There were 5 stations were sampled with only station 5 getting 2 casts to get a higher resolution in the mid-shallow water column.

Samples for pH analysis were collected first in the rosette sampling sequence. Seawater samples were collected from the Niskin bottles directly into 10-cm glass cylindrical optical cells (~30 mL volume) using a section of silicone tubing (~15 cm long). One end of the silicone tubing was first attached to the nipple of the Niskin bottle. The nipple was pushed in to initiate flow, and the silicone tubing was squeezed to eliminate air bubbles. The other end of the silicone tubing was attached to the optical cell, which was agitated to eliminate any residual bubbles. After ~15 seconds of sample flow, the cell was capped at one end with a Teflon stopper. The silicone tubing was then detached from the optical cell, and, with the water still flowing, the other cap was rinsed and used to seal the optical cell. Samples collected this way are not exposed to the atmosphere, and each cell was flushed with at least three cell volumes of seawater. The samples were collected and taken into the lab, where the outside of the cell was rinsed with tap water to eliminate salt build-up. The cells were dried thoroughly, and the optical windows were cleaned with Kimwipes immediately before measurement. Samples were thermostatted at 25 ºC (±0.05 °C) in a custom-made, 36-position cell warmer.

Seawater samples for total alkalinity (AT) were collected directly after pH samples from the Niskin bottles into clear 300 mL BOD borosilicate glass (Fisher Scientific) bottles as described in Section 3.4 of Liu et al. (2015).

Silicone tubing, connected directly to the Niskin, was inserted into the bottom of the bottle. Each bottle was rinsed three times with seawater with approximately one-third of the sample bottle volume. Sample bottles were overflowed for at least one complete volume (300 mL). The silicone tubing was squeezed to eliminate air bubbles, and care was taken to ensure no air bubbles were trapped in the bottle. While filling the bottle, the sample temperature was measured (ts) using a Fisher Scientific Traceable Digital Thermometer (±0.05℃). Sample temperature is required to ascertain sample volume and convert it to sample mass. Once all bubbles dissipated, sample caps were rinsed, and the tubing, while water still flowed, was slowly removed, and the sample was sealed by inserting the bottle cap. Seawater in the flange area was eliminated, and the bottleneck area was dried with a Kimwipe.

In the ship’s laboratory, the outside of each bottle was carefully rinsed with tap water, and the bottle surface was wiped clean with Kimwipes to remove any salt build-up. Samples were left to equilibrate to room temperature. Samples were not poisoned with HgCl2. When the samples warmed, seawater was forced into the flanged bottleneck. Care was taken to ensure no sample loss—the bottle was maintained upright, and water within the flanged neck was not wiped away as the sample warmed.

DIC and delta13C of DIC were analyzed with a Picarro Cavity Ring-Down Spectrometer (G2131-i) with Liaison autosampler; the detailed methodology is described in Subhas et al. (2015). These measurements were made on board the ship. Dickson seawater CRM was used as the standard for DIC; pre-weighed optical calcite powder was used as the standard for delta13C. Exetainer vials were pre-acidified and pre-weighed in the lab prior to the cruise. After the vials were filled with 3-5 mL of porewater and analyzed on the cruise, the stored vials were weighed again in the lab to obtain the sample mass. Results using this methodology were corrected by normalizing to measured values of Dickson seawater CRM. Uncertainty (1sigma) for replicate DIC and delta13C measurements were ±23 umol/kg and ±0.15 per mille (VPDB), respectively. DIC uncertainty was higher than reported in our previous studies (e.g., Subhas et al., 2015) likely due to mass determination: ship-board analysis necessitated weighing after analysis, but there was uncertainty in precisely how much sample mass may be removed during the analysis. Additionally, lower sample volume (5 mLs, instead of 7-8 mLs in our previous studies) may have added error.

The pH of each sample was determined on an Agilent 8453 spectrophotometer setup with a custom-made temperature-controlled cell holder. Only the tungsten lamp was turned on. The UV lamp was turned off to prevent photodegradation of organic matter in the samples by UV light. A custom macro program running on Agilent UV-Visible ChemStation Software Rev. B.04.01 was used to guide the measurements and data processing. The macro automated the procedures of sample input information, blank and sample scans, quality control, and data archiving. The quality control steps included checking the baseline shift after dye injection and monitoring the standard deviation of multiple scans. Absorbance blanks were taken for each sample, and 10 μL of purified m-cresol purple (10 mmol kg-1) was added for the analysis. pHT(total scale) was calculated according to Müller and Rehder (2018).

The salinity and temperature dependence is provided in Table 1 of Müller and Rehder (2018). The temperature and salinity dependence of e1 and e3/e2 are given in Eq. (6-7) of Muller and Rehder (2018). Sample temperature was then measured using a Fisher Scientific Traceable Digital Thermometer (±0.05℃). These equations are applicable for samples between temperature (278.15 ≤ T ≤ 308.15) and salinity (0 ≤ S ≤ 40). In all our measurements at sea, T ≈ 298.15 K. Water column salinity was measured using a Sea-Bird Electronics CTD sensor. Duplicate pH samples, collected from discrete samples taken from Niskin bottles (N=12), displayed a standard deviation of 0.001.

The automated spectrophotometric alkalinity system consists of a stir plate with an anchoring bottle holder and a Metrohm Dosimat 665. The bottle holder is configured so that optical fibers, connected either to an Ocean Optics LS-1 tungsten halogen light source with a blue filter or an Ocean Optics USB4000 spectrophotometer, are positioned about 1 cm above the bottom of the bottle (configuration shown in Liu et al. 2015, Fig 1). The spectrophotometer and Dosimat are connected to a portable computer equipped with custom software that controls acid delivery from the titrator and monitors absorbances. The software is programmed to control the rate and sequence of acid titration and record sample input, absorbance ratios, and acid volumes.

DIC and delta13C of DIC were analyzed with a Picarro Cavity Ring-Down Spectrometer (G2131-i) with Liaison autosampler.

925367_v1_watercolumn.csv