Assessment of modeling techniques for predicting Harmful Algal Bloom (HAB) outbreak using satellite data

Abstract

Since 1995, Harmful Algal Bloom (HAB) by Cochlodinium polykrikoides has frequently occurred in the South Sea of Korea (SSK). The HAB occurrence is related to various factors such as physical and biological, so it is very difficult to predict the HAB occurrence. To reduce the damage, it is essential to make a preliminary forecast of the HAB occurrence through correlation analysis of the HAB occurrence factors. The purpose of this study is to perform and assess the HAB occurrence prediction using modeling techniques. We used satellite data such as Photosynthetically Available Radiation (PAR) and Sea Surface Temperature (SST) which are physical factors of the HAB occurrence. In order to predict the HAB occurrence, we analyzed the correlation of PAR and SST data with past HAB occurrence time (1998-2018) and investigated the performance of the prediction models based on data curves, statistical analysis method, and learning accumulated database. These results are expected to be useful data for predicting future HAB occurrence.ict the HAB occurrence. To reduce the damage, it is essential to make a preliminary forecast of the HAB occurrence through correlation analysis of the HAB occurrence factors. The purpose of this study is to perform and assess the HAB occurrence prediction using modeling techniques. We used satellite data such as Photosynthetically Available Radiation (PAR) and Sea Surface Temperature (SST) which are physical factors of the HAB occurrence. In order to predict the HAB occurrence, we analyzed the correlation of PAR and SST data with past HAB occurrence time (1998-2018) and investigated the performance of the prediction models based on data curves, statistical analysis method, and learning accumulated database. These results are expected to be useful data for predicting future HAB occurrence.