A dynamic eco-hydrological drought framework integrating soil moisture memory, human pressure, and vegetation-atmosphere coupling under a non-stationary climate.

Drought is a complex environmental phenomenon emerging from interactions among climate variability, ecosystem processes, and human activities. However, widely used drought indices such as SPI and SPEI rely on meteorological anomalies and often neglect key eco-hydrological feedbacks that govern terrestrial water stress under climate change. To address this limitation, the Dynamic Eco-hydrological Drought Index (DEHDI) was introduced, a process-based framework that integrates vegetation-atmosphere coupling, recursive soil moisture memory, anthropogenic pressure, and CO2-mediated physiological regulation of evaporative demand. The framework combines vapor pressure deficit, normalized difference vegetation index, a Human Activity Proxy (HAP), and a CO2-adjusted potential evapotranspiration component within a recursive ecohydrological structure. Model parameters were optimized using Elastic Net regularization and evaluated across 25 climatically diverse stations in Iran during 1982-2024. 10-fold cross-validation against TerraClimate-derived soil moisture reference dataset demonstrated substantial improvements over conventional drought indicators. DEHDI achieved a mean Area Under the Receiver Operating Characteristic Curve (AUC) of 0.84, compared with 0.65 for SPEI and 0.61 for SPI, while also providing markedly higher categorical drought detection skill based on Matthews Correlation Coefficient (MCC) and Heidke Skill Score (HSS). The greatest improvements were observed in arid and semi-arid regions, where drought evolution is controlled by cumulative eco-hydrological stress and human-induced pressures rather than short-term meteorological anomalies alone. Sensitivity analyses identified soil moisture memory and anthropogenic pressure as dominant contributors to model performance. Furthermore, incorporating CO2-regulated physiological responses reduced drought overestimation associated with temperature-driven increases in atmospheric evaporative demand. The results demonstrate that drought should be viewed as an emergent eco-hydrological and socio-environmental process rather than solely a meteorological anomaly. By explicitly integrating ecological, hydrological, climatic, and anthropogenic controls, DEHDI provides a physically interpretable framework for drought monitoring, early warning, and environmental risk assessment in human-modified landscapes under accelerating climate change.