Loss of grasslands in the Great Plains has displaced native bee species through the conversion of natural habitat to cropland. The federally-supported Conservation Reserve Program (CRP) converts marginal cropland to restoration plantings that can serve as reservoirs for native species, but few studies have examined the success of these lands in supporting native bee communities. We sampled bee and floral communities on 108 CRP fields across sand prairie and short-, mixed-, and tall-grass habitats (spanning 650 km and 43-109 cm of precipitation annually). In addition to prairie habitat, our research included a grazing and a restoration treatment incorporating a low-diversity planting (CP2) and a higher-diversity, forb-focused planting targeting the development of pollinator and wildlife habitat (CP25). With regard to bee abundance and genus richness, we examined the effects of: 1) restoration type, grazing, and prairie habitat and 2) flowering forb cover, species richness and site-level floral community floristic quality (FQI). We predicted bee abundance and richness would be greater on fields restored with CP25 plantings as well as on grazed fields due to increased heterogeneity of the plant community. We used multi-model inference to determine the relative importance and effect of predictors on total bee abundance and genus-level richness.
Results/Conclusions
In 2018 we sampled fields twice over a 10-week period and collected 4299 bees in 33 genera associated with 136 forb species. For bee abundance, the relative cover of flowering forbs, CP, and prairie habitat occurred in all selected models (DAICc< 4) and had the highest relative importance values (0.91 for each). The best model included only these predictors, explained 33.0% of the variation in bee abundance, and was twice as likely to be the best model compared with the next most-supported model (ER=2.01) which did not include the effect of floral richness. All selected models predicting bee richness included the effect of relative floral cover, which was the most important predictor (importance=0.91). The top model included only relative forb cover and floral richness (importance=0.65) and explained 25.0% of the variation in bee richness. The evidence ratio (ER) comparing the top and second-most supported models was 1.2 and the penultimate model included the effect of CP, an equally important predictor as grazing (importance=0.40). Preliminarily, our results suggest that native bee communities would benefit from increased flowering forb cover in restored grasslands, thus enhancing pollination services to native species and adjacent croplands in predominately agricultural landscapes.