Development of study on AnnAGNPS model and its application | AnnAGNPS模型研究及应用进展

chinois. AnnAGNPS（Annualized Agricultural Non-Point Source Pollution Model）模型是由美国农业部开发的用于模拟评估流域地表径流、泥沙侵蚀和氮磷营养盐流失的连续型分布式参数模型。文章对AnnAGNPS的国内外研究应用进行了全面综述，以期为模型在我国的深入研究和应用提供参考。从AnnAGNPS模型的机理、结构、国内外应用现状及我国应用中存在的问题等方面作了综合性总结。模型应用情况表明，模型对径流和非点源污染的模拟基本在可接受精度范围内，对于地表径流的模拟能力要强于泥沙和氮磷营养盐，对总磷输出的模拟表现出了较大的不确定性；年、月时间尺度模拟精度高于单场降雨的模拟精度；单场降雨雨量越大，模拟效果越好；大尺度区域模拟精度低于小尺度区域；对洪峰流量估计过高。我国目前对模型的适应性和重要参数的敏感性分析方面研究较多，受相关数据获取困难等因素的影响，AnnAGNPS模型的应用缺乏系统性和连续性。对模型的修正或改进，以及非点源管理措施效果的模拟评价等方面的研究有待加强。

anglais. AnnAGNPS, a continuous and distributed model developed by USDA, can simulate and evaluate the yield quantities of runoff, sediment, and nutrients of a watershed on a daily, monthly, or yearly basis according to user’s specification. In this paper, study and application on AnnAGNPS model at home and abroad were analyzed in detail to offer reference for applying the model in our country. The principle, structure, present situation of appliance at home and abroad as well as problems of using the AnnAGNPS model in our country were presented. Application of the model showed that: Simulation results on runoff and non-point source pollution (NSP) of the model were accepted, of which, the simulation precision on runoff was higher than that on sediment, nitrogen and phosphorous nutrients. Simulation on total phosphorous took on some uncertainty. The simulation precision on year or month was higher than that on single rainfall. The amount of precipitation of single rainfall was bigger and the simulation effect better. The simulation precision on large-scale region was lower than that on small-scale one. The estimation on peak discharge of the model was too high. There have been a number of studies on the adaptability of the model and sensitivity of important parameters in our country, but affected by the difficulty to acquire relevant data, the application of AnnAGNPS model lacks of systemic and continuity. Modification or improvement on the model and effects simulation as well as evaluation of non-point source management measures need to be put into more studies.