Remote sensing of suspended sediment in high turbid estuary from sentinel-3A/OLCI: A case study of Hangzhou Bay

Abstract

The suspended sediment in water infers water quality, and directly reflects optical properties such as water transparency, turbidity, and water color. Thus, these physical properties provide a viable basis to rigorously retrieve for suspended sediment concentration (SSC) using satellite remote sensing water color measurements in estuaries. The contemporary Ocean and Land Color Instrument (OLCI) on Sentinel-3A, provides more waveband options for remote sensing of water color and an opportunity for retrieval of suspended sediment in estuarine coast. Yet, accurate retrieval of SSC in high turbid waters from OLCI is still challenging due primarily to the high uncertainty of atmospheric correction. Here, we use OLCI images to measure water quality in Hangzhou Bay, and construct a retrieval model of SSC, and cross-validated using Geostationary Ocean Color Imager (GOCI) data. The study shows that: (1) the atmospheric correction algorithm based on ultraviolet wavelengths (UV-AC) can achieve better results for both OLCI and GOCI data, and the overall correction accuracy for OLCI is higher than that for GOCI data; (2) the multi-band index model constructed by using Rrs(Oa16)/Rrs(Oa5) of OLCI data has higher retrieval accuracy and model stability, with R2 is 0.96, MRE is 17.52%, and RMSE is 69.10 mg/L; (3) the spatial distribution of SSC in the study area is complex, mainly showing that the SCC in the top of the bay is larger than the mouth of the bay, and the south shore is larger than the north shore; (4) whe distribution of SSC obtained from retrieving OLCI and GOCI data in general is consistent, with the OLCI SSC estimates with higher accuracy than GOCI data, and the numerical difference between the two retrieval results is more obvious in the ocean with high SSC; and (5) with appropriate atmospheric corrections and retrieval models, OLCI data can be used to estimate improved SSC observables in Hangzhou Bay. We conclude that the SSC retrieval models proposed here provide a good reference method for retrieval of water color observable in Hangzhou Bay coastal estuary.