Three CO2 concentrations were tested: 410 ppm, 750 ppm, and 1000 ppm respectively. For each CO2 concentration, four temperatures were tested: 15 degrees-C, 20 degrees-C, 25 degrees-C, and 30 degrees-C. Within each temperature, three light levels were tested: a sub-optimum light (SOL) intensity of 60 umol photons · m-2 · s-1, an optimum light (OL) intensity of 400 umol photons · m-2 · s-1 and an extreme light (EL) intensity of 800 umol photons · m-2 · s-1. All lights were set at a 12 h day: 12 h dark cycle. For logistical reasons, experiments were partially conducted in series, with all light treatments at two temperatures (either 15 degrees-C and 25 degrees-C or 20 degrees-C and 30 degrees-C) running simultaneously. This was repeated for each CO2 concentration.

Experiments were conducted in Multicultivator MC-1000 OD units (Photon Systems Instruments, Drasov, Czech Republic). Each unit consists of eight 85 ml test-tubes immersed in a thermostated water bath, each independently illuminated by an array of cool white LEDs set at specific intensity and timing. A 0.2um filtered CO2-air mix (Praxair Distribution Inc.) was bubbled through sterile artificial seawater, and the humidified gas mix was supplied to each tube via gentle sparging through a 2um stainless steel diffuser. Flow rates were gradually increased over the course of the incubation to compensate for the DIC uptake of actively growing cells, and ranged from <0.04 Liters per minute (LPM) at the start of the incubations to 0.08 LPM in each tube after 2 days. For each CO2 and temperature level, replication was achieved by incubating three tubes at sub-optimum light intensities, two tubes at optimum light intensity, and three tubes at extreme light intensities. Each experiment was split into two phases: An acclimation phase spanning 4 days, was used to acclimate cultures to their new environment. Pre-acclimated, exponentially-growing cultures were then inoculated into fresh media and incubated through a 3-day experimental phase during which assessments of growth, photophysiology, and nutrient cycling were carried out daily. All sampling started 5 hours into the daily light cycle to minimize the effects of diurnal cycles.

Experiments were conducted with artificial seawater (ASW) prepared using previously described methods (Kester et. al 1967), and enriched with nitrate (NO3), phosphate (PO4), silicic acid (Si[OH]4), at levels ensuring that the cultures would remain nutrient-replete over the course of the experiment. Trace metals and vitamins were added as in f/2 (Guillard 1975). The expected DIC concentration and pH of the growth media was determined for the different pCO2 and temperatures using the CO2SYS calculator (Pierrot et al. 2006), with constants from Mehrbach et al. (1973, refit by Dickson & Millero 1987), and inputs of temperature, salinity, total alkalinity (2376.5 umol · kg-1), pCO2, phosphate, and silicic acid. DIC levels in ASW at the start of each phase of the experiments were manipulated by the addition of NaHCO3, and was then maintained by bubbling a CO2-Air mix through the cultures over the course of the experiments. The pH of the growth media was measured spectrophometrically using the m-cresol purple method (Dickson 1993), and adjusted using 0.1N HCl or 0.1M NaOH. The media was distributed into 75 ml aliquots and each aliquot was inoculated with 5 ml of the T. pseudonana CCMP 1014 (TP1014) stock culture at the start of the experiments.

Flow cytometry:
Samples were fixed in Hexamethylenetetramine-buffered formaldehyde (final concentration 1% v/v) and stored at 4 degrees C in the dark for a maximum of 4 days. Cell counts were confirmed to be unaffected over storage for up to a week. Samples were analyzed on a Guava easyCyte HT Benchtop Flow Cytometer (Millipore-Sigma, USA). All data acquisitions were done with logarithmic signal amplification. Cytometer sample flow rates were kept low (0.24 uL · s-1) to accommodate high cell concentrations. Diatoms were identified based on size and chlorophyll autofluorescence using the forward scatter channel (FSC) and Red-FL (695/50 nm) channel respectively. Growth rates were derived by fitting an exponential curve to cell concentrations vs. time for a 48-hour period during which cells exhibited exponential growth in the experimental phase. Growth rates in treatments where cells did not grow, or declined in abundance were listed as 0. Particle sizes (equivalent spherical diameter in µm, ESD) were derived from FSC using size-calibration beads of known diameters ranging from 2 µm to 10 um (Particle Size standard kit, Spherotech Inc.).