由于长期以来养分的人为过量输入，许多国家和地区的土壤—地下水系统养分持续累积并已对非点源养分污染产生较高贡献。对这一累积效应，目前尚缺乏相应的定量研究，且现行的非点源污染模型和方法也难以满足其定量需求，阻碍了养分污染控制进程。本项目将基于流域空间统计模型的建模思路、物质质量守恒原理、计量经济学的考伊克转化方法，将不易测算的土壤—地下水系统累积养分量及其产生的养分入河量描述为相关流域属性变量和参数的函数，创建由若干个可测算变量和未知参数组成的考虑累积效应的流域非点源污染动态模拟模型；采用野外监测和调查统计、稳定性同位素分析/数字滤波、遥感影像解译、最值归一化、贝叶斯统计、净人为养分累积量计算等手段解决模型参数化、校正、多重验证及不确定性评价等问题；从而实现考虑累积效应的流域非点源污染分区、分类、分期定量，揭示各养分源滞后效应作用时间长短和养分累积的动态规律，为有效控制养分污染提供科学依据。

Due to long-term anthropological nutrient input in excessive pattern, the nutrient amount accumulated in soil-groundwater system has been increased continually and has yielded a considerable contribution proportion to riverine nutrient load. This accumulative effect has been recognized as one of major causes of the few decreases observed in riverine nutrient level after several years’ or decades’ implementation of nonpoint source (NPS) nutrient pollution control practices in many regions. However, few efforts have been performed to quantify this accumulative effect for guiding nutrient pollution control practices. Most important, this accumulative effect was rarely considered in many current NPS pollution models and approaches. Learning from the methodology in developing watershed spatially statistic models, law of mass conservation, and econometric Koyck conversion approach, accumulated nutrient amount in the soil-groundwater system and its resulting nutrient export load to the river, which are difficult to be measured and estimated, will be both skillfully described as the functions of several watershed attribute variables and parameters in this project. Based on this critical step, this project will develop an original dynamic watershed NPS pollution model considering the accumulative effect, which only contains several measurable or calculable variables and several unknown parameters . Combining the field experiment and survey, stable isotope analysis/digital filtering, remote sensing imagery interpretation, normalization, Bayesian statistic, and net anthropogenic nitrogen accumulation approach, the project will achieve the database establishment on watershed attribute variables and the model parameterization, calibration and multiple validations and uncertainty assessment. Relying on these efforts, watershed dynamic NPS nutrient pollution processes will be described quantificationally. The nutrient pollution contributions of different nutrient sources, their lagged effects’ duration lengths, and dynamic nutrient accumulation in response to long-term anthropological activities will be addressed explicitly. All of these results will provide necessary scientific evidences for advancing nutrient pollution controls.