Impact of sampling strategy and frequency of river water quality on nitrogen and phosphorus load estimation
In order to explore suitable sampling frequency and strategy for water quality monitoring in the basin and improve the accuracy of nitrogen and phosphorus pollutant load estimation,nutrient loads were estimated based on daily data of daily discharge and total nitrogen and total phosphorus concentrations in Fengyu River Basin,Eryuan County,Yunnan Province from 2011 to 2013,under different scenarios.Specifically,three different sampling scenarios were set up(scenario a:general random sampling,scenario b:sampling water samples with more than 10mm of precipitation added on the basis of random sampling,and scenario c:sampling water samples with more than 25mm of precipitation added on the basis of random sampling)with sampling frequency ranging from weekly to nearly monthly.Based on the combination of different sampling frequencies and sampling scenarios,the accuracy of load estimations by LOADEST model is evaluated by comparing the total nitrogen and total phosphorus load estimated by LOADEST with the measured value calculated by daily continuous water quality and streamflow data.The results show that:adding the concentration data during the rainfall period will improve the simulation accuracy of LOADEST model under the same sampling frequency;sampling during rainfall events increases the frequency of sampling(such as scenario b),resulting in the RMSE of total nitrogen increasing from 1.92 to 2.26,and the RMSE value of total phosphorus increased from 0.08 to 0.19,which will affect the simulation accuracy of the model by overestimating the nitrogen and phosphorus load.Under four different sampling frequencies,the RMSE values of total nitrogen and total phosphorus in scenario c ranged from 1.92 to 2.07 and 0.06 to 0.13,obtained the lowest RMSE value and the highest accuracy of model estimation;different frequencies and sampling scenarios have greater impact on total nitrogen load estimation results(RMSE ranging from 1.92 to 2.52)than total phosphorus load estimation(RMSE ranging from 0.06 to 0.18).The accuracy of load estimation results for scenario a and scenario b decreases as the sampling frequency declines.The estimated results of scenario c are not significantly affected by the reduction of the existing sampling frequency.The results of this study show that the estimation accuracy of nitrogen and phosphorus load can be obtained at a low sampling frequency of 28d/time by adding samplings during rainfall events based on weather forecast,which is consistent with the sampling frequency of environmental water quality in various regions of China.This study could provide reference for the water quality sampling strategy and frequency.
万方数据
维普
