Environmental conditions
We estimated the number of dry/ponded and flowing days at each intermittent site, using photo trap cameras, taking twice-daily picture of the river channel. From these pictures we determined the flow permanence (FP; proportion of days flowing during the study period) and the total duration of flowing days during each sampling campaign at each site. We also calculated spatial variables representing site connectivity using geographic information system analyses and digitized network maps that included information on the permanence regime of each reach (i.e. intermittent vs. perennial), based on long-term observations (Datry 2012, Gauthier et al. 2020). To estimate the site location within the network we calculate the distance to the source (DS). We measured distances to the nearest perennial site for intermittent reaches. To estimate upstream connectivity; i.e. the amount of perennial reaches potentially serving as colonization source, we calculated the length, and the percentage of all perennial reaches upstream (%UC) from each site (see e.g. Sarremejane et al. 2017). For graphical representation and analyses, sites were sometimes divided into flow regime categories: perennial (FP = 1; n=10) and intermittent (FP < 1; n=10) and into network location categories: headwaters (DS < 10km, n=14) and mainstem (DS > 10km, n=6).
We also measured a set of environmental variables at each site and in each season. When water was present, pH, conductivity and oxygen concentrations instream were measured using a multiprobe (Hach) and total dissolved nitrogen and soluble reactive phosphorus (SRP) were measured from filtered water samples in the laboratory. Dissolved nitrogen was determined in a TOC-Analyzer with Nitrogen-Analyzer (Shimadzu) and soluble reactive phosphorous (SRP) was determined with the molybdate blue reaction (spectrophotometer U-3900H, Hitachi). Flow velocity was determined using a flowmeter at 10 to 30 (depending on channel width) random locations at each reach. The wetted width was averaged across 10 transects. Instream bryophyte and substrate (bedrock, boulders, cobbles, gravel and sand) cover was determined visually. Riparian soil and instream sediment samples were dried (70⁰C) and burnt (550⁰C) to ashes in the laboratory to determine moisture and organic matter content, respectively. In summer, we estimated canopy cover in the instream and in the riparian area at each site, using a vertical tube (15 cm diameter) at 10 random locations. We also identified all ligneous plants located in the riparian area (<5m from the stream) and estimated their relative density and the total tree richness at each site. Temperature was recorded hourly for the duration of the study instream and in the riparian area using iButton dataloggers (Maxim Integrated Products, USA). Site altitude above sea level was also determined from a digitized elevation model.