基于PSUADE的VIC模型参数敏感性分析——以太湖西苕溪流域为例

Parameter Sensitivity Analysis of a Variable Infiltration Capacity Model Based on PSUADE

敏感性分析是增强水文模型率定的针对性、提高率定效率的有效方法。本文以太湖流域西苕溪为例,探究分布式水文模型VIC(Variable Infiltration Capacity)的参数敏感性及参数间的交互作用。为使本研究更具推广性,选取了关注高流量过程和关注整个流量过程线的两类目标函数。结果表明,可变下渗率曲线形状参数B、最大基流容量Dmax和最上层土壤层厚度d1对高流量过程较为敏感;当关注整个流量过程线时,除以上3个参数外,非线性基流开始时占最大基流的比例Ds也较为敏感。在参数的交互作用方面,当关注高流量过程时,参数d1与参数B的交互作用最为敏感;当注重整个流量过程时,参数d1与所有其它参数的交互作用均较为敏感。

The variable infiltration capacity(VIC)model is a distributed hydrological model widely used around the world and in China. Parameter sensitivity analysis of the VIC model is of great significance in its application and is an important procedure before model calibration,especially for manual calibration. The main purpose of sensitivity analysis is to determine parameters that have greater influence on model output to improve calibration efficiency and better understand hydrological processes. Here,we focused on parameter sensitivity analysis for the VIC model when applied to the Xitiaoxi catchment,Taihu Basin,eastern China. In order to comprehensively analyze parameter sensitivity,two kinds of objective functions were used:one kind paid more attention to high flows,while the other paid equal attention to the whole hydrograph. The results demonstrated that when the objective functions paid more attention to high flows,the variable infiltration curve shape parameter B,maximum velocity of baseflow Dmax,and thickness of soil moisture uppermost layer d1 were very sensitive. The objective functions paid equal attention to the whole hydrograph,despite the three parameters mentioned above,and parameter Ds which referred fraction of Dsmax where non- linear baseflow begins,was also sensitive. Regarding second order parameter sensitivity,interaction between d1 and B was the strongest when more attention was paid to high flow. Parameter interactions were stronger when paid equal attention to the whole hydrograph than high flow,especially interactions of d1 with all other parameters.