Remote Sensing Retrieving Turbidity And Its Application In Xiamen Bay

Turbidity is one of the important parameters of water quality evaluation.With the help of satellite remote sensing,sea surface turbidity data within a large area can be obtained continuously and quasi-synchronously.Among them,Landsat/Sentinel and other high-resolution satellites provide important data sources for analyzing and recognizing the turbidity variations within small-scale sea areas such as estuaries and harbors.Focusing on the Xiamen Bay and based on the satellite data from Landsat-8 and Sentinel-2,as well as turbidity data from buoys and the reflectance spectrum from in-situ measurements,the present study constructs a quantitative remote sensing inversion model of turbidity for the Xiamen Bay,after comparing and evaluating the applicability of two different atmospheric correction algorithms.In addition,preliminary temporal and spatial distribution characteristics of turbidity in the Xiamen Bay are provided.The results show that:Firstly,the consistency between the remote sensing reflectance spectrum from the in-situ measurements and that after the process of the atmospheric correction with EXP and DSF method is evaluated.The results show that the atmospheric corrected reflectance by DSF is closer to the in-situ measurements than that of EXP,with smaller verification error by DSF than that by EXP.In addition,the correlation analysis between the atmospheric corrected remote sensing reflectance of Landsat-8/Sentinel-2 by the two algorithms within the red band and the field-measured turbidity by buoys shows the better performance of DSF,indicating that DSF is more reasonable for Landsat-8/Sentinel-2 atmospheric correction process in the Xiamen Bay.Secondly,DSF-corrected remote sensing reflectance spectrum of Sentienl-2 and Landsat-8 shows relatively high consistency,and the linear fitting of Rrs from the two satellite images within the red band on the same day is very close to the 1:1 line.Furthermore,the connections from 563nm to 655nm bands of Landsat-8 and that from 561nm to 665nm Sentinel-2 almost overlap,presenting a very high consistency.Therefore,a matching dataset including the Rrs(665)transformed from the Rrs(655)of Sentinel-2 and Landsat-8 by linear relationship,as well as the turbidity data measured by buyos,is constructed.Referring to the Rrs(655)semi-analytical turbidity inversion algorithm constructed by Nechad and other scholars,the turbidity inversion model of Xiamen Bay is obtained from the dataset by regression fitting in the present study.Verifications show that the inversion model performs better than the empirical algorithm of single or multi band ratio with higher correlation coefficient.In addition,the algorithm can be applied to the landsat-8 and sentinel-2 turbidity inversion processes since the Rrs(655)used is from both satellites.As a result,turbidity in the Xiamen Bay can be monitored with higher sampling frequency and more complete time series with the use of Landsat-8 and sentinel-2A/2B satellites at present.Spatial distribution characteristics of turbidity within Xiamen Bay and its surrounding sea area are analyzed preliminarily using the turbidity time series data retrieved from Landsat-8 and Sentinel-2.The results show that areas with high turbidity are mainly distributed in the near shore areas,e.g.,Jiulong River Estuary,Dadeng Island,and Weitou Bay.Among them,the averaged turbidity of Jiulong River estuary is the highest,followed by Weitou Bay and Dadeng sea area,and the lowest one is near Tong'an Bay.In recent years,the turbidity around Dadeng Island is significantly affected by coastal construction.The reclamation activities around Dadeng Island and Nan'an Shijing lead to the loss of a large number of sand and soil into the sea,inducing 1.6-2.2 times higher averaged turbidity as well as a much larger affected area of high turbidity(>40NTU)during the construction period is than that during the period of no construction.