Forecasting green tide events in a semi-closed tidal flat using artificial intelligence and environmental big data

Abstract

Green tide is a phenomenon in which Ulva spp. grow abnormally in nearshore shallow water areas, resulting in degradation of the water quality and scenery, and generation of unpleasant odors. The Yatsu tidal flat, located in the inner part of Tokyo Bay, is among the most significant tidal flats remaining in Japan. In recent years, the occurrence of green tide events due to extraordinary growth of Ulva spp. has become problematic. In the present study, we used field observations conducted from 2003 to 2017 to construct a green tide forecasting model using artificial intelligence and a neural network. We found that forecast values could be accurately hindcast using field data. Next, we attempted to forecast the area to be affected by green tide using environmental big data for the part of Tokyo Bay near the Yatsu tidal flat. For the area selected, we showed that green tide forecasting 1 month ahead is achievable using environmental big data from the area near the tidal flat.

Keywords:

• Green tide
• Ulva spp
• Tokyo Bay
• tidal flat
• numerical simulation
• artificial intelligence

Disclosure statement

No potential conflict of interest was reported by the author(s).