Linking phytoplankton ecology and phytoplankton types with ocean carbon and physics: lessons from satellite data and CMIP5 models

Marine phytoplankton photosynthesis is a critical part of the biological carbon pump (BCP), which transfers to depth part of the resulting organic carbon (C), leading to C ingassing at the surface and oceanic C sequestration. Seasonal, interannual and multi-decadal environmental changes alter ocean nutrients, light and temperature, affecting in turn ocean plankton and BCP. We describe - and aim to unify - two separate approaches to understand the seasonal to interannual behavior of ocean phytoplankton ecology and its links with ocean C. Our focus here will be on the subpolar Pacific, subpolar North Atlantic and the subpolar Southern Ocean, regions with nutrient and light co-limited phytoplankton productivity and generally high seasonality in biology.

1.Satellite color datasets. We analyze the seasonal and interannual variability in observations of phytoplankton biomass and phytoplankton size groups from multiple satellite color products. Inspired by Fay and McKinley (2017) we analyze the connection between our phytoplankton indices and observed air-sea CO₂ fluxes, and compare them with those across CMIP5 models. We seek to understand the links and physics behind the observed variability of biomass and air-sea CO₂ fluxes. We explore a variety of physical indices, including turbulent mixing indices related to wind stirring and buoyancy forcing, indices for phytoplankton light limitation, and the rate of generation of turbulent kinetic energy.

2.Model intercomparison. We analyze extensively the output of multiple CMIP5 ESMs with explicit marine ecological modules, to identify the common physical mechanisms involved in the phytoplankton biomass, productivity, organic C export variability and changes over the historical period and 21st century (scenario RCP8.5). Across subpolar regions, we discuss coupling and decoupling between biomass, productivity, export, and air-sea C fluxes and present inter-model statistical significance. We also compare our CMIP5 and satellite-based results. For example, we find that CMIP5 models are poor at reproducing observed satellite-based phytoplankton biomass phenology and interannual variability, e.g. model diatoms are too seasonal in high latitudes but not seasonal enough at the subpolar-subtropical front, with interesting implications for air-sea CO2 fluxes.