streamlined subglacial bedforms have been described. One has been shown to extend north from PIB to within 100 km of the shelf edge along 107°W, while the other branches from it, trending NW across the outer shelf and reaching the shelf edge at 114°W.
New multibeam bathymetry data collected on RRS James Clark Ross Cruise JR179 in March 2008 show that the eastern trough continues NNE across the outer shelf. Across most of the seaward-shallowing outer shelf, subglacial bedforms have been obliterated by iceberg keel scours. However, new data in the axial region of the eastern trough reveal NNE-trending mega-scale glacial lineations (MSGL) 30 km further north than the previously interpreted southern limit of pervasive scouring. 40 km further NNE along the trough, another area of sea-floor containing NNE-trending lineations is encountered, although these are not sufficiently parallel to be classified as MSGL. Two narrow, seaward-inclined ramps in this area separate sets of lineations with slightly different trends, and we interpret these ramps as fronts of grounding zone wedges (GZWs) formed sequentially during retreat. Our new data also reveal two larger GZWs in the middle shelf section of the trough. These features indicate a step-wise grounding line retreat from the outer and middle shelf after the LGM, and a substantial subglacial sediment flux.
A 140-km long multibeam bathymetry and single-channel seismic profile was collected parallel to, and about 25 km south of the shelf edge on RRS James Clark Ross Cruise JR141 in February 2006. The profile shows that the eastern trough continues onto the outermost part of the shelf as a broad, low relief depression, with a maximum depth of 568 m at 105° 18′ W. This is considerably further east than where it has previously been suggested the trough might continue across the outer shelf, but is consistent with where a long-term sediment depocentre is suggested by a maximum in the continental margin free-air gravity anomaly. The part of the trough deeper than 530 m is more than 50 km wide on this profile, whereas outermost shelf depths are mostly 470-500 m between this trough and the western one.
In summary, available swath bathymetry data define two troughs that diverge from one another on the outer shelf off PIB. Each contains streamlined subglacial bedforms parallel to its overall trend, indicating that they were the paths of palaeo-ice streams. The precise timing of formation of the bedforms is not well constrained, so flow switching is a possibility. However, the seismic profile across the outer shelf does not show overlapping lenticular sediment units, which would be expected if flow-switching had been a common occurrence during previous glacial cycles. Alternative possibilities are that there was bifurcating flow or more widespread divergent flow on the outer shelf.

---

Oceanography of the Bellingshausen Sea and Implications for Antarctic Peninsula Ice Shelves

D.R. Shoosmith¹, A. Jenkins¹ and M.A. Brandon²

The Bellingshausen Sea, located in the eastern Pacific sector of the Southern Ocean, is a region of declining sea ice and land ice cover, partly driven by the warming regional climate. This region also differs from much of the rest of the continent as warm Circumpolar Deep Water (CDW), normally found only in the Antarctic Circumpolar Current, floods the continental shelf. CDW is around 3°C warmer than the surface freezing point and when it has access to the base of an ice shelf, high melt rates result. In early 2007, we undertook an intensive observational program to investigate the oceanographic regime of the Bellingshausen Sea. A large number of CTD sections were occupied, including across troughs that are thought to act as conduits for CDW on-shelf transport, and across the fronts of the floating ice shelves. Results will be presented and discussed in the context of CDW transport from the shelf break to the ice shelves, and the impact of this circulation on the ice shelves. In particular we look at the circulation under the Wilkins Ice Shelf using hydrographic data along its ice fronts.

---

Effects of basal-melting distribution on the retreat of ice-shelf grounding lines

Ryan Walker, Todd Dupont, Byron Parizek and Richard Alley

The stability of marine ice streams depends on the distribution as well as the magnitude of melting beneath the adjacent ice shelf, as shown by new model results. Recent observations of rapid retreat of ice-shelf grounding lines in the Amundsen Sea sector of West Antarctica have highlighted the need for understanding how basal melting of ice shelves by warm ocean waters affects ice dynamics and potentially contributes indirectly to sea-level rise. We apply two ice stream-ice shelf flowline models to investigate the effects of varying the spatial distribution of basal melting on grounding-line dynamics. For experiments with identical average melting, we find that retreat increases significantly as melting is concentrated near the grounding line, indicating that knowledge of the basal-melting distribution is likely necessary for accurate prediction of grounding-line migration.