Epifluorescence Microscopy Analysis

Phytoplankton images were captured with an Olympus Microscope DP72 Camera using an Olympus BX51 fluorescence microscope. 20 images were taken under the fluorescence of FITC (green fluorescence) n order to capture the fluorescence of cell proteins. Proflavine overstained cells to the point that we were not able to confidently differentiate heterotrophic cells from autotrophic cells, therefore only biomass and abundance values were able to be calculated. A 60x magnification lens was used to image slides with 0.8 µm filters while 20x magnification was used to image slides with 8 µm filters. Images were then processed using the ImageJ image analysis software (v 1.52a or 1.53c). Cells were manually outlined using the freehand tool and approximate feret cell length and cell area were calculated by ImageJ in pixels and then converted to microns using a calibration scale. To avoid biasing the examined cell size, cells that were roughly greater than 50% out of frame and cells that were broken or fragmented were not included in analysis. Conversions factors were applied to account for volume filtered and percentage of the filter area analyzed. To determine the true filtration diameter, a light microscope was used to examine a 25 mm glass fiber filter (GF/F) filter that had a small amount of dyed water filtered through. It was discovered that the filter funnel blocks roughly 12% of the 25-mm filter and the filtered region had a diameter of 22 mm. Equations 1-5 show the equations used to calculate cell width, biovolume, ESD (equivalent spherical diameter) and biomass, where * implies we assumed that cell height = cell width/2. ESD was used as a consistent measure of mean cell size since many plankton have an irregular shape. The height of a cell was assumed to be roughly equivalent to half of the cell width since cells are often flattened during filtration (Taylor et al., 2011) with the exception of diatoms. The biomass of diatoms (which were the only taxon we could conclusively identify) was estimated allometrically using equation 5 while all other cell biomass (non-diatoms) was estimated allometrically using equation 4. (Menden-Deuer and Lessard, 2000).

Equation 1: Cell width = (4/π) x (Area of the cell/Feret length of the cell)

Equation 2: Biovolume = (4/3)(π) x (Feret Length/2) x (Cell Width/2) x (Cell Height*/2)

Equation 3: ESD = 2 x (3 x Biovolume/4π)^(1/3)

Equation 4: Biomass (non-diatoms) = 0.216 x Biovolume^0.939

Equation 5: Biomass (diatoms) = 0.288 x Biovolume^0.811

For every cycle and depth, biomass size spectra were calculated for nanoplankton and microplankton biomass. Two different pore sizes were used to capture a range of cell sizes; 0.8 µm pore size filters captured 2 to 12 µm cells and 8 µm pore size filters captured >12 µm cells. First only cells that were in the size range of 2 to 12 µm were included for the 0.8 µm samples and cells >12 µm were included for the 8 µm samples. Next, biomass was calculated into units of pg C/mL by dividing biomass by the volume filtered and the conversion ratio. Then each biomass value was separated by ESD into different size bins and summed in each size bin. Size bins were created in octave spacing, e.g. 0.25 to 0.5 µm, 0.5 to 1 µm, 1 to 2 µm, etc., where the width of a size bin would be the difference of the values, e.g. size bin of the 1 to 2 µm would be 1 µm. A similar approach was done for abundance size spectra calculations for nanoplankton and microplankton abundance. Only cells in the appropriate ranges (2-12 µm for 0.8 µm; >12 µm for 8 µm samples) were sorted by ESD into different size bins, counted and then divided by the volume filtered and the conversion ratio. Next both sample sets would be combined for each size bin to give one set of values. Cells less than 2 µm were analyzed using flow cytometry (also shown in this project dataset).

Problems/Issues

Proflavine overstained cells to the point that we were not able to confidently differentiate heterotrophic cells from autotrophic cells, therefore only biomass and abundance values were able to be calculated. Several slides broke during travel in cycle 3 and 4 therefore there is data missing in depth profiles and only slides that had both 0.8 µm samples and 8 µm samples were included in the analysis. In addition, some 8 µm samples had higher biomass and slow filtration rates therefore a known amount of volume was removed from the sample to speed up filtration time and to avoid over-layering cells. This is why some volume filtering times are 325 mL in place of 400 mL. This was not an issue with 0.8 µm samples as 50 mL filtered at a normal rate for all samples.