Date of Award

Fall 2025

Project Type

Thesis

Program or Major

Natural Resources

Degree Name

Master of Science

First Advisor

Remington J. Moll

Second Advisor

Michael W. Palace

Third Advisor

Henry Jones

Abstract

Accurate estimates of wildlife densities are important for understanding population dynamics and informing management decisions. Recent technological developments such as improvements of unoccupied aerial systems (UASs) and sensors provide novel pathways to estimate wildlife densities. However, these methods require critical evaluation of efficacy. In this research, we evaluated the capability of UASs to predict moose (Alces alces) density across a large landscape (~12,000 km2) in northern New Hampshire, USA using longwave infrared imagery via UAS sampling. In Chapter 1, our objective was to account for imperfect detection by creating a sightability model for unmarked moose that includes covariates expected to affect large-scale UAS moose surveys. We conducted 35 UAS flights in northern New Hampshire, USA, during January and February 2023 and completed “sightability maneuvers” over 59 moose detections to collect images at various relative observation angles. From a Bayesian logistic regression based on a naïve observer analysis, we found that greater conifer coverage and sunnier sky conditions strongly reduced sightability of moose while ambient temperature had a weaker but also negative effect. This study provides the first successful quantification of sightability for a non-collared moose population, demonstrating a cost-effective approach for calibrating UAS sampling for additional locations and species while paving the way for future applications of this model to correct moose population sampling. In Chapter 2, our objectives were to 1) create a predictive model to produce spatially explicit moose densities in NH that accounted for imperfect detection (“sightability”), 2) compare resultant density estimates to the existing estimates derived from a hunter survey, and 3) compare model precision to management guidelines. We conducted 62 sampling flights that covered 64.6 km2 during January and February 2024. We fit a Bayesian N-mixture model to 33 adult moose detections, resulting in a density estimate of 0.64 moose/km2 (90% credible interval: 0.56 - 0.77 moose/km2). We incorporated geospatial variables into the model to make spatially explicit predictions of moose densities across the study area, resulting in an overall mean density of 0.67 moose/km2. This density was approximately twice that expected based on the deer hunter survey index. The precision of model estimates was within the benchmark range (90% CI within 25% of mean) typically targeted for effective management decision-making. Our work suggests that UASs offer a promising, albeit field-intensive, method for estimating density without tagged or GPS-collared individuals.

Download

Share

COinS