Spatially explicit abundance estimation of a rare habitat specialist: implications forSECRstudy design

Abstract

Abstract. Estimating abundance is an essential component of monitoring and recovery of rare species, and spatially explicit capture-recapture (SECR) models provide the means for robust density estimation. Previous work has elucidated principles of SECR study design for large, generalist carnivores, but less attention has been paid to study design considerations for smaller species, with less extensive home ranges. Here, we integrated data from an intensive pilot study with simulation modeling to evaluate the influence of survey sampling intensity on precision and accuracy in SECR abundance estimation for a rare lagomorph that specializes on patchily distributed early-successional habitats. Doing so, we obtained the first mark-recapture density estimates for the New England cottontail (Sylvilagus transitionalis). Capture probability and density on the landscape both impacted the required intensity of the sampling design. The optimal study design for robust estimation also required a greater number of traps relative to home range size or spatial extent than those recommended in prior SECR studies. This divergence emphasizes that SECR study design considerations will differ among organisms with varying spatial extent and habitat use. Demonstrating the appropriate sampling design for a study system is important prior to embarking in a SECR study. Integrating pilot empirical data with simulations provides a powerful means for optimizing SECR study design and for facilitating applicability of SECR approaches to a wider array of organisms with varying habitat and space use. This methodology may be employed in planning a monitoring program that maximizes effectiveness while minimizing cost and effort, as part of the adaptive management approach to monitor and recover rare or endangered species.

Fast evaluation of study designs for spatially explicit capture–recapture

“…The relative efficiency of detector configurations for SECR has been assessed by simulation for black bear Ursus americanus (Clark, ; Sollmann, Gardner, & Belant, ; Sun, Fuller, & Royle, ) and a few other species (Kristensen & Kovach, ; Tobler & Powell, ). Datasets are generated stochastically from the parameterized probability models for spatial distribution and the detection process, and estimates are obtained from each dataset by the chosen analysis method.…”

Monitoring a New England Cottontail Reintroduction with Noninvasive Genetic Sampling

“…This may explain some of the missed detections at Bellamy River WMA; however, surveys conducted at Wells Reserve did not follow heavy snow events, indicating that variation in individual cottontail movement ranges (smaller winter movement ranges for some females compared with males; M. L. Bauer, personal observation) in combination with factors such as deviation from 30‐m transect spacing by surveyors may account for decreased detection. Conducting 2 independent surveys adhering to 30‐m spacing between search transects (Kristensen and Kovach 2018) and avoiding surveying after heavy snowfall events could increase detection and improve the ability to track founder survival and reproduction in cottontail reintroductions.…”

“…Given the recurring challenges confronted across lagomorph reintroductions, more research is needed on the optimal number of New England cottontail individuals and releases necessary to combat postrelease mortality and establish a self‐sustaining population, and decisions should take into consideration both mortality rates of released individuals and the density of cottontails in the reintroduction landscape. Presuming an average winter density of 2 cottontails/ha (Barbour and Litvaitis 1993, Kristensen and Kovach 2018), it is unlikely that the reintroduced population in the 20 ha of habitat in the Bellamy River WMA had reached carrying capacity because the winter population never rose above 12 individuals. Cottontail densities are variable in the wild (Kristensen and Kovach 2018), and the expected density of 2 cottontails/ha may not be typical of all habitat types or possibly not attainable for relatively small, isolated patches.…”

“…Presuming an average winter density of 2 cottontails/ha (Barbour and Litvaitis 1993, Kristensen and Kovach 2018), it is unlikely that the reintroduced population in the 20 ha of habitat in the Bellamy River WMA had reached carrying capacity because the winter population never rose above 12 individuals. Cottontail densities are variable in the wild (Kristensen and Kovach 2018), and the expected density of 2 cottontails/ha may not be typical of all habitat types or possibly not attainable for relatively small, isolated patches. Further research on cottontail densities in relation to habitat and patch characteristics is needed to establish realistic target densities for reintroductions.…”

“…To monitor the reintroduced population, we collected cottontail fecal pellet samples during winter surveys conducted on 20 ha of available habitat surrounding the release site (the 10‐ha release patch and the additional 10‐ha managed shrubland patch). We collected pellets in a systematic, fine‐scale sampling scheme, in which the patch was searched with transects spaced 30 m apart and pellets were collected every 30 m along a transect when present, following methods of Kristensen and Kovach (2018). We recorded spatial information for search transects and pellets (Garmin GPSMAP 64 s, Olathe, KS, USA).…”