Spatial and Temporal Variability of Chlorophyll-a and the Modeling of High-Productivity Zones Based on Environmental Parameters: a Case Study for the European Arctic Corridor

Over the past 20 years, increasing temperature and receding ice-cover have led to changes in the Arctic ecosystem. Our study aims to create models that predict the position of high chlorophyll-a concentration (Chl-a) zones in the European Arctic Corridor (the Barents, Norwegian and Greenland Seas) to monitor these changes. Firstly, we use remotely sensed data to assess spatial and temporal changes in correlation between Chl-a and environmental parameters that could influence Chl-a in the region – Photosynthetically Active Radiation (PAR), Sea Surface Temperature (SST), Mixed Layer Depth (MLD) and Sea Surface Salinity (SSS) – over the 2010–2019 time period. We found significant correlation (∣r∣ = 0.6–0.8) between Chl-a and PAR and SST, and medium correlation (∣r∣ = 0.4–0.6) between Chl-a and SSS and MLD, correlation was highest during spring periods. Then, using a Random Forest Machine Learning algorithm in the Classifier modification, we created models for each sea to predict the position of high-productivity zones (Chl-a > 1 mg m−3) using environmental parameters. Our results suggested that Chl-a variability in the European Arctic Corridor is mostly determined by PAR (28–32% of Chl-a class variability), SST (25–29%), and SSS (26–31%); MLD played a lesser role (12–17%). According to validation, all the models showed high performance scores (F1-score = 66–95%) and slightly underestimated the total area of high productivity.