Assessing multiple data sets for habitat suitability modeling of bats across Illinois

Gaulke, Sarah M.

Permalink

https://hdl.handle.net/2142/115780 Copy

Description

Title

Assessing multiple data sets for habitat suitability modeling of bats across Illinois

Author(s)

Gaulke, Sarah M.

Issue Date

2022-04-29

Director of Research (if dissertation) or Advisor (if thesis)

Davis, Mark A

Committee Member(s)

• Larson, Eric
• O'Keefe, Joy
• Hohoff, Tara

Department of Study

Natural Res & Env Sci

Discipline

Natural Res & Env Sciences

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• bats
• habitat suitability modeling
• landscape ecology
• forest management
• hoary bat
• tri-colored bat
• eastern red bat
• Indiana bat
• bat conservation
• Lasiurus borealis
• Lasiurus cinereus
• Myotis sodalis
• Perimyotis subflavus

Abstract

North American bat populations have been severely and negatively impacted by numerous factors, including habitat loss and fragmentation, disease, and wind energy development. Yet bats provide critical ecosystem services, and are thus a focus of habitat conservation and management. As wide-ranging flyers, bats use habitats at a variety of scales, from small, isolated patches for roosting to large, contiguous corridors for migration. Landscape-level research is necessary to identify critical habitats, patches, and corridors to target management interventions. Habitat suitability models (HSMs) identify high quality habitat by predicting species occurrence at various spatial scales based on occurrence data and environmental variables. Bat occurrence data are mainly collected by mist netting or acoustics. The North American Bat Monitoring Program (NABat), a national monitoring protocol, provides a new data repository for developing HSMs. By combining NABat data with historical data, I can compare model performance by data type, which is essential for effective modeling. In this thesis, I seek to identify where suitable bat habitat is available across the state and compare the impact of different detection methods on HSM. First, I created Maxent HSMs for three bat species (hoary bat, eastern red bat, and tri-colored bat) across Illinois using species-specific landscape and climate variables. With the three models from this study and a previously published HSM for Indiana bats, I stacked the binary HSMs identifying priority conservation areas across Illinois I found that each species exhibited different distributions and habitat usage across Illinois. Stacking the HSMs highlights shared high-quality bat habitat in southern IL and along riparian areas. Identifying quality conservation areas allows managers to prioritize restoration and conservation and use available funds on the most effective habitat, especially as energy companies look for mitigation lands to purchase. Secondly, I sought to understand how different data types can influence HSM. I compared the overlap of models created from passive-only, active-only, and combined occurrences to identify the effect of multiple data types and detection bias. Passive data involves sensing the species remotely, while active detection involves handling the animal. For each species, the data type with the highest AUC value was the active-only model. By comparing the niche overlaps of HSMs between data types, I found a high amount of variation with no species having over 45% overlap among models. Passive models showed more suitable habitat in agricultural lands, while active models showed higher suitability in forested land, a reflection of sampling bias. Overall, this emphasizes the need to consider influences of detection and survey biases on modeling, especially when combining multiple data types. Biases from sampling, behavior at time of detection, and species life history intertwine to create striking differences among models. The biases and effects of each detection type should be considered in the final model output, particularly when the goal is to inform management decisions, as one data type may support very different interventions than another. Lastly, I created HSMs using only data from NABat’s acoustic protocol to compare to a model from all occurrences to understand if the data produced by the protocol were sufficient to create a robust HSM. I found that the NABat model was heavily impacted by acoustic data bias and did not create a robust enough model compared to a combined model, arguing for the inclusion of more occurrence records to create a model sufficiently robust enough to inform management decisions. Ultimately, I undertook a rigorous assessment of how various data types perform in the HSM ecosystem and provide recommendations for best practices of developing habitat models for bats using disparate data sources.

Graduation Semester

2022-05

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Sarah Gaulke

Owning Collections