Modeling Spatio-temporal Extremes via Conditional Variational Autoencoders

2025-12-06

AI Summary

This study addresses the modeling of spatiotemporal co-occurrence of extreme weather events under climate change. We propose the conditional variational autoencoder (cXVAE) framework, which integrates convolutional neural networks with climate indices (e.g., ENSO) to explicitly encode spatiotemporal dependence structures of extremes in the latent space, enabling counterfactual interventions and joint tail-risk quantification under climate-variable conditioning. Compared to conventional methods, cXVAE achieves high-fidelity reconstruction of extreme spatial fields at minimal computational cost, detects and attributes climate-driven shifts in dependence structure, and significantly improves return-period estimation and joint risk assessment. Applied to fire weather index data over eastern Australia, the model successfully identifies the systematic modulation of extreme dependence patterns by ENSO, offering a novel tool for climate attribution and adaptive risk management.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Extreme weather events are widely studied in fields such as agriculture, ecology, and meteorology. The spatio-temporal co-occurrence of extreme events can strengthen or weaken under changing climate conditions. In this paper, we propose a novel approach to model spatio-temporal extremes by integrating climate indices via a conditional variational autoencoder (cXVAE). A convolutional neural network (CNN) is embedded in the decoder to convolve climatological indices with the spatial dependence within the latent space, thereby allowing the decoder to be dependent on the climate variables. There are three main contributions here. First, we demonstrate through extensive simulations that the proposed conditional XVAE accurately emulates spatial fields and recovers spatially and temporally varying extremal dependence with very low computational cost post training. Second, we provide a simple, scalable approach to detecting condition-driven shifts and whether the dependence structure is invariant to the conditioning variable. Third, when dependence is found to be condition-sensitive, the conditional XVAE supports counterfactual experiments allowing intervention on the climate covariate and propagating the associated change through the learned decoder to quantify differences in joint tail risk, co-occurrence ranges, and return metrics. To demonstrate the practical utility and performance of the model in real-world scenarios, we apply our method to analyze the monthly maximum Fire Weather Index (FWI) over eastern Australia from 2014 to 2024 conditioned on the El Niño/Southern Oscillation (ENSO) index.

Problem

• Models spatio-temporal extremes using conditional variational autoencoders
• Detects climate-driven shifts in extreme event dependence
• Quantifies joint tail risk changes via counterfactual experiments

Innovation

• Conditional VAE integrates climate indices for extremes modeling
• CNN decoder convolves indices with latent spatial dependencies
• Enables counterfactual experiments to quantify climate intervention impacts