UVM Theses and Dissertations
Format:
Print
Author:
Farrell, Laura Emily
Dept./Program:
Biology
Year:
2012
Degree:
PhD
Abstract:
Landscape connectivity is integral to the persistence ofmetapopuhitions of carnivores and other terrestrial animals. This work investigates the landscape characteristics essential to use of areas by lynx and bobcats in northern New England, and explores the persistence ofconnectivity across areas ofthe region likely to experience future development pressure. A Mahalanobis distance analysis was conducted on telemetry data collected from 16 bobcats in western Vermont and 31 lynx in northern Maine to determine which variables remained most consistent across locations at local, daily distance moved, and home range scales. The model providing the widest separation between habitat and the landscape for bobcats suggests that they cueinto landscapefeatures suchasedge, availability of cover, and developmentrather than specific habitat types. The selected lynx model contained natural habitat, cover, elevation, and shrub scrub habitat at the daily distance moved scale-where their favored prey is most abundant.
Cross validation indicated that outliers had little effect on models for either species. Present connectivity within selected areas across the region is largely intact. A projection of future landscape change illustrates potential impacts of anthropogenic development on connectivity by approximately 2050. Reduction of connectivity is pronounced in areas where development pressures are concentrated, and indicates where conservation efforts might be focused in order to maintain connectivity across the landscape into the future. However, in order to plan a network of connective habitat knowledge of where species are located is critical, and this can only be achieved by monitoring targeted areas.
Detection probability should be assessed before designing and implementing species moniotring, yet there is no guideline on detection rates for many Northeastern mesomammal species, and no established regional survey methodology. To investigate detection rates and their variability across seasons and years, arrays of cameras and track plates were monitored at 21 sites over 7 seasons. Weather led to high rates of missing data during winter 2008-09, indicating that winter in this region would not always provide accurate data on detection probabilities. While cameras provided the bulk of detections for most species, a multimethod approach lowered latent time to detection and increased the number of sites where detected for 5 species through the study. Longer term data increased the number of detection sites for all species. Increasing camera survey duration from 3 to 14 days and the number of visits to 5 improved detection for each of the 13 species, and added detection sites for all species every season they were detected.
Long term field data were projected to determine optimum survey methodologies that obtain 95% probability of detection by simultaneously manipulating the number of cameras and track plates at a site over a varying number of visits. Results suggest that 95% probability of detection at a site is achievable for 11 species within 1 season, when either time to detection or expense are prioritized. Surveying over more than one season increased the number of sites where a1l species except for raccoons were detected. Results provide researchers with a guideline in planning further detection based surveys of northeastern mesomammals. Robust detection rates may increase the accuracy and precision of maximum likelihood based studies, and provide land use planners and managers with a tool to more reliably establish use of a site by a species.
Cross validation indicated that outliers had little effect on models for either species. Present connectivity within selected areas across the region is largely intact. A projection of future landscape change illustrates potential impacts of anthropogenic development on connectivity by approximately 2050. Reduction of connectivity is pronounced in areas where development pressures are concentrated, and indicates where conservation efforts might be focused in order to maintain connectivity across the landscape into the future. However, in order to plan a network of connective habitat knowledge of where species are located is critical, and this can only be achieved by monitoring targeted areas.
Detection probability should be assessed before designing and implementing species moniotring, yet there is no guideline on detection rates for many Northeastern mesomammal species, and no established regional survey methodology. To investigate detection rates and their variability across seasons and years, arrays of cameras and track plates were monitored at 21 sites over 7 seasons. Weather led to high rates of missing data during winter 2008-09, indicating that winter in this region would not always provide accurate data on detection probabilities. While cameras provided the bulk of detections for most species, a multimethod approach lowered latent time to detection and increased the number of sites where detected for 5 species through the study. Longer term data increased the number of detection sites for all species. Increasing camera survey duration from 3 to 14 days and the number of visits to 5 improved detection for each of the 13 species, and added detection sites for all species every season they were detected.
Long term field data were projected to determine optimum survey methodologies that obtain 95% probability of detection by simultaneously manipulating the number of cameras and track plates at a site over a varying number of visits. Results suggest that 95% probability of detection at a site is achievable for 11 species within 1 season, when either time to detection or expense are prioritized. Surveying over more than one season increased the number of sites where a1l species except for raccoons were detected. Results provide researchers with a guideline in planning further detection based surveys of northeastern mesomammals. Robust detection rates may increase the accuracy and precision of maximum likelihood based studies, and provide land use planners and managers with a tool to more reliably establish use of a site by a species.