Project Information
Collaborative Research: EAGER: Generation of high resolution surface melting maps over Antarctica using regional climate models, remote sensing and machine learning
Short Title:
Generation of high resolution surface melting maps over Antarctica using regional climate models, remote sensing and machine learning
Start Date:
End Date:
Surface melting and the evolution of the surface hydrological system on Antarctica ice shelves modulate the ice sheet mass balance. Despite its importance, limitations still exist that preclude the scientific community from mapping the spatio-temporal evolution of the surface hydrological system at the required resolutions to make the necessary leap forward to address the current and future evolution of ice shelves in Antarctica (Kingslake et al., 2019). Differently from Greenland, surface melting in Antarctica does not exhibit a dependency from elevation, with most of it occurring over ice shelves, at the sea level and where little elevation gradients exist. Therefore, statistical downscaling techniques using digital elevation models - as in the case of Greenland or other mountain regions - cannot be used. Machine learning (ML) tools can help in this regard. In this project, we address this issue and propose a novel method to map the spatio-temporal evolution of surface meltwater in Antarctica on a daily basis at high spatial (30 - 100 m) resolution using a combination of remote sensing, numerical modeling and machine learning. The final product of this project will consist of daily maps of surface meltwater at resolutions of the order of 100 m for the period 2000 - 2021 that will satisfy the following constraints: a) to be physically consistent with the model prediction and with the underlying governing dynamics for the melt processes; b) to capture the temporal dynamics of the model predictions, which include the temporal sequence of a set of past time steps which lead to the target prediction time, but could also include model predictions valid for a set of future time steps; c) to reconcile the higher spatial resolution of the input satellite measurements with the lower spatial resolution of the numerical model; d) to be consistent with previously generated surface melt products, so that temporal time series can be analyzed; e) to provide a measure of uncertainty to help with testing and validation.
Person Role
Tedesco, Marco Investigator and contact
Polar Cyberinfrastructure Award # 2136940
Polar Cyberinfrastructure Award # 2136939
Polar Cyberinfrastructure Award # 2136938
AMD - DIF Record(s)
Data Management Plan
None in the Database
Product Level:
0 (raw data)
Platforms and Instruments

This project has been viewed 16 times since May 2019 (based on unique date-IP combinations)