I wrote last month about our recent Line Transect Distance Surveys (LTDS) for Gopher Tortoises on the Orianne Indigo Snake Preserve. This survey technique has become the standard for the vast majority of the tortoise monitoring that occurs across the southeast and with good reason. The LTDS framework presents a relatively easy survey methodology that can be carried out reliably by a small group of observers. Furthermore, it accomplishes the general goal of monitoring populations for a species that can be difficult to actually encounter using other survey techniques (i.e., tortoises are underground a majority of the time). With all of this said, are there circumstances where an alternative survey approach may be desired to monitor tortoise populations?
In a recent paper published in the journal Herpetological Conservation and Biology, I, along with co-authors from Virginia Tech, examined the utility of using occupancy surveys to monitor the Gopher Tortoise population on Eglin Air Force Base. Despite abundant sandhill habitat (approximately 155,600 ha), Eglin actually has a small tortoise population, possibly because of historic habitat degradation (i.e., lack of fire) and/or heightened mortality rates for tortoises from human hunters (specialized equipment was historically used to forcibly remove tortoises from their burrows in this region; Taylor 1982). Regardless of the reason for the low density tortoise populations, monitoring them is extremely difficult because of the amount of habitat that would need to be surveyed using a normal LTDS approach.
Therefore, we decided to conduct a pilot study to assess whether or not occupancy surveys could be used to monitor tortoises on Eglin as has been done recently with other chelonian species (Zylstra et al. 2010; Erb et al. 2015). Occupancy surveys have become an increasingly common form of wildlife monitoring in recent years, relying on identifying whether or not some spatial unit is occupied by the species of interest. Over time, the number of occupied sites (the occupancy rate) can be tracked to get a better understanding of how a species is faring at a landscape scale. This is the approach that we use for our Eastern Indigo Snake surveys in southern Georgia. Occupancy surveys are also useful because they can take imperfect detection into account as part of the design and data analysis. Imperfect detection describes the phenomenon that, even if a species is present on a survey site, an observer is not guaranteed to find it.
To implement this pilot study, we divided the potentially suitable habitat on Eglin into a grid of 1-ha survey sites that were then grouped based on the dominant habitat type and the distance from the nearest known tortoise record. We surveyed sites at random from all available sites, while keeping the proportion of sites in each group similar to their overall availability across the landscape. Each site was surveyed with a small group of observers. Importantly, we used tortoise burrows as our metric of occupancy, which eliminated the need to conduct burrow scoping (a time consuming process). We also ended each survey after a single burrow was detected, allowing survey crews to conduct more surveys per day. A subset a sites were surveyed twice, allowing us to calculate the above mentioned detection probability.
Overall, we surveyed 469 sites (743 total surveys) in a 6-month period. Gopher Tortoise burrows were found at just 53 of these sites, and 39 of these sites were located with 60 m of a historic tortoise record. Our occupancy modeling results indicated that the detection probability for tortoise burrows was high (0.95). In other words, observers were likely to identify a burrow 95% of the time if a burrow was present on the survey site. The probability of tortoises occupying suitable habitat was highest at sites closer to historic records, decreasing from 0.42 to 0.01 from as the distance from records increased. These results highlight the clumped distribution of tortoises on Eglin despite the abundant upland habitat. Finally, we used these results to simulate the potential of future surveys to detect changes in the occupancy probability over time. Simulations suggested that this general methodology could be used to document a 3–5% annual decline in tortoise occupancy within 10 years but that a 1% annual decline would be unlikely to be detected within 20 years.
The results of our pilot study indicated that an occupancy monitoring approach could be useful on large landscapes with low tortoise densities like Eglin Air Force Base. We do stress that monitoring programs need clearly defined goals and managers should understand what metrics a particular monitoring program can produce (e.g., occupancy surveys do not generate estimates of population abundance or size). Finally, the clumped distribution of tortoises on Eglin suggests site-specific management actions could benefit the tortoise population. First, habitat management actions should be targeted to core areas where tortoises are known to occur because the likelihood of tortoises occupying other areas was extremely low. Second, translocations of tortoises (both on site and from external sources) could be used to increase the abundance and distribution of tortoises on base. This type of work is already underway (Kobilinsky 2017).
The full paper can be viewed at the following link:
Erb, L.A., L.L. Willey, L.M. Johnson, J.E. Hines, and R.P. Cook. 2015. Detecting long-term population trends for an elusive reptile species. Journal of Wildlife Management 79:1062–1071.
Kobilinsky, D. 2017. A mission to conserve wildlife: The unique challenges on military lands. Wildlife Professional 11:16–22.
Taylor, R.W., Jr. 1982. Human predation on the Gopher Tortoise (Gopherus polyphemus) in north-central Florida. Bulletin of the Florida State Museum, Biological Sciences 28:79–102.
Zylstra, E.R., R.J. Steidl, and D.E. Swann. 2010. Evaluating survey methods for monitoring a rare vertebrate, the Sonoran Desert Tortoise. Journal of Wildlife Management 74:1311–1318.