基于PEST自动校正的HSPF水文模拟研究  

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

作　　者：程晓光[1,2] 张静[1,2] 宫辉力[1,2] 

机构地区：[1]首都师范大学城市环境过程与数字模拟国家重点实验室培育基地,北京100048 [2]首都师范大学三维信息获取与应用教育部重点实验室,北京100048

出　　处：《人民黄河》2013年第12期33-36,共4页Yellow River

基　　金：国家自然科学基金资助项目(41271004;40901026);北京市科技新星项目(2010B046)

摘　　要：水文模型率定是评价模型适用性必不可少的部分。基于Levenberg-Marquardt算法的PEST自动率定程序具有优化速度快、模型健壮性好的优点,已被成功应用于许多流域。分别选取2005—2006年、2007—2008年作为校正期和验证期,结合人工试错法和PEST自动率定程序分析参数敏感性及95%置信区间,得到合适参数值,利用HSPF模型对北京延庆县妫水河流域进行水文模拟。人工率定纳什系数E=0.338、确定性系数R2=0.612;PEST自动率定校正期E=0.640、R2=0.787,验证期E=0.737、R2=0.859,并且各参数值均在95%置信区间。表明PEST程序在HSPF模型中具有较好的率定功能,且HSPF模型在研究区具有良好的适用性。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 eal- 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-Suteliffe 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.

关 键 词：HSPF模型 PEST 人工率定 径流 水文模拟 妫水河流域 

分 类 号：P333[天文地球—水文科学] P334.92[水利工程—水文学及水资源]