Dataset: Coralline algae experiment treatment conditions

Methodology: 

Sampling and analytical procedures: To test the response of the coralline algae Crusticorallina spp. and Bossiella orbigniana to future OA scenarios, we used an 18-aquaria indoor experimental system with flow-through seawater at the Sitka Sound Science Center to simulate three static pHT levels (current summer = 8.0, future summer/current winter = 7.7, future winter = 7.4) under two seasonal light regimes simulated with full-spectrum aquarium lights (AI Prime HD) (summer = PPFD 55μmol m-2 s-1, 13h d-1, winter = PPFD 40μmol m-2 s-1, 6h d-1). We had a total of 3 aquaria for each of the 6 treatment combinations. A full description of the pH control for this system can be found in Kroeker et al. 2021, but in short: pH was regulated using a relay system that controlled mixing of pre-equilibrated low-pH seawater (formed by bubbling pure CO2 gas into seawater: pH6.0) and ambient pH seawater into 9 header buckets (n=3 headers per pH treatment) that then flowed into the experimental aquaria. Each header bucket was equipped with a pH sensor (DuraFET, Honeywell) communicating with a controller (UDA 2152, Honeywell) to regulate flow of the low pH water through solenoid valves to maintain pre-programmed pH setpoints. Experimental pH levels were chosen to reflect current seasonal minimums of coastal pH measured at Harris Is. (57.032N, 135.277W) from 2016-2017, as well as end-of-century projections for Gulf of Alaska pH levels based on RCP 8.5 (-0.3 pHT from current levels). Experimental light regimes were defined using seasonal averages for day length and measured irradiance level at 10m depth at Harris Is.

Within each pH level and light treatment combination, half of the individual Crusticorallina spp. and B. orbigniana were randomly assigned to be paired in close proximity with the fleshy red alga Cryptopleura ruprechtiana (n=6 species treatment-1). All algal individuals were collected on Aug 5, 2017 at Harris Is. Total experimental duration was 45d (Aug 7-Sept 21, 2017). To monitor treatment conditions, we used a handheld meter (YSI) to take daily temperature readings in the replicate aquaria and measure salinity of incoming seawater daily just upstream of our experimental system. Additionally, discrete water samples were collected from replicate aquaria at four timepoints (Aug 18, 22, 25, and Sept 15) for determination of pH (total scale) and total alkalinity (TA). Discrete samples were collected without aeration in amber glass bottles, immediately poisoned with saturated HgCl2 (0.025% volume-1), and capped to prevent air exchange.

Discrete water samples for laboratory measurements of pH and/or TA were transported to UCSC for analysis within 8 months of collection. We measured pH spectrophotometrically (Shimadzu, UV-1800) using m-cresol purple dye following best practices (Dickson et al. 2017), with an average standard error of ± 0.0013 pH units among sample triplicates. TA measurements were performed using open cell titration (Metrohm, 905 Titrandro) and corrected against certified reference materials of CO2 in seawater (Dickson laboratory, Scripps Institution of Oceanography), with an average standard error of ±0.933μmol kg-1 SW-1 among sample triplicates. To calculate pHT in the replicate aquaria at the time of water sampling, we used our measurements of spectrophotometric pH, TA, temperature, and salinity, as well as the dissociation constants as inputs to the program CO2SYS. This dataset reflects all of the above-described data, including final carbonate chemistry calculations from CO2SYS.

Problem report:  

Discrete water samples had to be stored and shipped from Southeast Alaska to Santa Cruz, CA for analysis, and some samples were broken in transit and unusable. Additionally, we recognize that the delay between collection and laboratory analysis of water sample spec pH and TA may have resulted in a measurable amount of drift in these parameters, but we were unable to quantify any drift that may have occurred and thus these values are uncorrected. Lastly, calculated parameters for certain samples did not follow expected relationships (e.g., among spec pH and pCO2) and thus were flagged for potentially poor quality, which may have been caused by biological contamination in bottle samples.