Remote sensing of cyanobacteria in turbid productive waters

Abstract

Cyanobacterial algal bloom is a major water quality issue in inland lakes, reservoirs, and estuarine environments because of its scum and bad odor forming and toxin producing abilities. Health risks from cyanobacterial toxin can vary from skin irritations to fever, intestinal problems, and neurological disorders. Terminations of blooms also cause oxygen depletion leading to hypoxia and widespread fish kills. Adding to the problem, many species of cyanobacteria produce odorous compounds such as geosmin and 2-methylisoborneol (MIB) that cause "earthy-muddy" and "musty" odor in drinking water, which is also a serious issue in aquaculture and drinking water industry. Therefore continuous monitoring of cyanobacterial presence in recreational water bodies, surface drinking water sources, and water bodies dedicated for aquaculture is highly required for their early detection and subsequent issuance of a health warning and reducing the economic loss.

Remote sensing techniques offers the capability of identifying and monitoring cyanobacterial blooms in a synoptic scale. Over the years, the scientific community has focused on developing methods to quantify cyanobacterial biomass using phycocyanin, an accessory photosynthetic pigment, as a marker pigment. However, because of the confounding influence of chlorophyll- a and other photo pigments, remote retrieval of phycocyanin signal from turbid productive water has been a difficult task. This dissertation analyzes the potential of remote sensing techniques and develops empirical and quasi-analytical algorithms to isolate the phycocyanin signal from the remote sensing reflectance data using a set of radiative transfer equations and retrieves phycocyanin concentration in the water bodies. An extensive dataset, consisting of in situ radiometric measurements, absorption measurements of phytoplankton, colored dissolved organic matter, detritus, and pigment concentration, was used to optimize the algorithms. Validations of all algorithms were also performed using an independent dataset and errors and uncertainties from the algorithms were discussed. Despite the simplicity, an empirical model produced highest accuracy of phycocyanin retrieval, whereas, the newly developed quasi-analytical phycocyanin algorithm performed better than the existing semi-analytical algorithm. Results show that remote sensing techniques can be used to quantify cyanobacterial phycocyanin abundance in turbid and hypereutrophic waters.