Theses and Dissertations

ORCID

https://orcid.org/0000-0002-9653-9814

Advisor

Vance, Carrie K.

Committee Member

Kouba, Andrew J.

Committee Member

Caprio, Michael A.

Committee Member

Wang, Guiming

Committee Member

Santymire, Rachel M.

Date of Degree

5-10-2024

Original embargo terms

Embargo 1 year

Document Type

Dissertation - Open Access

Major

Life Sciences (Animal Physiology)

Degree Name

Doctor of Philosophy (Ph.D)

College

College of Agriculture and Life Sciences

Department

Department of Biochemistry, Molecular Biology, Entomology and Plant Pathology

Abstract

The Anthropocene epoch in which we are currently living, also known as the Holocene, has brought about unprecedented losses in planet Earth’s biodiversity. Numerous extirpations of floral and faunal species have been influenced by human encroachment and more specifically, the exploitation of such species and the respective habitats in which they reside. It is this notion that has propelled many scientists to take up intellectual arms in an effort to protect these invaluable resources. The purpose of this research was to develop technologies to measure and evaluate various variables that influence animal physiology, specifically in amphibians who represent the most threatened class of all animal taxa. Species-specific knowledge including life history and an understanding of evolutionary traits are often needed to effectively guide the management decisions surrounding any given animal population. Specific objectives of this project were to develop non-invasive methods, such as hormone monitoring, machine learning-aided ultrasonography, and near-infrared spectroscopy (NIRS), to assess vital physiological traits, such as biological sex, reproductive status, and chytrid fungus pathogen detection in threatened amphibian species. The novel technologies developed and applied in amphibians here may provide insights for addressing conservation related questions in other animal as well as plant species. Additionally, automation of physiological monitoring techniques through the use of machine learning methods reduces barrier to entry and enables these technologies to be operated by a larger practitioner base. This research also serves to advance methods surrounding chemometric analyses as it pertains to the discipline of wildlife spectroscopy, where large multivariate datasets require data manipulation strategies to produce robust prediction models for the physiological trait of interest for qualitative or quantitative assessment. To that end, a multi-model framework is provided for optimizing predictive outcomes to address questions relating to wildlife management and conservation initiatives.

Download

Available for download on Thursday, May 15, 2025

Share

COinS