An Improved Inherent Optical Properties Data Processing System for Residual Error Correction in Turbid Natural Waters

Being able to accurately estimate inherent optical properties (IOPs) at long time scales is key to comprehending the aquatic biological and biogeochemical responses to long-term global climate change. We employed the near-infrared band and combined it with four “common bands” at visible wavelengths (around 443, 490, 551, and 670 nm) to adjust the IOPs data processing system, IDAS
_v2
. We applied the IDAS
_v2
algorithm further to correct for the residual error in images of turbid waters. We evaluated the performance of the IDAS
_v2
algorithm using datasets covering a wide range of natural water types from clear open ocean to turbid coastal and inland waters. Due to the water-leaving signals’ sensitivity to the optically significant constituents of highly turbid waters, the near-infrared band was very important for retrieving IOPs from those waters. In our analysis, we found that the IDAS
_v2
algorithm provided IOPs data with <28.36% uncertainty for oceanic waters and <37.83% uncertainty for inland waters, which was much more effective than what a quasi-analytical algorithm provided. Moreover, the near-infrared band was better at removing the residual error and partial intermission bias in satellite remote sensing reflectance (
R
_rs
) data because of the strong absorption of pure water. We tested the IDAS
_v2
algorithm with numerically simulated and satellite observed data of turbid water. After applying IDAS
_v2
, the IOPs data were accurately determined from
R
_rs
data contaminated by the residual error. Furthermore, the mean intermission difference between Medium Resolution Spectral Imager 2 and Visible Infrared Imaging Radiometer
R
_rs
data at 443 and 551 nm decreased from 8%–25% to 1%–9%. These results suggest that we can accurately estimate IOPs data for natural waters including naturally clear and turbid waters.