The measurements were carried out during the POLAR-CHANGE cruise aboard the B.I.O. Hespérides between 14 February and 17 March 2023, in coastal waters around the Antarctic Peninsula and in the northern Weddell Sea. The researchers installed two Air-Sea Interface Tanks, each holding 0.9 cubic metres of seawater pumped from 4 metres depth, on the ship’s deck. One tank was covered with a lid transparent to all wavelengths, while the other was covered with a lid blocking UV radiation below 380 nanometres. Both headspaces were flushed continuously with particle-free air at 25 litres per minute, giving a 40-minute residence time, and the outgoing air was analysed with a proton transfer reaction quadrupole mass spectrometer for volatile compounds and with condensation particle counters for nanoparticles down to 1 nanometre in diameter. The seawater was pumped through a system designed to be gentle and non-toxic to microorganisms, and during two dedicated experiments the tanks were filled once and the water was incubated in place for two full days.

Of the 22 volatile compounds monitored, only DMS and MeSH were enriched in the tank headspaces relative to the incoming air. No enrichment of benzene, toluene, ethylbenzene and xylene, monoterpenes, isoprene or iodomethane was detected, which the researchers attribute to these compounds being undersaturated in the surface water. Sea-air fluxes of DMS and MeSH were always positive, meaning both gases degassed from seawater to air throughout the cruise. In the control tank exposed to natural light, the average DMS flux was 3.03 picomoles per square metre per second, peaking at 8.68, while the average MeSH flux was 0.64 picomoles per square metre per second, with a maximum of 2.65. Estimated aqueous DMS concentrations averaged 1.8 nanomolar in the control tank, reaching 4.5 nanomolar, and estimated MeSH concentrations averaged 0.8 nanomolar with a peak of 3.2 nanomolar. These DMS fluxes are comparable to those measured with the same tank system in subantarctic and subtropical waters of the Southwest Pacific during the earlier Sea2Cloud campaign.

The spatial pattern of dissolved DMS followed the oceanographic setting. Calculated aqueous concentrations were highest in the open Southern Ocean north-east of the Antarctic Peninsula, at 2.5 to 3 nanomolar, lowest in the Marginal Ice Zone of the Weddell Sea and in the Gerlache Strait, where values fell to around 1 nanomolar or below, and moderate at roughly 1.5 to 2 nanomolar along the western coast of the peninsula. The low values in the Gerlache Strait coincided with reduced salinity, reflecting the inflow of freshwater from melting ice. Fluxes did not simply mirror these concentrations, because they also depend on the DMS content of the incoming flushing air; fluxes were low over part of the Weddell Sea transect despite high seawater concentrations, and particularly high in coastal waters where ambient air DMS was low.

A striking diel rhythm emerged in the relative contribution of the two gases. DMS emission did not vary significantly between day and night, but the ratio of MeSH flux to the combined DMS and MeSH flux showed a clear maximum at night and declined during the day. The median ratio across the whole cruise was 0.17 in the control tank, consistent with values of roughly 0.10 to 0.40 reported in earlier studies from the North Atlantic, the Baltic and North Seas, the Southwest Pacific and previous enclosure experiments. The researchers attribute the daytime decline partly to the higher gas-phase reactivity of MeSH with hydroxyl radicals within the 40-minute headspace residence time, but the fact that the difference between the two tanks was largest at night points to additional processes.

Removing UV light had a statistically significant effect on both gases. DMS fluxes were on average 24 percent higher and MeSH fluxes 58 percent higher in the UV-shielded tank, confirmed by paired Wilcoxon signed-rank tests. Notably, the largest differences in air concentrations, especially for MeSH, appeared during the night, several hours after UV exposure had ended, which the authors interpret as a lag between daytime exposure and the biological or chemical response of the water. They suggest that aqueous-phase photochemistry driven by daytime UV, including DMS photooxidation and reactions with oxidants produced from chromophoric dissolved organic matter and nitrate photolysis, generates losses that persist into the night, since some water parcels remain in the tanks far longer than the nominal six-hour residence time. Ozone concentrations were also lower in the UV-shielded tank during the two enclosed-water experiments, by 38 percent in Experiment A and 50 percent in Experiment B, possibly reflecting enhanced consumption by the larger fluxes of reactive sulfur gases or photochemical differences in ozone production.

The most dramatic effect of UV removal was on new particle formation. In the control tank, particle concentrations above 1 nanometre reached only 38 particles per cubic centimetre, with formation rates peaking at 0.02 particles per cubic centimetre per second. In the UV-shielded tank, concentrations reached 300 particles per cubic centimetre and formation rates approached 0.15 particles per cubic centimetre per second, up to a factor of ten higher. These rates slightly exceed those reported for Antarctic regions and remain below Arctic values. The strongest particle formation occurred with coastal waters on both sides of the Antarctic Peninsula, and air MeSH concentrations showed the strongest correlation with particle numbers in the UV-shielded tank with coastal seawater, with a coefficient of determination of 0.56 for particles above 10 nanometres, compared with 0.31 for DMS.

Crucially, neither DMS nor MeSH alone could explain the new particle formation, and no measured biological variable differed between the tanks as sharply as the nanoparticle concentrations. This mirrors the conclusion of the earlier Sea2Cloud campaign in the Southwest Pacific, which used the same tank system, and points to a yet-unidentified precursor or process that, despite occurring during daylight hours, is enhanced when UV radiation is reduced. The timing also differs from the Pacific experiments, where new particles formed preferentially at the end of the night before dawn, whereas in the Antarctic tanks formation occurred around midday or at night. A significant positive correlation was found between particles above 1 nanometre and the abundance of 7 to 15 micrometre nanophytoplankton during the experiments, prompting the authors to call for further studies identifying the new-particle precursors related to nanophytoplankton.

Biological measurements inside the tanks during the two 48-hour experiments, one with coastal water and one with open ocean water, revealed how UV light affected the microbial community. Cell abundances of 1 to 2 micrometre picoplankton and of 2 to 7 and 7 to 15 micrometre nanophytoplankton were significantly lower under UV exposure, while the largest size class, 15 to 20 micrometres, showed no significant difference, indicating strongly size-dependent UV sensitivity. The photosynthetic efficiency of the phytoplankton, measured as the Fv/Fm ratio, followed the expected diel pattern with lower values at midday, and UV exposure caused a slight additional light stress at noon that negatively affected the short-term development of nanophytoplankton, particularly in open Southern Ocean waters. Bacterial concentrations correlated positively with calculated aqueous DMS and MeSH concentrations during the experiments, with Spearman coefficients between 0.43 and 0.63, consistent with the known bacterial cleavage and demethylation of dimethylsulfoniopropionate that produces the two gases. A relationship between the sulfur compounds and two of the virus groups was also observed during the experiments, although the authors caution that such correlations should not be read as direct cause-and-effect relationships.

The findings carry implications for how sulfur-driven aerosol formation is represented in models of the Southern Ocean, a region where biogenic emissions clearly exceed transported anthropogenic precursors during austral summer and where recent modelling suggests that including MeSH alongside DMS strengthens the aerosol cooling effect. The authors caution that the tank system generates fluxes equivalent to open-ocean conditions at a wind speed of about 2 metres per second, that aqueous DMS and MeSH concentrations were calculated from headspace measurements rather than directly measured, and that the mass spectrometer could not fully exclude interference from compounds sharing the same mass-to-charge ratio. Even so, the demonstration that UV light suppresses both sulfur emissions and nanoparticle production in natural Antarctic seawater marks a significant step in understanding the ocean’s role in regulating its own atmospheric chemistry.

Subject of Research: Effects of UV light on sea-air emissions of dimethyl sulfide and methanethiol and on new particle formation in Antarctic seawater

Article Title: DMS, MeSH and nanoparticles in semi-controlled deck-borne experiments using Antarctic seawater: on the effect of UV light

Article References: Chamba, G., Sellegri, K., Gros, V., Rose, C., Williams, L., Berdalet, E., Vaqué, D., Vila, M., Güell-Bujons, Q., Sà, E.-L., Castillo, Y. M., Rocchi, A., Colomb, A., Pichon, J.-M., Ribeiro, M., Wohl, C., Dall'Osto, M., & Simó, R. (2026). DMS, MeSH and nanoparticles in semi-controlled deck-borne experiments using Antarctic seawater: on the effect of UV light. Biogeosciences, 23(20), 7469-7492. https://doi.org/10.5194/bg-23-7469-2026

Image Credits: AI Generated

DOI: 10.5194/bg-23-7469-2026

Keywords: dimethyl sulfide, methanethiol, Antarctic Peninsula, Southern Ocean, ultraviolet radiation, new particle formation, air-sea flux, marine sulfur cycle, phytoplankton, bacteria, aerosols, POLAR-CHANGE campaign