Rainfall-driven agricultural non-point source pollution in reservoir catchments: Variability and reduction strategies.

Agricultural non-point source (NPS) pollution in reservoir catchments is a major contributor to water quality deterioration. Existing studies mainly focused on lowland or large-scale watersheds. However, due to the complex topography and underlying surface conditions of hill-type reservoir catchments, nitrogen transport mechanisms are difficult to characterize effectively using conventional monitoring approaches. Thus, the understanding of the effects of spatiotemporal heterogeneity in rainfall on nitrogen loss remains insufficient, particularly for fine-scale simulation and management at the subbasin scale. This study focuses on a large reservoir catchment in Dalian, China. Based on long-term data of precipitation, discharge, and inflow nitrogen concentration, a calibrated SWAT model was developed to systematically analyze the spatiotemporal effects of different hydrological year types and seasonal rainfall patterns on nitrogen loss. The reductions in nitrogen loss achieved through source control (fertilizer reduction) and ecological filtration (vegetative buffer strips) were quantified, and subbasin-based management strategies were proposed. Results indicate nitrogen loss peaks in August during wet years (≥870.55 mm), while peaks occur in May during normal (643.45-870.55 mm) and dry years (≤643.45 mm). Subbasins with steeper slopes (>11.71°) and higher nitrogen application rates (>123.30 t/y) exhibit higher nitrogen loss. Scenario analysis indicates that combined measures of fertilizer reduction (10%-30%) and buffer strips (1-5 m) effectively reduce nitrogen loss, with the greatest reduction observed in wet years. This study develops a subbasin-scale framework integrating hydrological variability, nitrogen dynamics, and management cost, providing scientific support for refined control of agricultural NPS pollution in hill-type reservoir catchments.