Dataset: HPLC data from field sampling sites in the Neuse River Estuary, Pamlico Sound, and Onslow Bay in the coastal North Atlantic, offshore from North Carolina, USA, in 2021 and 2022

Sampling summary: Discrete samples were all obtained from ~0.2m using a diaphragm pump and weighted, marked hose. All containers were kept in dark coolers at ambient temperature during transport to the laboratory.  All filtration was done within a few hours of collection and when conditions permitted, on board the research vessel.

Diagnostic phytoplankton photopigments were identified, separated and quantified by high performance liquid chromatography (HLPC) coupled to an in-line photodiode array spectrophotometer (Jeffrey et al.  1997):  Known volumes of water sample (500-1000 milliliters, enough to obtain color on the filter) were vacuum filtered (less than 25 kiloPascals) through 25 or 47 millimeter Whatman glass microfibre filters (GF/F) under reduced light conditions.  The filters were blotted dry, folded in half, wrapped in foil and then immediately frozen at -20 degrees Celsius until analysis.  The filters were placed in 15 milliliter centrifuge tubes containing 1.5-3.0 milliliters of 100% acetone (HPLC Grade), sonicated for 30-60 seconds using a Fisher Sonic Dismembrator 300 with microtip and extracted at -20 degrees Celsius for 12-24 hours.  After extraction the samples were centrifuged at 4500 rpm and the supernatant (i.e.- the combined extracted pigments) collected & filtered into amber glass autosampler vials using Millipex Millipore 0.45 micometer PTFE.  Two hundred microliters of extractant from each vial was injected into the HPLC system using a Spectra Physics (now Thermo Separations Products) AS3000 autosampler and SP8800 pump, running a non-linear, 55 minute, 2-solvent gradient adapted from Van Heukelem et.al. 1994 or 1995?.  The nonlinear, variable flow, binary gradient consisted of solvent A [80% methanol : 20% ammonium acetate (0.5 M adjusted to pH 7.2)] and B (80% methanol : 20% acetone).  The extractant was separated into individual pigments using a series of C18 reverse-phase columns to optimize photopigment separations:  The column order was a Rainin Microsorb guard column (0.46 x 1.5 centimeters, 3 micrometer packing) followed by a single monomeric reverse-phase C18 column (Rainin Microsorb-MV, 0.46 x 10 cm, 3 µm packing) followed by two polymeric reverse-phase C18 columns (Vydac 201TP5, 0.46 x 25 cm, 5 µm packing).  The columns were kept at a constant 52 degrees Celsius in an Alltech 330 column heater.  The separated pigments were then passed through an in line Shimadzu SPD-M10AV photodiode array detector which measured the absorbance of the sample/extractant, scanning the range of 350-800 nanometers every 2 seconds.  The data was collected and analyzed using Shimadzu's EZChrom software (Agilent Technologies, Inc (n.d.)).  Individual pigments are identified using a combination of peak retention time and absorbance spectrum shape.  Retention times and absorbance spectra are identified for each pigment by analyzing known pigments (either as pure standards or pigments or isolated from algal cultures).  Pigments are quantified from their peak areas, calculated at 440nm. A calibration curve is generated by injecting various volumes of a mixed standard composed of known quantities of seven pure pigment standards (fucoxanthin, zeaxanthin, bacteriochlorophyll a, canthaxathin, chlorophyll b, chlorophyll a, echinenone and ß-carotene) and calculating the peak areas of those pigments   The peak areas are regressed against the known quantities of each pigment to calculate the slope (Response Factor) for that pigment.  Response factors for pigments we do not have reference standards for are calculated using the ratio of absorbance coefficients of each pigment to its closest structurally related reference pigment, multiplying the known pigment's response factor by that ratio. Pigments extracted from the samples are then quantified by multiplying the peak areas of a chromatogram at 440nm by the response factors. Pigment values listed as below detection were below the software threshold for peak detection or had spectra below a similarity of 0.9 compared to library spectra. Technician expert judgment was used in difficult cases.

See the "Related Datasets" section for in-situ hydrography methods and data collected as part of the same study at the same field sampling locations.