A sampling strategy for habitat selection, mapping, and abundance estimation of deer by pellet counts

We propose a design-based strategy to exploit pellet group counts performed within plots of a prefixed size using the clearance count technique with the purposes of analyzing habitat selection, mapping the pellet group presence throughout the study area, and estimating the abundance of deer populations. As is customary in design-based inference, the strategy is free of model assumptions, and the precision and statistical consistency of the proposed estimators are determined by the probabilistic sampling scheme adopted to locate plots. The unique necessary assumptions are the absence of migratory movements during the survey period, the accurate recording of the number of pellet groups deposited within sample plots between the 2 visits, and a precise approximation of the daily defecation rate of the population. In addition to these assumptions, which can be attained by a suitable design of the survey, the statistical soundness of the strategy rests on the use of tessellation stratified sampling, a stratified sampling scheme that ensures an even distribution of plots throughout the study area. The scheme also allows for the estimation of the standard errors and the construction of confidence intervals without involving any other assumptions. We applied this strategy in summer 2019 in a protected area of a Mediterranean coastal region to estimate the density of a fallow deer (Dama dama) population. We estimated the corresponding standard error considering the uncertainty entailed by the estimation of the daily defecation rate, with the purpose of performing reliable monitoring. The proposed strategy provided precise estimates of deer abundance and is readily implementable in the field, standardized, and easily repeatable over time, thus allowing reliable monitoring and comparisons across time and space, which are fundamental attributes for management of deer populations.