Restoring Lepidopteran Communities to Oak Savannas: Contrasting Influences of Habitat Quantity and Quality

Authors


 Address correspondence to K. S. Summerville, email keith.summerville@drake.edu

Abstract

Ecological restoration is deemed important for the long-term conservation of biodiversity, but ecologists still lack an understanding of how habitat availability and habitat quality in a restored system interact to determine species diversity. This problem seems particularly apparent in Tallgrass Prairie and savanna ecoregions, where restored management units represent the majority of extant habitat. In this study, we tested three principal hypotheses, each stating that the diversity of Lepidoptera would be greater in (1) patches of savanna habitat that were larger; (2) patches that were of higher habitat quality; and (3) patches that had greater connectivity to management units of similar physiognomy. Lepidoptera were sampled in 2003 from 13 unmanaged woodland remnants within Neal Smith National Wildlife Refuge, a 2,292-ha prairie and savanna reconstruction project. We also measured 11 environmental variables within each site to assess variation in habitat quantity and quality. Principal components analysis (PCA) was used to identify major gradients of environmental variation among the 13 sites. Our PCA differentiated among woodlands along three environmental gradients, defined by (1) stand size, shape, topography, and oak dominance; (2) degree of disturbance; and (3) isolation. Total lepidopteran species richness, however, was only predicted by variation in the first principal component. Species richness of Lepidoptera known to be oak specialists was significantly affected by variation along all three PCA gradients. Surprisingly, more isolated woodland remnants contained a greater richness of oak feeders. Our results suggest that approaches to restoring oak savannas should emphasize aspects of both habitat quantity and quality. Beyond making individual management units larger, priority sites for restoration should possess a low importance of trees that are indicative of past habitat disturbance (e.g., Honey locust, White mulberry) even if canopy closure is substantial. Connectivity among restored habitats may benefit savanna moth communities only when habitat linkages contain a flora similar in composition to focal patches.

Ancillary