Numerical Simulation And Decision Support System Of Water Environment In The Min River Estuary Of China

Estuarine area is one of the most densely populated and the most active on human economic and society activity areas.With the rapid development of society and economy,the water environment of estuaries is also disturbed and destroyed by human activities,and hence affects the sustainable development of estuary area.Mathematical models are used to simulate and analyze the characteristics of estuarine water environment and its changing process,which can systematically reveal the changing rule of estuary water environment.On this basis,multiple heterogeneous water environment models were integrated uniformly to build the estuary water environment decision support system which can provide strong decision-making support for estuary water environment management and is an important non-engineering measure of estuary water environment management.Those have important theoretical and applied values.The Min River Estuary was selected as the study area and the following four aspects were included.(1)A two-dimension hydrodynamic and water quality model was established based on underwater topographic data of the Min River Estuary.The hydrological observation data of water level,velocity and the water quality observation data of CODMn and NH3-N concentration were adopted for model calibration.Multiple typical scenarios were designed to analyze the spatial and temporal distribution characteristics of hydrodynamic and water quality in the Min River Estuary and its change law of motion.(2)GIS technology was used to analyze the riverbed evolution characterstics of the Min River Estuary in recent years by the variation of the channel talweg and crossing section elevation.After that,the calibrated hydrodynamic and water quality model was applied to study the hydrodynamic and water quality respond to the evolution of riverbed under by the intensive disturbance of human activity.(3)By artificial intelligence method,a data-driven model was constructed based on genetic algorithm and support vector machine.The runoff,high tide and previous salinity were selected as the major influence factors of salinity variation.The estuary salinity prediction model based on GA and SVM was proposed,constructed and applied to the salinity prediction of the Min River Estuary.(4)Based on cloud computing,WebGIS and other advanced computer technologies,a system framework was built.For the heterogeneous problem of water environment models,an XML representation framework of model ontology based on the water environment domain knowledge was proposed.Multiple heterogeneous water environment models were uniformly integrated into estuary water environment decision support and applied to the water environment management of the Min River Estuary.The results of this study indicated that:(1)The two-dimension hydrodynamic and water quality model can well reflect the motion variance law of the water dynamics and water quality of the Min River Estuary.(2)The spatial difference of the Min River Estuary riverbed evolution is obvious.The South Branch is serious,next is the mainstream,least is the North Branch.(3)The the division ratio between the South Branch and the North Branch dramatically changed due to the violent riverbed evolution.Even during dry water period,the diversion ratio of the South Branch is significantly greater than that of the North Branch.Under the conditions of dry or normal period,the tidal zone and tidal current have traced back to the Shuikou Dam.The low tide level significantly decreases;the tidal range increases;and the runoff variability range increases.(4)The water quality of the South Branch is better,but that of the North Branch is poor.The pollution belt has the characteristics of wandering in a tidal cycle process,and the pollution belt in the middle and lower reaches of North Branch is more prominent.The simulation results of water quality response to riverbed evolution show that the water quality of South Branch improves slightly,but that of North Branch deteriorated.Pollutants concentrations raise,and the distance that pollutants traced increased.(5)Genetic algorithm can be used to optimize the parameters of support vector machine efficiently,and the constructed GA-SVM coupling model can successfully predict the salinity of the Min River Estuary based on the high tide level,runoff and previous salinity.The proposed model performs significantly better than the traditional SVM model and BPNN model in terms of prediction accuracy and computing time.(6)An XML representation framework of model ontology based on the water environment domain knowledge was adopted to make unified and abstract expression of heterogeneous water environment models.The heterogeneous models were uniformly integrated into the system platform and an estuary water environment decision-making support system was developed successfully,which was applied to the water environment management of the Min River Estuary.It simplifies the multifarious professional operation and strengthens the simulation and prediction ability of the models,and provides decision support for the water management department of the Min River.