Study On Remote Sensing Algorithm And The Temporal-spatial Distribution Of Turbidity In The East China Seas

Water turbidity (T) represents an optical determination of water clarity, which can be considered as a proxy for water quality. Investigating the turbidity is the key to understanding marine primary productivity, ecology, biochemistry process,hydrodynamic environment and material transport.One of the most significant problems of retrieving turbidity in the East China seas is the uncertainty of atmospheric corrections. Because of the complex optical environment in the East China seas, especially in the coastal areas, some operational algorithms are invalid causing unnecessary errors and a large amount of incorrect data masking.Note that the Rayleigh scattering correction can be accurately performed, the essential problem of atmospheric correction is to solve the estimation of the reflectance contributed by aerosols. Hence this research generated a turbidity remote sensing algorithm using Rayleigh-corrected reflectance (Rrc) data from GOCI instead of the conventional remote sensing reflectance (Rrc), which can avoid accurate atmospheric correction. We compared the Rrc algorithm with two other existing algorithms. The influence of atmospheric algorithms was also evaluated by comparing the results from using Rrc data and Rrs data (by UVAC). Differences among the results derived from Geostationary Ocean Color Imager (GOCI) and Moderate Resolution Imaging Spectroradiometer (MODIS) imagery data were compared, respectively. Then, a combined remote sensing algorithm was established to estimate turbidity from GOCI observations over the Eastern China Seas. In the end,the characteristics of the spatial-temporal distribution of turbidity in the Eastern China Seas during 2014-2015 were shown in this research. The main research results are list as follows:(1) The Rrc-based model was generated through a quadratic regression analysis between the band combination (Rrc(490) + Rrc(680)) / (Rrc(490) / Rrc(680)) and log10(T). This algorithm can estimate turbidity in highly turbid water without atmospheric correction. The R2 is larger than 80%. And the model is robust.(2) Through the comparative analysis, the performance of the Rrc-based model was better than two other models, which were empirical and semi-empirical algorithm, respectively. The Rrc-based model is suitable for the coastal waters of China, especially for highly turbid waters. However, the performance needs to be improved for clear waters. The comparison results also showed that the uncertainty of atmospheric correction would bring more error for results; and that the Rrc-based algorithm developed based on GOCI data could be extended to other sensors, for instance MODIS, by careful tuning of the relevant parameters.(3) Spatial-temporal distribution analysis was conducted by using a blended algorithm. The result showed that the turbidity in the East China seas is high nearshore, and low offshore; in the spring and autumn season, the turbidity values are unstable, while in the winter and summer season, the turbidity reached a maximum and a minimum, respectively, showing stable phases; In the study area, high turbidity values were mostly located in the Bohai Sea coast, Subei shoal, Yangtze river estuary and the west side of the Korean Peninsula. Zhejiang nearshore has a narrow high-turbidity zone. And the pattern of turbidity dues to comprehensive effects of many factors such as tide, monsoon, spring layers, some short-term weather phenomena, terrigenous input, and circulation.