Functional Linear Analysis of in situ Hyperspectral Data for Assessing CDOM in Rivers

Abstract

Turbidity and chlorophyll introduce high uncertainty in remote sensing of Chromophoric Dissolved Organic Matter (CDOM) in riverine and coastal water. To reduce the uncertainty, we developed a functional linear model (FLM) to analyze spectral responses to CDOM concentrations observed in a cruise along two rivers and a tidal bay. The analysis was supported with the measurement of high spatial resolution underwater CDOM concentrations and concurrent in situ above-surface hyperspectral remote sensing reflectance. The functional linear model is able to explain up to 91 percent of CDOM observations (H 2 = 0.91, RMSE = 0.0206). The dummy variables of local environmental factors included in the estimation improve CDOM assessment in coastal water. Our analysis suggests that the pattern changes of the FLM coefficient curves provide useful information for understanding the spectral signal interference from turbidity and chlorophyll. This work presents a feasibility study of in situ remote sensing of CDOM on a shipboard platform.