Variability of pCO(2) on diel to seasonal timescales in the Sargasso Sea near Bermuda

Continuous underway measurements of atmospheric and surface seawater pCO(2) were collected on numerous cruises in the Sargasso Sea (32 degrees N, 64 degrees W) near Bermuda from June 1994 to November 1995. We observed that seawater pCO(2) was highly variable on different timescales, ranging from diel to seasonal. On diel timescales, pCO(2) changes of 5-25 mu atm occurred in response to diurnal warming and cooling associated with solar heat fluxes. Over longer timescales, pCO(2) was influenced by atmospheric forcing and tropical cyclones. For example, a surface cooling of 3 degrees C and decrease in pCO(2) of 45-50 mu atm occurred after Hurricane Felix passed near Bermuda in August 1995. The decrease in pCO(2) was significant considering the annual change was 90-100 mu atm. Over all timescales, temperature was the dominant control on pCO(2) variability. We found that surface pCO(2) conditions were accurately predicted from temperatures with small errors (4-9 mu atm) if seasonal pCO(2)-temperature relationships were established. In future synthesis of regional p CO2 data it should be feasible to use surface temperature, remotely sensed from space, as a tool for extrapolation over wider spatial scales in the North Atlantic subtropical gyre. Net annual fluxes of CO2 for 1994 and 1995 (-0.25 to -0.6 mol CO2 m(-2) yr(-1)) were directed from atmosphere to ocean and were similar to values reported for 1989-1993 by Bares et al. [1996b]. We found that short-term variability of pCO(2) (diel warming and cooling or atmospheric forcing), frequency of sampling (every 3-4 days or monthly), or use of temperature-derived pCO(2) did not affect estimates of net yearly CO2 fluxes by more than 10-20%. However, strong winds associated with hurricanes decreased the net annual flux of CO2 into the ocean by 19-28% in 1995. The major sources of error for air-sea gas exchange was uncertainty associated with gas transfer-wind speed relationships and differences in the types of wind speed data used (daily averaged versus climatological). Such uncertainties make it difficult to quantify the contribution of gas exchange to the carbon cycle and the balance of carbon import and export terms in the upper ocean of the Sargasso Sea.