Understanding how the Earth's surface (i.e., ‘nature’s stage’) influences connections between biodiversity and ecosystem function (BEF) is a central objective in ecology. Despite recent calls to examine these connections at multiple trophic levels and at more complex and realistic scales, little is known about how landscape structure shapes BEF relationships among animal communities in nature. We coupled high-resolution habitat mapping with extensive field sampling to quantify connections among the geophysical habitat templet, invertebrate assemblages, and secondary production in two large North American riverscapes. Patterns of sediment size governed invertebrate assemblage structure, with particularly strong effects on composition, richness, and taxonomic and functional diversity. These relationships propagated to drive positive relationships between biodiversity and secondary production that were modified by scale, context-dependencies, and anthropogenic modification. Finally, leveraging spatially explicit descriptions of geophysical and biological properties, we uncovered distinct and nested spatial scales of biodiversity and secondary production, and suggest that multiple geophysical processes simultaneously influence these patterns at different scales. Together, our findings advance our understanding of relationships between the physical templet and patterns of BEF, and help to predict how perturbations to the Earth’s surface may propagate to influence biodiversity and energy flux through food webs.

This dataset contains habitat information from the Missouri and Yellowstone Rivers that was collected in 2013-2014 via side scan sonar, geoprocessed in ArcMap, and organized and analyzed in Microsoft Excel and R. Additionally, 25-30 invertebrate samples were collected from 6 reaches in the Missouri and Yellowstone Rivers throughout the year from a wide variety of habitats in 2014-2015. More specifically, the methods used to quantitatively sample invertebrate communities were as follows: main-channel sand habitat: a Ponar dredge sampler (0.052 m²) attached to a sounding reel on a boat; cobble/gravel: a Hess sampler (0.086 m²); wood: 250 μm mesh bags and scrubbed; depositional off-channels: a stovepipe core (0.031 m²); and large boulders: scrubbed in buckets while a D-frame dip net was held downstream. Samples were preserved with buffered formalin and rinsed onto nested sieves (1 mm and 250 um) in the laboratory. Invertebrates were enumerated and measured to the nearest mm to estimate abundance, biomass, and secondary production.

All results and figures can be recreated with Microsoft Excel and R.