The distribution, dominance patterns and ecological niches of plankton functional types in Dynamic Green Ocean Models and satellite estimates

Abstract. We compare the spatial and temporal representation of phytoplankton functional types (pPFTs) in four different Dynamic Green Ocean Models (DGOMs; CCSM-BEC, NEMURO, PISCES and PlankTOM5) to derived phytoplankton distributions from two independent satellite estimates, with a particular focus on diatom distributions. Global annual mean surface biomass estimates for diatoms vary between 0.23 mmol C m⁻³ and 0.77 mmol C m⁻³ in the models, and are comparable to a satellite-derived estimate (0.41 mmol C m⁻³). All models consistently simulate a higher zonal mean diatom biomass contribution in the high latitudes than in the low latitudes, but the relative diatom contribution varies substantially between models with largest differences in the high latitudes (20% to 100% of total biomass). We investigate phytoplankton distribution in terms of annual and monthly mean dominance patterns, i.e. the distribution of locations where a given PFT contributes more than 50% to total biomass. In all models, diatoms tend to dominate large areas of the high latitudes of both hemispheres, and the area of the surface ocean dominated by diatoms is significantly higher in the models than in the satellite estimates. We estimate the realized ecological niches filled by the dominant pPFT at each location as a function of annual mean surface nitrate concentration (NO₃), sea surface temperature (SST), and mixed layer depth. A general additive model (GAM) is used to map the probability of dominance of all pPFTs in niche and geographic space. Models tend to simulate diatom dominance over a wider temperature and nutrient range, whereas satellites confine diatom dominance to a narrower niche of low-intermediate annual mean temperatures (annual mean SST < 10 °C), but allow for niches in different ranges of surface NO₃ concentrations. For annual mean diatom dominance, the statistically modelled probability of dominance explains the majority of the variance in the data (65.2–66.6%). For the satellite estimates, the explained deviance is much lower (44.6% and 32.7%). The differences in the representation of diatoms among models and compared to satellite estimates highlights the need to better resolve phytoplankton succession and phenology in the models. This work is part of the marine ecosystem inter-comparison project (MAREMIP).