using camera traps and N-mixture models to estimate population abundance: model selection really matters

Estimating the abundance or density of wildlife populations is a critical part of species conservation and management, but estimates can vary greatly in precision and accuracy according to the data collection and statistical methods, sampling and ecological variation, and sample size. N-mixture models are a common method which has been applied to a wide range of taxa for estimating population abundance from non-invasive data representing the distribution of the species. We used population estimates from an aerial survey of moose and videos from camera traps to assess the sensitivity of N-mixture models to ecological conditions, the spatial scale at which they were measured, the criteria used to define independent detections, and model choice based on the common statistical criterion of parsimony. The most parsimonious N-mixture models were considerably biased, producing implausibly large and considerably imprecise estimates of the abundance of moose. Most of the other models produced es..., , , # Data from: using camera traps and N-mixture models to estimate population abundance: model selection really matters

---

## Description of the data and file structure

**Sheet 1: aerial survey data for three strata**\
total cells: number of cells in stratum\
sampled cells: number of cells in stratum that were surveyed\
total area: area (km2) of all cells in stratum\
sampled area: area (km2) of surveyed cells in stratum\
cell ID: unique identifier for each cell\
count: number of bulls, cows, and calves detected in each cell\
area: area (km2) of each cell

**Sheets 2-4: camera trap data**\
camera: unique identifier for each camera\
All other columns are the number of moose detections on each camera for each day.\
NAs represent days in which the camera was not active (dead battery or full memory card)