Translator Disclaimer
1 September 2012 Impacts of rural development on Yellowstone wildlife: linking grizzly bear Ursus arctos demographics with projected residential growth
Author Affiliations +
Abstract

Exurban development is consuming wildlife habitat within the Greater Yellowstone Ecosystem with potential consequences to the long-term conservation of grizzly bears Ursus arctos. We assessed the impacts of alternative future land-use scenarios by linking an existing regression-based simulation model predicting rural development with a spatially explicit model that predicted bear survival. Using demographic criteria that predict population trajectory, we portioned habitats into either source or sink, and projected the loss of source habitat associated with four different build out (new home construction) scenarios through 2020. Under boom growth, we predicted that 12 km2 of source habitat were converted to sink habitat within the Grizzly Bear Recovery Zone (RZ), 189 km2 were converted within the current distribution of grizzly bears outside of the RZ, and 289 km2 were converted in the area outside the RZ identified as suitable grizzly bear habitat. Our findings showed that extremely low densities of residential development created sink habitats. We suggest that tools, such as those outlined in this article, in addition to zoning and subdivision regulation may prove more practical, and the most effective means of retaining large areas of undeveloped land and conserving grizzly bear source habitat will likely require a landscape-scale approach. We recommend a focus on land conservation efforts that retain open space (easements, purchases and trades) coupled with the implementation of ‘bear community programmes’ on an ecosystem wide basis in an effort to minimize human-bear conflicts, minimize management-related bear mortalities associated with preventable conflicts and to safeguard human communities. Our approach has application to other species and areas, and it has illustrated how spatially explicit demographic models can be combined with models predicting land-use change to help focus conservation priorities.

Grizzly bears Ursus arctos are considered wilderness species requiring large undisturbed areas (Craighead et al. 1995). Historically, grizzly bears in North America ranged from Alaska to Mexico and California to the Dakotas, occupying numerous ecosystems. However, most grizzly bear populations currently occur in close proximity to humans and are considered conservation-reliant (Scott et al. 2005). Maintaining viability of these populations is a challenge for wildlife managers. In the continental United States, grizzly bears are listed as threatened under the Endangered Species Act (U.S. Fish and Wildlife Service 1993). The Yellowstone grizzly bear was delisted in April 2007 (U.S. Fish and Wildlife Service 2007b), but relisted by court order in November 2009, a decision currently under appeal. In British Columbia, Canada, there are concerns about long-term consequences of human changes to landscapes and the continued health of grizzly bear populations (Herrero 2005). The Alberta grizzly bear was formally listed by the province as threatened in June 2010 (Festa-Bianchet 2010). Few places exist where human land-use development has not adversely impacted grizzly bear habitats.

Long-term conservation of grizzly bears is directly related to human activity. This proximity between bears and humans has resulted in a source-sink dynamic (Knight et al. 1988, Schwartz et al. 2006e, Schwartz et al. 2010) in the Greater Yellowstone Ecosystem (GYE) where bears die at higher rates in and adjacent to areas with human activities. Schwartz et al. (2010) demonstrated that grizzly bear survival was negatively associated with increases in roads, human residences, other developed sites (e.g. campgrounds and lodges) and the time bears used areas open to ungulate hunting.

Rapidly accelerating growth of rural residential development (i.e. exurban sprawl) in some areas in Montana, Idaho and Wyoming has been identified as a factor impacting bear habitat (Schwartz et al. 2010) with the potential for an increase in grizzly bear-human conflicts and bear mortalities.

Human population growth in the Mountain West has exceeded growth in the rest of the nation. During 1970-1999, the GYE experienced a 58% increase in population size and a 350% increase in the area of rural land development (Gude et al. 2006). Land development exceeded population growth due to low-density (1 home/0.4-16.2 ha) exurban development (Gude et al. 2006). Gude et al. (2007) estimated that in 1980, about 3.1% of occupied grizzly bear habitat (Schwartz et al. 2002) had been impacted by exurban development, but projected that by 2020, 6.9% would likely be impacted under aggressive growth management and 10.7% under the boom growth scenario. In their biodiversity assessment of alternative future scenarios, Gude et al. (2007) did not estimate resulting changes in survival or reproduction of specific wildlife populations. Although other studies have done this (White et al. 1997, Schumaker et al. 2004), they felt that this step should be undertaken when sufficient data allowed for meaningful predictions. As a consequence, although the approach used by Gude et al. (2007) provided insight into the potential consequences of exurban development on grizzly bears, it did not quantify impacts to grizzly bear demographics. Additionally, bear numbers and bear distribution have continued to increase in the GYE, necessitating the need for a more rigorous analysis.

Here we build on Gude et al. (2007) and demonstrate how the distribution and extent of grizzly bear source and sink habitats may change under forecasted residential development scenarios. We define source habitats as those areas in the landscape within occupied grizzly bear range where predicted adult female survival was ≥ 0.91. We could have used a different rate, but chose 0.91 because Harris et al. (2006) demonstrated that with current GYE rates of reproduction (0.318 female cubs/female/year; Schwartz et al. 2006a) and survival of dependent young (cubs = 0.63 and yearlings = 0.817; Schwartz et al. 2006d), lambda (λ) ≥ 1.0 in 95% of stochastic simulations when adult female survival was 0.91. Schwartz et al. (2010) used this break point in female survival to illustrate the spatial extent of source and sink habitats in the GYE. In this article, we build on those projections and illustrate the spatial extent of increased sink habitats in the GYE associated with four projected build out scenarios developed by Gude et al. (2007). Our approach has application to other species and areas and illustrated how spatially exp