Almeida et al. (2025) Evaluation of eight satellite and reanalysis precipitation products over Angola: The value of targeted regional assessment for water management

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-11-20
• Authors: Manuel Almeida, Senlin Zhu, Pedro Coelho
• DOI: 10.1016/j.ejrh.2025.102955

Research Groups

• MARE - Marine and Environmental Sciences Centre, ARNET - Aquatic Research Network Associate Laboratory, NOVA School of Science and Technology, NOVA University Lisbon, Caparica, Portugal
• College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou, China

Short Summary

This study systematically evaluates eight satellite and reanalysis daily rainfall datasets over Angola from 2011 to 2023. It finds that daily precipitation is generally poorly captured by all datasets, with MSWEP v2.8 and IMERG v06B outperforming others, particularly in the arid southwest, underscoring the importance of targeted regional assessments for water management.

Objective

• To comprehensively evaluate eight satellite and reanalysis daily rainfall products across Angola, focusing on their performance at full-year, wet and dry seasons, extreme quantiles (5th and 95th percentiles), and annual totals.

Study Configuration

• Spatial Scale: Angola, southwestern Africa, covering diverse climatic and topographic environments. Datasets have spatial resolutions ranging from 0.0375° to 0.25°. Specific focus on the provinces of Namibe and Huíla in the arid southwest.
• Temporal Scale: Daily precipitation data from 2011 to 2023, analyzed across full-year, wet season (October-April), dry season (May-September), and extreme quantiles (5th and 95th percentiles).

Methodology and Data

• Models used: Eight satellite and reanalysis precipitation products: African Rainfall Climatology version 2 (ARC v2.0), Multi-Source Weighted-Ensemble Precipitation (MSWEP v2.8), Rainfall Estimates from Satellites (RFE v2.0), Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS v2.0), Tropical Application of Meteorology Using Satellite Data and Ground-Based Observations (TAMSAT v3.1), Fifth generation of European ReAnalysis (ERA5L), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks - Climate Data Record (PERSIANN-CDR), and Integrated Multi-satellite Retrievals for GPM (IMERG v06B).
• Data sources:
  • Ground observations: 45 in situ rain gauge observations from the Southern African Science Centre for Climate Change and Adaptive Land Management (SASSCAL) for the period 2011–2023.
  • Ancillary data: National Water Plan (NWP) mean annual precipitation map (1944–1974), Shuttle Radar Topography Mission (DEM), K¨oppen-Geiger climate classification.
  • Evaluation metrics: Five continuous metrics (Correlation Coefficient (R), Coefficient of Variation (CV), Bias, Kling-Gupta Efficiency (KGE), Root Mean Square Error (RMSE)) and four categorical metrics (Probability of Detection (POD), False Alarm Ratio (FAR), Frequency Bias Index (FBI), Critical Success Index (CSI)) were used.

Main Results

• All evaluated satellite and reanalysis precipitation datasets generally performed poorly in characterizing daily precipitation over Angola, with no single product consistently excelling across all temporal and spatial scales.
• MSWEP v2.8 and IMERG v06B consistently outperformed the other datasets across most timescales.
• All datasets reliably detected "no precipitation" events (less than 1 mm/day) but showed decreasing performance as rainfall intensity increased, struggling with light, moderate, heavy, and very high-intensity events.
• Performance was generally worse during the wet season compared to the dry season.
• MSWEP v2.8 and IMERG v06B demonstrated reasonable skill (KGE values up to 0.4) in specific regions, particularly the arid southwest provinces of Namibe and Huíla (Stations 22–35), especially during the 2014–2019 period.
• ERA5L significantly overestimated, and ARC significantly underestimated, mean annual precipitation in a transition region between Temperate (Cwb), Arid Steppe (BSh), and Arid Desert (BWh) climate classes.
• Higher spatial resolution (e.g., TAMSAT) did not necessarily correlate with better accuracy.

Contributions

• Provides the first comprehensive, multi-scale, and region-specific evaluation of eight widely used satellite and reanalysis precipitation products for Angola, addressing a significant data gap.
• Highlights the critical importance of targeted regional assessments for precipitation product validation, demonstrating that a one-size-fits-all approach is ineffective for effective water management in climatically diverse regions like Angola.
• Identifies the most reliable datasets (MSWEP v2.8 and IMERG v06B) and specific regions within Angola where their performance is reasonably good, offering practical guidance for hydrological applications.
• Recommends prioritizing MSWEP v2.8, IMERG v06B, CHIRPS v2.0, and PERSIANN-CDR for site-specific evaluations before their use in hydrological modeling and water resource management.

Funding

• Fundaç˜ao para a Ciˆencia e a Tecnologia (FCT, Portugal)
  • UIDB/04292/2020 (MARE - Marine and Environmental Sciences Centre)
  • UIDP/04292/2020 (MARE - Marine and Environmental Sciences Centre)
  • LA/P/0069/2020 (ARNET - Aquatic Research Network Associate Laboratory)

Citation

@article{Almeida2025Evaluation,
  author = {Almeida, Manuel and Zhu, Senlin and Coelho, Pedro},
  title = {Evaluation of eight satellite and reanalysis precipitation products over Angola: The value of targeted regional assessment for water management},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.102955},
  url = {https://doi.org/10.1016/j.ejrh.2025.102955}
}