Study On Growth Model Of Water Quality In The River

With the rapid development of social economy, the region of Beijing, Tianjin and Bohai Bay turn on many environmental problems, such as water resources shortage, environmental quality worsening, rivers contaminated (more than 90%), water quality risk etc., which have been an important bottleneck of the two cities developmental stratagem. Therefore,"the spatial-time variation and transportation law of primary contaminates"was regard as a key task of the National Basic Research Program (973 Program) of China.In view of this, this paper takes the city of Tianjin for an example, selects a representative rive—the DG River as the subject. The water quality predication models, the water quality growth models and piecewise linear water quality growth models were established.On the base of related literatures, the water quality parameters were determined, which include conventional water quality index, biogenic elements index and heavy metal elements. We conducted a total of nine samples, respectively in Mar.11, Mar. 18, Apr. 4, May 5, May 25, Aug. 4, Aug. 26, Oct.25 and Nov. 7. Water quality parameters were measured by the methods of water quality standard monitoring. At the same time, we conducted the pollution sources survey of the DG River. Combined with the results of pollution sources survey, CODcr and NH₃-N were chosen as the primary pollution factors of the DG River by the analysis of water quality.Based on the theory of contaminants transport, this paper established the transport prediction models of primary contaminants (CODcr, NH3-N). Among of them, the ARIMA(1,0,2)of CODcr and simple exponential smoothing model of NH₃-N have the better predicated precision (more than 80%)which can provide scientific foundations for the water quality predication and management of the rivers.In an innovative way, this paper applied the multilevel models to the water quality analysis of the river. Water quality growth model of primary pollution factors about the DG River were established. Among of them, the random intercept-slope growth model with time varying covariates for CODcr and the intercept-slope growth model for NH₃-N can better reflect the longitudinal change rule of water quality during low flow period, average flow period and high flow period as well as the temporal spatial change rule of pollutants. In order to more accurately learn the water quality variation along the different water period, this paper introduced the piecewise linear water quality growth model based on the water quality growth model. Respectively established two, three, four piecewise water quality growth models of the primary pollution factors . Among of them, three piecewise linear water quality growth models of CODcr and four piecewise linear water quality growth models of NH₃-H have the best goodness-of-fit, which can comparatively and accurately reflect growth trend of water quality along different water period. The three piecewise water quality growth model of CODcr divides the growth trend of CODcr into three phases: low flow period (Mar., Apr.), average flow period (May) and high flow period (Aug.), average flow period (Oct., Nov.). The four piecewise water quality growth model of NH₃-H separates the growth trend of NH₃-H into four phases: low flow period (Mar., Apr.), average flow period (May), high flow period (Aug.) and average flow period (Oct., Nov.). These models accurately expresse the law of water quality variation along the different water period. The successful application of these models can provide scientific foundation for general regulation and scientific management of rivers.