Phytoplankton traits from long-term oceanographic time-series

Trait values are usually extracted from laboratory studies of single phytoplankton species, which presents challenges for understanding the immense diversity of phytoplankton species and the wide range of dynamic ocean environments. Here we use a Bayesian approach and a trait-based model to extract trait values for four functional types and ten diatom species from field data collected at Station L4 in the Western Channel Observatory. We find differences in maximum net growth rate, temperature optimum and sensitivity, half-saturation constants for light and nitrogen, and density-dependent loss terms across the functional types. We find evidence of very high linear loss rates, suggesting that grazing may be even more important than commonly assumed and differences in densitydependent loss rates across functional types, indicating the presence of strong niche differentiation among functional types. Low half-saturation constants for nitrogen at the functional type level may indicate widespread mixotrophy. At the species level, we find a wide range of density-dependent effects, which may be a signal of diversity in grazing susceptibility or biotic interactions. This approach may be a way to obtain more realistic and better-constrained trait-values for functional types to be used in ecosystem modeling. Introduction Phytoplankton perform about half of global photosynthesis, form the base of the marine food web and are important drivers of biogeochemical cycles (Field et al. 1998). Model projections of changes in phytoplankton primary production with climate over the next century are extremely variable (Finkel et al. 2010, Finkel 2014). Projections of changes in communities and biogeochemical cycling usually depend on mechanistic models of phytoplankton productivity parameterized with traits of phytoplankton species (Le Quéré et al. 2005, Litchman et al. 2006). The traits used in models vary according to the research questions, but most commonly include maximum growth rate, Arrhenius-like temperature effects on growth rate, half-saturation parameters linking the growth rate to resource availability, and grazing susceptibility (Litchman et al. 2007, Irwin & Finkel 2016). At present, many of these parameters are not well constrained for phytoplankton communities (Anderson 2005, Irwin & Finkel 2016). Phytoplankton are evolutionarily and ecologically diverse and include many phyla and tens of thousands of species (Sournia et al. 1991, de Vargas et al. 2015). This complexity presents several challenges for trait-based modeling. Trait values measured in the lab are almost always determined for a few key species, while their applications in . CC-BY-NC-ND 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/148304 doi: bioRxiv preprint first posted online Jun. 9, 2017;