Data-driven assessment of nitrogen and phosphorus buffering capacity across 460 Chinese watersheds: Spatial patterns,...

Watershed nitrogen and phosphorus buffering capacity refers to the capacity of a watershed to retain nitrogen and phosphorus in soil, groundwater, and sediments, playing an important regulatory role in balancing human-induced nutrient inputs with downstream pollution outputs. However, its distribution, driving factors, and future trends have not been quantified at a national scale. Here, we analyzed data from 460 Chinese watersheds using geographic spatial analysis and explainable machine learning. The average nitrogen and phosphorus buffering indices are found to be 0.22 and 0.54 in China, with poorer buffering capacity (0.16 and 0.41) in Southeast Basin. Nitrogen buffering capacity is primarily driven by soil moisture, topography and soil properties, while phosphorus buffering capacity is mainly determined by precipitation and landscape pattern. Additionally, we found that environmental extremes had a stronger influence on buffering capacity than mean values, highlighting ecosystem sensitivity to hydrological and climatic extremes and the associated risk of nutrient loss under future climate change. In the future, nitrogen and phosphorus buffering capacity are projected to decline in approximately 41 - 55 % and 84 - 87 % of watersheds, respectively. This study reveals the spatial patterns and drivers of watershed nitrogen and phosphorus buffering capacity in China and proposes a transferable framework for other regions, demonstrating the potential of combining interpretable machine learning with causal inference for yielding robust and region-specific thresholds to guide locally tailored management interventions.