Divergent human impacts on multi-scale hydrological drought across the conterminous United States

Understanding emerging patterns of human impacts on hydrological drought is crucial for effective water resources management. Yet how human activities influence drought characteristics across time scales remains unclear. Here, we introduce a comprehensive multi-scale framework that integrates three hydrological models with a machine learning attribution approach on a continental scale to quantify human impacts on hydrological drought across the conterminous United States (CONUS). Long-term simulations from three national hydrological models were used to reconstruct natural streamflow during 1961-2010 and diagnose human-induced drought variations through comparison with observed drought events. Results derived from different models consistently indicate that human activities intensified hydrological drought in approximately one third of the CONUS, particularly in the southern Great Plains, while mitigating drought in 19%-25% of the land area, primarily in the west. Intra-annual analysis further reveals evident mitigation in winter but exacerbation in autumn. Multi-scale drought diagnostics demonstrate divergent shifts in spatial and population exposure: with increasing aggregation from monthly to annual scales, the area experiencing intensified drought expands from 28% to 34%, whereas the mitigated area contracts from 25% to 19%. Machine-learning attribution links this divergence to the increasing importance of land-use dynamics for longer-term drought, highlighting dam density as a key driver of monthly drought variability, while changes in forest and cropland dominate controls on annual drought. Our findings reveal scale-dependent and feature-mediated human impacts on hydrological drought, offering new insights into coupled socioeconomic-hydrological interactions and informing drought risk management under ongoing environmental change. • A novel multi-scale framework for quantifying human impacts on hydrological drought • Human-induced drought intensification reveal divergent changes across time scales • Dam and land management exerted different influence on short and long-term drought • Anthropogenic impacts mitigate drought in winter but exacerbate it in autumn