Discrete pCO₂ Measurements

Samples for discrete pCO₂ analyses were collected by overfilling 60-mL precalibrated serum bottles in the same manner as for oxygen and TCO₂. During Section P6 and following the static-headspace procedure of Johnson et al. (1990) for CH₄, a plastic pipette tip was inserted into the bottles making a water-tight seal at the bottle mouth. Then the bottles were inverted so that the volume of water displaced into the pipette was decanted. Next, the pipette was quickly withdrawn and the bottles were crimp-sealed leaving a headspace (gas phase) volume in each bottle of (nominally) 5 mL and a liquid phase of (nominally) 55 mL, constituting a closed or static system. The pipette method yields a highly reproducible headspace volume, and the headspace and water volume for each (numbered) serum bottle was determined gravimetrically prior to the cruise. The bottles were prepared and sealed outside on deck at the CTD site usually within 1 minute of collection. The atmospheric pressure was measured just prior to sealing so that the pre-equilibrated serum-bottle gas phase contained air at a known mixing ratio of CO₂ (determined regularly throughout the cruise) at a known total pressure (P). The initial liquid phase temperature was taken to be the potential temperature (T) of the sample. The bottles were laid in a thermostatted shaking water bath and equilibrated, by shaking, for 3 h at 20°C. After equilibration, the serum-bottle gas phase was displaced by a brine solution to flush and fill a gas sample loop whose contents were analyzed by gas chromatography. The mole fraction of CO₂ in the gas phase (xCO_2eq) was determined after the catalytic conversion of CO₂ to CH₄ with a flame ionization detector through comparison with a calibration curve based on CO₂ in air standards. These standards were subsequently intercalibrated with standards maintained by Taro Takahashi and Dave Chipman at the Lamont-Doherty Earth Observatory (LDEO).

Erroneously, in the original work (Johnson et al. 1990), no provision was made to measure the total gas phase pressure in the serum bottle after equilibration (P_eq). Because P_eq was not measured during WOCE Section P6, it had to be estimated. This was done by first calculating the moles of N₂, O₂, and Ar in the liquid phase prior to equilibration using potential temperature, the measured O₂ concentrations, and by assuming that each water sample was saturated at the surface with N₂ and Ar with moist air at 1 atm at the potential temperature of the sample. Next, the total number of moles of each gas in the introduced gas phase was phase was calculated. Hence the total number of moles of each gas present in the closed system (serum-bottle) was known. After equilibration, a small correction for glass expansion and the phase ratio volume change caused by the change in temperature during equilibration (usually warming) was applied. The partial pressure of each gas at the equilibration temperature (20°C) was then calculated from the total number of moles for each gas, and these gas partial pressures along with the equilibrium partial pressure of water vapor were summed to give P_eq in the headspace after equilibration. Then xCO_2eq and P_eq were multiplied to convert xCO _2eq to pCO₂ hereafter called pCO_2eq. Subsequent laboratory tests (C. Neill and D. Wallace, unpublished data) confirmed that the serum bottles were not subject to leakage and that the predicted pressure closely matched the actual headspace pressure. The close correspondence between measured and predicted headspace pressure has also been confirmed during extensive field tests (see Neill et al. 1997).

TCO₂ was measured on an unequilibrated duplicate sample, and the TCO₂ of the liquid phase after equilibration (repartioning of CO₂ between the gas and liquid phases) was calculated using a mass balance approach (hereafter designated TCO_2eq ). The carbonate alkalinity (CA) of the equilibrated sample was calculated using pCO_2eq and TCO_2eq with the thermodynamic constants of Roy et al. (1993) and software developed by Lewis and Wallace (1998). Because carbonate alkalinity is conserved during the equilibration, the derived CA (µmol/kg) is the in situ value prior to equilibration. Hence both the in situ TCO₂ (measured independently by coulometry) and CA are known for each sample prior to the equilibration, and these two parameters were used to calculate the sample in situ pCO₂ at the equilibration temperature using the Lewis and Wallace (1998) software and the Roy et al. (1993) constants. The pCO₂ in µatm is reported at the equilibration temperature and the equilibration temperature is also reported. Subsequently, the nutrient data became available and Total Alkalinity (TALK) was also calculated for each sample according to DOE procedures (1994) using the software given by Lewis and Wallace (1998). TALK values are not reported in this NDP.

The precision of the pCO₂ determination and the TALK calculated from pCO₂ and TCO₂ was assessed, when possible, according to the same procedures used for TCO₂. The precision of the pCO₂ determination and the derived TALK is given in Table 11 as follows: