| Deer numbers have been in decline across the West since the 1960s. Understanding the causes and assessing the success of management actions designed to counter them require robust estimates of abundance, which are difficult to obtain in forested regions, where aerial counts are not feasible. The use of non-invasive fecal DNA capture-recapture (CR) methods has been proposed to address this problem. Because landscapes are heterogeneous, scaling up estimates derived from localized sampling units to broader regions can pose challenges. We present an approach that uses habitat-based modeling from CR data collected from a black-tailed deer herd on their summer migratory range in the El Dorado National Forest. We selected four rectangular plots with habitat composition representative of the study area and randomly established 24 transects, 6 within each plot, for pellet collection. This design increased our ability to extrapolate density estimates to a larger study area of interest. Specifically, using the transect as the sample unit, we built and tested models relating CR-based estimates of abundance to habitat variables. Models resulted in continuous density surfaces extrapolated over the broader study area, providing region-wide estimates. Pitfalls and potential of this and other approaches to habitat modeling of density will be discussed.
|