Compound drought stressors drive vegetation decline in the African Great Lakes region: a multiscale causal analysis

Abstract Compound hydroclimatic stresses are increasingly threatening vegetation and food security in data-scarce regions like the African Great Lakes region (AGLR). Yet, their effects on vegetation are poorly quantified, limiting the development of effective drought response and food security policies. This study aims to identify the direct drivers and causal pathways of compound drought impacts on vegetation productivity across the region. To address this, this study conducted a comprehensive 25 year analysis (2000–2024) integrating satellite-derived normalized difference vegetation index with key hydroclimatic variables, including precipitation, temperature, soil moisture (SM), vapor pressure deficit (VPD), and teleconnections. Using a wavelet-based causal discovery Peter and Clark momentary conditional independence framework, we move beyond correlation to identify the direct drivers of vegetation stress. Our analysis reveals that croplands and shrublands experienced a 15% greater decline in greenness compared to forests during major climate events, such as the 2015–16 El Niño and the 2023–24 positive Indian Ocean Dipole. Importantly, the causal framework demonstrates that this severe vegetation stress is not primarily driven by precipitation or temperature anomalies alone. Instead, it is causally linked to the combined influence of coinciding VPD and SM deficits. This study provides clear, data-driven evidence that the combined effects of high atmospheric dryness and low SM are linked to significant vegetation productivity loss across the AGLR. These findings provide a scientific foundation for urgently integrating compound VPD-SM metrics into regional drought early-warning systems to improve the lead time and accuracy of food security interventions in the AGLR and similar climate-vulnerable regions.