The influence of sea ice on Antarctic ice core sulfur chemistry and on the future evolution of Arctic snow depth: Investigations using global models

Abstract

Observational studies have examined the relationship between methanesulfonic acid (MSA) measured in Antarctic ice cores and sea ice extent measured by satellites with the aim of producing a proxy for past sea ice extent. MSA is an oxidation product of dimethylsulfide (DMS) and is potentially linked to sea ice based on observations of very high surface seawater DMS in the sea ice zone. Using a global chemical transport model, we present the first modeling study that specifically examines this relationship on interannual and on glacial-interglacial time scales. On interannual time scales, the model shows no robust relationship between MSA deposited in Antarctica and sea ice extent. We show that lifetimes of MSA and DMS are longer in the high latitudes than in the global mean, interannual variability of sea ice is small (<25%) as a fraction of sea ice area, and sea ice determines only a fraction of the variability (<$30) of DMS emissions from the ocean surface. A potentially larger fraction of the variability in DMS emissions is determined by surface wind speed (up to 46%) via the parameterization for ocean-to-atmosphere gas exchange. Furthermore, we find that a significant fraction (up to 74%) of MSA deposited in Antarctica originates from north of 60&degS, north of the seasonal sea ice zone. We then examine the deposition of MSA and non-sea-salt sulfate (nssSO₄<super>2-</super>) on glacial-interglacial time scales. Ice core observations on the East Antarctic Plateau suggest that MSA increases much more than nssSO₄<super>2-</super> during the last glacial maximum (LGM) compared to the modern period. It has been suggested that high MSA during the LGM is indicative of higher primary productivity and DMS emissions in the LGM compared to the modern day. Studies have also shown that MSA is subject to post-depositional volatilization, especially during the modern period. Using the same chemical transport model driven by meteorology from a global climate model, we examine the sensitivity of MSA and nssSO₄<super>2-</super> deposition to differences between the modern and LGM climates, including sea ice extent, sea surface temperatures, oxidant concentrations, and meteorological conditions. We are unable to find a mechanism whereby MSA deposition fluxes are higher than nssSO₄<super>2-</super> deposition fluxes on the East Antarctic Plateau in the LGM compared the modern period. We conclude that the observed differences between MSA and nssSO₄<super>2-</super> on glacial-interglacial time scales are due to post-depositional processes that affect the ice core MSA concentrations. We can not rule out the possibility of increased DMS emissions in the LGM compared to the modern day. If oceanic DMS production and ocean-to-air fluxes in the sea ice zone are significantly enhanced by the presence of sea ice as indicated by observations, we suggest that the potentially larger amplitude of the seasonal cycle in sea ice extent in the LGM implies a more important role for sea ice in modulating the sulfur cycle during the LGM compared to the modern period. We then shift our focus to study the evolution of snow depth on sea ice in global climate model simulations of the 20th and 21st centuries from the Coupled Model Intercomparison Project 5 (CMIP5). Two competing processes, decreasing sea ice extent and increasing precipitation, will affect snow accumulation on sea ice in the future, and it is not known <italic>a priori</italic> which will dominate. The decline in Arctic sea ice extent is a well-studied problem in future scenarios of climate change. Moisture convergence into the Arctic is also expected to increase in a warmer world, which may result in increasing snowfall rates. We show that the accumulated snow depth on sea ice in the spring declines as a result of decreased ice extent in the early autumn, in spite of increased winter snowfall rates. The ringed seal (<italic>Phoca hispida</italic>) depends on accumulated snow in the spring to build subnivean birth lairs, and provides one of the motivations for this study. Using an empirical threshold of 20 cm of snow depth on level sea ice for ringed seal lair success, we estimate a decline of potential ringed seal habitat of nearly 70%.