Objective
The Department of Defense (DoD) expends considerable resources managing and conserving threatened, endangered, or at-risk snake species. Management for these species is often hampered by a lack of basic knowledge regarding their population size and trajectory. The low detectability of most snakes makes it difficult to determine their presence, or to employ traditional methods to estimate abundance. This project demonstrated a novel, simulation-based method, Innovative Density Estimation Approach for Secretive Snakes (IDEASS), for estimating snake density based on systematic road surveys, behavioral observations of snake movement, and spatial movement (radio telemetry) data.
The objective of this project was to demonstrate how IDEASS can integrate spatial movement, road crossing behavior, and systematic road survey data to estimate densities of secretive snakes on DoD installations. Specifically, the project team studied two cryptic species of conservation concern, the Eastern Diamondback Rattlesnake (EDR) and Southern Hognose (SHS), at Fort Stewart, Georgia and aimed to directly compare the effectiveness of IDEASS to traditional capturemark-recapture (CMR) techniques based on visual encounter surveys (VES). A second objective was to demonstrate how the IDEASS approach can make use of existing data sources to generate density estimates cost-effectively and retroactively, using the case study of Western Ratsnakes at Fort Hood, Texas.
Technology Description
The IDEASS approach combines behavioral observations of road crossing behavior (crossing speed), effort-corrected road survey data, and simulation-based modeling of spatial movement to estimate population densities. Radiotelemetric data are collected to quantify movement metrics (frequency, distance, and direction of movement in relation to home range center and roads). These movement data are then used to parameterize individual-based movement models in a biased correlated random walk framework (Turchin 1998, Crone and Schultz 2008) to estimate the frequency with which individuals cross roads. Next, information on survey vehicle speed and snake crossing speed are used to determine the probability of detecting a snake, given that it crosses the road transect during a survey. Snake encounter frequencies during systematic road surveys are then interpreted in light of detection probabilities and simulation model results to estimate snake densities and to assess various factors likely to affect encounter rates. By combining all of these pieces in an IDEASS model, one can calculate individual detection probabilities and population density for even very secretive, rarely encountered species that are not tractable to CMR approaches. Because road collecting is the primary method of capture for many rare and secretive snake species (Willson 2016), the adaptability and applicability of IDEASS makes it a robust and adaptable approach for snake conservation and management.
The field data collection at Fort Stewart consisted of four components: (1) extensive systematic diurnal road surveys of paved and unpaved road transects bisecting habitat suitable for the target species, (2) radio telemetry of EDR, (3) behavioral observations of natural road crossing events to quantify snake crossing speeds, and (4) extensive systematic VES surveys to attempt validation of the density estimates through traditional CMR. Field data was supplemented with radio telemetry data for EDR from similar habitats in South Carolina and with radio telemetry and road crossing speed data for SHS from North and South Carolina.
Demonstration Results
In all three cases, traditional density estimation via visual surveys and CMR failed completely due to lack of captures and recaptures, despite extensive field effort. A total of 598 person-hours across 230 VES yielded only a single EDR capture and zero detections of SHS. With no recaptures, CMR density estimation was not possible. Although the project team could not directly validate the density estimates generated with IDEASS against those from CMR, the estimates are not far below the only two comparable (inland locality) density estimates that are available for this species.
Implementation Issues
The project team concludes that IDEASS represents a powerful tool, and in some cases the only viable method, for estimating density of secretive snakes.
The biggest challenge during this demonstration was site access. Despite the fact that the project team selected study sites that were away from high-use areas of the base, access restrictions and training activities often prevented them from accessing radio-tagged snakes and surveying VES plots and pre-established road transects on a systematic or randomized basis. This problem was exacerbated by the fact that the road routes often bordered or transected several base compartments and on any given day it was common for some of those compartments to be open and others to have access restrictions. For such an approach to succeed, regular site access was absolutely critical. This was especially true for radio telemetry because fine-scale movement data were needed for IDEASS modeling.
The project team also chose to demonstrate this technique on a species that was far less common than anticipated. On southeastern maritime forests and barrier islands, EDR can occur at high densities and be relatively visible via VES (Means 2017). Because of the proximity of Fort Stewart to the coast and the high quality of its longleaf pine ecosystems, they expected that EDR would be relatively tractable to both road encounters and VES. This assumption was wrong. EDR on Fort Stewart occur at low densities, frequently exploit difficult to search habitats, such as palmetto thickets, and spend considerable time in Gopher Tortoise and Armadillo burrows, where they cannot be seen. Thus, the project team generated far fewer encounters than anticipated. They were still able to generate sufficient data to generate independent density estimates for two of the road transects with IDEASS, but they were unable to validate the estimates through direct comparison with those generated using CMR. Thus, a direct comparison of these approaches would still be highly beneficial but would need to focus on a situation where the focal species is more common or conspicuous.
The project team also generated less radio telemetry data that anticipated. This was a direct consequence of the limited EDR encounters. Robust radio telemetry data are integral to accurately parameterizing and validating movement models. The combination of fewer study animals and limited access to track them caused problems. Fortunately, the project team was able to elicit data from colleagues studying the same species at nearby sites with similar habitats. These data were very important in the ability to demonstrate the IDEASS approach.