Environmental correlates of activity and energetics in a wide-ranging social carnivore

Background: Environmental conditions can influence animal movements, determining when and how much animals move. Yet few studies have quantified how abiotic environmental factors (e.g., ambient temperature, snow depth, precipitation) may affect the activity patterns and metabolic demands of wide-ranging large predators. We demonstrate the utility of accelerometers in combination with more traditional GPS telemetry to measure energy expenditure, ranging patterns, and movement ecology of 5 gray wolves (Canis lupus), a wide-ranging social carnivore, from spring through autumn 2015 in interior Alaska, USA. Results: Wolves exhibited substantial variability in home range size (range 500–8300 k m ) that was not correlated −1 with daily energy expenditure. Mean daily energy expenditure and travel distance were 22 MJ and 18 km day , respectively. Wolves spent 20% and 17% more energy during the summer pup rearing and autumn recruitment seasons than the spring breeding season, respectively, regardless of pack reproductive status. Wolves were predomi- nantly crepuscular but during the night spent 2.4 × more time engaged in high energy activities (such as running) during the pup rearing season than the breeding season. Conclusion: Integrating accelerometry with GPS telemetry can reveal detailed insights into the activity and energet- ics of wide-ranging predators. Heavy precipitation, deep snow, and high ambient temperatures each reduced wolf mobility, suggesting that abiotic conditions can impact wolf movement decisions. Identifying such patterns is an important step toward evaluating the influence of environmental factors on the space use and energy allocation in carnivores with ecosystem-wide cascading effects, particularly under changing climatic conditions. Keywords: Alaska, Behavior, Canis lupus, Carnivore, Ecology, Energetics, Movement competition, predators, reproductive demands, and abi- Background otic factors. As the currency of ecosystem function, ener- Wildlife movement decisions while foraging are driven getic demand influences the behavioral decisions animals by a dynamic balance between maximizing energy make, dictating, where and how often they feed [1–4]. intake and minimizing costs. In addition to foraging and Mammalian apex carnivores in particular experience prey availability, wildlife movements are influenced by intrinsically elevated energy demands associated with large body size [5], endothermy [6] and carnivory [7, 8]. *Correspondence: calebmbryce@gmail.com To replenish the energy expended on vital functions (e.g., Department of Ecology and Evolutionary Biology, University metabolic work and activity, thermoregulation, growth, of California-Santa Cruz, Coastal Biology Building, 130 McAllister Way, reproduction, repair, waste; [9–11]), predators must Santa Cruz, CA 95060, USA Full list of author information is available at the end of the article locate, capture, and kill mobile prey. Hunting itself is an © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Bryce et al. Animal Biotelemetry (2022) 10:1 Page 2 of 16 energetically demanding activity with potential cascading on free-ranging male wolves. By calibrating these ACC- impacts across the ecological community [12]. Hunting GPS data on captive wolves and utilizing published esti- decisions of top predators and movement pathways may mates of wolf movement energetics (i.e., measures of trigger both density and behaviorally mediated trophic oxygen consumption [59, 60]), we compared daily energy cascades by directly decreasing prey populations and expenditure (DEE) estimated via accelerometry with DEE indirectly impacting the behavior of herbivores [13–16] derived from relationships between oxygen consump- and sympatric mesopredators [17–21]. Given the global tion and animal speed (determined using GPS telemetry decline in many top predator species [22–24], the quan- location data). We accounted for the potential effects of tification of free-ranging behaviors and resulting energy topography on both measures of DEE by measuring the demands is integral for defining resource requirements slope angle at which wolves travelled [61]. In addition, we −1 and understanding how movement patterns of these spe- estimated the movement rates (m  h ) and home range cies might propagate through the food web [25–29].size (km ) of these individuals to evaluate whether they As cursorial predators, gray wolves (Canis lupus) served as reliable proxies for energy expenditure and to expend immense energy resources in finding, pursuing, make ecological inferences about the movement patterns and capturing prey [30]. To obtain prey and maintain ter- of their packs. ritories, wolves roam widely on a daily basis [31], often We tested whether abiotic environmental variables, utilizing natural and anthropogenic linear travel cor- including ambient temperature, snow depth, and pre- ridors where available [32–35]. In some cases, wolves cipitation affected the movement rates of wolves and have been observed chasing prey for over 20 km [36] and their energy expenditure at hourly, daily, and seasonal covering nearly 80  km in 12  h [37]. Thus, in optimizing temporal scales. We defined seasons based on the known movement and hunting success, wolves are affected by breeding cycle of wolves in interior Alaska: breeding both abiotic (e.g., temperature, snow depth, precipita- (March–April), pup-rearing (May–July), and pup recruit- tion) and biotic (e.g., prey movement and vulnerability, ment (August–October). Given the wide-ranging move- proximity of rival packs) factors [38–42]. Given that cli- ments of this apex predator [8, 58, 62], we expected that mate change is rapidly warming northern latitudes and wolves would reduce movement rates during physiologi- impacting not only ambient temperature [e.g., 41] but cally suboptimal conditions (e.g., being active in ambient also the timing, type and location of precipitation [44, temperatures beyond the species’ thermoneutral zone), 45], it is especially important to understand how wolves with analogous DEE levels. We predicted that deeper currently respond to these variables. snow and higher temperatures, but not greater precipi- While numerous studies have estimated wolf energy tation, would reduce wolf movement rates and DEE. We intake (energy gain via consumption of prey; [46–48]), further examined whether these relationships varied few have quantified energy expenditure of free-rang - seasonally. Finally, we discuss ecological insights gained ing wolves, particularly at hourly scales across multiple by our efforts to quantify activity patterns and energy seasons. Continuous metabolic demands of free-rang- expenditure of these predators. ing animals are inherently difficult to estimate [49], but sophisticated biologgers can now provide detailed Results insights into how wolves adjust their movements and Behavior calibrations from nine captive wolves (Fig.  1) energy expenditure in response to environmental and resulted in clearly defined ODBA threshold values for seasonal factors. Daily energy requirements may be each behavior (Additional file  4: Fig. S1). From this, particularly high for breeders (reproductive adults that we defined five ODBA behavioral categories to use for are socially dominant given their size, behavior, and wolves in the wild as: < 0.1  g (resting), 0.1 < 0.25  g (sta- interactions with pack mates & rival packs [50–52]). As tionary), 0.25 < 0.75 g (walking), 0.75 < 1 g (highly active), pack leaders, breeders often assume energy-demanding and > = 1  g (running). We calculated the proportion of functions such as initiating prey attacks and breaking time each wolf spent conducting these behaviors. trail through high vegetation or deep snow [50, 53, 54]. Data were collected from four adult (ages 2–3  years) Despite the crucial role that dominant wolves play in male wolves (body mass 45.9 ± 1.4 kg) in Denali National pack persistence [55–57], free-ranging activity patterns Park and Preserve (DNPP, Alaska, USA) via ACC-GPS and associated energy budgets for these animals remain collars from March through October 2015 (208–211 days −1 poorly understood [58]. wolf ; Table  1). Data were additionally collected from Here, we describe an 8-month (March to October) one male wolf (age 2  years, body mass 45  kg) that was analysis of wolf movement in interior Alaska using infor- monitored from March until it was killed adjacent to mation collected by combined tri-axial accelerometer- DNPP in early May (50  days); our results, therefore, GPS radiocollars (hereafter ACC-GPS collars) deployed describe a total of 887 wolf-days. Collared individuals Br yce et al. Animal Biotelemetry (2022) 10:1 Page 3 of 16 Fig. 1 Wolf accelerometer-GPS collar calibration, showing A axis orientation, B a 4-min raw data sample depicting how distinct behaviors generate unique collar accelerometer signatures, and C associated overall dynamic body acceleration (ODBA) values were dominant wolves (known or suspected breeders) in Fig.  3) and was significantly lower during the breeding packs ranging in size from 2 to 14 individuals (5.4 ± 2.2 season than in the pup rearing season (EMMs p < 0.05) wolves/pack). Over the deployment period, home ranges but not the recruitment season (EMMs P = 0.31). There (95% utilization distributions (UD)) ranged from 510 to was no difference between the pup rearing and recruit - 8258  km with the largest UD used by the wolf in the ment seasons (EMMs P = 0.96). DEE calculated using western portion of the study area (Table 1, Fig. 2A). Ter- ODBA varied among individuals (F = 76.89, df = 4, rain heterogeneity is high in DNPP, so we measured the p < 0.001, R = 0.26; Table 1). elevations utilized (Fig. 2B) and slope angles of wolf paths When calculated using continuous time correlated (Additional file  5: Fig. S2) to account for the additional random walk (CTCRW) model-derived speed from the energy demands of navigating through mountainous GPS location data (see Methods), the mean DEE was −1 −1 terrain (see Methods). The median slope angle travelled 15.8 ± 0.1 MJ  day (range: 9.5–31.0  MJ  day ). This by all wolves was level (− 0.07° ± 0.04; Additional file  5: DEE estimate differed significantly from the mean DEE Fig. S2). The steepest uphill slope the wolves selected estimated from ODBA (Wilcoxon sign ranked test; was 35.6° and the steepest downhill slope observed was V = 2093, P < 0.001). Mean CTCRW-derived DEE was − 53.5°. Wolves used a mean slope angle of 1.9° ± 0.1, and not significantly correlated with home range size (95% slope angle varied among individuals (F = 384.98, df = 4, UDs, r = − 0.28, P = 0.75, n = 5) or daily distance trave- 2 2 p < 0.001, R = 0.29). led (r = 0.65, P = 0.06, n = 5; Table  1) based on linear regression. On average, the CTCRW movement-derived −1 Daily energy expenditure DEE was 6  MJ  day (95% CI 5.7–6.3) less than the −1 Mean (± SE) wolf DEE was 21.8 ± 0.2 MJ  day (range: DEE estimated from ODBA. The CTCRW movement- −1 9.9–50.3 MJ  day ; Table  1) when estimated from over- derived DEE varied among individuals (F = 141.98, df = 4, all dynamic body acceleration (ODBA). Mean DEE esti- p < 0.001, R = 0.39; Table 1). CTCRW movement-derived mated from ODBA was not significantly correlated with DEE also varied among seasons (χ = 19.08, df = 2, 2 2 home range size (95% UDs, r = − 0.26, p = 0.70, n = 5), p < 0.001, R = 0.74, Fig.  3B) and was marginally lower but was significantly correlated with daily distance trave - during the breeding season compared to the pup rear- led (r = 0.83, P = 0.02, n = 5; Table  1) based on linear ing (EMMs p = 0.076) but there was no statistical differ - regression. We tested the significance of linear mixed ence when compared to the recruitment season (EMMs effects models using chi-square tests and conditional R p = 0.99) or between the pup rearing and recruitment and we calculated estimated marginal means (EMMs) to season (EMMs p = 0.56). test for pairwise differences between seasons. DEE var - 2 2 ied among seasons (χ = 57.18, df = 2, p < 0.001, R = 0.56; Bryce et al. Animal Biotelemetry (2022) 10:1 Page 4 of 16 Table 1 Summary table for the five gray wolves monitored in and around Denali National Park and Preserve, Alaska Wolf ID Age (yrs) Weight Deployment Reproductive Pack size (start, Mean 95% UD 95% CI UD CTCRW daily Slope Daily energy expenditure 2 2 (kg; start, (days) status end) (km ) (km ) distance Angle during CTCRW DEE ODBA DEE end) travelled (km) incline −1 −1 (MJ day )(MJ day ) locomotion (°) 1502M 3 47, 48 212 Denned 2, 8 510 434–591 17.5 ± 0.6 0.79 15.3 23.0 ± 0.07 ± 0.1 ± 0.3 1501M 3 51, 52 211 Denned 4, 6 3983 2837–5321 22.4 ± 1.0 3.86 19.5 25.1 ± 0.14 ± 0.3 ± 0.4 1507M 2 45, N/A 50 N/A 14, 4 4107 1282–8547 13.5 ± 1.3 1.52 14.0 17.4 ± 0.13 ± 0.3 ± 0.4 1503M 3 41, 45 210 Denned 2, 6 2067 1482–274 14.5 ± 0.8 2.36 13.2 16.8 ± 0.09, ± 0.2 ± 0.3 1506M 2–3 46, 46 209 Did not den 5, 2 8258 4530–13,087 21.2 ± 0.9 1.02 15.6 23.3 ± 0.05 ± 0.2 ± 0.5 Mean ± SE 2–3 46, 47 178.0 ± 26.2 3 of 5 denned 5.4 ± 2.2, 3785 ± 1301 17.8 ± 1.8 1.9 15.8 21.8 5.2 ± 1.0 ± 0.1 ± 0.1 ± 0.2 Pack size includes any 2015 pups surviving to the end of the study period (early October). Where applicable, data are presented as mean ± SE UD: utilization distribution; CTCRW: continuous time-correlated random walk; DEE: daily energy expenditure; ODBA: overall dynamic body acceleration Br yce et al. Animal Biotelemetry (2022) 10:1 Page 5 of 16 Fig. 2 Study area depicting (A) accelerometer-GPS instrumented male wolf (n = 5) hourly relocations (colored points), core area (50% autocorrelated kernel density estimate (AKDE) utilization distribution; thick, inner colored contour) and home range (95% AKDE utilization distribution; thin, outer colored contour) in Denali National Park and Preserve, Alaska from March to October 2015. Triangles correspond to capture locations; squares depict den sites (n = 3). B The frequency at which each wolf used the elevations within their range. Colors correspond with the map colors for wolf ID. Minimum elevation: 141.14 m (ID: 1502 M), maximum elevation: 1848.81 m (ID: 1501 M) The ODBA and CTCRW movement-derived DEE were (EMMs p < 0.001 and E MMSs p = 0.084, respectively). correlated (r = 0.71, p < 0.001, n = 869) and increased lin- In contrast, movement rates did not vary seasonally. −1 −1 early by the equation: Movement rates were 644 ± 15 m  h , 876 ± 15 m  h , −1 and 764 ± 16 m  h for the breeding, pup rearing, and CTCRW movement − derived DEE recruitment seasons, respectively (EMMs p = 0.007– (1) = 5.43 + 0.48 ∗ ODBA DEE 0.99). Hourly movement rates and hourly ODBA were positively correlated (R = 0.40, p < 0.001). as shown in Additional file 6: Fig. S3. Environmental factors affecting wolf activity and movement rate Seasonal effects on wolf activity and movement Temperature Averaged across seasons, wolves travelled −1 −1 An interaction between ambient temperature and sea- 17.8 km  day (± 1.8 km  day ). During the breeding −1 son affected wolf mean hourly ODBA (χ = 350.84, season, wolves travelled on average 16.7 ± 0.9 km  day −1 df = 2, p < 0.001, R = 0.51) and hourly movement rate (range: 14.7–19.6 km  day ). In the pup rearing season, 2 2 −1 (χ = 293.98, df = 2, p < 0.001, R = 0.64). Mean hourly they travelled on average 21.0 ± 0.7 km  day (range: −1 ODBA increased marginally with increasing tempera- 9.4–23.8 km  day ), and during the recruitment sea- −1 ture in the breeding season (slope (β) = 0.006, t = 4.6, son wolves travelled on average 18.3 ± 0.7 km  day −1 p < 0.001) but decreased with increasing temperature (range: 7.9–23.9 km  day ). In doing so, wolves main- during pup rearing (β = − 0.04, t = − 18.6, p < 0.001) and tained expansive but variable home ranges (mean: −1 recruitment (β = − 0.02, t = − 7.4, p < 0.001; Fig.  5A). 3785 ± 1300  km , Table  1) with considerably smaller −1 Wolf hourly movement rate increased marginally with core areas of use (50% UDs, mean: 875 ± 301  km ; increasing temperatures during the breeding season Fig. 2A). (β = 4.5, t = 2.4, p = 0.02) but decreased with increasing We found a seasonal effect on wolf activity (mean temperatures during pup rearing (β = − 53.6, t = − 17.1, hourly ODBA; Fig.  4A). Mean hourly ODBA varied 2 2 p < 0.001) and recruitment (β = − 22.6, t = 3.9, p < 0.001) with season (χ = 226.41, df = 2, p < 0.001, R = 0.77) (Fig. 5B). and was 38% and 27% higher during the pup rearing and recruitment seasons than the breeding season Bryce et al. Animal Biotelemetry (2022) 10:1 Page 6 of 16 movements with increased snow depth during the breed- ing season (β = − 0.17, t = − 5.7, p < 0.001). Movements did not appear to be affected by snow depth in the pup rearing and recruitment seasons as little to no snow was present (Fig. 5D). Precipitation—rain and snowfall We found an interactive effect between hourly precipi - tation and season on the mean hourly ODBA of wolves 2 2 (χ = 6.45, df = 2, p = 0.04, R = 0.74). High levels of pre- cipitation reduced ODBA during the pup rearing season (β = − 0.05, t = − 2.5, p = 0.01), whereas ODBA was not significantly affected by hourly precipitation in the breed - ing or recruitment seasons (when precipitation was less; Fig.  5E). In contrast, hourly movement rates were unaf- fected by hourly precipitation (χ = 1.06, df = 1, p = 0.30, R = 0.60; Fig. 5F), and there was no interaction between season and precipitation (χ = 2.86, df = 2, p = 0.24, R = 0.60). Daily patterns in activity Mean hourly ODBA varied with hour of the day (0–23) in all seasons, and diel activity patterns also differed among seasons (Fig. 4b; see Additional file  2: Table S1 for GAMM results). Irrespective of season, the wolves exhibited cre- puscular activity patterns in both mean hourly ODBA and movement rates. On average, they moved at higher −1 rates at dusk (defined as 1  h before to 1  h after sunset, Fig. 3 Daily energy expenditure (DEE, MJ day ) of male wolves −1 mean: 1072 ± 33 m  h , breeding: 941 ± 58; pup rearing: (n = 5) in Denali National Park and Preserve, Alaska across 3 wolf biological seasons (breeding, pup-rearing, and recruitment), 1208 ± 55; recruitment: 1009 ± 60) and dawn (defined as −1 calculated from A Eqn. S1 using overall dynamic body acceleration 1  h before to 1  h after sunrise, mean: 1073 ± 37 m  h , (ODBA) derived from tri-axial accelerometers and B Eqn. S4 using breeding: 810 ± 61; pup rearing: 1371 ± 62; recruitment: speed derived from hourly continuous time-correlated random walk 909 ± 66). Wolves moved at lower rates during the day (CTCRW ) derived coordinates. Within each box, horizontal black −1 −1 (673 ± 11 m  h : breeding: 606 ± 20 m  h ; pup rear- lines denote median values; boxes extend from the 25th to the 75th −1 −1 percentile of each group’s distribution of values; vertical extending ing: 733 ± 17 m  h ; recruitment: 619 ± 21 m  h ) and at −1 lines denote adjacent values within 1.5 interquartile range of the 25th night (831 ± 18 m  h : breeding: 586 ± 25; pup rearing: and 75th percentile of each group 1174 ± 49; recruitment: 890 ± 29). Wolves were least active during the day in all sea- sons, predominantly crepuscular during the breeding Snow depth season, and most active at night during the pup rearing Snow depth also affected wolves’ mean hourly ODBA and recruitment seasons. These differences in the time with an interaction with season (χ = 23.31, df = 2, of day the wolves were active each season matches the p < 0.001, R = 0.53). Wolf ODBA was reduced with proportion of time the wolves spent with high or low increasing snow depth during the breeding season (when ODBA in each hour (Fig.  4). Fine-scale measurements the snow was deepest during our study) (β = − 0.008, of movements from the ACC show that the wolves spent t = − 6.6, p < 0.001). In the pup rearing and recruitment the majority of each hour resting in the breeding sea- seasons when the snow depth did not exceed 28 cm, wolf son (68.1% ± 0.4, totaling 16  h and 20  min of the day) ODBA was not affected by snow depth (Fig.  5C). Simi- and only 5.7% ± 0.1 (1  h and 22  min each day) running larly, there was an interaction between snow depth and (ODBA > 1  g; Additional file  7: Fig. S4). In contrast, dur- season affecting wolf hourly movement rate (χ = 32.87, ing the pup rearing and recruitment seasons, wolves df = 2, p < 0.001, R = 0.65), where the wolves reduced spent over two hours running each day (9.3% ± 0.2 and Br yce et al. Animal Biotelemetry (2022) 10:1 Page 7 of 16 Fig. 4 Daily activity patterns of male wolves (n = 5) in Denali National Park and Preserve, Alaska in three biological seasons: breeding, pup-rearing, and recruitment. A GAM smoothing of the distance moved between successive 1-h continuous time-correlated random walk (CTCRW ) derived locations (blue line, m), and the mean hourly overall dynamic body acceleration (ODBA, green line, g), as a function of hour of day (both with standard error shading). Day (white) and night (shaded) are indicated based on the average sunrise and sunset times for each season during collaring. Note the separate axis on the right for ODBA. B The proportion of each hour of the day when the ODBA (g) was within specific levels (see ODBA/Behavior scale) for each of the observed seasons. High ODBA are in yellow colors, low ODBA are in dark blues unaffected by all but the heaviest precipitation (Fig.  5). 8.4% ± 0.2, respectively) while spending 63% of the day Regardless of whether the pack was reproductively suc- resting (Table 2). cessful, collared wolves were more active in the pup-rear- During the night, wolves spent 2.4 × more time run- ing and recruitment seasons (Fig.  3) than in the spring ning during the pup rearing season than the breeding breeding season. season. During the pup rearing season, wolves spent We found wolves exhibited varying responses in activ- 1.6 × more time running during the night than in the day. ity due to ambient temperature. During the breeding sea- The distribution of high ODBA activities between night son, which was the coldest season of our study (mean: and day was more consistent during the breeding and − 3.4  °C ± 0.1, range: − 35.4–11.8  °C), activity rates mar- recruitment seasons, but wolves were consistently less ginally increased with temperature. During the pup rear- active in the breeding season compared to other seasons ing season, which was the warmest season of our study (Fig. 3, Table 2). (mean: 11.5  °C ± 0.1,  range: − 1.1–29.9  °C), activity rates Similar to ODBA, wolf hourly movement rate var- decreased with increasing temperatures. Similarly, dur- ied with hour of the day in all seasons (Additional file  3: ing the recruitment season (mean: 5.7  °C ± 0.1,  range: Table  S2) and wolves moved the shortest distances dur- − 9.7–22.8  °C) activity rates decreased with increasing ing the day. Movement patterns were predominantly cre- temperatures. Based on these findings, high ambient tem - puscular in the breeding season and nocturnal in the pup peratures appeared to have the strongest impacts on activ- rearing and recruitment seasons (Fig. 4a). ity rates. These results are similar to other cursorial canids including dingoes (Canis dingo) [63] and African wild dogs Discussion (Lycaon pictus) [64] that exhibited declines in activity rates We quantified the movement ecology of wolves equipped with increasing ambient temperatures. Wolves are cold- with ACC-GPS collars to estimate DEE and infer how adapted [31, 65, 66] but have higher maintenance costs several environmental factors (temperature, snow depth, (i.e., elevated basal metabolic rates) associated with large precipitation) and topography affect the behavior of organ masses to thermoregulate in the cold [67–69], which these wide-ranging carnivores in non-winter condi- would not be accounted for in either of our measures of tions. Wolves were primarily crepuscular (Fig.  4), were DEE [70]. The hottest observed temperatures occurred less active in high ambient temperatures, and largely during the day in the pup rearing season, and while this Bryce et al. Animal Biotelemetry (2022) 10:1 Page 8 of 16 −1 Fig. 5 Scatter plots of mean hourly overall dynamic body acceleration (ODBA (g), left panels) and hourly movement rate (m h , right panels) for male wolves (n = 5) in Denali National Park and Preserve, Alaska as a function of ambient temperature (°C, A, B; dashed line denotes 0 °C), snow depth (cm, C, D), and precipitation (cm, E, F). Colors correspond to wolf biological seasons and shading encompass 95% of the data Br yce et al. Animal Biotelemetry (2022) 10:1 Page 9 of 16 Table 2 Seasonal % of hour running and resting averaged across linked. Heterogeneity in the external environment 5 adult male wolves in Denali National Park and Preserve, Alaska (including slope, vegetation, substrate type) influences animal movement costs [86–88], and in turn these Season Running (% of hour) Resting (% of hour) movement costs impact how animals move through and Day Night Whole day Whole day interact with their environment [89–91]. Some DNPP wolf home ranges encompass mountainous terrain in Breeding 5.84 ± 0.16 5.42 ± 0.19 5.66 ± 0.12 68.06 ± 0.44 the Alaska Range and underscore the impact of the sur- Pup rearing 8.29 ± 0.16 13.22 ± 0.35 9.31 ± 0.15 61.73 ± 0.39 rounding environment on modulating transit costs. For Recruitment 7.37 ± 0.21 9.91 ± 0.27 8.42 ± 0.17 63.16 ± 0.45 example, wolf 1501  M routinely traversed high alpine For wolves, running corresponds to overall dynamic body acceleration passes (> 2000 m), while traveling between dens located (ODBA) > 1 g, while resting corresponds to ODBA < 0.1 g (see Additional file 5: on both sides of the range crest (Fig.  2). He conse- Fig. S2; mean ± SE) quently traversed the steepest average slopes of all the packs at 3.9° (compared to 0.8–2.4° for other packs) and averaged the farthest movements (Table 1). As a result, was the most active season overall, wolves were most wolf 1501  M had the highest associated DEE. Using mobile at dusk and dawn rather than during the heat of the an approach we established with pumas (Puma con- day. Similarly, moose (Alces alces) and caribou (Rangifer color) [61], our DEE analysis explicitly incorporates the tarandus), wolves’ primary prey in DNPP, are also heat- additional metabolic cost associated with locomotion sensitive [72–76]. Behavioral plasticity may be key for mit- up a slope in wolves traversing mountainous terrain. igating adverse effects of increasing diurnal temperatures Topographic slope has been shown to strongly influ - in wolves and other wide-ranging predators [77, 78]. ence large carnivore habitat selection and movement Snow was only present during the first and last few patterns at fine temporal scales [91– 93]. Optimizing weeks of our study (absent from May 5 until Sept. use of energy stores (e.g., via least-cost route selection 15; [79]). The presence and depth of snow can influ - [94]) may be critical for these species, which experi- ence both prey movements and their habitat selection ence higher absolute and relative net transport costs for [80, 81] as well as the energetic costs of wolf move- uphill locomotion and less downhill ‘reimbursement’ ment [82]. Snow can be an impediment to movement than lighter animals [95, 96]. in heavier-bodied herbivore prey due to higher foot As with other animals, wolf movement ecology is loading [39, 83], yet wolves were also less active and driven by seasonally variable internal and external fac- had lower movements during periods of deeper spring tors including hunger, fear, and habitat [66, 97, 98]. Our snow in the study area (Fig.  5C, D). Wolves selectively study’s aim was to quantify wolf behavioral and ener- travel through shallow, supportive snow in natural con- getic responses to environmental conditions, but nearby ditions, although their ability to behaviorally mitigate ungulate prey also respond to these same conditions the energetic costs imposed by snow are limited [84]. simultaneously. Given that measuring prey activity was No significant multi-decadal trend in annual snowfall beyond the scope of the study, it is difficult to differenti - at the park headquarters exists, but the winter snowfall ate whether wolves were responding dynamically to the leading into our study (i.e., 2014–2015) was lighter than activities of their prey, or directly to the environment. normal [85] and our spring results may, therefore, be The size of the study area and the associated heterogene - representative of the increasingly mild conditions pre- ity in local conditions also constrain our scope of infer- dicted with arctic climate change. ence. We sourced hourly temperature, precipitation, and Wolf activity, but not distance travelled, was affected snow depth from a single central weather station, but by an hourly precipitation–season interaction. Winter these parameters are inherently variable in mountainous snowfall has been shown to temporarily reduce wolf terrain, and data from one central site does not necessar- activity as it is thought to dampen hunting success [82]. ily reflect true conditions at the location of the wolves. As spring progressed into summer pup-rearing, wolves In addition, we collected data exclusively from dominant were less active during the instances of heavy rain that adult males within a pack (“breeders”) who likely exhibit occurred. These inverse effects across seasons suggest higher movement-related energy demands relative to that all but the most extreme precipitation is unlikely to other members of the pack [50, 53, 54]. significantly affect wolf movements. Across individuals and seasons, the average wolf mass- For highly mobile animals such as wolves, habitat −1 −1 specific DEE (454 kJ kg d ; Table 1) is comparable to a structure and metabolic transport costs are inextricably Bryce et al. Animal Biotelemetry (2022) 10:1 Page 10 of 16 relied on hourly GPS fixes to prolong collar battery life. previously reported wolf field metabolic rate (FMR, 474 kJ −1 −1 In addition to movements that result in changes in loca- kg d ; [99]) derived using the doubly labelled water tion, accelerometers are able to measure all body move- method, and energy requirement estimates based on food −1 −1 ment costs (e.g., scratching, interacting with conspecifics) consumption (473–715  kJ kg d ; [46, 58, 65]). The regardless of changes in animal location [70]. Our results DEE we derived may differ from prior estimates of wolf suggest that DEE was substantially impacted by body energy requirements via methodological or ecological movements that were not captured by changes in hourly variation. Rather than measuring FMR over several days locations. GPS location fixes have a spatial accuracy to weeks via doubly labelled water or estimating it by a of ≤ 31  m [106], which may have also introduced some multiple of basal metabolic rate, we used equivalent travel error in our movement-derived DEE. Infrequent loca- speeds to link mass-specific wolf oxygen consumption tion-derived measures of DEE inherently underestimate measurements to collar-derived ODBA values from wild animal movement paths, and, therefore, energy expendi- conspecifics. However, unlike doubly labelled water, our ture, and should be considered conservative [28, 107]. method for measuring DEE only accounts for changes in In addition, we found home range size poorly predicted movement-related costs and cannot account for changes mean DEE, which suggests that home range size alone in energy expenditure resulting from thermoregulation, may not serve as a reliable proxy for energy expenditure. reproduction, growth, specific dynamic action, or basal Yet, mean daily distance travelled did strongly correlate metabolic rate [70]. The accelerometer-equipped wolves with mean ODBA-derived DEE, suggesting that point-to- we monitored were highly mobile and active for an aver- point movements were important determinants of energy age of 36% or 8.6  h of each day (Table  2), comparable to expenditure in the individuals we monitored. previous estimates [100, 101]. Metabolic requirements for Once properly calibrated, accelerometers can provide wolves are approximately 25% higher than a typical euthe- fine-scale documentation of animal behavior [108] and rian mammal of similar body mass [102], suggesting that DEE [49, 109], as evidenced here through activity budg- to survive, wolves must consume considerably more calo- ets of wild wolves across multiple seasons. While we ries than would be predicted based on their body mass. averaged our accelerometer data over hourly intervals This elevated cost of carnivory translates into dispropor - to evaluate abiotic determinants of wolf DEE, high-fre- tionately high resource requirements [8]. quency accelerometer data can also be used to evaluate Our analyses focused on quantifying wolf activity instantaneous energetic costs, such as measuring the patterns and energy expenditure (rather than energy costs of individual kill events of prey [28, 110] and escape intake via prey consumption as well) in part due to the responses to disturbance [111, 112]. Our study monitored remoteness of the study area. The largely snow-free col - largely snow-free conditions for wolves, so additional lar deployment duration coupled with the outlying loca- studies are needed to reveal fine-scale wolf behavior tions of DNPP pack territories precluded our ability to and energy budgets in response to snow throughout the field-verify wolf kill remains from GPS clusters. However, course of the winter when it presents more of an impedi- field studies capable of investigating even a small num - ment to movement, although some recent work has been ber of GPS clusters stand to benefit from using acceler - conducted [82, 84]. Future studies are also warranted ometry in combination with GPS telemetry to detect kill to examine thermal effects on fine-scale wolf activity sites to estimate kill rates (and, therefore, energy intake) throughout winter, when temperatures are considerably for wolves. Preliminary assessments of this multi-sensor colder than what we observed [71]. approach were recently demonstrated for wolves and other terrestrial carnivores in a captive setting [103], and Conclusion the technique may prove to be critical in estimating kill Our study demonstrates the capacity of integrating acceler- rates in remote sites, such as Denali, where field-verifying ometry with GPS telemetry to reveal activity and energetic GPS clusters may be logistically challenging, cost pro- insights from carnivores in unprecedented detail. Such hibitive, or both. The combination of energetic intake and analyses offer a mechanistic approach for evaluating wolf expenditure could then be used to inform physiological travel patterns and resource requirements. As northern lat- landscape models of animal movement (e.g., [104, 105]). itudes continue to rapidly warm and change, the application When comparing accelerometer and movement- of these methods to future studies would enable research- derived metrics to estimate DEE in wild wolves, the two ers to track how fluctuations in parameters including snow - measures were strongly correlated (R = 0.71), but ODBA fall patterns and plant phenology and growth cascade up to estimates averaged 1.4 × greater than those obtained impact the spatial ecology and energetics of predators [113, from GPS fixes (Table  1). This difference is linked to the 114]. In lower latitudes, recovering gray wolf populations in distinct sampling intervals of the two sensors: acceler- the USA have recently been delisted from protection under ometers took near-continuous measurements, while we Br yce et al. Animal Biotelemetry (2022) 10:1 Page 11 of 16 the Endangered Species Act of 1973 [115]. Given the loss of energy expenditure associated with downhill travel can federal protection, insight into wolf foraging patterns and be either more or less costly than level costs depending prey requirements obtained via multi-sensor telemetry may on the down-slope angle travelled [95, 119, 120]. be invaluable for informing regionally specific management decisions and promoting the persistence of this keystone Wolf monitoring species throughout its range. In March 2015, male gray wolves were captured in the northern portion of DNPP (see Additional file  1; Fig. 2A) Methods using aerial darting by helicopter [121] and anesthe- Data collection tized with zolazepam–tiletamine (Telazol , Fort Dodge Wolf collar calibration Laboratories, Fort Dodge, IA, USA). Once anesthetized, We utilized a lab-to-field approach in which the routine wolves were weighed, measured, and fitted with the same behaviors and locomotor biomechanics of captive wolves ACC-GPS collars used during behavioral calibration (n = 9 adults, 4 male, 5 female; mass = 37.6  kg ± 0.7 SE) with captive wolves. We selected free-ranging adult male instrumented with ACC-GPS collars; model GPS Plus, wolves that were dominant (i.e., known or suspected to Vectronic Aerospace, Germany; approx. 960  g) were be breeders) so our results would not be confounded by measured in large (> 1 acre) outdoor enclosures prior sex or age-related variation in space use and energetics. to deployment on free-ranging conspecifics in the wild To address seasonal patterns of movement and energy (Fig.  1). ACC-GPS collars sampled acceleration continu- expenditure, we parsed the March–October collar ously at 32 Hz (± 8 g range) and took hourly GPS location deployment window into seasons based on the known fixes. We paired video-recorded (Sony HDR-CX290/B, breeding cycle of wolves in interior Alaska. These were 1080 HD, 60p) observations of captive wolves engaged defined as breeding (February–April), pup-rearing (May– in routine activities with collar accelerometer meas- July), pup recruitment (August–October), and nomadic urements to construct behavior and energy budgets for (November–January; [57]). Our March to October data free-ranging conspecifics. Five wolf behavioral categories collection, therefore, includes insights into all but the were identified for the purpose of this study: rest, station - nomadic winter movements of wolves in interior Alaska, ary, walk, highly active, and run. Behaviors and ODBA, which have been studied extensively (e.g., [62, 122]). a widely used proxy for animal energy expenditure [109, Collars recorded GPS locations hourly, and data were 116], were measured as each wolf was filmed moving downloaded directly from the collars upon retrieval. freely at known speeds behind a vehicle and along a fence During our 8-month study window, wolves were visually line between trainers in outdoor enclosures. Both speed observed from single-engine airplanes on 13 monitor- and metabolic rate are tightly linked to the dynamic com- ing flights to validate current wolf locations, wolf pack ponent of an animal’s body acceleration [109, 117, 118], size and composition, active den site locations and use, which allowed us to use wolf ODBA to translate sensor breeding status of individual wolves, and the timing and output from the collars into travel speed and the meta- suspected causes of mortality. ACC-GPS collars were bolic demands of various activities in the wild. removed at the conclusion of this study. We estimated the increase in DEE due to topography by measuring the slopes travelled by wolves from the Environmental variables change in elevation between consecutive location coor- All environmental variables were recorded at the Kan- dinates (see Additional file  1). Following Dunford et  al. tishna automated Snow Telemetry (SNOTEL) site [61], we then modelled the metabolic cost of travel on (63.53845, −150.98365, elevation: 509  m; https:// wcc. sc. slopes from previous studies of wolf energetics measured egov. usda. gov/ nwcc/ site? siten um= 1072). This station via open-flow respirometry on level and inclined tread - was selected because it is located near the geographic mills. Oxygen consumption (V O ) of wolves on the level center the study area, and therefore, its data may be rep- was measured by Taylor et  al. [59], and V O of wolves resentative of general trends throughout the study area moving on slopes up to 14° was provided by Weibel et al. on the north slope of the Alaska Range. However, it may [60]. The increased energetic cost of travel up a slope not reflect conditions at the locations of each wolf. The was, therefore, calculated as Kantishna SNOTEL site has a year-round precipitation gauge that measures snow and rain along with a mete- −1 V O deg. incline = 0.00743 + 0.028 orological station that records air temperature and other (2) weather parameters. Data were recorded and transmit- ∗ Speed n = 5, R = 0.98, p < 0.001 . ted from Kantishna hourly. These data were exported for −1 −1 −1 the study duration, converted to metric units, and uti- where speed is in m s and V O is in ml O kg  min . 2 2 lized in subsequent analyses. Additional details on the Decline (slope < 0°) costs were modeled as level given that Bryce et al. Animal Biotelemetry (2022) 10:1 Page 12 of 16 measurement of environmental variables are provided in using Shapiro–Wilk tests and the data were determined the Additional file 1. to be not normally distributed. Therefore, we used a non- parametric Wilcoxon signed-rank test for paired samples Wolf movement modelling to quantify whether the CTCRW movement-derived Wolf collars averaged a successful fix rate of 99.7% DEE differed significantly from the ODBA derived DEE. (± 0.08%), but to account for missing or irregularly timed Finally, we used the lm function to assess whether the location data, we used a continuous time-correlated ran- path angles the wolves travelled on varied between indi- dom walk (CTCRW) model (R package ‘crawl’ [123, 124]) viduals. Path angle was transformed using the natural to predict locations on hourly intervals based on the GPS logarithm. locations. We derived utilization distributions from the To examine the effect of season on DEE of wolves, we CTCRW locations using the full deployment period of constructed 2 linear mixed effects models (LMM, via the each wolf. We measured the area of the utilization dis- lmer function in the ‘lme4’ package [130]) with either tribution using 95% (home range) and 50% (core area) ODBA or the CTCRW movement-derived DEE as the of the autocorrelated kernel density estimation (AKDE) dependent variable and biological season and wolf ID method in the R package ‘ctmm’ [125]. Speeds of free- as the independent variables, with the season and wolf ranging wolves were calculated as the distance between ID as nested random variables to account for repeated consecutive CTCRW fixes using the Haversine formula measures per individual and allow variable intercepts divided by the elapsed time. Speed was further calcu- and slopes for season. The function emmeans (from the lated using ODBA from the accelerometers (Eqn. S3). ‘emmeans’ package [131]) was used to calculate EMMs This resulted in two different estimates of slope-informed and test for pairwise differences between seasons for energy expenditure for the wolves (Eqn. S1, Eqn., S4). In each model. both cases, the V O was converted to an hourly whole- Mean hourly ODBA was taken as the mean of ODBA body field energetic cost (in kilojoules) by multiplying by across each hour, transformed for normal Gaussian dis- − 1 20.1 J  ml , by each individual wolf ’s mass (in kg; [126]), tribution using Ordered Quantile normalization (via the and by 60 (for energy expenditure per hour). Hourly ‘bestNormalize’ R package [132]). We constructed an energy expenditures were summed to give DEE (in MJ). LMM with transformed hourly ODBA as the dependent These DEE measurements including slope-corrected variable and the season, and wolf ID as the independ- locomotion costs are used throughout the paper. Any ent variables, with wolf ID and season as nested random days with less than 20 h of ODBA data were excluded in effects. The LMM was fitted with a Bound Optimization the DEE estimates (n = 6). by Quadratic Approximation (‘bobyqa’) optimizer [133]. We also constructed this LMM model with CTCRW- Statistics measured distance as the dependent variable. The func - All analyses were conducted in the R statistical software tion emmeans was used to test for pairwise differences [127]. All Chi-square, F, and p values were obtained using between seasons for each model. the Anova function from the ‘car’ package [128] and con- Similarly, to examine whether the wolves adjusted their ditional R from the R package ’MuMIn’ [129]. Response activity level (mean hourly ODBA) in response to ambi- and explanatory variables of all models described below ent air temperature, we generated a LMM with hourly are summarized in Additional file 3: Table S2. ODBA as the dependent variable and ambient tempera- The lm function, from the base functions in R, was used ture, season, and wolf ID as explanatory variables, with to fit a linear model (LM) with DEE (calculated using an interaction between temperature and season. Wolf ID ODBA) as the dependent variable and wolf ID as the and season were included as nested random effects. Simi - independent variable to test for individual differences. larly, two further LMMs were constructed with the same We fit the same LM with DEE (calculated using CTCRW dependent, independent, and random variables except movement rate) as the dependent variable. We also tested temperature was replaced with either snow depth or pre- the strength and direction of the correlation between cipitation. Finally, three additional LMMs with the same the CTCRW movement-derived DEE and ODBA DEE independent and random variables were constructed using Pearson’s correlation. In addition, we tested for with the CTCRW-derived distances as the dependent correlations between home range size and mean daily variable. For the model of CTCRW distance by snow distance traveled with both measures of mean DEE for depth, CTCRW distance was square root transformed each wolf using linear regression to evaluate the abil- and a Nelder Mead optimizer was used. The two LMMs ity of home range size and movement to serve as prox- with precipitation as an independent variable were fitted ies for energy expenditure. We tested for normality in with a Nelder Mead optimizer. the CTCRW movement-derived DEE and ODBA DEE Br yce et al. Animal Biotelemetry (2022) 10:1 Page 13 of 16 Ecology and Evolutionary Biology (EEB) Department. Funding for wolf cap- To examine the daily activity patterns of wolves, tures and monitoring flights was provided by the US National Park Service. we constructed generalized additive mixed models (GAMMs) using the R package ‘mgcv’ [134]. We set mean Availability of data and materials The data sets generated and/or analyzed during for this study are available hourly ODBA (g) as a function of smoothed hour (0–23) from the corresponding author upon reasonable request. with an interaction with season. ID was included as a random variable. Models were fitted with a cyclic cubic Declarations regression spline and 20 knots. We also constructed the same model with CTCRW measured distance (m) as the Ethics approval and consent to participate This study was conducted in strict accordance with animal ethics including response variable rather than ODBA. All results are pre- capture and handling as approved by the United States Department of the sented as mean ± SE unless otherwise noted. We consid- Interior, National Park Service (Denali; Scientific Research and Collecting Per - ered p values ≤ 0.05 as significant. mit #DENA-2015-SCI-0001) and the University of California Santa Cruz Animal Care and Use Committee (IACUC Protocol #Willt1504). All human interven- tions including capture, administration of immobilizing drugs, radio collaring, and monitoring were conducted to minimize negative/adverse impacts on Abbreviations the welfare of the wolves. In addition, Wolf Park approved the collaring and ACC : Accelerometer; CTCRW : Continuous time-correlated random walk; DEE: observation of their animals. Daily energy expenditure; DNPP: Denali National Park and Preserve; GPS: Global positioning system; ODBA: Overall dynamic body acceleration. Consent for publication Not applicable. Supplementary Information Competing interests The online version contains supplementary material available at https:// doi. The authors declare that they have no competing interests. org/ 10. 1186/ s40317- 021- 00272-w. Author details Additional file 1. Supplementary methods. Department of Ecology and Evolutionary Biology, University of California- Santa Cruz, Coastal Biology Building, 130 McAllister Way, Santa Cruz, CA 95060, Additional file 2: Table S1. Results of GAMMs of mean hourly ODBA and USA. School of Biological Sciences, Institute of Global Food Security, Queen’s distance data analysis using Gaussian distribution with ’log’ link. Estimates University of Belfast, 19 Chlorine Gardens, Belfast BT9 7DL, Northern Ireland, and standard errors are reported as decimal numbers, edf and degrees of 3 4 UK. San Diego Zoo Wildlife Alliance, San Diego, CA 92101, USA. Present freedom are reported in whole numbers. Address: School of the Environment, Washington State University, Pullman, WA Additional file 3: Table S2. Response and explanatory variables of mod- 99164, USA. San Francisco Bay Bird Observatory, 524 Valley Way, Milpitas, CA, els. ‘*’ indicates inclusion of main effects and an interaction, ‘+’ indicates USA. Center for Integrated Spatial Research, Environmental Studies Depart- inclusion of main effects with no interaction. Where nested variables were ment, University of California, Santa Cruz, CA, USA. National Park Service, included, this is indicated as “|”. Denali National Park and Preserve, Central Alaska Inventory and Monitoring Network, P. O. Box 9, Denali Park, AK 99755, USA. U.S. Fish and Wildlife Service, Additional file 4: Figure S1.Captive wolf ODBA (g) for rest and locomo - Arctic National Wildlife Refuge, 101 12th Ave, Fairbanks, AK, USA. tion gaits. Additional file 5: Figure S2. Density plot showing the speed wild wolves Received: 17 August 2021 Accepted: 21 December 2021 utilized for travelling (ODBA >0.25 g) relative to the selected topographical slope. Additional file 6: Figure S3. The correlation between DEE derived from ODBA and CTCRW (i.e., GPS-derived) speed. Blue regression line (with References standard error shading) is described by Eqn. 1, black regression line is line 1. Stephens DW, Krebs JR. Foraging theory. Princeton: Princeton University of isometry (X=Y ). Press; 1986. Additional file 7: Figure S4. Seasonal variation in % of each hour 2. Brown JH, Gillooly JF, Allen AP, Savage VM, West GB. Toward a metabolic associated with behaviors (binned by ODBA, g) for gray wolves in Denali theory of ecology. Ecology. 2004;85:1771–89. National Park and Preserve, Alaska, USA. 3. Speakman JR. The cost of living: Field metabolic rates of small mam- mals. Adv Ecol Res. Academic Press; 1999. p. 177–297. https:// linki Acknowledgements nghub. elsev ier. com/ retri eve/ pii/ S0065 25040 86001 97 We thank T. Cambier & S. Hamilton for logistical support and for assistance 4. Humphries MM, McCann KS. Metabolic ecology. J Anim Ecol. with wild wolf capture in DNPP, as well as P. Goodmann & D. Drenzek and the 2014;83:7–19. other helpful staff at Wolf Park (Battle Ground, IN) for allowing us to observe 5. Peters RH. The ecological implications of body size. Cambridge: Cam- their captive wolves. We also thank J. Kamminga and members of the T.M. bridge University Press; 1983. Williams lab for valuable initial discussions and feedback. 6. Bennett AF, Ruben JA. Endothermy and activity in vertebrates. Science. 1979;206:649–54. Authors’ contributions 7. Carbone C, Pettorelli N, Stephens PA. The bigger they come, the harder CMB and TMW conceived the ideas and designed methodology; CMB, BLB, they fall: body size and prey abundance influence predator–prey ratios. and SMA collected the data; CED, AMP, and CMB analyzed the data; CMB and Biol Lett. 2011;7:312–5. CED led the writing of the manuscript with assistance from YW and AMP. All 8. Carbone C, Teacher A, Rowcliffe JM. The costs of carnivory. PLoS Biol. authors contributed critically to the drafts and gave final approval for publica- 2007;5:e22. https:// doi. org/ 10. 1371/ journ al. pbio. 00500 22. tion. All authors read and approved the final manuscript. 9. Kleiber M. The fire of life: an introduction to animal energetics. New York: Wiley; 1961. Funding 10. McNab BK. The significance of energetics for the population ecology We received support for this project from the National Science Foundation of vertebrates. Physiol Ecol Vertebr a view from energ. Ithaca: Cornell (DBI-1255913, GK-12 DGE-0947923), Mazamas, the ARCS Foundation, and University Press; 2002. p. 431–46. the University of California Santa Cruz Science Internship Program (SIP) and Bryce et al. Animal Biotelemetry (2022) 10:1 Page 14 of 16 11. Withers PC, Cooper CE, Maloney SK, Bozinovic F, Cruz-Neto AP. Physi- 37. Mech LD, Cluff HD. Movements of wolves at the northern extreme of ological characteristics of mammals. In: Withers PC, Cooper CE, Maloney the species’ range, including during four months of darkness. PLoS ONE. SK, Bozinovic F, Cruz-Neto AP, editors. Ecol Environ Physiol Mamm. 2011;6:e25328. https:// doi. org/ 10. 1371/ journ al. pone. 00253 28. Oxford: Oxford University Press; 2016. p. 146–289. 38. Peterson RO, Allen DL. Snow conditions as a parameter in moose–wolf 12. Williams TM, Peter-Heide Jørgensen M, Pagano AM, Bryce CM. Hunters relationships. Nat Can. 1974;101:481–92. versus hunted: new perspectives on the energetic costs of survival at 39. Huggard DJ. Eec ff t of snow depth on predation and scavenging by the top of the food chain. Funct Ecol. 2020;34:2015–29. https:// doi. org/ gray wolves. J Wildl Manage. 1993;57:382–8. 10. 1111/ 1365- 2435. 13649. 40. Thurber JM, Peterson RO. Eec ff ts of population density and pack size on 13. Ripple WJ, Beschta RL. Wolves and the ecology of fear: can predation the foraging ecology of gray wolves. J Mammology. 1993;74:879–89. risk structure ecosystems? Bioscience. 2004;54:755–66. 41. Dale BW, Adams LG, Bowyer RT. Functional response of wolves preying 14. Christianson D, Creel S. Ecosystem scale declines in elk recruitment on barren-ground caribou in a multiple-prey ecosystem. J Anim Ecol. and population growth with wolf colonization: a before-after-control- 1994;63:644–52. impact approach. PLoS ONE. 2014;9:e102330. 42. Hayes RD, Baer AM, Wotschikowsky U, Harestad AS. Kill rate by wolves 15. Ripple WJ, Beschta RL, Painter LE. Trophic cascades from wolves to on moose in the Yukon. Can J Zool. 2000;78:49–59. alders in Yellowstone. J Anim Ecol. 2015;354:254–60. https:// doi. org/ 10. 43. Bieniek PA, Walsh JE, Thoman RL, Bhatt US. Using climate divisions to 1016/j. foreco. 2015. 06. 007. analyze variations and trends in Alaska temperature and precipita- 16. Creel S, Winnie J, Maxwell B, Hamlin K, Creel M, Winnie J, et al. Elk alter tion. J Clim. 2014;27:2800–18. habitat selection as an antipredator responsen to wolves. Ecology. 44. Markon C, Gray S, Berman M, Eerkes-Medrano L, Hennessy T, Hun- 2015;86:3387–97. tington HP, et al. Alaska. In: Reidmiller DR, Avery CW, Easterling DR, 17. Prugh LR, Sivy KJ. Enemies with benefits: integrating positive and nega- Kunkel KE, Lewis KLM, May-cock TK, et al., editors. Impacts, Risks, tive interactions among terrestrial carnivores. Sih A, editor. Ecol Lett. Adapt United States Fourth Natl Clim Assessment, Vol II. Washington, 2020;23:902–18. doi: https:// doi. org/ 10. 1111/ ele. 13489. DC; 2018. p. 1185–241. https:// nca20 18. globa lchan ge. gov/ chapt er/ 18. Levi T, Wilmers CC. Wolves–coyotes–foxes: a cascade among carnivores. 26/. Ecology. 2012;93:921–9. 45. Wendler G, Gordon T, Stuefer M. On the precipitation and precipita- 19. Ripple WJ, Wirsing AJ, Wilmers CC, Letnic M. Widespread mesopredator tion change in Alaska. Atmosphere (Basel). 2017;8:253. effects after wolf extirpation. Biol Conserv. 2013;160:70–9. 46. Głowaciński Z, Profus P. Potential impact of wolves Canis lupus on 20. Newsome TM, Ripple WJ. A continental scale trophic cascade from prey populations in eastern Poland. Biol Conserv. 1997;3207:99–106. wolves through coyotes to foxes. J Anim Ecol. 2014;84:49–59. 47. Stahler DR, Smith DW, Guernsey DS. Foraging and feeding ecology of 21. Wang Y, Allen ML, Wilmers CC. Mesopredator spatial and temporal the gray wolf (Canis lupus): lessons from Yellowstone National Park, responses to large predators and human development in the Santa Wyoming, USA. J Nutr. 2006;136:1923S-1926S. Cruz Mountains of California. Biol Conserv Elsevier Ltd. 2015;190:23–33. 48. Metz MC, Smith DW, Vucetich JA, Stahler DR, Peterson RO. Seasonal 22. Morrison JC, Sechrest W, Dinerstein E, Wilcove DS, Lamoreux JF. Persis- patterns of predation for gray wolves in the multi-prey system of tence of large mammal faunas as indicators of global human impacts. J Yellowstone National Park. J Anim Ecol. 2012;81:553–63. Mammal. 2007;88:1363–80. 49. Halsey LG, Bryce CM. Proxy problems: Why a calibration is essential 23. Ripple WJ, Estes JA, Beschta RL, Wilmers CC, Ritchie EG, Hebblewhite for interpreting quantified changes in energy expenditure from M, et al. Status and ecological effects of the world’s largest carnivores. biologging data. Funct Ecol. 2021;35:627–34. https:// doi. org/ 10. 1111/ Science. 2014;343:151–63. https:// doi. org/ 10. 1126/ scien ce. 12414 84.1365- 2435. 13749. 24. Di Minin E, Slotow R, Hunter LTB, Montesino Pouzols F, Toivonen T, 50. Sand H, Wikenros C, Wabakken P, Liberg O. Eec ff ts of hunting group Verburg PH, et al. Global priorities for national carnivore conservation size, snow depth and age on the success of wolves hunting moose. under land use change. Sci Rep. 2016;6:1–9. Anim Behav. 2006;72:781–9. 25. Berger-Tal O, Polak T, Oron A, Lubin Y, Kotler BP, Saltz D. Integrating 51. MacNulty DR, Smith DW, Mech LD, Vucetich J, Packer C. Nonlinear animal behavior and conservation biology: a conceptual framework. effects of group size on the success of wolves hunting elk. Behav Behav Ecol. 2011;22:236–9. https:// doi. org/ 10. 1093/ beheco/ arq224. Ecol. 2012;23:75–82. https:// doi. org/ 10. 1093/ beheco/ arr159. 26. Wilmers CC, Wang Y, Nickel B, Houghtaling P, Shakeri Y, Allen ML, et al. 52. Cassidy KA, MacNulty DR, Stahler DR, Smith DW, Mech LD. Group Scale dependent behavioral responses to human development by a composition effects on aggressive interpack interactions of gray large predator, the puma. PLoS ONE. 2013;8:e60590. wolves in Yellowstone National Park. Behav Ecol. 2015;26:1352–60. 27. Laundré JW. How large predators manage the cost of hunting. Science. 53. Peterson RO, Jacobs AK, Drummer TD, Mech LD, Smith DW. Leader- 2014;346:33–4. https:// doi. org/ 10. 1126/ scien ce. 12601 70. ship behavior in relation to dominance and reproductive status in 28. Williams TM, Wolfe L, Davis T, Kendall T, Richter B, Wang Y, et al. Instanta- gray wolves. Canis lupus Can J Zool. 2002;80:1405–12. neous energetics of puma kills reveal advantage of felid sneak attacks. 54. MacNulty DR, Smith DW, Mech LD, Eberly LE. Body size and predatory Science. 2014;346:81–5. https:// doi. org/ 10. 1126/ scien ce. 12548 85. performance in wolves: is bigger better? J Anim Ecol. 2009;78:532–9. 29. Wilmers CC, Crabtree RL, Smith DW, Murphy KM, Getz WM. Facilitation 55. Mech LD. Alpha status, dominance, and division of labor in wolf by introduced trophic top predators: grey wolf in Yellowstone National packs. Can J Zool. 1999;77:1196–203. Park subsidies to scavengers. J Anim Ecol. 2003;72:909–16. 56. Mech LD, Boitani L. Wolf Social Ecology. In: Mech LD, Boitani L, edi- 30. Mech LD, Smith DW, Macnulty DR, editors. Wolves on the hunt. Chi- tors. Wolves Behav Ecol Conserv. University of Chicago Press; 2003. p. cago: University of Chicago Press; 2015. 1–34. 31. Mech LD. The Wolf: The Behavior and Ecology of an Endangered Spe- 57. Borg BL, Brainerd SM, Meier TJ, Prugh LR. Impacts of breeder loss on cies. New York: Natural History Press; 1970. social structure, reproduction and population growth in a social canid. J 32. James ARC, Stuart-Smith AK. Distribution of caribou and wolves in rela- Anim Ecol. 2015;84:177–87. tion to linear corridors. J Wildl Manage. 2000;64:154–9. 58. Peterson RO, Ciucci P. The wolf as a carnivore. In: Mech LD, Boitani L, 33. Whittington J, St. Clair CC, Mercer G. Spatial responses of wolves to editors. Wolves Behav Ecol Conserv. University of Chicago Press; 2003. p. roads and trails in mountain valleys. Ecol Appl. 2005;15:543–53. 104–30. 34. Latham ADM, Latham MC, Boyce MS, Boutin S. Movement responses 59. Taylor CR, Heglund NC, Maloiy GM. Energetics and mechanics of ter- by wolves to industrial linear features and their effect on woodland restrial locomotion 1. Metabolic energy consumption as a function of caribou in northeastern Alberta. Ecol Appl. 2011;21:2854–65. speed and body size in birds and mammals. J Exp Biol. 1982;97:1–21. 35. Dickie M, Serrouya R, McNay RS, Boutin S. Faster and farther: wolf move- 60. Weibel ER, Richard Taylor C, O’Neil JJ, Leith DE, Gehr P, Hoppeler H, et al. ment on linear features and implications for hunting behaviour. J Appl Maximal oxygen consumption and pulmonary diffusing capacity: a Ecol. 2016;54:253–63. https:// doi. org/ 10. 1111/ 1365- 2664. 12732. direct comparison of physiologic and morphometric measurements in 36. Mech LD, Korb M. An unusually long pursuit of a deer by a wolf. J Mam- canids. Respir Physiol. 1983;54:173–88. mal. 1978;59:860–1. 61. Dunford CE, Marks NJ, Wilmers CC, Bryce CM, Nickel B, Wolfe LL, et al. Surviving in steep terrain: a lab-to-field assessment of locomotor costs Br yce et al. Animal Biotelemetry (2022) 10:1 Page 15 of 16 for wild mountain lions (Puma concolor). Mov Ecol. 2020;8:34. https:// 86. Fancy SG, White RG. Energy expenditures for locomotion by barren- doi. org/ 10. 1186/ s40462- 020- 00215-9. ground caribou. Can J Zool. 1987;65:122–8. 62. Mech LD, Adams LG, Meier TJ, Burch JW, Dale BW. The wolves of denali. 87. Shepard ELC, Wilson RP, Rees WG, Grundy E, Lambertucci S, Vosper Minneapolis: University of Minnesota Press; 1998. SB. Energy landscapes shape animal movement ecology. Am Nat. 63. Tatler J, Currie SE, Cassey P, Scharf AK, Roshier DA, Prowse TAA. Accel- 2013;182:298–312. erometer informed time-energy budgets reveal the importance of tem- 88. Wilson RP, Quintana F, Hobson VJ. Construction of energy landscapes perature to the activity of a wild, arid zone canid. Mov Ecol. 2021;9:11. can clarify the movement and distribution of foraging animals. Proc R https:// doi. org/ 10. 1186/ s40462- 021- 00246-w. Soc B Biol Sci. 2012;279:975–80. 64. Rabaiotti D, Woodroffe R. Coping with climate change: limited 89. Halsey LG. Terrestrial movement energetics: current knowledge and behavioral responses to hot weather in a tropical carnivore. Oecologia. its application to the optimising animal. J Exp Biol. 2016;219:1424–31. 2019;189:587–99. https:// doi. org/ 10. 1007/ s00442- 018- 04329-1.https:// doi. org/ 10. 1242/ jeb. 133256. 65. Kreeger TJ, Mech LD. The internal wolf: Physiology, pathology & phar- 90. Pagano AM, Atwood TC, Durner GM, Williams TM. The seasonal ener- macology. In: Boitani L, editor. Wolves Behav Ecol Conserv. Chicago: getic landscape of an apex marine carnivore, the polar bear. Ecology. University of Chicago Press; 2003. p. 192–217. 2020;101:1–16. https:// doi. org/ 10. 1002/ ecy. 2959. 66. Theuerkauf J. What drives wolves: fear or hunger? Humans, diet, climate 91. Nickel BA, Suraci JP, Nisi AC, Wilmers CC. Energetics and fear of humans and wolf activity patterns. Ethology. 2009;115:649–57. https:// doi. org/ constrain the spatial ecology of pumas. Proc Natl Acad Sci. 2021;118: 10. 1111/j. 1439- 0310. 2009. 01653.x. e2004592118. https:// doi. org/ 10. 1073/ pnas. 20045 92118. 67. Daan S, Masman D, Groenewold A. Avian basal metabolic rates: their 92. Nisi AC, Suraci JP, Ranc N, Frank LG, Oriol-Cotterill A, Ekwanga S, et al. association with body composition and energy expenditure in nature. Temporal scale of habitat selection for large carnivores: balancing Am J Physiol Integr Comp Physiol. 1990;259:R333–40. https:// doi. org/ energetics, risk and finding prey. J Anim Ecol. 2021. https:// doi. org/ 10. 10. 1152/ ajpre gu. 1990. 259.2. R333.1111/ 1365- 2656. 13613. 68. Speakman JR, Thomas MD. Physiological ecology and energetics of 93. Singleton PH, Gaines WL, Lehmkuhl JF. Landscape permeability for large bats. In: Kunz TH, Fenton MB, editors. Bat Biol. University of Chicago carnivores in Washington: a Geographic Information System weighted- Press; 2003. p. 430–90. distance and least-cost corridor assessment. Portland, OR; 2002. https:// 69. Careau V, Morand-Ferron J, Thomas D. Basal metabolic rate of Canidae books. google. com/ books? hl= en& lr= & id= yI64KX- TrFcC & oi= fnd& pg= from hot deserts to cold arctic climates. J Mammal. 2007;88:394–400. PA1& dq= singl eton+ weigh ted+ dista nces& ots= DE6hh JtVYN & sig= https:// doi. org/ 10. 1644/ 06- MAMM-A- 111R1.1.R43Df JY- jMZcU 4lz_ lO98V idM2U#v= onepa ge&q= singl etonw eight 70. Pagano AM, Williams TM. Estimating the energy expenditure of free-eddis tance s&f= false. ranging polar bears using tri-axial accelerometers: a validation with 94. LaPoint S, Gallery P, Wikelski M, Kays R. Animal behavior, cost-based cor- doubly labeled water. Ecol Evol. 2019;9:4210–9. https:// doi. org/ 10. 1002/ ridor models, and real corridors. Landsc Ecol. 2013;28:1615–30. https:// ece3. 5053.doi. org/ 10. 1007/ s10980- 013- 9910-0. 71. Hooge BP, Adema G, Meier T, Roland C, Brease P, Sousanes P, et al. 95. Halsey LG, White CR. A different angle: comparative analyses of whole- Ecological overview of Denali National Park and Preserve. Alaska Park animal transport costs running uphill. J Exp Biol. 2016;220:161. Sci. 2006;5:6–13. 96. Reichman OJ, Aitchinson S. Mammal trails on mountain slopes: 72. Ditmer MA, Moen RA, Windels SK, Forester JD, Ness TE, Harris TR. Moose optimal paths in relation to slope angle and body weight. Am Nat. at their bioclimatic edge alter their behavior based on weather, land- 1981;117:416–20. scape, and predators. Curr Zool. 2018;64:419–32. 97. Nathan R, Getz WM, Revilla E, Holyoak M, Kadmon R, Saltz D, et al. A 73. Renecker LA, Hudson RJ. Seasonal energy expenditures and thermoreg- movement ecology paradigm for unifying organismal movement ulatory responses of moose. Can J Zool. 1986;64:322–7. research. Proc Natl Acad Sci. 2008;105:19052–9. 74. Street GM, Rodgers AR, Fryxell JM. Mid-day temperature variation 98. Morales JM, Moorcroft PR, Matthiopoulos J, Frair JL, Kie JG, Powell R, influences seasonal habitat selection by moose. J Wildl Manage. et al. Building the bridge between animal movement and population 2015;79:505–12. https:// doi. org/ 10. 1002/ jwmg. 859. dynamics. Philos Trans R Soc Lond B Biol Sci. 2010;365:2289–301. 75. Mörschel FM, Klein DR. Eec ff ts of weather and parasitic insects on 99. Nagy K. Field bioenergetics of mammals: what determines field meta- behavior and group dynamics of caribou of the Delta Herd. Alaska Can bolic rates? Aust J Zool. 1994;42:43–53. J Zool. 1997;75:1659–70. 100. Eggermann J, Gula R, Pirga B, Theuerkauf J, Tsunoda H, Brzezowska 76. Russell DE, Gunn A, Kutz S. Migratory Tundra Caribou and Wild Rein- B, et al. Daily and seasonal variation in wolf activity in the Bieszczady deer. NOAA Arct Rep Card. 2018. p. 67–73. Mountains. SE Poland Mamm Biol. 2009;74:159–63. 77. Muñoz AR, Márquez AL, Real R. An approach to consider behavioral 101. Theuerkauf J, Jędrzejewski W, Schmidt K, Okarma H, Ruczyński I, Śnieżko plasticity as a source of uncertainty when forecasting species’ response S, et al. Daily patterns and duration of wolf activity in the Białowieża to climate change. Ecol Evol. 2015;5:2359–73. Forest. Poland J Mammal. 2003;84:243–53. https:// doi. org/ 10. 1644/ 78. Noonan MJ, Newman C, Markham A, Bilham K, Buesching CD, Mac-1545- 1542(2003) 084% 3C0243: DPADOW% 3E2.0. CO;2. donald DW. In situ behavioral plasticity as compensation for weather 102. Nagy KA, Girard IA, Brown TK. Energetics of free ranging mammals, variability: implications for future climate change. Clim Change. reptiles, and birds. Annu Rev Nutr. 1999;19:247–77. 2018;149:457–71. https:// doi. org/ 10. 1007/ s10584- 018- 2248-5. 103. Petroelje TR, Belant JL, Beyer DE, Svoboda NJ. Identification of carnivore 79. Sousanes P. Snow surveys in Denali. US National Park Service; 2016. kill sites is improved by verified accelerometer data. Anim Biotelemetry. https:// www. nps. gov/ artic les/ denali- crp- snow- surve ys. htm. 2020;8:18. https:// doi. org/ 10. 1186/ s40317- 020- 00206-y. 80. Saperstein L. Winter forage selection by barren-ground caribou: effects 104. Hooten MB, Scharf HR, Morales JM. Running on empty: recharge of fire and snow. Rangifer. 1996;16:237. dynamics from animal movement data. Ecol Lett. 2018;22:377–89. 81. Lundmark C, Ball JP. Living in snowy environments: quantifying 105. Latombe G, Parrott L, Basille M, Fortin D. Uniting statistical and the influence of snow on moose behavior. Arctic, Antarct Alp Res. individual-based approaches for animal movement modelling. PLoS 2008;40:111–8. ONE. 2014;9:e99938. https:// doi. org/ 10. 1371/ journ al. pone. 00999 38. 82. Droghini A, Boutin S. The calm during the storm: snowfall events 106. D’Eon RG, Serrouya R, Smith G, Kochanny CO. GPS radiotelemetry error decrease the movement rates of grey wolves (Canis lupus). PLoS ONE. and bias in mountainous terrain. Wildl Soc Bull. 2002;30:430–9. 2018;13:1–15. https:// doi. org/ 10. 1371/ journ al. pone. 02057 42. 107. Brown DD, LaPoint S, Kays R, Heidrich W, Kummeth F, Wikelski M. Accel- 83. Telfer ES, Kelsall JP. Adaptation of some large North American mammals erometer-informed GPS telemetry: reducing the trade-off between for survival in snow. Ecology. 1984;65:1828–34. resolution and longevity. Wildl Soc Bull. 2012;36:139–136. 84. Droghini A, Boutin S. Snow conditions influence grey wolf (Canis lupus) 108. Brown DD, Kays R, Wikelski M, Wilson R, Klimley A. Observing the travel paths: the effect of human-created linear features. Can J Zool. unwatchable through acceleration logging of animal behavior. Anim 2018;96:39–47. Biotelemetry. 2013;1:1. 85. Central Alaska Network. Weather & Climate Summary: Denali National 109. Wilson RP, White CR, Quintana F, Halsey LG, Liebsch N, Martin GR, Park & Preserve Winter 2014-Fall 2015. 2015. et al. Moving towards acceleration for estimates of activity-specific Bryce et al. Animal Biotelemetry (2022) 10:1 Page 16 of 16 metabolic rate in free-living animals: the case of the cormorant. J Anim 134. Wood SN. Fast stable restricted maximum likelihood and marginal Ecol. 2006;75:1081–90. likelihood estimation of semiparametric generalized linear models. J R 110. Wilmers CC, Isbell LA, Saraci JP, Williams TM. Energetics-informed Stat Soc Ser B. 2011;73:3–36. https:// doi. org/ 10. 1111/j. 1467- 9868. 2010. behavioral states reveal the drive to kill in African leopards. Ecosphere. 00749.x. 2017;8:1–12. 111. Bryce CM, Wilmers CC, Williams TM. Energetics and evasion dynamics of Publisher’s Note large predators and prey: pumas vs. hounds. PeerJ. 2017;5:1–23. Springer Nature remains neutral with regard to jurisdictional claims in pub- 112. Williams TM, Blackwell SB, Richter B, Sinding MS, Heide-Jørgensen MP. lished maps and institutional affiliations. Paradoxical escape responses by narwhals (Monodon monoceros). Sci- ence. 2017;358:1328–31. https:// doi. org/ 10. 1126/ scien ce. aao27 40. 113. Bowyer T, Person DK, Becky MP. Detecting top-down versus bottom-up regulation of ungulates by large carnivores: Implications for conserva- tion of biodiversity. In: Ray JC, Redford KH, Steneck RS, Berger J, editors. Large Carniv Conserv Biodivers. 2005. p. 342–61. 114. Ripple WJ, Beschta RL. Large predators limit herbivore densities in northern forest ecosystems. Eur J Wildl Res. 2012;58:733–42. https:// doi. org/ 10. 1007/ s10344- 012- 0623-5. 115. Carroll C, Rohlf DJ, VonHoldt BM, Treves A, Hendricks SA. Wolf delisting challenges demonstrate need for an improved framework for conserv- ing intraspecific variation under the Endangered Species Act. Biosci- ence. 2020;71:73–84. https:// doi. org/ 10. 1093/ biosci/ biaa1 25/ 59418 53. 116. Gleiss AC, Wilson RP, Shepard ELC. Making overall dynamic body acceleration work: on the theory of acceleration as a proxy for energy expenditure. Methods Ecol Evol. 2011;2:23–33. https:// doi. org/ 10. 1111/j. 2041- 210X. 2010. 00057.x. 117. Qasem L, Cardew A, Wilson A, Griffiths I, Halsey LG, Shepard ELC, et al. Tri-axial dynamic acceleration as a proxy for animal energy expenditure; should we be summing values or calculating the vector? PLoS ONE. 2012;7:e31187. 118. Bidder OR, Qasem L, Wilson RP. On higher ground: how well can dynamic body acceleration determine speed in variable terrain? PLoS ONE. 2012;7:1–7. 119. Raab JL, Eng P, Waschler RA. Metabolic cost of grade running in dogs. J Appl Physiol. 1976;41:532–5. 120. Carnahan AM, van Manen FT, Haroldson MA, Stenhouse GB, Robbins CT. Quantifying energetic costs and defining energy landscapes experi- enced by grizzly bears. J Exp Biol. 2021;224:1–9. 121. Meier TJ, Burch JW. Wolf monitoring protocol for Denali National Park and Preserve, Yukon-Charley Rivers National Preserve and Wrangell-St. Elias National Park and Preserve. Fort Collins, CO; 2009. 122. Meier TJ, Burch JW, Adams LG. Tracking the movement of Denali’s wolves. Alaska Park Sci. 2006;5:30–5. 123. Johnson DS, London JM. crawl: an R package for fitting continuous- time correlated random walk models to animal movement data. Zenodo. 2018. https:// cran.r- proje ct. org/ web/ packa ges/ crawl/ crawl. pdf. 124. Johnson DS, London JM, Lea MA, Durban JW. Continuous-time correlated random walk model for animal telemetry data. Ecology. 2008;89:1208–15. 125. Calabrese JM, Fleming CH, Gurarie E. ctmm: an R package for analyz- ing animal relocation data as a continuous-time stochastic process. Methods Ecol Evol. 2016;7:1124–32. https:// doi. org/ 10. 1111/ 2041- 210X. 126. Schmidt-Nielsen K. Animal physiology: adaptation and environment. 5th ed. Cambridge: Cambridge University Press; 1997. 127. R Foundation for Statistical Computing. R: a language and environment for statistical computing. 2020. http:// www.r- proje ct. org/ Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : 128. Fox J, Weisberg S. An R Companion to Applied Regression. SAGE Publi- cations; 2019. fast, convenient online submission 129. Bartoń K. Package “MuMIn”: Multi-Model Inference. 2020. p. 75. thorough peer review by experienced researchers in your field 130. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67. http:// www. jstat soft. org/ v67/ rapid publication on acceptance i01/ support for research data, including large and complex data types 131. Lenth R V, Buerkner P, Herve M, Love J, Riebl H, Singmann H. Package • gold Open Access which fosters wider collaboration and increased citations ‘emmeans’: Estimated Marginal Means, aka Least-Squares Means. 2021. https:// github. com/ rvlen th/ emmea ns. maximum visibility for your research: over 100M website views per year 132. Peterson RA. Package ‘bestNormalize’’: Normalizing Transformation Functions’. 2021 https:// github. com/ peter sonR/ bestN ormal ize. At BMC, research is always in progress. 133. Powell M. The BOBYQA algorithm for bound constrained optimization Learn more biomedcentral.com/submissions without derivatives. NA Rep NA2009/06. 2009;39.