Study On Characteristics And Modelling Methods For Aquatic Ecological Environment Of Bohai Bay Using Hydroinformatics Technology

The Bohai Bay is a semi-enclosed sea bay with mild-slope mud beach and shallow water. The water exchange and self-purifying capacity of this sea bay are very weak. With the rapid development of economy and increasing pollutant emission along coastal area, marine ecological environment issues have become increasingly serious, such as serious eutrophication and frequently occured red tides. Research of coastal ecological environment characteristics and modelling methods has great significance for understanding of marine ecological processes, protecting coastal aquatic ecosystem and marine management.Due to the high complexity, the non-linearity and time-space specificity of coastal ecological environment, the internal mechanism of the system and its dynamic processes have not been fully understood, which results in the difficulty of using the conventional deterministic ecohydraulics models. Meanwhile, with the rapid increase of the quantity of coastal ecological environment data including romete sensing data, more and more information technologies are applied to the research of aquatic ecological environment. This dissertation studied the characteristics and modelling methods of coastal ecological environment using new hydroinformatics technologies.This dissertation firstly accounts for the problems of study on the characteristics and modelling methods for coastal ecological environment, including: 1) How to effectively extract some potential useful informations and knowledge from massive data, 2) How to availably reflect the spatial characteristics of aquatic ecological environment data, 3) How to availably simulate the high complexity and the non-linearity of coastal ecological environment, 4) How to availably reflect the effects of local action and spatial discrepancy during dynamic evolution, and 5) How to carry on the integration of deterministic hydrodynamic module with nondeterministic ecological module.Secondly, the ecological environment romete sensing data of Bohai Bay have been analyzed using data mining methods and spatial statistical analysis. Some potential useful knowledge for the aquatic ecological environment of Bohai Bay was extracted using clustering analysis, association analysis and decision tree analysis. The results of spatial autocorrelation analysis and spatial autoregressive model showed that there was a signifieant spatial autocorrelation in ecological environment index during the red tide and detected some anomalous changes of spatio-temporal distribution in ecological environment index.Thirdly, the high complexity and the non-linearity of coastal ecological environment were studied using nondeterministic soft computing, and an ecological model based on hybrid soft computing (Eco_HSC) was established. The results show that the Eco_HSC model can be preferably used to simulate and predict chlorophyll-a concentrations of each site in the Bohai Bay, and it has good generalization ability. Considering the effects of local action, spatial discrepancy and external factors, an integrated cellular automata and support vector machine model (CA-SVM) was developed to simulate the remote sensing chlorophyll-a concentrations in the Bohai Bay. The results show that the CA-SVM model can catch the spatio-temporal variation of chlorophyll-a concentrations.Finally, this dissertation tries to carry on discussion and research in the integration of deterministic method and nondeterministic method. An integrated ecohyrdraulics model has been developed by coupling the simulation results of hydrodynamic module with CA-SVM module. The results indicate that the intergrated model has better performace than CA-SVM, and can be applied to simulate the spatio-temporal variation of chlorophyll-a concentrations in the Bohai Bay.