Spatio-temporal Variational Analysis And Prediction Of Sea Water Quality In Shanghai Sea Area

A nearest neighbor interpolation pretreatment method was developed in high-dimension space in order to analyse the spatio-temporal variation of sea water quality in Shanghai sea area, based on the data of Environmental Monitoring Center of East China Sea Branch, State Oceanic Administration. The trends of sea water environmental factors were analyzed with the time series analysis methods based on traditional statistics and the time delay neural network based on artificial neural network theory. Results show that: (1) in diurnal period, physical factors and nutritional chemical factors like nitrogen and phosphorus, varied significantly; other chemical factors like heavy metals'variations are not obvious. Variations on surface and at bottom are different for some factors; (2) in seasonal period, factors varied significantly; variations are also different on surface and at bottom; (3) on spatial distribution, active phosphate, ammonia nitrogen, nitrite nitrogen and total Hg are diffused from sewage discharge outlets and have high concentrations near them; pH increased from inside to outside; chemical oxygen demand, nitrate nitrogen and bottom dissolved oxygen decreased from inside to outside of estuary; surface dissolved oxygen and other factors'distributions are not clear; (4) from 1985 to 2006, chemical oxygen demand and pH decreased; active phosphate, ammonia nitrogen, nitrite nitrogen and nitrate nitrogen increased, especially for active phosphate, the increase is very obvious; other factors have no prominent increase or decrease trends.