For each replicate, we selected one rock from a riffle habitat with cobbles of uniform shape. We covered the rock with aluminum foil before placing the rock in a sample bag with stream water. We transported the rock samples back to the lab in the dark on ice. In the lab, periphyton was removed by brushing the rocks with a toothbrush and rinsing with deionized water. Twenty mL subsamples from the slurry were preserved using 7% formalin solution for diatom community analysis.
To determine diatom densities and relative abundances, 5-10 mL aliquots were heated with 30% hydrogen peroxide for one hour and then rinsed six times with distilled water to remove oxygen by-products.We then mounted the coverslips to microscope slides with Naphrax mounting medium (Brunel Microscopes, Wiltshire, U.K.). We identified 500 valves per sample using a Zeiss Universal microscope (2005-2006 samples) and Leica DMRXE microscope (2018-2019 samples) under oil immersion at 1000x using brightfield optics.
In samples with extremely low densities, a maximum of 180 mm of transects were counted. We used standard taxonomic references (Patrick and Reimer 1966, 1975; diatoms.org) and algal floras from the southeastern U.S. (Camburn et al. 1979, Kociolek and Kingston 1999, Thomas et al. 2009, Furey et al. 2011, Bishop et al. 2017) for species-level identification. In order to ensure taxonomic harmonization of identifications between the 2005-2006 samples and the 2018-2019 samples, we defined “species complexes” for related species that were difficult to differentiate with light microscopy through extensive consultation between the researchers who identified the pre-die off (R. Bixby) and more recent post-die off (K. Solomon) sample sets. Species complexes included Encyonema minutum, Eunotia metamondon, Eunotia minor, Eunotia rhomboidea, Frustulia rhomboides, and Synedra rumpens (Appendix, Table A.1), and will hereafter be referred to by the species name.
To determine individual cell biovolumes, we measured approximately ten cells of each species and averaged the biovolumes of the ten individuals to determine a cell biovolume for each taxon (Hillebrand et al. 1999). In some cases where we were not able to measure 10 individual cells, we obtained biovolume values from the literature (diatom.ansp.org, Lowe and Pan 1996, Dye 2005). To determine cell density (cells mm-2) for each taxon, we divided the number of cells of each taxon by the fractional volume of the sample (i.e., volume measured/ total volume of slurry), and converted to a per-area measure by dividing the measured value by the proportion of the total rock area sampled. To determine cell biovolume (mm3 m-2) of each taxon, we multiplied the cell biovolume of each taxon by its density in the sample. To determine total cell density and total cell biovolume of each sample, we summed individual cell densities and biovolume across samples, and we considered total cell density and total cell biovolumes of each sample to be alternate measures of algal stand crop.
We assigned individual species to one of three functional groups -- low, high, motile -- based on the definitions of Rimet and Bouchez (2012), except in cases where our own or published observations of growth forms of individual species differed from that of Rimet and Bouchez (2012). Deviations from Rimet and Bouchez (2012) definitions were as follows: we classified Eunotia spp. with a cell biovolume less than 500 µm and a valve length to width ratio of less than eight as low instead of high, because we only observed these taxa singularly or in chains of two cells in our samples. We classified Meridion spp. as high because we observed these taxa erect on substrates and noted similar observations of these taxa by Lowe et al. (1986). We classified Frustulia spp. as motile instead of high because we only observed these taxa as single cells and not mucilaginous tubes. Finally, we classified Brachysira spp. as motile instead of low because of their “moderate motility” (Hamilton 2010) and classified Chamaepinnularia spp. as low instead of motile because of their “weak motility” (Tyree 2018).
