Relating temporal and spatial patterns of DMSP in the Barents Sea to phytoplankton biomass and productivity

Dimethylsulfoniopropionate (DMSP), produced by many marine phytoplankton, is the main precursor of the climate relevant gas dimethylsulfide (DMS). Currently, it is generally accepted that the relationship between DMSP and phytoplankton biomass (as chlorophyll a), while not representative of the absolute magnitude of the DMSP pool, is a good indicator of ecosystem structure. In this study we test the strength of the relationships between DMSP and various phytoplankton parameters in Arctic shelf waters of the Barents Sea. Our objective is to assess the predictive value that traditional phytoplankton carbon parameters have on DMSP. We discuss C:DMSP-S variability as a function of seasonality, water masses, grazing and nutrient limitation. For this purpose we analyze data from 5 cruises including winter, spring and summer conditions and across the seasonal ice zone at the time of the study. Highest phytoplankton DMSP concentration was usually measured at the ice edge. Marked seasonal variability was observed in phytoplankton carbon biomass and production but not necessarily in the particulate fraction of DMSP (DMSPp), resulting in seasonally varying C:DMSP-S. High winter DMSPp concentrations, when chlorophyll a and primary production were lowest and flagellates dominant, suggest a heterotrophic source. The production of extracellular carbon and the pool of dissolved DMSP (DMSPd) followed similar seasonal trends, with enhanced concentrations in spring, and we suggest that high dissolved primary production induced by nutrient limiting conditions resulted in high DMSPd concentrations. Mesoscale changes in total DMSP (particulate + dissolved) may be modeled from basin-wide total phytoplankton primary production (rather than from phytoplankton biomass) at seasonal and interannual scales. We conclude there is predictive power of DMSP concentrations in the Barents Sea based on seasonality, the position of the ice edge and the distribution of phytoplankton variables.