Dataset: RR2004 Bottle Data

R/V Roger Revelle (cruise ID RR2004) departed Honolulu, Hawaii on 26 December 2020. The ship transited south along the great circle route from Honolulu to 30°S x 150°W.

Sampling Overview:
Hydrographic profiles were performed along the transect with the CTD (the "full-water cast"), sampling for Freons, dissolved oxygen, DIC, alkalinity (and all other parameters of the carbonate cycle using the CO2-SYS program), extracted chlorophyll, nutrients, POC, PIC, biogenic silica, coccolithophore counts and FlowCAM analyses (for enumerating and classification of nanoplankton and microplankton species). Each CTD full-water cast was alternated with a "trip-on-fly" water cast. The latter casts involved tripping bottles at 24 depths "on the fly" as they passed the following 24 depth targets: 1000 meters (m), 900m, 800m, 700m,600m 500m, 450m, 400m, 350m, 300m, 250m, 200m, 150m, 120m, 110m, 100m, 90m, 80m, 70m, 60m, 50m, 40m, 25m, 5m. These later casts were used only to sample full water properties at the surface only, as well as DIC and nutrients at eight depths. These trip-on-fly casts served to provide greater resolution in hydrographic sections across the features. Once per day, typically pre-dawn, CTD casts included samples drawn for primary productivity and calcification. These measurements involved the use of 14C-bicarbonate and all manipulations were done in a portable radioisotope van on deck.

Sampling for trace metals was performed daily using nine Niskin-X bottles clamped to nonmetallic Aracom line, hung at depths to ~1000m and tripped with nonmetallic messengers. All trace-metal-clean manipulations were performed in a trace-metal-clean laboratory or a plastic bubble built within the ship's wet lab. Five carboy experiments were performed over the cruise, which involved incubating surface, trace-metal-clean water collected by a "Big Jon" surface sampler (towed from the side of the ship at 1-3 knots (kts), which maintained a distance of 5-10m from the side of the ship as it pumped surface water into two 200-liter (L) plastic tanks within the wet lab bubble). For three of the carboy experiments, the investigators conducted triplicate incubations of untreated control water plus five treatments (three replicates each) in plastic, acid-cleaned cubitainers with a) 5% dilution with subsurface water, b) 20-micromoles (uM) trace metal-clean nitrate, (c) 20 uM trace metal-silicate, (d) 1 nanomole (nM) of iron and (e) 1 nM of iron+20 um of silicate 6. The cubitainers were then sampled approximately every other day for 4-5 days while being incubated under surface light conditions in an on-deck incubator, with temperature maintained at T0 in-situ surface conditions. The carboys were sampled about every two days by the Balch group for chlorophyll, nutrients, PIC, POC, biogenic silica, quantitative coccolithophore counts, and quantitative FlowCAM samples (for enumeration of algal classes, cell volumes, and slope of the particle-size distribution). Later in the cruise, for two of the carboy experiments, due to time constraints with two simultaneous incubations, incubations could only be performed with a control and two treatments of (a) 5% subsurface water and (b) 2 nM iron).

Sampling Methods:
At sea collections: Water samples were collected using CTD casts from 103 stations encompassing Subtropical, Subantarctic and Polar waters in the Pacific Sector of the Southern Ocean. Discrete samples were taken from 10L Niskin bottles for the following measurements:

1. Chlorophyll - Water samples were filtered onto a 25-millimeter (mm) Millipore HA filter (mixed cellulose ester, 0.45-micrometer (µm) pore size). The filters were transferred to test tubes filled with chilled 90% acetone for extraction and vortexed until the filter dissolved. Tubes were stored in the dark in a freezer for 24 hours before analysis. Tubes were then re-vortexed and gently centrifuged (~1300g) for 5 minutes before being decanted into a glass cuvette for the fluorometer. A Turner Designs 10AU was used to read Fb of the sample and then, after adding 50 microliters (µl) of 10% HCL, to read Fa. The fluorometer was calibrated pre-cruise with a pure chlorophyll extract (Turner Designs part# 10-850) to determine Tau τ=(Fb/Fa pure chl a) and chlorophyll a was then calculated from: (Fb – Fa) * (τ/ τ-1) * (Vfiltered/Vextracted). Generally, all surface measurements were made in triplicate. The fluorometers (Turner 10-AUs) were calibrated using the calibration method defined by Turner Designs using standards purchased from Turner Designs. Additionally, for long cruises such as this cruise, a calibration was performed on the ship. References: Trees, et al.

2. Particulate organic carbon (POC) plus particulate organic nitrogen (PON) - Water samples were filtered onto 25 mm GF/F filters which were pre-combusted (450°, 5 hours). Filters were rinsed with filtered seawater (FSW) and then stored in individual petri-plates and dried (60°) for storage. Prior to analysis, the plates were opened and placed overnight in a sealed container like a dessicator with saturated HCL fumes to remove any particulate inorganic carbon (PIC). These samples were run by the Bigelow Laboratory Analytical Facility. The filters were packed into pre-combusted nickel sleeves and analyzed on a Perkin Elmer 2400 Series II CHNS/O for C, N, and H. The analyzer was calibrated using tin capsules as blanks and acetanilide to calibrate instrument response to carbon and nitrogen. NIST-certified check standards consisting of either low organic content soil or sediment are analyzed to determine accuracy of carbon detection. NIST-certified organic check standards such as corn flour or rice flour were analyzed to determine the accuracy of nitrogen detection. If values varied by more than 4% from stated values, the instrument was examined, any problems were addressed and the instrument was recalibrated and checked standards rerun until the error was within acceptable limits. Duplicate samples were run during each sample run to ensure results were reproducible. If duplicates could not be run on actual samples, as in the case of filter samples, duplicate check standards were analyzed. Duplicate samples typically varied less than 2%. One instrument blank was analyzed for every 12 samples run. One acetanilide standard was analyzed for every 15 samples run. If blank or acetanilide values differed significantly from previous values, a new series of standards and blanks were analyzed to recalibrate the instrument. The actual minimum detection limit (3 times the standard error) determined from the standard error of the instrument blanks is 2 micrograms for carbon and 4 micrograms for nitrogen. References: JGOFS (1996).

3. PIC (Particulate Inorganic Carbon) - Water samples were filtered through a 25mm, 0.4 µm pore size polycarbonate filter. The dry filter was rinsed with Potassium tetraborate (6.11 grams per liter K₂B₄O₇ · 4H₂O) buffer while still in the filter tower to remove as much seawater salt and also to maintain a high pH (~8.1) during sample storage and to preserve the CaCO₃ on the filter. Filters were placed into trace metal clean polypropylene centrifuge tubes and dried at approximately 60°. For analysis, the filters were sent to (a) the Sawyer Environmental Chemistry Laboratory at the University of Maine or to (b) the Department of Earth Sciences at Boston University. Filters were digested in a 5% nitric acid solution for 12 hours to dissolve all CaCO₃ and the solution was analyzed by ICP-AES (Inductively Couple Plasma – Atomic Emission Spectrometry) for Ca concentration. Filter and dissolution blanks were run as well as QC standards run with each batch of samples. The investigators also used the concentration of dissolved Na in the digestate to correct for any Ca present in sea salts left on the filter. PIC concentrations were calculated using the volumes of water filtered and the volume of the digestions, and assuming all Particulate Inorganic Carbon was in the form of CaCO₃.

4. Biogenic Silicas - To determine reactive silicate, 200 milliliters (mL) of seawater sample is filtered onto a 25 mm, 0.4um pore size polycarbonate filter. Filters are folded and placed in a super clear polypropylene centrifuge tube and dried in a drying oven at 60° Celsius (C) for 24 hours then tightly capped and stored until analysis. On shore, 0.2N NaOH is added and the sample is placed in a 95°C water bath. The digestions are then cooled and neutralized with 1N HCl. After centrifuging, the supernatant is transferred to a new tube and diluted with MilliQ water. Molybdate reagent is added and then a reducing agent is added to reduce silicomolybdate to silicomolybdous acid. The transmission at 810 nanometers (nm) is read on a Hitachi U-3010 spectrophotometer (SN 0947-010). Reactive silicate is calculated using a silicate standard solution standard curve prepared at least every 5 days or whenever new reagents are prepared. Readings are corrected using a reagent blank run at the same time as the standard curve and three tube blanks interspersed in each batch. References: Brzezinski & Nelson (1989); JGOFS (1996); Strickland & Parsons (1972).

5. Freon analysis - Sampling for the freons, CFC11 and CFC12, as well as analytical methods for their measurement were described previously (Bullister & Weiss, 1988; Bullister & Wisegarver, 2008), along with the equation of solubility of CFC11 and CFC12 as a function of temperature and salinity (Fine, 2011). A look-up table (Bullister, 2015) was used to convert CFC partial pressures measured in the Southern Hemispheric to the year of equilibration; the initial table covered up to the year 2015. The look-up table was extended to 2021 (M. Warner (Univ. Washington), personal communication).

6. Dissolved Inorganic Carbon and Total Alkalinity Measurements - The analytical method followed standardized protocols (Bates et al., 1996; Bates et al., 2001; Dickson et al., 2007; Knap et al., 1993). Samples for DIC and TA were collected in 250ml borosilicate glass bottles according to standard JGOFS methods. Milli‐Q cleaned bottles were rinsed out 3 times, bottom filled using silicone tubing, allowed to overflow at least 1X the bottle volume, ensuring no bubbles were in the sample and that it was sealed with a small headspace to allow for water expansion. Water samples were collected from all depths the CTD‐rosette sampled on full casts and from eight depths on the "trip‐on‐fly" casts. Two samples were collected from each Niskin bottle on the full casts. The first sample was poisoned with 100μl saturated mercuric chloride solution for analysis ashore. The second sample was not spiked and stored in the dark for no longer than 12 hours (to minimize any biological activity altering the sample) before being run aboard the ship, DIC first then TA. In addition to sampling from the rosette, samples were also collected and analyzed on board from the underway system. Both the underway and carboy samples were unpreserved, stored in the dark and analyzed on board the ship. Samples were processed at sea using a highly precise (0.02%; 0.4 millimoles per kilogram (mmoles kg-1)) VINDTA system (Bates, 2007; Bates et al., 1996; Bates & Peters, 2007). TA was measured on the VINDTA 3S by titration with a strong acid (HCl). The titration curve shows 2 inflection points, characterizing the protonation of carbonate and bicarbonate respectively, where consumption of acid at the second point is equal to the titration alkalinity. DIC was measured on the AIRICA by the extraction of total dissolved inorganic carbon content from the sample by phosphoric acid addition. The liberated CO₂ flowed with a N₂ carrier gas into a Li‐Cor non‐dispersive IR gas analyzer where the CO₂ levels were measured. For both instruments, within bottle replicates were run consecutively on start-up to check the precision, continuing once the instrument precision was ±2 micromoles per kilogram (μmol kg‐1) or better. These were followed by a combination of Certified Reference Materials (CRMs) produced by the Marine Physical Laboratory at UCSD and low nutrient surface water from the Bermuda Atlantic Time Series (BATS) site, which were run every 20‐24 samples on the VINDTA and every 6 samples on the AIRICA, to determine the accuracy and precision of the measurements and to correct for any discrepancies. The TA system CRM values did not vary more than 2 millimoles (mmol) within each batch of HCl acid. The AIRICA was more susceptible to drift and was affected by the lab temperature which is why CRMs were run much more often on the AIRICA, the system did not drift much and the lab temperature did not vary markedly. Both of the DIC and TA methods had a precision and accuracy of ~1 mmol kg-1 (precision estimates were determined from between-bottle and within-bottle replicates, and accuracy assessed using CRMs. The values for DIC and TA were used to calculate other parameters of the carbonate system using the software CO2sys (Lewis and Wallace, 1998). The calculated parameters were: pH, fCO₂, pCO₂, [HCO3‐], [CO3=], [CO2], alkalinity from borate; hydroxide ion; phosphate and silicate, Revelle Factor, plus the saturation states of calcite and aragonite.

7. Nutrients - Analyses of phosphate, silicate, nitrate+nitrite, nitrite, and ammonia were performed on a Seal Analytical continuous-flow AutoAnalyzer 3 (AA3). The methods used were described by Gordon et al. (1992), Hager et al. (1972), and Atlas et al. (1971). Details of modification of analytical methods used in this cruise are also compatible with the methods described in the nutrient section of the GO-SHIP repeat hydrography manual (Hydes et al., 2010).

8. Coccolithophore enumeration - Polarized microscopy was used to determine the concentration of coccolithophores and detached coccoliths in samples collected in the SW Pacific during the R/V Roger Revelle cruise from December 2020 to February 2021. A volume of 200mL was filtered onto 0.4μm-pore size, 25mm diameter polycarbonate filter and then processed according to Balch & Utgoff (2009).

9. Enumeration of major algal classes - A shipboard Yokogawa Fluid Imaging Technologies FlowCam imaging cytometer was used to enumerate the major microalgal classes and estimate the particle size distribution function. The instrument was keyed on particle backscattering and fluorescence properties. Samples were first filtered through 100um Nitex mesh to make sure the 100um diameter flow chamber did not clog. The instrument was run with a 10X objective in order to reliably count particles bigger than 4-5um diameter. Samples were processed according to Poulton and Martin (2010). Concentrations (per mL), percent contribution with respect to total particles, and biomass are presented. Carbon biomass was determined based on Menden-Deuer & Lessard (2000) method.

10. Primary Production and Calcification Carbon rates - Samples were also taken for measuring photosynthesis and calcification rates from 21 morning, full-CTD stations over the course of the trip (here called Productivity Stations). For these measurements, Niskin bottles were tripped at specific light depths throughout the euphotic zone (0.56%, 3.86%, 7.10%, 23.4%, 42.2% and 73.6%). During casts where there was sufficient light to measure PAR throughout the euphotic zone, these depths were calculated assuming a constant diffuse attenuation coefficient. For samples taken during the nighttime, estimation of those light depths was performed based on the assumption that the fluorescence maximum was located at the 1% light depth (Poulton et al., 2017). Water samples for incubation were transferred from Niskin bottles to incubation bottles, typically inside the ship's enclosed hanger, under subdued light conditions. Water samples were pre-filtered through 120mm nitex mesh to remove large grazers. Incubations were performed in 70 mL polystyrene tissue culture bottles that were previously acid-cleaned, rinsed with ethanol, reverse-osmosis water, then rinsed 5x with each sea water sample prior to filling. Photosynthesis and calcification were measured using the microdiffusion technique (Paasche & Brubak, 1994) with modifications by Balch et al. (2000) (see also Fabry (2010)). 14C bicarbonate (~30 mCi) was added for each water sample. Incubations were performed in triplicate (with an additional 2% buffered formalin sample (final concentration) used as a killed control) in simulated in situ conditions on-deck, corrected for both light quantity (extinction using bags made of neutral-density shade cloth) and quality (spectral narrowing) using blue acetate bag inserts. Bottle transfers between the incubators and radioisotope van were always done in darkened bags to avoid light shock to the phytoplankton. Deck incubators consisted of blue plastic tubs open to sky light, chilled using surface seawater from the ship's flowing seawater system. Calibration of those light levels in the bag were previously made using a Biospherical OSR2100 scalar PAR sensor inserted into each bag relative to a scalar PAR sensor outside the bag. All filtrations were performed using 0.4 mm pore-size polycarbonate filters. Following sample filtration, polycarbonate filters were rinsed three times with filtered seawater, then carefully given a "rim rinse" to make sure that all 14C-HCO3 in interstitial seawater in the filters was rinsed out. Filters and sample "boats" were placed in scintillation vials with 7mL of Ecolume scintillation cocktail. Samples were counted using a high-sensitivity Beckman Tricarb liquid scintillation counter with channel windows set for 14C counting. Counts were performed for sufficient time to reach 1% precision or 25 minutes for samples with lower counts. Blank 14C counts were always run for scintillation cocktail as well as the phenethylamine CO₂ absorbent. Standard equations were used for calculating primary production and calcification from the 14C counts with a 5% isotope discrimination factor assumed for the physiological fixation of 14C-HCO3 as opposed to 12C-HCO3. Specific intrinsic growth rates of organic matter were calculated by dividing daily photosynthetic carbon estimates by the concentration of POC. Carbon-specific intrinsic growth rates for PIC were calculated by dividing the calcification rate by the concentration of PIC.

Known Problems or Issues:
The investigators discovered the calcification blanks during the cruise had consistently higher DPMs than the photosynthesis blanks. They ran an extra experiment on the formalin blanks to see whether the buffer in the buffered formalin used to kill the cells, was causing the artificially high blanks. This experiment was performed using highly oligotrophic, 0.2mm filtered water found north of the Subtropical front in which there was no measurable phytoplankton fluorescence. The investigators filled 10 productivity bottles with this water, and added buffered formalin (buffered to pH 8.8) to half of them (leaving the other five bottles "live" despite the fact that all particles >0.2mm diameter had been filtered out), then incubated all bottles with 30uCi 14C bicarbonate in the dark for 24h. All 10 bottles were subsequently filtered onto 0.4um polycarbonate filters and subjected to the microdiffusion technique. The calcification blanks for the filtered, non-killed samples had radioactivity that was 46% lower than the blank samples "killed" with buffered formalin. Given the state of oligotrophy in the original water samples, and that they were incubated in darkness, the investigators conclude that the buffered formalin-killed calcification blanks caused a small chemical artifact. That is, that the buffer injected with the formalin into the incubation bottles was driving the carbonate equilibrium to precipitate a small amount of the 14C-bicarbonate, which was then caught on the filters for the killed blanks. For this reason, for all calcification blanks, the investigators subtracted the blank formalin values from filters that were acidified prior to counting (which drove off any residual 14C-carbonate precipitate (artifact) or residual 14C bicarbonate solution left in the interstices of filters).