MEMBER

Construction of prediction models for predicting the algal bloom occurrence in dam reservoirs

Dam reservoirs are essential water resources for water supply, but algal blooms, the rapid increase of algae, have been confirmed in many reservoirs in the world. Some algae can produce toxic substances and the algal scum may inhibit the process of water purification in a water treatment plant. Predicting the occurrence of algal blooms beforehand enables us to implement appropriate measures to minimize the damage caused by algae. Various factors affecting the occurrence of algal blooms have been investigated and it is widely known that the phenomenon has a site-specific nature. Thus, prediction models which can deal with the site-specificity of algal blooms should be developed. In this study, we constructed prediction models with a combination of variables selection algorithm and machine learning algorithm which can be applied in other reservoirs than the targeted reservoirs in this study. We aim to improve the model performances by integrating satellite and reanalysis atmospheric data using google earth engine in the next step.

1. Predictive models of algal bloom with sparse modeling and support vector machine
   Yohei Miura, Hiroomi Imamoto, Yosuhiro Asada, Michihiko Akiba, Osamu Nishimura, Daisuke Sano
   IWA World Water Congress & Exhibition 2022
   Bella Center, Copenhgen, Denmark (September 11-15, 2022)
2. Prediction of algal bloom in reservoir dams using sparce modeling and Support Vector Machine (poster presentation)
   Yohei Miura, Shota Yashima, Hiroomi Imamoto, Yasuhiro Asada, Michihiko Akiba, Daisuke Sano
   Water Convention 2022 at Singapore Internatinal Water Week (hybrid)
   Apr 17-22, 2022.