Using Particle Size Distribution and Habitat Models to Predict Phytoplankton Functional Types in the Atlantic Ocean

Phytoplankton community composition in the oceans is a critical determinant to both marine ecosystem structure and biogeochemical processes. A model was constructed to predict the dominant of five phytoplankton functional types (PFT) in the surface layer of the Atlantic Ocean as a fraction of biomass using a combination of particle size distribution (PSD) retrieved from satellite ocean color radiometric measurements and empirical niche models developed for each PFT. Presence of dominant PFTs were developed from a pigment data set from the Atlantic Ocean and co-located with gridded satellite, modeled and climatological data. The PFT model employs PSD, sea-surface temperature, photosynthetically available radiation, mixed-layer depth, wind and nutrients as inputs, and successfully classified the dominant groups 85% using data from the training set and 70% using an independent data set. The predicted monthly distribution patterns of the dominant PFT in the Atlantic during the SeaWiFS era appear reasonable yet require additional validation.