GIS-Based Study On Runoff Production And Sediment Yield Watershed Modeling In Two Types Of Scales And It's Application On Loess Hilly Region

On the basis of the principle for fluid mechanics, this paper firstly reviewed the mechanical balance condition for loess slope and gully slope area, then the influence to sediment yield from precipitation, slope, soil and vegetation was discussed in accordance with the local characteristic of loess hilly area. After that, the soil erosion and transport processes as well as interaction of loess slope and gully slope were investigated in order to provide necessary methodology for numerical modeling of runoff production, concentration and sediment yield.As for the distributed runoff and sediment yield model for small scale watershed, the primary purpose of this paper is aimed to improve our knowledge on distributed runoff production, concentration and sediment yield model coupling with RS and GIS tools. The Chabagou watershed and Xingzihe watershed, located in the middle stream of Yellow River, with extremely high soil loss rate in China, was selected as the test area to study the distributed runoff and sediment yield model in accordance with the local vegetation and land utilization characteristic on each grid. The model was combined with distributed runoff and sediment yield model, flow concentration and sediment transport model with the support of GIS technology. Excess runoff mode has been applied in runoff production model. The one dimesional kinematic wave theroy and the finite-difference method were used to simulate the overland flow. The calculation formulas of the soil erosion were deduced by means of energy balance principle. The runoff process coupled with sediment yield process at the watershed outlet can be routed from each cell respectively in Preissmann implicit scheme. According to the 17 storm and sediment events from 1970 to 2001 in both the two watersheds, simulated results are satisfactory. At last, the key and sensitive parameters of the model were selected for parameter uncertainty analysis based on GLUE(Generalized Likelihood Uncertainty Estimation) method in order to understand parameter calibration result.For the sediment yield modeling of meso-scale watershed, the Dalihe watershed, which as selected as the test area, can be classified as slope fields, gully fields, basic units and one channel for runoff concentration and sediment transport. The model includes two influence factors (Rainfall, landforms), four coefficients and thirteen relevant equations. The model, based on the physical process, was designated to estimate storm response in runoff and sediment yield. The prediction precision is above 55% according to model validation, so the model can be used for the purpose of soil erosion prediction in loess area with high and coarse sediment.