Modified Modularized Distributed Hydrological Model SWAT And Its Application In Lianshui Basin

In this paper, the following input data are collected through internet and field investigation: DEM data, landuse map, soil class map and weather data. The spatial database for runoff simulation and physical attribute database are also created. A distributed hydrological model--SWAT(Soil and Water Assessment Tool),developed by USDA ARS (Agricultural Research service), has been applied to simulate the runoff process in the Lianshui Basin (the upper reaches of Loudi). In the simulation, the study area is delineated into 11 subbasins and 119 HRUs. To make the SWAT become applicable in the study area, the experimental modification of SWAT model is also performed by replacing surface runoff module with Xin'anjiang model.In this paper, applying ASTER GDEM can extract the study area's hydro-information accurately, such as subbasin and river network. This is the foundation for model simulation in the study area.Simulating the daily flow and monthly average natural runoff in the study area indicates that the modified SWAT model is suitable for the Lianshui Basin. The 1990～1994 data series of daily and monthly average natural runoffs from Loudi gauging station are adopted for parameter calibration of the model, and the 1995～1999 daily and monthly average natural runoff s data from Loudi gauging station was used to calibrate the model. The results show that modified SWAT model could be applied to study area reasonably according to the high correlation coefficients and Nash–Suttclife coefficients. The simulation revealed the water resource variation accurately and provided theoretical support for water resource planning and management.Using modified model to simulate the monthly average runoff is also valid. This research shows that the modified model is also rational in simulating the monthly average runoff. When the monthly average runoff is bigger, the simulation value of modified model is more accurate and closer to measured data. Furthermore, when the monthly average runoff is smaller, the simulation value of SWAT model is relatively more accurate. This research indicates that the ground water simulation value is not so good and the module of ground water needs further improvement.This research also shows that it is not only necessary to build a distributed hydrological model suitable for the nature and geographical environment and lack data availability in China, but also feasible by adopting the established module of advanced hydrological model from foreign countries, and modifying or replacing the unsuitable module with new modules developed from the actual situation of our country.