Research Of Suspended Sediment Concentration And Size Distribution Impact On The Water Spectrum At The Maximum Turbid Zone Of Yangtze River Estuary

Suspended sediment concentration, spatial distribution and its transportation in the Yangtze River Estuary have become core indicators and important parameters in the researches of material fluxes, Material Cycle, changes of estuarine shoals, evaluation the impact of watershed projects on the estuary mouth and assessment of land resources in the Estuary. This has drawn the attention of academic circles, government departments and industry as well. Traditional single-point, multi-point measurement on board can only provide short-term partial data, lack of spatial distribution data and has a high cost, while remote sensing has an important advantage of covering large area and repeating observation in monitoring suspended sediment concentration.In this paper, following tasks were mainly carried out in the turbidity maximum area of Yangtze River Estuary:1) Field data acquisition and processing:Three in situ observations were performed at same site in the South Passage to collect spectral data, and surface water samples synchronously to measure sediment concentration and size distribution data in the laboratory. LISST-100x was used to measure floc particles as well..2) Analysis of the spectral reflectance characteristics above water surface and the relationship between suspended sediment concentration and the spectral reflectance. Giving theoretical explanation for the different sensitivity of the suspended sediment concentration at visible band and near infrared band. Using MODIS bands as a reference to establish several regression equations between spectral reflectance and suspended sediment concentration. Focus on the accuracy and reliability of statistical models, compares inversion equations with different band parameters and mathematical relationships of different forms at different times, and found by other researchers.3) Analysis the impact of suspended sediment particle size on the spectral reflectance characteristics and giving theoretical explanation and drawing the basic conclusion by analysis the measured data. The regression equations between spectral reflectance, the suspended sediment concentration and particle size was established, using of the mathematical relationships in literature. Further more, the impact of flocs on spectral characteristics was also considered and discussed.The primary conclusions obtained are as follows:1) Spectral reflectance at Near-infrared band has a good relation with suspended sediment concentration, which can be used for suspended sediment concentration inversion. Two-band ratio can effectively reduce the influence of other conditions (such as illumination, size, etc.) compare to parameters such as single band, and multi-band combination. For the model mathematical formulas, the polynomial model fit the data best, but without any excellence comparing to the exponential model and power model. 2) Through comparison of suspended sediment concentration inversion models founded by data from different periods and in the existing literature, it is found that models founded by different researchers are comparable, especially in a certain range of low suspended sediment concentration value indicating the statistical model is reliable. Strengthen the comparative study on the statistical models is an effective way to improve cognition of the reliability of the application.3) There is some correlation between sediment grain size and spectrum, but it cannot certify the impact of particle size on the spectrum, because sediment particle size has a good correlation with suspended sediment concentration which also has a high correlation with the spectrum. Theoretical analysis shows that sediment particle size distribution mainly through the effect of scattering to impact on the spectral reflectance; fine sediments have a stronger backscatter than coarse sediment.4) Field data showed that fine particle size would increase the spectral reflectance which was consistent with the theoretical analysis. Comparison of regression models founded using data from two different periods, it was found that the smaller the particle size distribution changes, the higher value the regression model R2 got.5) By establishing regression models regression equations between spectral reflectance, the suspended sediment concentration and particle size, it showed that it could not help to improve model accuracy when particle size was included as an parameter, but lead to a decline in accuracy.6) Grain size distribution of flocs particles had an important impact on water spectrum but it was difficult to express the effect quantitatively, when analysis by using suspended sediment mean diameter, median diameter and flocs mean diameter. When establish regression models regression equations between spectral reflectance, the suspended sediment concentration and flocs mean diameter, it also led to a decline in accuracy.