Modeling for multi-temporal cyanobacterial bloom dominance and distributions using landsat imagery

Abstract

Cyanobacterial blooms pose great risks to aquatic systems due to its increase in frequency and severity worldwide. This study developed multi-temporal satellite remote sensing models that accurately portray spatial distribution of freshwater cyanobacterial blooms, which traditional monitoring methods cannot accomplish. The models were developed using multiple linear regression involving in-situ measurements and Landsat imagery in Lake Champlain for cyanobacterial blooms and other water quality variables i.e. chlorophyll-a, total phytoplankton, chlorophyte, chrysophyta, pyrrophyta, secchi disk and temperature. The developed models with Landsat imagery successfully captured the distribution of cyanobacterial blooms and other variables (adjusted R2 = 0.6–0.98). Multi-temporal analysis of remote sensing highlighted the high bloom years with quantified cyanobacterial levels over the entire water body, associated with temperature, indicating the impact of precipitation on cyanobacterial bloom growth. The modeling approach in this study can provide information to improve our understanding of algal dominance and dynamics. This approach can be applied to other freshwater systems, providing a pathway for rapid response systems.