Choukri et al. (2026) Comparative analysis of CHIRPS and ERA5-Land for precipitation and drought assessment in Morocco (1987–2016)

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2026
• Date: 2026-03-21
• Authors: Majda Choukri, Mustapha Naimi, Mohamed Chikhaoui
• DOI: 10.1007/s00704-026-06180-4

Research Groups

• Department of Natural Resources and Environment, Hassan II Institute of Agronomy and Veterinary Medicine, Rabat, Morocco

Short Summary

This study comprehensively compared CHIRPS and ERA5-Land precipitation products against 114 ground stations in Morocco (1987–2016) to evaluate their performance in precipitation estimation and drought detection across various temporal scales and altitudes. It found ERA5-Land generally more accurate for precipitation variability and long-term drought, while CHIRPS showed limitations, especially in mountainous regions and for drought severity.

Objective

• To validate the CHIRPS and ERA5-Land precipitation datasets against conventional station observations.
• To assess the effect of altitude on the accuracy and performance of ERA5-Land and CHIRPS precipitation estimates.
• To examine the reliability of ERA5-Land and CHIRPS data for estimating the Standardized Precipitation Index (SPI) at different time scales (1, 3, 6, and 12 months) for drought detection.

Study Configuration

• Spatial Scale: Entire territory of Morocco (21° to 36° N latitude and 1° to 17° W longitude). CHIRPS at 0.05° (~5 km) spatial resolution; ERA5-Land initially at 9 km grid spacing, resampled to 0.05° × 0.05° for comparison.
• Temporal Scale: 30-year period from 1987 to 2016. Data aggregated into monthly, seasonal (spring, summer, autumn, winter, extended summer, extended winter), and annual periods. Drought assessment using SPI at 1, 3, 6, and 12-month time scales.

Methodology and Data

• Models used:
  • CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data)
  • ERA5-Land (ECMWF reanalysis land component)
  • Standardized Precipitation Index (SPI) calculated using a two-parameter gamma distribution.
  • Bicubic interpolation for resampling ERA5-Land to 0.05° spatial resolution.
  • Statistical metrics: Pearson's correlation coefficient (r), Mean Bias (MB), Standard Deviation (SD) ratio, Root-Mean-Square Error (RMSE).
  • Categorical drought detection indices: Probability of Detection (POD), False Alarm Ratio (FAR), Critical Success Score (CSI), Heidke Skill Score (HSS).
  • Run Theory for drought event identification (SPI < 0 for at least two consecutive months, minimum SPI ≤ -1).
  • Statistical tests: Wilcoxon signed-rank test (for product differences), Spearman’s rho coefficient (for altitudinal effects).
• Data sources:
  • Ground observations: Monthly precipitation data from 114 quality-controlled meteorological stations across Morocco (1987–2016), provided by the Directorate of Water Research and Planning.
  • Satellite data: CHIRPS v.2.0 Final (0.05° spatial resolution, 1987–2016).
  • Reanalysis data: ERA5-Land (initially 9 km, resampled to 0.05° spatial resolution, 1987–2016) from the Copernicus Climate Change Service (C3S) at ECMWF.

Main Results

• Precipitation Estimation:
  • ERA5-Land showed higher correlations (Pearson’s r: 0.61–0.86) and SD ratios near 1, indicating better representation of precipitation variability, especially during cold and transitional seasons. However, it tended to overestimate precipitation, particularly in mountainous areas (positive bias, annual RMSE of 150 mm).
  • CHIRPS displayed lower correlations (Pearson’s r: 0.47–0.77), systematically underestimated precipitation variability (SD ratio < 1), and underestimated precipitation during dry seasons. It had limitations at higher elevations (annual RMSE of 134 mm).
  • Seasonally, ERA5-Land performed significantly better in winter and extended winter, while CHIRPS outperformed ERA5-Land in summer, spring, and extended summer.
• Altitudinal Effects:
  • Both products showed performance degradation with increasing altitude.
  • ERA5-Land was more robust in mountainous regions, maintaining generally high correlations despite a gradual decline (Spearman's rho between -0.23 and -0.59). Its positive bias increased significantly with altitude.
  • CHIRPS was more sensitive to altitude, with a sharp drop in correlations beyond 1000 m (Spearman's rho between -0.35 and -0.68), and increasingly underestimated precipitation and variability at higher elevations.
• Drought Detection (SPI):
  • ERA5-Land reproduced long-term drought conditions (SPI-6, SPI-12) more accurately, with a higher probability of detection (POD > 60% for SPI-12) and lower false alarm ratio (FAR < 40% for SPI-12). It was closer to observed drought duration and severity across most time scales.
  • CHIRPS tended to exaggerate drought severity and duration, particularly at SPI-1 and SPI-12, and was accompanied by a high false alarm ratio (up to 92.5% for SPI-1).
  • Both datasets showed limited performance for short-term droughts (SPI-1 and SPI-3).
  • ERA5-Land systematically outperformed CHIRPS at all SPI scales, with stronger correlations and lower RMSE values.

Contributions

• Provided the first thorough national-level evaluation of CHIRPS and ERA5-Land precipitation accuracy and SPI-based drought detection across the entire Moroccan territory.
• Utilized a dense and long-term (30-year) observational dataset from 114 stations, addressing limitations of previous localized studies.
• Explicitly assessed the impact of altitude on the performance of both products for precipitation estimation and drought characterization.
• Evaluated the applicability of both products for drought characterization at multiple temporal scales (1, 3, 6, and 12 months).
• Offered critical guidance for natural resource managers in Morocco for selecting the most appropriate gridded precipitation products for robust drought assessments and mitigation strategies.

Funding

• Hassan II Academy of Science and Technology
• GISEC project

Citation

@article{Choukri2026Comparative,
  author = {Choukri, Majda and Naimi, Mustapha and Chikhaoui, Mohamed},
  title = {Comparative analysis of CHIRPS and ERA5-Land for precipitation and drought assessment in Morocco (1987–2016)},
  journal = {Theoretical and Applied Climatology},
  year = {2026},
  doi = {10.1007/s00704-026-06180-4},
  url = {https://doi.org/10.1007/s00704-026-06180-4}
}