The authors used decision and modeling analyses to evaluate management alternatives for a decision on whether to permit Cervus canadensis (elk) feeding on two sites on Bridger-Teton National Forest, Dell Creek and Forest Park. Supplemental feeding of elk could increase the transmission of chronic wasting disease (CWD) locally and disease spread regionally, potentially impacting elk populations over time with wider implications for Odocoileus hemionus (mule deer) and Odocoileus virginianus (white-tailed deer) populations and hunting, tourism, and regional revenue. Supplemental feeding is thought to improve overwinter elk survival and reduce the commingling of elk with cattle during months when brucellosis transmission risk is highest. We worked with the U.S. Department of Agriculture Forest Service to identify their fundamental objectives and associated performance metrics related to this feedground decision. We then developed disease and habitat selection models to quantify the effect of four management alternatives on select performance metrics. The four alternatives were to continue to permit feeding, phaseout permits to feed in three years, permit feeding on an emergency basis, or stop permitting feeding. In this report, we present methods and summarized results on disease and habitat selection models and summaries of other performance metrics analyzed by BIO-WEST, Inc. and Cirrus Ecological Solutions as part of an Environmental Impact Statement.

Data from Wyoming Game and Fish Department (WGFD) supported the assumption that supplemental elk feeding allows for larger elk populations in a region. We documented herd units (HU) with feedgrounds having 2.2 times higher densities of elk per area of winter range when compared against HUs without feedgrounds, after accounting for differences in sightability of elk counts on and off feedgrounds. Thus, throughout our analyses, we assumed feedground closures would reduce elk carrying capacity and previously fed elk population segments would decline by 50 percent because of feedground closures. We used a panel of experts to help estimate CWD transmission in fed and unfed elk population segments. In aggregate, the expert panel estimated that median values of direct and indirect transmission of CWD are expected to be 1.9 and 4 times higher, respectively, in fed elk populations compared to unfed elk. We used these disease transmission estimates in combination with local elk demographic rates and carrying capacity estimates to project disease and population dynamics.

In year 20, we predicted CWD prevalence would increase to 40 percent (5th and 95th percentiles = [32 percent, 51 percent]), and 14 percent (5th and 95th percentiles = [3.9 percent, 29 percent] on average for fed and unfed elk population segments, respectively, given a starting prevalence of 1.6 percent. The prevalence estimates for the unfed elk population segments are in the range of previous observations of CWD in elk in the western United States. The average CWD prevalence from 2016 to 2018 in the unfed elk population of Wind Cave National Park in South Dakota was 18 percent overall but up to 30 percent in some regions (Sargeant and others, 2021). Meanwhile, CWD prevalence in the Iron Mountain and Laramie Peak elk herds in Wyoming from 2016 to 2018 was 14 percent and 7 percent, respectively, despite being present since at least 2002 (Wyoming Game and Fish Department, 2020b).

Elk that are fed at Forest Park and Dell Creek constitute 14–19 percent of the total elk on their respective HUs. As a result, the differences between management alternatives are modest when considering the closure of only one feedground on an HU. The no feeding alternative for Forest Park resulted in a CWD prevalence of 16 percent in the Afton HU compared to 19 percent with continued feeding by year 20. In the Upper Green River HU, no feeding on Dell Creek resulted in a CWD prevalence of 30 percent compared to 34 percent with continued feeding. In terms of disease-associated mortality, we predicted the closure of Forest Park and Dell Creek feedgrounds would reduce the total number of CWD mortalities by 9 percent in the Afton HU and 20 percent in the Upper Green River HU during the 20-year timespan.

Our spatial analyses predicted that management alternative effects vary by HU as a function of private property and other wildlife winter ranges proximity relative to feedground location. The predicted number of elk abortions on private land, as a proxy for brucellosis risk to cattle, may increase by 8–10 percent in the absence of feeding at Dell Creek and Forest Park. On the Afton HU, the no feeding alternative resulted in the lowest use of mule deer and Alces alces (moose) winter range by elk. In the Upper Green River HU, the no feeding alternative resulted in less use of moose winter range compared to continued feeding, but higher elk use of mule deer winter range.

Eight feedgrounds are located on Bridger-Teton National Forest, all of which have permits that have expired or will expire prior to 2028. In addition, WGFD could change their management of feedgrounds in light of new information; therefore, we also assessed the cumulative effects of continued feeding, phaseout, and no feeding management alternatives across five HUs south of Jackson, Wyoming (Afton HU, Fall Creek HU, Piney HU, Pinedale HU, and Upper Green River HU). These five HUs ranged from about 60 to 95 percent of the elk herd using feedgrounds, which corresponded to a CWD prevalence at year 20 of 21–38 percent if all feedgrounds in those five HUs remained open relative to 13 percent if all feedgrounds were closed. For all management alternatives, we predicted a 30–58 percent decline in elk populations at the HU level over 20 years because of increasing CWD prevalence or from reductions in populations because of feedground closures. We predicted feedground closures may result in more immediate reductions in population size relative to alternatives that continue feeding (for example, continued feeding and emergency feeding alternatives); however, over longer periods of time, CWD-associated mortality leads to larger population reductions. The no feeding alternative resulted in higher elk population sizes compared to the continued feeding alternative after 10–15 years of implementation. Delayed action under a phaseout alternative resulted in roughly doubling CWD prevalence relative to no feeding on HUs with a large population of fed elk.

Summarizing our cumulative results across all five of the analyzed HUs, we predicted continued feeding will lead to fewer elk by year 20 (mean =10,400, standard deviation [SD] = 800) compared to no feeding at U.S. Department of Agriculture Forest Service sites (11,300, SD = 600). The closure of all feedgrounds was projected to result in the largest elk populations at year 20 (11,800, SD = 600). Continued feeding at all sites, however, resulted in the largest cumulative harvest of 55,400 (SD = 3,500) compared to 51,000 (SD = 1,600) for no feeding at all current feedground sites on the five HUs. Continued feeding also resulted in the lowest brucellosis costs to producers ($199,000, SD = $7,400) compared to phasing out all feedgrounds in three years ($218,000, SD = $7,500). Assuming moderate reductions in hunter interest because of increasing CWD prevalence in elk, we predicted that no feeding resulted in a regional revenue of $162 million (SD = $4.0 million) compared to $176 million (SD = $8.0 million) for continued feeding over the 20-year timeframe.

Recent CWD detections in mule deer and elk in Grand Teton National Park has elevated the importance of the current decision on whether, and how, to permit elk feeding on Forest Park and Dell Creek and the management of the other feedgrounds. Aggressive male harvest has slowed, but not stopped, the increasing prevalence of CWD in mule deer (Conner and others, 2021). It is unclear whether harvest management can be an effective tool to slow the spread of CWD in elk. There are also no effective treatments or vaccines for CWD, and it is unlikely that any will be developed that can be easily deployed in the near future. Thus, reducing artificial aggregations is one of the few management approaches suggested by the Western Association of Fish and Wildlife Agencies (Almberg and others, 2017).

Future surveillance and monitoring can be designed to resolve uncertainties that can improve future decision-making. If feedgrounds close, research could quantify elk population reductions in the absence of feeding, the redistribution of fed elk to other places, or the consequences of elk movement on private property. If feedgrounds remain open, research could assess how rapidly CWD spreads in artificial aggregations of elk; however, surveillance programs would need to be designed with sufficient power to detect initial changes of CWD prevalence. Delaying action on feedground management was projected to be costly. Results of the phaseout alternative relative to the no feeding alternative suggested a three-year delay was enough for substantial long-term changes in CWD prevalence. The long-term persistence of infectious CWD prions in the environment suggests that feedground management decisions may have long-lasting consequences.

Our results indicated tradeoffs in the ability of a management agency to achieve all their objectives, and all management alternatives resulted in significant reductions in elk population size. This report contains the foundational elements for formal decision analysis methods, which can be implemented to help decision makers transparently evaluate the consequences of decision alternatives and identify the set of actions that best achieve agency and stakeholder priorities.