IMPA-Net: Meteorology-Aware Multi-Scale Attention and Dynamic Loss for Extreme Convective Radar Nowcasting

2026-04-27

AI Summary

This study addresses the tendency of existing deep learning models to produce overly smoothed radar echo nowcasts due to pixel-wise loss functions, which often fail to capture critical severe convective signals. To overcome this limitation, the authors propose a deterministic 0–2 hour nowcasting framework that integrates meteorologically informed input organization, a parameter-free spatial mixer for structured reconfiguration of mesoscale channel features, and a multi-scale predictive attention module. The model is optimized end-to-end with a novel three-tier asymmetric dynamic loss function that adaptively adjusts according to training stage, storm intensity, and forecast lead time. Evaluated on an East China dataset, the method improves the Heidke Skill Score for echoes ≥45 dBZ from 0.049 to 0.143, significantly outperforming seven baseline approaches by achieving a better trade-off between severe convective event detection and false alarm suppression, while spectral analysis confirms more faithful preservation of mesoscale energy.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Learning & Optimization for NLP
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Short-range prediction of convective precipitation from weather radar observations is essential for severe weather warnings. However, deep learning models trained with pixel-wise error metrics tend to produce overly smooth forecasts that suppress intense echoes critical for hazard detection. This issue is exacerbated by insufficient multi-scale feature interaction and suboptimal fusion of heterogeneous geophysical inputs. We propose IMPA-Net (Integrated Multi-scale Predictive Attention Network), a deterministic 0-2 hour nowcasting framework that addresses these limitations through meteorologically-informed designs at the input, architecture, and loss function levels. A parameter-free Spatial Mixer reorganizes heterogeneous input channels at the mesoscale-$γ$ neighborhood (~2 km) via deterministic channel permutation, providing a structured cross-field prior. An integrated multi-scale predictive attention module serves as the spatiotemporal translator, capturing dynamics from mesoscale-$β$ to mesoscale-$γ$ scales. A Meteorologically-Aware Dynamic Loss employs three-level asymmetric weighting -- adapting across training epochs, storm intensity, and forecast lead time -- to counteract regression-to-the-mean. Evaluated against seven baselines on a multi-source radar dataset over eastern China, IMPA-Net raises the Heidke Skill Score at $\geq$45 dBZ from 0.049 (SimVP baseline) to 0.143 under matched settings. Relative to pySTEPS, it provides a better trade-off between severe-event detection and false-alarm control. Spectral analysis confirms preserved energy across mesoscale bands where competing methods show progressive smoothing. These improvements are shown within a single domain and convective regime; generalizability to other orographic and climatic regions remains to be tested.

Problem

• convective nowcasting
• extreme precipitation
• radar echo smoothing
• multi-scale feature interaction
• heterogeneous geophysical inputs

Innovation

• multi-scale attention
• meteorology-aware loss
• radar nowcasting
• spatial mixer
• convective precipitation