Unintended indirect effects limit elk productivity from supplemental feeding in the Greater Yellowstone Ecosystem

Abstract

The widespread practice of supplemental feeding, a bottom‐up forcing of resource availability, is intended to improve wildlife population health and survival. However, supplemental feeding could trigger indirect effects by altering predation rates and disease dynamics. We investigated the effects of feeding on three key elk (Cervus canadensis) population productivity metrics (calf:cow ratios, annual change in elk density, and harvestable surplus) across 13 regions in the Greater Yellowstone Ecosystem (GYE) over 26 years. Incorporating previous population size, climate, predator, and harvest data in a Bayesian regression framework revealed new insights about elk productivity metrics in the GYE. Supplemental feeding was associated with increased calf:cow ratios (4.9%) but was not substantially related to changes in elk density and harvestable surplus, which are both management targets. Notably, the feeding effect on calf:cow ratios appeared to be offset by increased wolf (Canis lupus) and grizzly bear (Ursus arctos horribilis) predation. We hypothesize that increased elk productivity resulting from supplemental feeding is primarily transferred to predator and pathogen trophic levels in this system with limited observed effects on elk abundance and harvestable surplus. Anthropogenic food resources may have unintended indirect consequences on other trophic levels that potentially limit the direct impacts of feeding.

INTRODUCTION

Supplemental wildlife feeding, a widespread practice ranging from backyard feeders to large-scale operations, aims to enhance wildlife populations and provide human enjoyment among other objectives (Murray et al., 2016). However, its ecological consequences are complex and challenging to predict (Griffin & Ciuti, 2023). A key complexity lies in understanding how supplemental feeding affects direct and indirect trophic interactions. Resource manipulations like supplemental feeding can initiate bottom-up forcing, but this intervention is often embedded within a complex ecological network (Sinclair & Krebs, 2002). For instance, whereas increased resource availability might intuitively be expected to increase herbivore populations (Lubow & Smith, 2004), the realized effect may be indirectly modified by top-down regulation of predators and/or enhanced disease transmission (Buck & Ripple, 2017). If predators readily consume the increased herbivore productivity, herbivore biomass may not necessarily increase (Ripple et al., 2016). This decoupling of productivity and abundance, observed as high as the ecosystem level in Marczak et al. (2007), highlights the complex interplay of bottom-up and top-down forces in shaping trophic interactions.

This altered trophic dynamic can have several downstream consequences. Supplemental feeding can lead to localized concentrations of herbivores, even if overall herbivore biomass does not dramatically increase, resulting in overgrazing and negative impacts on plant communities (Rinella et al., 2012). In addition to its direct trophic effects, supplemental feeding can alter predator populations (Oro et al., 2013), potentially intensifying competition with human hunters (Jonzén et al., 2013). This interplay raises two key questions: whether supplemental feeding decreases compensatory natural mortality (Bartmann et al., 1992) and whether hunting mortality is additive (Creel & Rotella, 2010) under these conditions. This trophic manipulation can also disrupt natural density-dependent population regulation, creating feeding dependencies that may negatively impact populations during periods of resource scarcity (Stewart et al., 2005). Supplemental feeding can also influence disease transmission dynamics (Becker et al., 2015; Cotterill et al., 2018) and contribute to more complex food web structures (Lafferty et al., 2008). Thus, supplemental feeding, while intended to benefit wildlife, can have multifaceted and potentially unintended consequences for ecosystem functioning.

Supplemental feeding of cervids is a widespread wildlife management practice employed in Europe and North America (Jones et al., 2014; Smith, 2001), often utilized to mitigate winter resource limitations (Ossi et al., 2015), but motivations can go further with diverse objectives ranging from mitigating agricultural damage to enhancing hunting opportunities (Milner et al., 2014). For example, in Europe, red deer (Cervus elaphus) are commonly provided with supplemental feed to reduce browsing pressure on forests and agricultural lands (Putman & Staines, 2004). The Greater Yellowstone Ecosystem (GYE), a large and relatively intact temperate ecosystem spanning parts of Wyoming, Montana, and Idaho in the western United States (Lynch et al., 2008), provides a compelling case study for examining these complex issues. The GYE is home to diverse ungulate populations, including the iconic elk (Cervus canadensis). Elk play a pivotal role in the GYE's food web (Christianson & Creel, 2014) and nutrient cycling (Peziol et al., 2023; Singer & Schoenecker, 2003), while holding important cultural and economic value through tourism and hunting (Maher et al., 2023).

Elk in the GYE have been supplementally fed in winter in some regions for over a century, providing a long-term view of its effects (Smith, 2001). This practice, primarily concentrated in the herd units (HUs) in the southern portion of the GYE (Figure 1), has been motivated by a variety of factors, including mitigating elk–livestock conflict and brucellosis transmission (Cross et al., 2007), minimizing damage on private lands (Smith, 2001), enhancing winter juvenile survival (Smith & Anderson, 1998), supporting higher elk densities for robust hunting opportunities (Smith, 2001), and compensating for winter range losses (Kahn, 2000). However, provisioned elk in the GYE exhibit altered migration patterns (Jones et al., 2014), increased aggregation and pathogen spread (Cross et al., 2007), heightened stress (Forristal et al., 2012), and reduced pregnancy rates (Cotterill et al., 2018). Furthermore, the long-term consequences of this practice are a subject of ongoing debate, with concerns intensifying amidst emerging threats like climate change (Rickbeil et al., 2019) and chronic wasting disease (CWD). This disease is a prion-based spongiform encephalopathy (Williams, 2005) that has recently been detected in an elk on a GYE feedground for the first time (Wyoming Game and Fish Department, 2025b). Crowding of elk on feedgrounds could enhance transmission (Galloway et al., 2021) and was modeled to result in smaller, more infected elk herds (Cook et al., 2023, 2025).

[IMAGE OMITTED. SEE PDF]

Supplemental elk feeding is an anthropogenic perturbation in the GYE that we can utilize to better understand trophic interactions in an intact ecosystem. We developed eight testable hypotheses regarding the influence of bottom-up (feeding and environmental) and top-down (predation, harvest) forces, as well as population dynamics, on key elk population productivity metrics—calf:cow ratios, elk density, and harvestable surplus—from 1995 to 2020 (Table 1). These hypotheses explore how supplemental feeding might influence elk survival, calf recruitment, and the balance between predator populations and elk productivity, while also assessing the consistency of feeding effects across varying climatic conditions. Specifically, we hypothesized that supplemental feeding would have a positive association with elk survival (H1, Table 1), whereas increased snow depth would either be negatively correlated (winter mortality) or positively associated (lusher summer vegetation) with survival (H2a,b, Table 1). We also predicted that increased summer rainfall may be positively correlated to elk survival (H3, Table 1), and that top-down forces, such as grizzly bear (Ursus arctos horribilis), wolves (Canis lupus), and harvest, may negatively correlate with elk survival (H4–6, Table 1). Furthermore, we hypothesized that increased elk density may be negatively linked with elk survival because of density dependence (H7, Table 1), whereas increased calf:cow ratios would have a positive association with recruitment (H8, Table 1). Finally, we explored how supplemental feeding could modify ecological responses, hypothesizing that feeding may synergize with environmental factors (I1–2, Table 2), subsidize predator effects on productivity in fed HUs (I3–5, Table 2), and lead to decreased productivity at high densities in fed populations (I6, Table 2).

TABLE 1 Hypotheses of elk productivity influenced by supplemental elk feeding and ecological regulation mechanisms in the Greater Yellowstone Ecosystem.

Hypothesis	Factor	Category	Predicted relationships	Rationale and previous work
H1	Feeding	Bottom-up	Supplemental feeding may be associated with increased elk survival.	Supplemental feeding can provide a direct nutritional benefit, potentially increasing survival rates, especially during harsh winters. However, Foley et al. (2015) found no relationship with feeding on calf:cow ratios.
H2a	Snow	Bottom-up	Increased snowpack may be associated with decreased winter elk survival.	Deep snow can hinder foraging, increase energy expenditure, and decrease winter survival (Proffitt et al., 2014). However, Foley et al. (2015) found a negative relationship of snow depth in fed HUs and no relationship in unfed HUs.
H2b	Snow	Bottom-up	Increased vegetation may be associated with increased summer elk survival.	Increased snowpack can lead to increased vegetation growth in the following summer (Potter, 2020) which can improve forage availability and thus elk survival.
H3	Rain	Bottom-up	Increased summer rainfall may be associated with increased vegetation and increased elk survival.	Summer rainfall can promote vegetation growth, increasing forage availability and potentially increasing pregnancy rates (Proffitt et al., 2014) and elk calf survival.
H4	Grizzly bear	Top-down	Increased grizzly bear density may be associated with decreased calf survival.	Grizzly bears are a major predator of elk calves (French & French, 1990). Increased grizzly bear abundance can lead to increased predation pressure and decreased calf survival (Foley et al., 2015; Proffitt et al., 2014).
H5	Wolf	Top-down	Wolf presence may be associated with decreased elk survival.	Wolves are a major predator of elk (Smith et al., 2020), and increased wolf presence can lead to increased predation risk and decreased calf recruitment (Christianson & Creel, 2014; Foley et al., 2015; Proffitt et al., 2014).
H6	Harvest	Top-down	Increased elk harvest may be associated with decreased elk population density.	Hunting had a higher reproductive impact than wolves (Wright et al., 2006).
H7	Count/density	Population dynamics	Increased elk density may be associated with decreased elk survival (density dependence).	Density-dependent competition for resources (e.g., forage, space) can decrease calf survival at high population densities (Lubow & Smith, 2004) and reduce yearling pregnancy and recruitment (Proffitt et al., 2014). However, Foley et al. (2015) found no relationship with abundance on calf:cow ratio in fed HUs but a positive one in unfed HUs.
H8	Calf:cow	Population dynamics	Increased calf:cow ratios may be associated with increased elk density and surplus.	Calf survival greatly influences population growth (Raithel et al., 2007).

TABLE 2 Potential interactions influenced by supplemental elk feeding in the Greater Yellowstone Ecosystem.

Interaction with feeding	Predicted response of calf:cow ratios	Rationale
I1 Snow	Higher calf:cow ratios are expected in fed HUs after harsh winters (high snowpack) compared to unfed HUs.	Supplemental feeding may mitigate the negative effects of harsh winters on calf survival.
I2 Rain	Higher calf:cow ratios are expected in fed HUs during years with high summer rainfall compared to unfed HUs.	High rainfall may enhance the positive effect of supplemental feeding on calf survival.
I3 Grizzly bear	Higher calf:cow ratios are expected in fed HUs despite the density of grizzly bears compared to unfed HUs.	Increased elk biomass may provide alternative prey options for predators, reducing predation pressure on calves.
I4 Wolf	Higher calf:cow ratios are expected in fed HUs despite the presence of wolves compared to unfed HUs.	Increased elk biomass may provide alternative prey options for predators, reducing predation pressure on calves.
I5 Harvest	Higher calf:cow ratios are expected in fed HUs despite higher harvest intensity compared to unfed HUs.	A larger elk population in fed HUs may reduce the proportion of calves harvested.
I6 Density	Lower calf:cow ratios are expected in fed HUs with increasing density compared to unfed HUs.	High density in fed HUs could cause resource limitations during certain times of the year (e.g., summer) and reduce calf survival due to competition for food and other resources.

METHODS

Study area

Our study encompassed seven unfed and six fed elk HUs within the GYE across Montana and Wyoming. Data for these HUs spanned from 1995 to 2020, encompassing a period that updates the analysis by Foley et al. (2015) to reflect more recent conditions in the GYE, including the continued increase of wolf and grizzly bear populations. We used historical HU data from Wyoming, whereas the Madison and Northern Yellowstone areas in Montana were represented by herd subunits (Figure 1). The Hoback HU in Wyoming was dissolved in 2020, so we apportioned elk count and harvest data between the Piney (64%) and Upper Green River (36%) elk HUs based on area. Descriptions of feedgrounds (Smith, 2001) and their habitat characteristics (Forristal et al., 2012) were previously documented and are referenced herein.

Dependent variables

We aggregated data from the Wyoming Game and Fish Department, National Park Service, and Montana Fish, Wildlife & Parks for this analysis (Dugovich et al., 2025). We used calf:cow ratios as a proxy for recruitment of juvenile elk into the population because of the difficulty of directly monitoring juvenile survival across large areas and timeframes (Harris et al., 2008; Lukacs et al., 2018). Wyoming Game and Fish Department, National Park Service, and Montana Fish, Wildlife & Parks annually estimated HU calf:cow counts during late winter or spring using fixed-wing aircraft and helicopters, targeting periods when animals concentrate on historical winter range with high visibility on snow. Abundance counts were performed similarly throughout the winter, mostly between December and January. Aerial surveys were not conducted over feedgrounds, so the total HU count was the summation of the ground survey at feedgrounds plus the aerial survey elsewhere in fed HUs.

We also compared HUs based on elk densities on winter range and harvestable surplus. Elk density on winter range (hereafter, elk density) was calculated by dividing the HU count by winter range area for each HU. Winter ranges were defined as the area where substantial numbers of elk were typically found between December 1 and April 30 for Wyoming (B. Scurlock, Wyoming Game and Fish Department, written communication, February 21, 2025) and based on 90% utilization during the five most average winter periods (first heavy snow to spring green-up) over 10 years for Montana (Montana Fish, Wildlife, & Parks, 1999). These observations relied on aggregated data from ground and aerial observations, as well as telemetry/GPS collar data. The Northern Yellowstone herd subunit spanned the jurisdictions of Yellowstone National Park and Montana. Winter range area for this population was calculated as 1526.63 km² (Lemke et al., 1998), and counts included the Yellowstone National Park portion. In some analyses, we assessed the change in elk density from one year to the next (elk density_t − elk density_t−1) as an index of population productivity, while controlling for the overall amount of habitat across HUs of different sizes. Harvestable surplus, defined as the number of elk that could be harvested annually while maintaining a stable population, was calculated by summing the previous fall harvest and the change in population size from the previous year (harvest_t−1 + count_t − count_t−1).

Covariates

The amount of supplemental feeding is strongly correlated with winter conditions (Cross et al., 2007). Therefore, we considered supplemental feeding as a binary in order to include winter severity as a potential confounding variable. Winter severity was quantified by averaging the maximum daily April snow water equivalent (SWE) from two to three SNOTEL (National Water and Climate Center, 2023) sites within each HU's boundaries. These same SNOTEL sites were used to calculate cumulative summer rainfall from June to September, recognizing their limitations in potentially misrepresenting the moisture regime experienced by elk at lower altitudes (Brennan et al., 2013). Grizzly bear density was derived from the updated GYE grizzly bear density index (Bjornlie et al., 2014; Corradini et al., 2023). Only a portion of the Afton HU overlapped the grizzly bear density index layer, but we assumed the same low index across the HU because the indexes in nearby HUs were similarly low (Appendix S1; Figure S1a). Due to limitations in accurate wolf density estimates across the study area, we used a coarse presence/absence approach based on wolf pack territory overlap with HUs in Wyoming and documented wolf counts in Montana elk herd subunits. Wolf presence began in the Northern Yellowstone herd subunit in 1995 with the initial release by the National Park Service (Smith et al., 2020). Wolf pack ranges in Wyoming were determined through capture, telemetry flights, and winter observations by the Wyoming Game and Fish Department and National Park Service. Finally, hunter surveys provided elk harvest estimates (Rich et al., 2013; Wyoming Game and Fish Department, 2025a). As a covariate, harvest estimates were divided by winter range area to create winter range harvest density. Given that our target productivity measures occurred in late winter to spring, all covariates were lagged by one year for analysis as explanatory variables (Figure 2).

[IMAGE OMITTED. SEE PDF]

Data analysis

One of our primary interests was to identify the effects of supplemental feeding on elk demographics in the GYE. Where supplemental feeding occurs, however, effects may be correlated with other covariates, such as snowpack and predator densities. We first analyzed the effect of supplemental feeding without any other covariates in an unconditional model to assess the relationship between feeding and elk productivity. We then attempted to control for confounding variables with models that included predation and environmental conditions. Because of moderate (>0.5) collinearity (Pearson's correlation coefficients) between covariates (total harvest and elk density = 0.581; grizzly bear density index and elk density = 0.557; wolf presence and year = 0.540; grizzly bear density index and elk calf:cow ratios = −0.526; wolf and grizzly bear density = 0.501), these variables were entered into separate models. Year was omitted from the analyses because of the increasing trend of predators and decreasing SWE and rainfall over time.

Productivity models

We conducted Bayesian hierarchical analyses using the rstanarm package (Goodrich et al., 2024) within the R statistical environment (version 4.4.3; R Core Team, 2025). To enhance model interpretability, we standardized continuous explanatory variables by centering (subtracting the mean) and rescaling (dividing by two SDs; Gelman, 2008). To account for population-level differences not fully captured by our covariates, we included a HU-level random effect in our models. Let the index t represent annual observations of different elk management units, h, for each year from 1995 to 2020 (lagged 1994–2019). The calf:cow ratios unconditional model (1) took the following general form with a HU-level random intercept, b_h, distributed normally with a mean of 0 and variance ν² (2) and error term, ϵ_th, distributed normally with a mean of 0 and variance σ² (3): 1 $$$ \mathrm{calf}:\mathrm{co}{\mathrm{w}}_{th}\sim {\upbeta}_0+{\upbeta}_1\mathrm{fe}{\mathrm{d}}_h+{b}_h+{\upvarepsilon}_{th}. $$$ 2 $$$ {b}_h\sim N\left(0,{\upnu}^2\right). $$$ 3 $$$ {\upvarepsilon}_{th}\sim N\left(0,{\upsigma}^2\right). $$$ The other multivariable models are provided in Table 3. We used the default priors provided by the rstanarm package. Specifically, this included weakly informative normal priors (mean = 0, SD = 2.5) for the regression coefficients, a normal prior for the intercept (mean = 0, scale determined by the scaled outcome variable), and a half-Cauchy prior (scale = 2.5) for the residual SD. All models were run with four Markov chains, each with 10,000 discarded warm-up iterations and 10,000 test iterations. Model diagnostics followed the procedure outlined by Muth et al. (2018), ensuring model validity: no divergent transitions were observed, effective sample sizes were adequate for all parameters (>9000), and the potential scale reduction factor (R̂) remained below 1.1 for all parameters. To describe the results, we used mean and 89% (significance; Makowski et al., 2019) and 50% (directionality) equal-tailed Bayesian credible intervals (BCIs).

TABLE 3 Multiple mixed models were used among productivity metrics due to collinearity.

Model category	Models	Hypotheses
Unconditional	aProductivity ~ fed + (1|HU)	H1
Multivariable	Productivity ~ fed + SWE + rain + grizzly + (1|HU)	H1, H2, H3, H4
Productivity ~ fed + SWE + rain + wolf + harvest + (1|HU)	H1, H2, H3, H5, H6
Calf:cow ~ fed + SWE + rain + density + (1|HU)	H1, H2, H3, H7
Δ elk density or surplus ~ fed + SWE + rain + calf:cow + (1|HU)	H1, H2, H3, H8

Projection prediction model

To account for the limitations of a binary feeding effect (Table 2) and address collinearity among variables, we deployed projection predictive feature selection, specifically to identify informative features and interaction effects (Piironen et al., 2020). This Bayesian approach iteratively searches for a subset of predictors that can effectively replicate the predictions of a full model while minimizing the risk of overfitting. We used the projpred package (Piironen et al., 2023) to select the optimal subset of predictor variables, evaluating all possible combinations to maximize predictive accuracy while minimizing model complexity. This process achieved two key goals: (1) it reduced collinearity among predictors, improving model robustness, and (2) it ensured that the selected features retained strong predictive power while also capturing potential interaction effects. We used a regularized horseshoe prior, T₀ (defined in Equation 4; Piironen & Vehtari, 2017), with D = 14 (number of regression coefficients), p₀ = 4 (prior guess for non-zero regression coefficients), and N = 299 (number of complete observations) for the calf:cow ratio response models: 4 $$$ {T}_0=\frac{p_0}{D-{p}_0}\times \frac{1}{\sqrt{N}}. $$$ We ran a reference model with all covariates, including year and feeding interactions (5), and QR decomposition, a matrix factorization technique, was employed to decompose the predictor matrix into an orthogonal matrix (Q) and an upper triangular matrix (R; Goodrich et al., 2024). This approach decomposed the design matrix for numerical stability to mitigate the influence of collinearity and provided a baseline for comparison with subsequent models. Because of model complexity and the higher predisposition of the regularized horseshoe prior to trigger divergent transitions, we omitted the HU-level random effect from the feature selection process. Lagged observations were defined as t − 1. The full reference model without the inclusion of the HU-level random effect was: 5 $$$ \mathrm{calf}:\mathrm{co}{\mathrm{w}}_{th}\sim {\upbeta}_0+{\upbeta}_1\mathrm{fe}{\mathrm{d}}_h+{\upbeta}_2{\mathrm{SWE}}_{\left(t-1\right)h}+{\upbeta}_3{\mathrm{rain}}_{\left(t-1\right)h}+{\upbeta}_4{\mathrm{grizzly}}_{\left(t-1\right)h}+{\upbeta}_5{\mathrm{w}\mathrm{olf}}_{\left(t-1\right)h}+{\upbeta}_6{\mathrm{harvest}\kern0.34em \mathrm{density}}_{\left(t-1\right)h}+{\upbeta}_7\mathrm{elk}\ \mathrm{densit}{\mathrm{y}}_{\left(t-1\right)h}+{\upbeta}_8{\mathrm{y}\mathrm{ear}}_t+{\upbeta}_9{\mathrm{fed}}^{\ast }{\mathrm{SWE}}_{\left(t-1\right)h}+{\upbeta}_{10}{\mathrm{fed}}^{\ast }{\mathrm{rain}}_{\left(t-1\right)h}+{\upbeta}_{11}{\mathrm{fed}}^{\ast }{\mathrm{grizzly}}_{\left(t-1\right)h}+{\upbeta}_{12}{\mathrm{fed}}^{\ast }{\mathrm{w}\mathrm{olf}}_{\left(t-1\right)h}+{\upbeta}_{13}{\mathrm{fed}}^{\ast }{\mathrm{harvest}\kern0.34em \mathrm{density}}_{\left(t-1\right)h}+{\upbeta}_{14}{\mathrm{fed}}^{\ast}\mathrm{elk}\ \mathrm{densit}{\mathrm{y}}_{\left(t-1\right)h}+{\upvarepsilon}_{th}. $$$ We used leave-one-out cross validation (Vehtari et al., 2017) for model selection and a predictive performance graph to choose the optimal number of covariates where the performance leveled off. The final submodel, which included the HU-level random effect, was selected based on this procedure. The reference and final models passed diagnostic tests as above in the productivity models section. Whereas we initially aimed to evaluate the effects of supplemental feeding and other ecological factors on overall elk productivity, data constraints limited our analysis to only calf:cow ratios.

RESULTS

Observational data

Fed elk HUs exhibited a less pronounced decline in calf:cow ratios (mean = 0.30, SD = 0.06) compared to unfed HUs (mean = 0.26, SD = 0.07; Figure 3a). This suggests potential differences in trophic interactions affecting HU productivity. Elk densities showed greater variability among fed HUs (mean = 5.68, SD = 4.48) compared to unfed HUs (mean = 3.74, SD = 2.24; Figure 3b). Specifically, Jackson and Fall Creek fed HUs maintained higher densities, whereas other fed HUs were comparable to unfed HUs. Annual surplus displayed similar variability between fed (mean = 943, SD = 793) and unfed HUs (mean = 777, SD = 948; Figure 3c). Investigating some of the underlying differences that may exist between these regions, we found that climate factors of SWE and summer rainfall displayed similar trends across all HUs (Appendix S1: Figure S2). The Jackson HU harbored the highest grizzly bear density index, followed by the Upper Green River HU, whereas other fed HUs only recently experienced grizzly bear recolonization (Appendix S1: Figure S1a). Wolf presence has expanded across the region since 1995 but appeared to have less persistence in some of the fed HUs (Appendix S1: Figure S1b). Jackson and Fall Creek HUs exhibited slightly higher harvest density estimates as fed HUs than other regions, although the Jackson HU experienced the greatest decline over time (Appendix S1: Figure S1c).

[IMAGE OMITTED. SEE PDF]

Productivity models

Model responses generally aligned with posterior predictive distributions, but the observed change in elk density and surplus models exhibited lower variability compared with the data modeled from the posterior distributions (Appendix S1: Figure S3). This is probably because these combined data responses represent a convolution of the original distributions (i.e., elk count and winter range area; elk count_t, elk count_t−1, and harvest), which often leads to smoother and narrower distributions. Supplemental elk feeding's unconditional relationship with population dynamics was a modest positive association with calf:cow ratios (H1; β = 0.049 [89% BCI = 0.006, 0.093]; Figure 4a1). This coefficient value represents a change of 0.049 calves per cow in fed HUs compared with unfed HUs. We found no significant association between supplemental feeding and change in elk density (H1; β = −0.055 [89% BCI = −0.197, 0.088]; Figure 4a2) or with harvestable surplus (H1; β = 164 [89% BCI = −254, 582]; Figure 4a3; Appendix S1: Table S1).

[IMAGE OMITTED. SEE PDF]

The feeding effect persisted across the calf:cow ratios multivariable feeding effect models (H1; mean β = 0.046 [mean 89% BCI range = 0.007, 0.085]; Figure 4b1–d1). In these models, favorable summer rainfall was associated with higher calf:cow ratios (H3; mean β = 0.014 [mean 89% BCI range = 0.003, 0.025]; Figure 4b1–d1). In the top-down models, grizzly bear density index (H4; β = −0.052 [89% BCI = −0.072, −0.031]; Figure 4b1; Appendix S1: Table S2) and wolf presence (H5; β = −0.048 [89% BCI = −0.059, −0.036]; Figure 4c1; Appendix S1: Table S3) were negatively related to calf:cow ratios, whereas harvest density (i.e., number of harvested elk per km² of winter range) had a suggestive negative relationship (H6; β = −0.015 [89% BCI = −0.031, 0.002]; Figure 4c1, Appendix S1: Table S3). Elk density (H7; β = −0.004 [89% BCI = −0.037, 0.032]; Figure 4d1; Appendix S1: Table S4) was not associated with calf:cow ratios.

A supplemental feeding effect was not associated with changes in elk density in the multivariable models (H1; mean β = −0.009 [mean 89% BCI range = −0.174, 0.157]; Figure 4b2–d2). Total harvest total showed a negative relationship with the change in elk density from year to year (H6; β = −0.165 [89% BCI = −0.316, −0.017]; Figure 4c2). In addition, SWE was also negatively associated with the change in elk density in the population dynamics model (H2a; β = −0.168 [89% BCI = −0.321, −0.017]; Figure 4d2).

Finally, the harvestable surplus models also did not find a significant relationship of supplemental feeding (H1; mean β = 162 [mean 89% BCI range = −360, 675]; Figure 4b3–d3). Across the multivariable harvestable surplus models, SWE had a consistent negative relationship (H2a; mean β = −249 [mean 89% BCI range = −468, −30]; Figure 4b3–d3), whereas the grizzly bear density index (H4; β = −537 [89% BCI = −969, −117]; Figure 4b3) and wolf presence (H5; β = −220 [89% BCI = −415, −24]; Figure 4c3) were negatively correlated with surplus. The calf:cow ratio was positively associated with harvestable surplus (H8: β = 216 [89% BCI = 15, 418]; Figure 4d3). We explored the potential influence of unmeasured annual variation by fitting additional models that included year as a random effect. These models yielded similar patterns to the above results.

Projection prediction model

To explore potential mechanisms underlying the feeding effect on calf:cow ratios, we used projection prediction. The best-fitting submodel included six covariates (Appendix S1: Figure S4) and incorporated predator interactions and elk density. Interaction plots revealed that fed elk populations experienced less pronounced declines in calf:cow ratios in response to increased grizzly bear abundance (I3; Figure 5a) and wolf presence (I4; Figure 5b) compared with unfed populations. Although the unconditional analysis presented previously suggested a positive overall association of feeding on calf:cow ratios, the multivariable submodel indicated that feeding interacted with predator impacts, partially offsetting grizzly bear (I3; β = 0.046 [89% BCI = −0.004, 0.095]) and wolf (I4; β = 0.051 [89% BCI = 0.022, 0.080]) negative influence (Figure 5c; Appendix S1: Table S5). Although the model without the HU-level random effect exhibited similar patterns, the influences of grizzly bears and elk density were more pronounced (Appendix S1: Figure S5a). Posterior predictive checks suggested our final submodel performed similarly to the reference model, validating our model selection strategy (Appendix S1: Figure S5b,c).

[IMAGE OMITTED. SEE PDF]

DISCUSSION

Food web research is replete with examples of unintended consequences arising from trophic manipulations (e.g., Cortés-Avizanda et al., 2016; O'Gorman & Emmerson, 2009; Oro et al., 2013). Our research on supplemental elk feeding in the GYE revealed how even seemingly beneficial interventions can disrupt natural ecological processes through indirect trophic routes. Whereas supplemental feeding is implemented to mitigate the effect of severe winters and improve population productivity while controlling elk distributions and limiting elk–human conflicts (Cross et al., 2007; Kahn, 2000; Smith, 2001; Smith & Anderson, 1998), we found its effects on elk productivity to be limited and context-dependent. Specifically, although calf:cow ratios were slightly higher in fed elk HUs (mean = 0.30) compared to unfed HUs (mean = 0.26), the magnitude of this feeding effect was modest (β = 0.049) but was consistent with H1. The stability of this response across calf:cow ratio models with different combinations of covariates underscores the robustness of this relationship (Oster, 2019). We found no significant effect of supplemental feeding on either elk density or harvestable surplus. The sensitivity of our models, as evidenced by the detection of significant correlations for all covariates besides density at some point in the analysis, suggests that a feeding effect would probably have been detected if present. This nuanced effect of supplemental feeding challenges the assumption of consistent and substantial improvements in elk productivity, underscoring the importance of considering broader ecological interactions—often neglected in research biased toward provisioned species (Shutt & Lees, 2021)—when implementing trophic manipulations to help achieve management goals.

Our results suggest that predation plays a key role as an indirect effect, limiting the benefits of supplemental feeding. Our coarse metric of wolf presence indicated a strong negative association between wolves and elk productivity, most likely due to their year-round predation pressure and documented preference for calves (Metz et al., 2012; Stahler et al., 2006). Meanwhile, grizzly bear predation concentrates on newborn calves (French & French, 1990). Despite these differing predation strategies, we observed consistent negative relationships between both wolf presence and grizzly bear densities on both calf:cow ratios and surplus elk (Figure 4), supporting hypotheses H4–5 and highlighting the potential challenges of utilizing supplemental feeding to enhance hunting opportunities in multi-predator systems. Moreover, predation negated the positive influence of feeding on calf:cow ratios (Figure 5, consistent with I3–4). We hypothesize that supplemental feeding can potentially increase elk productivity through bottom-up forcing, but this may not translate to increased abundance due to these top-down effects (Figure 6). Increased bottom-up resources may inadvertently subsidize predator populations by increasing prey availability (Roemer et al., 2001), potentially explaining the lack of a strong association between feeding and elk density and harvestable surplus. Whereas this study focused on wolves and grizzly bears, future research could investigate the potential impacts of other large predators, such as black bears and mountain lions, to gain a more complete understanding of these food web dynamics. Future changes to supplemental feeding regimes in the region (Cook et al., 2023, 2025) could provide valuable insights into the cascading effects of such management actions on the broader food web including predator and pathogen communities.

[IMAGE OMITTED. SEE PDF]

The potential for additive mortality from hunting, combined with environmental variability, further complicates our understanding of productivity perturbations deriving from supplemental feeding. We found no significant relationship between feeding and harvestable surplus, echoing previous research on limited feeding effects on hunter success (Milner et al., 2014). Harvest was associated with decreases in calf:cow ratios and elk density (H6), consistent with the idea that hunting mortality can be additive to other sources of mortality, such as predation (Brodie et al., 2013; Vucetich et al., 2005). Notably, the interaction between feeding and winter severity (I1) was not a strong predictor of calf:cow ratios, aligning with previous findings (Foley et al., 2015). This casts doubt on the justification of supplemental feeding as a means of mitigating winter severity, although we acknowledge that calf mortality occurring later in the year (after our January–March assessment period) could mask such an effect (but refer to Harris et al., 2008). Projected reductions in mid-century snowpacks (Hostetler et al., 2021) may further lessen the need for supplemental feeding. However, the potential trade-offs could be considered: the higher elk densities observed in some fed populations (e.g., Jackson and Fall Creek) could exacerbate forage shortages during a longer, drier, and more fire-prone warm season (Taper & Gogan, 2002) that promotes increased predation (Singer et al., 1997). These results emphasize the need to consider the long-term ecological implications of supplemental feeding, moving beyond short-term benefits to ensure the sustainability of elk populations and the health of the broader ecosystem.

Several methodological considerations warrant attention. Elk abundance estimates were derived from a combination of ground and aerial surveys. Whereas aerial surveys are susceptible to sightability bias due to factors like snow cover and vegetation (Anderson & Lindzey, 1996; Cogan & Diefenbach, 1998; Samuel et al., 1987; Unsworth et al., 1990), we used unadjusted counts to enable comparisons between Wyoming and Montana. Although this could lead to underestimates of absolute abundance, our focus on relative differences between fed and unfed areas probably mitigates this bias, assuming relatively consistent sightability. Other limitations include the inability to incorporate brucellosis prevalence due to insufficient data, despite its known negative impact on elk reproduction (Cotterill et al., 2018). Furthermore, harvest quantification methods varied between states and over time, and winter range classifications, while acknowledging their limitations as coarse measures, were derived from diverse observational techniques. Whereas we observed smaller winter ranges for fed units, probably due to elk aggregation around feedgrounds, this difference was not statistically significant (Appendix S1: Figure S6). Even with the limitations inherent in our data and methods, consistent patterns emerged revealing the complex and multifaceted effects of supplemental elk feeding within the GYE, highlighting the interconnectedness of ecological factors and supporting a holistic approach to management.

Further research could disentangle the complex interplay of supplemental feeding, predation, disease, climate, and elk population dynamics. Future studies employing techniques such as juvenile collaring to directly measure survival rates, coupled with advanced modeling approaches like structural equation modeling, would allow for more precise quantification of these relationships and help isolate the effects of confounding factors. Feeding is projected to increase the prevalence of CWD and suppress elk abundance (Cook et al., 2023, 2025). Like predators, disease may initiate a trophic cascade (Buck & Ripple, 2017), but the effect of CWD in elk on predator populations is unknown (Figure 6). Finally, exploring the potential for synergistic effects among multiple stressors, including climate change (Rickbeil et al., 2019) and human disturbance (Naylor et al., 2009), could facilitate the development of comprehensive and resilient elk management strategies that balance elk populations, hunting opportunities, and large carnivore conservation.

CONCLUSIONS

Whereas supplemental feeding is often implemented for various reasons, including mitigating property damage (Smith, 2001) and reducing disease transmission risk (Cross et al., 2007), our findings offer limited support for the justification of increased winter survival or enhanced hunting opportunities. Our research suggests that increased predation pressure, potentially linked to supplemental feeding, may decouple elk productivity from overall abundance, meaning that the intended excess elk from feeding are most likely being consumed by predators. The food web perspective we developed here has implications for future management decisions. If feeding is halted in advance of CWD taking hold, it may result in less drastic reductions in elk because it could reduce both subsidized predation and density-dependent disease.

AUTHOR CONTRIBUTIONS

Conceptualization: Paul C. Cross. Data collection and curation: Eric K. Cole, Sarah R. Dewey, Daniel R. MacNulty, Brandon M. Scurlock, and Daniel R. Stahler. Methodology: Brian S. Dugovich, Emma M. Tomaszewski, and Paul C. Cross. Analysis: Brian S. Dugovich. Visualization: Brian S. Dugovich and Emma M. Tomaszewski. Writing—original draft preparation: Brian S. Dugovich. Review and editing: All coauthors.

ACKNOWLEDGMENTS

We thank Montana Fish, Wildlife & Parks biologists, Kelly Proffitt, Michael Yarnall, and Julie Cunningham, for compiling and providing the Montana elk data. At the time of publication, these data were not publicly available. We also thank Wyoming Game and Fish Department biologist, Kenneth Mills, for assembling the Wyoming wolf data, which at the time of publication were not publicly available. We are grateful to Troy Koser, Wynne Moss, Will Rogers, and Tucker Russell for help with manuscript preparation. Finally, we thank our external reviewer, Frank van Manen. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT

Data (Dugovich et al., 2025) are available from the USGS ScienceBase-Catalog: .

References

Anderson, C. R., Jr., and F. G. Lindzey. 1996. “Moose Sightability Model Developed from Helicopter Surveys.” Wildlife Society Bulletin 24: 247–259.

Bartmann, R. M., G. C. White, and L. H. Carpenter. 1992. “Compensatory Mortality in a Colorado Mule Deer Population.” Wildlife Monographs 121: 3–39.

Becker, D. J., D. G. Streicker, and S. Altizer. 2015. “Linking Anthropogenic Resources to Wildlife‐Pathogen Dynamics: A Review and Meta‐Analysis.” Ecology Letters 18: 483–495.

Bjornlie, D. D., F. T. Van Manen, M. R. Ebinger, M. A. Haroldson, D. J. Thompson, and C. M. Costello. 2014. “Whitebark Pine, Population Density, and Home‐Range Size of Grizzly Bears in the Greater Yellowstone Ecosystem.” PLoS One 9: e88160.

Brennan, A., P. C. Cross, M. Higgs, J. P. Beckmann, R. W. Klaver, B. M. Scurlock, and S. Creel. 2013. “Inferential Consequences of Modeling Rather than Measuring Snow Accumulation in Studies of Animal Ecology.” Ecological Applications 23: 643–653.

Brodie, J., H. Johnson, M. Mitchell, P. Zager, K. Proffitt, M. Hebblewhite, M. Kauffman, et al. 2013. “Relative Influence of Human Harvest, Carnivores, and Weather on Adult Female Elk Survival across Western North America.” Journal of Applied Ecology 50: 295–305.

Buck, J. C., and W. J. Ripple. 2017. “Infectious Agents Trigger Trophic Cascades.” Trends in Ecology and Evolution 32: 681–694.

Christianson, D., and S. Creel. 2014. “Ecosystem Scale Declines in Elk Recruitment and Population Growth with Wolf Colonization: A Before‐After‐Control‐Impact Approach.” PLoS One 9: e102330.

Cogan, R. D., and D. R. Diefenbach. 1998. “Effect of Undercounting and Model Selection on a Sightability‐Adjustment Estimator for Elk.” The Journal of Wildlife Management 62: 269–279.

Cook, J. D., G. G. Cotterill, M. C. McEachran, T. A. Graves, E. K. Cole, and P. C. Cross. 2025. “Decision Framing Overview and Performance of Management Alternatives for Bison and Elk Feedground Management at the National Elk Refuge in Jackson, Wyoming (No. 2024‐5119‐A).” U.S. Geological Survey. https://pubs.usgs.gov/publication/sir20245119A.

Cook, J. D., P. C. Cross, E. M. Tomaszewski, E. K. Cole, E. H. Campbell Grant, J. M. Wilder, and M. C. Runge. 2023. “Evaluating Management Alternatives for Wyoming Elk Feedgrounds in Consideration of Chronic Wasting Disease (No. 2023–1015).” U.S. Geological Survey. https://pubs.usgs.gov/publication/ofr20231015.

Corradini, A., M. A. Haroldson, F. Cagnacci, C. M. Costello, D. D. Bjornlie, D. J. Thompson, J. M. Nicholson, K. A. Gunther, K. R. Wilmot, and F. T. van Manen. 2023. “Evidence for Density‐Dependent Effects on Body Composition of a Large Omnivore in a Changing Greater Yellowstone Ecosystem.” Global Change Biology 29: 4496–4510.

Cortés‐Avizanda, A., G. Blanco, T. L. DeVault, A. Markandya, M. Z. Virani, J. Brandt, and J. A. Donázar. 2016. “Supplementary Feeding and Endangered Avian Scavengers: Benefits, Caveats, and Controversies.” Frontiers in Ecology and the Environment 14: 191–199.

Cotterill, G. G., P. C. Cross, A. D. Middleton, J. D. Rogerson, B. M. Scurlock, and J. T. Du Toit. 2018. “Hidden Cost of Disease in a Free‐Ranging Ungulate: Brucellosis Reduces Mid‐Winter Pregnancy in Elk.” Ecology and Evolution 8: 10733–10742.

Creel, S., and J. J. Rotella. 2010. “Meta‐Analysis of Relationships between Human Offtake, Total Mortality and Population Dynamics of Gray Wolves (Canis lupus).” PLoS One 5: e12918.

Cross, P. C., W. H. Edwards, B. M. Scurlock, E. J. Maichak, and J. D. Rogerson. 2007. “Effects of Management and Climate on Elk Brucellosis in the Greater Yellowstone Ecosystem.” Ecological Applications 17: 957–964.

Dugovich, B. S., E. M. Tomaszewski, E. K. Cole, S. R. Dewey, D. R. MacNulty, B. M. Scurlock, D. R. Stahler, and P. C. Cross. 2025. “Elk Herd Unit Ecological Data for Northwestern Wyoming and Southwestern Montana in Relation to Elk Feedgrounds from 1995–2020.” U.S. Geological Survey Data Release. https://doi.org/10.5066/P137Y5UB.

Foley, A. M., P. C. Cross, D. A. Christianson, B. M. Scurlock, and S. Creel. 2015. “Influences of Supplemental Feeding on Winter Elk Calf:Cow Ratios in the Southern Greater Yellowstone Ecosystem.” The Journal of Wildlife Management 79: 887–897.

Forristal, V. E., S. Creel, M. L. Taper, B. M. Scurlock, and P. C. Cross. 2012. “Effects of Supplemental Feeding and Aggregation on Fecal Glucocorticoid Metabolite Concentrations in Elk.” The Journal of Wildlife Management 76: 694–702.

French, S. P., and M. G. French. 1990. “Predatory Behavior of Grizzly Bears Feeding on Elk Calves in Yellowstone National Park, 1986‐88.” Bears: Their Biology and Management 8: 335–341.

Galloway, N. L., R. J. Monello, D. Brimeyer, E. K. Cole, and N. T. Hobbs. 2021. “Supporting Adaptive Management with Ecological Forecasting: Chronic Wasting Disease in the Jackson Elk Herd.” Ecosphere 12: e03776.

Gelman, A. 2008. “Scaling Regression Inputs by Dividing by Two Standard Deviations.” Statistics in Medicine 27: 2865–2873.

Goodrich, B., J. Gabry, I. Ali, and S. Brilleman. 2024. “rstanarm: Bayesian Applied Regression Modeling Via Stan.” R Package Version 2.32.1. https://mc-stan.org/rstanarm.

Griffin, L. L., and S. Ciuti. 2023. “Should We Feed Wildlife? A Call for Further Research into this Recreational Activity.” Conservation Science and Practice 5: e12958.

Harris, N. C., M. J. Kauffman, and L. S. Mills. 2008. “Inferences about Ungulate Population Dynamics Derived from Age Ratios.” The Journal of Wildlife Management 72: 1143–1151.

Hostetler, S. W., C. Whitlock, B. Shuman, D. Liefert, C. Wolf Drimal, and S. Bischke. 2021. Greater Yellowstone Climate Assessment: Past, Present, and Future Climate Change in Greater Yellowstone Watersheds. Bozeman, MT: Montana State University, Institute on Ecosystems. https://scholarworks.montana.edu/handle/1/16361.

Jones, J. D., M. J. Kauffman, K. L. Monteith, B. M. Scurlock, S. E. Albeke, and P. C. Cross. 2014. “Supplemental Feeding Alters Migration of a Temperate Ungulate.” Ecological Applications 24: 1769–1779.

Jonzén, N., H. Sand, P. Wabakken, J. E. Swenson, J. Kindberg, O. Liberg, and G. Chapron. 2013. “Sharing the Bounty—Adjusting Harvest to Predator Return in the Scandinavian Human–Wolf–Bear–Moose System.” Ecological Modelling 265: 140–148.

Kahn, H. B. 2000. “Uses and Valuation of the National elk Refuge, Wyoming.” Yale School of Forestry and Environmental Studies Bulletin Series 104: 139–170.

Lafferty, K. D., S. Allesina, M. Arim, C. J. Briggs, G. De Leo, A. P. Dobson, J. A. Dunne, et al. 2008. “Parasites in Food Webs: The Ultimate Missing Links.” Ecology Letters 11: 533–546.

Lemke, T. O., J. A. Mack, and D. B. Houston. 1998. “Winter Range Expansion by the Northern Yellowstone Elk Herd.” Intermountain Journal of Sciences 4: 9.

Lubow, B. C., and B. L. Smith. 2004. “Population Dynamics of the Jackson Elk Herd.” The Journal of Wildlife Management 68: 810–829.

Lukacs, P. M., M. S. Mitchell, M. Hebblewhite, B. K. Johnson, H. Johnson, M. Kauffman, K. M. Proffitt, et al. 2018. “Factors Influencing Elk Recruitment across Ecotypes in the Western United States.” The Journal of Wildlife Management 82: 698–710.

Lynch, H. J., S. Hodge, C. Albert, and M. Dunham. 2008. “The Greater Yellowstone Ecosystem: Challenges for Regional Ecosystem Management.” Environmental Management 41: 820–833.

Maher, S. M. L., K. J. Barker, K. Kroetz, V. Butsic, B. Leonard, and A. D. Middleton. 2023. “Assessing the Ecosystem Services and Disservices Provided by Migratory Wildlife across the Greater Yellowstone Ecosystem.” Biological Conservation 283: 110090.

Makowski, D., M. S. Ben‐Shachar, and D. Lüdecke. 2019. “bayestestR: Describing Effects and Their Uncertainty, Existence and Significance within the Bayesian Framework.” Journal of Open Source Software 4(1541): 1–8.

Marczak, L. B., R. M. Thompson, and J. S. Richardson. 2007. “Meta‐Analysis: Trophic Level, Habitat, and Productivity Shape the Food Web Effects of Resource Subsidies.” Ecology 88: 140–148.

Metz, M. C., D. W. Smith, J. A. Vucetich, D. R. Stahler, and R. O. Peterson. 2012. “Seasonal Patterns of Predation for Gray Wolves in the Multi‐Prey System of Yellowstone National Park.” Journal of Animal Ecology 81: 553–563.

Milner, J. M., F. M. Van Beest, K. T. Schmidt, R. K. Brook, and T. Storaas. 2014. “To Feed or Not to Feed? Evidence of the Intended and Unintended Effects of Feeding Wild Ungulates.” The Journal of Wildlife Management 78: 1322–1334.

Montana Fish, Wildlife and Parks. 1999. “Montana Elk Winter Ranges, Summer Ranges, Calving Areas, and Migration Areas.” Data Basin. https://databasin.org/datasets/0eca4ea8f9b34555870599b815946d6e/.

Murray, M. H., D. J. Becker, R. J. Hall, and S. M. Hernandez. 2016. “Wildlife Health and Supplemental Feeding: A Review and Management Recommendations.” Biological Conservation 204: 163–174.

Muth, C., Z. Oravecz, and J. Gabry. 2018. “User‐Friendly Bayesian Regression Modeling: A Tutorial with rstanarm and shinystan.” The Quantitative Methods for Psychology 14: 99–119.

National Water and Climate Center. 2023. “Air and Water Database.” https://nwcc-apps.sc.egov.usda.gov/.

Naylor, L. M., M. J. Wisdom, and R. G. Anthony. 2009. “Behavioral Responses of North American Elk to Recreational Activity.” The Journal of Wildlife Management 73: 328–338.

O'Gorman, E. J., and M. C. Emmerson. 2009. “Perturbations to Trophic Interactions and the Stability of Complex Food Webs.” Proceedings of the National Academy of Sciences of the United States of America 106: 13393–13398.

Oro, D., M. Genovart, G. Tavecchia, M. S. Fowler, and A. Martínez‐Abraín. 2013. “Ecological and Evolutionary Implications of Food Subsidies from Humans.” Ecology Letters 16: 1501–1514.

Ossi, F., J. M. Gaillard, M. Hebblewhite, and F. Cagnacci. 2015. “Snow Sinking Depth and Forest Canopy Drive Winter Resource Selection More than Supplemental Feeding in an Alpine Population of Roe Deer.” European Journal of Wildlife Research 61: 111–124.

Oster, E. 2019. “Unobservable Selection and Coefficient Stability: Theory and Evidence.” Journal of Business and Economic Statistics 37: 187–204.

Peziol, M., L. M. Elbroch, L. A. Shipley, R. D. Evans, and D. H. Thornton. 2023. “Large Carnivore Foraging Contributes to Heterogeneity in Nutrient Cycling.” Landscape Ecology 38: 1497–1509.

Piironen, J., M. Paasiniemi, A. Catalina, F. Weber, and A. Vehtari. 2023. “projpred: Projection Predictive Feature Selection.” R Package Version 2.8.0. https://mc-stan.org/projpred/.

Piironen, J., M. Paasiniemi, and A. Vehtari. 2020. “Projective Inference in High‐Dimensional Problems: Prediction and Feature Selection.” Electronic Journal of Statistics 14: 2155–2197.

Piironen, J., and A. Vehtari. 2017. “Sparsity Information and Regularization in the Horseshoe and Other Shrinkage Priors.” Electronic Journal of Statistics 11: 5018–5051.

Potter, C. 2020. “Snowmelt Timing Impacts on Growing Season Phenology in the Northern Range of Yellowstone National Park Estimated from MODIS Satellite Data.” Landscape Ecology 35: 373–388.

Proffitt, K. M., J. A. Cunningham, K. L. Hamlin, and R. A. Garrott. 2014. “Bottom‐Up and Top‐Down Influences on Pregnancy Rates and Recruitment of Northern Yellowstone Elk.” The Journal of Wildlife Management 78: 1383–1393.

Putman, R. J., and B. W. Staines. 2004. “Supplementary Winter Feeding of Wild Red Deer Cervus elaphus in Europe and North America: Justifications, Feeding Practice and Effectiveness.” Mammal Review 34: 285–306.

R Core Team. 2025. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.

Raithel, J. D., M. J. Kauffman, and D. H. Pletscher. 2007. “Impact of Spatial and Temporal Variation in Calf Survival on the Growth of Elk Populations.” The Journal of Wildlife Management 71: 795–803.

Rich, L. N., R. E. Russell, E. M. Glenn, M. S. Mitchell, J. A. Gude, K. M. Podruzny, C. A. Sime, K. Laudon, D. E. Ausband, and J. D. Nichols. 2013. “Estimating Occupancy and Predicting Numbers of Gray Wolf Packs in Montana Using Hunter Surveys.” The Journal of Wildlife Management 77: 1280–1289.

Rickbeil, G. J. M., J. A. Merkle, G. Anderson, M. P. Atwood, J. P. Beckmann, E. K. Cole, A. B. Courtemanch, et al. 2019. “Plasticity in Elk Migration Timing Is a Response to Changing Environmental Conditions.” Global Change Biology 25: 2368–2381.

Rinella, M. J., R. Dean, M. Vavra, and C. G. Parks. 2012. “Vegetation Responses to Supplemental Winter Feeding of Elk in Western Wyoming.” Western North American Naturalist 72: 78–83.

Ripple, W. J., J. A. Estes, O. J. Schmitz, V. Constant, M. J. Kaylor, A. Lenz, J. L. Motley, K. E. Self, D. S. Taylor, and C. Wolf. 2016. “What Is a Trophic Cascade?” Trends in Ecology & Evolution 31: 842–849.

Roemer, G. W., T. J. Coonan, D. K. Garcelon, J. Bascompte, and L. Laughrin. 2001. “Feral Pigs Facilitate Hyperpredation by Golden Eagles and Indirectly Cause the Decline of the Island Fox.” Animal Conservation 4: 307–318.

Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. “Visibility Bias during Aerial Surveys of Elk in Northcentral Idaho.” The Journal of Wildlife Management 51: 622–630.

Shutt, J. D., and A. C. Lees. 2021. “Killing with Kindness: Does Widespread Generalised Provisioning of Wildlife Help or Hinder Biodiversity Conservation Efforts?” Biological Conservation 261: 109295.

Sinclair, A. R. E., and C. J. Krebs. 2002. “Complex Numerical Responses to Top‐Down and Bottom‐Up Processes in Vertebrate Populations.” Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 357: 1221–1231.

Singer, F. J., A. Harting, K. K. Symonds, and M. B. Coughenour. 1997. “Density Dependence, Compensation, and Environmental Effects on Elk Calf Mortality in Yellowstone National Park.” The Journal of Wildlife Management 61: 12–25.

Singer, F. J., and K. A. Schoenecker. 2003. “Do Ungulates Accelerate or Decelerate Nitrogen Cycling?” Forest Ecology and Management 181: 189–204.

Smith, B. L. 2001. “Winter Feeding of Elk in Western North America.” The Journal of Wildlife Management 65: 173–190.

Smith, B. L., and S. H. Anderson. 1998. “Juvenile Survival and Population Regulation of the Jackson Elk Herd.” The Journal of Wildlife Management 62: 1036–1045.

Smith, D. W., D. R. Stahler, and D. R. MacNulty, eds. 2020. Yellowstone Wolves: Science and Discovery in the World's First National Park. Chicago, IL: University of Chicago Press.

Stahler, D. R., D. W. Smith, and D. S. Guernsey. 2006. “Foraging and Feeding Ecology of the Gray Wolf (Canis lupus): Lessons from Yellowstone National Park, Wyoming, USA.” The Journal of Nutrition 136: 1923S–1926S.

Stewart, K. M., R. T. Bowyer, B. L. Dick, B. K. Johnson, and J. G. Kie. 2005. “Density‐Dependent Effects on Physical Condition and Reproduction in North American Elk: An Experimental Test.” Oecologia 143: 85–93.

Taper, M. L., and P. J. P. Gogan. 2002. “The Northern Yellowstone Elk: Density Dependence and Climatic Conditions.” The Journal of Wildlife Management 66: 106–122.

Unsworth, J. W., L. Kuck, and E. O. Garton. 1990. “Elk Sightability Model Validation at the National Bison Range, Montana.” Wildlife Society Bulletin 18: 113–115.

Vehtari, A., A. Gelman, and J. Gabry. 2017. “Practical Bayesian Model Evaluation Using Leave‐One‐Out Cross‐Validation and WAIC.” Statistics and Computing 27: 1413–1432.

Vucetich, J. A., D. W. Smith, and D. R. Stahler. 2005. “Influence of Harvest, Climate and Wolf Predation on Yellowstone Elk, 1961‐2004.” Oikos 111: 259–270.

Williams, E. S. 2005. “Chronic Wasting Disease.” Veterinary Pathology 42: 530–549.

Wright, G. J., R. O. Peterson, D. W. Smith, and T. O. Lemke. 2006. “Selection of Northern Yellowstone Elk by Gray Wolves and Hunters.” Journal of Wildlife Management 70: 1070–1078.

Wyoming Game and Fish Department. 2025a. “Annual Big Game HU Job Completion Reports.” Wyoming Game and Fish Department. https://wgfd.wyo.gov/hunting-trapping/job-completion-reports.

Wyoming Game and Fish Department. 2025b. “CWD Found in Elk Hunt Area 98.” Wyoming Game and Fish Department. https://wgfd.wyo.gov/news-events/cwd-found-elk-hunt-area-98.