LOADEST 모델을 이용한 영산강수계의 수질 경향성 분석

Abstract

In this study, water quality trends are analyzed using the measurement data of the Yeongsan River water quality forecasting point(Gwangjucheon2, Pungyeongjeongcheon, Hwangryonggang3-1, Pyeongdongcheon, Jiseokcheon4)which is a part of the water quality measurement network The water quality trend analysis of the target site was analyzed by applying the flow rate, BOD, and T-P among the measured data to the LOADEST model, and applying condition 1 (regression expression No. 9) and condition 2 (automatic selection No. 0). The difference in the actual load was analyzed by simulating the load through the regression equation of the LOADEST model, and the suitability of the regression equation of the LOADEST model for each point was analyzed. By analyzing the parameters according to the regression formula according to the simulation of the LOADEST model, the trends according to flow rate, season, and time were analyzed, and the significance was tested using a statistical method (P-value). The simulation results according to the conditions (1,2) of the LOADEST model showed that there was little difference between the actual load and the load. The maximum values ​​(outliers) at some points (Gwangjucheon2 Hwangryonggang3-1, Pyeongdongcheon, Jiseokcheon4)were calculated by underestimating the results of the LOADEST simulation. NSE, PBIAS, and RSR were analyzed to evaluate the suitability of the load simulated by the LOADEST model and the actual load. As a result of NSE analysis, except for the load regression analysis on the T-P of the Pyeongdongcheon, the remaining points were “satisfactory” or higher, indicating that the simulation results reflected the actual measurements well. As a result of BOD and T-P regression analysis, PBIAS was found to be “satisfactory” or higher. RSR was evaluated as high suitability as it showed a grade of “good” or higher at most points as a standard deviation ratio. The significance (P-value) of the parameters of the regression equation according to the conditions (1, 2) of the regression equation for each point by the LOADEST model was mostly analyzed as highly correlated and very significant. As a result of the parameter analysis, it was analyzed that the parameters (a0 ~ a2) indicating the trend of the flow rate had a large parameter value in common at the target point and were mostly significant. The parameters (a0 ~ a2) showed that the values ​​of parameters and P-values ​​increased at some points (Pungyeongjeongcheon, Hwangryonggang3-1, Pyeongdongcheon) under condition 2. It was analyzed that the regression parameters of the LOADEST model had a high correlation with the flow rate, and that the correlation between the seasonal cycle and the year-round change cycle was different depending on the location. As a result of analyzing the parameters (a3, a4) reflecting the change cycle according to the seasons as conditions (1, 2), most of the target points were negative (-) in the Sin cycle, and the significance level was usually analyzed, so the Sin cycle showed a tendency to change in water quality with a cycle opposite to that of Cos cycle was analyzed as positive (+), but mostly insignificant results were shown. As a result of the analysis according to condition 1 for the parameters (a5, a6) related to year-round time, the values ​​of the parameters at the target point except for Pyeongdongcheon were low, but most showed significant results. In the analysis according to condition 2, Pungyeongjeongcheon, Hwangryonggang3-1, and Pyeongdongcheon were calculated and analyzed by regression equations that do not include parameters (a5, a6), indicating that there is little correlation with time. As a result of the simulation of the LOADEST model, it was analyzed that the simulated value was underestimated at the specific flow rate (high flow rate) at the target point, so it was necessary to review and correct the measurement data. In this study, by using long-term measurement data to analyze the load and water quality trends of water quality factors, it is expected to be used as data for policies for using basic data and for water quality management.