A multi-scale analysis of species–environment relationships: rare moths in a pitch pine–scrub oak ( Pinus rigida–Quercus ilicifolia) community

Abstract

Community level analyses of species–environment relationships can provide critical information for conservation planners. A comprehensive analysis of the factors that influence species distributions should include variables measured at multiple scales because species–environment relationships are known to vary with the scale of observation. The pitch pine–scrub oak ( Pinus rigida–Quercus ilicifolia) communities, or pine barrens, of the northeastern USA are severely threatened by development and fire suppression. They also provide critical habitat for many species of rare moths. We used partial canonical correspondence analysis to assess the relative effects of three levels of environmental variables (plot, patch, and landscape) on the distribution and abundance of 10 species of rare moths in a pine barrens community in southeastern Massachusetts, USA. We also used a set of spatial variables to quantify and partial out the effects of spatial autocorrelation of species composition among sampling locations. All three levels of environmental factors combined, independent of spatial factors, accounted for virtually half (48.4%) of the total variation in the moth community. Sequential partitioning of the variance explained by each level of environmental factors indicated that landscape level factors explained more than twice as much variance as plot and patch level factors. Another environmental model that included only landscape level variables explained 53% of the total variation in the moth community. Patch density and percentage of the landscape comprised of open and sparse canopy, scrub oak habitats were the most significant variables. These results suggest that the presence of scrub oak habitat within relatively large, heterogeneous landscape mosaics may be more important for the maintenance of many rare pine barrens associated moth populations than plot or patch level characteristics.