Sightability Adjustment Methods for Aerial Surveys of Wildlife Populations

TLDR

Aerial surveys are routinely used to estimate wildlife abundance, but sightability bias from variable detection of animal groups remains a major source of error, prompting the development of models to predict sighting probabilities under different conditions. The study demonstrates how to incorporate sightability models into population size estimators by treating sighting probability as a response probability and derives formulas for cases where that probability must be estimated. An illustrative example applies the method to helicopter survey data of moose in Alberta, showing how the formulas are implemented in practice.

Abstract

Aerial surveys are routinely conducted to estimate the abundance of wildlife species and the rate of population change. However, sightability of animal groups is acknowledged as a significant source of bias in these estimates. Recent research has focused on the development of sightability models to predict the probability of sighting groups under various conditions. Given such models, we show how sightability can be incorporated into the estimator of population size as a probability of response using standard results from sample surveys. We develop formulas for the cases where the sighting probability must be estimated. An example, using data from a helicopter survey of moose in Alberta (Jacobson, Alberta Oil Sands Research Project Report, 1976), is given to illustrate the technique.