Enhancing Drought Early Warning in Arid Asia and Africa: Comparative Performance of the Evaporative Demand Drought Index under Climate Change

Drought represents a significant hydroclimatic hazard in arid regions of Asia and Africa, where climate change exacerbates evaporative demand, leading to intensified moisture stress, agricultural disruptions, and water insecurity. This study evaluates the Evaporative Demand Drought Index (EDDI) as a complementary tool to traditional precipitation-based indices, such as the Standardized Precipitation Index (SPI) and Standardized Precipitation-Evapotranspiration Index (SPEI), for monitoring drought dynamics, with a focus on flash droughts and evapotranspiration-driven anomalies. Utilizing high-resolution ERA5-Land reanalysis data (0.1° spatial resolution) from 1983 to 2023, EDDI was computed using the Penman-Monteith formulation across multiple timescales, including sub-weekly (1–3 weeks) and monthly (1–12 months) periods. The analysis encompassed diverse arid zones classified by the Intergovernmental Panel on Climate Change, including the Sahara (SAH), Mediterranean (MED), Arabian Peninsula (ARP), West Central Asia (WCA), Eastern Europe (EEU), West Siberia (WSB), East Siberia (ESB), East Central Asia (ECA), and Tibetan Plateau (TIB). Performance assessment involved modified Mann-Kendall trend tests, Sen’s slope estimation, Spearman rank correlations, run theory for drought characterization (duration, severity, intensity, peak, and frequency), and a case study of the 2010 drought event. Results revealed pronounced spatial and temporal heterogeneities in drought patterns. EDDI exhibited stronger drying trends compared to SPI and SPEI, driven by significant increases in reference evapotranspiration (ETo; 2.0–5.16 mm year⁻¹) and temperature (0.02–0.05°C year⁻¹) across most regions, except the TIB, where wetting trends predominated due to elevational effects. In hyper-arid areas such as SAH and ARP, EDDI identified significant drying in 45–60% of grid cells, in contrast to SPI's wetting signals, underscoring EDDI's sensitivity to atmospheric demand independent of precipitation. Spearman correlations between EDDI and SPEI were notably strong (ρ ranging from -0.83 to -0.91 at 1-month scales), exceeding those with SPI (-0.41 to -0.79), particularly in SAH (-0.91 for EDDI-SPEI). Drought frequency intensified post-2000 in all regions except TIB, with EDDI capturing higher severity in SAH and ARP due to elevated ETo. Run theory analysis showed that, at longer timescales, drought duration, severity, and intensity increased, while frequency decreased; EDDI consistently indicated more acute conditions than SPI in water-limited environments. During the 2010 drought, EDDI detected the onset 2–4 weeks earlier than SPI in ARP and SAH, highlighting its utility for rapid-onset events through sub-monthly sensitivity to ETo anomalies. Spatial progression revealed severe drought (category C4) expanding in SAH and EEU at 6–12-month scales, with EDDI and SPEI aligning more closely than SPI, reflecting the influence of vapor pressure deficits and land-atmosphere feedbacks grounded in the Budyko framework. These findings affirm EDDI's role as an indirect proxy for drought stress via evaporative demand, complementing precipitation-focused indices in arid settings where ETo dominates. Limitations include potential overestimation in irrigated areas and assumptions of stationarity under non-stationary climate conditions. Integration of EDDI into operational early warning systems could enhance proactive management, supporting adaptive strategies like water allocation and resilient agriculture amid projected aridification.