Description of Variables and Methods

Each station consisted of the lowering of a CTD/rosette package, upon which 10-L standard Niskin^® bottles were suspended. Pressure, temperature, and CTD salinities for the tabulated data were taken from the calibrated CTD data. The discrete water samples were drawn from the Niskin^® bottles upon retrieval of the rosette on deck. Samples were collected for analyses of oxygen, chlorofluorocarbons (CFCs), helium, tritium (Leg 1 only),TCO₂, ¹³C, pH (Leg 1 only), nutrients, and salinity. In addition, underway partial pressure of CO₂ (pCO₂) was measured throughout the cruise. This report addresses the temperature, salinity, TCO₂, and nitrate data.

CTD data were collected using a Neil Brown (Mark III) Instrument System. Pressure, temperature, and conductivity were recorded on the downtrace, and the discrete water samples were collected on the upcast with an electronically fired rosette sampler. The bottle salinity samples were analyzed with a Guideline Autosal, which was calibrated at the beginning of each day's run with a vial of Wormley standard seawater, and again after each case of samples was analyzed with another vial. An accuracy of 0.002 Practical Salinity Scale (PSS) and a precision of 0.001 PSS were achieved. A more detailed description of the CTD data processing has been published as a NOAA Data Report (McTaggart et al. 1993).

Nutrient analyses included nitrate, silicate, and phosphate. Due to accuracy and precision problems for silicate and phosphate, only nitrate data are included in this report.

Nitrate samples were collected from each Niskin^® bottle in aged 60-mL linear polyethylene bottles and analyzed for dissolved inorganic nitrate (NO₃--N). Analyses were performed on samples from all CTD casts with a five-channel Technicon Auto-Analyzer (AA-II) aboard ship. The analytical procedures and methodologies used in the analysis of nitrate are similar to those described by Armstrong et al. (1967), with modifications described in Atlas et al. (1971) and in a Technicon Corporation (1977) technical report. The detection limit for nitrate was 0.39 µmol/kg with a standard deviation of ±0.1 µmol/kg. The precision of duplicate measurements was ±0.25%, full scale. The accuracy was assumed to be 1% because no absolute standards were available. Nitrate measurements are reported in µmol/kg at 1 atm and an assumed laboratory temperature of 25^oC. Calibration standards ranged between 0 and 30 µmol/kg; therefore, only samples within that concentration range are reported in the data file.

Upon retrieval of the CTD-rosette package on deck, samples for TCO₂ were collected in 500-mL glass-stoppered bottles and poisoned with 0.5 mL of saturated HgCl₂ solution to decrease bacterial oxidation of organic matter prior to analysis. The samples were analyzed immediately when possible, but always within 24 h after collection.

The coulometric technique for TCO₂ analyses in seawater was originated by Ken Johnson (Johnson et al. 1985, 1987). UIC, Inc. supplied the coulometic/extraction system (Models 5011 and 5030, respectively). The following changes were made to the standard extraction system:

1. Both the pipet and seawater sample bottle were jacketed (the sample bottle was placed in a jacketed beaker) and were connected to a circulating bath set at 25^oC.
2. Ultra-pure N₂ was used for both carrier gas and sample delivery; prior to hook-up to glassware, the N₂ was sent through an in-line NaOH (Malcosorb) scrubber to remove any CO₂.
3. The condensing column was connected to a circulating bath filled with antifreeze and run at 1^oC.
4. An ORBO (activated Si gel from Supelco) tube was placed in-line between the glassware and titration cell to eliminate any excess moisture.
5. Standardization of instrumentation was modified (as described later in this section). The computer interfaced to the system was a Zenith ZBF-2339-BK.

The program supplied by UIC was modified significantly, to accommodate our particular needs. In coulometric analysis of TCO₂, all carbonate species in seawater (CO_2(aq), H₂CO₃, HCO₃^- and CO₃^2-) are converted to CO₂ by addition of excess acid. The evolved CO₂ is then moved into the titration cell by N₂ carrier gas where it is titrated potentiometrically by reacting quantitatively with ethanolamine to form hydroxyethyl carbamic acid; this is titrated with OH^- ions electrogenerated by the reduction of H₂O at a platinum cathode:

CO₂ + HO(CH₂)₂NH₂ --> HO(CH₂)₂NHCOO^- + H⁺ Ag(s) --> Ag⁺ + e^- H₂O + e^- --> 0.5H₂(g) + OH^- H⁺ + OH^- --> H₂0

The equivalence point is detected photometrically with thymolphthalein as an indicator. The cell solution is blue at the equivalence point of 10.5 pH and colorless at pH 9.3 after the addition of CO₂ in aqueous solutions. CO₂ drives down the pH and raises percent transmittance. As the acid is titrated, pH increases (hence, the blue color returns) and percent transmittance decreases, thus causing the titration current to pass from high to low to zero as the equivalence point is approached and sensed by the optical detector. The CO₂ level is calculated based on the quantity of electricity required to reach the equivalence point and the time of passage. The entire sequence takes from 8 to 11 min.

The volume of the pipet was ~50 mL, and was calibrated in the laboratory before and after the cruise. The pipet was cleaned by drawing a 25% solution of NaOH into the cell and allowing it to soak overnight. This eliminated any organic film inside the pipet and ensured a clean delivery. Pipet calibrations were conducted to be within the measured pipet temperature range during the cruise (24.5^oC-25.5^oC) utilizing a circulating bath. Milli-Q water was drawn into the pipet in the exact manner that a liquid standard or seawater sample was handled. The water was delivered by N₂ (flow rate is 200 mL/min) into a tared, ground-glass stoppered mixing flask, and drained for an additional 5 s (monitored by a stop-watch) to allow the droplets of water to be delivered; the flask was then immediately stoppered and weighed on a Mettler AE240 balance. Approximately 15-20 samples were collected per experiment. The following references were used to correct the weighings:

1. Volume Properties of Ordinary Water.
2. Reductions of Weighings in Air to Vacuo for Brass Weights and a Water Density of 1.00. Density of air used was 0.0012.
3. Temperature Correction for Glass Volumetric Apparatus.

The corrected volumes were then linearly regressed with temperature, and a calibration curve was established (typical r² = 0.80).

Schott-Duran glass 500-mL bottles were annealed at 450^oC for 1 h, then cleaned in a dishwasher with commercial grade dishwashing detergent. Prior to collection of samples, the solid ground-glass stoppers were coated with Type M Apiezon grease.

Acid used to convert carbonate species to CO₂ was a 1:10 solution of Baker reagent grade H₃PO₄. All coulometric chemicals (cathode solution, anode solution, and KI) were purchased from UIC, Inc.

Liquid standards were made up in a 0.7 M solution of KCl. The standards were treated just as a seawater sample and were delivered through the pipet under the same conditions. The standard used was Na₂CO₃ (Ultrex, Lot 935113); the KCl was reagent grade from Mallinkrodt.

The Na₂CO₃ was prepared in the laboratory by baking at 260-270^oC for 0.5 h, and desiccated overnight. The standards were weighed into pre-cleaned ground-glass stoppered vials with weights ranging from 0.20 to 0.25 g. They were immediately stored in an evacuated desiccator with fresh Si gel until prepared. The KCl was baked in a muffle furnace for 0.5 h at 260-270^oC, and cooled in a desiccator overnight.

The KCl solution and liquid standards were prepared in the following manner: Milli-Q water was boiled in a 3-L boiling flask for 20 min to drive off CO₂, then cooled overnight with a NaOH column attached to the neck of the flask. A glove-box was purged with ultra-pure N₂ for about 20 min; in the glove box, the KCl was mixed with the CO₂-free water in a clean 2-L volumetric. Half of this solution was stored in a 1-L sample bottle with siphon tube and clamp and was used to determine blank values for the KCl (see following discussion). The other half was used to make the Na₂CO₃ in a 1-L volumetric. After the standard equilibrated, it was poured into a 1-L bottle with siphon tube and clamp. This work was performed in an N₂ environment in the glove box.

The KCl solution was analyzed to determine a mean blank for the standard. Being careful not to expose the KCl solution to the atmosphere, it was drawn into the pipet in the same way as a sample. When handled in this way, the KCl blank was very constant, usually with a mean of around 6.0 ±0.3 µg C for an individual batch and an "over the cruise" mean of 6 ±1 µg C.

The standards yielded a mean calibration factor of 99.6588% ±0.0600 (n = 25).

A Certified Reference Material (CRM) was prepared and bottled by Dr. Andrew Dickson of the Scripps Institution of Oceanography (SIO). The TCO₂ concentration of the CRM was determined to be 2020 ±0.009 µmol/kg by manometric technique in the laboratory of Dr. Charles Keeling of the SIO. Bottles of the CRM were taken on the cruise and analyzed frequently to determine the accuracy and precision of the coulometric method; the results are reported in Table 1. Throughout the cruise, the accuracy was determined to be within 0.15% and the precision was within 0.12%. Replicates analyzed at three different stations throughout the cruise yielded a precision of < or = 0.05%. The reported TCO₂ data have been corrected to reflect the difference in accuracy to the CRM; the correction applied is +3 µmol/kg TCO₂.

Samples were analyzed with the same method as standards, that is, 4.5 mLs of acid were dispensed into the reaction vessel, and 2-3 minutes were allowed to pass to purge CO₂ from the acid. Following that, the pipet was rinsed twice with the sample, and the third fill was isolated and used for analysis. The sample was emptied into the reaction vessel, and allowed to drain for an additional 5 seconds (monitored via stopwatch) to allow droplets of water to be delivered.

Table 1. Certified Reference Material (batch 1) analyzed for total CO₂ during R/V Malcolm Baldrige CGC-90 Cruise

Date	TCO2(µmol/kg)
24 Feb. 1990	2012
25 Feb. 1990	2024
27 Feb. 1990	2014
02 Mar. 1990	2015
05 Mar. 1990	2012
05 Mar. 1990	2016
06 Mar. 1990	2019
07 Mar. 1990	2019
08 Mar. 1990	2019
10 Mar. 1990	2015
11 Mar. 1990	2016
12 Mar. 1990	2016
13 Mar. 1990	2017
15 Mar. 1990	2017
16 Mar. 1990	2017
20 Mar. 1990	2017
29 Mar. 1990	2017
30 Mar. 1990	2014
02 Apr. 1990	2018
03 Apr. 1990	2014
05 Apr. 1990	2016
06 Apr. 1990	2017
07 Apr. 1990	2018
07 Apr. 1990	2018
11 Apr. 1990	2020
12 Apr. 1990	2017
Mean	2017
Std. Dev	2.5

Sigma-t and sigma-theta were calculated using standard UNESCO algorithms (Fofonoff and Millard 1983), and the CTD measured in situ temperature and bottle salinities. When no bottle salinities were available or when they were defined as a questionable or unacceptable measurements, CTD salinities were used in the calculation.