Abdulahi et al. (2025) Impact of Climate Change on Drought Dynamics in the Ganale Dawa River Basin, Ethiopia

Identification

• Journal: Climate
• Year: 2025
• Date: 2025-11-11
• Authors: Mohammed Mussa Abdulahi, Pascal Egli, Anteneh Belayneh Desta, Yazidhi Bamutaze, Sintayehu W. Dejene
• DOI: 10.3390/cli13110231

Research Groups

• Africa Center of Excellence in Climate Smart Agriculture and Biodiversity Conservation, Haramaya University, Dire Dawa, Ethiopia
• Department of Geography, Norwegian University of Science and Technology, Trondheim, Norway
• School of Biological Science and Biotechnology, College of Natural and Computational Science, Haramaya University, Dire Dawa, Ethiopia
• Department of Geography, Geo-Informatics and Climatic Sciences, Makerere University, Makerere, Kampala, Uganda
• The International Center for Tropical Agriculture, Addis Ababa, Ethiopia

Short Summary

This study assessed the impact of climate change on agricultural and hydrological drought dynamics in Ethiopia's Ganale Dawa Basin using machine learning-enhanced CMIP6 projections and satellite-based indices. Findings reveal increasing variability in agricultural drought and continued recurrence of hydrological drought, especially under high-emission scenarios.

Objective

• To investigate how climate change influences agricultural and hydrological droughts in Ethiopia’s Ganale Dawa River Basin to provide evidence-based insights for adaptation and water management.

Study Configuration

• Spatial Scale: Ganale Dawa River Basin, southeast Ethiopia, covering approximately 171,050 square kilometers.
• Temporal Scale:
  • Historical baseline: 1982–2014
  • Future periods: 2015–2040 (near future), 2041–2070 (mid-future), 2071–2100 (far future)
  • Climate Scenarios: Shared Socioeconomic Pathway 2-4.5 (SSP2-4.5, moderate-emission) and SSP5-8.5 (high-emission).

Methodology and Data

• Models used:
  • Machine Learning: Random Forest (RF) model was selected as the best-performing algorithm (outperforming Light Gradient Boosting Machine, Extreme Gradient Boosting, Categorical Boosting, Long Short-Term Memory, and Multilayer Perceptron).
  • Climate Models: Multi-model ensemble mean (MME) of 10 top-performing Global Climate Models (GCMs) from the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP-CMIP6) dataset (CNRM-CM6-1, CNRM-ESM2-1, CanESM5, EC-Earth3, FGOALS-g3, GISS-E2-1-G, HadGEM3-GC31-LL, INM-CM4-8, INM-CM5-0, KACE-1-0-G).
  • Drought Indices: Standardized Soil Moisture Index (SSMI) for agricultural drought and Standardized Runoff Index (SRI) for hydrological drought.
  • Trend Analysis: Non-parametric Mann–Kendall (MK) test and Sen’s slope estimator.
• Data sources:
  • Historical Hydro-climatic: Famine Early Warning Systems Network Land Data Assimilation System (FLDAS) and ERA5-Land datasets.
  • Precipitation and Temperature Validation: Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) and Climate Hazards Center InfraRed Temperature with Station data (CHIRTS).
  • Future Climate Projections: NEX-GDDP-CMIP6 models.
  • Spatial Resolution: All datasets harmonized to a common 0.1° (approximately 10 km x 10 km) resolution.

Main Results

• The Random Forest model achieved high predictive accuracy (R² > 0.958 for SSMI and 0.873 for SRI).
• Agricultural Drought (SSMI):
  • Historically (1982–2014), 34 events occurred with a mean duration of 1.3 months and severity of 1.7.
  • Under SSP2-4.5, frequency declined to 10 events in the mid-future (2041–2070) before rising to 16 events in the far future (2071–2100).
  • Under SSP5-8.5, frequency showed increased variability with 33 events in the near future (2015–2040), dropping to 2 in the mid-future, and increasing again to 19 in the far future, projecting higher drought activity and persistence.
  • Precipitation was the dominant predictor (importance score 0.602).
• Hydrological Drought (SRI):
  • Historically, 31 events occurred with a mean duration of 2.8 months and severity of 3.4.
  • Under both SSP2-4.5 and SSP5-8.5, hydrological droughts were more persistent, with frequencies ranging from 26 to 36 events across future periods, consistently showing longer durations and higher severities than agricultural droughts.
  • Air temperature was the most important predictor (importance score 0.411).
• Spatial Trends: Historical soil moisture conditions exhibited both wetting and drying tendencies. Future projections under SSP5-8.5 showed a gradual shift toward wetter conditions for SSMI but greater spatial heterogeneity and stronger changes (both wetting and drying) for SRI compared to SSP2-4.5.
• Overall: The study reveals increasing variability in agricultural drought and continued recurrence of hydrological drought, with hydrological droughts being more persistent and severe, particularly under the high-emission SSP5-8.5 scenario in the far future.

Contributions

• Quantified the distinct impacts of climate change on both agricultural and hydrological drought dynamics at a basin scale in the data-scarce Ganale Dawa River Basin, Ethiopia.
• Utilized machine learning-enhanced CMIP6 projections and satellite-based drought indices to improve the accuracy and reduce uncertainties in future drought assessments for tropical regions.
• Identified the differing primary drivers (precipitation for agricultural drought, air temperature for hydrological drought) and characteristics (short-term variability vs. long-term persistence) of these drought types.
• Provided evidence-based insights crucial for developing integrated, dual-pathway adaptation strategies, combining immediate agricultural responses with sustained water management and climate mitigation efforts.

Funding

• NORAD through the NORHED II program for the MERIT project [grant number 60683, 2021].

Citation

@article{Abdulahi2025Impact,
  author = {Abdulahi, Mohammed Mussa and Egli, Pascal and Desta, Anteneh Belayneh and Bamutaze, Yazidhi and Dejene, Sintayehu W.},
  title = {Impact of Climate Change on Drought Dynamics in the Ganale Dawa River Basin, Ethiopia},
  journal = {Climate},
  year = {2025},
  doi = {10.3390/cli13110231},
  url = {https://doi.org/10.3390/cli13110231}
}