Delineating bio-optical characteristics to enhance spatial-based quantification of CDOM in clear to turbid waters

Abstract

Accurate remote sensing estimates of inland water colored dissolved organic matter (CDOM) are dramatically challenged by the dynamics of bio-optical characteristics, in which water biochemical driving mechanisms are particularly complex and highly variable. Actually, many of the empirical relationships in the inland water CDOM algorithms are only applicable to specific waters, and they are extremely susceptible to changes in phytoplankton and sediment. This study obtains a classification of bio-optical properties based on spectral shape and constructs a methodology for optically heterogeneous datasets to dynamically monitor CDOM in water. The reliability of the classification algorithm is validated using measured dataset and Global Reflectance community dataset for Imaging and optical sensing of Aquatic environments (GLORIA dataset) (ag(443) varying from 0.001 m−1 to 25.14 m−1). The results show that retrieved CDOM absorption has an improved performance with a mean absolute percentage error (MAPE) of 38.69 % for measured dataset and 43.29 % for GLORIA dataset, respectively. Drivers of CDOM in inland waters can be illustrated on the basis of bio-optical types, CDOM levels, and CDOM source characteristics produced by satellite sensors.

Keywords

• Remote sensing
• Colored dissolved organic matter
• Optical dominant component classification
• Inland water