Basin risk explains patterns of macroinvertebrate community differences across small streams in the Fayetteville Shale, AR

Understanding the relationship between natural landscape characteristics and human activities is crucial to predicting conditions under which loss of freshwater biodiversity occurs. For example, region-specific human activities, like unconventional natural gas (UNG) development, are often absent from models because of their limited spatial extent, yet their effects may alter stream ecosystems. Unconventional natural gas development is an expanding human activity that requires land-clearing, water withdrawal, and chemicals for hydraulic fracturing that could alter stream habitat and water quality, yet the degree of alteration may depend on basin natural landscape characteristics. We modified a basin-specific multi-metric risk model and used it to predict differences in macroinvertebrate communities in small stream basins (range 5.71-90.96 km2) in the Fayetteville Shale, Arkansas. We modeled basin risk as the interaction between basin natural sensitivity to alterations and basin exposure to human activities. We predicted that macroinvertebrate communities in 18 basins exposed to a gradient of unconventional natural gas (UNG) in a pasture-dominated landscape would experience greater differences in macroinvertebrate metrics across a risk gradient than 22 basins with a similar gradient of pasture land. Macroinvertebrate metrics of diversity declined by as much as 88% across the basin risk gradient. In contrast, macroinvertebrate metrics of biomass and density increased by 28% across the same basin risk gradient. Basin risk explained more variation in macroinvertebrate communities than sensitivity or exposure alone, suggesting an interaction between basin natural landscape characteristics and human activities. In contrast to our prediction, all macroinvertebrate metrics responded similarly in basins with and without UNG, which suggests either there was no added stressor-effect of UNG or UNG alterations resulted in the same biological effects as pasture. However, macroinvertebrate metrics in basins with and without UNG responded differently across a sediment-specific risk model where dams were modeled as sediment traps instead of structures that alter flow and temperature. As land alterations continue in the Fayetteville Shale, our basin-specific risk model can be used as a tool to identify at-risk stream communities.