APPLICATION OF VIS-NIR SPECTRAL RADIOMETRY TECHNIQUES TO ESTIMATE CHLOROPHYLL A IN TURBID WATERS

The complex composition and distribution of CPAs (color producing agents) in turbid waters such as the Western Basin of Lake Erie (WBLE) present a challenge to the application of remote sensing data for differentiating among various in-water constituents and estimating their concentrations independently. In this paper, multivariate procedures and VIS-NIR (visible - near infrared) spectral analysis techniques are applied to a lab based spectrophotometer and MERIS (Moderate Resolution Imaging Spectrometer) data to estimate the concentration of chlorophyll a in the WBLE. Principle component analysis (PCA) of the first-derivative hyper-spectral data extracted three major spectral components. Matching using reference spectra indicated that the extracted spectral patterns represent signatures of in-water constituents that govern the optical properties of the WBLE namely, cyanobacteria, diatoms associated with phycocyanin from the blue-green algae and inorganic clays. The Labspec data clearly revealed known spectral features of phytoplankton, such as absorption minima near 560 and 700 nm which can be attributed to the minimum absorption capacity and fluorescence effects, respectively. The signal also extracted the red absorption peak due to chlorophyll a near 675 nm. Spectra recorded by the MERIS sensor were able to detect the in-vivo absorption and fluorescence peaks; however, percentage reflectance is much lower in the case of MERIS data due to absorption by water, dissolved organic matter and atmospheric interference. Principle component (PC) regression of chlorophyll a on the PC scores indicated that 68.5 % of variation of chlorophyll a in the WBLE can be explained by PC score 1. We applied semi-analytical band algorithms and an empirical absorption line height approach to the spectra obtained from both platforms. The techniques have proven to be efficient in isolating detectable patterns that represent chlorophyll a. The procedures have improved the down-welling and up-welling irradiance relationship increasing the sensitivity of the bands to selected CPAs. The minimum RSQ between the various models applied and in-situ chlorophyll a measurements was 0.57, illustrating the potential of the remote sensing algorithms in accounting for chlorophyll a variation in turbid Case 2 waters.