Abstract
dc:description.abstractBat populations across eastern North America have experienced substantial declines due to white-nose syndrome, habitat alteration, and ongoing forest management pressures, increasing the need for accurate and locally informed approaches to monitoring and conservation. Acoustic surveys have become one of the most widely used tools for assessing bat distribution and habitat use; however, differences in sampling effort, automated call identification filtering, and statistical modeling approaches may substantially influence ecological inference and management recommendations. In this dissertation, I evaluated landscape-scale habitat associations, temporal activity patterns, and methodological considerations for acoustic monitoring of bats in managed forest landscapes of Pennsylvania and Maryland to provide evidence-based recommendations for improving acoustic survey design, data processing, and management decisions. First, I compared multiple approaches for modeling acoustic detections using generalized linear mixed models and eight acoustic data handling techniques for a rare species, the northern long-eared bat (Myotis septentrionalis), and a common species, the eastern red bat (Lasiurus borealis). I found that filtering acoustic detections using maximum likelihood estimate (MLE) thresholds from automated identification software improved prediction accuracy for rare species models relative to unfiltered data or highly conservative match-ratio thresholds. In contrast, the most conservative filtering approach produced the greatest predictive accuracy for eastern red bats, demonstrating that optimal acoustic processing approaches may differ among species according to detectability and prevalence. These findings suggest that managers should adopt species-specific acoustic data processing strategies rather than relying on a single filtering approach. For rare or difficult-to-detect species, such as federally listed bats, nightly maximum likelihood estimator (MLE)-retained detections provided the most reliable balance between minimizing false-positive identifications and retaining biologically meaningful observations. For common, readily detected species, more conservative filtering approaches may further improve predictive performance. Because species abundance and call characteristics vary geographically, practitioners should incorporate cross-validation during model development to identify the most appropriate data handling approach for their monitoring objectives. Second, I investigated multi-year landscape-scale habitat associations for nine bat species across Pennsylvania and Maryland. Several species exhibited distinct habitat relationships, including greater big brown bat (Eptesicus fuscus) activity along field edges and greater Lasiurus borealis activity in xeric conditions relative to hydric or mesic sites. Most species demonstrated relatively stable activity over the survey period despite low detections for several cave-obligate species affected by white-nose syndrome. Seasonal variation in acoustic activity suggested relationships with reproductive phenology (e.g., pregnancy, lactation, and juvenile volancy periods). Contrary to expectations, McNab's Landform Index was not strongly associated with activity for most species, including M. septentrionalis. Results also indicate that acoustic surveys conducted throughout the currently recommended survey window (May 15-August 15) better capture seasonal changes associated with pregnancy, lactation, and juvenile volancy than shorter sampling periods. Surveying both early and late portions of the season reduces the likelihood of incorrectly concluding that species are absent because life-stage-specific activity patterns were missed. Something USFWS should consider is requiring sampling in both early and late seasons to account for intraseasonal variation of activity. Finally, I assessed the influence of sampling effort and survey design on habitat association models by standardizing acoustic detector placement within a landscape-scale sampling grid. Results demonstrated that poorly matched survey effort and inconsistent detector deployment can generate uninformative or misleading habitat models, potentially biasing management recommendations and conservation actions. Collectively, this work demonstrates the importance of standardized sampling design, species-specific acoustic data filtering, and locally focused multi-year monitoring for improving the reliability of bat habitat association models and informing conservation and forest management decisions in eastern North America. I recommend that practitioners adopt MLE-retained detections for rare species, though further evaluations should be conducted to determine if these finding are appropriate range-wide. I also recommend to maintain acoustic studies across the full May 15-August 15 survey period (or beyond to capture true phenology) and increase sampling intensity where feasible to improve inference regarding species-environment relationships.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Fisheries and Wildlife Science
- Department dc:contributor.department
- Fish and Wildlife Conservation
- Grantor dc:publisher
- Virginia Tech
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Isenhour, Zackary William
- Chair dc:contributor.committeechair
-
- Ford, William Mark
- Committee members dc:contributor.committeemember
-
- Rijal, Santosh
- Hunter, Elizabeth Ann
- Karpanty, Sarah M.
Subjects
dc:subject × 7Rights
dc:rights- Statement dc:rights
-
- In Copyright
- Licence dc:rights.uri
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Other
- vt_gsexam:47513