二维水质模型参数的敏感性分析

Analysis of parameter sensitivity of two-dimensional water quality model

建立了二维水量-水质耦合模型,并增加了随机函数生成模块与Regionalized Sensitivity Analysis(RSA)功能。以太湖总氮为例,运用RSA方法分析该模型参数的敏感性。结果表明:有机物沉降速率VS3最为敏感,对总氮模拟结果的影响较大;太湖各分区的参数敏感性不尽相同,如监测点平台山的可接受参数的敏感性顺序为:VS3>Kpzd>K12>K71>K2>VS4>K2D>GP1,而东太湖34参数的敏感性顺序则为:VS3>VS4>Kpzd>GP1>K2D>K12>K71>K2。说明模型参数的敏感性差异主要由于各分区的水质条件及水环境变化规律不同。本文确定的模型参数敏感性分析方法,可为水质模型的参数选取提供依据。

Two-dimensional coupling model of water quality and water quantity was established. Then a random function module and Regionalized Sensitivity Analysis（ RSA） function were added into the model. Taking total nitrogen of Taihu lake for example,the paper used RSA method to analyze the parameter sensitivity of established model. The results show that organic sedimentation rate VS3is the most sensitive and has greater influence on the simulation result of total nitrogen; the parameter sensitivity of each subarea in Taihu Lake is different,the sensitivity order of acceptably parameters for Pingtaishan monitoring point is VS3＆gt; Kpzd＆gt; K12＆gt; K71＆gt; K2＆gt; VS4＆gt; K2D＆gt; GP1,and the parameter sensitivity order of East Lake 34 is： VS3＆gt; VS4＆gt; Kpzd＆gt; GP1＆gt; K2D＆gt; K12＆gt; K71＆gt; K2. The research indicated that the difference of model parameter sensitivity is mainly owing to water quality conditions and variation of water environment in different subareas,which means that the sensitivity of parameters is also different with the environment changes of simulated subarea. The article determined a set of sensitivity analysis method of model parameter and can provide a reliable basis of parameter selection for application of water quality model in governance and plan of water environmental pollution.