Date Approved
8-18-2025
Graduate Degree Type
Thesis
Degree Name
Biology (M.S.)
Degree Program
Biology
First Advisor
Paul W. Keenlance
Second Advisor
Michael J. Joyce
Third Advisor
Alexandra B. Locher
Fourth Advisor
Joseph J. Jacquot
Academic Year
2024/2025
Abstract
American martens (Martes americana) are closely associated with complex forest structure. Current and historical forest alterations have significantly reduced or destroyed critical habitat for many wildlife species. Previous studies on marten foraging have primarily focused on prey abundance, while relatively little is known about habitat factors that affect marten foraging. Mapping fine-scale complex forest elements provides critical insight into the importance of different habitat elements.
In Chapter 2, we first define size classes of coarse woody debris (CWD) associated with marten foraging efforts, then we test the capability of automated Light Detection and Ranging (lidar) methods in detecting CWD in those size classes. We found that martens used larger pieces of CWD than what is randomly available on the landscape (p < 0.05). Our newly developed methods for CWD detection using lidar yielded a 47% detection rate of CWD in marten functional size classes. Lidar’s ability to detect CWD was negatively affected by percent canopy cover (β = 0.96, p < 0.001), and positively affected by size of CWD (length – β = 1.08, p < 0.001; diameter – β = 1.08, p < 0.01).
In chapter 3, we used third-order resource selection functions (RSFs) to investigate marten foraging habitat selection using lidar-derived habitat covariates. Our candidate models represented different ecological hypotheses and included canopy cover, horizontal cover, distance to edge, and percent ground covered by CWD. Our top model was the prey capture model, which indicated that martens selected foraging sites with greater canopy cover, taller trees, greater horizontal cover, and closer proximity to forest edges relative to availability. CWD exhibited a nonlinear relationship, with the strongest selection for areas of intermediate density (~15%). A significant interaction between CWD and canopy cover indicated that selection for CWD was strongest under closed canopies.
The top model had modest discrimination between used and available sites (AUC = 0.62), likely due to broadly distributed suitable habitat. Nonetheless, our findings underscore the value of lidar data in modeling fine-scale habitat selection for species like the American marten.
ScholarWorks Citation
Tisdale, James Michael, "Defining and Characterizing American Marten Habitat Using Airborne Light Detection and Ranging (lidar) in Northeastern Minnesota" (2025). Masters Theses. 1160.
https://scholarworks.gvsu.edu/theses/1160
