My research

My research centers on complex dynamical systems and their uncertainties. I am particularly interested in stochastic modeling and data-based methods (including machine learning) for modeling and simulation of multiscale systems. Another major research topic for me is the quantification of uncertainties and rare/extreme events that occur in these systems. Two important application domains for my research are atmosphere-ocean-climate science as well as power grids and renewable energy generation.

IAS fellowship (Jan 2020 - Jun 2020)

How can we track the progress towards extreme events and regime switches in complex dynamical systems?

Extreme events (e.g. heat waves, power blackouts) can have major impact, however it is challenging to study them with models and simulation. Regime switches are similarly difficult to simulate when, like extreme events, they have only a small chance of occurring in a given time period. Models that can capture these events are often of highly complex, involving nonlinear interactions, multiple scales and many degrees of freedom. This makes model simulations slow and computationally expensive.

Being able to monitor and quantify the progress towards an extreme event or transition can be tremendously helpful for speeding up the sampling of extreme events (so-called rare event simulation). For some systems, insight from e.g. physical intuition or prior knowledge of model properties can be used to find suitable quantities to track. However, a generic approach that can be used for a wide class of problems isn’t available yet. My aim is to work towards such an approach.