Research On Remote Sensing Based Water Quality Monitoring Models In Huangpu River Upstream Waters

Water quality monitoring is the primary basis of water quality evaluation and water pollution prevention. As a necessary complement to conventional water quality monitoring methods, remote sensing based water monitoring has the advantages of large scale, speediness, cost-effective and capable of making long-term dynamic water quality monitoring. Nowadays, water pollution of Huangpu River is regarded as one of the most significant environmental problems in Shanghai. It has importmant research and application value for exploiture and environmetal protection in upstream Huangpu River through establishing water quality dynamic monitoring method based on remote sensing technology.Historical data of water quality in the up stream area of Huangpu River were comprehensively analyzed and evaluated to obtain the water quality trend after the filtration of the water quality parameters. The measurement methods of spectral reflection above the surface of water and related data processing methods were studied, after which water spectrum and the water quality paramentrs were measured in the lab and on the site. Based on the analysis of these data, water quality models for parameters such as DO, Turbidity, and Chl-a were established. After the study of characteristics and processing methods of the satellite remote sensing data, TM data were analyzed in emphasis. The correlation coefficient between the water quality parameters and the TM data were obtained by using regression methods. DO, COD, and TP's estimation models were established based on the correlation coefficient. The universality of these models was verified.Following results were obtained: the spectral transmittance ranged from 550nm to 840nm was suitable for estimating turbidity with high precision;the reflectance ratio of 701 nm to 675nm was fit for estimating the concentrations of chlorophyll-a;the TM4/TM2 ratio had the highest correlation coefficient with concentration of DO, whereas the TM5/TM1 ratio was better for COD.Some of the water quality models with their precision established in this paperwere:Turbidity = 354.58xT700+338.19, r=0.95;Chl-a=3.4724xR701/R675-1.5378,r=0.82;DO=6.5088xexp(-0.2322*TM4/TM2),r=0.83;COD=17.007xTM5/TMl+13.393, r=0.84;The universality of these models except Chl-a were verified and the error between the assessed value and the measured value were acceptable: the average deviation of the Turbidity, DO and COD were less than 6%, 15% and 17%, which indicated these models could meet the needs of remote sensing based water quality monitoring.