Tracey D. Tuberville¹, Kimberly M. Andrews¹, James D. Westervelt², Harold E. Balbach², John Macey³, and Larry Carlile³

¹Savannah River Ecology Lab, University of Georgia, Drawer E, Aiken, SC 29802, USA
²Construction Engineering Research Laboratory, US Army Engineer Research and Development Center, Champaign, IL, USA
³Fort Stewart Army Installation, Fish and Wildlife Branch, 1177 Frank Cochran Drive,
Fort Stewart, GA 31314-4940, USA

Background: One of the challenges with conserving rare species is identifying the most effective management targets; that is, the demographic traits most likely to positively influence population persistence through either manipulation of the habitat or the wildlife population. Furthermore, these targets should represent the most efficient use of limited resources, especially given that resource managers need to balance multiple, often complex issues (Reed et al. 2009). Population models can often aid managers in this process, and such models are frequently used to rank relative threats to specific populations, evaluate effects of proposed management actions or regulations, determine which demographic or ecological variables have greatest influence on extinction risk, and identify information gaps and research priorities (Tuberville et al. 2009 and references therein).
Population viability analysis (PVA) models represent a traditional modeling approach that has been used to support management decision-making for both game and nongame species. Unfortunately, robust PVA models require extensive population-level data for accurately estimating demographic parameters. Developing PVAs for rare species can be difficult, therefore, because complete life history information and long-term population trend data often are not available. For many rare species, however, detailed information is known about their natural history and the behavior of individuals, including how they interact with each other and the landscape and how they respond to environmental cues. For these species, individual-based models (IBMs) may be more appropriate than PVA models for performing demographic sensitivity analysis; IBMs have the added advantage of imposing a spatially explicit landscape context.
The gopher tortoise (Gopherus polyphemus) is an example of a species whose life history data are incomplete, but whose natural history and individual behavior are well characterized. This species is considered to be declining throughout its range (Smith et al. 2006). It is federally listed as threatened in the western portion of its range (USFWS 1987) and is currently under consideration for listing throughout the remainder of its range (USFWS 2009) . Gopher tortoise populations occur on many military installations throughout the southeastern USA (Wilson et al. 1997) and the species has been identified for management under the Army’s Species at Risk (SAR) program. The SAR program seeks to develop proactive management strategies to ensure long-term viability of imperiled species that currently reside on military installations (NatureServe 2004).
Fort Stewart is the largest Army installation within the range of the gopher tortoise. Given the current and anticipated increase in training demands at Fort Stewart—in terms of both intensity and spatial extent—
resource managers are challenged to maintain viable populations of rare species within a limited or even diminishing footprint. One of the most practical ways to address this challenge is by improving demographic conditions for “at-risk” species through improvement of their existing habitat. Population models, used with demographic sensitivity analysis in particular, can help to determine the extent that habitat management alone can influence demographic parameters of rare species (such as the gopher tortoise) so that their abundance and likelihood of persistence will increase.

SREL Reprint #3217