AI Framework Uses GANs to Predict Climate Change Impacts on Drought

A new paper introduces an artificial intelligence framework based on Conditional Generative Adversarial Networks (Conditional GANs) to generate future spatio-temporal trajectories of climatic indices, focusing on drought prediction. The research, available on arXiv, aims to aid risk management and insurance strategies in the face of climate change ¹.

The study, titled "A Wasserstein GAN-based climate scenario generator for risk management and insurance: the case of soil subsidence," highlights the increasing costs associated with natural disasters. According to the United Nations Office for Disaster Risk Reduction (2025), the average annual cost of natural catastrophes increased significantly between 1970 and 2020 ¹.

The research emphasizes the need for the insurance sector to adapt to the evolving climate context. Organizations such as the IFOA and the WWF have highlighted the necessity for medium- to long-term strategies that extend beyond the one-year horizon of current prudential regulations ¹.

The AI framework, named SwiGAN, focuses on the Soil Wetness Index (SWI), a key indicator used in France to assess drought severity. Droughts account for approximately 30% of the indemnities paid under the French natural catastrophe insurance scheme ¹.

SwiGAN simulates plausible drought propagation patterns up to 2050 for a region of France particularly exposed to this hazard. By generating realistic sequences of SWI maps, the model provides insights into drought dynamics under climate change scenarios ¹.

The generated data supports the design of adaptive risk management and insurance strategies. The methodology is also generalizable to other climate-related perils and actuarial applications, such as economic scenario generation ¹.

The authors of the paper are Antoine Heranval, Olivier Lopez, Didier Ngatcha, and Daniel Nkameni ¹.

The paper is categorized under Machine Learning (cs.LG), Risk Management (q-fin.RM), and Applications (stat.AP) ¹.