Iterative near-infrared atmospheric correction scheme for global coastal waters

• DOI: 10.1016/j.isprsjprs.2021.07.005
• Publication Date: 2021-08-03T21:54:15Z
• Authors: Cheng Xue; Shuguo Chen; Zhongping Lee; Lianbo Hu; Xinhao Shi; Mingsen Lin; Jianqiang Liu; Chaofei Ma; Qingjun Song; Tinglu Zhang
• Keywords: SeaWiFS; Atmospheric correction; Ocean color; Imaging spectrometer

Abstract Atmospheric correction (AC) is a key step in ocean color remote sensing for obtaining oceanic and coastal water ecosystem products. However, traditional NIR-based AC scheme ( AC NIR ) often fails in turbid coastal waters dominated by high suspended particulate matter (SPM). SWIR-based AC is valid for turbid waters, but it cannot be applied to some satellite sensors (e.g., Sea-viewing Wide Field-of-view Sensor (SeaWiFS), MEdium Resolution Imaging Spectrometer (MERIS), and future small satellite sensors) that have no SWIR bands. To obtain reliable data of turbid coastal waters by these sensors without SWIR bands, an improved AC NIR ( AC NIR-PRO ) scheme is proposed herein based on the frame of traditional AC NIR ( AC NIR-STA ). The synthesized and in situ datasets are analyzed to extend bio-optical models of absorption and backscattering coefficients in AC NIR-STA to highly turbid waters. AC NIR-PRO utilizes concentrations of both chlorophyll-a ([Chla]) and SPM ([SPM]) to estimate water-leaving contributions at the NIR bands. Similar to NIR-SWIR AC, AC NIR-PRO can obtain the information of global coastal waters with remote sensing reflectance at 862 nm ( R r s ( 862 ) ) up to 0 . 025 ∼ 0 . 03 sr − 1 . R r s of AC NIR-PRO is also validated by AERONET-OC datasets (mean absolute percent difference (MAPD) of R r s ( 551 ) is approximately 16%) for various coastal waters including high colored dissolved organic matter. Further, the magnitude and spatial patterns of [SPM] and [Chla] derived from the R r s products after performing AC NIR-PRO are significantly better than those from AC NIR-STA , especially for turbid coastal waters.