Evidential Deep Learning for Probabilistic Modelling of Extreme Storm Events
Ayush Khot, Xihaier Luo, Ai Kagawa, Shinjae Yoo
TL;DR
This work tackles uncertainty quantification in extreme storm nowcasting by proposing Evidential Deep Learning (EDL) integrated with a spatiotemporal backbone. EDL outputs parameters of a Normal-Inverse-Gamma distribution from a single forward pass, enabling calibrated aleatoric and epistemic uncertainties without multiple model runs. The approach shows favorable computational efficiency relative to ensembles and MC dropout, with competitive predictive performance and improved uncertainty calibration on the SEVIR dataset. The findings suggest practical benefits for real-time forecasting and climate risk assessment, while highlighting areas for further improvement and distributional flexibility in future work.
Abstract
Uncertainty quantification (UQ) methods play an important role in reducing errors in weather forecasting. Conventional approaches in UQ for weather forecasting rely on generating an ensemble of forecasts from physics-based simulations to estimate the uncertainty. However, it is computationally expensive to generate many forecasts to predict real-time extreme weather events. Evidential Deep Learning (EDL) is an uncertainty-aware deep learning approach designed to provide confidence about its predictions using only one forecast. It treats learning as an evidence acquisition process where more evidence is interpreted as increased predictive confidence. We apply EDL to storm forecasting using real-world weather datasets and compare its performance with traditional methods. Our findings indicate that EDL not only reduces computational overhead but also enhances predictive uncertainty. This method opens up novel opportunities in research areas such as climate risk assessment, where quantifying the uncertainty about future climate is crucial.