Times are displayed in (UTC+02:00) Brussels, Copenhagen, Madrid, ParisChange
We present a study on using artificial intelligence (AI) to detect phytoplankton community composition in the Great Lakes from space. The AI model was trained using a dataset of EPA field-measured phytoplankton groups collected from 2012 to 2019. The input features for the model include VIIRS remote-sensing reflectance, surface temperature, as well as geospatial and temporal information. The outputs of the model are the total cell volume concentration and volume fractions of eight different phytoplankton groups, including diatoms, Chlorophytes, Chrysophytes, Cryptophytes, Cyanophytes, Euglenoids, Pyrrhophytes, and others. The testing results show a significant relationship between the model-derived phytoplankton groups and the measured phytoplankton groups. This study highlights the potential of AI in learning patterns from historical data and reconstructing the history of desired water quality parameters in large aquatic systems, such as the Great Lakes.