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.

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.

Improving Surface Melt Estimation over Antarctica Using Deep Learning: A Proof-of-Concept over the Larsen Ice Shelf

Accurately estimating surface melt volume of the Antarctic Ice Sheet is challenging, and has hitherto relied on climate modelling, or on 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 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 = 0.95 mm w.e. per day, mean absolute error = 0.42 mm w.e. per day, and coefficient of determination = 0.95). Moreover, the deep MLP model outperforms conventional machine learning models (e.g., random forest regression, XGBoost) 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, likely due to the heterogeneous drivers of melt 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. On the other hand, more work is required to refine the method, especially for complicated and heterogeneous terrains.

Estimating Surface Melt on the Larsen Ice Shelf Using a Deep Neural Network: Opportunities and Challenges

<p>Presently, surface melt over Antarctica is estimated using climate modeling or remote sensing. However, accurately estimating surface melt remains challenging. Both climate modeling and remote sensing have limitations, particularly in the most crucial areas with intense surface melt.  The motivation of our study is to investigate the opportunities and challenges in improving the accuracy of surface melt estimation using a deep neural network. The trained deep neural network uses meteorological observations from automatic weather stations (AWS) and surface albedo observations from satellite imagery to improve surface melt simulations from the regional atmospheric climate model version 2.3p2 (RACMO2). Based on observations from three AWS at the Larsen B and C Ice Shelves, cross-validation shows a high accuracy (root mean square error = 0.898 mm.w.e.d<sup>−1</sup>, mean absolute error = 0.429 mm.w.e.d<sup>−1</sup>, and coefficient of determination = 0.958). The deep neural network also outperforms conventional machine learning models (e.g., random forest regression, XGBoost) and a shallow neural network. To compute surface melt for the entire Larsen Ice Shelf, the deep neural network is applied to RACMO2 simulations. The resulting, corrected surface melt shows a better correlation with the AWS observations in AWS 14 and 17, but not in AWS 18. Also, the spatial pattern of the surface melt is improved compared to the original RACMO2 simulation. A possible explanation for the mismatch at AWS 18 is its complex geophysical setting. Even though our study shows an opportunity to improve surface melt simulations using a deep neural network, further study is needed to refine the method, especially for complicated, heterogeneous terrain.</p>

Microbial Fe(III) reduction as a potential iron source from Holocene sediments beneath Larsen Ice Shelf

Recent recession of the Larsen Ice Shelf C has revealed microbial alterations of illite in marine sediments, a process typically thought to occur during low-grade metamorphism. In situ breakdown of illite provides a previously-unobserved pathway for the release of dissolved Fe2+ to porewaters, thus enhancing clay-rich Antarctic sub-ice shelf sediments as an important source of Fe to Fe-limited surface Southern Ocean waters during ice shelf retreat after the Last Glacial Maximum. When sediments are underneath the ice shelf, Fe2+ from microbial reductive dissolution of illite/Fe-oxides may be exported to the water column. However, the initiation of an oxygenated, bioturbated sediment under receding ice shelves may oxidize Fe within surface porewaters, decreasing dissolved Fe2+ export to the ocean. Thus, we identify another ice-sheet feedback intimately tied to iron biogeochemistry during climate transitions. Further constraints on the geographical extent of this process will impact our understanding of iron-carbon feedbacks during major deglaciations.

Clay Mineralogical Characteristics of Sediments Deposited during the Late Quaternary in the Larsen Ice Shelf B Embayment, Antarctica

Variations in grain size, clay mineral composition, and stable isotopes (δ13C and δ15N) are closely linked to the sedimentary facies that reflect mineralogical and geochemical modification during the retreat and advance of the Larsen ice shelf. A whole round core of marine sediment (EAP13-GC17, 236 cm below the sea floor) was collected on the northwestern Larsen B embayment of the Antarctic Peninsula during a marine geological expedition (the ARA13 Cruise Expedition by the Korea Polar Research Institute, 2013). Four sedimentary facies (U1–U4) were clearly distinguishable: bioturbated sandy mud (open marine, U1), laminated sandy mud (sub–floating ice shelf, U2), sandy clay aggregates (deglacial, U3), and muddy diamictons (sub-glacial, U4), as well as interbedded silty. Clay minerals, including smectite, chlorite, illite, and kaolinite, were detected throughout the core. An increase in the clay mineral ratio of smectite/(illite + chlorite) was clearly observed in the open marine condition, which was strongly indicated by both a heavier isotopic composition of δ13C and δ15N (−24.4‰ and 4.3‰, respectively), and an abrupt increase in 10Be concentration (~30 times). An increase in the average values of the crystal packet thickness of illite (~1.5 times) in U1 also indicated sediments transported in open marine conditions. Based on the clay mineral composition in U1, the sediments are likely to have been transported from the Weddell Sea. The clay mineralogical assessments conducted in this region have significant implications for our understanding of paleodepositional environments.