The phytoplankton groups differ in maximum growth rates, sinking rates, nutrient requirements, and optical properties. The 4 nutrients are nitrate, regenerated ammonium, silica to regulate diatom growth, and iron. Three detrital pools provide storage of organic material, sinking, and eventual remineralization. Carbon

Nutlin-3a order cycling involves dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC; Fig. 2). DOC has sources from phytoplankton, herbivores, and carbon detritus, and a sink to DIC. DIC has sources from phytoplankton, herbivores, carbon detritus, and DOC, and is allowed to exchange with the atmosphere, which can be either a source or sink. The ecosystem sink for DIC is phytoplankton, through photosynthesis. This represents the biological pump portion PI3K inhibitors ic50 of the carbon dynamics. The solubility pump portion is represented by the interactions among temperature, alkalinity

(parameterized as a function of salinity), silica, and phosphate (parameterized as a function of nitrate). The alkalinity/salinity parameterization utilizes the spatial variability of salinity in the model adjusted to mean alkalinity TA=TA̲S/S̲where TA is total alkalinity and S is salinity. The underscore represents global mean values. TA is specified as 2310 μE kg−1 (Ocean Model Intercomparison Project (OCMIP; www.ipsl.jussieu.fr/OCMIP) and S as 34.8 PSU (global model mean). Since the model contains nitrate but not phosphate, we estimate phosphate by multiplying nitrate by 0.1. This is derived from the global mean ratio of nitrate to phosphate from NODC for their top three Florfenicol standard levels. The calculations for the solubility pump follow the standards set by the Ocean Model Intercomparison Project (reference above). We recognize that this approximation for alkalinity is not optimal, but the surface results compare favorably with data (see Gregg et al., 2013). The difference between the model and GLODAP global surface alkalinity is 2.7 μEq l−1

(=0.1%) with basin correlation of 0.95 (P < 0.05) ( Gregg et al., 2013). We consider this sufficient for the present purpose of intercomparing model results from forcing by different reanalysis products. We employ a locally-developed lookup table valid over modern ranges of DIC, salinity, temperature, and nutrients for computational efficiency, at little cost to accuracy. Air–sea CO2 exchange as a function of wind uses the Wanninkhof (1992) formulation, as is common in global and regional ocean carbon models (e.g., McKinley et al., 2006). A more complete description of NOBM can be found in Gregg et al. (2013). NOBM is spun-up for 200 years under climatological forcing from each reanalysis. Initial conditions for DIC are derived from the Global Data Analysis Project (GLODAP; Key et al., 2004). DOC initial conditions are set to 0 μM. Subsequent tests with non-zero DOC initial conditions showed negligible differences. Other initial conditions are described in Gregg and Casey (2007).