Methods of Computation

The seawater pCO2 data listed in LDEO database are based on direct measurements of seawater pCO2 made using equilibrator-CO2 analyzer systems. Many of the data listed are from semi-continuous pCO2 systems with flow-through water; many others are measurements for discrete water samples made at hydrographic stations along with measurements for other chemical and physical properties. Although different types of equilibrators (e. g., shower type, bubbling type, membrane type, and rotating disks in flow-through or closed circulation systems) and CO2 gas analyzers (non-dispersive infrared analyzers and gas chromatographs of various designs) were employed, the results from different systems are accepted as long as analyzers were properly calibrated using validated CO2-air gas mixtures and the carrier gas was equilibrated with seawater samples. Because of the diversity of methods used, it is impossible to present details of the method used by each contributing research group. Detailed methodology may be obtained directly from the investigators listed in Table 1, or from the CDIAC reports for specific expeditions.

It is important to point out that the methods used for computing CO2 concentrations in equilibrated gas have been varied among groups. For example, some groups computed a least-square fit of readings for three or more standard gas mixtures to a quadratic equation and used it to calculate concentrations in samples. Some groups used four or five standard gas mixtures for calibrations, and fitted the data to a 4th order polynomial equation. And some groups used an output from linearization circuits of infrared analyzer, and linearly regressed three or more standard gas readings to obtain sample CO2 concentrations. The outputs from a gas chromatograph are a linear function of CO2 concentration, and hence a linear regression was used for calibration. These different data reduction methods yield CO2 concentrations varying ± 1.5 ppm (or ± 1.5 µatm in pCO2). However, we did not recompute the CO2 values using a single uniform algorithm but instead, accepted CO2 concentration values as reported to us. Since different analyzers and different numbers of standard gas mixtures were used by respective groups, no single uniform data reduction scheme can be applied, and hence we relied on the judgment of each group for selecting the data reduction scheme most suited for their operational modes and skills. Measurements that were made using only one calibration gas mixture were judged unreliable and were not included in this database.

Using the reported CO2 concentration values, the pCO2 value in sample seawater at the equilibration temperature, (pCO2)eq, has been recomputed with the relationship:

(pCO2)eq = Vco2 (Peq - Pwater),

where Vco2 is the mole fraction concentration of CO2 in carrier gas (Vco2 is same as Xco2, which is often used in literature, and these qualities may be used interchangeably); Peq is the total pressure of gas in the equilibrator; and Pwater is the equilibrium water vapor pressure at temperature of equilibration, Teq, and salinity. Since some equilibrators were operated open to the room air, Peq values may be equal to the ship's interior pressure or to the barometric pressure outside the ship depending on the location of the equilibrator. When an equilibrator is located in an enclosed shipboard laboratory and is open to the room air, Peq is the ambient pressure in the laboratory. While an equilibrator operated in an enclosed space, only the barometric pressure at sea surface was reported in some data sets, but not Peq. In such cases, Peq is assumed to be the reported barometric pressure at sea surface plus 3 mb, that represents an overpressure normally maintained inside a ship. This correction increases the (pCO2)sw value by about 1 µatm. When the pressure was not reported, we used the climatological value in the nearest box from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) Reanalysis II Project file for the month of the observation. The pCO2 at in-situ seawater temperature is computed using an integrated form of the temperature effect for isochemical seawater, (δln pCO2/δT)Sal, Alk, TCO2 (Takahashi et al., 1993):

(pCO2)sw @Tin situ = [(pCO2)sw @ Teq] Exp{0.0433 (Tin situ – Teq) – 4.35 x 10-5 [(Tin situ)2 – (Teq)2 ] },

where the "sw" and "eq" indicate the in situ and equilibrator conditions respectively. Throughout the computation, CO2 gas is assumed to behave as an ideal gas that mixes with air and water vapor ideally. Although CO2 fugacity is used in a number of published papers and data reports, we refrained from using the fugacity since it is computed differently from an investigator to another. Although we do not list the sea-air pCO2 differences in this report, we recommend the formula below for the computation of atmospheric pCO2 and the corresponding value for sea-air pCO2 difference.

(pCO2)air = (VCO2)air (Pbaro – Psw),

where Pbaro is the barometric pressure at sea surface, and Psw is the equilibrium water vapor pressure at the temperature and salinity for mixed layer water. The subscript "air" indicates the value for atmosphere samples.

The sea-air pCO2 difference, ΔpCO2, is then computed using:

ΔpCO2 = (pCO2)sw – (pCO2)air.

Since CO2 is assumed to be an ideal gas for both (pCO2)sw and (pCO2)air, the small effects of non-ideality should cancel due to differencing for pCO2. Positive ΔpCO2 values indicate that the sea is a source for atmospheric CO2, whereas negative values indicate that the sea is a sink.

Values for the fugacity of CO2 in seawater, fCO2, have been submitted to us by some investigators. However, often the fugacity is not clearly defined as to whether only the non-ideality arising from CO2-CO2 molecular interactions is considered and/or that from CO2-H2O interactions is also included. Because of these ambiguities, we have chosen not to list fCO2 values. Since the mole fraction concentrations of CO2 in equilibrated gas samples are also included in the reported data, we have computed pCO2 using the reported temperature, pressure, and other data and listed in this report using the ideal gas law as explained earlier. Since fCO2 values are always smaller than the corresponding pCO2 values by 1 to 2 µatm and the differences are large enough with respect to the precision of measurements and the mean global sea-air pCO2 difference of about 10 µatm, they should not be used interchangeably with pCO2.

Beginning with this version (V2007) we have added a column reporting the partial pressure of CO2 in seawater in units of Pascals.


CDIAC   |  CCSI   |  ESD   |   ORNL   |   Security   |   Contact Us   |   maintained by Alex Kozyr   |   Last modified: 18 Aug 2010