Improving surface melt estimation over the Antarctic Ice Sheet using deep learning: a proof of concept over the Larsen Ice Shelf

Accurately estimating the surface melt volume of the Antarctic Ice Sheet is challenging and has hitherto relied on climate modeling or observations from satellite remote sensing. Each of these methods has its limitations, especially in regions with high surface melt. This study aims to demonstrate the potential of improving surface melt simulations with a regional climate model by deploying a deep learning model. A deep-learning-based framework has been developed to correct surface melt from the regional atmospheric climate model version 2.3p2 (RACMO2), using meteorological observations from automatic weather stations (AWSs) and surface albedo from satellite imagery. The framework includes three steps: (1) training a deep multilayer perceptron (MLP) model using AWS observations, (2) correcting Moderate Resolution Imaging Spectroradiometer (MODIS) albedo observations, and (3) using these two to correct the RACMO2 surface melt simulations. Using observations from three AWSs at the Larsen B and C ice shelves, Antarctica, cross-validation shows a high accuracy (root-mean-square error of 0.95 mm w.e. d−1, mean absolute error of 0.42 mm w.e. d−1, and a coefficient of determination of 0.95). Moreover, the deep MLP model outperforms conventional machine learning models and a shallow MLP model. When applying the trained deep MLP model over the entire Larsen Ice Shelf, the resulting corrected RACMO2 surface melt shows a better correlation with the AWS observations for two out of three AWSs. However, for one location (AWS 18), the deep MLP model does not show improved agreement with AWS observations; this is likely because surface melt is largely driven by factors (e.g., air temperature, topography, katabatic wind) other than albedo within the corresponding coarse-resolution model pixels. Our study demonstrates the opportunity to improve surface melt simulations using deep learning combined with satellite albedo observations. However, more work is required to refine the method, especially for complicated and heterogeneous terrains.

Generation and fate of basal meltwater during winter, western Greenland Ice Sheet

Basal sliding in the ablation zone of the Greenland Ice Sheet is closely associated with water from surface melt introduced to the bed in summer, yet melting of basal ice also generates subglacial water year-round. Assessments of basal melt rely on modeling with results strongly dependent upon assumptions with poor observational constraints. Here we use surface and borehole measurements to investigate the generation and fate of basal meltwater in the ablation zone of Isunnguata Sermia basin, western Greenland. The observational data are used to constrain estimates of the heat and water balances, providing insights into subglacial hydrology during the winter months when surface melt is minimal or nonexistent. Despite relatively slow ice flow speeds during winter, the basal meltwater generation from sliding friction remains manyfold greater than that due to geothermal heat flux. A steady acceleration of ice flow over the winter period at our borehole sites can cause the rate of basal water generation to increase by up to 20 %. Borehole measurements show high but steady basal water pressure rather than monotonically increasing pressure. Ice and groundwater sinks for water do not likely have sufficient capacity to accommodate the meltwater generated in winter. Analysis of basal cavity dynamics suggests that cavity opening associated with flow acceleration likely accommodates only a portion of the basal meltwater, implying that a residual is routed to the terminus through a poorly connected drainage system. A forcing from cavity expansion at high pressure may explain observations of winter acceleration in western Greenland.

Improved representation of the contemporary Greenland ice sheet firn layer by IMAU-FDM v1.2G

The firn layer that covers 90 % of the Greenland ice sheet (GrIS) plays an important role in determining the response of the ice sheet to climate change. Meltwater can percolate into the firn layer and refreeze at greater depths, thereby temporarily preventing mass loss. However, as global warming leads to increasing surface melt, more surface melt may refreeze in the firn layer, thereby reducing the capacity to buffer subsequent episodes of melt. This can lead to a tipping point in meltwater runoff. It is therefore important to study the evolution of the Greenland firn layer in the past, present and future. In this study, we present the latest version of our firn model, IMAU-FDM (Firn Densification Model), with an application to the GrIS. We improved the density of freshly fallen snow, the dry-snow densification rate and the firn's thermal conductivity using recently published parameterizations and by calibrating to an extended set of observations of firn density, temperature and liquid water content at the GrIS. Overall, the updated model settings lead to higher firn air content and higher 10 m firn temperatures, owing to a lower density near the surface. The effect of the new model settings on the surface elevation change is investigated through three case studies located at Summit, KAN-U and FA-13. Most notably, the updated model shows greater inter- and intra-annual variability in elevation and an increased sensitivity to climate forcing.

Impact of radiation penetration on Antarctic surface melt and subsurface snow temperatures in RACMO2.3p3

This study investigates the sensitivity of modeled surface melt and subsurface heating on the Antarctic ice sheet to a new spectral snow albedo and radiative transfer scheme in the Regional Atmospheric Climate Model (RACMO2), version 2.3p3 (Rp3). We tune Rp3 to observations by performing several sensitivity experiments and assess the impact on temperature and melt by changing one parameter at a time. When fully tuned, Rp3 compares well with in situ and remote sensing observations of surface mass and energy balance, melt, temperature, albedo and snow grain specific surface area. Furthermore, the introduction of subsurface heating in Rp3 significantly improves the snow temperature profile. Near surface snow temperature is especially sensitive to the prescribed fresh snow specific surface area and fresh dry snow metamorphism. These processes, together with the refreezing grain size and subsurface heating, are important for melt around the margins of the Antarctic ice sheet. Moreover, small changes in the albedo and the aforementioned processes can lead to an order of magnitude overestimation of melt, locally leading to runoff and a reduced surface mass balance.

Clouds drive differences in future surface melt over the Antarctic ice shelves

Recent warm atmospheric conditions have damaged the ice shelves of the Antarctic Peninsula through surface melt and hydrofracturing, and could potentially initiate future collapse of other Antarctic ice shelves. However, model projections with similar greenhouse gas scenarios suggest large differences in cumulative 21st century surface melting. So far it remains unclear whether these differences are due to variations in warming rates in individual models, or whether local surface energy budget feedbacks could also play a notable role. Here we use the polar-oriented regional climate model MAR to study the physical mechanisms that will control future melt over the Antarctic ice shelves in high-emission scenarios RCP8.5 and SSP585. We show that clouds enhance future surface melt by increasing the atmospheric emissivity and longwave radiation towards the surface. Furthermore, we highlight that differences in meltwater production for the same climate warming rate depend on cloud properties and particularly cloud phase. Clouds containing a larger amount of liquid water lead to stronger melt, subsequently favouring the absorption of solar radiation due to the snow-melt-albedo feedback. By increasing melt differences over the ice shelves in the next decades, liquid-containing clouds could be a major source of uncertainties related to the future Antarctic contribution to sea level rise.

Projected increases in surface melt and ice loss for the Northern and Southern Patagonian Icefields

The Northern Patagonian Icefield (NPI) and the Southern Patagonian Icefield (SPI) have increased their ice mass loss in recent decades. In view of the impacts of glacier shrinkage in Patagonia, an assessment of the potential future surface mass balance (SMB) of the icefields is critical. We seek to provide this assessment by modelling the SMB between 1976 and 2050 for both icefields, using regional climate model data (RegCM4.6) and a range of emission scenarios. For the NPI, reductions between 1.5 m w.e. (RCP2.6) and 1.9 m w.e. (RCP8.5) were estimated in the mean SMB during the period 2005–2050 compared to the historical period (1976–2005). For the SPI, the estimated reductions were between 1.1 m w.e. (RCP2.6) and 1.5 m w.e. (RCP8.5). Recently frontal ablation estimates suggest that mean SMB in the SPI is positively biased by 1.5 m w.e., probably due to accumulation overestimation. If it is assumed that frontal ablation rates of the recent past will continue, ice loss and sea-level rise contribution will increase. The trend towards lower SMB is mostly explained by an increase in surface melt. Positive ice loss feedbacks linked to increasing in meltwater availability are expected for calving glaciers.

Central tropical Pacific convection drives extreme high temperatures and surface melt on the Larsen ice shelf

The northern parts of the Larsen ice shelf, eastern Antarctic Peninsula (AP) have experienced dramatic break-up since the early 1990s as a result of strong summertime surface melt, which has been linked to stronger circumpolar westerly winds. Here we show extreme summertime surface melt and high temperatures over the eastern AP and Larsen C ice shelf occur because of enhanced deep convection in the central tropical Pacific, which produces cyclonic conditions across the middle and high-latitude South Pacific, and a strong high pressure anomaly over Drake Passage. Together these transport extreme heat and moisture from low latitudes to the AP, at times in the form of "atmospheric rivers", producing strong surface warming and melt on the Larsen ice shelf by the Foehn effect. Therefore, variability in central tropical Pacific convection is crucial for interpreting past and projecting future AP surface mass balance and extreme temperature events.

Influence of Arctic sea-ice loss on the Greenland ice sheet climate

The Arctic is the region on Earth that is warming the fastest. At the same time, Arctic sea ice is reducing while the Greenland ice sheet (GrIS) is losing mass at an accelerated pace. Here, we study the seasonal impact of reduced Arctic sea ice on GrIS surface mass balance (SMB), using the Community Earth System Model version 2.1 (CESM2), which features an advanced, interactive calculation of SMB. Addressing the impact of sea-ice reductions on the GrIS SMB from observations is difficult due to the short observational records. Also, signals detected using transient climate simulations may be aliases of other forcings. Here, we analyze dedicated simulations from the Polar Amplification Model Intercomparison Project with reduced Arctic sea ice and compare them with preindustrial sea ice simulations while keeping all other forcings constant. In response to reduced sea ice, the GrIS SMB increases in winter due to increased precipitation, driven by the more humid atmosphere and increasing cyclones. In summer, surface melt increases due to a warmer, more humid atmosphere providing increased energy transfer to the surface through the sensible and latent heat fluxes, which triggers the melt-albedo feedback. Further, warming occurs throughout the entire troposphere over Baffin Bay. This deep warming results in regional enhancement of the 500 hPa geopotential heights over the Baffin Bay and Greenland, increasing blocking and heat advection over the GrIS’ surface. This anomalous circulation pattern has been linked to recent increases in the surface melt of the GrIS.