arXiv subjects

Congyi Nai

Publications and source records attributed to Congyi Nai.

5 recordsLinked to original sources

arXiv2026

Cast3: Translating numerical weather prediction principles into data-driven forecasting

Data-driven weather models have made rapid advances in recent years, reaching and in some metrics surpassing the large-scale forecast skill of operational numerical weather prediction. This progress, however, has been built almost entirely on the reanalysis data that NWP produced, while the methodological knowledge that the NWP community distilled over decades of multi-scale atmospheric modelling remains largely unused. Here we present Cast3, a generative forecasting framework that systematically absorbs NWP meta-knowledge to close this gap. Cast3 operates on variable-resolution cubed-sphere grids for scale-aware representation and constructs structurally diverse super-ensembles that sample the complementary biases of different grid discretizations, delivering state-of-the-art ensemble prediction. It further introduces generative nudging, a posterior-sampling strategy that distils the collective information of the full ensemble into a single forecast possessing both the large-scale accuracy of the ensemble mean and the mesoscale realism of a high-resolution member. Evaluated across synoptic-scale skill, spectral fidelity, station-level surface verification, and tropical cyclone prediction, Cast3 outperforms established deterministic and generative baselines across various dimensions. More broadly, these results demonstrate that the design principles embedded in computational atmospheric science offer a rich and largely untapped foundation for the next generation of data-driven Earth system modelling.

physics.ao-ph↗

arXiv2025

Fusion of multi-source precipitation records via coordinate-based generative model

Precipitation remains one of the most challenging climate variables to observe and predict accurately. Existing datasets face intricate trade-offs: gauge observations are relatively trustworthy but sparse, satellites provide global coverage with retrieval uncertainties, and numerical models offer physical consistency but are biased and computationally intensive. Here we introduce PRIMER (Precipitation Record Infinite MERging), a deep generative framework that fuses these complementary sources to produce accurate, high-resolution, full-coverage precipitation estimates. PRIMER employs a coordinate-based diffusion model that learns from arbitrary spatial locations and associated precipitation values, enabling seamless integration of gridded data and irregular gauge observations. Through two-stage training--first learning large-scale patterns, then refining with accurate gauge measurements--PRIMER captures both large-scale climatology and local precision. Once trained, it can downscale forecasts, interpolate sparse observations, and correct systematic biases within a principled Bayesian framework. Using gauge observations as ground truth, PRIMER effectively corrects biases in existing datasets, yielding statistically significant error reductions at most stations and furthermore enhancing the spatial coherence of precipitation fields. Crucially, it generalizes without retraining, correcting biases in operational forecasts it has never seen. This demonstrates how generative AI can transform Earth system science by combining imperfect data, providing a scalable solution for global precipitation monitoring and prediction.

physics.ao-ph↗

arXiv2025

Generative assimilation and prediction for weather and climate

Machine learning models have shown great success in predicting weather up to two weeks ahead, outperforming process-based benchmarks. However, existing approaches mostly focus on the prediction task, and do not incorporate the necessary data assimilation. Moreover, these models suffer from error accumulation in long roll-outs, limiting their applicability to seasonal predictions or climate projections. Here, we introduce Generative Assimilation and Prediction (GAP), a unified deep generative framework for assimilation and prediction of both weather and climate. By learning to quantify the probabilistic distribution of atmospheric states under observational, predictive, and external forcing constraints, GAP excels in a broad range of weather-climate related tasks, including data assimilation, seamless prediction, and climate simulation. In particular, GAP is competitive with state-of-the-art ensemble assimilation, probabilistic weather forecast and seasonal prediction, yields stable millennial simulations, and reproduces climate variability from daily to decadal time scales.

cs.LG↗

arXiv2024

DRUM: Diffusion-based runoff model for probabilistic flood forecasting

Extreme floods pose escalating risks in a changing climate, yet forecasting remains challenging due to peak flow underestimation and high uncertainty. We introduce DRUM, a diffusion-based probabilistic deep learning approach that advances extreme flood forecasting across representative basins in the contiguous United States. DRUM outperforms state-of-the-art benchmarks, enhancing nowcasting skill for the top 0.1% of flows in 72.3% of studied basins. Under operational scenarios, DRUM extends reliable lead times by nearly a full day for 20- and 50-year floods. When evaluated with measured precipitation, an ideal condition, recall improves by 0.3-0.4 and the early warning window extends by 2.3 days for 50-year floods. The enhancement potential varies regionally, with precipitation-driven flood zones in the eastern and northwestern U.S. benefiting most, gaining 3-7 days in lead time. These findings highlight the transformative potential of diffusion models as a cutting-edge generative AI technique for advancing hydrology and broader Earth system sciences.

physics.geo-ph↗

arXiv2024

Boosting weather forecast via generative superensemble

Accurate weather forecasting is essential for socioeconomic activities. While data-driven forecasting demonstrates superior predictive capabilities over traditional Numerical Weather Prediction (NWP) with reduced computational demands, its deterministic nature and limited advantages over physics-based ensemble predictions restrict operational applications. We introduce the generative ensemble prediction system (GenEPS) framework to address these limitations by randomizing and mitigating both random errors and systematic biases. GenEPS provides a plug-and-play ensemble forecasting capability for deterministic models to eliminate random errors, while incorporating cross-model integration for cross-model ensembles to address systematic biases. The framework culminates in a super-ensemble approach utilizing all available data-driven models to further minimize systematic biases. GenEPS achieves an Anomaly Correlation Coefficient (ACC) of 0.679 for 500hPa geopotential (Z500), exceeding the ECMWF Ensemble Prediction System's (ENS) ACC of 0.646. Integration of the ECMWF ensemble mean further improves the ACC to 0.683. The framework also enhances extreme event representation and produces energy spectra more consistent with ERA5 reanalysis. GenEPS establishes a new paradigm in ensemble forecasting by enabling the integration of multiple data-driven models into a high-performing super-ensemble system.

physics.ao-ph↗