SWAT分布式流域水文物理模型的改进及应用研究

Improvement and Application of SWAT—A Physically Based,Distributed Hydrological Model

SWAT(Soil and W ater Assessm ent Tool)模型是一个集成遥感(RS)、地理信息系统(G IS)和数字高程模型(DEM)技术的先进的分布式流域水文物理模型。为了推动该模型在中国的适应性研究及应用,并改进模型以提高水文模拟的精度,针对模型在中国西北寒旱区的黑河流域和中西部温润的汉江流域的水文模拟中发现的问题进行了扩充和改进,增加了土壤粒径转换模块和天气发生器(WGEN)数据预处理模块,改进了模型中的WGEN算法、潜在蒸散量模拟算法以及气象参数的空间离散方法。利用扩充和改进后的模型对汉江褒河上游江口流域的降雨-径流过程进行了系统的研究。结果表明,不仅模型的使用效率有明显提高,而且改进后模型的效率系数和相关系数也比改进前有较大改善。

SWAT （ Soil and Water Assessment Tool） model is an advanced, physically based, distributed hydrological model that is integrated with Remote Sensing （ RS）, Geographic Information System （GIS） and Digital Elevation Model （DEM） techniques. In order to impel its application study in China and improve the simulation precision of the model in hydrological process studies, some expansions and improvements were made to the modeling system for the real conditions on arid, semi-arid Heihe River Basin in the northwestern China and humid Hanjiang River Basin in middle-western China. Firstly, soil grain size transferring module and data pre-processing module for weather generator （ WGEN ） were added to original SWAT model to encounter the inconsistent problems existed in data standard utilized in Euramerican and Chinese systems. And then the further improvements to the model were done focusing on： 1 ） three new spatial interpolation methods, including Inverse Distance Weighing method （ IDW）, Thiessen polygonal method and DEM-based PRISM method, were used as a new choice to replace the original interpolation method that takes the meteorological data from the nearest weather station as the representative weather data for each sub-basin. 2） An advanced approach was proposed to modify the original method in daily maximum and minimum temperature estimations in SWAT for avoiding the logical mistakes in temperature simulations. 3） An areal evapotranspiration estimation module based on improved Penman-Monteith method was developed to replace the original approaches used in SWAT. At last the offline experiment on rainfall-runoff simulations were conducted using the meteorological, land use/cover and soil data from the Jiangkou Basin, the Hanjiang River by means of the improved SWAT model and its original version. The detailed study suggested a significant improvement of modeling performance by the fact that in model calibration pe- riod, the Nash-Sutcliffe Criterion （NSC） and correlation coeffcient were improved from 0.47 and 0.79 to 0.92 and 0.98 respectively, while in validation period, they were improved from 0.08 and 0.68 to 0.94 and 0.97 accordingly. The improved SWAT model and its software packs will impel the application studies in the relevant disciplines in China.