Evaluating Snake Density Using Passive Integrated Transponder (PIT) Telemetry and Spatial Capture–Recapture Analyses for Linear Habitats
Many snake species are elusive and difficult to study in field settings. As such, little is known about their population ecology despite conservation needs for many species. Advances in field techniques and statistical methods can improve our understanding of snake ecology. We used passive integrated transponder (PIT) telemetry to track Nerodia sipedon (Northern Watersnakes, n = 94) and Regina septemvittata (Queensnakes, n = 119) in six low-order streams in central Kentucky, USA from June to October 2016. We assessed snake density, spatial scale of detection, and detection probability using PIT tag relocations and spatial capture–recapture methods for linear habitats. Specifically, we modeled population density as a function of individual stream and land cover type, spatial scale of detection as a function of sex, and detection probability as a function of sex and time-varying covariates. Individual streams were a better predictor of snake density than land cover type; density estimates ranged from 6 ± 3 N. sipedon/km (mean ± standard error) to 107 ± 17 N. sipedon/km and 6 ± 5 R. septemvittata/km to 63 ± 10 R. septemvittata/km. Female R. septemvittata had a larger spatial scale of detection (55 ± 4 m) than male R. septemvittata snakes (40 ± 4 m). Spatial scale of detection did not differ between sexes for N. sipedon (females: 40 ± 4 m; males: 35 ± 3 m). The combination of PIT telemetry and spatial capture–recapture analyses can effectively estimate population densities and other population parameters for snakes and other reptiles and amphibians associated with linear habitats.Abstract
Location of six streams surveyed for Nerodia sipedon and Regina septemvittata in central Kentucky, USA.
Location of snake detections for Nerodia sipedon and Regina septemvittata along a stream (A), creation of “traps” placed approximately every 15 m in ArcGIS (B), and snapping detections to the trap locations in ArcGIS (C). Data are shown for one stream in Kentucky, USA.
Density estimates for Nerodia sipedon and Regina septemvittata in six central Kentucky streams. Results were based on trap spacing of 15 m between traps.
Detection probabilities of Nerodia sipedon and Regina septemvittata on the basis of days since last rain, rain accumulation, day of year, and water temperature in central Kentucky between 6 June and 12 October 2016. Solid black lines indicate mean and 95% confidence intervals for female snakes; dashed gray lines indicate mean and 95% confidence intervals for male snakes. Differences among sexes were not significant for N. sipedon. Rain accumulation and water temperature were not included in the best model for N. sipedon and are not shown. Results were based on trap spacing of 15 m between traps.
Detection probabilities of Nerodia sipedon and Regina septemvittata based on days since last rain, rain accumulation, day of year, and water temperature in central Kentucky between 6 June and 12 October 2016. Solid black lines indicate mean and 95% confidence intervals for female snakes; dashed gray lines indicate mean and 95% confidence intervals for male snakes. Differences among sexes were not significant for N. sipedon. Rain accumulation and water temperature were not included in the best model for N. sipedon and are not shown. Results were based on trap spacing of 30 m between traps. Figure A2 corresponds to Fig. 4, which used a trap spacing of 15 m between traps.
Contributor Notes
