Understanding the influence of habitat fragmentation on butterfly occurrence patterns: Developing and testing models across global ecosystems

Background/Question/Methods

Human activities such as agriculture, forestry and urbanization have profound effects on natural environments, the most widespread of which is the fragmentation of landscapes. Communities occupying fragmented landscapes experience altered landscape structure, reduced habitat availability and biodiversity losses, potentially resulting in unique community compositions across fragments, depending upon how individual species respond to patch characteristics. Variation in such responses among species highlights the importance of both life-history traits and habitat characteristics in determining community responses to fragmentation. Most studies that attempt to model fragmentation effects, however, are performed on a species-specific basis and often fail to consider the influence of life-history traits. My research investigates fragmentation effects on butterfly distribution patterns across three continents, and addresses the question: Are fragmentation effects consistent across geographically wide-spread locales and distinct evolutionary and ecological contexts? Data from Prague (CZ), Perth (AU), and Denver (CO, USA) are used to 1) statistically model how fragment characteristics and life-history traits govern community membership across patches, and 2) test whether each model predicts community distributions in the other fragmented ecosystems. Good model performance between regions would indicate that fragmentation influences butterfly occurrence patterns similarly, while poor performance would suggest that the influences are unique among communities.

Results/Conclusions

Initial modeling results, using the Prague data, have shown that butterfly distributions are most heavily influenced by the traits of 1) generations per year, 2) diapause strategy, and 3) flight period. Furthermore, these three traits also best explain variation in butterfly occurrences at Colorado sites, suggesting that they may be important drivers in the responses of species to fragmentation, regardless of the community’s geographical location. Here I present the results of initial model development, using Prague data, as well as test results, using Colorado data. Here, I report on model construction, and the preliminary findings for model transferability between study areas.

Results from this work have applied value to the fields of restoration and conservation ecology. First, the documentation of which features influence responses to fragmentation at each study location may contribute to the development of targeted and effective management or reintroduction plans by local organizations. Furthermore, this work explores the possibility of constructing broadly applicable models that transfer among independent communities inhabiting unique settings. The continued development of such generalizable models might allow for the rapid assimilation of information across diverse regions, and contribute to a broad-scale assessment of the impacts of continued fragmentation on communities around the globe.