Xingtao Song | arXiv Science

arXiv2026

Composable multi-satellite precipitation estimation for evolving observing systems

Rapid and spatially continuous precipitation monitoring is critical for flood, landslide, and other hydrometeorological hazard warnings, particularly in regions where rain-gauge and weather-radar networks are sparse. The coordinated use of heterogeneous satellite observations, including geostationary infrared, passive microwave, and spaceborne radar measurements, is therefore a key pathway toward more accurate and spatially refined precipitation monitoring. Recent deep-learning methods have substantially improved multi-source satellite precipitation estimation, but most remain tied to predefined combinations of satellite inputs. As satellite observing systems evolve, incorporating new instruments often requires substantial model retraining and maintenance. We propose PRISMA, a generative framework for precipitation retrieval from multi-source observations. The framework separates the training of the precipitation prior from sensor-specific observational constraints, enabling sensor branches to be flexibly composed or extended without retraining the precipitation generative backbone. We successively integrate FY-4B/AGRI, GPM/GMI, F16-F18 SSMIS, and GPM/DPR-Ka observations within the PRISMA framework, achieving consistent improvements in precipitation-estimation accuracy. Matched-footprint experiments further confirm the effective use of complementary information from coincident sensors. Independent station validation shows that PRISMA outperforms IMERG Final in both CRPS and RMSE while providing positive fair Brier skill across all precipitation thresholds. PRISMA enables flexible composition of heterogeneous satellite observations and rapid generation of accurate ensemble precipitation estimates, strengthening satellite-based precipitation monitoring.

physics.ao-ph↗

arXiv2026

Nowcast3D: Reliable precipitation nowcasting via gray-box learning

Reliable nowcasting of extreme precipitation remains difficult because convective systems are strongly nonlinear, multiscale, and nonstationary in 3D. Radar is the backbone of nowcasting, yet existing methods struggle to predict extremes: physics-based extrapolation cannot capture growth and decay, deterministic learning tends to oversmooth and underestimate peaks, and purely generative models often lack physical consistency. Hybrid schemes help but are mostly limited to 2D composite reflectivity, collapsing the atmosphere into one layer and discarding vertical structure critical for height-dependent dynamics. We introduce Nowcast3D, a gray-box, fully 3D framework that works directly on volumetric radar reflectivity. The end-to-end model couples physically constrained neural operators (advection, local diffusion, and microphysics) with a conditional diffusion model to generate ensemble forecasts with quantified uncertainty. Trained on provincial-scale 3D volumes over a $10.24^\circ \times 10.24^\circ$ region and fine-tuned on a $2.56^\circ \times 2.56^\circ$ city region ($0.01^\circ \approx 1$ km), Nowcast3D provides near-real-time forecasts up to 3 h and outperforms competitive baselines in cross-region and temporal out-of-sample tests. It can also infer wind fields without labeled supervision, supporting physically plausible transport. In a nationwide blind evaluation by 160 meteorologists, Nowcast3D ranked first and was preferred in 57% of post-hoc assessments, surpassing the leading baseline (27%). These results highlight its reliability and operational value for extreme precipitation nowcasting.

cs.LG↗