Event-Level Probabilistic Prediction of Extreme Rainfall over India Using Physics-Gated Latent Dynamics
Arun Govind Neelan · 2026
Read Event-Level Probabilistic Prediction of Extreme Rainfall over India Using Physics-Gated Latent Dynamics by Arun Govind Neelan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-03166.txt.
Description
Arun Govind Neelan, writing alone, studies whether large-scale weather fields can predict daily rainfall extremes over India. Extreme monsoon rainfall threatens lives and infrastructure, but stochastic convective triggering and multiscale interactions make daily prediction hard. Environmental context from reanalysis data helps, yet its ability to localize individual extreme events is fundamentally limited by spatial displacement errors in gridded forecasts and observations alike.
Neelan compares an changing ConvLSTM earlier method with a physics-gated latent differential formula system. The latter models atmospheric evolution as a continuous-time latent process, modulated by physics-based gating under unstable cases. Pixel-wise scores were limited for both models. Event-level verification told a clearer story. The earlier method detected only about a quarter of extremes, while the gated model achieved near-complete finding with a better success index.
Catalog record
| CREATOR | Arun Govind Neelan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-03166