StormMind: Disentangled Layerwise Modeling for Convective Weather Systems

Jun CHEN, Minghui Qiu, Lin CHEN, Yan Fang, SHUXIN ZHONG, Binghong CHEN, Kaishun Wu

Abstract:

Timely nowcasting is critical for public safety during fast-evolving storms, where even short delays can trigger cascading failures—as in the October 2024 Spain flash flood that claimed over 90 lives within minutes.
While radar offers reliable real-time sensing of atmospheric structure, models that collapse 3D volumes into 2D slices inevitably discard vertical information essential for capturing storm growth, phase transitions, and collapse.
We introduce \N, a physically grounded framework that forecasts convective evolution by modeling causal interactions across stratified atmospheric layers.
\N\ addresses two fundamental challenges:
(1) the nonlinear, asynchronous coupling between low-, mid-, and high-level processes;
and (2) reflectivity uncertainty, where storms with distinct vertical structures may appear deceptively similar on radar, masking their true phase and intensity.
To tackle these issues, \N\ designs:
i) a \textit{\ComponentB} that models storm evolution from two complementary perspectives—horizontal morphology, capturing the spatial organization of physical processes within individual atmospheric layers, and vertical coupling, modeling energy exchanges across layers;
and
ii) a \textit{\ComponentC} that adaptively fuses intra- and inter-layer signals, conditioned on the evolving storm state, to infer phase transitions (e.g., initiation, intensification, dissipation).
Evaluated on the large-scale 3D-NEXRAD dataset (2020–2022, U.S.),~\N\ outperforms strong baselines, achieving a 14.71\% gain in CSI$_{40}$.
In real-world deployment with the Guangzhou Meteorological Bureau (Mar–May 2025), it improves CSI$_{40}$ by 9.39\% and boosts early-warning accuracy (98.33\%).