A Plethodontid Perspective: Responding to Disturbance —From Hourly Weather to Historical Settlement and Modern Fire
Andrew John Wilk, MS
Advisor: William E. Peterman
Understanding the drivers of “where” and “how many” is fundamental to wildlife ecology and conservation. For decades, species distribution modelers have drawn predictions of where species will occur from presence-only or presence/assumed-absence data and climate norms, topography, and vegetation while ignoring the rich history that land has to offer. Despite the recognized importance of incorporating detection processes into wildlife analyses and how land use history can affect present day populations, they largely are ignored for simplicity or a lack of data. This thesis begins by evaluating weather downscaling methods int the recent microclimc package to understand if the outputs are accurate and then passes those that are to distribution models for five species of plethodontid salamander which incorporates imperfect detection and a variety of historical disturbance. The resulting fit, predictions, and validation are compared between models that incorporate detection, disturbance, or both in addition to topographic features. Finally, it ends by using the robust toolset available through abundance modeling to examine “how many” individuals are present across a range of recent wildfire severity.