MambaRain: Multi-Scale Mamba-Attention Framework for 0-3 Hour Precipitation Nowcasting

2026-05-14

AI Summary

This study addresses the challenge that existing deterministic nowcasting methods struggle to effectively model long-range spatiotemporal dependencies in radar data, leading to a significant performance drop in 2–3 hour precipitation forecasts. To overcome this limitation, the authors propose a multiscale encoder-decoder architecture that innovatively integrates Mamba state space models with self-attention mechanisms: the former captures long-range temporal dynamics efficiently at linear computational complexity, while the latter explicitly models spatial correlations. Additionally, a spectral loss function is introduced to suppress blurring artifacts inherent in chaotic precipitation systems. Experimental results demonstrate that the proposed method substantially outperforms current deterministic approaches across 0–3 hour precipitation nowcasting, achieving particularly notable improvements in the critical 2–3 hour forecast window.

Abstract

Accurate precipitation nowcasting over extended horizons (0-3 hours) is essential for disaster mitigation and operational decision-making, yet remains a critical challenge in the field. Existing deterministic approaches are predominantly constrained to shorter prediction windows (0-2 hours), exhibiting severe performance degradation beyond 90 minutes owing to their inherent difficulty in capturing long-range spatiotemporal dependencies from radar-derived observations. To address these fundamental limitations, we propose MambaRain, a novel multi-scale encoder-decoder architecture that synergistically integrates Mamba's linear-complexity long-range temporal modeling with self-attention mechanisms for explicit spatial correlation capture. The core innovation lies in a hybrid design paradigm wherein Mamba blocks leverage selective state space mechanisms to model global temporal dynamics across extended sequences with computational efficiency, while self-attention modules explicitly characterize spatial correlations within precipitation fields - a capability inherently absent in Mamba's sequential processing paradigm. This complementary synergy enables comprehensive spatiotemporal representation learning, effectively extending the viable forecasting horizon to 2-3 hours with substantial accuracy improvements. Furthermore, we introduce a spectral loss formulation to mitigate blurring artifacts characteristic of chaotic precipitation systems, thereby preserving fine-scale motion details critical for nowcasting accuracy. Experimental validation demonstrates that MambaRain substantially outperforms existing deterministic methodologies in 0-3 hour nowcasting tasks, with particularly pronounced performance gains in the challenging 2-3 hour prediction range.

Problem

• precipitation nowcasting
• long-range spatiotemporal dependencies
• radar-derived observations
• forecasting horizon
• deterministic approaches

Innovation

• Mamba
• self-attention
• spatiotemporal modeling
• precipitation nowcasting
• spectral loss