Unified landslide hazard assessment using hurdle models

Tue, 25 Oct 2022 15:32:00 BST

This model represents the first attempt to simultaneously address two out of the three requirements of the landslide hazard definition, namely, the probability of observing a landslide and its size.

Landslide size matters: A new data-driven, spatial prototype

Tue, 25 Oct 2022 16:00:00 BST

This work presents the first statistically-based model in the literature that estimates how large a landslide or multiple landslides together may be within a specific areal unit. We use our model to fit the most complete global dataset of landslides induced by earthquakes, which shows that our model is able to suitably generate maps of expected landslide sizes.

Practical strategies for regression models for extremes

Tue, 25 Oct 2022 16:15:00 BST

Inspired by the usually overlooked restrictions inherited by finite-sample maxima distributions when they are approximated by the generalised extreme-value distribution, we here make three main contributions for regression models for extremes.

Space-time gap filling for extreme hot-spots

Tue, 25 Oct 2022 16:17:00 BST

We provide probabilistic predictions of summary statistics of a spatio-temporal data process, focusing on episodes with simultaneous extreme values. we propose a two-step approach, where we first model marginal distributions with a focus on accurate modelling of the right tail and then, after transforming the data to a standard Gaussian scale, we estimate a Gaussian space-time dependence model defined locally in the time domain for the space-time sub-regions where we want to predict.

Complex tail dependence structures in space

Tue, 25 Oct 2022 16:21:00 BST

We propose a divide-and-conquer strategy to model complex spatial extremes. Contrary to classical models based on generalised Pareto processes, our model allows the tail dependence structure to flexibly change across space.

Forecasting strong wind speeds

Tue, 25 Oct 2022 16:52:00 BST

In this project, we develop a flexible spliced Gamma-Generalized Pareto model to forecast extreme and non-extreme wind speeds simultaneously. Our model belongs to the class of latent Gaussian models, for which inference is conveniently performed based on the integrated nested Laplace approximation method.

Evolution of extreme losses in stock markets

Tue, 25 Oct 2022 17:00:00 BST

We introduce a regression model that allows us to assess how extremal dependence evolves over time. We apply our model to three leading European stock markets: FTSE100 (UK), CAC40 (France) and DAX30 (Germany).