The Application Of Distributed Hydrologic Model MIKESHE To Simulate Hydrologic Responses Of Small Watershed In North China

It has been well recognized that the physically-based distributed hydrological models could be used for in-depth analysis for watershed hydrology and watershed management decision making. Therefore, development and application of those models has been becoming more and more popular. Features of physically-based distributed hydrological model in terms of application, scale issues, calculation domain, integration with GIS and RS, calibration, validation have been reviewed in the paper.Taking the MIKESHE model for example, the functional modules were introduced separately, which make up of the DHM and based on hydrologic process. Applying MIKESHE to the Tumen experimental watershed and Maoshigou watershed in Miyun county,Beijing. Rainfall-runoff simulation files were made by using data such as precipitation and ET observed in 2004 and 2005 and the DEM, landuse in 2002. The spatial grid of all the DEMs were 10m×10m, 30m×30m, and 50m×50m, the step of the temporal data were 5 or10 minutes, 1 or 2 hours.Firstly, parameters were calibrated by the simulation of 2004-7-21 rainfall-runoff in tumen watershed, after which a list of parameters and initial conditions were confirmed. And then, other rainfall-runoff simulations (2004, 2005) of the two watersheds were used to model validation. Combining the former research on soil, vegetation, runoff, the simulation result was analyzed and the reason that causes the faults was found. After parameters calibration and model validation, test the sensitivity of MIKESHE model to parameters and the impact of temporal and spatial resolution on model simulation result, evaluate the effects of 3 hypothetical scenarios on discharge processes on watershed scale in the 2005-6-28, 2004-7-21, 2005-7-9 rainfall- runoff simulation.Simulation results show that: the calibration and validation results are reasonable; the impact of temporal and spatial resolution on model simulation is, the higher the temporal and spatial resolution is, the better the simulation results are, and the longer the model runs. Sensitivity test of MIKESHE model to parameters suggested that: the initial water depth (overland flow threshold water depth) significantly influences the peakflow occurring time. The bigger the initial water depth is, the later the runoff peak appears. The overland Manning coefficients and river bed resistance determine the overland flow and river conflux time, the bigger the manning number (m) and river bed resistance are, the faster the flow...