WeatherDiagFlow: Evidence-Grounded Radar Nowcasting with Diagnostic Flow Refinement

Chunlei Shi; Yufeng Zhu; Yixiao Liang; Dan Niu; Yongchao Feng; Qiliang Wu; Jiong Wang

cs.LGSep 24, 2026

WeatherDiagFlow: Evidence-Grounded Radar Nowcasting with Diagnostic Flow Refinement

Authors: Chunlei Shi, Yufeng Zhu, Yixiao Liang, Dan Niu, Yongchao Feng, Qiliang Wu, Jiong Wang

Organizations: Department of Automation, Southeast University, Nanjing, China · Beijing Fengyun Meteorological Science and Technology Development Co., Ltd., Beijing, China · State Key Laboratory of Virtual Reality Technology and Systems, Beihang University, Beijing, China · Department of Information Science and Technology, Fudan University, Shanghai, China

Abstract

Radar nowcasting is essential for short-term warning and emergency response, yet conventional systems mainly return future radar fields and provide limited support for operational communication and post-event verification. We formulate radar nowcasting as an evidence-grounded forecast--bulletin--audit task, in which a numerical forecaster produces both future radar fields and structured diagnostic evidence. Forecast-time bulletins use only model-available evidence, whereas post-event audits incorporate future radar truth only after the forecast horizon is observed. Based on this task formulation, WeatherDiagFlow predicts motion, growth and decay, heavy-echo risk, and uncertainty to condition rolling flow refinement, while frozen-scaffold residual calibration improves long-lead strong-echo preservation. A multi-agent layer converts the structured evidence into operational bulletins and independently generates verification audits without feeding textual outputs back into the forecaster. Experiments on FJRADAR demonstrate competitive overall performance and improved strong-echo event skill. WeatherDiagFlow therefore connects numerical prediction, evidence-grounded reporting, and auditable verification under a leakage-controlled protocol.

Figures & tables

Figure 1: WeatherDiagFlow separates forecast-time bulletin generation from post-event audit. Radar history and a coarse forecast produce predicted diagnostic fields that condition rolling flow refinement and a three-stage 0–3 hour weather bulletin. After future radar truth becomes available, verification metrics and error areas are added only to the audit product and reliability label. Textual agents package and check evidence but do not feed back into, or alter, the numerical forecast.

Method	SSIM $\uparrow$	LPIPS $\downarrow$	CSI10 $\uparrow$	CSI20 $\uparrow$	CSI30 $\uparrow$	POD20 $\uparrow$	POD30 $\uparrow$	HSS20 $\uparrow$	HSS30 $\uparrow$
SmaAt-UNet [ 18 ]	0.3283	0.4040	0.2901	0.2004	0.0608	0.2374	0.0666	0.3154	0.1128
Rolling pMF + coarse [ 9 ]	0.2108	0.4587	0.1210	0.0638	0.0169	0.0679	0.0181	0.1104	0.0320
DiffCast [ 19 ]	0.3538	0.4401	0.2566	0.1565	0.0612	0.2123	0.0756	0.2458	0.1120
CasCast [ 7 ]	0.3318	0.4050	0.3110	0.2322	0.0894	0.2929	0.1060	0.3560	0.1611
exPreCast [ 16 ]	0.3018	0.4118	0.2773	0.1783	0.0485	0.2068	0.0517	0.2852	0.0910
WDF-RC-AD (Ours)	0.3730	0.2146	0.2883	0.2448	0.1351	0.3811	0.3384	0.3757	0.2273

Table 1: Event-oriented and perceptual nowcasting results on the common 2,125-case FJRADAR test set. WDF-RC-AD (Ours) denotes the validation-selected preference-calibrated anti-decay variant.

Motion	Growth	Risk	Risk	Risk	U–err.
resid.	MAE	CSI	POD	FAR	corr.
0.0151	2.2935	0.2078	0.2603	0.3624	0.3706

Table 2: Diagnostic-field checks for the WeatherDiagFlow diagnostic branch.

Figure 2: Lead-time CSI30 on the common FJRADAR test set, pooled over the valid pixels and all forecast cases at each lead time. The 30-dBZ threshold emphasizes strong-echo event detection, while the 0–180 min curves show how skill changes as temporal uncertainty accumulates. The comparison includes the diagnostic-flow baselines and the retained WDF-RC and WDF-RC-AD variants.

Figure 3: Representative large-area case with forecast-time evidence and post-event audit separated.

Variant	D	RC	AD	CSI10 $\uparrow$	CSI20 $\uparrow$	CSI30 $\uparrow$	POD30 $\uparrow$	FAR30 $\downarrow$	HSS30 $\uparrow$
WDF	$\checkmark$	$-$	$-$	0.0645	0.0471	0.0274	0.0298	0.7401	0.0519
WDF-RC	$\checkmark$	$\checkmark$	$-$	0.2599	0.2266	0.1284	0.3005	0.8169	0.2200
WDF-RC-AD	$\checkmark$	$\checkmark$	$\checkmark$	0.2883	0.2448	0.1351	0.3384	0.8164	0.2273

Table 3: Component ablation on the common FJRADAR test manifest. D, RC, and AD denote diagnostic rolling flow, residual calibration, and anti-decay preference calibration, respectively. Best completed results are in bold ; FAR is lower-is-better.

Explore similar work

FreCast: Refining Radar Echo Intensity via Phase-Preserving Amplitude Residual Diffusion for Precipitation Nowcasting
Aug 9, 2026Heping Fang, Zihuai Yin, Kaicheng Mao +2Precipitation Nowcasting Radar
Department of Statistics and Data Science, Southern University of Science and Technology, Shenzhen 518055, China · Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China · Guangdong Provincial Key Laboratory of Brain-Inspired Intelligent Computation, Department of Computer Science and Engineering and the Department of Statistics and Data Science, Southern University of Science and Technology, Shenzhen 518055, China
Physics-Based Deep Spatiotemporal Hyperlocal Radar Nowcasting with a Multi-Variable U-Net for High-Resolution Precipitation Forecasting
Jul 17, 2026Akshay Sunil, Muhammed Rashid, Raja Sekhar Sivaraju +2Precipitation Nowcasting Radar
Centre for Climate Studies, Indian Institute of Technology Bombay, Mumbai, India · Regional Meteorological Centre (RMC) in Colaba, Mumbai, India
Pointwise is Pointless? A Multimodal Ablation Study for Precipitation Nowcasting with Graph Neural Networks
Jun 16, 2026Ophélia Miralles, Máté Mile, Christoffer Artturi +2Precipitation Nowcasting Radar
Norwegian Meteorological Institute, Oslo, Norway

Aug 9, 2026, cs.CVJ/K move · Enter open · S save

FreCast: Refining Radar Echo Intensity via Phase-Preserving Amplitude Residual Diffusion for Precipitation Nowcasting

Heping Fang, Zihuai Yin, Kaicheng Mao, Peiguang Zhang+1

Department of Statistics and Data Science, Southern University of Science and Technology, Shenzhen 518055, China · Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China · Guangdong Provincial Key Laboratory of Brain-Inspired Intelligent Computation, Department of Computer Science and Engineering and the Department of Statistics and Data Science, Southern University of Science and Technology, Shenzhen 518055, China

Precipitation Nowcasting Radar Residual Conditional Diffusion Model Frequency Spatiotemporal Deep Learning

Open paper AskPDF arXiv

Precipitation nowcasting predicts the spatiotemporal evolution of future radar echoes from historical radar echo sequences, thereby estimating the occurrence, development, and movement of precipitation over the near term. In recent years, deep learning has become an important approach to precipitation nowcasting. Although state-of-the-art models can generally capture the overall spatial distribution of future precipitation, their predictions still exhibit substantial biases in radar echo intensity at individual locations. This observation motivates a more targeted strategy for reducing forecast errors. Instead of regenerating an entire radar echo sequence without spatial constraints, the predicted precipitation structure can be used to guide the refinement of echo intensities at individual locations. This structure-guided refinement directly targets echo intensity biases. Accordingly, we propose FreCast, a two-stage framework for radar echo prediction. The first stage generates an initial forecast of future radar echoes. The second stage uses the spatial structure of the initial forecast as a constraint to further correct intensity biases at individual locations in the first-stage prediction. Experiments on three datasets demonstrate that FreCast achieves consistent improvements across forecast skill metrics. Qualitative results further show that FreCast better preserves rainband continuity and intense precipitation structures at longer lead times.

Abstract

Explore similar work

FreCast: Refining Radar Echo Intensity via Phase-Preserving Amplitude Residual Diffusion for Precipitation Nowcasting

Physics-Based Deep Spatiotemporal Hyperlocal Radar Nowcasting with a Multi-Variable U-Net for High-Resolution Precipitation Forecasting

Pointwise is Pointless? A Multimodal Ablation Study for Precipitation Nowcasting with Graph Neural Networks