Combining deep generative models with extreme value theory for synthetic hazard simulation: a multivariate and spatially coherent approach

TL;DR

This paper introduces a novel approach combining generative adversarial networks with extreme value theory to efficiently simulate realistic, spatially coherent multivariate climate hazard events for risk assessment.

Contribution

It presents a new method that integrates GANs with extreme value theory to model dependence structures in high-dimensional climate data, enabling realistic hazard simulation.

Findings

Successfully models dependence in multivariate climate data

Generates thousands of realistic hazard events efficiently

Flexible approach transferable to other climate datasets

Abstract

Climate hazards can cause major disasters when they occur simultaneously as compound hazards. To understand the distribution of climate risk and inform adaptation policies, scientists need to simulate a large number of physically realistic and spatially coherent events. Current methods are limited by computational constraints and the probabilistic spatial distribution of compound events is not given sufficient attention. The bottleneck in current approaches lies in modelling the dependence structure between variables, as inference on parametric models suffers from the curse of dimensionality. Generative adversarial networks (GANs) are well-suited to such a problem due to their ability to implicitly learn the distribution of data in high-dimensional settings. We employ a GAN to model the dependence structure for daily maximum wind speed, significant wave height, and total precipitation over the Bay of Bengal, combining this with traditional extreme value theory for controlled extrapolation of the tails. Once trained, the model can be used to efficiently generate thousands of realistic compound hazard events, which can inform climate risk assessments for climate adaptation and disaster preparedness. The method developed is flexible and transferable to other multivariate and spatial climate datasets.