Montana Wildlife Project
The Next Frontier: D&E Tools to estimate density via cameras to examine population impacts of CWD/Effects of Management
August 2021 - June 2023
We will jointly develop a novel Bayesian model-based estimator using a time-to-event framework based on the inter-arrival times of individuals in remote camera images (Kalbfleisch and Prentice, 1980). We will evaluate this model and compare it to the design-based models (Moeller et al. 2018) using computer simulations, and determine what method is superior. Next, we will examine how camera deployment affects the performance of the estimators (e.g., placement and spacing of cameras, time-lapse vs. motion-activated, number of cameras needed, etc.) also using computer simulation studies. From these results, we will develop guidelines for designing remote camera grids for estimating deer density. We will use machine-learning techniques (Tabak et al. 2020) to classify deer from camera imagery collected from the Midwest CWD-affected regions. Using this classified imagery, we will apply the appropriate statistical models, based on our simulation studies, to estimate deer density. We will work with managers to implement these statistical tools and survey design recommendations to estimate deer densities to assess risk of CWD growth and spread in new regions, and to measure the effectiveness of CWD control strategies based on reducing the density of deer (i.e., the most common recommended management response to CWD – AFWA BMPs [Gillin and Mawdsley 2018]).