Grassland butterfly responses to prairie restoration : effects of prescribed fire and grazing

Fire and grazing are common methods used for prairie restoration and conservation. However, effects of restoration on grassland invertebrates have been evaluated with mixed results. We examined the effects of prairie restoration through fire and grazing and the relative contributions of the direct and indirect effects of time since fire on the grassland butterfly community of the Loess Hills of Iowa, USA. Both total and habitat-specialist (prairie-dependent) butterfly abundance were highest on prairies that were managed with grazing and burning, and lowest on those that were only burned. Butterfly species richness did not differ among any of the management types. Responses of individual butterfly species to management practices were variable. In the best predictive models, both habitat-specialist and total butterfly abundance were negatively associated with the percent cover of bare ground, total butterfly abundance was positively associated with the percent cover of forbs, and habitat-generalist butterfly abundance was positively associated with floral resource availability. Areas managed with fire, grazing, or a fire/grazing combination all maintained equally species rich, yet compositionally different, butterfly communities. Butterfly abundance increased as time since burn increased. The percent cover of warm season grasses and bare ground decreased while the cover of cool season grasses, forbs, and litter depth increased with time since burn. We used path analysis to examine direct and indirect effects of burning. For habitat-specialist species abundance path models, the total indirect effects of time since burn through floral resources, warm season grass cover, or bare ground were stronger relative to the direct effect of time since burn. The indirect pathway through bare ground had higher relative strength than other indirect paths in the models for habitat-specialist species. For the habitat-generalist species path models, the direct effect of time since burn was stronger relative to the indirect effects. Because of this variation in butterfly species responses to different management practices, we do not recommend a single type of management that would benefit all species or even all species within habitat-specialist or habitat-generalist habitat guilds. Our data illustrate the profound effects, both direct and indirect, of fire on grassland butterfly abundance