Ahmadi et al. (2026) Evaluation of the accuracy of six satellite precipitation products and their spatial-temporal variability patterns using ground station data (case study: Mazandaran province, Iran)

Identification

• Journal: Natural Hazards
• Year: 2026
• Date: 2026-01-01
• Authors: Sedigheh Bararkhanpour Ahmadi, Reza Norooz Valashedi, Mehdi Nadi, Khalil Ghorbani, Karim Solaimani
• DOI: 10.1007/s11069-025-07747-6

Research Groups

• Department of Water Engineering, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
• Department of Water Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran
• Department of Watershed Science and Engineering, Sari Agricultural Sciences and Natural Resources University, Sari, Iran

Short Summary

This study evaluated the accuracy and spatial-temporal variability of six gridded satellite precipitation products (GSPPs) against ground station data in Mazandaran Province, Iran, from 2007 to 2023. The results indicated that CPC, IMERG, and CMORPH generally exhibited the highest accuracy, with performance varying significantly across temporal scales and topographic conditions.

Objective

• To evaluate and compare the accuracy and spatial-temporal variability patterns of six gridded satellite precipitation products (CHIRPS, CMORPH, IMERG, CPC, SM2RAIN, PERSIANN-CDR) against ground station data in Mazandaran Province, Iran.

Study Configuration

• Spatial Scale: Mazandaran Province, northern Iran (approximately 23,842 km²), characterized by significant elevation differences (21 m to 5595 m above sea level). Analysis was conducted at regional and point-to-grid levels.
• Temporal Scale: January 2007 to December 2023 (17 years), with evaluations at daily, monthly, seasonal, and annual time scales.

Methodology and Data

• Models used:
  • CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data, version 2.0)
  • CPC (Climate Prediction Center Global Unified Daily Gauge-Based Precipitation Dataset)
  • CMORPH (Climate Prediction Center Morphing Technique, daily data from CDRs)
  • IMERG (Integrated Multi-satellitE Retrievals for GPM)
  • PERSIANN-CDR (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks – Climate Data Record)
  • SM2RAIN-ASCAT (Soil Moisture to Rain from ASCAT data)
• Data sources:
  • Ground-based observations: Daily precipitation time series from 15 synoptic meteorological stations in Mazandaran Province.
  • Satellite precipitation products: CHIRPS (0.05° spatial resolution), CPC (0.05° spatial resolution), CMORPH (8 km spatial resolution), IMERG (GPM mission), PERSIANN-CDR (0.25° spatial resolution), SM2RAIN-ASCAT (1 km or 0.1° spatial resolution).
  • Evaluation indices: Statistical (Root Mean Square Error (RMSE), Spearman’s Correlation Coefficient (CC), Bias (BIAS), Mean Absolute Error (MAE), Nash-Sutcliffe Efficiency (NSE), Kling-Gupta Efficiency (KGE)) and categorical (False Alarm Ratio (FAR), Probability of Detection (POD), Critical Success Index (CSI)).
  • Quantile regression was used to investigate the pattern of changes in different quantiles of GSPPs versus observed precipitation.

Main Results

• Overall Performance: CPC, IMERG, and CMORPH products generally showed the highest accuracy across most evaluation indices. PERSIANN-CDR consistently exhibited the lowest accuracy (e.g., CC = -0.3, KGE = -2, RMSE = 6 mm).
• Temporal Scale Accuracy: The accuracy of monthly precipitation data was generally higher than annual and seasonal data for most GSPPs.
• Spatial Variability: GSPPs performed better in the eastern and coastal areas of Mazandaran Province compared to mountainous regions. Products tended to underestimate precipitation in coastal and low-altitude areas but overestimate in mountainous areas.
• CPC Product Specifics: CPC demonstrated the best overall performance, with high correlation (CC = 0.89), low RMSE (2 mm), and favorable BIAS ((-0.15)–0.4), KGE (0.7), NSE (0–1), POD (0.77), FAR (0.3), and CSI (0.7). Its annual, monthly, and quantile precipitation estimates were closest to observed data.
• Quantile Analysis: CPC showed the greatest similarity to observed precipitation across various quantiles. Observed precipitation exhibited minimal change in lower quantiles but a significant positive upward slope in extreme upper quantiles (0.9–0.99).
• Topographic Influence: Elevation changes significantly impacted GSPP accuracy, with higher uncertainty and lower performance observed at high-altitude stations compared to coastal areas.

Contributions

• Provided a comprehensive, multi-product, multi-scale (spatial and temporal) evaluation of satellite precipitation products in a topographically complex region of Iran.
• Identified the best-performing GSPPs (CPC, IMERG, CMORPH) for Mazandaran Province, offering crucial guidance for hydrological modeling and water resource management in data-scarce areas.
• Quantified the spatial and temporal biases of each GSPP, highlighting their varying performance based on topography (underestimation in coastal areas, overestimation in mountainous areas).
• Utilized a wide array of statistical and categorical evaluation indices, along with quantile regression, for a robust and multi-faceted assessment of product accuracy and variability patterns.
• Emphasized the importance of local validation and the need for bias correction of GSPPs before their direct application in regional studies.

Funding

Not explicitly stated in the provided text.

Citation

@article{Ahmadi2026Evaluation,
  author = {Ahmadi, Sedigheh Bararkhanpour and Valashedi, Reza Norooz and Nadi, Mehdi and Ghorbani, Khalil and Solaimani, Karim},
  title = {Evaluation of the accuracy of six satellite precipitation products and their spatial-temporal variability patterns using ground station data (case study: Mazandaran province, Iran)},
  journal = {Natural Hazards},
  year = {2026},
  doi = {10.1007/s11069-025-07747-6},
  url = {https://doi.org/10.1007/s11069-025-07747-6}
}