[1] Measurements of atmospheric O 2 /N 2 ratios and CO 2 concentrations can be combined into a tracer known as atmospheric potential oxygen (APO % O 2 /N 2 + CO 2 ) that is conservative with respect to terrestrial biological activity. Consequently, APO reflects primarily ocean biogeochemistry and atmospheric circulation. Building on the work of Stephens et al. (1998), we present a set of APO observations for the years 1996-2003 with unprecedented spatial coverage. Combining data from the Princeton and Scripps air sampling programs, the data set includes new observations collected from ships in the low-latitude Pacific. The data show a smaller interhemispheric APO gradient than was observed in past studies, and different structure within the hemispheres. These differences appear to be due primarily to real changes in the APO field over time. The data also show a significant maximum in APO near the equator. Following the approach of Gruber et al. (2001), we compare these observations with predictions of APO generated from ocean O 2 and CO 2 flux fields and forward models of atmospheric transport. Our model predictions differ from those of earlier modeling studies, reflecting primarily the choice of atmospheric transport model (TM3 in this study). The model predictions show generally good agreement with the observations, matching the size of the interhemispheric gradient, the approximate amplitude and extent of the equatorial maximum, and the amplitude and phasing of the seasonal APO cycle at most stations. Room for improvement remains. The agreement in the interhemispheric gradient appears to be coincidental; over the last decade, the true APO gradient has evolved to a value that is consistent with our time-independent model. In addition, the equatorial maximum is somewhat more pronounced in the data than the model. This may be due to overly vigorous model transport, or insufficient spatial resolution in the air-sea fluxes used in our modeling effort. Finally, the seasonal cycles predicted by the model of atmospheric transport show evidence of an excessive seasonal rectifier in the Aleutian Islands and smaller problems elsewhere.

A B S T R A C T The atmospheric Ar/N 2 ratio is expected to undergo very slight variations due to exchanges of Ar and N 2 across the air-sea interface, driven by ocean solubility changes. Observations of these variations may provide useful constraints on large-scale fluxes of heat across the air-sea interface. A mass spectrometer system is described that incorporates a magnet with a wide exit face, allowing a large mass spread, and incorporates an inlet with rapid (5 s) switching of sources gases through a single capillary, thus achieving high precision in the comparison of sample and reference gases. The system allows simultaneous measurement of Ar/N 2 , O 2 /N 2 and CO 2 /N 2 ratios. The system achieves a short-term precision in Ar/N 2 of 10 per meg for a 10 s integration, which can be averaged to achieve an internal precision of a few per meg in the comparison of reference gases. Results for Ar/N 2 are reported from flasks samples collected from nine stations in a north-to-south global network over about a 1 yr period. The imprecision on an individual flask, as estimated from replicate agreement, is ±11 per meg. This imprecision is dominated by real variability between samples at the time of analysis. Seasonal cycles are marginally resolved at the extra-tropical stations with amplitudes of 5 to 15 per meg. Annual-mean values are constant between stations to within ±5 per meg. The results are compared with a numerical simulation of the cycles and gradients in Ar/N 2 based on the TM2 tracer transport model in combination with air-sea Ar and N 2 fluxes derived from climatological air-sea heat fluxes. The possibility is suggested that Ar/N 2 ratios may be detectably enriched near the ground by gravimetric or thermal fractionation under conditions of strong surface inversions.

The atmospheric Ar/N 2 ratio is expected to undergo very slight variations due to exchanges of Ar and N 2 across the air-sea interface, driven by ocean solubility changes. Observations of these variations may provide useful constraints on large-scale fluxes of heat across the air-sea interface. A mass spectrometer system is described that incorporates a magnet with a wide exit face, allowing a large mass spread, and incorporates an inlet with rapid (5 s) switching of sources gases through a single capillary, thus achieving high precision in the comparison of sample and reference gases. The system allows simultaneous measurement of Ar/N 2 , O 2 /N 2 and CO 2 /N 2 ratios. The system achieves a short-term precision in Ar/N 2 of 10 per meg for a 10 s integration, which can be averaged to achieve an internal precision of a few per meg in the comparison of reference gases. Results for Ar/N 2 are reported from flasks samples collected from nine stations in a north-to-south global network over about a 1 yr period. The imprecision on an individual flask, as estimated from replicate agreement, is ±11 per meg. This imprecision is dominated by real variability between samples at the time of analysis. Seasonal cycles are marginally resolved at the extra-tropical stations with amplitudes of 5 to 15 per meg. Annual-mean values are constant between stations to within ±5 per meg. The results are compared with a numerical simulation of the cycles and gradients in Ar/N 2 based on the TM2 tracer transport model in combination with air-sea Ar and N 2 fluxes derived from climatological air-sea heat fluxes. The possibility is suggested that Ar/N 2 ratios may be detectably enriched near the ground by gravimetric or thermal fractionation under conditions of strong surface inversions.

No abstract