WaveOp-LiteFM: Lightweight Neural-Operator Flow Matching for Satellite-to-Radar Precipitation Retrieval

WaveOp-LiteFM is a lightweight neural-operator flow matching method designed for satellite-to-radar precipitation retrieval, enabling precipitation monitoring in regions with limited radar coverage. The approach addresses the critical trade-off in existing generative flow matching models between pixel-space formulations with prohibitive computational costs and latent-space modeling that sacrifices fine precipitation detail. By leveraging neural operator architectures, the method achieves computational efficiency while maintaining retrieval quality for ground-based radar precipitation estimation from geostationary satellite observations.