"The Antarctic and Greenland ice sheets are major contributors to today’s sea level budget
and their future evolution constitute the largest source of uncertainty when projecting global
sea level change under future emission and socio-economic pathways. Both ice sheets
interact strongly with neighbouring systems, reacting and impacting with the atmosphere and
ocean, affecting circulation, energy budget or bio-chemical cycles.
Understanding and replicating this complex and integrated system, and projecting its
behaviour under changing conditions, requires a vast quantity of observations, coupled with
powerful algorithms and numerical simulation combining artificial intelligence and physics.
It is in this context that the 4DAntarctica, 4DGreenland, Digital Twin Antarctica and Digital
Twin Greenland projects have been conceived, seeking to advance the current state of
knowledge on the hydrology of both the Greenland and Antarctic Ice Sheet, by capitalising
on the latest advances in Earth Observation data, AI, and numerical simulation. The digital
twin component of the two 4D projects are seeking to demonstrate how these various
elements can be consolidated into a dynamic, digital replica of our planet which accurately
mimics Earth’s behaviour and responds to the challenges set by Europe’s on the Destination
Earth initiative.
In this presentation we will synthesize the work performed as part of these 4 projects and
discuss specific avenues to further develop digital twins of the ice sheets, further exploiting
the synergies between observations, models and artificial intelligence."

"Ice temperature within the ice is a crucial characteristic to understand the Antarctic ice sheet evolution because temperature is coupled to ice flow. Since temperature is only measured at few locations in deep boreholes, we only rely on numerical modelling to assess ice sheet-wide temperature. However, the design of such models leads to a number of challenges. One important difficulty is that the temperature field strongly depends on the geothermal flux which is still poorly known (Burton-Johnson et al, 2020). Another point is that up to now there is no fully suitable model, especially for inverse approaches: i)analytical solutions are only valid in slowly flowing regions; ii)models solving only the heat equation by prescribing geometry and ice flow do not take into account the past changes in ice thickness and ice flow and do not couple ice flow and temperature. Conversely, 3D thermomechanical models that simulate the evolution of the ice sheet take into account all the relevant processes but they are too computationally expensive to be used in inverse approaches. Moreover, they do not provide a perfect fit between observed and simulated geometry (ice thickness, surface elevation) for the present-day ice sheets and this affects the simulated temperature field. In order to speed-up the simulation of present-day ice temperature a numerical emulator based on deep neural network (DNN), of the thermomechanically coupled ice sheet models GRISLI (Quiquet et al. 2018) has been developed. We use GRISLI outputs that come from 4 simulations, each covers 900000 years (8 glacial-interglacial cycles) to get rid of the initial configuration influence. The simulations differ by the geothermal flux map used as boundary condition. Finally, a database is built where each ice column for each simulation is a sample used to train the DNN. For each sample, the input layer (precursor) is a vector of the present-day characteristics: ice thickness, surface temperature, geothermal flux, accumulation rate, surface velocity and surface slope. The predicted output (output layer) is the vertical profile of temperature. In the training, the weights of the network are optimized by comparison with the GRISLI temperature.
The first results are very encouraging with a RMSE of ~ 0.6 °C .The computational time of GRISLI-DNN for generating temperature field of whole Antarctica (16000 columns) is about 20 s.
The first application of the emulator is to associate it to ESA’s SMOS satellite observations to infer the 3D temperature field and improve our knowledge of geothermal flux. Indeed, it has been shown that SMOS data, coupled with glaciological and electromagnetic models, give an indication of temperature in the upper 1000 m of the ice sheet. Studies to use the same concept to lower microwave frequencies (i.e. 0.4 GHz) and expand the temperature retrieval capabilities are also on-going. Moreover, the emulator could also be used for initialization of computationally expensive ice sheet models."

Beneath the ice sheets of Greenland and Antarctica lies an extensive hydrological system, which includes networks of often highly dynamic and interconnected subglacial lakes. These lakes have the capacity to store, and episodically release, meltwater, thereby modulating the flow of water beneath the ice and, at times, altering the dynamics of the overlying ice sheet itself. Subglacial lakes beneath the Greenland Ice Sheet are of particular interest because, unlike their Antarctic counterparts, they are likely to be more closely connected to the surface hydrological system. Thus, as Earth’s climate warms, generating increasing fluxes of surface meltwater, it may be expected that the distribution and dynamics of Greenland’s subglacial lakes may also evolve too. Whilst the network of Antarctic subglacial lakes is relatively well studied, very little is known about the existence and dynamics of lakes beneath the Greenland Ice Sheet. Despite theoretical predictions suggesting that more than 1600 lakes may exist, only a tiny number (64) have been identified to date, with less than 10 of these having been observed to actively discharge water. This paucity of observations is primarily due to the smaller size of Greenlandic subglacial lakes, which presents an observational challenge for traditional altimetry-based satellite techniques, that offer relatively coarse resolution or spatial sampling. Here we assess the use of new streams of high resolution satellite data to reveal insight into Greenland’s subglacial lakes, and consider the potential for these datasets to inform subglacial hydrological models within a future Digital Twin of Greenland. Specifically, we present the results and lessons learned from a pilot study that analysed 35,000 super high resolution (2 metre) Digital Elevation Models to search for signatures of subglacial lake drainage and filling. Additionally, we consider the value added by complementary data streams, including estimates of surface deformation derived from Synthetic Aperture Radar imagery and altimetry, to consider how a future Greenland Digital Twin may leverage large and diverse datasets in order to inform models and, in turn, deliver new insight into Greenland’s elusive subglacial hydrological system.

The Antarctic Ice sheet is a key component of the Earth system, impacting on global sea level, ocean circulation and atmospheric processes. Meltwater is generated at the ice sheet base primarily by geothermal heating and friction associated with ice flow, and this feeds a vast network of lakes and rivers creating a unique hydrological environment. Subglacial lakes play a fundamental role in the Antarctic ice sheet hydrological system because outbursts from ‘active’ lakes can trigger, (i) change in ice speed, (ii) a burst of freshwater input into the ocean which generates buoyant meltwater plumes, and (iii) evolution of glacial landforms and sub-glacial habitats. Despite the key role that sub-glacial hydrology plays on the ice sheet environment, there are limited observations of repeat sub-glacial lake activity resulting in poor knowledge of the timing and frequency of these events. Even rarer are examples of interconnected lake activity, where the draining of one lake triggers filling of another. Observations of this nature help us better characterise these events and the impact they may have on Antarctica’s hydrological budget, and will advance our knowledge of the physical mechanism responsible for triggering this activity. In this study we analyse 9-years of CryoSat-2 radar altimetry data, to investigate a newly identified sub-glacial network in the Amery basin, East Antarctica. CryoSat-2 data was processed in ‘swath mode’, increasing the density of elevation measurements across the study area. The plane fit method was employed in 500 m by 500 m grid cells, to measure surface elevation change at relatively high spatial resolution. We identified a network of 10 active subglacial lakes in the Amery basin. 7 of these lakes, located below Lambert Glacier, show interconnected hydrological behaviour, with filling and drainage events throughout the study period. We observed ice surface height change of up to 6 meters on multiple lakes, and these observations were validated by independently acquired TanDEM-X DEM differencing. We then use these observations in conjunction with simulation of the melt water production and transport at the base of the ice sheet and across the grounding line to constrain sub-glacial fluxes and explore its impact on ocean circulation and ice-ocean interaction. This case study is an important decade long record of hydrological activity beneath the Antarctic Ice Sheet which demonstrates the importance of high resolution swath mode measurements. In the future the Lambert lake network will be used to better understand the filling and draining life cycle of sub-glacial hydrological activity under the Antarctic Ice Sheet.

"Around the periphery of the Greenland and Antarctic Ice Sheets, networks of supraglacial lakes and streams form each summer, in response to seasonal surface melting. The nature, extent and dynamics of these surface hydrological systems are important because it affects the transport of freshwater towards the coast, and can impact upon factors such as ice dynamics. With the launch of operational missions carrying optical sensors, such as Sentinel-2, there is the opportunity to monitor this system at weekly periodicity around the entirety of the ice sheet margins each summer. However, conventional mapping approaches require extensive manual post-processing to remove false positives, which makes them infeasible within the context of a Digital Twin.
Here, we consider more automated approaches, which have been investigated within the 4D Greenland and 4D Antarctica studies, and that are better suited for implementation within a future Digital Twin. Specifically, we evaluate the potential of Machine Learning approaches, including a Random Forest algorithm, which are trained to classify surface water from non-water features in a pixel-based classification. We will assess their performance relative to conventional approaches, including their spatial and temporal transferability, and investigate performance across the margin of both the Greenland and Antarctic Ice Sheets. Finally, we shall consider the computational requirements and lessons learnt during this work, within the context of a future Ice Sheet Digital Twin."

"The ESA 4DGreenland project has the objective of performing an integrated assessment of Greenland’s hydrology through maximizing the use of Earth observation (EO) data. Not all aspects of Greenland hydrology are observable from EO, and models are still required to close the integrated assessment. Here, we give initial thoughts on how we can build on the vast observational datasets generated within 4DGreenland and progress towards a digital twin of the Greenland ice sheet to bridge the gap between models and EO data.
The diversity of the gathered EO dataset provides an ideal playground for investigating hidden features within the data using AI or providing needed standardized training data for feature modeling. One such example, is the use of auxiliary EO datasets in the conversion of radar altimeter derived volume change of the Greenland ice sheet into mass balance. This study showed the prospects of data fusion to elevate our knowledge of ice sheet behavior and relay less on numerical earth system models. In this presentation we outline the requirements for such a digital twin, and how we foresee that EO data and models can be combined to fulfil those requirements.
"