The Research On Comprehensive Evaluation Of Surface Water Quality And Simulation Of Non-point Source Pollution Within Tiaoxi Watershed

Despite the government has made major advances in the control of industrial and domestic point sources of water pollution, water pollution is still a critical problem in China as nonpoint source (NPS) pollution from runoff of fertilizer, pesticides, and discharges from intensive animal production facilities is becoming a major source of water pollution. Development and implementation of a viable and scientific action plan is required to effectively control nonpoint source pollution. Surface water quality assessment and quantitative evaluation of nonpoint source loads can provide an important support for decision-making. The goal of this study is to assess the water quality of surface water and to quantify the pollution loadings coming from nonpoint source. Tiaoxi watershed within Taihu was selected as the study area, and the main research work was made up of:(a) comprehensive evaluation for the surface water quality using self-organizing map (SOM) and Hasse Diagram Technique (HDT),(b) establishment and optimization of the classification models based on support vector machine (SVM), and (c) establishment of HSPF model for quantitative simulation and evaluation of nonpoint source pollution loads for Tiaoxi watershed. The main conclusions of the study were as follows:(1) SOM and HDT were used to classify, model, interpret and evaluate the monitoring data for surface water quality, which displayed the water quality characteristics within Western Tiaoxi watershed. The water quality parameters like permanganate index, ammonia nitrogen (NH3-N), total nitrogen (TN) and total phosphorus (TP) were selected for HDT comprehensive analysis and the results showed that water quality was better in the upstream, but relatively worse in the midstream and downstream. Sensitivity analysis showed that CODMn, TN and TP were important factors with high environmental risk.(2) Support vector machine (SVM) classification models were constructed using a radial basis function (RBF). These models were used for classification according to dissolved oxygen (DO), permanganate index, NH3-N, TN, or TP, and cross-validation and grid-search were applied to find satisfactory parameters for RBF for the improved models. Then the improved models were made used of assess water quality within Tiaoxi watershed. The results showed that the SVM classification (SVC) models performed relatively better than the Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) models for classification.(3) The spatial and attribute database for Tiaoxi watershed were established by using GIS and BASINS, and HSPF model was built. In the process, HSPF model was calibrated and validated for hydrology, sediment and nutrient respectively. Relative error (RE) and Nash-Suttcliffe efficiency coefficient (Ens) were used to assess the model. The results showed that the hydrology parameters were suitable; the model had a good capability for the flow, but only general capability for sediment and nutrient like nitrogen and phosphorus. In general, the simulation was acceptable. The accuracy could basically meet the requirements, and the model can be used for nonpoint source pollution within Tiaoxi watershed.(4) By using of HSPF model, quantitative evaluation of the loads of TN and TP from nonpoint source pollution in Tiaoxi watershed in2007was carried out, and the values were compared with baseline surveys afterwards. It showed that the simulated TN load was10698t, and the surveys was9003t, where the relative error was18.83%; the simulated TP load was1695t, and the surveys was1499t, with a relative error of13.8%. Nonpoint source pollution accounted for51.5%of all nitrogen pollution and42.7%of all phosphorus, and was a major pollution source in Tiaoxi watershed.(5) The HDT sensitivity analysis for the agricultural nonpoint sources in Tiaoxi watershed indicated that the main factors of nonpoint source pollution loads were the pollution from farming and livestock poultry industry and aquaculture. According to HDT analysis based on the output of TN and TP in sub-basin in Tiaoxi watershed using HSPF model, combined with the cultivated soil loads for TN and TP from nonpoint sources, the key control area were determined. Scenario analysis based on the fertilizer application showed that reducing the amount of chemical fertilizer could reduce the output of nonpoint source, but not significantly. Scenario analysis based on controlling the nonpoint source pollution from livestock pollution showed that TN and TP reduction rate reached13.56%and18.73%when reducing the livestock emission by50%. Finally, some suggestions were given for controlling the nonpoint source pollution within Tiaoxi watershed.