Year of Award

2019

Document Type

Thesis

Degree Type

Master of Science (MS)

Degree Name

Wildlife Biology

Department or School/College

W. A. Franke College of Forestry and Conservation

Committee Chair

Dr. Paul M. Lukacs

Commitee Members

Dr. Creagh W. Breuner, Dr. E. Frances Cassirer, Dr. Victoria J. Dreitz

Keywords

Mountain Goat, Abundance, Occupancy, Non-invasive, Double-observer, Closure assumption

Publisher

University of Montana

Subject Categories

Biology | Ecology and Evolutionary Biology | Population Biology | Research Methods in Life Sciences | Statistical Models

Abstract

Abundance and occupancy are two parameters of central interest to the field of ecology. Furthermore, accurate (both precise and unbiased) estimates are key pieces to the puzzle of effective wildlife management decision-making. While there exist a variety of sampling techniques and statistical models for effectively estimating population parameters for frequently encountered and large mammals, methods for sampling unmarked and rare species are few and far between. The first step to acquiring usable parameter estimates is through the use of sampling theory and incorporation of probabilistic sampling designs to collect count-data and occurrence-data. Often, it is assumed that probabilistic sampling designs will be ineffective in surveying for rare species due to insufficient encounters with the species of interest. However, many of these probabilistic sampling methods remain untested, both with respect to modern statistical models and in the context of low-density species. The consequences of not incorporating probability-based sampling designs and not meeting field sampling assumptions are not well understood in the field of ecology and can thus provide uncertainty when making management decisions. In this paper, we test disparate field methods and statistical models that apply a complete random sampling design for estimating unbiased occupancy and abundance of mountain goats (Oreamnos americanus) – a low-density and difficult-to-study species. In doing so, we developed a novel data analysis approach that directly solves the problem of approximating the closure assumption in addition to successfully producing a method and modelling technique that yields unbiased estimates of mountain goat abundance.

Share

COinS
 

© Copyright 2019 Molly McDevitt