RainSeer: Fine-Grained Rainfall Reconstruction via Physics-Guided Modeling

TL;DR

RainSeer addresses the challenge of reconstructing high-resolution rainfall fields by reinterpreting radar reflectivity as a physically grounded, evolving prior. It introduces a two-stage architecture: a Structure-to-Point Mapper that aligns mesoscale radar structures with ground-based rainfall using hierarchical encoders and a Bidirectional Physics-Aware Attention, and a Geo-Aware Rain Decoder that models descent and microphysical evolution via a Causal Spatiotemporal Attention mechanism. The approach is trained with a combination of MSE and GeoLoss to enforce urban-spatial consistency and physical plausibility. Evaluations on MeteoNet and RAIN-F show RainSeer achieves substantial improvements in MAE, RMSE, NSE, and CC over baselines, highlighting its potential for accurate, structure-preserving rainfall reconstruction and improved flood forecasting.

Abstract

Reconstructing high-resolution rainfall fields is essential for flood forecasting, hydrological modeling, and climate analysis. However, existing spatial interpolation methods-whether based on automatic weather station (AWS) measurements or enhanced with satellite/radar observations often over-smooth critical structures, failing to capture sharp transitions and localized extremes. We introduce RainSeer, a structure-aware reconstruction framework that reinterprets radar reflectivity as a physically grounded structural prior-capturing when, where, and how rain develops. This shift, however, introduces two fundamental challenges: (i) translating high-resolution volumetric radar fields into sparse point-wise rainfall observations, and (ii) bridging the physical disconnect between aloft hydro-meteors and ground-level precipitation. RainSeer addresses these through a physics-informed two-stage architecture: a Structure-to-Point Mapper performs spatial alignment by projecting mesoscale radar structures into localized ground-level rainfall, through a bidirectional mapping, and a Geo-Aware Rain Decoder captures the semantic transformation of hydro-meteors through descent, melting, and evaporation via a causal spatiotemporal attention mechanism. We evaluate RainSeer on two public datasets-RAIN-F (Korea, 2017-2019) and MeteoNet (France, 2016-2018)-and observe consistent improvements over state-of-the-art baselines, reducing MAE by over 13.31% and significantly enhancing structural fidelity in reconstructed rainfall fields.

Figures (4)

• Figure 1: Modeling Strategy of Abrupt Rainfall Field.
• Figure 2: The Framework of RainSeer.
• Figure 3: Performance(RMSE/MAE/NSE/CC) Comparison w.r.t. #Neighbors, Time Window, and Mask Ratios.
• Figure 4: Visualization of Reconstruction Rainfall.