Each site was visited in the spring (April-May), summer (June-July), and fall (September) periods. Exceptions include MIS and OAK, which were only sampled during the summer period. During each visit, a YSI multiprobe (Yellow Springs Instruments, Inc., Yellow Springs, OH, USA) was used to measure temperature, specific conductance, and percent dissolved oxygen. Due to multiprobe malfunction, data from a summer 2021 YSI deployment was used to characterize MIS water parameters. In addition to YSI parameters, 250 mL acid-washed Nalgene bottles were used to collect water samples for nutrient analyses at each sampling point. Each water sample was subsampled into two vials, of which one was refrigerated, and one was frozen within 24 h of collection. The refrigerated subsample was used to determine orthophosphate (SRP) concentrations using USEPA method 365.1 (O’Dell 1996). The frozen subsample was used to determine dissolved silica concentrations using USEPA method 370.1 (USEPA) and chloride, sulfate, and nitrate using USEPA method 300.0 (Pfaff 1993).
Mats from wadable sites were collected using a suction device and placed in sterile Whirlpak® bags, then put on ice for transport to the Annis Water Resources Institute (AWRI, Muskegon, MI, USA). Three replicate mat samples were collected from each habitat type at each site during each sampling event. Mats from MIS were collected by NOAA divers using a coring device, and transported to AWRI as cores in plastic tubes on ice. Plankton tow samples were also collected at GSS and ECB to determine taxa that may be considered part of the surrounding planktonic community, rather than active members of the microbial mat community. Each mat sample collected was subsampled, with one subsample used for generating unialgal cultures and the other for metabarcoding.

Subsamples for metabarcoding were frozen at -80 °C within 36 h of collection, except for MIS samples which were stored at 10 °C for 72 hours prior to harvesting, then frozen at -80 °C, due to logistical limitations. DNA was extracted from the metabarcoding subsamples using the Qiagen PowerSoil DNA Extraction Kit (Qiagen, Crawley, UK) according to the manufacturer’s protocol, with a negative control consisting of autoclaved nanopore water included for each subset of extractions and for each primer to assess potential processing contamination. To prepare samples for Illumina amplicon sequencing, a two-step PCR approach was employed. The initial PCR was completed to amplify the two barcode markers (rbcL and 16S) in individual reactions using specific primers with the attached Illumina adapter. The primary PCR amplification was completed in 25 μl reactions using 12.5 μl of Q5 High-Fidelity2X Master Mix (New England BioLabs Inc., Ipswich, MA, USA), 1.0 μl of each primer (1μM), 9.5 μl RNase-free H2O, and 1 μl DNA. For the 16S marker, the primer pair and thermocycler protocol from Walters et al. (2015) were employed. For the rbcL marker, we targeted a 312 bp region of the rbcL plastid gene using an equimolar mix of the three forward and two reverse degenerate primers from Vasselon et al. (2017), along with the thermocycler protocol.

Following PCR amplification, samples were sent to the University of Tennessee, Knoxville for processing and sequencing. PCR products were cleaned with Agencourt AmPure XP beads (Beckman Coulter Inc., Indianapolis, IN, USA) and quantified using a Qubit Fluorometer (v.2.0; ThermoFisher Scientific, Waltham, MA, USA). Samples were normalized, and a second PCR reaction (50 μl) enriched with Q5 High-Fidelity 2X Master Mix was performed to apply indexing primers, following cycling conditions: 95° C for 3 min followed by 10 cycles of 95° C for 30 s, 55° C for 30 s, 72° C for 30 s, with a final extension of 72° C for 5 min, modified from the 16S protocol (Illumina 2013). A second PCR clean-up was performed, and samples were quantified using a Qubit Fluorometer. Libraries were loaded with 25% PhiX clustering control on the Illumina MiSeq platform for 300 bp × 2 paired end reads using the V3 kit.