Carbon dioxide in northern high latitude oceans: Anthropogenic increase and air-sea flux variability

The aim of this thesis is to further our knowledge of carbon dioxide in the northern high latitude oceans (northern North Atlantic, Barents Sea, and Arctic Ocean) by studying the anthropogenic change in the oceanic CO2, the inter-annual variability of the air-sea CO2 flux, and the relationship between this variability and changes in other oceanic processes. An introductory chapter and four papers are presented. Descriptions of the seawater carbonate system parameters, air-sea exchange of CO2, and related processes are given in the introduction chapter. The anthropogenic increase in partial pressure of CO2 (pCO2) in the surface water of the Barents Sea is evaluated in paper I. The effect of alternations of the Barents Sea climate between cold and warm modes on the annual cycles of seawater fugacity and air-sea flux of CO2 is investigated in paper II. Oceanic uptake of atmospheric CO2 associated with the seasonal formation of sea ice in Storfjorden and the implication for the entire Arctic Ocean is studied in paper III. An assessment of the variations of the air-sea flux of CO2 in the northern North Atlantic for 20 winters (1981-2001) is carried out in paper IV. PCO2 in the surface water of the Barents Sea is shown to have increased parallel with the atmospheric pCO2 between 1967 and 2000-2001 (paper I). This was determined by comparing seawater pCO2 from 1967 with that from 2000-2001. The former was estimated from surface seawater temperature (SST) while the latter was computed from data of total dissolved inorganic carbon and alkalinity. A procedure which accounts for the natural variability was applied and the difference between seawater pC02 of 1967 and that of 2000-2001 is attributed to the uptake of excess CO2. In the Atlantic sector of the Barents Sea, the surface seawater fugacity of CO2 (fCO s''w) is shown to be lower than the atmospheric fCO2 throughout the year, implying that the area is an annual sink of atmospheric CO2 (paper II). Additionally, changes in the meltwater pool during summer, associated with changes in temperature of the inflowing Atlantic Water, is found to give rise to variations in the annual cycle of fCO2s''w. During cold years, sea ice extends south of the polar front and melts in the Atlantic sector. This results in a reduced heating and earlier stratification of the surface water during summer, with the latter triggering a rapid and strong phytoplankton bloom. As a consequence, cold years are characterized by a stronger and shorter lived fCO2s''w drawdown during summer. However, significant differences are not found between the annual mean air-sea CO2 fluxes computed for cold and warm years. In this study, fCO2''s''w was calculated by applying an empirical relationship to a 23-year time series of apparent oxygen utilization, seawater temperature, salinity, and phosphate, which have been grouped into cold and warm years. The result was combined with data of atmospheric mole fraction of CO2 and wind speed to calculate the air-sea flux of CO2. Seasonal sea ice formation and subsequent brine rejection produce high density brine-enriched Shelf Water (BSW) in Storfjorden. It is shown that sea ice formation is accompanied by a seaward flux of atmospheric CO2 (paper III). This was inferred by using inorganic carbon and auxiliary hydrographic and nutrient data collected during four cruises from 1999 to 2002. The inferred flux is 12 times higher for the part of the fjord where open water and/or thin ice conditions prevail throughout the winter. By extrapolating this result to the entire Arctic Ocean, it is estimated that sea ice formation during winter can account for an uptake of atmospheric CO2 of around 50x 10''1''2 g C yr ''-''1. Further, it is speculated that changes in the areal extent of sea ice formation in the Arctic Ocean at the end of this century may triple this uptake. The mean winter time air-sea CO2 flux in the northern North Atlantic is found to be 0.1 G ton, with an interannual variability of about +- 10 % (paper IV). This was studied by using an empirical relationship between seawater fCO2 and SST. Gridded data of SST, sea level pressure, and wind speed were used in combination with data for atmospheric mole fraction of CO, to calculate the air-sea flux in the time period 1981 until 2001. Locally, and on a monthly time scale, the interannual variability is found to be higher, typically 20 - 40 %. Changes in wind speed and fCO2 in the atmosphere account for most of the interannual variations. (Author)