Dataset: Carbonate chemistry data for shipboard ocean alkalinity enhancement (OAE) experiments in the North Atlantic on R/V Atlantic Explorer cruise AE2320 in September 2023

Experiments involved the addition of NaOH solution prepared by weighing ACS grade NaOH in the lab prior to the cruise (R/V Atlantic Explorer AE2320) in a plastic Falcon tube that was capped and sealed with parafilm tape. During the cruise, DI water was added to make up stock NaOH solutions with a final concentration of 1 M. The NaOH solution was pipetted into the seawater filled bags through the Luer-fitted stopcock. Because NaOH contributes only alkalinity but not DIC, seawater in the experiments was out of equilibrium with the atmosphere, which was intended to simulate conditions immediately following alkalinity addition to seawater during OAE deployments.

In total, 5 experiments were conducted. The first experiment (experiment A) was a control with no alkalinity addition. In the second, third, and fourth experiments (B, C, and D), alkalinity was enhanced by 500, 1000, and 2000 micromoles per kilogram (µmol/kg) respectively. The fifth experiment (E) represents a set of "sacrificial" time series experiments whereby 9 bags were prepared similar to other experiments and alkalinity was enhanced by 1000 µmol/kg in each one of them, but each bag was sequentially opened and filtered in order to evaluate the precipitate mineralogy through time. In experiment E, water samples for TA and DIC measurements were taken only at the end of the experiment. The experiments were run for approximately 5 days.

Two separate 12-milliliter (mL) seawater samples were taken from bags through time, one for DIC and one for TA. Each of these samples was subsequently modified in order to test recently proposed best practices for carbonate chemistry sampling techniques (Schulz et al., 2023). These proposed techniques were designed to retain the original DIC and TA values at the time of sampling while decreasing Ω in the sample container to avoid mineral precipitation during sample storage. For DIC samples, adding an acid to a sample in a completely sealed vessel with no headspace neutralizes a proportion of the previously added alkalinity and thus decreases Ω while retaining all the DIC inside the vial. Similarly, for TA samples, bubbling CO2 into the sample increases the DIC, and thus decreases Ω without changing the TA. As such, Ω can be lowered in both samples to prevent mineral precipitation during sample storage in a way that allows for the accurate determination of DIC and TA (Schulz et al., 2023). We note that these techniques only work for conservative carbonate system parameters (i.e. DIC and TA), and not for non-conservative parameters such as pCO2 or pH.

The 12 mL aliquot taken for DIC was passed through a 0.2 micrometer (µm) filter into a gas-tight borosilicate vial (CHROMONE, NJ, USA), poisoned with 2.4 microliters (µL) of saturated HgCl2, and then acidified by adding a pre-calculated volume of 0.075 M HCl using a glass syringe through the plastic vial septum to titrate the initially added alkalinity. The amount of HCl added was 80, 160, 400, and 160 µl for samples taken from experiments B, C, D, and E, respectively. The 12 mL TA aliquot was filtered (0.2 µm filter), and then bubbled with pure CO2 using a nylon tubing with a stainless steel needle for 30 seconds to increase its DIC without changing TA, followed by poisoning with HgCl2. A gas regulator was used to maintain a constant CO2 flow rate and to prevent over-bubbling. The DIC and TA samples were returned to the lab where they were kept in cool and dark conditions until analysis, which took place within 2 months.

TA was determined using an open-system Gran titration on weighed 5 mL samples in duplicate using a Metrohm 805 Dosimat, with a 1 mL burette, and an 855 robotic Titrosampler. An 0.04 M HCl titrant was used to first acidify the sample to a pH of 3.9 before continuing to a pH of 3.25, dosing at 0.02 mL increments. The analyses were calibrated using in-house seawater standards that were run every 15 samples, to assess titrant and electrode drifts throughout the day. A nonlinear least-squares method was used to determine TA as outlined in the Best Practices guide (Dickson et al., 2007).

DIC was determined using an Apollo LI-5300A connected to a Li-COR CO2 analyzer, with CO2 extracted from a 1.5 mL sample volume by adding 0.8 mL of 3% phosphoric acid. Once opened, the sample lines were inserted to the base of the vial and sealed with parafilm tape to limit gas exchange. Before each analysis, 0.75 mL of sample and 0.8 mL of acid is drawn into the sample syringe to flush out any prior remnants from the system. After the flush, the 1.5 mL sample is drawn into the calibrated syringe and injected into the reaction chamber, where resulting CO2 is carried by a zero CO2 air stream to the Li-COR CO2 analyzer. Samples were run in triplicates. The instrument was calibrated twice daily against an in-house seawater standard that were intercalibrated against seawater Certified Reference Materials (Dickson batch #187).

The saturation state with respect to aragonite (Ω_A) throughout the experiments was calculated using PyCO2 1.8.1 (Humphreys et al., 2022), the Python version of the original CO2SYS program (Lewis et al., 1998) using the carbonic acid dissociation constants of Mehrbach et al. (1973) refitted by Dickson and Millero (1987). The Ω_A calculations used a corrected concentration of Ca to account for changes induced by CaCO3 precipitation.