Modeling coccolithophores in the global oceans

Abstract

Abstract Coccolithophores are important ecological and geochemical components of the global oceans. A global three-dimensional model was used to simulate their distributions in a multi-phytoplankton context. The realism of the simulation was supported by comparisons of model surface nutrients and total chlorophyll with in situ and satellite observations. Nitrate, silica, and dissolved iron surface distributions were positively correlated with in situ data across major oceanographic basins. Global annual departures were +18.9% for nitrate (model high), +5.4% for silica, and +45.0% for iron. Total surface chlorophyll was also positively correlated with satellite and in situ data sets across major basins. Global annual departures were −8.0% with Sea-viewing Wide Field-of-view Sensor (SeaWiFS) (model low), +1.1% with Aqua, and −17.1% with in situ data. Global annual primary production estimates were within 1% and 9% of estimates derived from SeaWiFS and Aqua, respectively, using a common primary production algorithm. Coccolithophore annual mean relative abundances were 2.6% lower than observations, but were positively correlated across basins. Two of the other three phytoplankton groups, diatoms and cyanobacteria, were also positively correlated with observations. Distributions of coccolithophores were dependent upon interactions and competition with the other phytoplankton groups. In this model, coccolithophores had a competitive advantage over diatoms and chlorophytes by virtue of a greater ability to utilize nutrients and light at low values. However, their higher sinking rates placed them at a disadvantage when nutrients and light were plentiful. In very low nutrient conditions, such as the mid-ocean gyres, coccolithophores were unable to compete with the efficient nutrient utilization capability and low sinking rate of cyanobacteria. Comparisons of simulated coccolithophore distributions with satellite-derived estimates of calcite concentration and coccolithophore blooms showed some agreement, but also areas of departure. Vast blooms observed in the North Atlantic were well-represented by the model. However, model coccolithophores were nearly absent in the North Pacific, while calcite estimates suggested widespread abundance in summer. In situ observations supported the satellite calcite, suggesting a deficiency in the model. New satellite estimates of phytoplankton groups indicated good agreement of diatoms in one case, and poor agreement in general in another. Comparisons of phytoplankton group primary production with other models showed wide disparity. The divergence among models and satellite estimates is common for such an emerging field of research. The quantitative comparisons with in situ observations are encouraging, but disparities with model and satellite estimates suggested that further research is needed.