Extreme events such as environmental hazards, network failures, financial shocks or outbreaks of health crises have a profound impact on societies and ecosystems. Mitigation and adaptation strategies need quantitative information about the likely nature (frequency and severity) of such events; their statistical modelling is considered in extreme value analysis. This session in particular puts an emphasis on looking at high-dimensional situations that are naturally encountered, when we consider spatial stochastic processes, space-time dynamics of stochastic processes, network structures at different scales, or even a combination of those. Depending on the type of application in mind, there are different recent developments on the statistical side. 

Each speaker (2 female, 2 male) represents a recent novel direction.

1. Phyllis Wan (Erasmus School of Economics, Netherlands), is an early career researcher in the area. Her latest research considers graphical modelling and sparse structures of extremes, a topic that is currently attracting a lot of research attention with considerable potential in large scale network modelling.
(With the WSC on her doorstep, her participation will also support the inclusion of a young aspiring researcher after returning from maternity leave.)

2. Thomas Opitz (Directeur de Recherche, MathNum division of INRAE, Avignon) is one of the leading scientists when it comes to the extreme value modelling of spatial or spatio-temporal environmental stochastic processes. Such processes are inherently high-dimensional. His latest work includes applications to understand multiple risks under climate change and modelling of wildfire danger, and linking extreme events to deep generative stochastic models.

3. Claudia Neves (King's College London) has a considerable track record in non-parametric statistics of extremes. Most recently, she has been advancing knowledge exchange with risk modelling related to the energy sector. She has also advanced statistical testing for space-time trends in extreme events, a topic she will present here.

4. Marco Oesting (University of Stuttgart) is Tenure-Track Professor for Computational Statistics with more than 10 years of experience in statistical modelling of spatial extremes. His interests include mathematically founded model construction, development of statistical inference for these, and applications to environmental sciences. He will bring a Bayesian perspective on multi-scale modelling to this session.