Summary The maintenance of habitat heterogeneity in agricultural landscapes has been promoted as a key strategy to conserve biodiversity. Animal response to grassland heterogeneity resulting from spatiotemporal variation in disturbance is well documented; however, the degree to which edaphic variation generates heterogeneity detectable by grassland wildlife has proven more difficult to study in natural settings. We conducted a field experiment to study how soils directly affect vegetation structure and composition and indirectly affect bird and butterfly assemblages using plantings of tallgrass prairie species managed as agroenergy crops in Iowa, USA. The experimental design included four vegetation treatments of varying species richness replicated on three soil types. Habitat characteristics varied widely among soils. Crops on sandy loam, the driest, most acidic soil with the lowest nutrient content, developed shorter, less dense vegetation with sparse litter accumulation and more bare ground compared to crops on loam and clay loam. Birds and butterflies responded similarly to soil‐induced variation in habitat characteristics. Their abundance and species richness were similar on all soils, but their assemblage compositions varied among soils in certain vegetation treatments. In low‐diversity grass crops, bird assemblages using sandy loam were dominated by species preferring open ground and sparse vegetation for foraging and nesting, whereas assemblages using loam and clay loam were dominated by birds preferring tall, dense vegetation with abundant litter. In high‐diversity prairie crops, the species composition of forbs in bloom varied among soils and strongly influenced butterfly assemblages. Synthesis and applications. Prairie agroenergy crops established with identical management practices developed variable habitat characteristics due to natural edaphic variation, and this heterogeneity influenced the spatial distribution of bird and butterfly assemblages due to differential habitat use among species. This finding suggests that if unfertilized prairie crops were grown for agroenergy by land managers large‐scale, soil‐induced habitat heterogeneity would promote wildlife diversity within and among fields, further increasing the habitat value of these crops compared to the fertilized, annual monocultures that currently dominate the agricultural landscape. Our study also highlights the need for managers to consider soil properties when selecting sites to restore grassland habitat for species of conservation concern.

Landscape homogenization and the removal of riparian areas have altered stream ecosystems worldwide. Numerous conservation programmes attempt to improve water quality and increase instream habitat heterogeneity to elicit desired biological responses. However, the effectiveness of many conservation efforts on isolated stream fragments remains unknown, especially in grassland regions. The effects of grassland conservation practices and the re‐establishment of riparian corridors in the James River basin, South Dakota (USA) on stream water quality, habitat availability and aquatic macroinvertebrate and fish assemblages were studied in an agriculturally dominated prairie landscape. Grassland conservation efforts may have repaired riparian condition, reduced turbidity and created more diverse instream habitat complexes at conservation sites based on comparisons with paired reference reaches. Reference sites were relatively homogeneous, with prevalent siltation, bank erosion and disturbances to the riparian vegetation. Owing to significant riparian vegetation development, overhanging and aquatic vegetation, benthic detritus and woody materials were significantly more common at conservation reaches. Restoration efforts that assume ‘if you (re‐)build it, they will come’ (i.e. the ‘field of dreams’ hypothesis) underestimate other important barriers to biodiversity restoration in dynamic, grassland riverscapes. Although aquatic organisms in grassland ecosystems are adapted to rapidly inhabit available habitats, the development of niche space at conservation reaches did not directly result in colonization by aquatic life. Grassland management actions did not address stream connectivity issues or overcome land use influences elsewhere in the riverscape that may govern the responses of aquatic macroinvertebrates and fish. Stream fragmentation and continuing, damaging land use patterns seemed to exceed the positive effects of restoring isolated stream reaches in these heavily degraded catchments. Catchment‐scale management strategies that combine reach‐level restoration actions with efforts to improve connectivity are likely to be more successful in degraded riverscapes.

No abstract