arXiv subjects

Zhiqiu Gao

Publications and source records attributed to Zhiqiu Gao.

2 recordsLinked to original sources

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

StormDiT: A generative AI model bridges the 2-6 hour 'gray zone' in precipitation nowcasting

Accurate short-term warnings for extreme precipitation are critical for global disaster mitigation but are hindered by a persistent predictability barrier at the 2-6 hour horizon -- the "nowcasting gray zone." In this window, traditional observation-based extrapolation fails due to error accumulation, while numerical weather prediction is computationally too slow to resolve storm-scale dynamics. Recent generative AI approaches attempt to bridge this gap by decomposing precipitation into separate deterministic advection and stochastic diffusion components. However, this decomposition can sever fundamental causal links between entangled atmospheric processes, such as the dynamic initiation of convection triggered by boundary advection. Here we present StormDiT, a unified generative model that treats weather evolution as a holistic spatiotemporal problem, learning the coupled physics of the gray zone without human-imposed structural priors. Trained on a massive dataset of 7,720 precipitation events from China, our model achieves a breakthrough in long-horizon stability. On a heavy-rainfall test set, it maintains skillful prediction for strong convection ($\ge$ 35 dBZ) with a Critical Success Index (CSI) near 0.2 across the full 6-hour forecast at 6-minute resolution. Crucially, the model exhibits superior probabilistic calibration, accurately quantifying operational risks. On the public SEVIR benchmark, our unified paradigm more than doubles the state-of-the-art 1-hour performance for heavy rain and establishes the first robust baseline for 3-hour forecasting. Furthermore, interpretability analysis reveals that the model attends to non-local physical precursors, such as outflow boundaries, explicitly validating its emergent understanding of convective organization.

physics.ao-ph↗