Evaluation of the distance sampling technique to monitor gregarious species: A simulation exercise

Abstract

AbstractWildlife count data are used to provide estimates of population abundance. To quantify changes in the size of populations, it is necessary to determine the precision of abundance estimates. Ground or aerial survey methods are often intended to be a census and thus do not provide estimates of precision. One solution is to switch approaches to a different sampling and estimation procedure, such as distance sampling. Using an example of the long‐term waterfowl censuses in the Camargue, a major wintering area in southern France, we used simulation to quantify the precision and cost effectiveness of transitioning to a distance sampling approach. Based on abundance and distribution data for 14 species of waterbirds (Anatidae and coot [Fulica atra]) in the Camargue, we evaluated 5 distance sampling scenarios with parallel line transects corresponding to 5 different sampling intensities. Our results suggested that, if the objective of counts is to estimate the abundance of all waterfowl together, switching from pseudo‐total to distance sampling counts would be realistic on both statistical and financial grounds. The most advantageous scenario would correspond to a sampling effort with transect lines spaced 1,000 m apart, which would allow a coefficient of variation of 0.20 for total waterfowl abundance with a 27% increase in flight costs compared to the current method. Heterogeneity in the size of detected bird groups contributed to the variability in our results, however, adjusting sampling effort from 1,750 m spaced transect to 760 m allowed reduced variation by 55%. We recommend proceeding to test flights which would allow some site‐specific parameters to be refined before changing methods. However, the distance sampling method as simulated here would prove impractical to evaluate the abundance and distribution of one species alone.