Deep Learning for Simulating Harmful Algal Blooms Using Ocean Numerical Model

In several countries, the public health and fishery industries have suffered from harmful algal blooms (HABs) that escalated to become a global issue. Though computational modeling offers an effective means understand mitigate adverse effects of HABs, it is challenging design models adequately reflect complexity HAB dynamics. This paper presents method involving application deep learning ocean model for simulating Alexandrium catenella . The classification regression convolutional neural network (CNN) are used blooms. CNN determines bloom initiation while estimates density. GoogleNet Resnet 101 identified as best structures CNNs, respectively. corresponding accuracy root square error values determined 96.8% 1.20 [log(cells L –1 )], results obtained in this study reveal simulated distribution follow bloom. Moreover, Grad-CAM identifies salinity temperature contributed whereas NH 4 -N influenced growth