Research on multi-parameter stepwise coupling inversion method for marine integrated element water quality model based on data-driven and genetic calculation

Abstract

In this method, there are two inversion phases step by step, namely, preliminary inversion and precise inversion. Preliminary inversion is implemented by coupling data-driven model with Marine Integrated Element Water Quality Model (MIEWQM), to determine the relationship between the multi-parameter and the pollution concentration values in gauge stations. Field data are inputted the relationship to obtain the initial values of multi-parameter matching. MIEWQM is embedded into Genetic Algorithms (GAs) to search global multi-parameter optimal values in the precise inversion. Preliminary inversion provides the initial values of multi-parameter matching to the precise inversion for shortening the value-taking range of population in GAs. Case studies in Bohai Bay, China. An “identical twin experiment” method and a practical application with Dissolved Inorganic Nitrogen (DIN) are used to verify the present method.