Dataset: Discrete sample measurements
Deployment: AR69-03

Dissolved Oxygen:

Samples were collected into volume-calibrated flasks and preserved for Winkler dissolved oxygen analysis following standard protocols (Langdon, 2010). For AR69-03, in 2022, all samples were titrated onboard the ship within 24-48 hours of collection using a custom-built Winkler titrator with automated potentiometric end point detection (control software available here: Nicholson et al., 2023). For AR45, in 2020, COVID-19 pandemic restrictions precluded the ability to conduct titrations onboard the ship and samples were instead preserved following the procedures of Zhang et al. 2002 and titrated on land at the end of the cruise.

Precision of the sample collection and analysis procedure is determined by agreement between replicate measurements from the same Niskin bottle. Due to the need to preserve samples for land-based analysis, all samples from AR45 were collected in triplicate. After removal of measurements where outlier data indicated that samples had not been successfully preserved (4 of 63 total samples), median agreement among replicates was 0.15% (0.4 μmol/kg). For AR69-03, all samples were collected either in duplicate or in triplicate and median agreement among replicates was 0.11% (0.3 μmol/kg). In some cases, larger discrepancies among replicates reflect errors in sample collecting and/or preservation; these results are reported for completeness but are flagged as questionable based on outlier analysis conducted in using these samples to calibrate the SBE43 oxygen sensor included in the ship’s sensor package (see Related Dataset: Fogaren and Palevsky, 2024).

Accuracy of sample measurements depends on standardization of the sodium thiosulfate titrant based on a reference standard. The sodium thiosulfate titrant used on each cruise was determined by standardization with a 0.01N potassium iodate reference solution from Ocean Scientific International Ltd (OSIL). Lab-prepared potassium iodate standards, measured routinely throughout AR69-03 and before, during, and after land-based analysis of the AR45 samples to verify titration accuracy and stability, were verified and adjusted by measurements against the OSIL standard.

Dissolved Inorganic Carbon and Total Alkalinity:

Samples were collected for dissolved inorganic carbon (DIC) and total alkalinity (TA) analysis following standard protocols (Dickson et al., 2007). Samples were collected into either 250 mL or 500 mL borosilicate glass bottles and preserved with saturated mercuric chloride (100 μL in 250 mL bottles, 200 μL in 500 mL bottles) for later analysis.

Samples were analyzed at the Boston College Marine Biogeochemistry Laboratory. DIC was analyzed using an Apollo SciTech AS-C6L DIC Analyzer and TA was analyzed using an Apollo SciTech AS-ALK2 TA Analyzer. Both DIC and TA were measured from each sample bottle. All DIC measurements were made on the day the bottle was opened for analysis, and TA measurements were made within the same week. DIC and TA instruments were calibrated daily and monitored throughout each analysis session by measuring Certified Reference Materials (Andrew Dickson, UCSD).

Analytical replicates were measured for all samples such that after analytical outliers (sigma) were removed, all samples for both DIC and TA retained at least two replicate measurements (median number of replicates for DIC = 3; median number of replicates for TA = 4). Analytical precision was determined for each sample as the standard deviation of analytical replicates. Mean analytical precision for all DIC samples in this dataset is 0.6 µmol/kg. Mean analytical precision for all TA samples in this dataset is 1.8 µmol/kg. Individual samples are flagged as questionable (QC flag = 3) if the analytical precision is >8 µmol/kg. Individual samples for DIC are also flagged as questionable (QC flag = 3) if CRMs run prior and subsequent to the sample in question differ by >8 µmol/kg.