Abstract. Phytoplankton community composition profoundly influences patterns of nutrient cycling and the structure of marine food webs; therefore predicting present and future phytoplankton community structure is of fundamental importance to understanding how ocean ecosystems are influenced by physical forcing and nutrient limitations. In this paper, we develop a mechanistic model of phytoplankton communities that includes multiple taxonomic groups, test the model at two contrasting sites in the modern ocean, and then use the model to predict community reorganization under different global change scenarios. The model includes three phytoplankton functional groups (diatoms, coccolithophores, and prasinophytes), five nutrients (nitrate, ammonium, phosphate, silicate and iron), light, and a generalist zooplankton grazer. Each taxonomic group was parameterized based on an extensive literature survey. The model successfully predicts the general patterns of community structure and succession in contrasting parts of the world ocean, the North Atlantic (North Atlantic Bloom Experiment, NABE) and subarctic North Pacific (ocean station Papa, OSP). In the North Atlantic, the model predicts a spring diatom bloom, followed by coccolithophore and prasinophyte blooms later in the season. The diatom bloom becomes silica-limited and the coccolithophore and prasinophyte blooms are controlled by nitrogen, grazers and by deep mixing and decreasing light availability later in the season. In the North Pacific, the model reproduces the low chlorophyll community dominated by prasinophytes and coccolithophores, with low total biomass variability and high nutrient concentrations throughout the year. Sensitivity analysis revealed that the identity of the most sensitive parameters and the range of acceptable parameters differed between the two sites. Five global change scenarios are used to drive the model and examine how community dynamics might change in the future. To estimate uncertainty in our predictions, we used a Monte Carlo sampling of the parameter space where future scenarios were run using parameter combinations that produced adequate modern day outcomes. The first scenario is based on a global climate model that indicates that increased greenhouse gas concentrations will cause a later onset and extended duration of stratification and shallower mixed layer depths. Under this scenario, the North Atlantic spring diatom bloom occurs later and is of a smaller magnitude, but the average biomass of diatoms, coccolithophores and prasinophytes will likely increase. In the subarctic North Pacific, diatoms and prasinophytes will likely increase along with total chlorophyll concentration and zooplankton. In contrast, coccolithophore densities do not change at this site. Under the second scenario of decreased deep-water phosphorus concentration, coccolithophores, total chlorophyll and zooplankton decline, as well as the magnitude of the spring diatom bloom, while the average diatom and prasinophyte abundance does not change in the North Atlantic. In contrast, a decrease in phosphorus in the North Pacific is not likely to change community composition. Similarly, doubling of nitrate in deep water does not significantly affect ecosystems at either site. Under decreased iron deposition, coccolithophores are likely to increase and other phytoplankton groups and zooplankton to decrease at both sites. An increase in iron deposition is likely to increase prasinophyte and diatom abundance and decrease coccolithophore abundance at both sites, although more dramatically at the North Pacific site. Total chlorophyll and zooplankton are also likely to increase under this scenario at both sites. Based on these scenarios, our model suggests that global environmental change will inevitably alter phytoplankton community structure and potentially impact global biogeochemical cycles.
A simple optical index shows spatial and temporal heterogeneity in phytoplankton community composition during the 2008 North Atlantic Bloom Experiment