Research

Speaker: Dr. Tasha Snow, Colorado School of Mines

The Antarctic Ice Sheet’s contribution to sea level rise is a wildcard in future climate projections. All of Antarctica’s ice loss occurs along the coastline from state-sized glaciers and half of that loss arises as a result of melting by the ocean, making a robust understanding of ocean heat circulation around the continent critical. Field observations have been historically sparse in this harsh and remote area of the world, forcing researchers to lean predominantly on remote sensing to investigate processes driving ice loss. A recent expansion of Antarctic field observations and explosion of satellite data collection may help to answer new and longstanding ice-ocean questions, but the big data volumes have rendered standard research pipelines outclassed and data storage unfeasible on local systems. Here, we show new applications of remote sensing and state-of-the-art statistical methods that harness the power of open-source cloud-computing infrastructure to make multi-decadal fine- (sub-kilometer) or Antarctic continent-scale investigations using disparate and often sparse datasets tractable. We use visible and thermal infrared imagery fused with field observations to track the seasonal and interannual variability in the Antarctic Coastal Current, and a physics-featurized deep learning method to characterize glacier meltwater plumes that can impact ice shelf stability. We also showcase the open science principles, cloud-computing infrastructure, and new collaborative research standards we have started to build for conducting research in complex, often unobservable Antarctic systems. The specialized social and technological innovations that we have developed for glacier-ocean research are portable and generalizable to other scientific communities to help democratize science and accelerate discovery for the broader research community and public.