Alternative D Description

If following alternative D, the FWS, along with the other cooperating agencies—NPS, FS, and WGFD, will work to reduce the number of elk that overwinter on the NER to about 5,000 animals, prior to stopping feeding, by increasing elk harvest opportunities. The goal is to reduce the population over 10 years under the new “Bison and Elk Management Plan,” having most of the herd reduced prior to year 8. Then, after 10 years of active elk herd reductions, the FWS will stop feeding elk on the Refuge. If, within those first 10 years, the prevalence of CWD is measured as 7 percent or higher in the Jackson Elk Herd Unit (JHU), the FWS will stop feeding elk immediately. Seven percent was selected as the threshold because previous work by Galloway and others (2021) indicated that the Jackson elk herd, mostly contained with the JHU, will decline at CWD prevalence equal to or greater than 7 percent. Regardless of the CWD prevalence, the FWS will end its supplemental feeding of elk after year 10. The NER will continue to feed bison over the next 20 years to minimize the potential of human-wildlife conflicts and maintain a minimum population of 500 bison.

General Approach

The evaluation of alternative D followed mostly the same process as the other five initial alternatives, whereby the actions described by the alternative were translated into quantitative relationships within a series of linked computer models. First, the alternative components associated with increasing harvest rates or stopping feeding in response to CWD were considered within the matrix elk population and CWD simulation model described in chapter B (herein CWD model; Cross and others, 2025, this volume). The CWD model allowed for repeat simulations of the elk population (fig. F1) and showed how the population changed over time in response to natural aging, reproduction, survival, and CWD processes, and also as a function of actions related to harvest and feedground management. The results of the simulations were used to summarize performance metrics that focused on elk population size, elk suffering, number of harvested elk, and changes in CWD prevalence (Performance metrics [PMs] 1a–c, 5a; table F4).

To model elk space-use patterns under alternative D, the raw elk counts of the number of elk with CWD were then projected onto a set of spatially explicit probabilistic maps that corresponded to monthly space use patterns for elk under conditions that matched seasonal variations in habitats and feeding operations at State and Federal feedgrounds (hereafter, spatial model; Cotterill and others, 2025, this volume, chap. C). The results from the spatial model produced metrics associated with elk use of sensitive habitats (PMs 2a–c, table F4) and private lands (PM 6a, table F4). A subset of the private lands where cattle overwinter were further used, in combination with counts of adult females, pregnancy rates, and brucellosis seroprevalences to derive a brucellosis abortion hazard for elk as a measure of risk to cattle producers (PM 6c, table F4; Cotterill and others, 2025, this volume, chap. C).

Finally, the outputs from the CWD and spatial models were used in three separate socioeconomics models to estimate metrics associated with either NER visitor counts (PM 5b, table F4) and their spending at local businesses (PM 8b, table F4), hunting revenues from elk tag sales (PM 8a, table F4) and hunting-related contributions to local economies (PM 8b, table F4), or outfitter revenues (PM 8c, table F4). Other metrics associated with FWS management costs and invasive species risk were covered in chapter A (Cook and others, 2025). We assumed that the effects to bison populations (PM 1d, table F4), human-bison conflict (PM 6b, table F4), and bison harvest (PM 9a, table F4) under alternative D would be the same as under the continue feeding alternative because alternative D continues feeding bison over the entire 20 years of evaluation. The models are written in R (version 4.4.0; R Core Team, 2024) and are available online (Cook, 2024; Cotterill and Graves, 2025; Cross, 2025; McEachran and Cook, 2024).

Modeling alternative D differed from the other alternatives in that four distinct scenarios of alternative D were considered. The variables were whether the 7 percent CWD trigger and increased elk harvest were included or not, and in what combination with other management strategies. The first scenario did not include the 7 percent CWD trigger or increased elk harvesting and assumed that elk feeding was ended after year 10. The second scenario increased elk harvesting to achieve the goal of 5,000 elk using the NER by year 8 but did not include the 7 percent CWD trigger. The third scenario included the 7 percent CWD trigger to halt elk feeding, but did not increase elk harvesting. The fourth scenario included both the CWD trigger and increased elk harvesting options. These four scenarios helped to understand the range of outcomes that might happen given the inclusion and success of these different actions.

Results are presented across two spatial scales to facilitate broad-scale and fine-scale comparisons among the alternatives. The broader spatial scale (analysis area, fig. F1) included all areas where the Jackson elk herd currently inhabit, but also those areas where they are expected to relocate under management alternatives that stop feeding. This area included the JHU and the northern part of the Fall Creek Herd Unit. The finer spatial scale was defined as the JHU and is the scale at which most elk management decisions are typically made (fig. F1).

Modeling Details Under Alternative D

CWD and Elk Population Size

The inclusion of the 7 percent CWD trigger resulted in a lower overall CWD prevalence over the 20 years (table F1). At year 10, if there were no increased hunting and no CWD trigger, the CWD prevalence in the JHU averaged 38 percent (standard deviation [SD]=9.7 percent) before declining back to 25 percent (SD=8.6 percent) by year 20 (fig. F2). In contrast, when there were both increased hunting and a CWD trigger, the CWD prevalence by year 10 in the JHU was, on average, 22 percent (SD=7.5 percent) and, by year 20, reached a CWD prevalence of 26 percent (SD=10 percent). Including the CWD trigger also resulted in fewer winter CWD mortalities over the 20 years compared with when a 7 percent CWD trigger was not included in the alternative. At the spatial extent defined by the full analysis area, we found similar trends as in the JHU with lower CWD prevalence under scenarios of alternative D that included the 7 percent CWD trigger (table F1). All scenarios of alternative D had lower CWD prevalence compared to the alternative that continued elk feeding for 20 years (table F1).

The alternative D scenario with the 7 percent CWD trigger resulted in a larger elk population size by year 20 compared with an alternative D with no 7 percent CWD trigger (table F1). Alternative D without the CWD trigger resulted in a Jackson elk herd population size at year 20 of 4,421 (SD=1,278) without an increase in hunting or 4,754 (SD=1,461) with an increase in hunting (table F1, fig. F1). In contrast, the inclusion of a CWD trigger and an increase in hunting resulted in a Jackson elk herd population size of 6,143 (SD=1,791) by year 20. An alternative D with a CWD trigger and an increase in hunting resulted in a higher elk population size by year 20 and a lower CWD prevalence than the continued feeding alternative (table F1). At the spatial extent defined by the full analysis area, we found similar trends as in the JHU with larger elk populations under scenarios that included the CWD trigger (table F1). The scenarios of alternative D that had no CWD trigger had lower elk population sizes compared to the alternative that continued feeding for 20 years, whereas alternative D scenarios with the CWD trigger had larger population sizes at year 20 (table F1).

The alternative D scenarios that focus on increasing initial hunting rates resulted in more cumulative elk harvested over the 20-year timeframe (11,041 [SD=1,970] elk harvested with CWD trigger and increased hunting compared to 9,845 [SD = 2,068] harvested with a CWD trigger but no increase in hunting). The increased hunting rates led to more animals harvested at the full analysis area, having 14,752 (SD=2,260) elk hunted over 20 years with the CWD trigger and increased harvest compared to 13,573 (SD=2,453) with the trigger but without increased harvest (table F1).

Elk Use of Sensitive Habitats and Private Lands

In general, all four scenarios of alternative D resulted in fewer elk-use days on sensitive aspen, willow, and cottonwood habitat types on and adjacent to the Refuge compared to the continue feeding alternative (table F2). The exceptions for lower elk use were for cottonwood and willow under the alternative D scenario that included the 7 percent CWD trigger but not increased hunting, and willow under the alternative D scenario that included both the 7 percent trigger and had increased hunting (table F2). The scenario without the 7 percent CWD trigger, but with increased hunting, tended to perform the best across all metrics associated with elk use of sensitive habitats among the four scenarios of alternative D. Elk use of private lands and brucellosis risk to livestock was lower than the continue feeding alternative under alternative D scenarios without the 7 percent CWD trigger and higher under the two scenarios with the CWD trigger (table F2).

Socioeconomic Effects

The two alternative D scenarios without increased harvest showed similar cumulative revenues from elk harvest tags, hunting-related expenditures, and outfitter revenues compared to the continue feeding alternative over the 20-year period (table F3). These patterns result from the increased number of elk harvested under the alternative scenarios with increased harvest rates (table F1). In contrast, the alternative D scenarios with increased harvest had higher harvest-tag revenues, hunting-related expenditures, and outfitter revenues than the continue feeding alternative over the 20-year period (table F3).

Tradeoffs Among Alternatives

The full suite of nineteen performance metrics can be found in table F4. No single alternative or scenario among continued feeding, no feeding, and the four alternative D scenarios performed the best on all metrics simultaneously. Notably, there is a consistent tradeoff whereby stopping feeding earlier results in lower CWD prevalence and higher elk population sizes after 20 years but produces higher elk use of private lands. The no feeding alternative results in one of the lowest estimates of CWD in elk and the largest elk population in year 20; however, that same alternative resulted in high elk use of private lands and sensitive habitats. Under the no feeding alternative, the bison population was lower, human-bison conflicts increased, and the estimated bison harvest was lower (table F4). In contrast, continuing to feed elk resulted in higher CWD prevalence in elk and lower elk population size in year 20, but less use of private lands, and some sensitive habitats, particularly willow. For bison, continued feeding resulted in more bison, fewer human-bison conflicts, and more estimated harvest (table F4).