A count model approach on the occurrences of harmful algal blooms (HABs) in Ambon Bay
Mohammad Mahmudi, Lukas Giovani Gonzales Serihollo, Endang Yuli Herawati, Evellin Dewi Lusiana, Nanik Retno Buwono
Abstract
Climate change and increased anthropogenic activities have resulted in an imbalance in the aquatic ecosystem and have triggered the appearance of harmful algae in Ambon Bay, Indonesia. This study aims to identify the phytoplankton community structure, measure physicochemical water quality (temperature, salinity, pH, DO, nitrate and phosphate) in Ambon Bay, and create a prediction model to estimate the occurrence of harmful algal bloom based on these water quality measures. The results of the statistical count model (Poisson regression) showed that three phytoplankton divisions were observed: Bacillariophyta, Dinophyta and Cyanophyta. Of these, Bacillariophyceae were the most abundant. The only species of the Cyanophyta division identified was Trichodesmium, a type of harmful algae that can produce high biomass that may clog fish gills and generate low oxygen. Our Poisson regression model suggested that all water quality factors measured affected the abundance of Trichodesmium in Ambon Bay and that, moreover, rising levels of nitrate and salinity will cause a surge in Trichodesmium.