Department Systems Analysis, Integrated Assessment and Modelling

Mathematical models help us increase our understanding and our predictive and control capabilities of complex environmental systems. Our team of physicists and engineers is engaged in the following two areas: 

Modeling of complex systems

We employ a diverse array of techniques from statistical physics, nonlinear sciences, and machine learning to model complex natural and engineered environmental systems. These systems, with their numerous degrees of freedom and intricate interactions, often necessitate significant simplification resulting in uncertain predictions. We employ stochastic models to address and quantify this uncertainty. 

Development of algorithms

Most of our models need to be calibrated to data. We use Bayesian statistics to infer model parameters and quantify their uncertainty. This process is computationally challenging, especially for slow or stochastic models. To address this, we develop and apply both general-purpose algorithms and tailored solutions for specific models.