River microbiome composition reflects macroscale climatic and geomorphic differences in headwaters but not in lower reaches

Maintaining the quality and quantity of water resources in light of complex changes in climate, human land use, and ecosystem composition requires detailed understanding of ecohydrologic function within catchments, yet monitoring relevant upstream characteristics can be challenging. In this study, we investigate how variability in riverine microbial communities can be used to monitor the climate, geomorphology, land-cover, and human development of watersheds. We collected streamwater DNA fragments and used 16S rRNA sequencing to profile microbiomes from headwaters to outlets of the Willamette and Deschutes basins, two large watersheds prototypical of the U.S. Pacific Northwest region. In the temperate, north-south oriented Willamette basin, microbial community composition correlated most strongly with geomorphic characteristics (mean Mantel test statistic r = 0.19). Percentage of forest and shrublands (r = 0.34) and latitude (r = 0.41) were among the strongest correlates with microbial community composition. In the arid Deschutes basin, however, climatic characteristics were the most strongly correlated to microbial community composition (e.g., r = 0.11). In headwater sub-catchments of both watersheds, microbial community assemblages correlated with catchment-scale climate, geomorphology, and land-cover (r = 0.46, 0.38, and 0.28, respectively), but these relationships were weaker downstream. Development-related characteristics were not correlated with microbial community composition in either watershed or in small or large sub-catchments. Our results build on previous work relating streamwater microbiomes to hydrologic regime and demonstrate that microbial DNA in headwater streams additionally reflects the structural configuration of landscapes as well as other natural and anthropogenic processes upstream. Our results offer an encouraging indication that streamwater microbiomes not only carry information about microbial ecology, but also can be useful tools for monitoring multiple upstream watershed characteristics.