水资源供需之间的矛盾导致我国面临前所未有的水资源危机。为了缓解这种危机，需要建立以完善技术工具为基础的水资源管理新方法。分布式水文模型能够满足新的水资源管理方式的技术需求。SWAT（Soil and Water Assessment Tool）模型是美国农业部农业研究局开发的流域尺度分布式水文模型。该模型在北美和欧洲地区得到广泛应用，在我国具有广阔的应用前景。SWAT模型的输入数据种类繁多且在大多数研究中都是海量数据。对输入数据的合理处理及管理是决定最终研究成功与否的重要因素之一。本研究是解决我国模型用户对于SWAT模型输入数据预处理需求的一种探索。在对SWAT模型的数据流进行归纳和总结的基础上提出输入数据预处理的整体工作框架，以此提高模型初次使用者从数据流角度对SWAT模型的认识程度，减少用户熟悉模型需要的时间并指导用户进行输入数据预处理工作。从系统集成角度对输入数据准备工作的需求进行分析，提出建立SWAT模型输入数据预处理系统并进行设计及开发，以期提高数据准备工作的效率、准确度、精确度以及模型使用者对研究区域的熟悉程度。论文的主要成果如下：1．提出在收集的原始数据到最终模型读取的数据文件的转化过程中用户将会面对原始数据、输入数据及输入文件三种数据概念。在对AVSWAT将输入数据处理成输入文件的过程进行分析的基础上，提出建立适用于研究区域模型所需的基本输入数据的内容。2．提出输入数据预处理的整体工作框架，重点归纳了建立适用于研究区域模型所需要的基本输入数据的准备工作。空间数据的预处理包括GIS图层及点状地图要素的预处理，属性数据的预处理包括对土地利用数据、土壤属性数据、点源数据、水库数据、气象数据及管理措施概化信息的预处理。3．提出用户从系统集成角度对SWAT模型输入数据的处理需求并对SWAT模型输入数据预处理系统进行设计。根据需求分析和系统应用目标设计系统的总体逻辑结构及开发模式。在此基础上对系统的应用层和数据层进行设计。该系统以数据为中心，应用层的核心模块包括数据导入模块、数据分析模块、数据处理模块、数据输出模块及GIS模块。数据层包括关系数据库及空间数据，两者通过关键字段进行关联。4．以全球环境基金项目—“基于SWAT模型的漳卫南运河子流域水量过程模拟与非点源污染估算”为例，对SWAT模型输入数据预处理系统进行开发及测试。重点实现实测气象数据的空间插值及天气生成器输入文件的自动生成。

The big difference between water demand and water supply leads to water resource crisis in China. In order to solve the problem, some new water resource management policies which are based on top sciences and techniques have to be implemented. Distributed hydrologic model can fulfill the requirements for making reasonable water resource management policies. SWAT (Soil and Water Assessment Tool) is a watershed level distributed hydrologic model which is developed by USDA-ARS. The model has been widely used in North America and Europe and has a great potential in solving the problem that China confronts. Input data of SWAT model are very diverse. And in most cases, the input data are mass data. Consequently, proper management and preprocessing for input data of SWAT model is one of the most important factors for a successful application of SWAT model. This research is an exploration of fulfilling the requirements of model users in China for preprocessing input data of SWAT model. In order to make beginners be familiar with the model and guide the process of preparing input data, the data flow of SWAT was reviewed and a framework of preprocessing the input data was proposed. The requirements for preprocessing input data of the model in the aspect of software development was analyzed. Then the idea of developing preprocessing system for input data of the model was proposed. And the system which aims at improving efficiency and accuracy of the preparation and making model users be more acquainted with river basin they are studying was designed and developed. The main contributions of this thesis include:1. In the process of converting measured data collected into the files that model use, there are three kinds of data that the model user will deal with, which are raw data, input data and input files. Based on analysis of the process of converting input data into input files which is executed by model user interface-AVSWAT, the list of input data to build a model that can preliminarily reflect the water budget in a specific research area was proposed.2. The framework of input data preprocessing, which focuses on the preparation for the input data by which SWAT can preliminarily simulating the hydrologic cycle in a research area, was established. Spatial data preparations include GIS layers and point map elements preprocessing. Property data will be preprocessed are land use data, soil property data, point source data, reservoir data, weather data and management information.3. According to requirements analysis and objectives of applying the preprocessing system for input data of SWAT model, the general logic structure of the system and system development method were designed. Based on these, designs for system functions and background database were proposed. The core of the system is data. The nuclear functions of the system are data import, data analysis, data processing, data export and GIS functions. The data layer of the system includes GIS data and a relation database, which are linked by key fields.4. The system was developed and tested based on a GEF project-Hydrological Simulation and Assessment of Non-point Source Pollution for the Zhangweinan River Sub-watershed Using SWAT Model. In this case, spatial interpolation of measured weather data and generation of weather generator input file were the main functions being developed.