PhD project: Assessing future changes in Greenland runoff using machine learning and climate models

University: Aarhus University Department: Department of Environmental Science

Section: Atmospheric Emissions & Modelling

Supervisor: Peter L. Langen

Co-supervisors: Ruth H. Mottram (DMI), Andreas Peter Ahlstrøm (GEUS)

Project term: 15.12.2023 – 14.12.2026

Master’s degree: MSc in Mathematics, University of Graz, Austria

Experience: Data Scientist with 5 years experience in Research

How much is the Greenland Ice Sheet melting in the next 100 or 150 years? The physcial models used for creating future projections of the ice sheet are computationally expensive and thus we have only a limited number of possible future outcomes. In my Phd project I am developing emulators using Deep Learning, which allow for creating future projections significantly faster. With those emulators I will build large ensembles and assess the likely range of future runoff.

My teaching includes introductory courses on Python programming that I teach at my institute. Also I created a Workshop on Machine Learning for Climate Downscaling, that specifically targets my colleagues I am working with closely. I plan to teach similar courses on Machine Learning in the future.

In my Phd project I am co-supervised by the Danish Meteorological Institute (DMI), where I collaborate closely with other Phd students and Post-docs working on Machine Learning techniques for climate modelling in the Arctic and Antarctic. In addition, I am also co-supervised by the Geological Survey of Denmark and Greenland (GEUS), which allows me to work closely with researchers who operate a multitude of measurement stations on the Greenland ice sheet.