Statistical Characteristics And Analysis Of Surface Water Quality In Yiluo River

The ecological system of the Yellow River is fragile,the economic and social development of the upper and middle reaches and the downstream beach areas are relatively lagging,and water pollution has not been completely resolved.In recent years,the state has taken the ecological conservation and high-quality development of the Yellow River Basin as a major national strategy,effectively coordinating the relationship between ecological protection and economic development in the Yellow River Basin.Based on this,In order to study the trend of water quality and future changes in surface water of Yi Luo River.In this paper,Seasonal Kendall test,Pollutant Transport Rate test and Partial Mann-Kendall test were used to analyze the water quality data of chemical oxygen demand and ammonia-nitrogen which measured in the nine years from 2009 to 2017 at qilipu of Gongyi and the confluence of Yiluo in the Yi Luo River.The results show that: the surface water quality data of the Yi Luo River is seasonal;Comparing the results of Pollutant Transport Rate test with Seasonal Kendall test,it can be concluded that the water pollution of these two sections belongs to point source pollution;In the Partial Mann-Kendall test,excluding the effect of flow fluctuation,the chemical oxygen demand and ammonia-nitrogen’s concentrations of the two sections show a decreasing trend,which means that the water quality of the two sections is getting better and better,the trend of residual term of EMD model also proves the result.In the meantime,The Kolmogorov-Smirnov test,Anderson-Darling test and Chi-Squared test show that the COD and ammonia-nitrogen of the monitored section are in Normality test.Therefore,the two monitoring sections are compared using Auto-regressive Integrated Moving Average(ARIMA),Exponential Smoothing State Space model(ETS),Cubic Exponential Smoothing(Holt-Winters),Support Vector Machine(SVM)and a aggregators of Empirical Mode Decomposition and nonlinear auto-regression with exogenous input neural network(EMD+NARX),determining the optimal model in this paper.The results show that: By comparing the root mean square error,mean absolute error,etc.of five prediction models,it is concluded that the time series of chemical oxygen demand,ammonia-nitrogen at qilipu of Gongyi and chemical oxygen demand at the confluence of Yi Luo are bested predicted by EMD+NARX neural network model.the ARIMA model is better for predicting ammonia-nitrogen at the confluence of Yiluo,which is probably due to the better smoothness of the ammonia nitrogen’s fluctuation;In general,the combined model of EMD+NARX neural network can well predict the change of water quality in a period of future,and the use of the combined model can well make up for the shortcomings of the single model.This study enriched the related research on the trend and prediction model of surface water quality in Yiluo River,The feasibility of the combined model of EMD+NARX neural network in the prediction of the surface water of the Yiluo River has been verified.Meanwhile,the water quality management of the Yiluo River reaches has been strengthened.