Dataset: FlowCam Particle Sizes

Samples were collected during October and November 2018 on R/V Tangaroa cruise TAN1810 ("SalpPOOP" cruise) in the Chatham Rise (subtropical and subantarctic waters off the coast of New Zealand). Samples were collected from Niskin bottles of CTD deployments to the base of the mixed layer and the deep chlorophyll maximum. 250 mL subsamples were concentrated by gravity filtration to 10 mL over a 2 µm 47 mm filter, and 2 mL of this concentrate was imaged using a FlowCam's 10X objective lens to quantify the larger (>3 µm) phytoplankton (Sieracki et al. 1998).

FlowCam image analyses were conducted using FlowCam's dedicated classification software VisualSpreadsheet (v. 4.18.5). First, duplicate images resulting from parabolic flow within the flow cell were manually removed. The particles within the remaining images were then classified based on the quality with which VisualSpreadsheet detected their outlines. For particles where the outline appeared to provide good estimates of length and width (classes "Other", "Pennates", "Chaeto", and "Ciliate"), size was calculated for a prolate spheroid using the minimum feret as width and maximum feret as length.

Four categories of particles imaged by the FlowCam proved problematic due to poorer outline quality and had to be treated specially. For such pennate diatoms (class "BadPen"), we manually measured the caliper width of every such pennate diatom in the first CTD cast of each of the four cycles to derive the average pennate width for that cast, and applied this average retroactively to all poorly detected pennates within that cycle. Equivalent spherical diameter (ESD) and biovolume (BV) for the resulting prolate spheroid were then calculated as normal. For semi-transparent dinoflagellates (class "ClearDino"), we recalculated the width using the recorded length and mean aspect ratio for well-detected dinoflagellates in the first CTD cast of each cycle and then calculated ESD and BV normally. For Chaetoceros (class "BadChaeto") we recalculated the length and width of these particles using the recorded aspect ratio and area. Each Asterionelopsis colony (class "Aster") was saved as an individual image file, and representative single cells within a colony that were parallel to the camera’s field of view were manually analyzed using ImageJ (v. 1.52a). This procedure was repeated for all colonies in the first cast of each cycle and the mean width and length taken to produce an "average" Asterionelopsis cell. We then manually counted the number of individual cells present in each image and applied them to the averaged cell sizes to estimate the total biovolume.

The biomass of ciliates (class "Ciliate") was estimated as 0.19 pg C µm⁻³ Putt and Stoecker 1989). The biomass of diatoms (classes "Chaeto", "BadChaeto","Pennate", "BadPen", and "Aster") was estimated allometrically as 0.288*Biovolume^0.811 while other protists and unidentified particles (classes "Other" and "ClearDino") were estimated using 0.216*Biovolume^0.939 (Menden-Deuer and Lessard 2000).