Pang et al. (2025) Comprehensive evaluation of the spatiotemporal distribution characteristics of multi-source precipitation products: a case study of an extreme climate event in Henan, Central China

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2025
• Date: 2025-10-14
• Authors: Zihao Pang, Junxia Gu, Yu Zhang, Pan Yang, Zheng Wang, Shuai Han, Zhiwei Zhu
• DOI: 10.1007/s00704-025-05747-x

Research Groups

• National Meteorological Information Center, Beijing, China
• The Key Laboratory of Cloud-Precipitation Physics and Weather Modification, China Meteorological Administration, Beijing, China
• Henan Meteorological Observation Data Center, Zhengzhou, China

Short Summary

This study comprehensively evaluates the spatiotemporal characteristics of multi-source precipitation products (radar, satellite, model, and merged) during an extreme rainstorm in Henan, Central China, revealing that the merged product (CMPAS) significantly outperforms single-source products in accurately capturing fine precipitation features.

Objective

• To evaluate the accuracy and spatiotemporal distribution characteristics of multi-source precipitation products (RADAR, IMERG, GSMAP, ERA5, CMPAS) during the extreme "7.20" rainstorm in Henan, Central China, focusing on precipitation frequency-intensity distribution structure, diurnal variation, and spatiotemporal evolution.

Study Configuration

• Spatial Scale: Henan, Central China (overall extent: 31°N–40°N, 109°E–112°E; precipitation core area: 34°N–36.7°N, 112.5°E–114.8°E).
• Temporal Scale: 00:00 on July 17, 2021 to 23:00 on July 22, 2021 (UTC) (6 days).

Methodology and Data

• Models used:
  • RADAR (Radar quantitative precipitation estimation product from China Meteorological Administration)
  • IMERG (Integrated Merged Multi-satellite Retrievals V06 Final Run from NASA GPM Science Team)
  • GSMAP (Global Satellite Mapping of Precipitation-Gauge-NRT V8 from JAXA)
  • ERA5 (Fifth generation of Global Atmospheric Data Reanalysis from ECMWF)
  • CMPAS (CMA Multi-source Merged Precipitation Analysis Product, a ground-radar-satellite fusion product)
• Data sources:
  • Ground rain gauge observations (Gauge): Approximately 11,000 stations in Henan and surrounding areas, used as reference.
  • Radar, Satellite (IMERG, GSMAP), Reanalysis (ERA5), and Merged (CMPAS) precipitation products.
  • Methodology:
    • Data pre-processing: Temporal matching (IMERG converted to hourly), spatial matching (grid-to-point analysis for comparison with gauge).
    • Precipitation characteristics analysis: Hourly-scale changes in precipitation amount, frequency, intensity, peak time, start time, and end time.
    • Quantitative evaluation of precipitation amount-intensity distribution (DAI) using a two-parameter exponential fitting method.
    • Quantitative evaluation of precipitation frequency-intensity distribution (DFI) using a two-parameter double exponential fitting method.
    • Composite analysis of precipitation evolution around peak time to assess asymmetry.

Main Results

• Overall Performance: All products capture the precipitation area. CMPAS shows the highest agreement with gauge observations in location and extreme value accuracy, exhibiting better spatial continuity due to its 0.01° resolution.
• RADAR: Significantly overestimates precipitation amount and intensity, particularly in the core area and mountain fronts, with heavy precipitation contributing disproportionately to total precipitation and frequency. It accurately reflects peak time but shows earlier precipitation start times.
• Satellite Products (IMERG, GSMAP): Generally underestimate precipitation, especially heavy precipitation. They exhibit a significant peak time phase lag and reduce the observed asymmetry of the precipitation process. IMERG performs well for weak precipitation but underestimates heavy precipitation. Both IMERG and GSMAP indicate a more easterly precipitation center.
• Model Product (ERA5): Performs the worst, significantly overestimating precipitation frequency (e.g., 90% in core area vs. 60% observed) and underestimating precipitation intensity. It shows a pronounced peak time phase lag, with peaks occurring in the early morning, and reflects a more symmetrical precipitation evolution.
• Diurnal Variation: CMPAS and RADAR accurately reproduce the observed asymmetry (rapid increase to peak, then slow decline). Satellite and model products show peak time phase lags and reduce this asymmetry.
• Frequency-Intensity Distribution: CMPAS's frequency-intensity and amount-intensity distributions are most similar to gauge observations. RADAR shows an excessive contribution from strong precipitation frequency and amount. IMERG, GSMAP, and ERA5 show an overly high contribution from weak precipitation frequency and amount.
• Extreme Values: RADAR significantly overestimates extreme precipitation intensity (e.g., 99th percentile at 52 mm/h vs. 32 mm/h observed). Satellite and reanalysis products significantly underestimate extreme intensities (e.g., 99th percentile less than 20 mm/h and 10 mm/h respectively).

Contributions

• Provides a comprehensive, fine-scale evaluation of multi-source precipitation products (radar, satellite, model, merged) during an extreme rainstorm, extending beyond traditional cumulative metrics.
• Introduces quantitative methods (exponential and double exponential fitting) to assess structural biases in precipitation frequency-intensity and amount-intensity distributions.
• Highlights the superior stability and accuracy of multi-source merged products (CMPAS) for monitoring extreme precipitation events, recommending their use.
• Offers insights into the limitations of single-source satellite and reanalysis products in capturing high spatiotemporal variability and intense local heavy precipitation.
• Provides a scientific basis for improving model algorithms and monitoring accuracy for extreme precipitation, while acknowledging the value of satellite and reanalysis data for long-term climate trend analysis.

Funding

• Special fund of China Meteorological Administration Rainstorm Fine Analysis and Forecast Youth Innovation Team (Grant CMA2023QN05)
• National Science Foundation of China (Grant 42205049)
• Innovation Foundation of Key Laboratory of Cloud-Precipitation Physics and Weather Modification of China Meteorological Administration (Grant 2023CPML-B08)

Citation

@article{Pang2025Comprehensive,
  author = {Pang, Zihao and Gu, Junxia and Zhang, Yu and Yang, Pan and Wang, Zheng and Han, Shuai and Zhu, Zhiwei},
  title = {Comprehensive evaluation of the spatiotemporal distribution characteristics of multi-source precipitation products: a case study of an extreme climate event in Henan, Central China},
  journal = {Theoretical and Applied Climatology},
  year = {2025},
  doi = {10.1007/s00704-025-05747-x},
  url = {https://doi.org/10.1007/s00704-025-05747-x}
}