In this study we analyzed spatial and temporal variability of chlorophyll-a (Chl-a) concentration as an indicator
of ocean productivity in the European Arctic Corridor (the Barents, Norwegian and Greenland Seas), and its
dependence on the environmental parameters: euphotic layer depth (Zeu), Photosynthetically Active Radiation
(PAR), Sea Surface Temperature (SST), Mixed Layer Depth (MLD) and Sea Surface Salinity (SSS). We used
monthly 4x4 km remote sensing data for 2010-2019 years from: Ocean Colour Climate Change Initiative (OC
CCI) database (Chl-a and Zeu), NASA’s Ocean Biology Processing Group’s MODIS images (PAR), MUR SST
database (SST), ESA’s SMOS images (SSS), while MLD was calculated from in situ EN4 Hadley Center and
ARMOR data. Using the Random Forest Machine Learning algorithm in the Classifier modification we created
models describing the relationship between Chl-a and environmental parameters, and retrieving the position of
high-productivity waters (Chl-a > 1 mg m-3) on the basis of these relationships. The results suggested that Ch-a
variability is mostly determined by light availability (PAR and Zeu) and intrusions of dynamically active warm,
saline Atlantic waters and low-salinity river output and melting ice (SST and SSS as indicators).
According to validation, all the models showed a good performance (f1-score = 74-95%) and slightly underestimated
actual Chl-a values and positions of high-productivity waters.