Mapping the phytoplankton spring bloom using an Imaging FlowCytobot and machine learning on FerryBox ship of opportunity in the eastern North Sea

Summary

During the spring of 2024 (February to May), we equipped a ship of opportunity (MS Color Fantasy) with an Imaging Flow Cytobot (McLane Research Laboratories, Inc.) to observe the spring phytoplankton bloom community composition, spatial distribution, and phenology between Oslo, Norway and Kiel, Germany covering these three areas of interest: Oslofjord, Skagerrak and Kattegat. We trained a convolutional neural network (CNN) using species relevant to the spring bloom to identify and semi-quantitatively detect the development of the phytoplankton community. We will show the results and discussion about the classifier, machine learning training steps and identification and quantification. We will also assess the physical and biogeochemical influences contributing to the phytoplankton bloom development.