<<

3560 The triggers of the disintegration of Voyeykov Shelf (2007), Wilkes , East , and its subsequent evolution Jennifer Arthur, Chris Stokes, Stewart Jamieson, Bertie Miles, Rachel Carr, Amber Leeson Corresponding author: Jennifer Arthur Corresponding author e-mail: [email protected]

Understanding the controls on the dynamics and disintegration of ice shelves and tongues is crucial to predicting future inland ice loss from the and its contribution to sea-level rise. Voyeykov (VIS) in Wilkes Land, , lost ~2445 km2 of ice during a disintegration event in late March/early April 2007. Here, we investigate changes in its structure, thickness and velocity during the seven years preceding its disintegration and its evolution until April 2020 using Landsat, Envisat ASAR and MODIS imagery and annual velocity mosaics. We also investigate potential external controls by comparing the timing and scale of changes in a range of datasets including air temperature and wind-field anomalies from climate reanalysis, surface melt predicted by a regional and concentrations. We show that disintegration was preceded by widening and propagation of major rifts extending to the calving front, but that following disintegration, an immediate speed-up did not occur, indicating the disintegrated portion of VIS exerted minimal buttressing. Our analysis suggests ice shelf disintegration was likely triggered by mélange removal linked to an atmospheric warm anomaly and a less extensive latent heat polynya, which triggered multiyear landfast sea ice (MYLI) breakout and the intrusion of warmer surface . Katabatic wind-driven ponding weakened the MYLI adjacent to VIS and primed it for breakout. The example of VIS highlights the need for ice shelf and ice sheet models to account for mélange-ice shelf interactions in the context of predicted future warming. An important implication is that surface - driven hydrofracturing is not always a necessary precursor to ice shelf collapse. 3661 A 3-D Model of Antarctic Ice Shelf Surface Sammie Buzzard, Alex Robel Corresponding author: Sammie Buzzard Corresponding author e-mail: [email protected]

The formation of surface has been linked with the disintegration of many ice shelves in the over the last several decades. Despite the importance of surface meltwater production and transport to ice shelf stability, knowledge of these processes is still lacking. Understanding the surface hydrology of ice shelves is an essential first step to reliably project future from ice sheet melt. In order to better understand the processes driving meltwater distribution on ice shelves, we present results from case studies using a new 3-D model of surface hydrology for Antarctic ice shelves. It is the first comprehensive model of surface hydrology to be developed for Antarctic ice shelves, enabling us to incorporate key processes such as the lateral transport of surface meltwater. Recent observations suggest that surface hydrology processes on ice shelves are more complex than previously thought, and that processes such as lateral routing of meltwater across ice shelves, ice shelf flexure and surface debris all play a role in the location and influence of meltwater. Our model allows us to account for these and is calibrated and validated through both remote sensing and field observations. Here we present results from in depth studies from selected ice shelves with significant surface melt features. This community-driven, open-access model has been developed with input from observations, and allows us to provide new insights into surface meltwater distribution on Antarctica’s ice shelves. This enables us to answer key questions about their past and future evolution under changing atmospheric conditions and vulnerability to meltwater driven hydrofracture and collapse. 3762 Analysing palaeocirque glacier equilibrium line altitudes as indicators of palaeoclimate across the southern Rachel Oien, Matteo Spagnolo, Brice Rea, Iestyn D. Barr, Robert G. Bingham Corresponding author: Rachel Oien Corresponding author e-mail: [email protected]

The equilibrium line altitudes (ELAs) of palaeocirque are used to obtain quantitative palaeoclimatic information from alpine environments. The dimensions of these glaciers, and their ELAs, are reconstructed based on the topography of, now ice-free, . Typically, this approach is used to derive palaeoclimatic data for a particular time period, typically glossing over the fact that cirques are time- transgressive landforms, shaped over multiple glacial cycles. In this study, we test the validity of using palaeocirque ELAs as indicators of palaeoclimate by comparing them to modern cirques and their relationship to climate (Oien et al., 2020). To achieve this, we reconstruct ELAs from 800 palaeocirques across the southern Scandinavian Mountains. The cirques are mapped in GIS, and their ELAs calculated using the floor altitude and also via a more sophisticated approach, using the GlaRe and ELA GIS tools. The population of cirques is analysed to investigate whether sub-divisions can be made on the basis of floor altitude, aspect, and links to palaeoclimatic patterns. This study will then be compared to 255 extant cirques. 3863 Calving multiplier effect controlled by geometry Donald Slater, Doug Benn, Tom Cowton, Joe Todd Corresponding author: Donald Slater Corresponding author e-mail: [email protected]

Quantifying the impact of submarine melting on calving is absolutely central to understanding the response of marine-terminating glaciers to ocean forcing. Modeling and observational studies have shown the potential for submarine melting to amplify calving (the calving multiplier effect), but there is little consensus as to under what conditions this occurs. Furthermore, the process has not yet been parameterized so that it can be included in the large-scale models used for sea level projection. Here, with help from full-Stokes Elmer/Ice simulations, we propose an analytical basis for understanding the presence or absence of the calving multiplier effect. We show that as a calving front becomes undercut it becomes more susceptible to both failure (calving only of ice that is undercut) and rotational failure (full thickness calving of ice behind the grounding line). By deriving analytical thresholds for these two forms of calving based on the depth-mean shear and terminus pressure moment, respectively, we suggest that the dominant calving style is determined by the terminus geometry and by the shape of melt undercutting in particular. Uniform undercutting extending from the grounding line to the waterline promotes serac failure and no multiplier effect, while glaciers experiencing linear undercutting that is deepest at the grounding line are more likely to experience a multiplier effect. Our study offers a quantitative framework for understanding where and when the calving multiplier effect occurs, and, therefore, a route to parameterizing the effect in ice sheet-scale models. 3964 Seasonally stable temperature gradients through supraglacial debris in the Everest region of , Central Himalaya Ann Rowan, Lindsey Nicholson, Duncan Quincey, Morgan Gibson, Tris Irvine-Fynn, Scott Watson, Patrick Wagnon, David Rounce, Sarah Thompson Corresponding author: Ann Rowan Corresponding author e-mail: [email protected]

Rock debris covers about 30% of glacier areas in the Central Himalaya and modifies the impact of atmospheric conditions on mass balance. The thermal properties of supraglacial debris are diurnally variable but remain poorly constrained for monsoon-influenced glaciers over the timescale of the ablation season. We measured vertical debris profile temperatures at 12 sites on four glaciers in the Everest region with debris thickness ranging from 0.08–2.8 m. Typically, the length of the ice ablation season beneath supraglacial debris was 160 days (15 May to 22 October) – a month longer than the monsoon season. Debris temperature gradients were approximately linear (r2 > 0.83), measured as –40°C m– 1 where debris was up to 0.1 m thick, –20°C m–1 for debris 0.1–0.5 m thick, and – 4°C m–1 for debris greater than 0.5 m thick. Our results demonstrate that the influence of supraglacial debris on the temperature of the underlying ice surface, and therefore melt, is stable at a seasonal timescale and can be estimated from near- surface temperature. These results have the potential to greatly improve the representation of ablation in calculations of debris-covered and projections of their response to . 3572 94 GHz radar backscatter: measurements of glacier terrain William Harcourt, David Macfarlane, Duncan Robertson, Brice Rea, Matteo Spagnolo Corresponding author: William Harcourt Corresponding author e-mail: [email protected]

Ground-based remote sensing measurements are capable of yielding new insights into complex glacier processes due to their ability to generate measurements at high spatial and temporal resolution. Here, we will report on the development of a novel ground-based millimetre-wave radar system that is capable of mapping radar backscatter and generating Digital Elevation Models (DEMs) of glacier surfaces at high angular resolution and in most weather conditions. The sensor, called AVTIS2, is a 94 GHz (~3 mm wavelength) real-aperture Frequency Modulated Continuous Wave (FMCW) radar that mechanically scans across a scene of interest to generate a 3D data cube of radar backscatter at each range bin, azimuth and elevation angle. Point are derived from these measurements by calculating the range to the maximum power along each Line of Sight (LOS), from which gridded DEMs are constructed. This study aims to quantify the Normalized Radar Cross Section (NRCS) of glacier ice from measurements of radar backscatter. This quantity normalizes the backscatter measurements by correcting for the size of the area illuminated by the radar beam. It quantitatively describes the glacier ‘clutter’ environment at 94 GHz, which is a complex mixture of debris, meltwater, ice of varying sizes and macro- and micro-scale topographical variations. To quantify the NRCS over glacier ice, we deployed AVTIS2 near the terminus of Rhônegletscher, , in September 2019. Quantifying the NRCS relies on accurate knowledge of the area illuminated by the radar beam. However, the size of the beam area projected on a surface depends on the local incidence angle between the radar LOS and the terrain. We calculate this angle using a DEM generated from AVTIS2 measurements. We will discuss the range of NRCS values derived using this approach and consider the implications for future field deployments and possible terrain classification methods. 3574 WACSWAIN project: isotope and chemical record from Skytrain , Antarctica Mackenzie Grieman, Helene Hoffmann, Jack Humby, Robert Mulvaney, Christoph Nehrbass-Ahles, Isobel Rowell, Eric Wolff Corresponding author: Mackenzie Grieman Corresponding author e-mail: [email protected]

The motivation for the WArm Climate Stability of the West in the last (WACSWAIN) project is to investigate the possible collapse of the (WAIS) and its surrounding ice shelves during the Last Interglacial (~120 000 years ago). As part of this project, a 651 m ice core was drilled to bedrock at Skytrain Ice Rise in Antarctica during the 2018/2019 field season. In this part of the study, isotope and marine ion signals from this ice core since the (~20,000 years ago) are presented. Ice core water isotope and marine ion signals are indicators of changes in ice sheet and ice shelf extent. Changes in ice sheet and ice shelf extent are linked to regional climate patterns. Water isotope ice core signals have the potential to show changes in these patterns. Variability in ice core levels of marine ions indicate changes in the proximity of Skytrain Ice Rise to open ocean or sea ice. Water isotopes and marine ions (sodium and magnesium) were analysed continuously using cavity ring down spectroscopy and inductively coupled plasma mass spectrometry, respectively. 18O was - -233 ± 46‰ in this ice core. Mean levels of sodium and magnesium were 55 ± 57 ppb and 7 ±During 6 ppb, the respectively. Holocene, meanSodium δ and magnesium28 ± 8‰ levels (± 1σ) decline and mean from δD was into the early Holocene. They then increase 2-fold and become more variable from the early to late-Holocene. This increase could indicate a retreat of the ice shelf to its current position. 3575 Age–depth stratigraphy of inferred from airborne radar and ice-core chronology Julien Bodart, Robert Bingham, David Ashmore, Nanna Karlsson, Andy Hein, David Vaughan Corresponding author: Julien Bodart Corresponding author e-mail: [email protected]

Understanding the contribution of the West Antarctic Ice Sheet (WAIS) to past and future sea level has emerged as a scientific priority over the last three decades. In recent years, observed thinning and ice-flow acceleration of the marine-based Pine Island Glacier has demonstrated that dynamic changes are central to the long-term stability of the WAIS. However, significantly less is known about the evolution of the catchment during the Holocene. Internal Reflecting Horizons (IRHs) provide a cumulative record of accumulation, basal melt and ice dynamics that, if dated, can be used to inform ice flow models to project spatial and temporal mass changes. Here, we use airborne radar data sets to trace four consistent IRHs spanning the late Quaternary across the Pine Island Glacier catchment. We use the WAIS Divide ice- core chronology to assign ages to three IRHs: 4.72 ± 0.08, 6.94 ± 0.11, and 16.50 ± 0.62 ka. We use a 1-D model, constrained by observational and modelled accumulation rates, to produce an independent validation of our ice-core-derived ages and provide an age estimate for our shallowest IRH (2.31-2.92 ka). We find that significantly older ice is present below our deepest reflector, but the absence of continuous radar-observed reflectors detected in our dataset at depth currently limits our ability to trace ice older than the Last Glacial Maximum. The clear correspondence between our IRH package and the one previously identified over Institute , altogether representing ~20% of the WAIS, suggests that a unique set of stratigraphic markers spanning the Holocene exists widely across . 3576 GIV: a new open-source toolbox for calculating glacier velocity time series from optical . Maximillian Van Wyk de Vries, Andrew D. Wickert Corresponding author: Maximillian Van Wyk de Vries Corresponding author e-mail: [email protected]

Ice surface velocities provide insight into the dynamics of glaciers, including their stability, thickness, and hazard potential. Full two-dimensional velocity fields may be derived from satellite imagery using a technique known as feature tracking, in which the displacement of distinctive patterns are traced from one image to the next. Recent increases in the availability of high-resolution, short-repeat-time satellite imagery amplify the impact of these methods. We present a new toolbox, Glacier Image Velocimetry (GIV), which enables rapid processing of large image datasets through an easy to use interface. GIV implements novel image pre- and post-processing to extract velocity fields even from partially clouded, shadowed and otherwise complex glacier surfaces. We present three examples of how this model may be used: to complement a field campaign (Glaciar Perito Moreno, Argentina), calculate the velocity field of a large, surging (Vavilov ice cap, ), and invert for the ice volume present within a tropical ice cap (Volcán Chimborazo, Ecuador). 3577 Towards supervised classification of open water and across Antarctic ice shelves Rebecca Dell, Ian Willis, Alison Banwell, Neil Arnold, Anna Ruth Halberstadt, Hamish Pritchard Corresponding author: Rebecca Dell Corresponding author e-mail: [email protected]

The areal extent and volume of surface meltwater (open water and slush) on Antarctic ice shelves has implications for ice shelf (in)stability. For example, the collapse of the Larsen B ice shelf in 2002 followed the development and subsequent hydrofracture-induced drainage of > 2000 surface . Since this break-up event, numerous studies have investigated the areal extent and volume of surface lakes across Antarctic ice shelves. However, very few studies have investigated the changing distribution of saturated (slush) within and between melt seasons. Saturated firn can store a significant volume of surface meltwater on an ice shelf, and also acts as a precursor for the ponding of open water. Therefore, it is important to quantify the extent and distribution of all surface meltwater (open water and slush) across Antarctica’s ice shelves. To do this, we develop a Random classifier in Google Engine using the Landsat 8 satellite record, capable of detecting surface meltwater on ice shelves on a pan-Antarctic scale. We train and test our classifier on six ice shelves: (1) George VI, (2) Nansen, (3) Shackleton, (4) Amery, (5) Roi Baudouin, (6) Nivlisen. A k-means clustering algorithm is first used to generate output clusters from a range of training image scenes, which are then manually interpreted to form the final training classes (e.g. open water, slush, , dirty ice). These training classes are then used to train the Random Forest classifier. We test the classifiers performance using expert elicitation. Early results across the six training sites are promising, as the Random Forest classifier is able to accurately classify slush and open water at various stages in the melt season. 3578 Sensitivity of tidewater glaciers to lateral variation in submarine melting Tom Cowton, Joe Todd, Doug Benn Corresponding author: Tom Cowton Corresponding author e-mail: [email protected]

The response of tidewater glaciers to ocean warming remains a key uncertainty in sea level rise predictions. Here we use a 3-D numerical model to examine the response of an idealized tidewater glacier to spatial variations in submarine melt rate. Whilst melting towards the centre of the terminus causes only a localized increase in mass loss, melting near the lateral margins triggers increased calving across the width of the glacier, causing the terminus to retreat at several times the width-averaged melt rate. This occurs because melting near the margins has a greater disruptive impact on the compressive stress arch that transfers resistance from the side walls to the body of the glacier. We suggest that the rate of terminus advance or retreat may thus be governed by the difference between ice velocity and submarine melting in the slow flowing zones away from the glacier centre. 3580 and radars reveal retreat repertoires Richard Hindmarsh, Howard Conway, Nancy Bertler, Tollý Aðalgeirsdóttir Corresponding author: Richard Hindmarsh Corresponding author e-mail: [email protected]

Sea-level has risen by 120 m since 20 ka BP due to the disappearance of ice caused by warming of global climate. The rate of sea-level rise is now 3 mm/a, in part due to the melting of ice and increased flow across the grounding line in localized (10-30 km) areas. The dynamics of ice sheets are not well enough understood to make reliable predictions of SLR in the coming centuries; this has led to an interest in ice- rises as pinning-points that provide buttressing. Two significant developments have occurred; (i) drilling of five ice cores into Antarctic ice-rises; and (ii) the forecast and verification of occurrence of anticlines, known as the ‘Raymond Arches’, under ice-rise divides, whose formation relies on the shear-thinning non-linear rheology of ice. This was first proposed in the 1980s, and verified by radar sounding in the 1990s. Raymond Arches have been observed in ice rises surrounding Antarctica, and their amplitudes used to infer the date where the divide repositioned to the present location and deduce the subsequent thinning of ice. The pRES has been used to measure vertical motion and confirm that the ice is flowing in a way that produces arches. These data can be combined with geological data to infer the geography and rates of ice-sheet retreat and illustrate that the retreat of marine ice- sheets shows complex patterning. Isotopic analysis of the cores gives information about the climate and ice=mass geography when the ice was deposited as snowfall. We focus on the RICE core, drilled into Roosevelt Island (RI) by an eight-nation team led by Nancy Bertler (VUW), and on Fletcher Promontory (FP). We have used two types of radars to explore ice rises and their particularly strong presentation of the Raymond Effect; pulse radars to explore the isochrone anticlines under divides, and phase-sensitive radars to measure the horizontal variation in vertical velocity that produces these anticlines. Radar work shows RI divide reached its current location 3ka BP, and thinned 300m, while FP divide commenced thinning of 500m at 5ka BP. The RICE ice-core analyses provide further detail. We present results from the ice- cores of RI and the radar work on both RI and FP. These studies illuminate the complex retreat histories and raise further questions about the extent to which the retreat was forced by global warming or determined by internal ice-sheet dynamics; in other words, questions about the retreat repertoire. 3581 Constraining till permeability via tidal velocity variations Katarzyna Warburton, Duncan Hewitt, Jerome Neufeld Corresponding author: Katarzyna Warburton Corresponding author e-mail: [email protected]

Many marine-terminating glaciers show a tidal signature in their surface velocity. While some accelerate and decelerate at the frequency of the daily tides, others show a fortnightly variation, indicating that a non-linear mechanism acts between the tidal forcing and the velocity response. Glacier speed depends sensitively on subglacial water pressure, so to model the link between the upstream hydrological network and the ocean tides, we look at water transport across the grounding line. Starting from a minimal mathematical model of the flow of water under an elastic ice sheet, we show that the incoming tide drives the grounding line to rapidly migrate inland, while the downstream migration is limited by the rate of fluid drainage through the subglacial till. The asymmetry between the up‐ and downstream motion allows the grounding line to act as a non‐linear filter on the tidal forcing. The degree of filtration depends on the permeability of the hydrological network and the amplitude of the tidal forcing, providing a novel constraint on till permeability that reconciles observations from across Antarctica. 3584 -wide inventory of ice marginal lakes using Sentinel-1, Sentinel- 2 and ArcticDEM Penelope How, Alexandra Messerli, Eva Mätzler, Maurizio Santoro, Andreas Wiesmann, Rafael Caduff, Kirsty Langley, Mikkel Høegh Bojesen, Frank Paul Corresponding author: Penelope How Corresponding author e-mail: [email protected]

Ice marginal lakes are a dynamic component of terrestrial water storage, currently holding up to 0.43 mm of sea level equivalent globally along with supraglacial and proglacial lakes. They form when meltwater is stored at the margins of glaciers and ice sheet outlets, with the potential for catastrophic flooding that can alter downstream and ecosystems including suspended sediment flux, water salinity, and enhanced and deposition. There are many known ice marginal lakes around the margin of the and ice caps, however, an exclusive and complete record of all Greenland ice-marginal lakes currently does not exist. Here, we present a comprehensive inventory of ice marginal lakes in Greenland for the year 2017, classified using three well-established and independent methods – multi-temporal backscatter classification from SAR imagery, multi-spectral indices classification from optical imagery, and sink detection from DEMs. Overall, 3347±15% ice-marginal lakes (over 0.05 sq km) were identified, with an average area of 0.88 sq km. The highest proportion of lakes are found around Greenland’s ice caps and the southwest margin of the ice sheet. Through analysis with comparative studies, it is suggested that there is an increasing trend in the abundance of ice marginal lakes, with a 75% increase over the west margin since 1985. It is likely that these changes will alter the dynamics of this terrestrial store of water and its effect on future sea level budget. 3585 Investigating fracture patterns along the lateral shear margins of the Rebecca Fletcher, Brice Rea, Matteo Spagnolo, Andrew Newton, David Healy Corresponding author: Rebecca Fletcher Corresponding author e-mail: [email protected]

Lateral shear margins (LSMs) are zones of water-saturated and porous temperate ice that separate zones of inactive ridge-ice from the rapidly flowing stream-ice. Lateral shear margins result from the sharp transition in basal conditions from a zone of sliding to a zone of little-to-no slip. Along with higher mechanical strength till and topographic high points, LSMs provide the dominant resistive stress to ice stream flow. LSMs can be identified and mapped from surface velocity data, as velocity is shown to decrease in the order of 2 to 3 magnitudes in a few kilometres across the LSM. LSMs are also identifiable from the landform (palaeo)record, as (LSM-moraines) constructed at the ice bed interface, thus helping to delineate former ice stream geometry. This project initially aims to investigate the deformation characteristics of contemporary ice streams, with a focus on lateral propagation. The propagation angle of in biaxial stress fields (both tensile stress and shear stress), can fluctuate in orientation by up to 20º from the optimum orientation, but is overall a function of the initial crevasse orientation and the ratio of the two stresses (van der Veen, 1999). This study utilizes LANDSAT 8 imagery at 15 m resolution of the Bindschadler Ice Stream (BIS) on the Siple of Antarctica. LSMs were identified using the MEaSUREs version 2 dataset, and the length, width, and orientation of each crevasse was measured. Analyses of crevasse orientations will be undertaken to determine the fracture characteristics of the marginal crevasses and how these fit with the lateral propagation crevassing theory of Van der Veen (1999). 3586 Expansion of surface lakes into Greenland Ice Sheet interior enhanced by dynamically-evolving surface relief Adam Igneczi, Andrew Sole, Stephen Livingstone, Felix Ng, Chris Clark, Miren Vizcaino Corresponding author: Adam Igneczi Corresponding author e-mail: [email protected]

Surface lakes play a key role in the drainage system of the Greenland Ice Sheet (GrIS) as they facilitate the formation of surface-to-bed meltwater conduits. This process affects the seasonal and inter-annual evolution of the subglacial drainage system, and consequently ice flow velocity and sediment transport. Surface lakes have expanded significantly towards higher elevations since the 1970s, and the few available projections predict that this trend is likely to continue during the 21st century. Less is known, however, about the changing spatial distribution of surface lakes beyond 2100, and how changing surface relief modulates the trend. Local topographic relief, produced as undulations (i.e. bumps) on the surface that form by transmission of bed variations to the surface in flowing ice, has been demonstrated to control the spatial distribution of surface depressions and lakes on the contemporary GrIS. Here we use an extended mathematical model of this transmission, along with ice sheet model outputs to estimate the changing surface relief and spatial distribution of surface lakes of the GrIS between 1980 and 2300. Our results show that the surface relief of the GrIS will increase very modestly by 2100 – with an ice sheet-wide mean increase of 1.9-2.5% compared to 1980-1999. However, the increase will be an order of magnitude larger by 2300 – with an ice sheet-wide mean increase of 32%. In agreement with previous projections, we predict a strong increase in the total ice sheet-wide volume of surface lakes during the 21st century – with a 47.5-87.4% increase compared to 1980-1999. Although the expansion of surface lakes will continue beyond 2100, the rate of increase will slow down, with the total volume increasing to only 120.7% (compared to 1980- 1999) by 2300. This is due to the relative scarcity of surface depressions in the areas affected by the rising ELA after 2100 – these regions are usually far from the ice sheet margin at high elevations. According to our estimations, neglecting the effects of the changing surface relief would underestimate the ice sheet-wide total volume of surface lakes by 11.7% by 2300, whereas it would only cause a 1.5-2.7% underestimation by 2100. Hence, surface lake projections assuming constant surface topography are valid for the 21st century, but projections beyond that should consider the effects of the changing surface relief. 3587 Characterizing and quantifying shear margin ice fabric anisotropy from radar sounding Tun Jan Young, Dustin M. Schroeder, Poul Christoffersen, Slawek M. Tulaczyk Corresponding author: Tun Jan Young Corresponding author e-mail: [email protected]

It has been known since the 1970s that polar ice is birefringent as a result of dielectric anisotropy associated with the bulk alignment of ice , known as the crystal orientation fabric, which is manifested in radar sounding as interference patterns from differences in orthogonal wave propagation speeds. Data collected in 2018/19 at the Eastern Shear Margin (ESM) of using the CReSIS Accumulation-C airborne radar show, in addition to traditional radiostratigraphic features such as surface and englacial layering, periodic layering in received power that cross-cuts these traditional features, the latter which we interpret as arising from birefringent interference. Previous model-based studies show that the wavelength of birefringent interference is inversely proportional to the strength of fabric asymmetry. Using this relationship, we quantify the bulk-fabric strength of ice along six separate transects and reveal large-scale trends in birefringence and fabric asymmetry as they cross the ESM. From these results, we infer the evolution of the crystal orientation fabric across the shear margin, where ice is subjected to varying levels of both pure and simple shear. Our findings suggest the potential of the upper reaches of the ESM having undergone recent inward migration. Together with ground-based polarimetric radar experiments, our study highlights the potential of radar sounding to observe and infer variations in fabric strength from regions of complex flow at multiple spatial scales. 3590 Detecting ice slabs in firn using seismic full waveform inversion (FWI) Emma Pearce, Adam Booth, Sebastian Rost, Paul Sava Corresponding author: Emma Pearce Corresponding author e-mail: [email protected]

The transformation of into ice is a fundamental process in glaciology. The yearly accumulation of fresh snowfall increases overburden pressure, changing the snow’s properties such that it transitions into firn and pure glacier ice thereafter. However, a smooth firn gradient can be interrupted by episodic melt, which introduces refrozen layers or more extensive ice slabs into the firn column. These layers can have significant implications for the surface hydrology of glaciers hence detecting and resolving them is a valuable component of mass balance. The development of ice slabs on Antarctic ice shelves has also been invoked as a mechanism for collapse. Given the complexity of firn columns, and the processes they undergo (e.g., settling, sublimation, recrystallization) it has not been possible to develop a theoretically-based model that accurately predicts firn properties with depth. Therefore, methods of characterizing firn often use direct ice core sampling, or seismic data to obtain velocity-depth profiles from which firn density may be inferred. Firn models are typically built from seismic data using a Herglotz-Wiechert (HW) inversion of first-break travel-times, but the output trends can be over- simplified. Furthermore, while it may detect the top of a refrozen slab, HW inversion is less suited to resolving its base. Full Waveform Inversion (FWI) methods can mitigate these limitations, both detecting and resolving the vertical extent of refrozen slabs. Using synthetic models that simulate ice slabs (5m to 80 m thickness, 5m to 80m depth, and 3800 m s–1 velocity) within a background firn trend (Herron- Langway firn densification model), we show that FWI is a promising method for characterizing complex firn gradients. Current models simulate seismic data from 3Hz to 60 Hz, which is lower than a real-data acquisition, but on-going work is exploring the stability of the inversion as the starting frequency is increased. Nevertheless, the current FWI strategy both detects a sharp increase in ice velocities as waves enter the slab and can resolve the velocity anomaly with a high degree of accuracy. 3591 Seismic quality factor measured for compressional and shear waves in the firn column of Korff Ice Rise, West Antarctica Ronan Agnew, Roger Clark, Adam Booth, Alex Brisbourne Corresponding author: Ronan Agnew Corresponding author e-mail: [email protected]

Seismic attenuation, expressed by the Quality Factor Q, is a valuable property in seismic investigations of glaciers, ice sheets and other settings. This value arises firstly because Q is used to correct measurements of seismic amplitude for wavelet propagation effects, as in reflection amplitude-versus-angle (AVA) studies. Secondly, Q is also an indicator of ice properties such as temperature, with higher Q values associated with colder ice. There is evidence from laboratory and field studies of , and unconsolidated and solid geological materials, that the ratio of the compressional- and shear-wave quality factors, Qp : Qs, may be indicative of fluid content or saturation. Thus, a measurement of Qp and Qs in the firn column may also be a useful tool to help determine the hydrological structure of the firn. Despite its importance, few studies appear to have measured Qp in firn columns and none appear to have measured Qs in firn. Doing so for either compressional- or shear-wave arrivals is challenging, due to the raypaths followed by the diving-wave first arrivals and the proper application to them of attenuation measurement tools. In preparation for an AVA study of bed properties at Korff Ice Rise, West Antarctica, we have used diving wave amplitudes and a modified spectral-ratio method to measure Qp as a function of depth in the firn column, represented as 6 discrete constant-Q layers with thicknesses between 6 and 17 m. Shot gathers with vertically oriented geophones at offsets of 2.5–1000m were used to measure Qp. For detecting the shear component, the geophones were oriented horizontally; in this configuration, diving and reflected shear phases were recorded with high signal-to-noise ratios. Qp shows progressive increases in depth from 43 ± 8 in the uppermost 12 m, to 317 ± 43 between 74 and 80 m depth. At 74 m depth, firn is indistinguishable from the underlying ice column, which has a Q of 275 ± 60. The shear-wave record is currently being analysed, and profiles of Qs and the implied Qp : Qs ratio will be obtained. 3592 Subglacial lakes and hydrology across the Ellsworth subglacial highlands, West Antarctica Felipe Napoleoni, Stewart S. R. Jamieson, Neil Ross, Michael J. Bentley, Andrés Rivera, Andrew M. Smith, Martin J. Siegert, Guy J. G. Paxman, Guisella Gacitúa Corresponding author: Felipe Napoleoni Corresponding author e-mail: [email protected]

Subglacial water plays an important role in ice sheet dynamics and stability. Subglacial lakes are often located at the onset of ice streams and have the hypothesized to enhance ice flow downstream by lubricating the ice-bed interface. The most recent inventory of Antarctica mapped nearly 400 lakes, of which ~14% are found in West Antarctica. Despite the potential importance of subglacial water for ice dynamics, there is a lack of detailed subglacial water characterization in West Antarctica. Using radio-echo sounding data, we analyse the ice-bed interface to detect subglacial lakes. We report 33 previously uncharted subglacial lakes and present a systematic analysis of their physical properties. This represents a ~40% increase in subglacial lakes in West Antarctica. Additionally, a new of basal topography of the Ellsworth subglacial highlands was built and used to create a hydropotential model to simulate the subglacial hydrological network. This allows us to characterize basal hydrology, determine subglacial water catchments and assess their connectivity. We show that the simulated subglacial hydrological catchments of , Pine Island Glacier and Thwaites Glacier do not correspond to their ice surface catchments. 3593 What is stirring at a lacustrine glacier calving front? Adrian Dye, Joseph Mallalieu, Fran Falcini Corresponding author: Adrian Dye Corresponding author e-mail: [email protected]

Studies in Patagonia, Nepal and have shown that where glaciers terminate in proglacial lakes glacier mass loss is accelerated through thermal and mechanical processes, particularly through the formation of thermal notches in the ice front. Despite this there are limited studies into the thermal regime of proglacial lakes and observations of temperature directly at the ice water contact point. The abundance and temperature of these proglacial lakes in glacier systems has received relatively little attention, which needs to be addressed given recent extreme heat events and the Arctic Amplification of increased air temperatures. We present temperatures and a calving mechanism record, including analysis of movement at the ice front throughout the 2017 and 2019 melt season at an Arctic lacustrine glacier. Previous melt models for lacustrine terminating glaciers have been compromised by a lack of data from the hazardous water to ice contact point and assume a uniform temperature (e.g. 1°C). Using in situ and spatial surveys, we recorded proglacial lake thermal regime. During the 2017 melt season, winds were blowing away from the glacier and water temperatures of 3 C were recorded at the ice front. Whereas during the 2019 melt season winds were predominantly blowing towards the glacier, resulting in water temperatures of ~4 C directly at the ice front (recorded over 12 days). Time lapse imagery was used to analyse terminus geometry change in both field seasons at the front of an actively calving Arctic glacier (67.954878°N, 18.561535°E), which rapidly lost 10 523m2 of ice (0.67% of area in RGI, 2008) between 2014 and 2018. We present a calving mechanism record from 2017 and 2019, with the majority of events being driven by melt undercutting. The analysis of iceberg movement reveals complex circulation patterns at the ice front, which play a strong role in subaqueous mass loss and changing geometry of the ice front during a period of rapid retreat. We work towards developing a time series of iceberg calving volume from SfM analysis of time lapse imagery and sonar analysis to determine subaqueous geometry and melt rates (Mallalieu et al., 2017). 3594 Baffin Bay sea ice thickness and the impact of snow depth products and processing methods Isolde Glissenaar, Jack Landy, Alek Petty, Nathan Kurtz Corresponding author: Isolde Glissenaar Corresponding author e-mail: [email protected]

The ice cover of the is increasingly becoming dominated by seasonal sea ice. It is important to focus on the processing of altimetry ice thickness data in thinner seasonal ice regions to understand seasonal sea ice behaviour better. This study focusses on Baffin Bay as a region of interest to study seasonal ice behaviour. We aim to reconcile the spring sea ice thickness derived from multiple satellite altimetry sensors and sea ice charts in Baffin Bay and produce a robust long-term record (2003–19) for analysing trends in sea ice thickness. We investigate the impact of choosing different snow depth products (the Warren climatology, a passive microwave snow depth product and modelled snow depth from reanalysis data) and snow redistribution methods (a sigmoidal function and an empirical piecewize function) to retrieve sea ice thickness from satellite altimetry sea ice freeboard data. Snow depth redistribution schemes have been applied to laser altimetry data but have not yet been tested for radar altimetry data. We attempt to create a redistribution function for snow depth over radar altimetry freeboard data to check if it is required for reliable ice thickness estimates. The choice of snow depth product can result in differences in mean sea ice thickness across Baffin Bay of up to 20%. Moreover, the choice in snow redistribution method can result in differences in mean sea ice thickness of up to 25%. Previous studies have shown a possible long-term asymmetrical trend in sea ice thinning in Baffin Bay. The present study shows that whether a significant long-term asymmetrical trend was found depends on the choice of snow depth product and redistribution method. The satellite altimetry sea ice thickness results with different snow depth products and snow redistribution methods show that different processing techniques can to different results and can influence conclusions on total and spatial sea ice thickness trends. Further processing work on the historic radar altimetry record is needed to create reliable sea ice thickness products in the marginal ice zone. 3595 Subglacial meltwater channel dimensions and processes at the ice sheet scale: 3-D morphometry of a large sample of Rob Storrar, Andrew Jones, Frances Butcher, Nico Dewald, Chris Clark, Cathy Delaney, David Evans, Emma Lewington, Stephen Livingstone Corresponding author: Rob Storrar Corresponding author e-mail: [email protected]

Meltwater exerts an important control on ice sheet dynamics and is the subject of a burgeoning literature. However, the active subglacial environment remains difficult to study mainly because of its inaccessibility. Present understanding of the dimensions, pattern, and extent of subglacial meltwater channels at the ice sheet scale is limited to sporadic observations and numerical modelling. We address this gap by using the Quaternary geomorphological record as a to quantify in detail the dimensions and pattern of subglacial meltwater channels beneath portions of two Quaternary ice sheets. We present the results of a high-resolution (2 m), large sample (n >50 000) study of three-dimensional morphometry at sample locations in SW Finland and , . Detailed mapping of esker outlines and crestlines permits the quantification of a number of parameters, including: length; width; height; cross-sectional area; volume; sinuosity; cross- sectional symmetry; uphill/downhill trends. Whilst the dimensions of eskers are not always directly proportional to the size of the meltwater channel, they offer a powerful proxy for understanding the size and shape of meltwater conduits and the configuration of subglacial drainage systems across large areas. The results may be used to: (1) inform numerical models of subglacial meltwater drainage; (2) inform process models of esker formation; and (3) provide a dataset of esker morphometry against which other features may be compared (e.g. sinuous ridges on ). 3596 Structural controls on the hydrology of crevasses on the Greenland Ice Sheet Thomas Chudley, Poul Christoffersen, Samuel Doyle, Thomas Dowling, Robert Law, Charlotte Schoonman, Marion Bougamont, Bryn Hubbard Corresponding author: Thomas Chudley Corresponding author e-mail: [email protected]

Surface crevasses on the Greenland Ice Sheet deliver significant volumes of meltwater to the englacial and subglacial environment, but the topic has received little attention compared to supraglacial lake and drainage. Here, we explore relationships between crevasse hydrology, topography, and surface-parallel stress regime at a fast-flowing, marine-terminating sector of the Greenland ice sheet. Regional-scale observations of , crevasses, and stress were made across a 3000 km2 region using Sentinel-2, ArcticDEM, and MEaSUREs velocity data. Across the 2017–19 melt seasons, on average 25.9% of crevassed ice surface was observed to fill with water, with warmer years experiencing more extensive ponding. The spatial distribution of water-filled crevasses was consistent between years, but whilst the distribution of supraglacial lakes is controlled by surface topography (i.e. ponding in basins), this is not true for crevasses, with less than 10% of water-filled crevasses occurring in topographic minima. Instead, we suggest that mean surface-parallel stresses exert a significant control over the distribution of ponded crevasses. We found that dry crevasse fields, where no ponded meltwater was observed through the entire melt season, were more likely to exist in tensile mean stress regimes, which we interpret to be due to meltwater draining continuously into the englacial system. Conversely, wet crevasse fields, hosting ponded meltwater, were more likely to exist in compressive mean stress regimes, which we interpret to be a result of closed englacial conduits. Using contemporaneous high spatio-temporal resolution observations obtained from uncrewed aerial vehicle surveys on Store Glacier, we show that these ponded crevasses can drain through episodic rapid drainage events (i.e. hydrofracture). Mean stress regime may therefore inform spatially heterogeneous styles of meltwater delivery through crevasses to the bed of ice sheets, with distinct consequences for basal processes such as subglacial drainage efficiency and cryo- hydrologic warming. 3597 Firn structure and its variability across Larsen C Ice Shelf, Antarctic Peninsula, from multimodal layered transdimensional inversion (MuLTI) of seismic dispersion curves and borehole density logs Siobhan Killingbeck, Bernd Kulessa, Adam Booth, Bryn Hubbard, Alex Brisbourne, Suzanne Bevan, Adrian Luckman, David Ashmore, Phil Livermore Corresponding author: Bernd Kulessa Corresponding author e-mail: [email protected]

Rising surface temperatures are causing firn layers on Antarctic ice shelves to melt and compact more rapidly, a process strongly implicated in ice shelf disintegration through ice column warming and support for meltwater-forced crevasse propagation. On Larsen C Ice Shelf, Antarctic Peninsula, a compacting firn layer has been inferred from airborne radar and satellite data, with strongly reduced air content in the shelf’s north and north-west sectors, yet the structural evolution and hydrological processes governing firn compaction have so far remained uncertain. Here we derive Rayleigh-wave dispersion curves for all applicable seismic survey sites Larsen C, as key inputs to our Multimodal Layered Transdimensional Inversion (MuLTI) scheme, a Markov chain Monte Carlo implementation of Bayesian inversion to derive the probability distribution of shear wave velocity (Vs) with depth. In addition to dispersion curves, MuLTI incorporates independent depth constraints, here derived from density logs acquired in five boreholes co-located with seismic surveys in Cabinet and Whirlwind Inlets. Once calibrated in this manner, we apply MuLTI to all other available seismic data sets, allowing us to map the spatial and depth distribution of Vs, compressional wave velocity (Vp), Poisson’s ratio and shear modulus across the whole ice shelf and compare it with existing maps of firn air contents. We find that MuLTI generates more realistic depth profiles of seismic velocities and density than the widely-adopted continuous refraction method of Kirchner and Bentley (1990), since shear waves are more sensitive than compressional waves to vertical density contrasts. MuLTI thus emerges as a powerful method for identifying the melt and compaction history of firn layers, with important implications for Antarctic ice shelf stability and ice sheet and glacier hydrology and surface mass balance more generally. 3598 High resolution 3-D RES imaging of subglacial lineations under the Rutford Ice Stream, West Antarctica Rebecca Schlegel, Adam Booth, Tavi Murray, Andy Smith, Alex Brisbourne, Roger Clark, Ed King, Steph Cornford Corresponding author: Rebecca Schlegel Corresponding author e-mail: [email protected]

Understanding the present-day glacier dynamics of the West Antarctic Ice Sheet is essential for reconstructing its past dynamics. Subglacial lineations, such as mega- scale glacial lineations and , are known to be indicative of fast ice flow. A better understanding of the mechanisms driving fast ice flow, and therefore basal topography and properties (e.g. water content, roughness), are important inputs for models to predict future dynamics and losses of the West Antarctica Ice Sheet. The Rutford Ice Stream (more than 2 km thick, of which 1.4 km is below sea level) is a fast-flowing glacier in West Antarctica: the ice surface speed at the grounding line is >1 m/day, stable over the past 30 years. The ice-bed interface is at the pressure- melting point. Legacy radio-echo sounding (RES) and seismic 2-D profiles revealed highly elongated lineations, up to ~14 km long, up to 150 m high, and 50–500 m wide, aligned in the ice-flow direction. In one location, the deposition of sediment, arranged as a , was observed over a period of <10 years. To study the detailed architecture of the lineations, three 3×3 km grids of 3.5 MHz zero-offset RES data were acquired over 10 weeks in 2017/18, with crossline and inline spacings of 20m and ~1.5m respectively, dense enough to be compatible with 3D migration to improve the resolution of lineations. We present the high resolution basal topography of the bed, including the downstream end of landforms. Furthermore, reflected amplitudes in the dataset show spatial variations that indicate changes in bed properties across that region. Areas with different reflectivity seem to contain landforms at different spatial scale, which coincides with a change in the assumed glacier dynamics interpreted from previous studies. Compared to conventional 2-D acquisition strategies, our 3-D imaging does not change the overall topographic model of the landforms, but greatly benefits the resolution of small-scale features at their downstream end. The subtle features can be significant indicators of particular dynamic processes, hence targeted 3-D imaging can be a valuable component of a RES acquisition strategy. 3599 Highly temporally and spatially variable Antarctic ice flux throughout the 21st century Bertie Miles, Chris Stokes, Stewart Jamieson, Hilmar Gudmundsson, Adrian Jenkins, Jim Corresponding author: Bertie Miles Corresponding author e-mail: [email protected]

It has been widely reported that ice flux from the Antarctic Ice Sheet has increased over the preceding decades. The largest increases in ice flux have occurred in the Sector, with a much more limited change in East Antarctica. However, much less attention has been focused on the temporal and spatial variations of ice flux in Antarctica over the observational period. In this study we combine existing velocity products (ITS_LIVE and MEaSUREs) to create 12 timestamped velocity mosaics between 1999 and 2018 to investigate both overall trends in ice flux and the variability across the observational period. At an ice sheet scale we report a 50 GT yr–1 increase in ice discharge in West Antarctica and no overall change in East Antarctica. However, at an individual catchment scale we observe considerable temporal and spatial variability. For West Antarctica, despite the overall increase in discharge clear periods of deceleration are observed in most individual catchments. In East Antarctica, despite overall consistency, 3-10% variations in ice discharge are observed at several major outlet glaciers (e.g. Denman, Totten, , Cook, Matusevitch, Rennick). These variations in discharge are primarily controlled by the ocean, but are also strongly influenced by localized factors such as stochastic calving events, fast ice induced calving events/advance, bed topography and pinning points; and in some cases resulting in opposing trends in neighbouring catchments. Improving our understanding the processes driving these short term variations will be important in improving the accuracy of future sea level contributions from Antarctica. 3600 ~200 years of change in mountain glacier extent of Troms and county, northern Joshua Leigh, Chris Stokes, David Evans, Rachel Carr, Liss Andreassen Corresponding author: Joshua Leigh Corresponding author e-mail: [email protected]

Glaciers are important indicators of climate change, and recent observations worldwide document increasing rates of mountain glacier recession. Here we present approximately 200 years of change in mountain glacier extent in Troms and Finnmark county, northern Norway. This was achieved through (1) mapping and lichenometric dating of major systems within a subset of the main study area (the Rotsund ) and (2) mapping recent (post-1980s) changes in ice extent from remotely sensed data. Lichenometric dating reveals that the Little (LIA) maximum occurred approximately 1814 CE (±41 years), which is before the early twentieth-century LIA maximum proposed on the nearby Lyngen Peninsula but younger than LIA maximum limits in southern and central Norway (mid-eighteenth century). Between LIA maximum and 1989 CE, a small sample of measured glaciers (n = 15) shrank a total of 3.9 km2 (39 %), and those that shrank by more than 50% are fronted by proglacial lakes. Between 1989 and 2018 CE, the total area of glaciers within the study area (n = 219 in 1989) shrank by approximately 35 km2. Very small glaciers (<0.5 km2) show the highest relative rates of shrinkage, and 90% of mapped glaciers within the study area were less than 0.5 km2 in 2018 CE. 3601 GlacierMap: developing a citizen science glacier mapping tool for the Peruvian Caroline Clason, Sally Rangecroft, Gina Kallis, Shaun Lewin, Tom Mullier, Will Blake, Iain Stewart Corresponding author: Caroline Clason Corresponding author e-mail: [email protected]

Glaciers provide clear, visual evidence of environmental change, and are retreating in response to warming temperatures around the . In the tropical Andes of , glaciers act as buffers to available water supplies, essential to water, , and security downstream, and are especially important during the dry season. These direct, and indirect, impacts of glacier change receive considerably less attention from the global media in comparison to the melting of ice sheets and sea level rise, and the impacts of some other environmental issues. We have developed a web-based citizen science glacier mapping tool, GlacierMap, to help to raise awareness of these issues, particularly amongst secondary school pupils, and to contribute to increasing public backing for tackling, adapting to, and mitigating the impacts of climate change. Users of GlacierMap undertake an interactive learning experience by mapping a glacier from two different periods (1984 and 2018) from freely available Landsat data, visually demonstrating glacier retreat within the Cordillera Blanca, while learning more about the impacts of this retreat from information provided by the project. GlacierMap also allows us to assess the value of ‘crowd-sourcing’ of mapping data for the purposes of glacier monitoring and data generation, while providing an accessible teaching at a time where remote learning has become increasingly important. 3602 Is there a need to redefine the concept of for glaciated basins? Sally Rangecroft, Caroline Clason, Will Blake Corresponding author: Sally Rangecroft Corresponding author e-mail: [email protected]

Peak water is a term that has been increasingly used over the last decade to highlight concerns around changing freshwater . The term refers to a tipping point in time for a , where water availability reaches its maximum and thereafter is in continual decline. Using an example glaciated basin in the Peruvian Andes, the Rio Santa, we demonstrate that there is a need to rethink and redefine the approach and terminology used to communicate the study of peak water, particularly in glaciated basins. We propose that the location, timing, and scale of the discharge time series analysed to estimate peak water are extremely important, and require careful consideration to ensure that changes observed are from glacier melt alone, and are not intertwined with changes due to human activities or non-glacier fed freshwater inputs from the wider catchment. To address this, we suggest two new terms to help bring clarity to future research in this topic: i) ‘glacial peak water’ in the upper basin to represent changes in discharge specific to glacier meltwater only; and ii) ‘basin peak water’ further downstream to represent changes in discharge on the basin level, which intrinsically also includes human activities and inputs from precipitation and . It is important for management to know whether peak water has been reached or not in order to properly plan the apportionment of water in future, however, the variations in peak water from different locations and times may be relevant to different end users and stakeholders. Our proposed additions to the peak water terminology may help bring clarity to better inform water resource users and managers in the face of changing and challenging issues. 3603 Glaciers, rock glaciers and mountain landsystems Brian Whalley Corresponding author: Brian Whalley Corresponding author e-mail: [email protected]

The of rock glaciers has been in dispute for many years; what they are, what they represent and how they relate to climate change. The two views, ‘glacier origin’ or ‘permafrost origin’ still divide opinion but knowing which is ‘correct’ for paleoclimatic interpretation (and on Mars) but specifically for water resources in a warming climate. Recent developments in Google Earth can give a sufficiently good resolution to show even small rock glaciers and related features. This poster shows how Google Earth can be used to map, via a ‘landsystem’ approach, and distinguish between the models. Examples from Switzerland, and will be used as illustrations. 3604 Incorporating dh/dt measurements into ice flow model inversions: a case study on the Larsen C ice shelf Tom Mitcham, Hilmar Gudmundsson, Jonathan Bamber Corresponding author: Tom Mitcham Corresponding author e-mail: [email protected]

To model ice dynamics in real-world settings, inverse methods are often used to estimate the free parameters in the governing equations. Within the Úa ice flow model, used here, these parameters are the rate factor in Glen’s flow law, A, and the basal slipperiness coefficient, C, in a ‘Weertman’ sliding law. Typically, the inverse problem is solved by iteratively adjusting these parameters to minimize the mismatch between observations of ice velocity and modelled ice velocity. However, discrepancies between the velocity measurements and other data sets used in the model, such as ice thickness or bedrock topography, can result in unphysical, large amplitude, short wavelength dh/dt (rate of change of ice thickness) values in the subsequent forward runs. Existing techniques to combat this behaviour, and to provide initial model conditions close to a steady state, include model spin-up or relaxation runs and applying synthetic mass balances. Here, in a case study on the Larsen C ice shelf and its tributaries, dh/dt measurements are incorporated into the inversions such that the model output is tuned to fit both velocity and dh/dt observations. The resulting A and C fields, and the initial transient model behaviour, are compared with the those from inversions using velocity observations alone. The suitability of the resulting initial conditions for subsequent perturbation experiments are compared with those from other commonly used model initialisation techniques. 3605 Future mass balance of the Jim Jordan, Hilmar Gudmundsson, Adrian Jenkins, Bertie Miles, Stewart Jamieson, Chris Stokes Corresponding author: Jim Jordan Corresponding author e-mail: [email protected]

The East Antarctic Ice Sheet (EAIS) is the single largest potential contributor to future global mean sea level rise, containing a water mass equivalent of 53 m. Recent work has found the overall mass balance of the EAIS to be approximately in equilibrium, albeit with large uncertainties. However, changes in oceanic conditions have the potential to upset this balance. This could happen by both a general warming of the ocean and also by shifts in oceanic conditions allowing warmer water masses to intrude into ice shelf cavities.We use the Úa numerical ice-flow model, combined with ocean-melt rates parameterized by the PICO box mode, to predict the future contribution to global-mean sea level of the EAIS. Results are shown for the next 100 years under a range of emission scenarios and oceanic conditions on a region by region basis, as well as for the whole of the EAIS. 3606 Buoyant plume based melt rate parametrizations accounting for stratified ambient ocean and non-constant slope basal geometry Alexander T. Bradley, Robert Arthern, C. Rosie Williams Corresponding author: Alexander T. Bradley Corresponding author e-mail: [email protected]

Accurate projections of the future evolution of ice sheets rely on an adequate description of the underlying ocean to provide melt rates on the ice shelf base. Fully- coupled ice-ocean simulations are, however, computationally expensive, and are therefore often limited in their spatial detail or temporal extent. Simplified basal melt rate parametrizations are often used to alleviate this problem, removing the need for the ocean component of the model, which makes up the vast majority of the computational expense. The current state of the art basal melt rate parametrizations are based on the theory of two-dimensional buoyant plumes, but these parametrizations are limited since they do not account for any temperature or salinity stratification of the ambient ocean beyond the plume, and assume a linear ice shelf base geometry (or account for non-constant slopes in an ad-hoc fashion). We present details of a new basal melt rate parametrization, based on an asymptotic analysis of the equations describing two dimensional buoyant plumes, that accounts for non-constant slopes of the ice shelf base and two-layer stratification of the ambient ocean, where warm, salty bottom water is separated from , fresh upper water by a relatively thin pycnocline. Comparison with numerical solutions of the plume equations demonstrates that our parametrization represents an improvement on the current state of the art. 3608 Looking inside a mid-latitude glacier on Mars Frances E. G. Butcher, Neil S. Arnold, Dan C. Berman, Susan J. Conway, Joel M. Davis, Matt R. Balme Corresponding author: Frances E. G. Butcher Corresponding author e-mail: [email protected]

Putative debris-covered glaciers in Mars’ mid-latitudes formed ~2 Ma to 1 Ga. Their internal structure should record their flow histories, and thus information about the environmental conditions under which they evolved. We present observations of internal structures within a glacier in the Nereidum Montes region of Mars’ southern mid-latitudes (51.24°W, 42.53°S). The structures are visible in the wall of a gully, which has incised through the full length of the glacier. The structures dip up- glacier, and appear to connect flow-transverse arcuate foliations on the ice surface to the deep interior (and possibly the bed) of the glacier. We analyse the internal structures using a 1 m/pixel digital elevation model derived from 25 cm/pixel High Resolution Imaging Science Experiment (HiRISE) stereo-pair images, and a false- colour HiRISE image. We also input the DEM and an inferred glacier bed topography into the Ice Sheet System Model to simulate glacier flow assuming no and present-day mean-annual surface temperature (210K). The internal structures dip up-glacier at ~20° from the bed, and are spectrally ‘redder’ than the bulk glacier, which appears ‘bluer’. This could result from differences in debris-concentration and/or surface roughness. Our ice flow modelling experiments show that the internal structures occur in a zone of ice flow compression approaching the glacier terminus. We therefore suggest that the internal structures are analogous to fold or thrust-fault features found in compressional shear zones near to the termini of glaciers on Earth. Such features can be debris-rich, transporting englacial and/or subglacial debris to the surface. Flow-transverse surface foliations, such as those connected to the glacier-internal structures identified here, are common on the surfaces of glaciers across Mars’ mid-latitudes. Our observations suggest that foliations formed in zones of ice flow compression on Mars could contain a component of basal and/or englacial debris, which could be sampled (e.g., for astrobiological research), without the requirement for deep-drilling. 3609 The transferability of adjoint inversion products between different ice flow models Jowan Barnes, Thiago Dias dos Santos, Daniel Goldberg, Hilmar Gudmundsson, Mathieu Morlighem, Jan De Rydt Corresponding author: Jowan Barnes Corresponding author e-mail: [email protected]

Thwaites Glacier in West Antarctica is the subject of much investigation, due to rapid changes in the area over recent decades. Improving the understanding and functionality of modelling techniques used to make future predictions is a large part of this. Among the most important challenges faced by ice flow models is how to represent basal and rheological conditions which are challenging to obtain from direct observations. A common practice is to use numerical inversions to calculate estimates for the unknown properties, but there are many possible methods and no one standardized approach. As such, every ice flow model has a unique initialisation procedure. Here we compare the outputs of inversions from three different ice flow models, each employing a variant of adjoint-based optimisation to calculate basal sliding coefficients and flow rate factors using the same observed surface velocities and ice thickness distribution. We seek to answer two related questions; how much are the underlying physics represented in inversion outputs compared to an individual model’s numerical processes, and can the products of inversions be used outside their model of origin? We find that our inversions produce very similar distributions of basal sliding across all models, despite using different techniques, implying that the methods used are highly robust and represent the physics without much influence by individual model behaviours. Transferring the products of inversions between models results in time-dependent simulations displaying variability on the order of or lower than existing model intercomparisons and process studies. While the successful transfer of inversion outputs from one model to another requires some extra effort and technical knowledge of the particular models involved, it is certainly possible and could indeed be useful for future intercomparison projects. 3610 Quantifying uncertainty in future projections of ice loss from the Filchner- Ronne basin Emily Hill, Sebastian Rosier, Hilmar Gudmundsson, Mat Collins Corresponding author: Emily Hill Corresponding author e-mail: [email protected]

Mass loss from the Antarctic Ice Sheet is the main source of uncertainty in projections of future global sea-level rise. Despite advances in our understanding of feedbacks in the ice sheet-ice shelf-ocean system, future projections of ice loss remain poorly constrained in many parts of Antarctica. In particular, there is ongoing debate surrounding the future of the Filchner-Ronne (FR) basin. The FR basin has remained relatively unchanged in recent decades, but an increase in air and ocean temperatures in the neighbouring , could force rapid retreat in the near future. Here, we use an uncertainty quantification approach and the ice flow model Úa to understand the key physical processes and model variability in future projections of global mean sea level contribution (GMSL) from the FR basin. We first perform an ensemble of model simulations for four RCP warming scenarios using extensive random sampling of uncertain input parameters. Model responses (GMSL) were then used to create a surrogate model, which allow us to extensively sample the input parameter space to derive probabilistic projections of GMSL contribution from the FR basin. Our results indicate that the FR basin will have a predominantly negative contribution to sea level rise by the year 2300, owing to increases in accumulation with warming. However, under certain parameter combinations, and greater warming (RCP 8.5), there is the potential for positive contributions to sea level rise. We find that the main drivers of uncertainty in our projections are related to our forcing rather than model parameters, particularly related to percentage changes in precipitation and the scaling of atmospheric to ocean temperatures. Future work is needed to better constrain these parameters in order to produce more accurate projections of future mass loss from this region of Antarctica. 3615 Coupled ice-ocean simulations of the Amundsen Sea glaciers Jan De Rydt Corresponding author: Jan De Rydt Corresponding author e-mail: [email protected]

Ice-ocean interactions are regarded as one of the key processes that control the evolution of glaciers and ice streams in Antarctica. A careful representation of basal melting in numerical simulations of ice flow is therefore essential. The most comprehensive representation of ice-ocean interactions is through the coupling with a 3d ocean circulation model, which allows to capture the complex feedbacks between glacier dynamics and ambient ocean conditions. Results from one such model, Ua-MITgcm, are presented here, based on a regional configuration of the Amundsen Sea and its neighbouring glaciers in West Antarctica. We compare numerical simulations of ice-ocean dynamics between 1996 and 2016 to observations and explore the sensitivity of our results to the initial state of the ice sheet and a range of uncertain physical processes such as basal sliding. 3617 From steady streaming to oscillations: the role of subglacial drainage and temperate ice in ice stream dynamics Marianne Haseloff, Ian Hewitt, Richard Katz Corresponding author: Marianne Haseloff Corresponding author e-mail: [email protected]

The majority of Antarctic ice is discharged by fast-flowing ice streams. Some of these ice streams exhibit variations in velocities and ice stream discharge on decadal to centennial time scales, but the factors controlling these variations are still insufficiently understood. Observations indicate that the fast-flow of ice streams is enabled by meltwater lubricating the ice stream bed, and models suggest that this lubrication is the result of a between fast flow, heat dissipation at the ice stream bed and within the ice, and meltwater production. In particular, recent studies have shown that heat dissipation in temperate ice stream margins, which are regions of high lateral strain, can contribute significantly to the subglacial water balance. However, the role of this meltwater flux in ice stream dynamics remains unclear. Here, we investigate the roles of subglacial drainage and feedbacks between fast flow and heat dissipation in ice-stream evolution. The ice is modeled as a vertically uniform plug flow. Water flow at the bed is modelled as a diffusive system whose transmissivity increases with decreasing effective pressure. Dynamical feedbacks in the energy balance include both frictional heating along the bed and heat dissipation within the ice. Within our model, two distinct dynamic regimes can be identified: if the hydraulic permeability of the bed is sufficiently high to evacuate all meltwater produced at the ice stream bed and in its margins, a moderately-fast steadily flowing ice stream forms. Conversely, `binge-purge’ oscillations between fast and stagnant flow emerge when the hydraulic permeability of the bed is too low to evacuate the meltwater produced by the ice stream. Topogrophic controls can suppress this oscillatory behaviour, while the inclusion of temperate ice dynamics amplifies it