Water Resources Research, Год журнала: 2025, Номер 61(5)

Опубликована: Май 1, 2025

Abstract Human activities profoundly impact the terrestrial water cycle and spatiotemporal dynamics of blue green resources. Distributed hydrological models are essential for simulating resources within a basin. However, neither process‐based nor data‐driven have fully captured effects human on distribution in space time. Here we construct distributed machine learning model monthly resources, which is trained calibrated Yellow River Basin (YRB) China, validated tested transferability to similar climatic zones case Colorado (CRB) United States. The modeling thoroughly accounts influence activities, incorporating 5 scales (grid, county, city, province, cluster), 4 algorithms, 2 integration methods (Stacking Bayesian). R values reached 0.84 0.97 models, respectively, during test period YRB. corresponding high accuracy maintained with 0.72 when transferred CRB. performed better regions higher activity intensity. Precipitation spatial encoding respectively most sensitive feature variables while nighttime lights population density significant activity‐related features. study highlights non‐negligible impacts socioeconomic factors feasibility modeling.

Язык: Английский