Dimethylsulphide production in the Southern Ocean using a nitrogen-based flow network model and field measurements from ACE-1

Dimethylsulphide (DMS) has been implicated in climate change as a possible negative feedback to global warming, and several Models have been developed that simulate the production of DMS in the marine environment. The focus of this study is to improve the nitrogen based Gabric Model, using field data collected during the Southern Hemisphere First Marine Aerosol Characterisation Experiment (ACE-1) in the Southern Ocean in 1995. Two Model Runs (Series A and B) were carried out with six simulations of varying biotic and abiotic inputs applied over the voyage transect (41-48°S), reflecting Model default values or field values from the experiment. The abiotic inputs were time-step, dissolved dimethylsulphoniopropionate (DMSP) and DMS, and the biotic nitrogen inputs were from phytoplankton, bacteria, zooflagellates, large protozoa, micro and mesozooplankton. The focus of the abiotic assessment was nutrient (nitrate) uptake and dissolved DMSP and DMS output. Model output of the biotic compartments was assessed for congruence with predicted ecological patterns of succession.

Despite a limited data set the study provides a good insight into the utility of the Model, which functioned as a heuristic rather than predictive tool. In simulation 1 (Series A) where the only field value was nitrate, all latitudes from 41-48°S concurred with the ecological succession predicted by the Model authors and the successional pattern predicted by other researchers, with a double phytoplankton peak indicating remineralisation of nitrogen via the microbial loop. In many simulations the Model produced lower values of dissolved DMS than were measured, and production of DMS in the Model appears constrained. However, in simulation 5 (Series A) DMS model outputs were closest to the mean dissolved DMS levels reported on RV Discoverer. In this simulation, field values were used for phytoplankton, nitrate, dissolved DMSP and DMS, with bacterial abundance and micro and mesozooplankton increased over their Gabric default values. Also, the phytoplankton double peak occurred earlier, as did the peaks in bacteria, zooflagellates, and large protozoa. Simulations that deviated more significantly from the predicted successional patterns were characterised by single peaks in phytoplankton growth and delayed bacterial growth. Series C simulations at latitude 43°S, in an attempt to reduce phytoplankton predation by bacteria, increased DMS output reasonably successfully. However, significant recalibration of the Model is recommended in conjunction with field studies to gather vital background biological data?-?particularly in the areas of nutrient limitation, phytoplankton speciation, and the cellular content of the DMS precursor compound, DMSP.