Vertical Structure in Phytoplankton Growth and Productivity Inferred From Biogeochemical-Argo Floats and the Carbon-Based Productivity Model

Abstract

Estimates of marine net primary production (NPP) commonly rely on limited in situ 14C incubations or satellite-based algorithms mainly constrained to the surface ocean. Here we combine data from biogeochemical Argo floats with a carbon-based productivity model (CbPM) to compute vertically resolved estimates of NPP. Inferred NPP profiles derived by informing the CbPM with float-based, depth-resolved, bio-optical data are able to qualitatively reproduce the vertical structure in NPP inferred from in situ 14C incubations at various ocean regions. At station ALOHA, float-based estimates agree within uncertainty with productivity observations at depth, but underestimate surface NPP. We test the ability of the CbPM to infer the depth-resolved structure in NPP from bio-optical properties in the mixed layer, in similar fashion as how remote sensing algorithms of ocean productivity operate. In Southern Ocean waters, the depth-reconstructing implementation of the CbPM overestimates phytoplankton division rates and Chl:C below the mixed layer, resulting in artificially high subsurface NPP when compared with the fully float-informed implementation of the model. The CbPM subsurface extrapolation of phytoplankton Chl, Chl:C, division rates, and NPP improves by accounting for deep nutrient (iron) stress impacts on photoacclimation in the Southern Ocean. This improvement is also observed in vertically integrated NPP, where the mean bias between model implementations in depth-integrated productivity south of 30°S is reduced by 62% when account for deep iron limitation. Our results demonstrate that profiling data from biogeochemical Argo floats can serve to inform regional adjustments that lead to the improvement of marine productivity algorithms