Measurement of CO₂ in the Atmosphere and in Surface Seawater

Carbon dioxide measurements in the atmosphere and in surface water were made on these cruises by Kimberly Kelly. Mixing ratios of CO₂ were analyzed with an automated, temperature-controlled gas chromatographic system similar to that described by Weiss (1981), and Bates et al. (1993). Every 12 minutes one of four gases (atmospheric sample, standard 1, equilibrator vapor sample, or standard 2) was injected for analysis. Details of sample collection, gas standard calibration, and data reduction follow.

Air was pumped through 3/8-inch O.D. plastic-coated aluminum tubing (Dekoron) from the jackstaff on the bow of the ship, 10 meters above the sea surface, to the oceanographic laboratory for analysis. This line was continuously flushed at 8-10 L/min. Aliquots from this air stream were dried using phosphorous pentoxide and 2 milliliters were injected into a gas chromatograph (GC) where CO₂ was separated from the other gases, catalytically reduced to methane, and routed to the flame ionization detector. The instrument response was compared with the response of standard gases for quantification. The standard gases were dried and treated in exactly the same manner as the air and equilibrator samples.

Surface seawater entered the ship's seachest via a forward intake line located approximately 5 meters below the water level. From the seachest, the water was pumped up to a showerhead-type Plexiglas equilibrator designed by R. Weiss (Butler et al. 1988). Water rains through the equilibrator at 15-20 L/min, and the gases dissolved in the seawater partition between the aqueous and vapor phases according to their solubilities at the temperature and salinity of the seawater in the equilibrator. Aliquots of the equilibrator vapor phase were sampled, dried, and analyzed as described above for atmospheric samples.

The gas standards were dried, whole-air mixtures contained in aluminum cylinders. The working standards were calibrated against the primary standards before and after the 1987, 1988, and 1989 cruises. The primary standards (CC48232 - 352.47 ppm, CC48302 - 355.97 ppm, CC48314 - 353.39 ppm) were filled and calibrated by the National Oceanic and Atmospheric Administration/Climate Monitoring and Diagnostics Laboratory (NOAA/CMDL) in February 1987. Subsequent calibrations of these primary standards by CMDL in November 1987, June 1990, and August 1990 have shown that the CO₂ mixing ratios remained constant during this 3.5 year period. The CO₂ mixing ratios in the working standards ranged from 228.15 ppm to 460.43 ppm. The working standards were calibrated by CMDL in 1990. These calibrations agreed to within 0.5% of the values assigned to each tank by PMEL based on the primary standards and a linear FID detector response. The data reported here have been reduced using the CMDL calibrations of the working standards. The mixing ratios reported here are in the World Meteorological Organization (WMO) X85 scale (Thoning et al. 1987). The accuracies of the CO₂ standards are given by CMDL as ± 0.6% by comparison with the manometric analyses of C. D. Keeling at the Scripps Institution of Oceanography. Instrument precision was determined by the average percent standard deviation of the standard response over a 6-hour period. The precision varied between cruises but was always less than 1.2% and was more typically around 0.4% (Table 3).

Table 3. Instrument precision in percent standard deviation from the standard value.

The CO₂ mixing ratios in the air and equilibrator samples were calculated as follows. The data were first visually filtered to eliminate any episodes of ship contamination. These episodes were quite evident from the extremely high carbon monoxide values that were measured simultaneously. The mixing ratios of CO₂ in both the air and equilibrator samples were then computed based on peak area and either a 4-hour or 6-hour running mean single-point standard. These dry-air mixing ratios were then binned into hourly values based on the measurements made 30 minutes before and after the hour.