Influence of dimethyl sulfide on the carbon cycle and biological production

Abstract

Dimethyl sulfide (DMS) is a significant source of marine sulfate aerosol and plays an important role in modifying cloud properties. Fully coupled climate simulations using dynamic marine ecosystem and DMS calculations are conducted to estimate DMS fluxes under various climate scenarios and to examine the sign and strength of phytoplankton-DMS-climate feedbacks for the first time. Simulation results show small differences in the DMS production and emissions between pre-industrial and present climate scenarios, except for some areas in the Southern Ocean. There are clear changes in surface ocean DMS concentrations moving into the future, and they are attributable to changes in phytoplankton production and competition driven by complex spatially varying mechanisms. Comparisons between parallel simulations with and without DMS fluxes into the atmosphere show significant differences in marine ecosystems and physical fields. Without DMS, the missing subsequent aerosol indirect effects on clouds and radiative forcing lead to fewer clouds, more solar radiation, and a much warmer climate. Phaeocystis, a uniquely efficient organosulfur producer with a growth advantage under cooler climate states, can benefit from producing the compound through cooling effects of DMS in the climate system. Our results show a tight coupling between the sulfur and carbon cycles. The ocean carbon uptake declines without DMS emissions to the atmosphere. The analysis indicates a weak positive phytoplankton-DMS-climate feedback at the global scale, with large spatial variations driven by individual autotrophic functional groups and complex mechanisms. The sign and strength of the feedback vary with climate states and phytoplankton groups. This highlights the importance of a dynamic marine ecosystem module and the sulfur cycle mechanism in climate projections.

E3SMv0‐HiLAT: A Modified Climate System Model Targeted for the Study of High‐Latitude Processes

“…Changes in SWCF (middle panels) are evident in several regions, with opposite responses seen between piControl and 4xCO2 states over portions of the high latitudes of both hemispheres and also over much of the (split) ITCZ of the Pacific. Stippling indicates much of this difference to be significant, consistent with the finding of Wang, Maltrud, Elliott, et al () of a statistically significant impact of DMS on cloud forcing under RCP8.5 (Representative Concentration Pathway 8.5, a greenhouse gas concentration trajectory resulting in an approximate 8.5 W/m 2 radiative forcing by the year 2100) conditions, which has a greenhouse gas forcing similar to that in our 4xCO2 scenario. Changes in total cloud fraction (bottom panels) are again opposite in sign between piControl and 4xCO2 states over much of the ITCZ, with stippling there indicating significance.…”

“…In general, the simulated phytoplankton biomass were found to have lower variability than observational data, likely due to the coarseness of model resolution (the same as used in this model), which cannot resolve local spikes of nutrient concentrations. The simulated annual mean DMS concentration of the surface ocean was, in these previous studies, found to be 2.3 nM under the contemporary climate scenario (Wang et al, ; Wang, Maltrud, Elliott, et al, ), in close agreement with the observation‐based estimate of 2.34 nM (Lana et al, ). Simulated zonal mean DMS was also found to be in good agreement with the observational estimate.…”

“…The most prominent differences between the DMS fluxes of 4xCO2 and piControl simulations are an overall increase in emissions south of 30° S and in the high Arctic and a decrease in fluxes between 30° S and 40° N but with large geographic variations (Figure ). The increased fluxes in the Southern Ocean, the sea ice zone south of 65° S, and the Arctic are consistent with increases in surface DMS concentration from oceanic phytoplankton activity, while the decreases elsewhere are consistent with decreased surface DMS concentrations (Wang, Maltrud, Elliott, et al, ). The changes in DMS concentrations occur despite relatively modest changes in phytoplankton biomass, as DMS surface concentrations reflect the phytoplankton community structure and not strictly the chlorophyll biomass (Wang et al, ; Wang, Maltrud, Burrows, et al, ).…”

“…An issue we do take up here, standing as a notable deficiency of the current generation of climate models even as it represents a significant control in the climate system (e.g., Wang, Maltrud, Burrows, et al,), is the impact of high‐latitude ocean/sea ice ecosystems on cloud nucleation. Our effort aims to improve on this situation through changes in our model formulation that are detailed in the subsections below, including (i) modifications to the atmospheric model to allow for prognostic marine aerosol precursors and to reduce SWCF biases, (ii) replacing version 4 of the sea ice model with a more recent version that includes biogeochemistry, and (iii) extending the ocean model's biogeochemistry.…”

“…This coupling enables high‐latitude aerosols and clouds to respond to simulated changes in oceanic DMS production associated with sea ice retreat and to shift in the simulated distribution of Phaeocystis and other phytoplankton species associated with warming and freshening of the surface ocean, increasing ocean stability, and injection of nutrients into the ocean surface from melting sea ice. Shifts in community composition under future climate change have the potential to change the regional distribution of DMS production, with possible regional‐scale impacts on atmospheric surface temperature and sea ice (Wang, Maltrud, Elliott, et al, ).…”