3.1 Effect of sampling day, sampling position, and environment  
There was a significant relationship between sampling day and both species (F = 5.933, p-value = <0.001, ndf (numerator degrees of freedom) = 4.055, ddf (denominator degrees of freedom) = 90.945, R-sq. (adj.) = 0.336) and OTU (F = 2.765, p-value = 0.030, ndf = 3.32, ddf = 91.68, R-sq. (adj.) = 0.103) richness. In contrast, neither temperature, discharge, nor sampling position had an effect on the number of species or OTUs detected. Highest species and OTU richness was detected in spring with lowest richness in the summer months for all three sampling positions (Figure 2 and Figure S3). The vast majority of macroinvertebrate species was detected at all three sampling positions (280), also with over 99% of the reads assigned to these (Figure 2). Although some species were exclusively found at one of the three sampling positions, relative occurrence of these species was low in contrast to shared species (Figure S4). We found no significant differences between surface, riverbed or riverbank samples based on OTU richness (Kruskal-Wallis rank sum test: chi-squared = 0.085, p-value = 0.957, ndf = 2, ddf = 97).