Modeling For Suspended Sediment Concentration Retrieval And Analyzing The Applicability In Bohai Bay Offshore Area

Offshore waters which has wide continental shelf, high biological productivity, abundant natural resources, and strongly influenced by people and terrigenous material, is a typical area of understanding the impact of suspended sediment on the ocean and environment. Bohai Bay as an important gulf of our country’s natural resources and social economy, it has very important research significance and application value on monitoring the suspended sediment, understanding the aquatic environment, exploiting and using the natural resources, reducing the problems of ecological environment and constructing the offshore. Ocean color remote sensing is an important technical means of monitoring the offshore waters environment, it can monitor ocean in real time, continuous and large scale, it is a effective means to monitoring Bohai Bay suspended sediment concentration(SSC).For the purpose of using the remote sensing in the study of SSC, based on analyzing the spectral characteristics of water, built the statistical regression model, principal component analysis model and the semi-analysis model to connect the RS data and the SSC data. Analyzed the characteristics of SSC in the study area, the precision of every model, and the applicability of model in area, time and concentration level. Main research contents in this paper are the following.(1) According to the measured spectral data from 2008 to 2011, analyzed the spectral characteristics of water in the study area. The spectral characteristics of study area is similar to the characteristics of case Ⅱwaters, and it has two obvious peaks. The first peak is in the region of 550~600nm, the second peak is in 780~830nm. With the increase of the SSC of water, the second refletion peak increased faster than the first one, and showed a phenomenon of “red shift”.(2) On the basis of analyzing the relationship between the spectral data and the SSC data, the statistical regression model, principal component analysis model and the semi-analysis model were built to inversing the SSC in the study area. Among them, the ratio of TM4/TM1 was selected as a sensitive factor to establish the quadratic polynomial model. The principal component analysis model was the index model which was built by the two selected principal components. Calculated the inherent optical properties by the quasi-analytical algorithm, analyzed the value of f(μ0) and Q, established the the semi-analysis model based on the medial inherent optical properties.(3) By analyzing the SSC distributed map of the study area, the results showed that the high concentration of SSC was concentrated in the south west offshore waters and the south offshore waters, the SSC of southern waters was higher than the northern waters, offshore waters was higher than the deep sea.(4) Used the Matlab software to calculate the R, IA, MAE, MRE and RMSE between the measured SSC and the inversed SSC of the statistical regression model, principal component analysis model and the semi-analysis model. Results showed that Rs and IAs were higher than 0.9, MAEs were less than 6mg/L, MREs were below 30%, RMSEs were different. The accuracy of semi-analysis model was slightly higher than the others, its R, IA, MAE, MRE and RMSE were 0.983, 0.988, 4.282mg/L, 27.85% and 5.967mg/L respectively.(5) Analyzed the applicability of area, time and concentration, and the results showed that, the statistical regression model was more suitable for the offshore areas of Huanghua Port and Tianjin Port, August and September, or less than 50mg/L of the SSC. The principal component analysis model was more suitable for the offshore area of Caofeidian, June and August, or more than 50mg/L of the SSC. The semi-analysis model was more suitable for the offshore area of Caofeidian, June and August, or all the SSC.