Firn aquifers are areas of subsurface meltwater storage that form in glaciated regions where intense summer surface melting and high annual snowfall occur. Firn aquifers can induce hydrofracturing and thereby trigger instability, potentially accelerating ice flow in glaciers, or inducing rapid calving in ice shelves. Although they have been known on mountain glaciers for many decades, widespread firn aquifers, 5 to 50 m below the surface, have only recently been discovered in Greenland. A mapping algorithm using a new satellite microwave remote sensing technique correctly identified the presence of firn aquifers in Greenland, and was used to map their full extent there. Antarctica’s potential for firn aquifers is at present unconfirmed, yet application of a similar technique indicates that they likely exist in coastal and ice shelf regions that have climate conditions similar to firn aquifer areas in Greenland.

Here we will present initial evidence identifying firn aquifers on several ice shelves in Antarctica. We have generated high spatial resolution (~1 km) ice sheet-wide firn aquifer maps over the continent using several spaceborne sensors. Our retrieval approach combines a simple radiative transfer model that integrates ice sheet geophysical parameters derived from past field observations with time series of multi-frequency microwave brightness temperature (37 GHz, 1.4 GHz) and radar backscatter (1.2 GHz) collected from the active and passive microwave instruments aboard NASA’s Soil Moisture Active Passive (SMAP) satellite, and the Advanced Microwave Scanning Radiometer (AMSR-E) aboard NASA’s Aqua satellite.

Two key sites in the Antarctic Peninsula will be surveyed during a November 2018 exploratory field campaign: 1) the Wilkins Ice Shelf, and (2) the southern George VI Ice Shelf. Field observations to be attempted include snow and firn characteristics derived from shallow firn cores (~60 m), and snow pits (~2 m), and radio echo soundings from a dual-frequency (30 MHz, 250 MHz) Ground Penetrating Radar (GPR) and an Autonomous phase-sensitive Radio Echo Sounder (ApRES) operating in the frequency range of 200–400 MHz. Two years of weather data over the inferred firn aquifer sites are anticipated from the installation of an Automated Meteorology-Ice-Geophysics Observing System (AMIGOS) that includes a thermistor string, camera system, Global Positioning System (GPS), and albedo-based surface melting detection.

Previous studies have suggested that the key reservoirs initiating hydrofracturing on ice shelves are meltwater ponds formed on impermeable ice layers at the surface. However, if present, firn aquifers represent alternate reservoirs capable of storing significantly larger volumes of meltwater year-round within porous layers of recrystallized snow. Thus, the formation of firn aquifers – rather than meltwater ponds - may rapidly induce ice shelf instability in high-melt, high-accumulation, costal and ice shelf regions of Antarctica.

The Greenland Ice Sheet has lost mass at an accelerating rate over the past two decades. Approximately 60% of this mass loss can be attributed to surface melt which rose to record levels in 2012. Each summer, these surface meltwaters drain from the ice-sheet surface to the bed via moulins and crevasses. Initial research postulated that as these seasonal meltwaters access the bed, they would enhance glacier sliding, indicative of a positive feedback between surface melt and ice velocity. However, recent findings from a land-terminating sector of the ice sheet contend that enhanced surface melt causes expansion of a channelised sub-glacial drainage system, causing a reduction in glacier velocity as a result of a regional decrease in sub-glacial water pressure. Conversely, additional work has argued that for the period 2000/01-2016/17, little significant change in ice velocity occurred, raising the possibility that the previously observed slowdown may be related to local processes rather than regional-scale surface melt forcing.

Since the record surface melt of 2012, a period of relative mass balance stability has been observed across Greenland, with 2017 having the lowest maximum surface melt extent since 1996. This is coincident with positive Arctic and North Atlantic Oscillations, promoting cyclonic conditions and so reduced incoming solar radiation and enhanced precipitation. Given that the proposed mechanism for a long-term velocity slowdown requires a continual increase in surface melt, ice motion would be expected to stabilise during a period of decreasing surface melt.

Here we present observations of ice velocity, derived from feature tracking of over 2500 pairs of Landsat scenes, from 1994-2018 for an area of ~11,000 km² in SW Greenland, extending along ~350 km of the margin and inland to an elevation of ~1000m. Our data shows a period of dynamic stability in the mid-late 1990s, followed by a clear slowdown trend (mean -1.66 m/yr²; p < 0.05) for the period 2002/03 to 2011/12 which is spatially pervasive across the entire study area. From 2011/12 to 2017/18, ice velocities remain relatively stable, coincident with recent reductions in surface melt production and less negative surface mass balance. These observations are consistent with a response of the basal drainage system to variable meltwater forcing on a multi-annual scale, whereby reduced surface melt production halts the expansion of subglacial channelisation allowing ice velocities to stabilise.

One of the major difficulties in predicting sea-level rise in a warming climate is the dynamic process of mass loss due to calving of ice shelves and glaciers. Although this process accounts for almost all the mass loss from the Antarctic Ice Sheet and almost half that from the Greenland Ice Sheet, it is still inadequately represented in ice-sheet models.
The NERC funded project Calving Laws for Ice Sheet Models (CALISMO) aims to improve the parameterisation of calving in ice-sheet models by using a hierarchy of numerical models. The discrete nature of calving means that calving cannot be represented explicitly in continuum ice-sheet models, but is instead parameterised using, for example, empirical crevasse depth laws, strain thresholds or ice-cliff failure criteria.
Although these calving laws succeed in replicating much of the observed calving behaviour, they require tuning and therefore introduce large uncertainties when extrapolated into the future. One of the objectives of CALISMO is to gain an in-depth understanding of the physical processes of calving by periodically exchanging glacier geometry, velocity fields and friction coefficients between a coupled ocean circulation--continuum ice-flow model (MITgcm—ELMER/Ice) and a discrete element fracture mechanics model (HiDEM), which simulates the calving.
The ready availability and quality of repeat satellite observations of the cryosphere is playing an important role in initialising and validating the modelling process. We aim to observe and model a wide range of glacier and ice-shelf geometries, in different oceanographic and climate settings, from large floating Antarctic ice shelves to fast-flowing Greenland glaciers. We use ESA Sentinel 1 imagery to record ice front behaviour, to feature track surface velocity and to monitor calving rates. We also use DLR TanDEM-X data to create DEMs of the calving fronts. SAR and optical imagery are also used to inspect the formation and break-up of fjord ice and mélange, to identify meltwater plumes adjacent to the glacier fronts, and to resolve crevassed shear bands within glacier calving fronts.
Examples will be presented of many of these observations together with their inclusion and influence on model development and behaviour.

We present a method to combine satellite altimetry (CryoSat-2) and satellite gravimetry (GRACE) observations to obtain a high-resolution ice-mass balance for Antarctica. We take into account the different, to some extent complementary signal and noise characteristics of each satellite observation with respect to the spatial wavelength. We include uncertainties introduced by the processing and corrections, such as the choice of the re-tracking algorithm and the snow/ice volume density model (CryoSat-2) or the filtering of correlated errors and the correction for glacial-isostatic adjustment (GRACE). We apply our method to the Antarctic ice sheet and the time period 2011-2017, in which CryoSat-2 and GRACE were simultaneously operational. We show that our map of ice-mass trends is consistently evaluable across all spatial scales and more accurate than estimates based on a single satellite mission, yielding a data set potentially useful for data assimilation into ice-sheet models.

Land ice is the largest single contributor to eustatic sea level rise, with an increasingly large contribution over the last 15 years due primarily to increased mass losses from the Greenland Ice Sheet (GrIS). Additionally, future model projections of ice sheet change form the largest uncertainties in projecting future sea level. As a result, it is imperative to have accurate and robust estimates of contemporary ice sheet mass balance in order provide closure to the sea level budget and for use as initial conditions in future forward model projections.

Three independent approaches currently exist for determining ice sheet mass balance: altimetry, gravimetry and the mass budget technique. Each technique measures ice sheet mass balance at differing spatial and temporal resolutions using diverse observational datasets. Additionally, each technique relies on geophysical forward model output (e.g. GIA and SMB models) to resolve for unobserved processes, which may introduce sources of uncertainty and biases into the results that are hard to quantify.

To overcome these challenges, we present current progress on the application of a Bayesian Hierarchical Model (BHM) approach to determine ice sheet mass balance trends for both Greenland and Antarctica since 2003. The BHM approach allows for the determination of basin scale mass trends in addition to the relative contribution by each driving process: namely, surface mass balance, ice dynamics, glacial isostatic adjustment and firn compaction. The framework uses a novel source separation approach to attribute a wide range of observations including: altimetry (ENVISAT, ICESat and CryoSat-2, NASA Laser ATM), GRACE and in-situ observations (GPS uplift rates and SMB measurements) to the driving processes causing the observed mass change. Forward geophysical models are only used to aid source separation, making the solutions predominantly data-driven. As a result, the output from the BHM approach can be used as an independent validation of these geophysical models. A key advantage of this approach is its flexibility in being able to combine diverse and discontinuous observations in space and time to produce a single reconciled estimate of ice sheet mass change.

Additionally, an overview of future opportunities for the framework will be discussed. This will include how the framework can utilize observations from the recently launched ICESat-2, GRACE-FO and Sentinel-3 missions in order to extend the time series into the next decade. Doing so will providing a continuous record of ice sheet mass change reconciling observations from past and future satellite missions. These long-term observational records are essential for understanding the multi-decadal ice sheet response to climate forcing.

This work is part of the wider GlobalMass (www.globalmass.eu) European Research Council five-year project, based at the University of Bristol.

The glaciers flowing into West Antarctica’s Amundsen Sea Embayment (ASE) have been characterized by significant changes during the last decades including acceleration, grounding line retreat and a positive mass imbalance. Here we measure Thwaites glacier draining the Amundsen Sea sector of West Antarctica grounding line retreat, ice shelf thinning and melting at unprecedented spatial and temporal resolution using COSMO-SkyMed, TanDEM-X satellite radar interferometry and World-View high resolution optical imagery spanning 2011 -2017. Comparison with previous ERS data from 2011 show a maximum retreat of 5-7 km across the main Thwaites glacier tongue and Thwaites Eastern ice shelf (TEIS) corresponding to an increased retreat rate of 0.5 and 1.5 km/yr respectively. Grounding line retreat has been fueled by the enhanced intrusion of warm, salty, subsurface ocean water of circumpolar deep water origin onto the continental shelf, beneath the floating ice shelf, to reach the glacier grounding zone and melt it from below at rates varying from 50 to 190 m/yr. The retreat rate varies heterogeneously depending on the magnitude of ice melt by the ocean, the rate of ice thinning and the shape of the glacier surface and bed topography. Overall we delineate a detailed mechanism of retreat where: basal melting occurs at the grounding line extending few km upstream where salty water infiltrates due to the tidally induced grounding line migrations. Heterogeneous basal melting triggers non-uniform grounding line retreat following the bed topography slopes. The opening of small channels (<1.2 ± 0.2 km) at the ocean-ice interface triggers 7±0.3 km grounding line retreat for a total ungrounded area of about 33±3 *106 m2. The ungrounding is not instantaneous and in the case of TEIS sees the formation of salty subglacial lake first that further boost the opening of subglacial channels. At Thwaites main trunk instead we record instantaneous grounding line retreat on areas where the glacier lies below the height above flotation subsequent to nearby melting at the grounding line.