Algal blooms have a negative impact on ecosystems, and in the case of cyanobacterial blooms, they pose a threat to human health, either through direct exposure or contamination of seafood. The latter is particularly relevant in aquaculture and fishery regions. This study focuses on the timely detection and prediction of bloom events.
Multispectral satellite images were used to monitor estuaries with two different types of algal blooms — cyanobacteria and macroalgae. They photosynthesize, meaning they contain chlorophyll, which could be detected by spectral signature. To detect blooms using satellite data, the Floating Algae Index (FAI) was applied in both cases, as it effectively highlights the spectral signature of chlorophyll. This method proved effective in identifying both floating macroalgae (e.g., Ulva spp.) and cyanobacteria suspended in the water column.
To analyse the spatiotemporal dynamics of algal blooms, a logistic regression model was developed. The model was validated against direct field observations and demonstrated higher accuracy compared to a simple classification approach. Additionally, environmental variables such as temperature, water flow rate, is estuary mouth open or closed, salinity and other factors were incorporated into the model.
The developed model can be used to predict the likelihood of blooms based on environmental factors.