Event-Level Probabilistic Prediction of Extreme Rainfall over India Using Physics-Gated Latent Dynamics

TL;DR

This paper introduces a physics-gated latent dynamics model that significantly improves event-level probabilistic prediction of extreme rainfall over India by effectively leveraging large-scale atmospheric data and modeling atmospheric evolution continuously.

Contribution

The study proposes a novel Physics-Gated Latent Ordinary Differential Equation framework that outperforms traditional models in predicting extreme rainfall events at the event level.

Findings

PG-LODE detects nearly all extreme rainfall events with higher accuracy.

ConvLSTM detects only 27% of extreme events, showing limited skill.

Physics-gated continuous-time modeling enhances extreme rainfall risk assessment.

Abstract

Extreme rainfall over the Indian monsoon region poses severe societal and infrastructural risks but remains difficult to predict at daily time scales due to stochastic convective triggering and multiscale atmospheric interactions. While large-scale atmospheric fields provide important environmental context, their ability to localize extreme rainfall events is fundamentally limited. In this study, we examine how large-scale atmospheric information from ERA5 reanalysis can be leveraged for event-level probabilistic prediction of daily rainfall extremes over India. We compare an adaptive ConvLSTM baseline with a proposed Physics-Gated Latent Ordinary Differential Equation (PG-LODE) framework, which models atmospheric evolution as a continuous-time latent process whose dynamics are explicitly modulated by a physics-based gating mechanism under convectively unstable conditions. Extreme events are defined using the local 95th percentile of the India Meteorological Department gridded rainfall dataset during the June to September monsoon season. Pixel-wise evaluation shows limited skill for both models due to spatial displacement errors, whereas event-level tile-based verification reveals a clear performance contrast. The ConvLSTM remains highly conservative, detecting only 27 percent of extreme events, while PG-LODE achieves near-complete detection with a substantially higher critical success index and a moderate false alarm rate. These results demonstrate that physics-gated continuous-time latent dynamics offer a robust pathway for translating large-scale atmospheric predictability into reliable assessments of extreme rainfall risk.