Flow cytometry: Samples for picophytoplankton abundance were collected pre-dawn from Niskin bottles mounted on a 24-place rosette system equipped with a Seabird SBE911 CTD and a Seapoint fluorometer. Samples (2-mL) were preserved (0.5% paraformaldehyde) and frozen in LN2, then stored at -80°C until shore-based analyses. Flow cytometry samples were thawed and stained for 1 h with the DNA stain Hoechst 33342 (1 µg/ml, Monger and Landry, 1993), then analyzed with a Beckman Coulter EPICS Altra flow cytometer (Selph et al., 2011). Listmode data were processed using FlowJo (version 9.7.7, Treestar, Inc.) to delineate Prochlorococcus (PRO), Synechococcus (SYN), and eukaryotic phytoplankton (PEUK).
Chlorophyll a contributions for Prochlorococcus (PRO) and Synechococcus (SYN) were assigned from normalized chlorophyll (red) fluorescence (NCF) from flow cytometry as follows:
(1) PRO NCF/L = PRO NCF × PRO (cells/L)
(2) SYN NCF/L = SYN NCF × SYN (cells/L)
Assuming that PRO NCF/L was directly proportional to the pigment divinyl chlorophyll a (DVCHLa, ng/L) since DVCHLa is only found in PRO, we estimated the monovinyl chlorophyll a (MVCHLa) associated with SYN as:
(3) SYN MVChla = (SYN NCF/L)/(PRO NCF/L)× DVChla
Microscopy for Trichodesmium: Trichodesmium (TRICH) abundances were assessed from 6.6-L samples taken from 6 depths (2-50 m) in daily (~noon) CTD casts. Water was gravity filtered directly from the Niskin bottle onto 8-µm, 47-mm Millipore TETP filters, preserved (2% paraformaldehyde), mounted on glass slides and frozen (-80°C). Trichome chlorophyll, carbon and nitrogen (CN) contents to biovolume ratios were determined from 6.6-L samples collected as above on the same noon casts but onto 20-µm, 47-mm filters and frozen (-80°C).
TRICH abundance was based on microscopical analyses of preserved, frozen slides (see methods below). Thawed filters were scanned using a dissecting microscope (10X-30X) with a NightSea SFA adaptor and Royal Blue light head (EX 440-460 nm, EM >500 nm) to find all orange-glowing trichomes and colonies. TRICH were digitally imaged (OMAX camera) using ToupLite (Touptec.com), counted, and trichome lengths measured. Trichome widths were determined with an Olympus BX-41 epifluorescence microscope (200X, EX 450-480 nm, dichroic 500 nm, EM >515 nm). These data comprised the background contribution of trichomes to HPLC samples.
For chlorophyll a contents, duplicate samples of TRICH from unpreserved, frozen samples (see methods below) were suspended in salt water, filtered onto GF/F filters, extracted (90% acetone), and fluorescence determined with a 10AU fluorometer using the acidification method (Strickland and Parsons, 1972).
HPLC pigments: Samples (2.2-L) for pigment analyses by high-pressure liquid chromatography (HPLC) were collected pre-dawn from Niskin bottles mounted on a 24-place rosette system equipped with a Seabird SBE911 CTD and a Seapoint fluorometer. They were filtered onto GF/F filters, frozen in LN2 and stored at -85°C. On shore, samples were sent to Horn Point Analytical Services Laboratory (University of Maryland Center for Environmental Science). There they were extracted, and analyzed using an automated 1100 HPLC system with Agilent temperature-controlled autosampler, Peltier temperature-controlled column oven compartment, PDA detector and ChemStation software. The HPLC method uses a C8 column and a reversed phase, methanol-based solvent system (Van Heukelem and Thomas, 2001; Hooker et al., 2012). MVCHLa and DVCHLa are detected at 665 nm. Carotenoid and xanthophyll accessory pigments are detected at 450 nm.
The pigments used for phytoplankton taxonomic identification were monovinyl chlorophyll a (MVCHLa), divinyl chlorophyll a (DVCHLa), monovinyl chlorophyll b (MVCHLb), divinyl chlorophyll b (DVCHLb), chlorophyll c3 (CHLc3), zeaxanthin (ZEAX), fucoxanthin (FUCO), 19’-hex-fucoxanthin (HEX), 19’-but-fucoxanthin (BUT), allophycocyanin (ALLO), peridinin (PER), neoxanthin (NEO), and prasinoxanthin (PRAS).
SYN and TRICH MVCHLa was subtracted from the total MVCHLa, and the remaining MVCHLa was used for all eukaryotic taxa in CHEMTAX analyses (v. 1.95, Wright, 2008). For CHEMTAX, initial pigment ratios (accessory pigment:MVCHLa) were those of oceanic species (Higgens et al., 2011) and indicative of the following groups: chlorophytes (CHLOR), diatoms (DIAT), prymnesiophytes - type 6 (PRYM), pelagophytes (PELAG), cryptophytes (CRYPT), prasinophytes - type 3 (PRAS3), and dinoflagellates (A-DINO). Data were divided into 2 groups: shallower and deeper than 60 m, since some of the accessory pigments were only present in deep samples (NEO and ALLO) and the general pattern of pigments showed a different community at depth. The initial ratio matrix was randomized into 60 matrices (0.7 x random number between -0.5 and +0.5), which were then applied to the data sets (Selph et al., 2021, Supp. Table 1).