水质模型的全局参数敏感性分析

Global parameter sensitivity analysis on water quality model

LH-OAT全局参数敏感性分析方法集成了LH抽样与随机OAT方法的优点,在随机OAT的基础上融入LH抽样法,保证参数敏感性分析的强壮性及有效性。但LH-OAT方法必须依靠复杂的计算机程序来实现,因为没有通用程序而限制了其推广应用。试验研究了LH-OAT方法的通用程序设计,并成功用于WASP富营养化水质模型中浮游植物动力学模型的全局参数敏感性分析。结果表明在充足供氮、磷时,浮游植物生物量变化率最敏感的参数是水温,其次是浮游植物的内源呼吸速率和温度调节系数,其三是浮游植物中叶绿素a与碳的比值,处于第四位的是浮游动物的数量;各参数合理的取值范围是成功使用LH-OAT方法的关键。

The LH-OAT sensitivity analysis method for the global parameters integrates the advantages of the LH sampling method with a random OAT method to ensure the robustness and effectiveness of the parameters.LH-OAT method must rely on complex computer program to achieve,however its application is limited because there is no generic procedures currently available in computer software.This paper investigated the generic procedures to apply LH-OAT method,and successfully used it for the phytoplankton dynamics model which was the core of the WASP water quality model.The results showed that in the condition of adequate supply of nitrogen and phosphorus,the most sensitive parameter was water temperature,followed by the endogenous respiration rate of phytoplankton and the carbon to chlorophyll ratio,and the fourth was zooplankton predation rates on phytoplankton.The reasonable range of each parameter was the key for the successful use of LH-OAT method.