Part 1: Information About the Numeric Data Package
Surface Water and Atmospheric Carbon dioxide and Nitrous oxide Observations by Shipboard automated Gas Chromatography: Results from Expeditions between 1977 and 1990
R. F. Weiss, F. A. Van Woy, P. K. Salameh
Scripps Institution of Oceanography
University of Calfornia, San Diego
La Jolla, California
Carbon dioxide (CO2); gas chromatography; marine atmospheric concentrations; nitrous oxide (N2O); oceanography; surface seawater dissolved gasess.
The surface water and atmospheric carbon dioxide (CO2) and nitrous oxide (N2O) data reported here were obtained by direct shipboard gas chromatographic measurement. These data include results from 11 diferent oceanic surveys for a total of 41 expedition legs. The represented oceanic surveys include the following: (1) the Indomed expedition, 1977-1979 [Indomed legs 4 and 5 are also part of the Geochemical Ocean Sections (GEOSECS) Indian Ocean expedition]; (2) the North pacific Experiment (NORPAX) Hawaii-Tahiti Shuttle Experiment, 1979-1980; (3) and (4) the Transient Tracers in the Ocean, North Atlantic and Tropical Atlantic Studies (TTO/NAS, TTO/TAS), 1981-1983; (5) the Hudson 82-001 expedition, 1982; (6) the Ajax expedition, 1983-1984; (7) and (8) the Trans-pacific Secions expeditions along 24 degrees North and 47 degrees North (TPS24 and TPS47), 1985; (9) the fifth &qout;Antarktis&qout; expedition (Ant V) of the R/V Polarstern, as part of the Winter Weddell Sea Experiment, 1986; (10) the South Atlantic Ventilation Experiment (SAVE), 1987-1989 (SAVE legs 4-6 are also designated as legs 2-4 of R/V Melville's Hydros expedition); and (11) the 1990 expedition of the National Oceanic and Atmospheric Administration's Climate and Global Change series (CGC-90). Table 1 presents a track list showing the dates, ports of deprature and arrival, regions surveyed, and cruise ship names for each of the 41 expedition legs that contributed data. In addition, a series of maps showing the tracks of the expeditions and the N2O and CO2 results for each expedition leg is presented in Figs. 1-83.
this document describes the results of surface water and atmospheric CO2 and N2O measurements carried out by shipboard gas chromatography over the period of 1977-1990. The measurements were made by an automated high-precision shipboard gas chromatographic system developed during the late 1970s and used extensively over the intervening years. This instrument, which is described by Weiss (1981), measures CO2 by flame ionization after quantitative reaction to methane in a stream of hydrogen using a reduced nickel catalyst preceded by a palladium catalyst to protect the nickel from oxygen and other atmospheric oxidants. Nitrous oxide is measured by a separate electron capture detector. (Methane is also measured by the flame ionization detector, although the system is not optimized for this gas. The methane results are not included in this data set because methane's equlibration time constant is long, and the results are therefore subject to contamination by biological activity in the ship's seawater pumping system.)
The chromatographic system measures 196 dry-gas samples a day, divided equally among the atmosphere, gas equilibrated with surface water, a low-range gas standard, and a high-range gas standard. Thus, the atmosphere and the ocean are each measured every half-hour, or 48 times a day. This corresponds to a spatial resolution of about 10 km when the ship is under way and gives several replicate measurements at each hydrographic station. The typical relative standard deviation of a single determination is about 0.04% for CO2 and 0.3% for N2O, but precision is occasionally affected adversely by shipboard operating conditions. The measurements are calibrated with dry-air secondary standards stored in Spectra-Seal aluminum cylinders. these standards are periodically calibrated for CO2 against the Scripps Institution of Oceanography (SIO) manometric scale in the laboratory of C. D. Keeling and for N2O against the calibration scale developed by Weiss et al. (1981). In addition to its accuracy, the chromatographic method for CO2 offers the benefits over other commonly used techniques of being independent of oxygen concentration and using small amounts of sample and standard.
Surface seawater is pumped continuously from the bow of the ship (nominal depth approximately 3 m) at a rate of about 100 liters/min. this high pumping rate and the use of plastic polyvinylchloride (PVC) piping assure a minimal change in temperature and a minimal opportunity for chemical alteration of the water. The equilibrator is constructed of heavy acrylic plastic (for visibility and temperature insulation) and has an internal gas volume of about 20 liters. The equilibrator design consists of 2 concentric cylindrical stages, with a drain at the center to minimize volume changes as a result of ship motion. The water "rains" through the 2 stages of the equilibrator at a combined rate of about 20 liters/min, and a low 0.2-atm pressure head minimizes spraying, bubble entrapment, and other dynamic pressure effects. The 20-liter gas space is circulated by an air pump through a closed loop which provides the pressurized gas required by the chromatograph. Each half-hourly analysis removes about 75 ml of gas from this pumped loop. The first stage of the quilibrator is vented to the outside atmosphere so that the gas used for the analysis is replaced by clean marine air.
The temperature of the equilibrator is monitored and compared with the surface ocean temperatures measured at hydrographic stations to determine the thermal effect of the ship's pumping system as a function of intake temperature. The maximum amount of change, found for the coldest surface waters, is typically a warming of <1°C. As expected, this temperature difference decreases to zero when the water temperature reaches the mean inside temperature of the ship. The measured CO2 values are corrected for this thermal effect (roughly 4% per degree) using an empirical equation (Weiss et al. 1982) which is dominated by the temperature dependence of the CO2 solubility coefficient (Weiss 1974). The measured N2O values are also corrected for the solubility effect (Weiss and Price 1980).
The response time of the equilibrator has been evaluated theoretically and experimentally. For unbuffered gases such as nitrous oxide, oxygen and nitrogen, the theoretical response time (assuming complete exchange betweenwater and gas) is given as FS/V, where F is the flux of water through the system, V is the volume of the equilibrated gas phase, and S is the Oswald solubility coefficient. For the equilibrator used in the measurements presented here, this gives a characteristic (1/e) response time of about 1 min for N2O, about 0.5 hr for oxygen and about 1 hr for nitrogen. For CO2 the response time would be similar to N2O if there were no chemical bufering, but with chemical buffering (see gas exchange discussion in Broecker and Peng 1982) the response time is enhanced by an order of magnitude to about 0.1 min. Laboratory experiments by Weiss et al. (1982) and by scientists at the National Oceanic and Atmospheric Administration (NOAA), Pacific Marine Environmental Laboratory (PMEL) and Climate Monitoring and Diagnostics Laboratory (CMDL), who have adopted this equilibrator design, have confirmed that the actual equilibration times are close to these theoretical values.
These differences in exchange times are important in understanding the performance of an equiibrator that is vented to atmospheric pressure. Since the major components of equilibrated gas -- nitrogen, oxygen, and argon -- have equilibration times that are much longer than those of the measured species, N2O and CO2 the effect will be for the equilibrium partial pressures of these latter two gases to be prevent in a water-saturated gas phase at a total pressure equal to the barometric pressure. This is exactly the condition that is satisfied by the actual atmosphere when it is in equilibrium with the ocean, since the gas-phase boundary layer is always saturated with water vapor and at the total barometric pressure. Since the chromatographic system measures the dry-gas mole fractions of these constituents, xCO2 and xN2O, in both the atmosphere and the equilibrated gas, and the total pressure is the same in both cases, the differences in xCO2 between these phases are a close measure of the differences in CO2 partial pressure (pCO2), as long as the total pressure is near 1 atm:
delta(pCO2) = delta(xCO2)(P - pH2O),
where delta signifies the difference between sea and air, P is the barometric pressure, and pH2O is the water vapor pressure. Because the temperature and the barometric pressure are routinely recorded, the system is effectively completely constrained, but even without these variables delta(xCO2) is a very good approximation of delta(pCO2). The argument is, of course, the same for the partial pressure of N2O.
Another question which is more difficult to answer is whether the equilibrium reached by the equilibrator is the true thermodynamic equilibrium that we wish to measure. This type of question is always very dificult to answer, but there is indirect evidence that it is very close. The discrete equilibrator pCO2 measurements carried out by T. Takahashi's group at Lamont-Doherty Geological Observatory during the SAVE expeditions have shown agreement with the equilibrator values presented here to within 1 or 2 ppm (T. Takahashi, personal communication), even though their measurements are made at a fixed temperature and must be corrected to the surface water temperature. Also, the measurements of pN2O in the central gyres of the major oceans presented here are generally within 1% of atmospheric saturation. If this were not the correct equilibrium value, one could not explain the constancy of these values over many thousands of kilometers in many different central gyres. Through continued use of the same equilibrator design during the World Ocean Circulation Experiment (WOCE), it is hoped to obtain further verification that the measurements are being made at true equilibrium through comparisons with the discrete pCO2 and carbon system measurements being carried out by the other laboratories.
Concentrations of CO2 and N2O are calculated by fitting detector peak area response to a quadratic polynomial forced through the origin (zeroth order term is zero). The calculation is performed with the assumption that the linearity of the detector varies slowly compared with changes in detector sensitivity. Accordingly, the second order term of the quadratic polynomial (linearity of the detector) is dteremined from a running mean of the high standard to low standard response ratio over a range of plus and minus 20 runs, and the first order term (sensitivity of the detector) is determined from the immediately bracketing high and low standard runs. Further details concerning the methods of sample measurement and analysis are provided in Weiss (1981), a copy of which is provided in the appendix of this document.
The magnetic tape (or floppy diskettes) that accompanies this document includes a descriptive information file (File 1 on the magnetic tape or NDP044.DES on the floppy diskettes), a file (File 2 on the magnetic tape or TRACK.LST on the floppy diskettes) containing a list of the expedition legs on which measurements were made, and a file (File 3 on the magnetic tape or DATA.LST on the floppy diskettes) containing a list of the corresponding data filenames. The tape or diskettes also contain two data files for each expedition leg: one file containing the atmospheric results and one containing the surface seawater results for xCO2 and xN2O [in parts per million (ppm) and parts per billion (ppb), respectively].
The data in this package constitute one of the most extensive records available of xCO2 and, particularly, xN2O in marine air and surface seawater. These data will be useful in modeling applications dealing with the oceans's role in the global biogeochemical cycles of carbon and nitrogen. The combination of atmoshperic and surface seawater sampling represented in these data should also make them useful in studies of ocean-atmosphere dynamics. In addition, since determinations of pCO2 in the past were usually derived indirectly, these shipboard gas chromatographic analyses are especially valuable in that they represent direct measurements of seawater CO2 and will be useful in studies evaluating other methodologies for determining pCO2.
The locations, surface water temperatures, and barometric pressures presented in the surface water and atmospheric CO2 and N2O data set are all interpolated from the discrete values recorded on the ship, and therefore must be taken only as approximations. As noted in the list of expeditions presented in Table 1, only N2O was measured on these first four expedition legs, but this should not be a significant impediment to the use of the data for most applications, as was discussed in the methodology seciton (Section 5). The reader should note that measurements on NORPAX leg 7 and before were made using anhydrous calcium sulfate drying agent to dry the measured samples, which may bias the CO2 results slightly due to acid-base reactions. This problem is discussed by Weiss (1981) and affects only these early legs. During the TPS-24 and TPS-47 expeditions, nitrogen carrier gas was used instead of argon-methane carrier gas for the N2O determinations. Subsequent comparisons with flask atmospheric samples suggested that this may have produced a small bias of 1 or 2 ppb in the N2O results from these legs, but no corrections have been applied.
The primary purpose of these data is to describe large-scale distributions of CO2 and N2O. The times of the measurements are accurate to within 1 min, but the ancillary data such as position, sea surface temperature, and barometric pressure were interpolated from observations of extremely variable frequency and accuracy. The uncertainties resulting from these interpolations do not significantly increase the errors in the CO2 and N2O results. ancillary data are reported for the sake of completeness but should not be used in their own right without a more thorough investigation of measurement and interpolation erors.
The Carbon Dioxide Information Analysis Center (CDIAC) endeavors to provide quality assurance (QA) of all data before their distribution. To ensure the highest possible quality in the data, CDIAC conducts extensive reviews for resonableness, accuracy, completeness, and consistency of form. Although the reviews have common objectives, the specific form must be tailored to each data set; this tailoring process may involve considerable programming efforts. The entire AQ process is an improtant part of CDIACS's effort to ensure accurate, usable CO2-related data for researchers.
It is important to emphasize that the data were edited by the authors before submission to CDIAC to remove serious outliers and contaminated samples and to correct gross numerical errors. However, not all the data have yet been subjected to the level of scrutiny associated with careful interpretive work. Readers are therefore requested to draw to the attention of the authors any suspected inconsistencies in these data. Readers who have obtained this report directly from CDIAC will automatically receive notification of updates and corrections. The authors wish to encourage scientific collaborations with readers for the purpose of interpreting the results of these observations.
The following summarizes the QA checks performed on the surface water and atmospheric CO2 and N2O data by CDIAC.
- The format of all informtion, includeing header items, was checked to ensure consistency throughout each data file.
- All numeric values were inspected for logical inconsistencies (e.g., values of DATEDA <1 or >31; YEAR <1977 or >1990; TIME <0 or >2359 LAT <-90.0 or > 90.0; LON <-180.0 or >180.0) and for the presence of outliers (e.g., PRESS <900 or >1100; H2OTMP <-5 or >32)
The data distributed in this package are identical to the original data received by CDIAC. However, in order to enhance the ease of use of these data, the following alterations in format were made.
- The data filenames were modified to conform to the two-level naming convention of DOS-base systems; for example, 05.INDOMED.LEG11A.WATER was changed to INDOM11A.H2O. A complete listing of the filenames and the corresponding expedition legs is given in Section 11 of this document.
- Within each data file, all header material was condensed into a single line. This involved removing blank lines and abbreviating the designations for sample type (i.e., atmposhperic data or surface seawater data). In addition, all descriptive column titles were removed.
- Values of latitude and longitude were converted from degrees and minutes to decimal degrees, and signs were added to denote the hemisphere (Northern and Eastern Hemispheres were assigned "+" values; Southern and Western Hemispheres were assigned &qout;-&qout; values).
- The designation for missing values, given as blanks in the orignal files, were chagned to the following: -999.9 for missing values of barometric pressure; 99.99 for missing values of surface water temperature; and -99.9 for missing values of xN2O and xCO2.
- Broecker, W. S., and T.-H. Peng. 1982. Tracers in the Sea. Eldigio Press, Lamont-Doherty Geological Observatory of Columbia University, Palisades, New York.
- Weiss, R. F. 1974. Carbon dioxide in water and seawater: The solubility of a non-ideal gas. Marine Chemistry 2:203-215.
- Weiss, R. F., and B. A. Price. 1980. Nitrous oxide solubility in water and seawater. Marine Chemistry 8:347-359.
- Weiss, R. F. 1981. Determinations of carbon dioxide and methane by dual catalyst flame ionization chromatography and nitrous oxide by electron capture chromatography. Journal of Chromatographic Science 19:611-616.
- Weiss, R. F., C. D. Keeling, and H. Craig. 1981. The determination of tropospheric nitrous oxide. Journal of Geophysical Research 86:7197-7202.
- Weiss, R. F., R. A. Jahnke, and C. D. Keeling. 1982. Seasonal effects of temperature and salinity on the partial pressure of carbon dioxide in seawater. Nature 300:511-513.
this document describes a data set consisting of surface water and atmospheric CO2 and N2O measurements carried out by shipboard gas chromatography over the period 1977-1990. The data are available witout charge upon request on nine-track magnetic tape, on floppy diskettes (IBM PC format, high- or low-density, 5.25- or 3.5-in. diskettes), or through File Transfer Protocol (FTP) from CDIAC. Requests for magnetic tapes should include any specific instructions for transmitting the data as required by the user to access the data. Requests not accompanied by specific instructions will be filled on nine-track, 6250 BPI, standard=labeled tapes with characters written in Extended Binary Codes Decimal Interchange Code (EBCDIC), and files will be formatted as gien in Section 11. Requests should be addressed to the following:
Carbon Dioxide Information Analysis Center
Oak Ridge National Laboratory
Post Office Box 2008
Oak Ridge, Tennessee 37831-6290
The tape and documentation can be ordered by telephone, fax, or through electronic mail.
Telephone: (615) 574-0390
Fax: (615) 574-2232