Abstract: Phytoplankton exhibit diverse physiological responses to temperature which influence their fitness in the environment and consequently alter their community structure. Here, we explored the sensitivity of phytoplankton community structure to thermal response parameterization in a modelled marine phytoplankton community. Using published empirical data, we evaluated the maximum thermal growth rates (µmax) and temperature coefficients (Q10; the rate at which growth scales with temperature) of six key Phytoplankton Functional Types (PFTs): coccolithophores, cyanobacteria, diatoms, diazotrophs, dinoflagellates, and green algae. Following three well-documented methods, PFTs were either assumed to have (1) the same µmax and the same Q10 (as in to Eppley, 1972) (2) a unique µmax but the same Q10 (similar to Kremer et al. 2017) or (3) a unique µmax and a unique Q10 (following Anderson et al. 2021). These trait values were then implemented within the MIT biogeochemistry and ecosystem model (called Darwin) for each PFT under a control and climate change scenario.

Our results suggest that applying a µmax and Q10 universally across PFTs (as in Eppley, 1972) leads to unrealistic phytoplankton communities, which lack diatoms globally. Additionally, we find that accounting for differences in the Q10 between PFTs can significantly impact each PFT’s competitive ability, especially at high latitudes, leading to altered modeled phytoplankton community structures in our control and climate change simulations. This then impacts estimates of biogeochemical processes, with, for example, estimates of export production varying by ~10% in the Southern Ocean depending on the parameterization.

Our results indicate that the diversity of thermal response traits in phytoplankton not only shape community composition in the contemporary and future, warmer ocean, but that these traits have significant feedbacks on global biogeochemical cycles.