Advancing basin-scale drought monitoring: Development of a regional combined drought index using precipitation, soil moisture, and vegetation data
Article
Abebe, AK, Zhou, X, Lv, T et al. (2025). Advancing basin-scale drought monitoring: Development of a regional combined drought index using precipitation, soil moisture, and vegetation data
. AGRICULTURAL WATER MANAGEMENT, 318 10.1016/j.agwat.2025.109734
Abebe, AK, Zhou, X, Lv, T et al. (2025). Advancing basin-scale drought monitoring: Development of a regional combined drought index using precipitation, soil moisture, and vegetation data
. AGRICULTURAL WATER MANAGEMENT, 318 10.1016/j.agwat.2025.109734
Drought remains a critical challenge in Ethiopia's Awash River Basin (ARB), where rainfed agriculture is highly sensitive to climate variability. This study presents a regional Combined Drought Index (rCDI), integrating the Standard Precipitation Index (SPI-3), soil moisture anomaly (SMA), and Vegetation anomaly (VA) using a Principal Component Analysis (PCA)-based weighting approach. Monthly gridded data from 2001 to 2023 were used to generate dynamic, grid-specific weights, capturing spatiotemporal drought variability across the basin. The rCDI was validated against independent station-based SPI-3 data, detrended crop yields (maize and sorghum), and documented drought events for both the Belg (short rainy) and Kiremt (long rainy) seasons. Analysis combined Google Earth Engine (GEE) with Python via Google Colab. Strong correlations (r > 0.70) were observed with SPI-3 were observed in most areas, though weaker (r > 0.45) in arid Belg Zones. Crop yield analysis revealed stronger rCDI sensitivity to maize in the upper ARB and sorghum in upland/northwestern areas, reflecting crop-climate adaptation. The rCDI effectively captured major droughts (2002/2003, 2008–2012, 2015, and 2022), consistent with reported socio-economic impacts. Seasonal patterns showed Belg experiencing more frequent and severe droughts than Kiremt. Statistical trend analysis confirmed rCDI's strength in monitoring evolving drought conditions, supporting early warning and sustainable resource management. A statistical downscaling using Artificial Neural Network (ANN) enhanced soil moisture resolution from 10 km to 1 km, improving rCDI's accuracy. By integrating meteorological, agricultural, and ecological dimensions, the rCDI provides a comprehensive tool for basin-scale drought assessment and monitoring in data-scarce, climate-sensitive regions.
