Water Quality Assessment Of Tuojiang River Based On Nitrogen And Phosphorus Load Characteristics From Agricultural Non-point Sources

Under the development concept of "lucid waters and lush mountains are invaluable assets",comprehensive assessment of water quality and research on water pollution characteristics are important prerequisites for water pollution control.Tuojiang River,located in central Sichuan Province,is a major tributary of the upper reaches of the Yangtze River.As an important natural resource of the Sichuan Basin,the water quality of the Tuojiang River has deteriorated to varying degrees due to the influence of industrial and agricultural activities.To elucidate the characteristics of water pollution in typical industrial and agricultural interleaved,long-distance and transboundary river basins,31 sampling sites containing 12 water quality parameters were selected in the Tuojiang River in this work.Based on the water quality data of 2018-2019 and statistical yearbook data of each district and county of the Tuojiang River Basin,principal component analysis,correlation analysis,pollution discharging coefficient method and cluster analysis were used.Based on APCS-MLR model,PMF model and BP artificial neural network model,the present situation of water quality and its spatiotemporal variation characteristics and pollution characteristics of Tuojiang River were studied.The results of this study can provide theoretical basis and technical support for the development of water resources,water environment control and planning of the Tuojiang River basin.Additonally,it can provide scientific basis for the study of water environment assessment and pollution characteristics of typical industrial and agricultural interleaved,long-distance and transboundary basins..The main conclusions obtained are as follows:1)The main pollution sources and their quantitative contributions to pollutants of the Tuojiang River were determined.Bsed on the optimized APCS-MLR and PMF receptor models,four pollution sources were identified of the Tuojiang River,which were agricultural non-point sources,industrial wastewater,domestic sewage and soil weathering.The results of APCS-MLR showed that most of the water quality parameters were mainly affected by agricultural non-point sources and industrial wastewater,with the mean values of 39.90% and 18.18%,respectively.The PMF model showed that about 89.78% of TN comes from agricultural non-point sources,while about 82.00% of TP comes from industrial sources.Overall,both PMF and APCS-MLR models indicate that agricultural non-point sources are the most important sources of pollution of the Tuojiang River.2)The characteristics of nitrogen and phosphorus pollution load of agricultural non-point sources in the Tuojiang River Basin were obtained.In the Tuojiang River Basin,TN load of planting industry decreased from 38000 tons in 2018 to 36000 tons in 2019,while TP load increased from 15000 tons in 2018 to 18000 tons in 2019.The grey water footprint and WPL of planting industry indicate that planting industry exerts great pressure on the water environment of the basin,and the absorption capacity of surface fresh water to nitrogen and phosphorus pollutants in each region is generally saturated.The nitrogen and phosphorus loads of the livestock industry in the Tuojiang River Basin decreased from 11412 tons and 2798.1 tons in 2018 to 9860.4 tons and2,417.7 tons in 2019,respectively.The analysis of the composition of nitrogen and phosphorus load in livestock industry showed that pig and poultry farming were the main sources of nitrogen and phosphorus pollution in the Tuojiang River basin.The nitrogen and phosphorus loads of domestic sources decreased from 11712.67 and1774.8 tons in 2018 to 11217.60 and 1699.78 tons in 2019,respectively.Planting industry was the main contributor to the nitrogen and phosphorus pollution load of agricultural non-point sources in the Tuojiang River Basin,and its contribution to the nitrogen and phosphorus pollution load was 61.43% and 77.39%,respectively.The average contribution of living sources to TN and TP loads was only 20.23% and 9.15%.The average contribution of livestock industry to TN and TP load was 18.34% and13.46%.3)The temporal and spatial variation of water quality of Tuojiang River was clarified.The concentration of water quality parameters were generally higher in the middle reaches of the Tuojiang River,and the seasonal variation analysis shows that the concentration of water quality parameters in summer is significantly higher than that in the other three seasons.The water quality of the Tuojiang River was comprehensively evaluated based on WQI and fuzzy comprehensive assessment method.The results showed that the water quality of the Tuojiang River was generally between "good" and "poor",and 93.50% of the sampling sites were classified as "medium" water quality,indicating that the water body of the Tuojiang River was slightly damaged.In general,the water quality of the Tuojiang River showed significant temporal and spatial difference,with poorer water quality in the middle reaches of the Tuojiang River and in summer due to intensive agricultural activities.4)The construction principles,methods and steps of the MIWQI water quality evaluation model are proposed,which has been successfully applied to Tuojiang River water quality assessment.The results showed that MIWQI model divided the water quality of Tuojiang River into "medium" to "very poor" grade.Additionally,MIWQI model and WQI method have the same water quality evaluation results in 71.00%sampling sites.It shows that the MIWQI model proposed in this work can be used to evaluate Tuojiang River water quality.5)The response relationship between nitrogen and phosphorus load from agricultural non-point sources and water quality of the Tuojiang River was obtained.The results showed that the temporal and spatial variation patterns of TN and TP concentrations and water quality of the Tuojiang River were significantly correlated with the nitrogen and phosphorus loads from agricultural non-point sources of the Tuojiang River Basin(p < 0.05).The nitrogen and phosphorus loads from agricultural non-point source in the Tuojiang River affected the water quality of the Tuojiang River to a certain extent.6)The BP artificial neural network model for predicting TN concentration of the Tuojiang River was established.The input parameters of BP ANN the model were determined by correlation analysis.The training function and the number of hidden layer nodes are optimized by trial and error method.According to correlation coefficient,mean relative error and root mean square error,the optimal training function is trainlm,and the optimal number of hidden layer nodes is 4 and 9,respectively.According to the result,the optimized BP artificial neural network can predict the TN concentration of the Tuojiang River with satisfactory results.In the first case,the mean relative error of the BP-ANN model is 1.28%,the optimal calibration error is 0.013,and the total correlation coefficient reaches 0.991.In the second case,the mean relative error of the BP-ANN model is 8.40%,the optimal calibration error is0.200,and the global correlation coefficient is 0.933.