Statistics and Machine Learning for Assessing Climate Risk
Contact person: Thordis L. Thorarinsdottir
Keywords: Climate risk, extremes, machine learning, statistical modelling
Research groups: Statistics and Data Science, Hydrology and Water Resources
Department of Mathematics, Department of Geosciences
Climate risks are increasingly felt in all regions of the world, and they are especially visible in the growing number of disasters that are driven by climatic events. Physical climate risks result from (increasing frequency/occurrence of) acute weather events, such as flooding, wildfires, extreme heat, storm surges and droughts and chronic climatic events/changes like increasing aridity and sea level rise. These extreme events are, by definition, rare, meaning that typically there are only a few examples of past events at any given location. Understanding and modelling these events and the associated risks thus calls for advanced statistical and machine learning models that combine statistical theory and environmental process understanding to overcome data deficiencies. We are interested in methodological research in statistics and machine learning towards these aims. Research proposals may cover several methodological approaches within this scope.
Methodological research topics:
- Non-stationary modelling of extremes in space and/or time
- Multivariate and compound extremes
- Extended range climate risk prediction
- Decision-making under climate risk
- Decision-theoretic model evaluation
External partners:
- Norwegian Computing Center (NR)
- Norwegian Meteorological Institute (met.no)