环江流域径流模拟研究

Simulated Runoff Processes in the Huanjiang Drainage Basin

黄土高原绝大部分地区降水稀少、径流量小,水资源总体相当匮乏,准确掌握区域内径流情况对科学配置稀缺的水资源至关重要。本研究选取黄土高原中部马莲河上游的环江流域为研究对象,采用SWAT模型进行径流模拟研究,并结合1980-1990年的径流实测数据校准和验证模型。分析显示:1采用30m精度的高分辨率数字高程模型(DEM)数据比用90m分辨率DEM数据获得的河网更精细,而模拟误差较小;2子流域划分水平对径流模拟有一定的影响,当子流域划分增加到一定水平时径流量趋于稳定;3通过调整模型中敏感性较高的参数率定模型,使得校准期和验证期月均径流深Re均小于10%,R2分别为0.893和0.822,Ens分别为0.89和0.82,SWAT模型在该地区具有较好的适用性。

The Chinese Loess Plateau is area with low precipitation and small runoff. Accurate understanding of the regional runoff situation is of great significance to the scientific allocation of water resources. The SWAT model is a distributed hydrological model based on good physical processes. Here we utilize the SWAT model as the technical means to simulate runoff in the Huanjiang drainage basin in the upper Malian River,northwestern Loess Plateau. We compared simulation results using 90 m resolution DEM data and 30 m resolution DEM data and found that the latter displays more precise channels,longer channels,steeper average slope and smaller error.The runoff simulation is influenced by sub-watershed division levels whereby when the numbers of sub-watershed division have achieved a certain level,the impact of the division level on the model is ignored. Parameters such as ALPHA_BF,CN2,SOL_AWC,ESCO and SOL_K are more sensitive in the sensitivity analysis and were adjusted in order to obtain better results. These parameters yield results which show relative simulated errors less than 10%,and a relative coefficient and Nash-Suttcliffe coefficient higher than 0.80 in the calibration period 1980-1985 and the validation period 1986-1990. The SWAT model can therefore simulate runoff in the study area and has huge potential for application in the Loess Plateau and northwest China. This model is capable of integrating existing land use,soil type and meteorological data to predict future changes of runoff.