2. Analytical Methods
Several significant advances in the chemical analysis of inorganic carbon system parameters occurred immediately prior to and during the inorganic carbon survey on the WOCE/WHP and OACES cruises. Many of the DIC measurements were performed by using a coulometer, often connected to a single-operator multiparameter metabolic analyzer (SOMMA), which has greater precision than previous techniques. For the U.S. cruises and many of the (later) foreign cruises, standard operating protocols were applied following the techniques outlined in the Handbook of Methods for the Analysis of the Various Parameters of the Carbon Dioxide System in Sea Water (DOE 1994). Certified reference materials (CRMs) were used on all U.S. cruises and on many non-U.S. cruises as secondary standards for DIC analyses. Discussion of the preparation and use of CRMs for DIC and TAlk is presented in detail in Dickson, Afgan, and Anderson (accepted) and Dickson, Anderson, and Afgan (submitted) (see also http://andrew.ucsd.edu/co2qc/). The CRMs consisted of natural seawater that was filtered, sterilized, and poisoned. Certification of the CRMs for DIC is based on manometric analyses in the shore-based laboratory of C. D. Keeling of Scripps Institution of Oceanography (SIO) over a period of several years. The CRMs were bottled in large batches into 500-mL borosilicate glass containers, sealed to prevent contamination, and shipped to the users. They were then analyzed at sea interspersed with samples over the course of each of the cruises as a way of verifying accuracy. Because CRMs were routinely analyzed for DIC during most cruises, groups analyzing TAlk on those cruises often measured CRMs for their alkalinity content. This enabled post-cruise corrections to be made to TAlk data based on archived CRMs that were analyzed at C. D. Keeling′s laboratory and at the laboratory of F. Millero in Miami.
Insufficient metadata precluded a determination of which cruises had adjustments applied for TAlk based on CRM analyses. CRMs were not available for the other two carbon parameters discussed in this report (CO2 and pH). However, for some of the cruises the CRMs were used to normalize the pH values by calculating the pH of the CRMs from the certified TAlk and DIC values. Analyses of salinity, nutrients, and O2 followed WOCE WHP protocol (WOCE 1994).
Analytical procedures differed between cruises. SOMMAs were used for all of the cruises that had U.S. investigators performing the DIC measurements (Johnson et al. 1993). For non-U.S. investigators, other procedures were used, ranging from use of coulometers with manual or automated pipettes to analysis by gas chromatography (Table 2). For some cruises, DIC was determined from TAlk and pH. The TAlk was measured by potentiometric titration logging of either a full titration curve (referred to as "Full" in Table 2) (Millero et al. 1993) or single-point titrations ("1-point") (Perez and Fraga 1987). For cruises A6 and A7, we were unable to find the appropriate metadata describing how the alkalinities were measured. The pH measurements were done by electrode, spectrophotometric, or potentiometric determination (Byrne and Breland 1989). Discrete CO2 was measured by equilibration of a discrete sample with a headspace with known initial concentration and by subsequent IR analysis (Wanninkhof and Thoning 1993) or gas chromatography (Chipman, Marra, and Takahashi 1993, Neil et al. 1997).