https://cpw.cvlcollections.org/files/original/8c52666ab9f00ce4e016d21dfc1725f7.pdf 48f5ddfe4d3261e0be03a2c52c1b85d0 PDF Text Text The research in this publication was partially or fully funded by Colorado Parks and Wildlife. Dan Prenzlow, Director, Colorado Parks and Wildlife • Parks and Wildlife Commission: Marvin McDaniel, Chair • Carrie Besnette Hauser, Vice-Chair Marie Haskett, Secretary • Taishya Adams • Betsy Blecha • Charles Garcia • Dallas May • Duke Phillips, IV • Luke B. Schafer • James Jay Tutchton • Eden Vardy �Received: 5 March 2021 | Revised: 10 October 2021 | Accepted: 13 October 2021 DOI: 10.1002/jwmg.22175 RESEARCH ARTICLE Effects of willow nutrition and morphology on calving success of moose Forest P. Hayes1 | Joshua J. Millspaugh1 Ragan M. Callaway1 | Chad J. Bishop1 1 Wildlife Biology Program, University of Montana, Missoula, MT 59812, USA 2 Colorado Parks and Wildlife, Fort Collins, CO 80526, USA Correspondence Forest P. Hayes, Department of Fish, Wildlife, and Conservation Biology, 1474 Campus Delivery, Colorado State University, Fort Collins, CO 80523, USA. Email: forest.hayes@colostate.edu Present address Forest P. Hayes, Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins, CO 80523, USA. | Eric J. Bergman2 | Abstract Across much of North America, populations of moose (Alces alces) are declining because of disease, predation, climate change, and anthropogenic‐driven habitat loss. Contrary to this trend, populations of moose in Colorado, USA, have continued to grow. Studying successful (i.e., persistent or growing) populations of moose can facilitate continued conservation by identifying habitat features critical to persistence of moose. We hypothesized that moose using habitat with higher quality willow (Salix spp.) would have a higher probability of having a calf‐at‐heel (i.e., calving success). We evaluated moose calving success using repeated ground observations of collared in- Funding information dividuals with calves in an occupancy model framework to Colorado Parks and Wildlife account for detection probability. We then evaluated the impact of willow habitat quality and nutrition on moose calving success by studying 2 spatially segregated populations of moose in Colorado. Last, we evaluated correlations between willow characteristics (browse intensity, height, cover, leaf length, and species) and willow nutrition (dry matter digestibility [DMD]) to assess the utility of using those characteristics to assess willow nutrition. We found willow height and cover had a high probability of being positively associated with higher individual‐level calving success. Willow DMD, browse intensity, and leaf length were not predictive of individual moose calving success; however, the site with higher mean DMD consistently had higher mean estimates of calving success for the same year. Our results suggest surveying DMD is likely not a useful metric for assessing differences in calving J Wildl Manag. 2022;1–15. wileyonlinelibrary.com/journal/jwmg © 2022 The Wildlife Society | 1 �2 | HAYES ET AL. success of individual moose but may be of use at population levels. Further, the assessment of willow morphology and density may be used to identify areas that support higher levels of moose calving success. KEYWORDS Alces alces, calving success, Colorado, moose, nutrition, reproduction, Salix, willow Understanding the relationships between habitat and reproductive success is a key aspect of successfully managing wildlife populations. Moreover, interpreting this relationship is especially valuable for species in which direct annual evaluation of reproductive success is not feasible. Reproductive success of ungulates is closely linked to nutrition intake (Belovsky 1978, Heard et al. 1997, Testa and Adams 1998, Parker et al. 2009, Cook et al. 2013). In their broad‐scale assessment, Cook et al. (2013) noted the importance of adequate summer nutrition in driving reproductive performance of female elk (Cervus elaphus) and their offspring in the Northwest and Rocky Mountains. Similarly, Proffitt et al. (2016) noted differences in pregnancy rates of elk related to differences in nutritional conditions. Further, given the impact of maternal condition, which is directly attributed to nutritional condition, offspring survival can be affected (Bishop et al. 2009, Shallow et al. 2015). In conjunction, these relationships suggest that monitoring the condition of nutritional resources may be a suitable surrogate for measuring reproductive success. For moose (Alces alces), willow (Salix spp.) is often a key forage resource that likely affects maternal condition and, subsequently, calving success. As a low‐density and difficult to monitor species, moose have proven challenging and costly to survey with low bias (Månsson et al. 2011). When willow forage is available, moose will preferentially select willow over other forage (Renecker and Schwartz 2007). In areas with high concentration of willow, willow can comprise 90% of the diet of moose (Dungan and Wright 2005). The dietary importance of willow in these areas in conjunction with the well‐established links between nutrition intake, body condition, and reproduction of moose (Edwards and Ritcey 1958, Testa and Adams 1998, Ruprecht et al. 2016) suggest that the quality and quantity of willow forage are likely key factors affecting calving success. Summer nutrition intake is especially important to moose as they build fat reserves for times of year, such as winter, when energy consumption is insufficient to meet maintenance requirements (Schwartz et al. 1987, Renecker and Hudson 1989). Populations of moose with low nutrition availability have been previously associated with delayed maturation, pauses in reproduction, and low twinning rates (Boertje et al. 2019). Colorado, USA, lies at the southern extent of the range of moose (Timmermann and Rodgers 2017). Unlike many moose populations in the lower 48 states, populations of moose in Colorado have been relatively robust over the past 2 decades (Timmermann and Rodgers 2017). In Colorado, predation pressure on moose is minimal and the prevalence of parasites and disease is relatively low (Colorado Parks and Wildlife, unpublished data). As such, the primary factors limiting moose populations in Colorado are the ability to consume adequate nutrition and survive at southern latitudes (Van Ballenberghe and Ballard 2007). Researchers at Colorado Parks and Wildlife (CPW) reported moose pregnancy rates to differ between 2 spatially segregated, geographically proximate populations of moose in Colorado (E. J. Bergman, CPW, unpublished data). Similarities between populations in landscape composition, habitat availability, and geography may reduce the set of alternative explanatory factors. Forage quality and quantity remain poorly understood for these 2 populations and are logical variables to investigate that may explain differences between populations. Our objectives of this research were 2‐fold. First, we sought to understand the correlation between forage quality, quantity, and browse intensity on the probability of individual moose having a calf‐at‐heel, which we term calving success, by evaluating 2 spatially separated populations of moose in Colorado. We hypothesized that use of �MOOSE CALVING SUCCESS | 3 higher quality willow habitat during summer months would be positively associated with higher calving success. Our second objective was to evaluate correlations between willow characteristics and willow nutrition. The strong relationship between nutrition intake and calving success (Belovsky 1978, Heard et al. 1997, Testa and Adams 1998) makes nutrition desirable to monitor. That said, we recognize that extensive monitoring of willow nutrition may not be possible for future studies given the time and expense associated with these surveys. Thus, we evaluated relationships between willow characteristics and willow nutrition that may be useful to future conservation‐ and management‐related studies in which analyzing nutrition directly is not feasible. STUDY AREA We conducted research in 2 study areas located in northern Colorado from 2015–2019. The first study site, North Park, was located south of the Colorado‐Wyoming border 100 km west of the Rocky Mountain Front and near the town of Walden. Moose were first translocated to the area during the 1970s and have been there since then. The North Park study area was a wide (14–46 km), high‐elevation valley (2,400–2,750 m) composed of a mixture of rolling sagebrush (Artemisia spp.) hills, irrigated agricultural fields, and riparian corridors dominated by willow. Riparian willow communities were largely composed of Geyer's willow (S. geyeriana), mountain willow (S. monticola), planeleaf willow (S. planifolia), and Booth's willow (S. boothii). The hills surrounding North Park were primarily composed of lodgepole pine (Pinus contorta) with Englemann spruce (Picea engelmannii) intermixed. The second study site was along the Laramie River roughly 40 km northeast of the North Park study area and was separated from North Park by the Rawah Mountains (3,200–3,840 m). Moose were translocated to Laramie River during the early 1980s and have been there since then. The Laramie River Study area was composed of a narrow valley floor (3.0–8.5 km wide) and characterized by riparian willow communities along the river with similar species composition to the North Park study site. Uplands around Laramie River were characterized by rangeland co‐dominated by sagebrush and a grass‐forb mix but occurred with less frequency in comparison to North Park. Larger hills located within the valley were dominated by quaking aspen (Populus tremuloides), while slopes to the east and west were dominated by lodgepole pine and Engelmann spruce. Geographic features and climate of both study areas were similar. Elevation of riparian areas ranged from 2,400 m to 2,800 m. Temperatures ranged from highs in July of 26.5°C to lows of −15°C in January (U.S. Department of Agriculture National Resources Conservation Service 2020). North Park tended to have slightly higher maximum temperatures than Laramie River throughout the year by about 2.2°C. Mean low temperatures were more similar, although North Park tended to be slightly colder during the winter months. Average annual precipitation was around 40.4 cm for both sites. Moose in both study areas were managed by CPW. Management actions included limited male and female moose hunting. Predator assemblages were also similar between study sites. American black bears (Ursus americanus), mountain lions (Puma concolor), and coyotes (Canus latrans) were present in both sites, whereas wolves (Canis lupus) and grizzly bears (Ursus arctos) were absent. Black bears and mountain lions infrequently predated upon moose, resulting in minimal predation pressure (Bergman et al. 2020b). METHO DS Capture and observation of moose We captured moose in the North Park and Laramie River study areas as a part of a broader research initiative from 2015 to 2019. Between 20 December and 27 January of each winter (2015–2019), we captured adult (≥2 years old, as determined by body size) female moose via helicopter darting. We sedated moose using 1 of 3 different drug �4 | HAYES ET AL. combinations: BAM (54.6 mg of butorphanol, 18.2 mg of azaperone, and 21.8 mg of medetomidine) in combination with ketamine (200 mg), carfentanil (3 mg) in combination with xylazine (100 mg), or thiafentanil (10 mg) in combination with xylazine (25 mg). After handling, we antagonized capture drugs with naltrexone (100 mg, antagonist for carfentil and thiafentenil), tolazoline (500 mg, antagonist for azaperone and xylazine), or atipamezole (100–150 mg, antagonist for medetomidine and xylazine). Once sedated, we blindfolded moose to minimize stress. Moose received oxygen, via nasal cannula, to minimize risks of adult and fetal hypoxia. We subsequently fitted moose with satellite and global positioning system (GPS)‐equipped very high frequency radio‐collars (Vertex Plus, Vectronics Aerospace GmbH, Berlin, Germany; G5‐2D, Advanced Telemetry Systems, Isanti, MN, USA), and uniquely numbered ear tags. Following the first year of capture, we recaptured some previously captured moose, but only on an opportunistic basis. Moose captured in previous years retained satellite collars and remained available for observation in subsequent years. We captured 214 moose between 2015 and 2019. In total, there were 145 unique individuals, 80 from North Park and 65 from Laramie River. We initiated ground observations of collared moose beginning in mid‐May of each year when the probability of parturition was <0.50 (Bergman et al. 2020a). We attempted to observe each collared individual ≥1 time/week. Typically, 1 observer completed ground observations by radio‐tracking the collared moose to document the presence or absence of a calf. Recent GPS locations of moose expedited ground observations. When a single observer failed to gain an observation after 4 repeated efforts, we used a 2‐observer approach. During these scenarios, the second observer was positioned along the exit route that the moose was expected to take, and the first observer radio‐tracked the moose in the same manner as a single‐observer approach. We recorded an observation when we sighted an identifiable female moose and 1 of 2 conditions was met: we observed a calf in the immediate vicinity of the identifiable female or the surrounding 1–2‐m area was visible with no calf present. Repeated observations continued through the end of August. We initially prioritized moose for observation based on whether or not they had been captured the previous winter (i.e., moose with known pregnancy status were a higher priority for observation). Once individual moose had been observed ≥1 time, we prioritized animals based on timing of the most recent observation (i.e., individual moose observed most recently were lowest priority for upcoming observations and animals who had not been recently observed were a higher priority). In addition to observations that occurred as a part of the formal study process, we recorded a small number of opportunistic observations by the public (n = 9) during autumn. We only recoded opportunistic observations when individual identification of moose (possible because of unique ear tags) was available. Willow surveys We collected willow samples from the North Park and Laramie River study sites from 2017 to 2019 during the month of July to assess spatial variation in willow quality, quantity, and utilization (i.e., spatial samples). We used GPS locations from collared moose during June of the same year to represent areas used by moose during summer (Jun to Aug). We then randomly selected sample points from those classified as willow (Simpson et al. 2013). For the pilot year, 2017, we randomly selected 40 points from each study area from all points in June after filtering for land cover type. For 2018 and 2019, we selected 60 points from each study area and additionally stratified the random points by individual moose (i.e., we selected either 2 or 3 points from the used locations of each moose). By stratifying random locations by moose, we avoided biasing observation effort towards moose with more recorded used locations because of collar type or geographic location. We restricted random locations to those ≥20 m apart. If a point was too close to another location, did not have any willow present within 10 m, or was inaccessible (i.e., inaccessible private property), we selected alternative random locations (representing ~15% of sampled locations) from the same individual sequentially from a list of replacement points. �| MOOSE CALVING SUCCESS 5 At each spatial point, we flagged the location and identified the closest willow. We identified each willow to species and collected leaves and twigs from the closest branches to the marked point. We collected samples from the apical 15–20 cm of branches, which generally represent the current year's growth and, depending on the growth stage, previous year's growth. We restricted samples to branches between 0.5 m and 2.5 m above ground level, which is consistent with observed use by moose in our area. The effective browse range of moose reported in literature varies among studies (Stickney 1966, Bergström and Guillet 2002, Burkholder et al. 2017) but typically falls within a similar range. We collected leaf samples by stripping leaves from branches by hand to simulate moose foraging behavior and clipped twigs from areas where we had stripped leaves. For leaf and twig samples, we collected ≥10 g (wet weight) from each plant. We randomly selected 5 leaves from the sample and recorded the length of each. We placed leaf and twig samples in open paper bags and allowed them to air‐dry in a low humidity environment for ≥2 months. Each year, once dry, we transferred samples into sealed plastic bags to prevent further moisture exchange. We established 4 10‐m transects at each spatial sample location, 1 in each cardinal direction. We laid a tape reel along each transect and recorded each willow intercepting the tape, noting the intercepts, species, and plant height. We did not record willow intercepts outside of the typical browse range of moose (i.e., willow plants <0.5 m or with no leaves below 2.5 m). We measured browse use of the willow closest to the spatial point and of the willows closest to 5 m and 10 m along each transect. For the 5‐m and 10‐m samples, we selected the closest willow within 2.5 m (to avoid overlapping samples) of the transect for each distance and direction. We followed the sampling protocol developed by Stickney (1966) and modified by Burkholder et al. (2017) for evaluating browse use. In brief, we sampled the closest branch to the sample location within the browse range of moose and evaluated ≥20 twigs. We define twigs as an unbranched portion of a branch with apical or lateral growth consisting of current and previous year's growth (Burkholder et al. 2017). We classified each twig as belonging to 1 of 5 categories: unbrowsed (Nub ), twigs that do not show any evidence of foraging; browsed (Nb ), twigs that have had the apical portion within 15–20 cm removed from foraging; leaf stripping (Ns ), twigs that have not been browsed but that have leaves stripped off; browsed and leaf stripping (Nbs ), twigs that have been browsed and have leaf stripping farther along the branch; and heavily browsed (Nhb ), twigs that were browsed at a diameter of ≥0.5 cm. We expanded upon the adjusted estimator of browse use presented in Burkholder et al. (2017) to incorporate additional weighted classifications of browse intensity to reflect increased browse pressure (e.g., Nb receives a weight of 1 and Nhb receives a weight of 3). We calculated adjusted browse intensity (Badj ) for each plant as: Badj = Nb + Ns + (Nbs × 2) + (Nhb × 3) Nub + Nb + Ns + (Nbs × 2) + (Nhb × 3) In 2017, we selected 5 willow plants spatially representative of each study site and proximate to vehicle access to enable monitoring of leaf quality as a function of phenology through weekly sampling (i.e., phenology samples). For each plant, we collected leaf and twig samples following the willow spatial sample methodology every week from May to August of each year (2017–2019). We dried phenology samples under the same conditions as the spatial samples. For spatial and phenology samples, we standardized sample composition to include a 2:1 ratio of forage and browse by dry weight. We sent samples to Dairy One Forage Laboratory (Dairy One, Ithaca, NY, USA) for analysis. We had all samples analyzed for neutral detergent fiber (NDF), acid detergent fiber (ADF), ash, and acid detergent lignin. We did not evaluate crude protein because pilot data (Colorado Parks and Wildlife, unpublished data) indicated low variability between the 2 study sites. For the pilot study year (2017), we analyzed all spatial samples in addition to bi‐weekly phenology samples. For 2018 and 2019, we combined phenology samples from each study site in equal proportions prior to analysis to represent average study site nutritional conditions. For spatial samples, �6 | HAYES ET AL. we analyzed each sample individually. We calculate percent dry matter digestibility (DMD) using ash and NDF values for each sample as follows (Robbins et al. 1987a, b):  −0.0451× lignin ×100    × NDF + ((−16.03 + 1.02 × (100 − NDF)) − 0) NDF  − 0.03 × (ash × 100) DMD = 0.9231 × e    We did not analyze concentrations of secondary compounds (i.e., tannins) because previous researchers suggested estimations of digestibility in woody forage may not need to be adjusted for tannins for moose because they may benefit from ingesting tannins (Spaeth et al. 2002). Further, although tannin concentrations may influence forage selection, they are not a driving factor in selection by moose (Stolter et al. 2005). Last, the saliva of moose binds to some types of tannins, reducing negative effects (Hagerman and Robbins 1993, Juntheikki 1996). Analytical methods For each willow sample at each spatial point, we modeled the effect of willow species, browse intensity, leaf length, and willow intercept on DMD. We did not consider the effect of year on nutrition as the intent of this analysis is to explain DMD through the lens of willow characteristics. We used a Bayesian framework for the analysis and evaluated the effect of each covariate based on 95% credible intervals using the following equation: DMD = species + browse intensity + leaf + willow intercept Because we collected nutrition samples from a single willow at each spatial point (i.e., we did not use 5‐m and 10‐m samples for nutrition analysis), we used the species, browse intensity, and leaf lengths from 1 plant per point. We limited analysis of species effect on DMD to species with ≥10 samples. For leaf length, we averaged the 5 leaf lengths recorded during collection of the vegetation sample. We averaged the amount of willow intercept (i.e., willow overlapping transects) for each point. We chose to evaluate only a global model, containing all covariates, and interpreted effects based on the strength of the effect and the 95% credible intervals. We selected this approach based on the objective of evaluating the utility of each covariate as opposed to building the most parsimonious model to describe willow nutrition. We estimated the proportion of female moose with a calf for each year using multiple repeated observations of calf presence using an occupancy model as described by Bergman et al. (2020b). An occupancy model framework uses a flexible model structure and enables estimation of occupancy (ψ) while accounting for imperfect detection (p). In this case, ψ represents calving success (i.e., the probability of a female moose having a calf‐at‐heel). We chose to model ψ varying for each year and site and p varying each month. This structure allowed us to evaluate differences in calving success between study sites and the effect of additional covariates on ψ for each study site. To evaluate the effect of spatial willow covariates on ψ, we used average values from each point and evaluated which points overlapped with individual moose core home ranges. For each spatial point, we calculated the mean value of adjusted browse intensity (Badj ), leaf length, and the proportion of amount of transect intersected by willow. We then calculated the 75% home range for each moose and year using a kernel density estimate (KDE) with the adehabitatHR R package (Calenge 2006). We chose the 75% KDE to represent the expanded core home range of each animal in which the majority of habitat use occurs. We then intersected spatial samples with each home range estimate. Last, we assigned covariate values from intersecting points to the respective year and moose. We used a linear model to estimate the effect of individual moose on the covariate of interest as follows: cov = μ + moose effecti , �| MOOSE CALVING SUCCESS 7 where cov is the covariate of interest (e.g., DMD), μ is the mean value for all individuals, and moose effect varies by individual. We used the estimate (moose effect) for each individual moose to evaluate the effect of each covariate on ψ. This approach has 2 benefits. First, because the core home ranges of moose overlap with differing numbers of spatial samples, the precision of observed covariates varies between individual moose. The linear model accounts for this through varied precision of the estimated effect. Second, this approach allows uncertainty from observed covariates to be correctly propagated in the calf occupancy model. We used a Bayesian framework for analysis of calving success as it offers the benefit of accurate error propagation across estimated variables and simplifies interpretation of estimated probabilities. We tested the effect of each covariate (DMD, leaf length, willow intercept, browse intensity) on calf occupancy in a univariate model using a sin link function as follows: ψ = (sin (b0 + moose effecti × covariate value) + 1)/2, where moose effect represents the modeled value from the previous equation and covariate value represents the measured value for each covariate (e.g., DMD). The sin link function is used to link the beta parameter (moose effect) to the real parameter (ψ) and constrain the real parameters to the [0, 1] interval, the possible range of values of occupancy. The sin link function is poorly documented but commonly used (White and Burnham 1999) instead of the logit link function when estimating parameters close to 0 or 1. We ran all Bayesian models for 100,000 iterations with 50,000 iterations of burn‐in on 6 chains at which point all parameters had reached convergence (r̂ < 1.01) as assessed by the Gellman‐Rubin diagnostic (Gellman and Rubin 1992). We conducted all analyses using the Program R software system (R Core Team 2020). RESULTS Analysis of spatial samples resulted in 319 measurements of DMD (Table 1). We censored 3 samples from North Park in 2017 because of poor drying. A linear model comparing sample week of phenology samples and DMD percent for each study site and year combination showed no trend in quality over time (|r| ≤ 0.01). Spatial willow covariates (DMD, browse intensity, leaf length, average willow height, willow cover) sampled during July of each year at each spatial point displayed variance across both site and year (Figure 1). Mean DMD was 3.7 percentage points higher in Laramie River than North Park in 2017 and 2.1 percentage points lower in 2018. In 2019, mean T A B L E 1 Numeric summary of spatial willow samples by species in Laramie River and North Park, Colorado, USA, 2017–2019. Willow species are Bebb's willow (S. bebbiana; SABE2), Booth's willow (SABO2), Drummond's willow (S. drummondiana; SADR), strapleaf willow (SAERL), Geyer's willow (SAGE2), Pacific willow (S. lasiandra; SALAC), mountain willow (SAMO2), and planeleaf willow (SAPL2) Willow species Year SABE2 SABO2 SADR SAERL 2017 0 7 0 1 2018 2 8 1 2019 1 8 Total 3 23 SAGE2 SALAC SAMO2 SAPL2 Total 52 0 22 1 83 7 56 4 29 10 117 0 16 53 3 16 22 119 1 24 161 7 67 33 319 �8 | HAYES ET AL. F I G U R E 1 Measurements of willow dry matter digestibility (A), browse intensity (B), leaf length (C), willow height (D), and willow cover (E) from willow samples within 2 moose population home ranges in Laramie River and North Park, Colorado, USA, 2017–2019 DMD was functionally identical between study areas, separated by 0.5 percentage points. Browse intensity at both sites was highly variable with mean values increasing slightly from 2017 to 2019. Leaf length was similar in both study sites and for all years. Average willow height had high sample variance, was similar across years, and had a higher mean in North Park for all years. Willow cover was also highly variable with a wide distribution of values and similar means for both sites with slightly lower cover on transects conducted in 2019. A Pearson's pairwise correlation test showed low correlation between all combinations of covariates (| r| < 0.25) except for willow height and willow, cover which had a positive correlation (r = 0.47). Analysis of DMD relative to willow browse intensity, height, cover, leaf length, and species revealed few of these factors had a strong influence on DMD with high certainty (Figure 2). Willow species was largely not a good predictor of DMD with wide credible intervals for each species. Strapleaf willow (Salix ligulifolia) had the strongest effect of any species evaluated and was the only species with credible intervals not overlapping zero. Estimates for Geyer's willow, mountain willow, and planeleaf willow were similar with a slight positive effect on DMD. Booth's willow had a slight negative effect on DMD, but the credible interval substantially overlapped zero. Browse intensity and willow height both had mean effect estimates that were close to zero and substantially overlapping credible intervals and thus were uninformative. Willow cover had a very slight negative effect on DMD with credible intervals substantially overlapping zero. The effect of leaf length also had a negative effect on DMD and was the only non‐species covariate with credible intervals not overlapping zero. Ground observation efforts resulted in 352 observations of moose (North Park = 148, Laramie River = 204) recorded during the study (Table 2). Mean time between observations was 32.8 ± 29.7 (SD) days. Of the observations, 201 (North Park = 88, Laramie River = 113) had no calf observed, 144 (North Park = 58, Laramie River = 86) had one calf observed, and 7 had twin calves observed (North Park = 2, Laramie River = 5). The observations consisted of 108 unique individuals, 60 in North Park and 48 in Laramie River. �| MOOSE CALVING SUCCESS 9 F I G U R E 2 Effect of willow browse intensity (proportion), height (m), cover (proportion), leaf length (mm), and species on percent dry matter digestibility in Laramie River and North Park, Colorado, USA, 2017–2019. Willow species are Booth's willow (SABO2), Geyer's willow (SAGE2), strapleaf willow (SAERL), mountain willow (SAMO2), and planeleaf willow (SAPL2). Bold lines and thin lines represent 50% and 95% credible intervals, respectively T A B L E 2 Numeric summary of observations of female moose in Laramie River and North Park, Colorado, USA, 2015–2019. Observations are categorized by month and year, and were used to model detection probability (p) and calf presence (ψ). Fall observations were opportunistic and include observations during September–December Observation period Year May Jun Jul Aug Fall Total 2015 19 40 19 6 1 85 2016 20 20 8 11 13 72 2017 12 18 8 12 2 52 2018 18 19 13 17 7 74 2019 2 49 8 5 5 69 Total 71 146 56 51 28 352 Mean calf occupancy estimates (ψ) for both study areas were highest in 2015 and 2016 ( x̅ range = 0.78–0.93, SD range = 0.07–0.11; Figure 3). Estimates were most disparate between populations in 2017 with means of ψ = 0.40 ± 0.13 for North Park and ψ = 0.72 ± 0.14 for Laramie River. In 2018, the ψ estimate was higher for North Park (ψ = 0.70 ± 0.12) than for Laramie River (ψ = 0.59 ± 0.10). In 2019, ψ was estimated at 0.50 for both populations (SD range = 0.11–0.12). For each year with nutrition data, when mean values of DMD differed, the site with higher DMD had higher estimates of ψ. Detection probability was lowest for May, during parturition, with an estimate of 0.48 ± 0.09 (Figure 4). Detection probability was highest during summer (Jun, Jul, Aug), increasing across that period from 0.77 to 0.90 (SD range = 0.06–0.09). Estimated detection probability of a calf in the fall was lower ( x̅ = 0.70 ± 0.11) relative to summer but remained higher than the detection probability in May. The univariate models testing covariate effects on occupancy showed mixed effects characterized by large credible intervals. The effect of DMD ( x̄ = 41.16 ± 5.80%) was functionally zero (β = 0.02) with a wide credible interval (−0.92–0.97). The credible interval for the effect of browse intensity ( x̄ = 0.40 ± 0.15) substantially overlapped zero (β = −0.59, credible interval = −1.88–0.79) and was therefore uninformative. �10 | HAYES ET AL. F I G U R E 3 Predicted moose calving success (i.e., the probability of a moose having ≥1 calf‐at‐heel; ψ), using an occupancy model in a Bayesian framework, for 2 populations of moose in Laramie River and North Park, Colorado, USA, 2015–2019. The center point represents the mean estimate for each combination of population and year. Bold and thin vertical lines represent 50% and 95% credible intervals, respectively F I G U R E 4 Estimates of moose calf detection probability (i.e., the probability of observing uniquely identifiable female moose with a calf‐at‐heel; p) in Laramie River and North Park, Colorado, USA, 2015–2019. We used an occupancy model framework to estimate a separate probability for each survey period (May, Jun, Jul, Aug, fall). The center point represents the mean estimate. Bold and thin vertical lines represent 50% and 95% credible intervals, respectively F I G U R E 5 Effect size of willow dry matter digestibility (%; DMD), browse intensity (proportion), willow height (m), willow cover (proportion), and leaf length (mm) on moose calf occupancy in Laramie River and North Park, Colorado, USA, 2015–2019. Points represent the mean estimate for each covariate. Bold lines and thin lines represent 50% and 95% credible intervals, respectively �MOOSE CALVING SUCCESS | 11 Of the 5 univariate models, willow cover ( x̅ = 0.44 ± 0.23) was the only covariate with a credible interval not overlapping zero (Figure 5). Willow cover showed a moderately strong positive effect on ψ with a mean estimate of 1.32 (credible interval = 0.15–1.88). In biological terms, willow cover of 45% was correlated with a calving success of 0.53, whereas willow cover of 65% was correlated with a calving success of 0.75. Willow height ( x̅ =2.47 ± 0.90 m) had a positive effect on ψ with an estimate of 0.79 (credible interval = −0.20 –1.90) but included a small probability of a negative effect (1.5%). Leaf length ( x̅ = 49.69 ± 12.70 mm) substantially overlapped zero with the widest credible interval (β = −0.42, credible interval = −1.93–1.75). DISCUSSION The lack of trend in willow phenology (i.e., DMD through time) for both study sites supports 2 of our key assumptions when designing this project. First, nutrition at both study sites is consistent across the survey period. Second, sampling willow nutrition during July provides a representative sample of nutrition throughout summer. Evidence that these 2 assumptions have been met supports the subsequent analyses of DMD and willow covariates gathered during July of each year. Site‐level DMD varied substantively across site and year without a consistent pattern during the study period (Figure 1). Of particular interest, mean DMD was higher in Laramie River than North Park in 2017, lower in 2018, and practically equal in 2019. These site differences in measurements of DMD closely track site differences in estimates of ψ in which Laramie River had higher (2017) then lower (2018) then equal (2019) estimates of ψ in comparison to North Park. While credible intervals substantially overlap between study areas, the observed pattern is consistent with the established relationship between nutrition and reproduction in ungulates (Parker et al. 2009) and is supported by studies showing suppressed reproduction in populations with low nutritional availability (Severud et al. 2019). For moose, pregnancy rates have been positively correlated with body condition (Testa and Adams 1998), suggesting prior‐year nutrition intake may influence calving success. To meet the energetic demands of reproduction, most ungulates depend on prior accumulation of energy reserves, although females rely, at least in part, on energy gained during the breeding season (Mysterud et al. 2005). Further, moose display relatively strong site fidelity, especially during summer (Ofstad 2013, Morrison et al. 2021). Moreover, site fidelity is strongest in predictable landscapes, such as our study sites, where vegetative greening occurs at regular intervals (Morrison et al. 2021). This suggests that observed individual‐level nutrition availability may be similar to prior‐year nutrition availability. While prior‐year nutrition intake is a critical component of an animal's current‐year condition, numerous other climatic variables can also have dramatic carryover impacts on condition. For example, delayed impacts of temperature, which may extend for multiple years, and spring‐summer precipitation have been observed to influence recruitment of moose (Monteith et al. 2015). Additionally, snow depth, a measure of winter severity, influences population growth rates over a multiple‐year period (Mech et al. 1987, Post and Stenseth 1998). As a result, such factors confound the effect of prior‐year nutrition and reduce its utility for evaluating calving success. Credible intervals overlapped zero for most covariates in the analysis of DMD relative to willow characteristics (Figure 2), which suggests that the covariates considered in this study are not sufficient to accurately predict DMD. Prior researchers identified differences in nutrition based on willow species (Stolter et al. 2005, Stumph and Wright 2007); however, these differences may be overshadowed by temporal, geographic, or morphological characteristics (Stumph and Wright 2007). If the objectives of future survey efforts are to assess the relative nutrition quality of willow, simply monitoring the covariates used here is not sufficient. That said, these data can still provide valuable information when used to evaluate calving success (i.e., the probability of female moose having a calf‐at‐heel). The general trends in ψ (Figure 3) align closely with Bergman et al. (2020b) who evaluated 4 years of data and did not differentiate between populations. Of note, when compared to previously reported estimates of ψ, the decrease in 2017 can largely be ascribed to a single population (North Park) rather than a decrease in ψ across both populations. �12 | HAYES ET AL. Additionally, ψ appears to be more variable between sites through time than previously thought. Estimates for 2015 to 2017 support observations by CPW biologists that a higher proportion of female moose in Laramie River had calves each year than in North Park. This pattern, however, does not hold for 2018 and 2019 in which Laramie River had lower ψ estimates (2018) and then functionally identical estimates to North Park (2019). Notably, 2017 was the only year in which the mean estimate of ψ for each population fell outside of the 95% credible interval of the other. Thus, the results presented suggest that ψ did not differ significantly between these populations with the exception of the year 2017. The greatest advantage of employing a Bayesian framework in these analyses was the ability to use an integrated model evaluating the effect of each covariate on ψ with appropriate error propagation. In this model, the uncertainty in calf occupancy for each individual moose was carried into the estimate of covariate effects on ψ. Although this results in broader credible intervals for estimates, it provides a more truthful assessment of covariate uncertainty. This uncertainty is readily apparent in the estimate of covariate effects on calf occupancy (Figure 5). While credible intervals of all but 1 estimate ended up overlapping zero, these estimates can still provide biologically relevant trends when there is a high probability of either a positive or negative correlation. The high probability of a positive effect of willow height and the positive effect of willow cover suggest that these covariates result in an increase in ψ. Overall, female moose were more likely to have a calf when they had access to large (i.e., tall), high density (i.e., high ratio of willow cover) willow. Although a substantial body of work has documented moose preference for willow communities and forage (Dorn 1970, Stevens 1970, Dungan and Wright 2005, Renecker and Schwartz 2007), our findings provide evidence that high density, mature willow is even more important for female moose with calves. These findings are consistent with previous research showing selection by female moose for dense habitats providing both food and protection from predation (Dussault et al. 2005). Thermoregulatory demands may also drive habitat selection and contribute to increased use of areas with taller willows (Street et al. 2015). Our results provide the foundation for further investigation of this relationship and draw to light the possibility of incorporating remote‐sensed vegetation surveys (e.g., vegetation height from light detection and ranging [lidar]) to evaluate habitat quality for reproducing female moose. The high degree of uncertainty in the effect of browse intensity on ψ (Figure 5) is likely due to high variance in the raw parameter estimates (Figure 1). The estimated effect of browse intensity trended negative but is inconclusive (P > 0.87). Two possible biological explanations of a negative effect are that female moose are more likely to have a calf either in areas with lower densities of moose or in areas with sufficiently dense willow that it results in lower browse intensity. The first explanation is supported by literature that has documented female moose to calve in areas that are less frequently used by solitary moose (Dussault et al. 2005). Further, some evidence suggests that individual moose exhibit different habitat selection strategies (Poole et al. 2007), which may explain the uncertainty in this estimate. Because of the lack of a clear relationship between browse intensity and DMD (Figure 2) it is unlikely that this effect is being driven by selection for more nutritious plants. Leaf length was the least informative of the covariates considered for calf occupancy, having the widest credible interval (Figure 5). This is likely due to relatively low variation in the parameter between site, year, and individual sample (Figure 1). These results indicate that leaf length is not a useful parameter for explaining moose calving success. The approach we employed, evaluating spatial samples based on the overlap with individual home ranges, proved to be a useful technique to evaluate sparse measures of nutrition and vegetation quality at landscape scales. In using this method, we were able to leverage available spatial use data from individual moose to provide estimates for each individual rather than distilling all analyses to the site level. This approach enables a much more robust analysis of effects than solely looking at effects at the site level. Despite this benefit, limitations in the amount of fine‐resolution data available led to reduction in the precision of estimates. Collecting and analyzing DMD at the scale required to estimate individual nutrition for multiple populations of animals is likely not feasible without spatial averaging. High‐resolution data can be collected through the direct observation of feeding behavior and counting of individual bites (Dungan and Wright 2005), but inference is limited by the number of individuals that can be observed simultaneously. Other researchers focused on creating �MOOSE CALVING SUCCESS | 13 landscape‐level models of digestible energy (Rowland et al. 2018) but typically focused on forecasting habitat use rather than nutrition consumption for individuals. The methodology we employed has the benefit of optimizing nutrition sampling effort by targeting known areas of use. Future studies addressing this question may wish to analyze composite samples consisting of multiple willow plants to reduce sampling variability. At a site level, when mean values of DMD differed, the site with higher DMD had higher estimates of ψ, suggesting that variation in willow DMD may be biologically significant to moose. That said, the predictive utility of this relationship is likely limited because of sampling variability and model uncertainty in ψ. Our findings suggest a comparison of site‐level covariates would only be useful with large differences in ψ between populations or years. Further, given the small observed differences in DMD between study sites, future studies evaluating this relationship may benefit from including analysis of secondary compounds, like tannins, in evaluation of forage nutrition. When comparing populations with similar estimates of ψ (e.g., study populations in 2015, 2016, 2019), the variance in DMD measurements will likely be greater than the difference in occupancy, thereby reducing the utility of the covariate. Annually monitoring wildlife populations can be very expensive but is recognized to be of greater importance for species near critical thresholds in population abundance (Hauser et al. 2006). As southern populations of moose may be demographically more vulnerable than northern populations (Ruprecht et al. 2016), management would benefit from frequent populations assessment. While our results suggest monitoring willow characteristics is not sufficient to detect small‐scale changes in calving success, they offer a coarse assessment that may be more easily conducted than direct observation. If substantive changes are observed in willow habitat availability or nutrition, our results suggest increased effort in population monitoring should be conducted. M A N A G E M E N T I M P L I C A TI O N S Our results provide evidence that an assessment of willow height and cover may serve as a good indicator of spatial areas that support higher levels of calving success. In practice, this may be applied as a coarse measure of habitat suitability when making conservation or land management decisions. Additionally, if managers wish to determine fine‐scale differences or changes in calving success, assessment of willow quality is likely ineffective, and resources are better allocated towards obtaining direct estimates of calving success. A C KN O W L E D G M E N T S We are grateful to S. S. Peterson, A. J. Howell, S. T. Boyle, B. C. Smith, K. L. Yeager and A. R. P. McLean who spent many hours tracking and observing moose. We are indebted to J. R. Runge, K. M. Proffitt, and 2 anonymous reviewers for improving this manuscript through critical review. Financial and logistical support for this research was provided by Colorado Parks and Wildlife. Additional research funding was provided in part by a United States Fish and Wildlife Service Federal Aid Research Grant and by the Philip L. Wright Memorial Research Award. CO NFL I CT OF INTERES T S The authors declare that there are no conflict of interests. ETHICS STATEME NT All animal capture, handling, and monitoring was conducted in accordance with approved Institutional Animal Care and Use Committee (IACUC) protocols (University of Montana IACUC file 032‐17CBWB‐060517 and CPW ACUC 08‐2013). REFERENCES Belovsky, G. E. 1978. Diet optimization in a generalist herbivore: the moose. Theoretical Population Biology 14:105–134. �14 | HAYES ET AL. Bergman, E. J., F. P. Hayes, and K. Aagaard. 2020a. Estimation of moose parturition dates in Colorado: incorporating imperfect detections. Alces 56:127–135. Bergman, E. J., F. P. Hayes, C. J. Bishop, and P. M. Lukacs. 2020b. Moose calf detection probabilities: quantification and evaluation of a ground‐based survey technique. Wildlife Biology 2020:wlb.00599. Bergström, R., and C. Guillet. 2002. Summer browsing by large herbivores in short‐rotation willow plantations. Biomass and Bioenergy 23:27–32. Bishop, C. J., G. C. White, D. J. Freddy, B. E. Watkins, and T. R. Stephenson. 2009. Effect of enhanced nutrition on mule deer population rate of change. Wildlife Monographs 172:1–28. Boertje, R. D., G. G. Frye, and D. D. Young, Jr. 2019. Lifetime, known‐age moose reproduction in a nutritionally stressed population. Journal of Wildlife Management 83:610–626. Burkholder, B. O., N. J. DeCesare, R. A. Garrott, and S. J. Boccadori. 2017. Heterogeneity and power to detect trends in moose browse utilization of willow communities. Alces 53:23–39. Calenge, C. 2006. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling 197:516–519. Cook, R. C., J. G. Cook, D. J. Vales, B. K. Johnson, S. M. McCorquodale, L. A. Shipley, R. A. Riggs, L. L. Irwin, S. L. Murphie, B. L. Murphie, et al. 2013. Regional and seasonal patterns of nutritional condition and reproduction in elk. Wildlife Monographs 184:1–45. Dorn, R. D. 1970. Moose and cattle food habits in southwest Montana. Journal of Wildlife Management 34:559–564. Dungan, J. D., and R. G. Wright. 2005. Summer diet composition of moose in Rocky Mountain National Park, Colorado. Alces 41:139–146. Dussault, C., J.‐P. Ouellet, R. Courtois, J. Huot, L. Breton, and H. Jolicoeur. 2005. Linking moose habitat selection to limiting factors. Ecography 28:619–628. Edwards, R. Y., and R. W. Ritcey. 1958. Reproduction in a moose population. Journal of Wildlife Management 22:261–268. Gelman, A., and D. B. Rubin. 1992. Inference from iterative simulation using multiple sequences. Statistical Science 7: 457–472. Hagerman, A. E., and C. T. Robbins. 1993. Specificity of tannin‐binding salivary proteins relative to diet selection by mammals. Canadian Journal of Zoology 71:628–633. Hauser, C. E., A. R. Pople, and H. P. Possingham. 2006. Should managed populations be monitored every year? Ecological Applications 16:807–819. Heard, D., S. Barry, G. Watts, and K. Child. 1997. Fertility of female moose (Alces alces) in relation to age and body composition. Alces 33:165–176. Juntheikki, M. R. 1996. Comparison of tannin‐binding proteins in saliva of Scandinavian and North American moose (Alces alces). Biochemical Systematics and Ecology 24:595–601. Månsson, J., C. E. Hauser, H. Andrén, and H. P. Possingham. 2011. Survey method choice for wildlife management: the case of moose Alces alces in Sweden. Wildlife Biology 17:176–190. Mech, L. D., R. E. McRoberts, R. O. Peterson, and R. E. Page. 1987. Relationship of deer and moose populations to previous winters' snow. Journal of Animal Ecology 56:615–627. Monteith, K. L., R. W. Klaver, K. R. Hersey, A. A. Holland, T. P. Thomas, and M. J. Kauffman. 2015. Effects of climate and plant phenology on recruitment of moose at the southern extent of their range. Oecologia 178:1137–1148. Morrison, T. A., J. A. Merkle, J. G. C. Hopcraft, E. O. Aikens, J. L. Beck, R. B. Boone, A. B. Courtemanch, S. P. Dwinnell, W. S. Fairbanks, B. Griffith, et al. 2021. Drivers of site fidelity in ungulates. Journal of Animal Ecology 90:955–966. Mysterud, A., E. J. Solberg, and N. G. Yoccoz. 2005. Ageing and reproductive effort in male moose under variable levels of intrasexual competition. Journal of Animal Ecology 74:742–754. Ofstad, E. 2013. Seasonal variation in site fidelity of moose (Alces alces). Thesis, Norwegian University, Trondheim, Norway. Parker, K. L., P. S. Barboza, and M. P. Gillingham. 2009. Nutrition integrates environmental responses of ungulates. Functional Ecology 23:57–69. Poole, K. G., R. Serrouya, and K. Stuart‐Smith. 2007. Moose calving strategies in interior montane ecosystems. Journal of Mammalogy 88:139–150. Post, E., and N. C. Stenseth. 1998. Large‐scale climatic fluctuation and population dynamics of moose and white‐tailed deer. Journal of Animal Ecology 67:537–543. Proffitt, K. M., M. Hebblewhite, W. Peters, N. Hupp, and J. Shamhart. 2016. Linking landscape‐scale differences in forage to ungulate nutritional ecology. Ecological Applications 26:2156–2174. R Core Team. 2020. R: a language and environment for statistical computing. Version 3.6.0. R Foundation for Statistical Computing, Vienna, Austria. Renecker, L. A., and R. J. Hudson. 1989. Ecological metabolism of moose in aspen‐dominated boreal forests, central Alberta. Canadian Journal of Zoology 67:1923–1928. �MOOSE CALVING SUCCESS | 15 Renecker, L. A., and C. C. Schwartz. 2007. Food habits and feeding behavior. Pages 403–439 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and management of the North American moose. Second edition. University Press of Colorado, Boulder, USA. Robbins, C. T., T. A. Hanley, A. E. Hagerman, O. Hjeljord, D. L. Baker, C. C. Schwartz, and W. W. Mautz. 1987a. Role of tannins in defending plants against ruminants: reduction in protein availability. Ecology 68:98–107. Robbins, C. T., S. Mole, A. E. Hagerman, and T. A. Hanley. 1987b. Role of tannins in defending plants against ruminants: reduction in dry matter digestion? Ecology 68:1606–1615. Rowland, M. M., M. J. Wisdom, R. M. Nielson, J. G. Cook, R. C. Cook, B. K. Johnson, P. K. Coe, J. M. Hafer, B. J. Naylor, D. J. Vales, et al. 2018. Modeling elk nutrition and habitat use in western Oregon and Washington. Wildlife Monographs 199:1–69. Ruprecht, J. S., K. R. Hersey, K. Hafen, K. L. Monteith, N. J. DeCesare, M. J. Kauffman, and D. R. MacNulty. 2016. Reproduction in moose at their southern range limit. Journal of Mammalogy 97:1355–1365. Schwartz, C. C., W. L. Regelin, and A. M. Franzmann. 1987. Protein digestion in moose. Journal of Wildlife Management 51: 352–357. Severud, W. J., T. R. Obermoller, G. D. DelGiudice, and J. R. Fieberg. 2019. Survival and cause‐specific mortality of moose calves in northeastern Minnesota. Journal of Wildlife Management 83:1131–1142. Shallow, J. R. T., M. A. Hurley, K. L. Monteith, and R. T. Bowyer. 2015. Cascading effects of habitat on maternal condition and life‐history characteristics of neonatal mule deer. Journal of Mammalogy 96:194–205. Simpson, R., T. Curdts, J. Deak, S. Campbell, J. Carochi, and A. Cade. 2013. Colorado Parks and Wildlife ‐ Basinwide Layer Package. <https://www.arcgis.com/home/item.html?id=893739745fcd4e05af8168b7448cda0c>. Accessed 1 Feb 2020 Spaeth, D. F., R. T. Bowyer, T. R. Stephenson, P. S. Barboza, and V. Van Ballenberghe. 2002. Nutritional quality of willows for moose: effects of twig age and diameter. Alces 38:143–154. Stevens, D. R. 1970. Winter ecology of moose in the Gallatin Mountains, Montana. Journal of Wildlife Management 34: 37–46. Stickney, P. F. 1966. Browse utilization based on percentage of twig numbers browsed. Journal of Wildlife Management 30: 204–206. Stolter, C., J. P. Ball, R. Julkunen‐Tiitto, R. Lieberei, and J. U. Ganzhorn. 2005. Winter browsing of moose on two different willow species: food selection in relation to plant chemistry and plant response. Canadian Journal of Zoology 83: 807–819. Street, G. M., A. R. Rodgers, and J. M. Fryxell. 2015. Mid‐day temperature variation influences seasonal habitat selection by moose. Journal of Wildlife Management 79:505–512. Stumph, B. P., and R. G. Wright. 2007. Effects of willow quality on moose distribution in a montane environment. Alces 43: 129–142. Testa, J. W., and G. P. Adams. 1998. Body condition and adjustments to reproductive effort in female moose (Alces alces). Journal of Mammalogy 79:1345–1354. Timmermann, H. R., and A. R. Rodgers. 2017. The status and management of moose in North America‐circa 2015. Alces 53: 1–22. United States Department of Agriculture National Resources Conservation Service. 2020. Colorado SNOTEL sites. <https://www.wcc.nrcs.usda.gov/snow/>. Accessed 1 Feb 2020. Van Ballenberghe, V., and W. B. Ballard. 2007. Food habits and feeding behavior. Pages 223–245 in A. W. Franzmann and C. C. Schwartz, editors. Ecology and management of the North American moose. Second edition. University Press of Colorado, Boulder, USA. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46:S120–S139. Associate Editor: Kelly Proffitt. How to cite this article: Hayes, F. P., J. J. Millspaugh, E. J. Bergman, R. M. Callaway, and C. J. Bishop. 2022. Effects of willow nutrition and morphology on calving success of moose. Journal of Wildlife Management 1–15. https://doi.org/10.1002/jwmg.22175 � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Journal Articles Description An account of the resource CPW peer-reviewed journal publications Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Effects of willow nutrition and morphology on calving success of moose Description An account of the resource <span>Across much of North America, populations of moose (</span><i>Alces alces</i><span>) are declining because of disease, predation, climate change, and anthropogenic-driven habitat loss. Contrary to this trend, populations of moose in Colorado, USA, have continued to grow. Studying successful (i.e., persistent or growing) populations of moose can facilitate continued conservation by identifying habitat features critical to persistence of moose. We hypothesized that moose using habitat with higher quality willow (</span><i>Salix</i><span> spp.) would have a higher probability of having a calf-at-heel (i.e., calving success). We evaluated moose calving success using repeated ground observations of collared individuals with calves in an occupancy model framework to account for detection probability. We then evaluated the impact of willow habitat quality and nutrition on moose calving success by studying 2 spatially segregated populations of moose in Colorado. Last, we evaluated correlations between willow characteristics (browse intensity, height, cover, leaf length, and species) and willow nutrition (dry matter digestibility [DMD]) to assess the utility of using those characteristics to assess willow nutrition. We found willow height and cover had a high probability of being positively associated with higher individual-level calving success. Willow DMD, browse intensity, and leaf length were not predictive of individual moose calving success; however, the site with higher mean DMD consistently had higher mean estimates of calving success for the same year. Our results suggest surveying DMD is likely not a useful metric for assessing differences in calving success of individual moose but may be of use at population levels. Further, the assessment of willow morphology and density may be used to identify areas that support higher levels of moose calving success.</span> Bibliographic Citation A bibliographic reference for the resource. Recommended practice is to include sufficient bibliographic detail to identify the resource as unambiguously as possible. <a href="https://doi.org/10.1002/jwmg.22175" target="_blank" rel="noreferrer noopener">https://doi.org/10.1002/jwmg.22175</a> Creator An entity primarily responsible for making the resource Hayes, Forest P. Millspaugh, Joshua J. Bergman, Eric J. Callaway, Ragan M. Bishop, Chad J. Subject The topic of the resource <em>Alces alces</em> Calving success Colorado Moose, nutrition Reproduction <em>Salix</em> Willow Extent The size or duration of the resource. 15 pages Date Created Date of creation of the resource. 2022-01-07 Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/InC-NC/1.0/" target="_blank" rel="noreferrer noopener">In Copyright - Non-Commercial Use Permitted</a> Format The file format, physical medium, or dimensions of the resource application/pdf Language A language of the resource English Is Part Of A related resource in which the described resource is physically or logically included. The journal of wildlife management Type The nature or genre of the resource Article