A spatially explicit simulation of phosphorus loads in the Poyang Lake Sub-basins: An improved export coefficient model with landscape resistance

Study region The Poyang Lake Basin (12.21 × 10⁴ km²) in China, encompassing the five major tributaries—Ganjiang, Fuhe, Xinjiang, Lean, and Liaohe Rivers—was selected as the study area. This region experiences a subtropical humid monsoon climate with annual precipitation ranging from 1387 to 1795 mm. Despite its hydrological abundance, the basin faces increasing eutrophication pressure primarily from non-point source (NPS) phosphorus loads delivered by its tributaries. Study focus To overcome the limitation of the classical Johnes export coefficient model (ECM) in reflecting environmental heterogeneity, this study developed an improved ECM incorporating landscape resistance (LR-ECM). Landscape resistance—a concept characterizing spatial heterogeneity in pollutant transport—offers a novel framework to enhance lumped models for simulating NPS phosphorus loads in large watersheds. We systematically quantified nine key environmental factors (relative elevation, slope, surface roughness, vegetation coverage, rainfall, topographic wetness index, soil erodibility, soil leaching index, and flow length) and integrated them via spatial principal component analysis and the minimum cumulative resistance model to construct a comprehensive landscape resistance surface. This surface was then normalized and coupled with the traditional ECM to simulate the 2020 spatial distribution of phosphorus loads, with results validated using monitoring data at the tributaries' outlets. New hydrological insights for the region The study revealed that hydrological and topographic factors dominated the landscape resistance structure. The LR-ECM achieved markedly superior accuracy, reducing the maximum relative error across the five tributaries from 126.26% to below 9%. Simulations identified livestock farming as the primary phosphorus source, with high-load areas concentrated near watershed outlets and in zones of agricultural intensification. The LR-ECM successfully delineated critical source areas at a hydrologically meaningful (sub-basin) scale, providing a scientific basis for targeted NPS pollution control. This work validates that landscape resistance effectively characterizes the constraints of environmental heterogeneity on pollutant transport, offering a robust tool for precise NPS modeling and management in data-scarce large watersheds.