Gebresellase et al. (2025) Projected impacts of climate and land use changes on streamflow extremes in the upper awash Basin, Ethiopia

Identification

• Journal: Weather and Climate Extremes
• Year: 2025
• Date: 2025-09-27
• Authors: Selamawit Haftu Gebresellase, Zhiyong Wu, Wada Idris Muhammad, Gebremedhin Gebremeskel Haile
• DOI: 10.1016/j.wace.2025.100806

Research Groups

• College of Hydrology and Water Resources Engineering, Hohai University, Nanjing, China
• College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing, China
• Department of Earth and Environmental Sciences, Wesleyan University, Middletown, CT, United States

Short Summary

This study investigated the projected impacts of climate and land use/land cover (LULC) changes on streamflow extremes in the Upper Awash Basin, Ethiopia. It found that climate change is the dominant driver, significantly increasing high-flow extremes and decreasing low-flow extremes, while LULC changes had statistically non-significant effects.

Objective

• To examine the projected effects and relative contributions of climate change and Land Use and Land Cover (LULC) changes on streamflow extremes (high-flow and low-flow events) in the Upper Awash Basin, Ethiopia.

Study Configuration

• Spatial Scale: Upper Awash Basin (UAB), central Ethiopia, covering approximately 11,607 square kilometers (between 8°16′ N to 9°18′ N latitude and 37°57′ E to 39°17′ E longitude).
• Temporal Scale:
  • Historical: 1985–2014 (GCM training/reference), 1988–2014 (LULC scenarios with baseline climate), 1990–2013 (historical streamflow trend validation), 1990–1997 (SWAT calibration), 1998–2001 (SWAT validation).
  • Future: 2031–2060 (referred to as 2030s) and 2061–2090 (referred to as 2060s) for climate scenarios; 2030 and 2060 for LULC scenarios.

Methodology and Data

• Models used:
  • Hydrological model: SWAT (Soil and Water Assessment Tool)
  • Calibration and Uncertainty Analysis: SWAT-CUP (Sequential Uncertainty Fitting version 2 - SUFI-2)
  • GCM selection: Advanced envelope-based method
  • Downscaling: Perfect Prognosis (PP) downscaling using Generalized Linear Regression Models (GLMs)
  • Land Use/Land Cover (LULC) projection: Cellular Automata-Markov (CA-Markov) model
• Data sources:
  • Climate data:
    • CMIP6 Global Climate Models (GCMs) ensemble mean (selected skilled models: CanESM5, FGOALS-g3, MPI-ESM1-2-LR, IPSL-CM6A-LR, ACCESS-CM2, INM-CM4-8).
    • Reference/Predictand data: CHIRPS (Climate Hazards Group Infrared Precipitation with Stations) for precipitation, ERA5 (fifth generation of atmospheric reanalysis data) for maximum and minimum temperature.
    • Predictor variables for downscaling (from ERA5): Geopotential (Z), Specific humidity (Q), Temperature (T), Mean sea level pressure (MSL), Total precipitation (TP), Westerly wind component (U), Southerly wind component (V), 2-meter temperature (2T).
  • Spatial data:
    • Digital Elevation Model (DEM): 30-meter spatial resolution from United States Geological Survey (USGS).
    • Soil data: Food and Agriculture Organization (FAO) soil database.
    • Land Use/Land Cover (LULC): 15-meter Landsat 7 ETM+ imagery (historical maps for 1985, 2000, and 2015).
  • Observed streamflow: Daily discharge data from the Homble outlet station (Awash Basin Authority).
  • Future scenarios: Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5) for climate, Business-As-Usual (BAU) and Governance (GOV) for LULC.

Main Results

• Climate change significantly impacts streamflow extremes, with high-flow indices showing pronounced increases and low-flow indices exhibiting substantial declines.
  • Under SSP8.5 in the 2060s, Maximum High Flow (MHF) increased by 63.16 %, counts of High-Flow Pulses (HPC) by 26.85 %, and duration of High-Flow Pulses (HPD) by 14.96 %.
  • Under SSP8.5 in the 2060s, Minimum Low Flow (MLF) decreased by 67.11 %, counts of low-flow pulses (LPC) by 34.40 %, and duration of low-flow pulses (LPD) by 5.95 %.
• Land Use and Land Cover (LULC) changes demonstrated statistically non-significant effects on both high- and low-flow indices across all scenarios and periods.
• Climate change is identified as the dominant driver of future hydrological extremes in the Upper Awash Basin.
• LULC projections:
  • Business-As-Usual (BAU) scenario projects urban areas expanding to 1,196.78 square kilometers (10.15 %) and cropland to 9,159.21 square kilometers (77.71 %) by 2060, with significant declines in forest and shrubland.
  • Governance (GOV) scenario emphasizes sustainable land management, controlling urban sprawl (665.80 square kilometers or 5.65 % by 2060) and increasing forest and grassland cover.
• The SWAT model showed good performance during calibration (R²=0.79, NSE=0.77, PBIAS=15.4 %, KGE=0.79) and validation (R²=0.74, NSE=0.71, PBIAS=18.8 %, KGE=0.54), with P-factor values > 0.5 and R-factor values < 1.5.
• Downscaling methods: Method M1 was optimal for precipitation, and Method M3 was optimal for temperature, based on consistent performance and accuracy in representing extreme values.

Contributions

• This study fills a critical gap by being the first to directly assess the relative contributions of climate change and LULC changes to hydrological extremes (high-flow and low-flow events) in the Upper Awash Basin.
• It provides essential insights for guiding future water resource planning and climate adaptation strategies in the UAB, emphasizing the need for climate-focused adaptation due to the overwhelming influence of climate change on streamflow extremes compared to LULC changes.

Funding

• National Natural Science Foundation of China (U2240225)

Citation

@article{Gebresellase2025Projected,
  author = {Gebresellase, Selamawit Haftu and Wu, Zhiyong and Muhammad, Wada Idris and Haile, Gebremedhin Gebremeskel},
  title = {Projected impacts of climate and land use changes on streamflow extremes in the upper awash Basin, Ethiopia},
  journal = {Weather and Climate Extremes},
  year = {2025},
  doi = {10.1016/j.wace.2025.100806},
  url = {https://doi.org/10.1016/j.wace.2025.100806}
}