Research on the HSPF Hydrologic Simulation based the PEST Automatic Calibration-Case Study:Guishui River Basin,Beijing

宫辉力,程晓光,张静
DOI: https://doi.org/10.3969/j.issn.1000-1379.2013.12.012
2013-01-01
Abstract:Parameter calibration is the essential step for successful application of hydrologic models. The PEST,based on Levenberg-Marquardt,a nonlinear and automatic method to calibrate hydrologic models,has been widely used in many basins successfully. Guishui River,located at north-west of Beijing and faced with scarce water resources,is selected as the study area. The years of 2005-2006 and 2007-2008 were selected as cal-ibration year and validation year respectively. Manual calibration and automatic calibration were used simultaneously in the paper. First,the suit-able initial values of parameter were selected as the order of yearly runoff,monthly runoff and flow hydrographs by manual calibration. Then param-eter sensitivity and 95% confidence limits were analyzed with PEST and all the estimated values of parameters were within the limits indicated the result was reliable. Goodness-of-fit of the PEST calibration model uses the Nash-Sutcliffe coefficient (E)and coefficient of determination (R2 ). Statistical comparisons between simulated data and observed date for the calibration year give a reasonable agreement with E equal to 0. 640 and R2 equal to 0. 787 though the data of 0. 338 and 0. 612 by manual calibration. These values are respectively 0. 737 and 0. 859 for validation year. O-verall,PEST is a reasonable method and HSPF model has the good applicability on the Guishui River basin.
What problem does this paper attempt to address?