Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
Antarctica has unique areas that expose blue ice, which contrast to most of the continent (~98%) that is covered by snow. Some of these blue ice areas (BIAs) contain meteorite concentrations and (very) old ice, making them very valuable for understanding our Solar System and the climate of the past, respectively. Meteorites and old ice become accessible through ablative processes that remove upper layers of ice and leave embedded material exposed on the surface. As a result, very old ice has been found in blue ice areas (>2 million years old), and >60% of meteorites retrieved on Earth come from Antarctica. However, not all BIAs act as figurative gateways to travel in time and space. Different processes need to combine favorably to find meteorites and old ice. To understand where to go in Antarctica, we accessed and combined various remote sensing data in a deep-learning framework to identify BIAs. By using a multi-sensor approach, we could also detect blue ice under temporary snow covers. Moreover, we created a map of potential meteorite collection sites using machine learning. The analyses showed that the ice flow velocity, the exposure of ice, the surface slope, and the extreme surface temperature are the most important indicators for the presence of meteorites. The machine-learning framework also allowed us to project how the presence of meteorites is to change in the future. Next, we will use data-driven techniques to direct old ice sampling efforts in BIAs.
Varun Sharma, Michael Lehning, Franziska Gerber
Reto Georg Trappitsch, Xuan Li