Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

Megan C Milligan; Lorelle I Berkeley; Lance B McNew

doi:10.1371/journal.pone.0233756

. 2020 Jun 4;15(6):e0233756. doi: 10.1371/journal.pone.0233756

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

Megan C Milligan ^1,^*, Lorelle I Berkeley ², Lance B McNew ¹

Editor: W David Walter³

PMCID: PMC7272000 PMID: 32497098

Abstract

Habitat selection links individual behavior to population abundance and dynamics, so evaluation of habitat selection is necessary for conservation and management. Land management can potentially alter both the structure and composition of habitats, thus influencing habitat selection and population size. Livestock grazing is the dominant land use worldwide and, while overstocking has been linked to the decline of many wildlife species, properly managed grazing could improve habitat quality and maintain native rangeland habitats. We evaluated breeding season habitat selection of female sharp-tailed grouse, an indicator species for grassland ecosystems, in relation to grazing management and landscape features in eastern Montana and western North Dakota. At broad spatial scales, females selected for multiple landscape features, including grassland, but exhibited no selection for either landscape or management variables when selecting habitat at smaller spatial scales. Females selected for pastures managed with rest-rotation grazing when choosing a home range, but selection did not equate to improved fitness. Moreover, we observed strong individual variation in both home range size and third-order habitat selection. While the high variability among individuals makes specific management recommendations difficult, selection for grassland habitats at broad scales suggests that strategies that maintain intact native rangelands are important for the conservation of sharp-tailed grouse.

Introduction

Habitat selection, especially for reproduction, is an important individual decision-making process that links individual behavior to population abundance and dynamics and determines the spatial distribution of both species and individuals [1–3]. The process of habitat selection represents a trade-off in which individuals balance competing demands, such as acquiring resources and avoiding predators, to maximize fitness [4, 5]. Thus, habitat selection is a key behavior that allows individuals to respond to spatial and temporal variation in their environment [6], and research increasingly suggests that both demography and habitat selection of wildlife populations vary spatially [7–10].

Land management has the potential to alter both the structure and composition of habitats and thus can influence the habitat selection of individuals and populations. Livestock grazing is the dominant land use worldwide and can influence the structure, composition, and productivity of habitats [11–14]. While overgrazing has been implicated in the decline of many wildlife species [13], properly managed grazing could benefit populations both by keeping native grasslands intact and providing a mosaic of habitats in different stages of disturbance, which may mimic historic disturbance regimes [11, 15, 16]. Specialized grazing systems that focus on enhancing structural and compositional heterogeneity in vegetation are being increasingly promoted and some systems, most notably patch-burn grazing, have been shown to benefit both wildlife and cattle production in tallgrass prairie ecosystems [16–21]. However, the effects of grazing on both wildlife and vegetation vary across broad spatial scales and are often strongly influenced by mediating factors such as precipitation and soil conditions [11, 22–25].

Rest-rotation grazing is a specialized grazing system that could function similarly to patch-burn grazing [26, 27] in areas like the northern Great Plains where fire is not a socially acceptable management technique [28], although this hypothesis has not been adequately tested. Originally developed to improve range condition [29], the rest-rotation system, developed by Hormay and Evanko [30], is based on the idea that grazing during consecutive growing seasons reduces plant vigor and that rest from grazing is necessary to allow plants to recover [29, 30]. By altering the timing of grazing for individual pastures each year and incorporating an additional period of rest, rest-rotation grazing could also create a patchwork of habitats on the landscape, with rested pastures having the most residual cover [26, 27]. By utilizing a patchwork of habitats, individuals may be able to better balance foraging activities with predator avoidance. The effects of grazing system, however, are also influenced by stocking rate, which is a measure of grazing intensity, and high stocking rates can have negative effects, particularly on grouse [17, 18, 25, 31].

Recognized as an indicator species for grassland ecosystems [32], sharp-tailed grouse (Tympanuchus phasianellus) are a model species to evaluate the effects of livestock grazing on wildlife. Throughout their life history, sharp-tailed grouse have diverse habitat requirements, including short, bare areas for lekking, denser herbaceous cover for nesting, and deciduous shrubs for winter cover and food. Identifying management strategies to conserve grouse populations could improve conservation of a variety of other grassland species [33, 34]. However, very little is known about the general spatial ecology of sharp-tailed grouse and specifically the effects of livestock grazing on their space use.

Habitat selection by prairie-grouse (Tympanuchus spp.) is driven in part by predator avoidance, so having sufficient cover is important to conceal both nests and adults [35]. Therefore, grazing, which can influence both the structure and composition of habitats, could have important indirect effects on grouse selection behavior. Patch-burn grazing, a management strategy that increases heterogeneity in tallgrass prairies, improved habitat for greater prairie-chickens (T. cupido) and lesser prairie chickens (T. pallidinctus) relative to management that incorporates annual spring burning and intensive early stocking [36, 37]. Beyond the effects of patch-burn grazing on prairie-chickens, however, the effects of livestock grazing on prairie-grouse are not well understood [38].

Other factors, such as landscape configuration and anthropogenic development, can also influence habitat selection. Grouse have been shown to minimize predation risk at multiple spatial scales by selecting for habitats providing horizontal and vertical cover [39–41], sites with more grassland on the landscape [42–44], and less cropland [45, 46, but see 39]. Other studies, however, suggest that landcover does not have a large influence on selection or that selection for different habitat types varies among sites [43, 46, 47]. Anthropogenic development generally has negative effects on grouse. Greater sage-grouse (Centrocercus urophasianus) selected for lower densities of oil and gas development, sharp-tailed grouse avoided roads and distribution lines [48], and greater and lesser prairie-chickens avoided anthropogenic structures and expanded home ranges in proximity to wind energy development [47, 49, 50]. Home range size was not related to road density, however, and selection for or against roads varied among study areas for prairie-chickens [43]. Further complicating relationships, aspects of habitat selection can change from year to year with different weather conditions [36], and can vary across spatial scales, with home range size for prairie-chickens, for example, related to the amount of precipitation received at different sites spread across multiple states [43]. Furthermore, habitat selection can vary with the availability of a resource, termed a functional response, where individuals experiencing different conditions may respond differently [51]. Taken together, the lack of information for sharp-tailed grouse and the differing results for related species across time and space make generalized habitat management recommendations inappropriate.

Our objective was to evaluate the effects of livestock grazing management on the breeding season habitat selection of female sharp-tailed grouse while considering other habitat features at multiple orders of selection. Habitat selection is a hierarchical process and studies that evaluate selection at multiple spatial scales can improve understanding of wildlife-habitat relationships [52, 53]. We evaluated both second- and third-order habitat selection of female grouse, defined as the selection of habitat for an individual’s home range within the larger study area and the selection of habitat within an individual’s home range, respectively [52]. Livestock grazing has the potential to maintain grassland habitats [54] and we hypothesized that grouse would select for large grassland patches at all orders of selection. Furthermore, rest-rotation grazing could influence grouse habitat selection by creating a patchwork of habitats that are periodically rested from disturbance. Therefore, we hypothesized that if rest-rotation grazing increases heterogeneity in grassland habitats, then females would select for rest-rotation pastures and have smaller home ranges when using those potentially higher-quality pastures due to increased availability or proximity of important resources.

Study area

This study was conducted during 2016–2018 in southern Richland and McKenzie Counties in eastern Montana and western North Dakota, U.S.A., respectively (centered on 47.52°N, -104.06°W). The study area was composed of Great Plains mixed-grass prairie interspersed with Great Plains badlands and wooded draws and ravines [55] and was primarily managed for cattle production. Vegetation was a mixture of mid and short grasses, with western wheatgrass (Pascopyrum smithii), little bluestem (Schizachyrium scoparium), needle-and-thread (Hesperostipa comata), Kentucky bluegrass (Poa pratensis), blue grama (Bouteloua gracilis), and crested wheatgrass (Agropyron cristatum) being the dominant graminoids. The three study years differed drastically in the amount of precipitation received. We obtained daily precipitation data from the National Oceanic and Atmospheric Association (NOAA) station in Sidney, MT, and calculated the amount of precipitation received annually (1 January–31 December) and during the sharp-tailed grouse breeding season (15 March–15 August). Annual precipitation was 419.3 mm in 2016, 216.4 mm in 2017, and 341.5 mm in 2018. Total precipitation during the breeding season was 268.7 mm in 2016, 105.2 mm in 2017, and 312.1 mm in 2018.

The study area was centered on an ~3,300-ha Upland Gamebird Enhancement Program (UGBEP) project established by the Montana Department of Fish, Wildlife and Parks in 1993 that included four separate 3-pasture Hormay rest-rotation systems (Hormay and Evanko 1958). In a given year, cattle were stocked in one pasture from May—July (growing season), then moved to a second pasture during August—October (post-growing season), while the third pasture was rested and the order of rotation was shifted within each 3-pasture rest-rotation system the next year. Therefore, no pasture was grazed during the same season in consecutive years and pastures rested in the previous year theoretically should have had the most residual cover. Average pasture size in the four rest-rotation systems was 292 ± 116 ha. Pastures of surrounding ranches, which included both private land and 4 pastures located on U.S. Forest Service National Grasslands were managed with more commonly used livestock grazing systems, including both season-long systems (19 pastures, ~4,800 ha) and 2- and 3-pasture summer rotation systems (25 pastures, ~5,200 ha). Grazing occurred in season-long pastures from approximately May to early November, while cattle were stocked in the same pastures in summer rotation systems each year for the same 6–8-week period (approximately April–June, June–July or Aug–Nov). Average pasture sizes in the season-long and summer rotation systems were 242 ± 312 ha and 238 ± 335 ha, respectively. Stocking rates were controlled by landowners and lessees and averaged rates were 0.93 animal unit month (AUM) ha^-1, 1.46 AUM ha^-1, and 0.76 AUM ha^-1 for rest-rotation, season-long, and summer rotation pastures, respectively. The range of stocking rates for grazed pastures was 0.38–3.25 AUM ha^-1, 0.17–4.28 AUM ha^-1, and 0.21–4.45 AUM ha^-1 for rest-rotation, season-long, and summer rotation pastures, respectively, and included similar distributions within each grazing system [56]. Average stocking rates did not exceed the range of rates (1.11–1.48 AUM ha^-1) recommended by the Natural Resources Conservation Service (NRCS) for the most common ecological site (R058AE001MT) in the study area. Environmental variables including topography, average vegetation productivity, soil type, vegetation canopy greenness as measured by the Normalized Difference Vegetation Index (NDVI) in June 2018, and the variation in small-scale vegetation cover and structure were similar among grazing systems [56].

Methods

We captured grouse using walk-in funnel traps at 12 leks (5 in rest-rotation pastures, 3 in summer rotation pastures, and 4 in season-long pastures) during March—May in 2016–2018. Females were fitted with very high frequency (VHF) radio-transmitters (model A4050; Advanced Telemetry Systems, Isanti, MN). Radio-marked females were located by triangulation or homing ≥ 3 times/week during the breeding season (15 March– 15 August). Coordinates for triangulated locations were calculated using Location of a Signal software (LOAS; Ecological Software Solutions LLC, Hegymagas, Hungary) and examined for spatial error. All locations with low estimation precision (> 200 m error ellipse) were discarded. All animal handling was approved under Montana State University’s Institutional Animal Care and Use Committee (Protocol #2016–01) and permits for field studies were obtained from both Montana Fish, Wildlife and Parks and North Dakota Game and Fish.

We analyzed location data for the breeding season (15 March– 15 August) and defined a home range as the space an individual needed to forage, reproduce, and survive. Previous studies have found that small sample sizes can bias home range estimates [57], so analyses were restricted to birds with ≥ 30 locations and ≥ 20 locations not associated with a nest site. We used the fixed kernel method [58] with the default smoothing parameter to calculate home ranges as the 95% utilization distribution for the breeding season (April—August) using the adehabitatHR package in Program R v3.5.1.

We identified nine landscape metrics a priori that could influence sharp-tailed grouse space use. Three of those metrics were related to rangeland management: grazing system and stocking rate (AUM ha^-1) during either the current or previous year. Two landscape metrics represented anthropogenic disturbance, including oil pads and roads. Four additional landscape variables were related to landcover: % grassland, % wooded draws, % cropland, and the density of edge habitat (total landcover edge length / polygon area). We collected information on grazing management for every pasture in the study area by interviewing landowners to determine the number and class of animals stocked and the timing of stocking to determine the grazing system (rest-rotation, summer rotation, season-long) and stocking rate (AUM ha^-1) during the current and previous year. Stocking rate is a measure of the number of animals in a unit area (e.g., pasture) during the entire grazing season. We digitized the location of oil pads and roads in the study area and roads were defined as paved and dirt state and county roads and did not include ranch two-tracks. We utilized the 30-m resolution LANDFIRE data depicting landcover type for habitat classifications [55]. A habitat patch edge was defined as any area where the landcover type (grassland, wooded draws, or cropland) of adjacent pixels was different and edge density was defined as the length of patch edge divided by home range size.

We examined second order selection, or an individual’s selection of a home range within the larger study area, using resource selection functions to compare used and available home ranges following Design II of Manly et al. [59]. We characterized grouse resource use with estimated home ranges for each individual for each year. If an individual was monitored in multiple years, we randomly selected one home range to include in analyses. To sample availability, we randomly placed 1,000 circular home ranges across the study area that were equal in area to the average grouse home range size (~500 ha). The study area was defined as the 99% kernel utilization distribution estimated using locations from all collared individuals. Using the spatial layers described above, we calculated the following variables within each used and available home range: proportion grassland, proportion wooded draws, proportion cropland, total edge density, average distance to oil pad, average distance to road, and the proportion of each grazing system. We then examined correlations for each pair of explanatory variables (r > |0.6|), and excluded proportion cropland and edge density, which were both highly collinear with proportion grassland. We then used logistic regression to compare used and available home ranges with available home ranges weighted (w = 1000) to improve convergence [60]. We first evaluated all combinations of habitat and anthropogenic variables, and then built a final candidate model set including a top habitat and anthropogenic model in combination with all combinations of grazing management variables. We compared models based on Akaike’s Information Criterion for small sample sizes (AIC_c) and models representing the majority of model weight (w_i) were considered the most important [61]. We considered variables to be significant if 85% confidence intervals did not overlap zero and variables were considered uninformative if a model was <2 ΔAIC_c units from the top model but only differed in a single parameter [62].

We used linear models to evaluate the relationship between home range size and each metric described above, as well as the effects of year; nest outcome; and distance to nearest lek. We evaluated all single-variable models using Akaike’s Information Criterion corrected for small sample sizes (AIC_c). Models that were within 2 ΔAIC_c of the top model and represented a majority of model weight (w_i > 0.6) were considered important.

To evaluate third-order habitat selection, or the selection of habitat within individual home ranges, we used resource selection functions to compare used and available points following Design III of Manly et al. [59], where individual telemetry locations were classified as used points and available points were randomly sampled for each individual within their home range. We evaluated the nine landscape metrics described above at both used and available points. For the habitat variables, we used FRAGSTATS 4.2 [63] to conduct a moving window analysis to calculate the proportion of each landcover type and the density of edge habitat within 8 buffer distances representing various spatial scales of influence (30, 75, 125, 200, 500, 750, 1000, 1300 m) and evaluated the scale for each landcover type that best predicted grouse space use [64]. We chose scales across a continuum, with 30 m representing the minimum size imposed by our spatial data and 1,300 m approximating the average size of the breeding season home range of a female sharp-tailed grouse in our study area. A scale of 200 m represents the average distance moved daily by female sharp-tailed grouse during the breeding season in our study. The remaining scales represent intermediate distances between the minimum imposed by our spatial data and the average size of a breeding season home range.

We conducted 1,000 simulations for each variable and each scale of landcover variables to determine the number of available points required for coefficient estimates to converge [S1–S5 Figs; 60]. Based on the simulations, available points were sampled at a 15:1 available:used ratio within each individual bird’s home range to balance coefficient convergence and computational efficiency. For all models, we used binomial linear mixed models in a Bayesian framework with both random intercepts and slopes to account for potential autocorrelation among sampling points and individual variation in selection [65, 66]. For the four landcover covariates, we first selected the spatial scale at which selection was the strongest. We compared the 8 buffer distances using calculated deviance information criteria (DIC) to identify a top model sensu Laforge et al. [64], and we considered > 5 DIC units to be a substantial difference in model fit [66].

After assessing collinearity for each pair of explanatory variables (r ≥ 0.6) and selecting the variable with the most support based on calculated DIC, we then evaluated support for all management and landscape variables in a full model. We centered and scaled all predictor variables to calculate standardized coefficients of fixed effects to make population-level inferences about each habitat variable and improve model convergence. Coefficients with 95% credible intervals that did not overlap zero were considered important. To determine the degree of variation in selection among individuals, we examined variation in individual-specific slopes for each predictor variable and calculated the number of individuals that were significantly selecting for or against each variable based on 95% credible intervals. To evaluate whether selection varied with resource availability, we calculated the mean value of each covariate at used and available points for each individual female [67, 68] and plotted the use of a variable against its availability [51].

We fit all binomial selection models with random intercepts and slopes using integrated nested Laplace approximation (INLA) using the R-INLA package in Program R. This approach is a computationally efficient alternative to existing algorithms because it circumvents Markov chain Monte Carlo (MCMC) sampling by providing efficient approximations of marginal posterior distributions and it has been shown to be useful for fitting generalized linear mixed models used to calibrate resource selection functions [69–71]. Following recommendations from Muff et al. [69], we used independent priors with large prior variance for all model components and used penalized complexity priors for the variances of random slopes (see example code in S1 Appendix). Because individual-specific intercepts are not of interest in resource selection functions, we treated them as random effects with large, fixed variance (10⁶) following Muff et al. [69].

Results

During the 2016–2018 breeding seasons, we collected a total of 7,178 locations and calculated 142 home ranges for 118 individual females (40 in 2016, 53 in 2017, 49 in 2018). Home range size was estimated without bias relative to sampling effort (S1 Table). Mean breeding season home range size for all females was 489 ± 41 ha but varied from 58–3,717 ha (Table 1). Home range sizes were less variable within pastures managed with summer rotation grazing compared to those in other systems (Fig 1), but grazing system did not have a significant effect on average size of home ranges (Table 2). Density of edge habitat within the home range was the best predictor of home range size (Table 2) and was negatively related to the size of breeding season home ranges (β = -5.26 ± 1.48; Fig 2).

Table 1. Home range size (95% volume contour) for radio-marked female sharp-tailed grouse monitored in the 3 grazing systems during the breeding seasons of 2016–2018.

Females were assigned to the grazing system containing ≥ 60% of their home range or were considered to use multiple systems if no one system accounted for ≥ 60% of their home range.

Grazing System	# Females	Mean area (ha) ± SE	Min. area (ha)	Max area (ha)
Rest-rotation	47	557 ± 94	63.81	3717.45
Summer rotation	44	361 ± 39	86.13	1198.89
Season-long	36	408 ± 43	57.51	1103.58
Multiple systems	15	838 ± 179	191.43	2265.66

Open in a new tab

Fig 1 — An individual female was assigned to a grazing system according to the system containing ≥ 60% of the individual’s home range.

Table 2. Support for candidate models predicting the home range size of female sharp-tailed grouse during the breeding seasons of 2016–2018.

The percent of a home range containing either the rest-rotation or summer rotation system are measured in relation to the season-long system. The number of parameters (K), AIC_c values, ΔAIC_c values, model weights (w_i), and log-likelihoods are reported.

Model	K	AIC_c	ΔAIC_c	AIC_c w_i	LogLik
Edge density	3	2157.27	0.00	0.93	-1075.55
Dist. to grassland edge	3	2165.05	7.78	0.02	-1079.44
Nest outcome	3	2165.25	7.98	0.02	-1079.54
Null	2	2166.80	9.53	0.01	-1081.36
Year	3	2167.47	10.20	0.01	-1080.65
% Rest-rotation	3	2167.71	10.43	0.01	-1080.77
Stocking rate	3	2168.12	10.84	0.00	-1080.97
% Summer rotation	3	2168.14	10.87	0.00	-1080.98
Dist. to lek	3	2168.65	11.38	0.00	-1081.24
Dist. to road	3	2168.73	11.46	0.00	-1081.28
Dist. to oil pad	3	2168.84	11.57	0.00	-1081.33
Prop. grassland	3	2168.88	11.61	0.00	-1081.36

Open in a new tab

At the second order, breeding season home range selection was best predicted by the proportion grassland (β = 0.48 ± 0.13), the proportion wooded draws (β = 0.30 ± 0.11), distance to oil pad (β = 0.32 ± 0.11), and the proportion rest-rotation (β = 0.24 ± 0.10) within the home range (Table 3 and S2 Table). The proportion grassland had the strongest effect based on scaled coefficients, but all variables had positive effects on home range selection. The relative probability of home range selection increased with more grassland, more wooded draws, further from oil pads, and in pastures managed with rest-rotation grazing (Fig 3).

Table 3. Support for candidate models predicting second order selection, or home range selection, of female sharp-tailed grouse during the breeding seasons of 2016–2018.

Model	K	AIC_c	ΔAIC_c	AIC_c w_i	LL
% Grassland + % wooded draws + dist. to oil pad + % rest-rotation	5	2322.10	0.00	0.44	-1156.02
% Grassland + % wooded draws + dist. to oil pad + % rest-rotation + current stocking rate	6	2323.92	1.83	0.18	-1155.92
% Grassland + % wooded draws + dist. to oil pad + % rest-rotation + previous stocking rate	6	2324.08	1.98	0.17	-1156.00
% Grassland + % wooded draws + dist. to oil pad	4	2326.07	3.98	0.06	-1159.02
% Grassland + % wooded draws + dist. to oil pad + % summer rotation	5	2326.27	4.17	0.06	-1158.11
% Grassland + % wooded draws + dist. to oil pad + current stocking rate	5	2327.64	5.54	0.03	-1158.79
% Grassland + % wooded draws + dist. to oil pad + previous stocking rate	5	2328.07	5.97	0.02	-1159.01
% Grassland + % wooded draws + dist. to oil pad + % summer rotation + current stocking rate	6	2328.07	5.97	0.02	-1158.00
% Grassland + % wooded draws + dist. to oil pad + % summer rotation + previous stocking rate	6	2328.28	6.19	0.02	-1158.10
% Rest-rotation	2	2334.22	12.12	0.00	-1165.10
% Rest-rotation + previous stocking rate	3	2335.96	13.87	0.00	-1164.97
% Rest-rotation + current stocking rate	3	2335.99	13.89	0.00	-1164.98
Null	1	2336.38	14.29	0.00	-1167.19
% Summer rotation	2	2338.20	16.10	0.00	-1167.09
Current stocking rate	2	2338.38	16.28	0.00	-1167.18
Previous stocking rate	2	2338.39	16.29	0.00	-1167.19
% Summer rotation + current stocking rate	3	2340.17	18.07	0.00	-1167.07
% Summer rotation + previous stocking rate	3	2340.20	18.10	0.00	-1167.09

Open in a new tab

At the third order, preliminary analyses suggested that a spatial scale of 1,300 m for grassland, 1,300 m for wooded draws, 500 m for cropland, and 1,000 m for edge density represented the scale of strongest female habitat selection (S3 Table). However, the proportion of grassland was correlated with both the proportion of cropland and the density of edge habitat (S4 Table), so only proportion grassland was used in the full model. In the full analysis, 95% credible intervals for all variables overlapped zero, suggesting no significant selection at the population level (Fig 4). However, variability in selection as measured by the variation in individual-specific slopes for each predictor variable was high, indicating large differences in individual habitat selection (Fig 5). Individuals were selecting for and against habitat variables in equal numbers (Fig 6), resulting in no population-level selection. Nevertheless, selection varied linearly with availability suggesting that habitat use was proportional to availability (Fig 6).

Symbols represent individual females that selected against (gray circle), selected for (gray triangle), or displayed no significant selection (white square) for each variable and the diagonal represents proportional resource use.

Discussion

High individual variability in both home range size and third-order habitat selection of female sharp-tailed grouse outweighed any potential population-level trends. When selecting home ranges, females strongly selected for multiple landscape features, whereas third-order selection within home ranges was highly variable among individuals but proportional to availability, which suggests highly plastic habitat use within the population at this scale. While grouse selected for pastures managed with rest-rotation grazing when selecting a home range, we found no evidence for selection based on grazing management when choosing locations within home ranges.

Home range sizes in our study were on average larger and more variable than those previously reported for sharp-tailed grouse, although previous studies were limited by sample size and often included male grouse [41, 72, 73]. Previous estimates of home ranges for sharp-tailed grouse have come primarily from shrub-steppe or forested regions and our home range estimates are more in line with those from prairie-chickens in the Great Plains that had larger home ranges with more variation among individuals [36, 43, 74]. Home range size was negatively related to the density of edge habitat, suggesting that females could use a smaller area to meet their basic needs in more diverse habitats. At this scale, females strongly selected for grassland, which is consistent with previous studies finding both general selection for grassland [42–44] and that increased cropland on the landscape decreased adult survival in our study area [75], although the negative relationship between home range size and edge density and selection for wooded draws may suggest that other habitat types are important to female grouse during the breeding season. In addition, females selected home ranges that were further from oil pads, which supports previous research that has found consistently negative effects of energy development on grouse [47, 49, 50, 76].

When choosing a home range, most females selected for pastures managed with rest-rotation grazing but showed no selection for either grazing system or stocking rate when selecting habitat within the home range. Our results corroborate previous research finding that greater prairie-chickens strongly selected for areas managed with a heterogeneity-based fire-grazing management system [36]. However, previous research in our study area found that grazing system was not strongly linked to nest survival or adult female survival, important demographic parameters influencing grouse populations [56, 77]. Thus, selection for rest-rotation pastures did not equate to improved fitness for nesting sharp-tailed grouse. We found no evidence that space use of sharp-tailed grouse was influenced by stocking rate, which conflicts with previous studies that have documented consistently negative effects of high stocking rates on prairie-chickens [17, 18, 31]. Stocking rates in our study area were considered light to moderate by NRCS standards though [78] and it is possible that selection may only be apparent across a broader range of stocking rates.

In contrast to selection of home ranges, we found no evidence of selection for habitat features within the home range and our results conflict with previous studies observing small-scale selection based on vegetation features [39, 41, 79]. Our habitat variables consisted only of remotely-sensed data, however, and did not include fine-scale measures of vegetation structure or composition; a related study found that small-scale vegetation was critical to nest survival and selection [56]. Furthermore, grouse habitat selection based on both landcover and anthropogenic disturbance such as roads has been shown to vary among studies and even sites within a single study [36, 39, 42, 43, 46, 47], which can complicate population-level interpretation of effects.

While there was no evidence for population-level selection at the third order, there was significant individual variation in habitat selection within the home range, suggesting that fine-scale habitat selection may be flexible or less important after a home range has been selected. Significant individual variation is consistent with previous work suggesting that habitat selection can vary by year or weather conditions and can vary across spatial scales [36, 43]. Taken together, this suggests that generalized habitat recommendations across sites and related species may not be appropriate. Given that habitat use did not vary with availability, the variation in habitat selection behavior suggests a high degree of plasticity in the population [80]. If individual differences are consistent across time, then those differences can represent alternative approaches that evolved to respond to a variable environment [81, 82]. Regardless, if individual differences are correlated with fitness, individual variation can have ecological and evolutionary implications [83, 84]. Future research should explore both the consistency in individual differences in resource selection across time and the link between individual differences and fitness.

Conclusions

At a broad scale, female sharp-tailed grouse exhibited strong selection, particularly for grassland, when choosing a home range, but showed no selection for habitat or management variables when selecting locations within their home ranges. Females did select home ranges in pastures managed with rest-rotation grazing, but selection was not related to improved reproductive success or survival [56, 77]. Given observed individual variation, the choice of grazing system may not have a significant influence on sharp-tailed grouse populations in the northern mixed-grass prairie when stocking rates are low to moderate. Importantly, female sharp-tailed grouse exhibited strong individual differences in both home range size and third-order habitat selection that outweighed any potential population-level trends, suggesting that specific management recommendations are inappropriate, particularly across large spatial scales. Collectively, our results suggest that maintaining large intact grasslands on the landscape will have higher conservation value for sharp-tailed grouse than prescriptive livestock grazing systems.

Supporting information

S1 Appendix. Example code for Bayesian logistic regression model evaluating third order habitat selection using the R package R-INLA.

(DOCX)

Click here for additional data file.^{(16.2KB, docx)}

S1 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of grassland measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S2 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of wooded draws measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S3 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of row crop agriculture measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S4 Fig. Simulation results evaluating the number of available points necessary for convergence of edge density measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S5 Fig. Simulation results evaluating the number of available points necessary for convergence of variables measured at a single scale.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S1 Table. Support for candidate models predicting the relationship between the number of locations per female and home range size of female sharp-tailed grouse during the breeding seasons of 2016–2018.

The number of parameters (K), AIC_c values, AIC_c values, model weights (w_i), and log-likelihoods are reported.

(DOCX)

Click here for additional data file.^{(16.6KB, docx)}

S2 Table. Support for candidate models predicting the relationship between habitat and anthropogenic variables and home range selection of female sharp-tailed grouse during the breeding seasons of 2016–2018.

The number of parameters (K), AIC_c values, AIC_c values, model weights (w_i), and log-likelihoods are reported.

(DOCX)

Click here for additional data file.^{(15.4KB, docx)}

S3 Table. Support for models predicting the spatial grain of each landcover variable that best predicts sharp-tailed grouse habitat selection during the breeding seasons of 2016–2018, based on Deviance Information Criteria (DIC).

(DOCX)

Click here for additional data file.^{(17.5KB, docx)}

S4 Table. Multicollinearity results for management and landscape variables in the full third order resource selection analysis evaluating habitat selection within the home range for sharp-tailed grouse during the breeding seasons of 2016–2018.

(DOCX)

Click here for additional data file.^{(17.4KB, docx)}

S1 Data

(ZIP)

Click here for additional data file.^{(10.5MB, zip)}

Acknowledgments

Our study was made possible by the generous cooperation of private landowners who allowed us access to their land and the help of many field technicians who collected data. Dr. Jeff Mosley and Dr. Jay Rotella, and Montana Fish, Wildlife and Parks (MT FWP) staff John Ensign and Melissa Foster, provided useful guidance throughout the study. Associate Editor Dr. Walter, Dr. Joseph Smith, and an anonymous reviewer provided comments that improved the manuscript.

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

This research was supported by the general sale of hunting and fishing licenses in Montana and matching funds under Federal Aid in Wildlife Restoration grant W-161-R1 awarded to LIB and LBM.

References

1.Johnson MD. Measuring habitat quality: a review. The Condor. 2007;109(3):489–504. [Google Scholar]
2.Jones J. Habitat selection studies in avian ecology: a critical review. The Auk. 2001;118(2):557–62. [Google Scholar]
3.Boyce MS, Johnson CJ, Merrill EH, Nielsen SE, Solberg EJ, Van Moorter B. Can habitat selection predict abundance? J Anim Ecol. 2016;85(1):11–20. [DOI] [PubMed] [Google Scholar]
4.Beyer HL, Haydon DT, Morales JM, Frair JL, Hebblewhite M, Mitchell M, et al. The interpretation of habitat preference metrics under use–availability designs. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010;365(1550):2245–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.McLoughlin PD, Morris DW, Fortin D, Vander Wal E, Contasti AL. Considering ecological dynamics in resource selection functions. J Anim Ecol. 2010;79(1):4–12. [DOI] [PubMed] [Google Scholar]
6.Hebblewhite M, Merrill EH. Trade-offs between predation risk and forage differ between migrant strategies in a migratory ungulate. Ecology. 2009;90(12):3445–54. [DOI] [PubMed] [Google Scholar]
7.Wiens JA, Milne BT. Scaling of ‘landscapes’ in landscape ecology, or, landscape ecology from a beetle’s perspective. Landscape Ecol. 1989;3(2):87–96. [Google Scholar]
8.Hagen CA, Sandercock BK, Pitman JC, Robel RJ, Applegate RD. Spatial variation in lesser prairie-chicken demography: a sensitivity analysis of population dynamics and management alternatives. J Wildl Manage. 2009;73(8):1325–32. [Google Scholar]
9.McNew LB, Gregory AJ, Sandercock BK. Spatial heterogeneity in habitat selection: Nest site selection by greater prairie-chickens. J Wildl Manage. 2013;77(4):791–801. [Google Scholar]
10.Allen AM, Månsson J, Jarnemo A, Bunnefeld N. The impacts of landscape structure on the winter movements and habitat selection of female red deer. European Journal of Wildlife Research. 2014;60(3):411–21. [Google Scholar]
11.Krausman PR, Naugle DE, Frisina MR, Northrup R, Bleich VC, Block WM, et al. Livestock grazing, wildlife habitat, and rangeland values. Rangelands. 2009;31(5):15–9. [Google Scholar]
12.Alkemade R, Reid RS, van den Berg M, de Leeuw J, Jeuken M. Assessing the impacts of livestock production on biodiversity in rangeland ecosystems. Proc Natl Acad Sci. 2013;110(52):20900–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Fleischner TL. Ecological costs of livestock grazing in western North America. Conserv Biol. 1994;8(3):629–44. [Google Scholar]
14.Boyd CS, Beck JL, Tanaka JA. Livestock grazing and sage-grouse habitat: impacts and opportunities. Journal of Rangeland Applications. 2014;1:58–77. [Google Scholar]
15.Fuhlendorf SD, Engle DM. Restoring heterogeneity on rangelands: Ecosystem management based on evolutionary grazing patterns. Bioscience. 2001;51(8):625–32. [Google Scholar]
16.Coppedge BR, Fuhlendorf SD, Harrell WC, Engle DM. Avian community response to vegetation and structural features in grasslands managed with fire and grazing. Biol Conserv. 2008;141(5):1196–203. [Google Scholar]
17.McNew LB, Winder VL, Pitman JC, Sandercock BK. Alternative rangeland management strategies and the nesting ecology of greater prairie-chickens. Rangeland Ecol Manage. 2015;68(3):298–304. [Google Scholar]
18.Winder VL, McNew LB, Pitman JC, Sandercock BK. Effects of rangeland management on survival of female greater prairie-chickens. J Wildl Manage. 2018;82(1):113–22. [Google Scholar]
19.Engle DM, Fuhlendorf SD, Roper A, Leslie DM Jr. Invertebrate community response to a shifting mosaic of habitat. Rangeland Ecol Manage. 2008;61(1):55–62. [Google Scholar]
20.Limb RF, Fuhlendorf SD, Engle DM, Weir JR, Elmore RD, Bidwell TG. Pyric–herbivory and cattle performance in grassland ecosystems. Rangeland Ecol Manage. 2011;64(6):659–63. [Google Scholar]
21.Fuhlendorf SD, Townsend DE II, Elmore RD, Engle DM. Pyric-herbivory to promote rangeland heterogeneity: evidence from small mammal communities. Rangeland Ecol Manage. 2010;63(6):670–8. [Google Scholar]
22.Lipsey MK, Naugle DE. Precipitation and soil productivity explain effects of grazing on grassland songbirds. Rangeland Ecol Manage. 2017;70(3):331–40. [Google Scholar]
23.Holechek JL, Gomez H, Molinar F, Galt D. Grazing studies: what we’ve learned. Rangelands. 1999;21:12–6. [Google Scholar]
24.Schieltz JM, Rubenstein DI. Evidence based review: positive versus negative effects of livestock grazing on wildlife. What do we really know? Environmental Research Letters. 2016;11(11):113003. [Google Scholar]
25.Briske DD, Derner J, Brown J, Fuhlendorf SD, Teague W, Havstad K, et al. Rotational grazing on rangelands: reconciliation of perception and experimental evidence. Rangeland Ecol Manage. 2008;61(1):3–17. [Google Scholar]
26.Frisina MR. Grazing private and public land to improve the Fleecer Elk Winter Range. Rangelands. 1991;13(6):291–4. [Google Scholar]
27.Montana Department of Fish Wildlife and Parks. Upland game bird enhancement program strategic plan. Helena, MT, USA: 2011. [Google Scholar]
28.Sliwinski M, Burbach M, Powell L, Schacht W. Factors influencing ranchers’ intentions to manage for vegetation heterogeneity and promote cross-boundary management in the northern Great Plains. Ecol Soc. 2018;23(4):45–62. [Google Scholar]
29.Hormay AL. Principles of rest-rotation grazing and multiple-use land management: US Department of the Interior, Bureau of Land Management; 1970. [Google Scholar]
30.Hormay AL, Evanko AB. Rest-rotation grazing: a management system for bunchgrass ranges: California Forest and Range Experiment Station; 1958. [Google Scholar]
31.Kraft JD. Vegetation characteristics and lesser prairie chicken responses to land cover types and grazing management in western Kansas [Thesis]. Manhattan, KS, USA: Kansas State University; 2016.
32.Roersma SJ. Nesting and brood rearing ecology of plains sharp-tailed grouse (Tympanuchus phasianellus jarnesi) in a mixed-grass/fescue ecoregion of southem Alberta [Thesis]. Winnipeg, MB, Canada: University of Manitoba; 2001.
33.Hillman CN, Jackson WW. The sharp-tailed grouse in South Dakota: South Dakota Department of Game, Fish and Parks; 1973. [Google Scholar]
34.Poiani KA, Merrill MD, Chapman KA. Identifying conservation-priority areas in a fragmented Minnesota landscape based on the umbrella species concept and selection of large patches of natural vegetation. Conserv Biol. 2001;15(2):513–22. [Google Scholar]
35.Bergerud AT, Gratson MW. Survival and breeding strategies of grouse In: Bergerud AT, Gratson MW, editors. Adaptive strategies and population ecology of northern grouse. Minneapolis, MN, USA: University of Minnesota Press; 1988. p. 473–577. [Google Scholar]
36.Winder VL, McNew LB, Pitman JC, Sandercock BK. Space use of female greater prairie-chickens in response to fire and grazing interactions. Rangeland Ecol Manage. 2017;70(2):165–74. [Google Scholar]
37.Lautenbach JM, Plumb RT, Robinson SG, Hagen CA, Haukos DA, Pitman JC. Lesser prairie-chicken avoidance of trees in a grassland landscape. Rangeland Ecol Manage. 2017;70(1):78–86. [Google Scholar]
38.Dettenmaier SJ, Messmer TA, Hovick TJ, Dahlgren DK. Effects of livestock grazing on rangeland biodiversity: A meta-analysis of grouse populations. Ecology and Evolution. 2017; 10.1002/ece3.3287 [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Goddard AD, Dawson RD, Gillingham MP. Habitat selection by nesting and brood-rearing sharp-tailed grouse. Canadian Journal of Zoology. 2009;87(4):326–36. [Google Scholar]
40.Marks JS, Marks VS. Habitat selection by Columbian sharp-tailed grouse in west-central Idaho: USDI Bureau of Land Management; 1987. [Google Scholar]
41.Saab VA, Marks JS. Summer habitat use by Columbian sharp-tailed grouse in western Idaho. Great Basin Nat. 1992;52(2):166–73. [Google Scholar]
42.Ryan MR, Burger LW, Jones DP, Wywialowski AP. Breeding ecology of greater prairie-chickens (Tympanuchus cupido) in relation to prairie landscape configuration. Am Midl Nat. 1998;140:111–22. [Google Scholar]
43.Winder VL, Carrlson KM, Gregory AJ, Hagen CA, Haukos DA, Kesler DC, et al. Factors affecting female space use in ten populations of prairie chickens. Ecosphere. 2015;6(9):1–17. [Google Scholar]
44.Niemuth ND. Identifying landscapes for greater prairie chicken translocation using habitat models and GIS: a case study. Wildl Soc Bull. 2003;31:145–55. [Google Scholar]
45.Manzer DL. Sharp-tailed grouse breeding success, survival, and site selection in relation to habitat measured at multiple scales [Dissertation]. Edmonton, AB, Canada: University of Alberta; 2004.
46.Cope MG. Distribution, habitat selection and survival of transplanted Columbian sharp-tailed grouse (Tympanuchus phasianellus columbianus) in the Tobacco Valley, Montana [Thesis]. Bozeman, MT, USA: Montana State University; 1992.
47.Winder VL, McNew LB, Gregory AJ, Hunt LM, Wisely SM, Sandercock BK. Space use by female greater prairie-chickens in response to wind energy development. Ecosphere. 2014;5(1):1–17. [Google Scholar]
48.Stonehouse KF, Shipley LA, Lowe J, Atamian MT, Swanson ME, Schroeder MA. Habitat selection and use by sympatric, translocated greater sage-grouse and Columbian sharp-tailed grouse. J Wildl Manage. 2015;79(8):1308–26. [Google Scholar]
49.Dinkins JB, Conover MR, Kirol CP, Beck JL, Frey SN. Greater Sage-Grouse (Centrocercus urophasianus) select habitat based on avian predators, landscape composition, and anthropogenic features. The Condor: Ornithological Applications. 2014;116(4):629–42. [Google Scholar]
50.Hagen CA, Pitman JC, Loughin TM, Sandercock BK, Robel RJ, Applegate RD. Impacts of anthropogenic features on habitat use by Lesser Prairie-Chickens. Studies in Avian Biology. 2011;39:63–75. [Google Scholar]
51.Mysterud A, Ims RA. Functional responses in habitat use: Availability influences relative use in trade-off situations. Ecology. 1998;79(4):1435–41. [Google Scholar]
52.Johnson DH. The comparison of usage and availability measurements for evaluating resource preference. Ecology. 1980;61(1):65–71. [Google Scholar]
53.McDonald L, Erickson W, Boyce M, Alldredge J. Modeling vertebrate use of terrestrial resources In: Silvy NJ, editor. The Wildlife Techniques Manual Volume 1: Research. Baltimore, MD, USA: Johns Hopkins University Press; 2012. p. 410–29. [Google Scholar]
54.Curtin CG, Sayre NF, Lane BD. Transformations of the Chihuahuan Borderlands: grazing, fragmentation, and biodiversity conservation in desert grasslands. Environmental Science & Policy. 2002;5(1):55–68. [Google Scholar]
55.LANDFIRE. LANDFIRE Existing Vegetation Type layer. U S Department of Interior, Geological Survey http://wwwlandfiregov/indexphp [2016, November 18]. 2013.
56.Milligan MC, Berkeley L, McNew LB. Effects of rangeland management on the nesting ecology of sharp-tailed grouse. Rangeland Ecol Manage. 2020;73:128–37. [Google Scholar]
57.Seaman DE, Millspaugh JJ, Kernohan BJ, Brundige GC, Raedeke KJ, Gitzen RA. Effects of sample size on kernel home range estimates. J Wildl Manage. 1999;63:739–47. [Google Scholar]
58.Worton BJ. Kernel methods for estimating the utilization distribution in home-range studies. Ecology. 1989;70(1):164–8. [Google Scholar]
59.Manly B, McDonald L, Thomas D, McDonald T, Erickson W. Resource selection by animals: statistical analysis and design for field studies. Kluwer, Nordrecht, The Netherlands: 2002. [Google Scholar]
60.Northrup JM, Hooten MB, Anderson CR Jr, Wittemyer G. Practical guidance on characterizing availability in resource selection functions under a use–availability design. Ecology. 2013;94(7):1456–63. [DOI] [PubMed] [Google Scholar]
61.Burnham KP, Anderson DR. A practical information-theoretic approach. New York, NY, USA: Springer; 2002. [Google Scholar]
62.Arnold TW. Uninformative parameters and model selection using Akaike’s Information Criterion. J Wildl Manage. 2010;74(6):1175–8. [Google Scholar]
63.McGarigal K, Cushman SA, Ene E. FRAGSTATS v4: Spatial pattern analysis program for categorical and continuous maps. University of Massachusetts, Amherst: 2012;http://www.umass.edu/landeco/research/fragstats/fragstats.html. [Google Scholar]
64.Laforge MP, Vander Wal E, Brook RK, Bayne EM, McLoughlin PD. Process-focused, multi-grain resource selection functions. Ecol Model. 2015;305:10–21. [Google Scholar]
65.Gillies CS, Hebblewhite M, Nielsen SE, Krawchuk MA, Aldridge CL, Frair JL, et al. Application of random effects to the study of resource selection by animals. J Anim Ecol. 2006;75(4):887–98. [DOI] [PubMed] [Google Scholar]
66.Thomas DL, Johnson D, Griffith B. A Bayesian random effects discrete-choice model for resource selection: Population-level selection inference. J Wildl Manage. 2006;70(2):404–12. [Google Scholar]
67.Laforge MP, Brook RK, van Beest FM, Bayne EM, McLoughlin PD. Grain-dependent functional responses in habitat selection. Landscape Ecol. 2016;31(4):855–63. [Google Scholar]
68.Holbrook JD, Squires JR, Olson LE, DeCesare NJ, Lawrence RL. Understanding and predicting habitat for wildlife conservation: the case of Canada lynx at the range periphery. Ecosphere. 2017;8(9):e01939. [Google Scholar]
69.Muff S, Signer J, Fieberg J. Accounting for individual-specific variation in habitat-selection studies: Efficient estimation of mixed-effects models using Bayesian or frequentist computation. J Anim Ecol. 2019;89(1):80–92. [DOI] [PubMed] [Google Scholar]
70.Fong Y, Rue H, Wakefield J. Bayesian inference for generalized linear mixed models. Biostatistics. 2010;11(3):397–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Wang X, Ryan YY, Faraway JJ. Bayesian regression modeling with INLA. Boca Raton, LA, USA: Chapman and Hall/CRC; 2018. [Google Scholar]
72.Boisvert JH, Hoffman RW, Reese KP. Home range and seasonal movements of Columbian sharp-tailed grouse associated with conservation reserve program and mine reclamation. West N Am Nat. 2005;65:36–44. [Google Scholar]
73.Christenson CD. Nesting and brooding characteristics of sharp-tailed grouse (Pedioecetes phasianellus jamesi) in southwestern North Dakota [Thesis]. Grand Forks, ND, USA: University of North Dakota; 1970.
74.Patten MA, Pruett CL, Wolfe DH. Home range size and movements of greater prairie-chickens. Studies in Avian Biology. 2011;39:51–62. [Google Scholar]
75.Milligan MC, Berkeley LI, McNew LB. Effects of rangeland management on the survival of adult sharp-tailed grouse. J Wildl Manage. In review. [Google Scholar]
76.Holloran MJ, Kaiser RC, Hubert WA. Yearling greater sage-grouse response to energy development in Wyoming. J Wildl Manage. 2010;74(1):65–72. [Google Scholar]
77.Milligan MC. Effects of grazing management on sharp-tailed grouse ecology in mixed-grass prairies [Dissertation]. Bozeman, MT, USA: Montana State University; 2019.
78.Montana UNRCS-. Technical Guide—R058AE001MT. 1983.
79.Matthews TW, Tyre AJ, Taylor JS, Lusk JJ, Powell LA. Habitat selection and brood survival of greater prairie-chickens. Studies in Avian Biology. 2011;39:179–91. [Google Scholar]
80.Sih A, Bell AM, Johnson JC, Ziemba RE. Behavioral syndromes: an integrative overview. The Quarterly Review of Biology. 2004;79(3):241–77. [DOI] [PubMed] [Google Scholar]
81.McLoughlin PD, Boyce MS, Coulson T, Clutton-Brock T. Lifetime reproductive success and density-dependent, multi-variable resource selection. Proceedings of the Royal Society B: Biological Sciences. 2006;273(1593):1449–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Leclerc M, Dussault C, St-Laurent M-H. Behavioural strategies towards human disturbances explain individual performance in woodland caribou. Oecologia. 2014;176(1):297–306. [DOI] [PubMed] [Google Scholar]
83.Réale D, Dingemanse NJ, Kazem AJ, Wright J. Evolutionary and ecological approaches to the study of personality. Philosophical Transactions of the Royal Society B. 2010;365:3937–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Sih A, Cote J, Evans M, Fogarty S, Pruitt J. Ecological implications of behavioural syndromes. Ecol Lett. 2012;15(3):278–89. [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0233756.r001

Decision Letter 0

W David Walter

13 Feb 2020

PONE-D-19-33753

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

PLOS ONE

Dear Dr. Milligan,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Mar 29 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

W. David Walter, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

In addition to suggestions provided by the reviewers, please pay particular attention to those comments on the details of your second-order selection. Please present those details requested, and, in your potential resubmission, please address why 2 different methods were used for second-order selection (compositional analysis) and third-order selection (RSF) because they present 2 very different attempts at assessing habitat selection.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Our internal editors have looked over your manuscript and determined that it is within the scope of our Biodiversity Conservation Call for Papers. This collection of papers is headed by a team of Guest Editors for PLOS ONE (https://collections.plos.org/s/biodiversity). The Collection will encompass a diverse range of research articles on biodiversity conservation, including conservation of endangered species. Additional information can be found on our announcement page: https://collections.plos.org/s/biodiversity

If you would like your manuscript to be considered for this collection, please let us know in your cover letter and we will ensure that your paper is treated as if you were responding to this call. If you would prefer to remove your manuscript from collection consideration, please specify this in the cover letter.

3. In your Methods section, please provide additional location information of the study area, including geographic coordinates for the data set if available.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Review of “Habitat selection of female” by Milligan et al.

This is well-written manuscript on a topic and species of considerable management interest. Specific comments follow, referenced by line number.

Line 14: Habitat selection affects much more than abundance, as the authors point out in the introduction. It might be helpful to add or change a word in this line to acknowledge that (something along the lines of performance, viability, or persistence).

Line 16: Consider “…thus influencing habitat selection and population size” (or performance, viability, or persistence).

Line 20: Mention where the research took place.

Line 29: I suggest deleting the word “moderate,” simply stating that we want stocking rates that maintain native rangelands.

Line 65: Here and throughout the paper, many of the cited references are about greater prairie-chickens, not sharp-tailed grouse. I realized they are congenerics and that more information is available about chickens than sharptails, but I feel like there is a bit of a bait-and-switch by mentioning sharptails in some instances and grouse (accompanied by chicken papers) in others. Many solid papers about sharptail habitat selection and reproduction are not cited in this manuscript. In addition, management, environment, and social factors differ considerably between the authors’ study area and the Flint Hills, and the paper would be improved if more background was provided for the study area.

Line 107: Clarification on some of the metrics would be good. The paper mentions “unfragmented grassland patches,” but that’s a bit of an oxymoron to me. Landscape ecologists think of habitat patches that, depending on patch size and arrangement, constitute a fragmented or unfragmented landscape. If patchiness is determined by grazing regime, then it would be more clear to refer to heterogeneity or something similar.

Line 199: A major shortcoming to compositional analysis is that it only considers composition of the landscape and doesn’t consider things like proximity to roads, wells, edge, etc. I think an RSF-based approach (similar to the next section but with a different set of availabilities) would be much more informative.

Line 214: How the available points were determined should be described, as their number and spatial distribution can substantially affect results.

Line 220: Perhaps “scale” instead of “spatial grain,” which is frequently used to describe the resolution of landcover data.

Line 222: Good!

Line 234: I’m confused, as line 210 indicates that Design II of Manly et al. was used. Manly et al. states that “With design III studies the use and availability of resource units is measured separately for each animal.” Random effects as espoused by reference 62 (Gillies et al.) are used when “individual animals are monitored and pooled to estimate population level effects.” I’ll readily admit that I’m easily confused, but I think some clarification as to what is being assessed and how it was done is necessary.

Lines 279-280: Table 2 also confuses me, as it provides cumulative wi, but all the models (other than the null) only have one variable as evidenced by a k of 3. I suggest deleting the column with the cumulative wi or explaining the purpose of that column.

Lines 289-290: The “Total” row should be eliminated. Simply state that there was a total of 142 females. The other values appear to be means, except for Max area, which is not the mean but a copy of the value for rest-rotation.

Lines 292-293: The compositional analysis doesn’t really provide a whole lot of information that would be helpful for management.

Line 302: As mentioned before, “grain size” should be replaced. Here it suggests an area of 1300m x 1300m, but in reality, 1300 m is the radius of a circle with an area 3.14 times larger than the implied square.

Line 315: What about the signs of the relationships in Figure 3?

Line 319: Figure 4 indicates that females select for proximity to roads, which is quite different from the abstract, which states that females select for large intact grasslands.

Lines 348-350: Good point.

Lines 388-401: This paper focuses on habitat selection, but it would be useful if the authors provided more information from other aspects of their research regarding how reproductive success was influenced by habitat, and if the same factors popped out in both analyses.

Reviewer #2: Overall, this is a nicely set-up study that seeks to answer some pertinent, long-unanswered questions about effects of livestock management on a sharp-tailed grouse and their breeding habitat. The paper is generally well written and the conclusions appear well-supported by the results. I do, however, have a couple substantial issues with the analysis and its interpretation. These concerns are related and may be cleared up fairly quickly.

First, the third-order resource selection model seems to be misspecified. Random effects seem to have been included in a way that makes no biological or mathematical sense.

Second, the considerable space is given to interpreting these random effects as ‘individual variation’ in behavior, which is an interpretation that the model structure—if I’m seeing it right—doesn’t support.

I’m having a hard time knowing what I’m looking at in the results because the methods lack important details about the model. I would like to see the methods clarified by including full model statements (including priors—the reader is told that the priors are vague or uninformative, but has no way of verifying that for themselves because they are not provided) and, ideally, JAGS model code in Supplemental Materials.

If I am correct in my understanding that random intercepts, but not random slopes, were included in this RSF model, then I would like to see this corrected by either a) removing the random intercepts and any references to variation among individuals in their behavior, since this wasn’t a feature of the model, or b) adding random slopes and interpreting those parameters as individual variation.

Option A would take hardly any time at all, whereas Option B would entail a bit of work. I’m not set on either one, but I do tend to think a random-intercept/random-slope model is more appropriate for this study design.

Best regards,

Joe Smith

Line-by-line comments are provided as an attachment.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Joseph Smith

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: milligan sharptail nest selection review.pdf

Click here for additional data file.^{(121KB, pdf)}

PLoS One. 2020 Jun 4;15(6):e0233756. doi: 10.1371/journal.pone.0233756.r002

Author response to Decision Letter 0

19 Mar 2020

Reviewer #1: Review of “Habitat selection of female” by Milligan et al.

This is well-written manuscript on a topic and species of considerable management interest. Specific comments follow, referenced by line number.

Author response: We agree and have added additional wording to convey that habitat selection affects more than just abundance (line 18).

Line 16: Consider “…thus influencing habitat selection and population size” (or performance, viability, or persistence).

Author response: We have changed the wording as suggested (lines 20-21).

Line 20: Mention where the research took place.

Author response: We have added the location of our study area (lines 25-26).

Line 29: I suggest deleting the word “moderate,” simply stating that we want stocking rates that maintain native rangelands.

Author response: We feel that it is important to include the word moderate to avoid overstating our conclusions, particularly given the number of studies that have found negative effects of high stocking rates (Briske et al. 2008, McNew et al. 2015, Kraft 2016, Winder et al. 2018). Stocking rates throughout our study area were generally moderate and it is possible that wildlife responses may be different at higher stocking rates. Therefore, we have kept the wording to qualify our conclusions.

Author response: We agree that it would be better to have more citations from sharp-tailed grouse papers and we have tried to include them wherever possible (e.g., Christenson 1970, Hillman and Jackson 1973, Marks and Marks 1987, Cope 1992, Saab and Marks 1992, Roersma 2001, Manzer 2004, Boisvert et al. 2005, Goddard et al. 2009, Stonehouse et al. 2015). However, there is no previous research on sharp-tailed grouse habitat selection in relation to grazing management, which is why we had to rely on the prairie-chicken literature, particularly studies in the Flint Hills, which have evaluated the relationship between grouse and grazing in depth.

Author response: We agree that our use of the term “fragmented” is confusing because it is scale-dependent and grasslands are naturally heterogeneous. In this context, we were referring to grassland fragmentation caused by other land uses, such as cropland. We have clarified that we hypothesized that grouse would select for large grassland patches and removed reference to “unfragmented patches” (line 111).

Author response: We agree that one major limitation of compositional analysis is that it cannot accommodate continuous predictors. However, compositional analysis has several advantages, particularly that it assesses proportional habitat use and so avoids the unit-sum problem and that it correctly uses the individual home range as the sample unit. Although it was originally proposed in 1993, compositional analysis is still widely used in published articles today (over 50 articles citing the original paper on Google Scholar since 2019, for example). In addition, it was an appropriate approach given our objectives to evaluate home range selection relative to landcover type and grazing system, both of which are categorical variables. We have clarified both the approach and its usefulness given our objectives in the methods section (lines 214-219).

Line 214: How the available points were determined should be described, as their number and spatial distribution can substantially affect results.

Author response: We have clarified that available points were sampled randomly through an individual’s home range (lines 222-224).

Line 220: Perhaps “scale” instead of “spatial grain,” which is frequently used to describe the resolution of landcover data.

Author response: We have changed the wording throughout to use “scale” instead of “spatial grain”.

Line 222: Good!

Author response: We stated that we used Design III of Manly et al (2002), where the “animals in a sample are radio-collared, and the relocations of an animal identify used resource units for that animal. . . The collection of available or unused resource units within an animal’s home range is sampled or censused.” We have included additional language to clarify that individual telemetry locations were used locations and we randomly sampled available locations within each individual’s home range (lines 220-225).

Author response: This is a good point and we have removed the column showing cumulative weights.

Author response: We have removed the “Total” row.

Lines 292-293: The compositional analysis doesn’t really provide a whole lot of information that would be helpful for management.

Author response: Given the strong selection for mixed-grass prairie, we feel that the compositional analysis provides valuable evidence for the importance of large grassland patches. We have clarified the language to stress that we mean large patches and not unfragmented patches, given that the term “fragmented” could mean multiple things in this context. This should clarify that our results suggest that conserving large grassland patches from conversion to other land uses is an appropriate and useful management strategy for the conservation of sharp-tailed grouse.

Author response: We have changed the wording throughout to use “scale” instead of “spatial grain”.

Line 315: What about the signs of the relationships in Figure 3?

Author response: We have redone our third order analysis and now included a graph showing the mean and 95% credible intervals for each population-level selection coefficient in the analysis which shows the direction and strength of the relationships (Fig 3).

Line 319: Figure 4 indicates that females select for proximity to roads, which is quite different from the abstract, which states that females select for large intact grasslands.

Author response: Although this relationship was very weak, it was no longer significant after we reanalyzed the data following the second reviewer’s comments.

Lines 348-350: Good point.

Author response: We have included references to other aspects of our research evaluating both adult survival and reproduction to place our results in context and inform the reader how other aspects of grouse ecology were influenced by the same or different habitat variables (lines 342-344, lines 349-350, lines 371-372).

First, the third-order resource selection model seems to be misspecified. Random effects seem to have been included in a way that makes no biological or mathematical sense.

Author response: Thank you for the excellent review and recommendations. We have followed Option B and redone our analysis with random slopes, which allows us to interpret those parameters as individual variation. We have also included model code in the Supplemental Materials so that a reader can fully understand how the model was specified.

Line-by-line comments:

Line 25: May need to remove reference to individual-level variation in habitat selection; see

detailed comments on methods below.

Author response: Given that we have now included random slopes in our model, we can make inferences regarding individual-level variation in habitat selection.

Line 39: I’m being nitpicky here, but this is such a vague, fluffy statement that it could mean

almost anything. Perhaps a specific example of how ‘conservation and management actions’

could ‘consider’ this information would imbue this statement with some needed substance.

Author response: We have removed this sentence.

Line 54: Make the speculative/hypothetical nature of this claim perfectly clear here. The two

provided citations provide no actual evidence that they function like patch-burn grazing. The

UGBEP makes a pretty unequivocal statement that they benefit upland birds, but provides no

evidence backing that up whatsoever. Frisina’s paper only deals with whether rotational grazing

can reduce conflict between cattle and elk on winter range, and never mentions heterogeneity.

Just make sure you don’t accidentally help propagate misinformation by leading the reader

(most of whom don’t check citations) to believe that there are studies that support this.

Line 61: Same comment here as above. Yes, they will have increased residual cover, but I don’t

think I’ve ever seen any evidence that RGS increase structural heterogeneity or create a

‘patchwork’ relative to other grazing systems (I haven’t searched exhaustively, though). We

actually found the opposite effect in Roundup, and it looks like heterogeneity in VOR was lowest

on RGS ranches in your study, too (Fig 6, random plots, in your recent REM paper--you could

even cite that). You could simply add a statement like, “but this hypothesis has not been

adequately tested.”

I think the purpose of RGS may have been lost in translation between range managers and

wildlifers. In Bailey and Brown (2011) Rangeland Ecol Manage 64:1–9, they write, “To alleviate

selective grazing across multiple scales, many rangeland managers have implemented

rotational grazing systems (RGS)...The potential benefits are projected to derive from a

decrease in area available to livestock at any given time (improved distribution); an increase in

stock density (increased uniformity of defoliation across species and communities)...”

So, rangeland managers may be quite aware that these systems are having the exact opposite

effect, at least in some settings.

Author response: We have changed the wording as suggested to emphasize that this is a hypothetical relationship (line 58).

Lines 70–71: This sentence is also very vague. What kind of ‘implications?’

Author response: We have changed the wording to be specific about the kind of implications (lines 73-74).

Line 185: Define an ‘abrupt change.’

Author response: We have changed the wording to more clearly define how we measured edge density (lines 189-191).

Line 200: I’d like to see more detail on your methods for the second-order resource selection

analysis, rather than just referring the reader to a couple citations. Which variables were

considered, how was model fitting and model selection accomplished? There’s no way I could

re-create your analysis with the details you give here.

Author response: We have included additional details on both variables and models considered in the analyses assessing home range size and second order selection. We included additional information on which models were evaluated and our model selection procedure for the home range size analysis and we have included further details on compositional analysis to make the process more transparent (lines 181-219).

Lines 232–234: “For all models, we used generalized linear mixed models in a Bayesian

framework with a logit-link and female ID as a random intercept to account for potential

autocorrelation among sampling points [62, 63].”

I’m pretty sure this isn’t doing what you intended. Given your particular availability sampling design, these random intercepts are unnecessary and shouldn’t affect your inference at all. You later (lines 253–255) describe σ 2 as representing individual-level variation in selection for various habitat variables, which is a bit different than the way you justify inclusion of random

intercepts here (autocorrelation in the response variable--not sure what that means in this case),

but this is not what random intercepts do, either.

Consider what the intercept represents in a simple, univariate logit-link resource selection

function where you’ve centered your covariate at zero. If you convert it to the probability scale,

it’s the predicted probability that a case/sample is a used case (1) rather than an available case

(0) at the mean value of the covariate. In a use-availability RSF, its value is going to reflect the

prevalence of use cases—your ratio of used to available samples. You determined that when

you decided how many random samples to include for each used sample. The intercept has no

biological meaning in an RSF, it is simply a reflection of the way the investigator sets up the

analysis.

Unless you included both random intercepts and random adjustments to the slope (coefficient for the covariate—e.g., ,where the j subscript indexes individuals, βj ~ N(β, σ2) from your citation (#69), then all you are doing by allowing the intercept to vary among individuals is allowing random adjustments to the overall (i.e., regardless of the covariate value) probability of use.

Given this is a design parameter, not a biological one, and in your study design you’ve fixed it at 1:15 = 0.0625 for all individuals in your dataset (this is why your y-axis tops out at about 0.07 in Fig. 4), there is no reason to expect that there will be any variation among individuals at all. By design, individuals do not vary in mean P(use) (i.e., prevalance), and you’ve fixed the regression coefficients, so they don’t vary in their responses to covariates. So what is the variation you’re modeling with these random effects? You can prove this to yourself by running this R code with your data:

require(here)

require(tidyr)

require(dplyr)

require(lme4)

df <- read.csv(here('data/Pts.15.FULL.csv'), stringsAsFactors = FALSE)

# check that prevalence is equal across individuals

df %>%

select(id, use) %>%

group_by(id) %>%

summarize(prevalence = mean(use)) %>%

pull(prevalence)

# scale and center continuous covariates

df.scaled <- df %>% mutate_if(is.double, scale)

# fit a simple univariate fixed-effects glm

m.fixed <- glm(use ~ 1 + roaddist, data=df.scaled, family=binomial(link = ‘logit’))

# fit the same model, but with random intercepts for individuals

m.rand.intercept <- glmer(use ~ 1 + roaddist + (1|id), data=df.scaled,

family=binomial(link = ‘logit’))

# note warning about boundary effect

# check estimates and SE's of fixed effects (your inference)

summary(m.fixed)

summary(m.rand.intercept)

The SEs and coefficient estimates are equal for these two model specifications. Your inference is exactly the same. Also look at the estimate of individual-level variation for the glmer--it’s effectively zero (the boundary effect).

The random intercept takes on a very different role when you also include random slopes. If individuals vary in their selection for a covariate, then you need random intercepts because it doesn’t make sense to force P(use) through a single point along the x-axis. If some respond positively and others negatively to distance to roads, for example, there’s no reason to expect they’ll all converge on the exact same probability of selection at the mean distance to roads in your dataset. Random intercepts give some needed flexibility there.

Your random intercepts, however, have no clear biological interpretation that I can think of, which is probably why your results show they’re all about the same, may in fact be zero, and, if you plotted their densities instead of means and CRIs, you’d probably find them to closely reflect their priors (which I’m guessing were Uniform(0,2) based on Fig. 5). This is a really long-winded way of saying you need random slopes, not just random intercepts, if you want to know anything about individual-level variation in selection. If you don’t, you should take the random intercepts out because prevalence is a design parameter that you’ve fixed to be constant across individuals (Line 231). They are not affecting your inference.

Author response: While random intercepts were necessary given our unbalanced sampling design (individuals had different numbers of relocations), we have redone this analysis to now incorporate random slopes in addition to random intercepts.

Line 243: “Regression coefficients for each variable were the product of binary indicator

variables and both continuous and categorical covariates.”

Hmmm, this doesn’t make sense the way it’s worded. They should be the product of the

indicator variable and an effect size parameter (the ‘raw’ coefficient). Multiplying the indicator by the covariate (the whole vector? I’m having a hard time picturing what you mean here) will give you either a vector of zeros or your original vector of covariates. It won’t give you a regression coefficient.

Author response: Now that we have redone this analysis, we have removed this statement.

Line 246: “We assumed that all variables with high inclusion probability based on the posterior

distributions of their indicator variables influenced habitat selection and variables with inclusion

probabilities ≤ 0.25 were unimportant [67]”

Mutshinda et al. used ≤0.25 to identify unimportant variables because they used a fixed prior probability of 0.5, so that posterior probability coincides with a Bayes factor of 0.33, or 3:1 against the variable. Because you have a distribution of prior odds, rather than a fixed value, 0.25 doesn’t have any real meaning here. If, for example, your model converged on an inclusion probability of 0.1, then 0.25 would actually represent a Bayes factor of 3, or 3:1 evidence for the variable. If anything, I could see taking the median posterior inclusion probability and using that to calculate the cutoff, but that seems a little circular, too. I’d argue that variable selection is simpler if you just pick what you think is an appropriate prior probability and consider it fixed. With the amount of data you have, I doubt this value is going to have much of an influence on which variables are identified as important, but you could always try running it with a few values to check that sensitivity. Unless mean variable inclusion probability is a parameter you’re interested in making inference on (which it doesn’t seem like it is, given that you don’t report it in the results), then I don’t quite see the value in putting a prior on it and estimating it.

Author response: We have redone the analysis and are no longer using indicator variables.

Line 250: Define “standardized coefficients.”

Author response: We have clarified that we centered and scaled all predictor variables to calculate standardized coefficients (lines 255-257).

Line 252: So are you using credible intervals or the posterior inclusion probability to determine

which variables are important? Is one given higher priority than another?

Author response: We are no longer using inclusion probabilities and have clarified that we used 95% credible intervals to determine which variables were important (lines 257-258).

Line 253: I’m not exactly sure what these σ2 are, but given that you did not include random

slopes, they cannot represent individual variation in behavior.

Author response: We have now included random slopes and make inferences on individual variation using the variation in individual-specific slopes for each variable (Fig 4) and the number of individuals that were selecting for and against each variable (Fig 5).

Line 259: “Vague uniform or normal priors were used for all model parameters related to

covariates and their measures of error.”

This statement is itself a little vague. I would like to see either a) complete model statements in your methods section or b) your JAGS model code (in supplemental info?) so I can see exactly what priors were placed on each parameter. Ideally, provide both.

Author response: This is a good point and we have now included model code in our Supplemental Material so that the reader can fully understand our methods.

Line 269: Insert ‘Gelman-Rubin statistic’ before ‘values’ or just refer to them as ‘potential scale

reduction factors’ so we know what values you’re referring to.

Author response: Given our new analysis approach (fitting a model with integrated nested Laplace approximation), this statistic is no longer relevant as it applies to MCMC approaches.

Line 270: I’m glad to see you assessed model fit, but it would be nice to know which specific

‘attribute of the data’ you used to calculate the Bayesian p -value.

Author response: We have removed this statement given our new analysis approach.

Line 290: Were any models with >1 predictor considered? Again, more details on the 2nd order

resource selection analysis are needed.

Author response: We have included additional details on both variables and models considered in the analyses assessing home range size and second order selection. We included additional information clarifying that only single-variable models were evaluated and our model selection procedure for the home range size analysis and we have included further details on compositional analysis to make the process more transparent (lines 181-219).

Lines 292–293: This sentence and Table 3 seem to provide exactly the same information, so I

would just include one or the other. I find Table 3 hard to interpret (isn’t the info on either side of the diagonal redundant? And why isn’t there a #2-ranked landcover?), so I’d suggest axing it.

Author response: This is a good point and we have now removed Table 3.

Line 308: Again, I don’t think 0.25 is a meaningful cutoff. Try to come up with a criterion more

applicable to your particular model.

Author response: With our revised analysis, we are no longer using indicator variables and so no longer require a cutoff.

Line 311: This doesn’t indicate individual variation in habitat selection unless you fit random

intercept and random slope models. So am I right to assume these variances indicate variation

among individual-level adjustments to the intercept? If so, why is there more than one? Are

these from separate, univariate models? I’m generally very confused about your model

structure, so I’ll reiterate here that it would really help if you provided model statements and

code so readers know exactly what parameters these are. Also (and this is a personal preference thing unless the journal has specific guidelines about it), I’m much more accustomed to thinking of variation on the scale of standard deviations rather than variances. Seems easier to interpret since it’s on the same scale as the regression coefficients and data. Is there a specific reason you report σ 2 rather than σ?

Author response: Now that we have included random slopes in our model, we make inferences about individual variation using the variation in individual-specific slopes for each variable (Fig 4). We have also included model code in our Supplemental Material so readers know exactly how the model was specified.

Line 377: Given that you have the ability to fit random-slope, random-intercept models (because

you sampled availability at the individual level), there is no reason you shouldn’t. In fact, I would argue that assuming the slopes are fixed across individuals is inappropriate given that

availability differed across your study area, as you state on Line 348. At the individual level and

the third-order, you are only likely to detect avoidance of, say, oil and gas features, if there are

oil and gas features within the home range to avoid. By forcing all the responses to be the same

among individuals, you’re not allowing your models to pick up on those potential functional

responses to availability. Moreover, fitting random-slope, random-intercept models does not preclude making inference at the population level. The beauty of these multi-level models is that you can make inference at multiple levels of the hierarchy—population or individual. Going back to the example of citation #69, this model provides not only estimates of the individual-level σ j ’s, but also the population-level β .

Author response: This is a good point and we have now included random slopes.

Fig. 1. Might this look a little better on the log(ha) scale?

Author response: We feel it is more intuitive if the variable is not transformed, but can defer to the associate editor.

Fig. 2. Since both variables are continuous, why not plot the data? It’s a more honest way of

showing the strength and predictive capability of a relationship than just plotting the regression

line for the mean and its confidence interval.

Author response: We have now included the raw data in our plot of the regression results (Fig 2).

Fig. 3 and Fig. 5: Both of these figures depict information that seems better suited to a table.

Just a suggestion.

Author response: We have redone our analysis, so Fig 3 and Fig 5 now present different information.

Attachment

Submitted filename: Response to Reviewers.pdf

Click here for additional data file.^{(163.1KB, pdf)}

PLoS One. doi: 10.1371/journal.pone.0233756.r003

Decision Letter 1

W David Walter

22 Apr 2020

PONE-D-19-33753R1

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

PLOS ONE

Dear Dr. Milligan:

We would appreciate receiving your revised manuscript by 22 May 2020. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

W. David Walter, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

We thank the authors for addressing the reviewer comments in great detail. I would request the authors address the remaining concerns of the same 2 authors as well as my own concerns regarding second-order selection.

In the previous draft, a reviewer requested more details on second-order selection, particularly on the variables and reason for not conducting RSF for second-order as done with third-order. I don’t believe the authors adequately addressed the previous concerns but more importantly the results seem confusing to me. Let’s start with the justification and the response from the authors in the previous version:

Line 200: I’d like to see more detail on your methods for the second-order resource selection

analysis, rather than just referring the reader to a couple citations. Which variables were

considered, how was model fitting and model selection accomplished? There’s no way I could

re-create your analysis with the details you give here.

AE’s comments: The authors state on Lines 210: “We used compositional analysis to compare used versus available landcover types and grazing systems separately.” However, in the results the authors identify results only for landcover types (Lines 295-301) in the order of preference. They then claim in Lines 301-303 that “There was no evidence that selection of home ranges in relation to grazing system was different from random (p = 0.20), suggesting that females were not differentiating between pastures in the different grazing systems.” This appears to be relating the linear models in the preceding section linking size of home range to grazing system, not compositional analysis? If it were a separate analysis on grazying system, I would expect a similar sentence of the 3 systems as presented for the landcover types in Lines 295-301 (i.e., no preference for grazing system). I am not sure if I am missing something here?

More importantly, your justification for second-order selection that “others are still using it” is poor and I recommend the authors delete the entire concept of second order selection in your manuscript. Although the authors present compositional analysis in the Methods and Results, the authors then combine both second- and third-order selection in a statement in Lines 357-358 and second-order selection is never mentioned again. The details of second-order selection does not add much obvious value to the manuscript and the authors do not even spend any time on its contribution in the Discussion?

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Reviewer #1: Lines 32-33: Very minor point of subject-verb agreement: “…strategies…are important for the conservation of sharp-tailed grouse. “

Line 69: Similar to a comment in my initial review, I find this type of citation to be misleading. Sharp-tailed grouse are “Recognized as an indicator species for grassland ecosystems” by an unpublished master’s thesis from Kansas on lesser prairie chickens? Yes, sharptails are recognized as indicator/keystone/umbrella species, but there are certainly more appropriate entities to cite, including state wildlife management plans, joint venture implementation plans, and other plans developed by groups involved with on-the-ground conservation in the study area.

Lines 218-219: The objective might have been to evaluate landcover type and grazing system, but in many areas those are influenced by or correlated with continuous variables such as proximity to road, proximity to water, or topographic variation, to name a few. I still maintain that the compositional analysis leaves out important information. Variables quantifying anthropogenic disturbance are included in the within-home-range analysis (line 227); it only makes sense that if birds avoid such disturbance within their home range that they might avoid such disturbance when selecting a home range.

Line 289: Final decision is up to the editor, but most people find informative figure headings easier to understand than descriptive, i.e., the heading for Figure 2 would convey more information and be easier to assimilate if it read “Home range size was negatively related to density of edge habitat within the home range.” Other headings could be similarly revised.

Lines 349-352: But habitat selection across the landscape did not consider these factors, and the reason that some individuals had few to no roads or oil wells within their home range might be because they selected areas with few to no roads.

Lines 371-373: It was stated earlier that topography and soil type didn’t vary among grazing systems, but if there was substantial variation across the landscape (regardless of grazing system), might they have affected habitat selection?

Reviewer #2: Overall, I'm very happy with the improvements that were made in response to the first round of reviews. I'm especially glad the authors chose to re-analyze the third-order resource selection data with a varying-intercept and varying-slope model. There are just a few minor interpretation issues that I think should be addressed. These are discussed below.

Line-by-line comments on PONE-D-19-33753R1:

Line 190: This is a little too vague to reproduce. Is it edge length divided by home range size?

Line 209: You might consider using a term other than ‘home range’ in this context, since it doesn’t really apply to a group of individuals. Maybe ‘95% utilization distribution’ or something like that.

Line 275: I didn’t catch this before, but you must have had several females that contributed multiple home ranges to this analysis. In keeping with the rest of your analysis, you should account for this non-independence with random effects or subsample so each individual contributes just one home range.

Figure 1: Point taken re: interpretability, and I’ll defer to the AE’s judgement on this, too, but it’s awfully hard to see any differences among the means on this scale.

Figure 2: Need units on the x-axis label (m/km2, or whatever these are).

Line 308: Strike ‘the variable of’.

Figure 3: The y-axis label could be a lot more informative than ‘effect size.’ Maybe, ‘population level selection coefficient,’ or ‘global mean selection coefficient.’

Figure 4: Comparing this figure ot Figure 3, it’s apparent that these are not the individual-level slopes per se, but the individual-level adjustments to the population-level means shown in Figure 3. This should either be explained more clearly in the figure caption, or (and I’d prefer this) you could show boxplots of the actual individual-level slope estimates. This would show the location and scale of these slope parameters in one place.

Line 313 and Figure 5: This figure, and Figure 4, are interesting in that they show there is substantial individual-level variation in selection coefficients. But they immediately raise the question: Are we actually seeing individuals that vary widely in their habitat preferences, or are we just seeing more-or-less constant 3rd-order habitat use among home ranges with highly variable resource availability (i.e., functional responses, sensu Mysterud and Ims)? You could answer that with your data, but you kind of leave us hanging. And the statement on Line 313 that you’re seeing ‘no population-level selection’ kind of sweeps this issue under the rug.

Line 329: What you’ve shown here hasn’t convince me we are seeing ‘highly plastic habitat use.’ Keep in mind the distinction between resource use and resource selection. You’ve demonstrated highly variable selection coefficients, but you have not shown that individuals are highly variable in what they’re actually using. This comes back to my comment above re: functional responses to availability. Because you don’t test for functional responses, you haven’t convinced me these birds are actually displaying divergent behaviors/preferences when it comes to third order resource selection.

I’m not saying you’re interpreting your data wrong, just that you haven’t quite provided us, the readers, with enough evidence. I went ahead and plotted mean use (y-axis) against mean availability (x-axis) for several of your variables among home ranges, and this is what it looks like:

[see pdf version of comments!]

This tells me you’re probably right, they are very plastic in what they use. In fact, use seems to be more variable than availability in some cases. I’d suggest including something like this, or (even better) selection coefficients plotted against mean availability, if you want to make your argument of highly plastic habitat preference/use stronger. You could even combine the information in Figure 5 with this type of plot by color coding the home ranges according to positive, null, or negative individual-level selection coefficients.

Line 351: This sentence seems like a holdover from the last version of the manuscript, before you incorporated individual-level random slopes. Given what the plot above shows, I don’t think variation in availability among home ranges is obscuring some real response to oil pad distance—they just don’t seem to be avoiding them at all.

Line 361: Reconsider the word ‘inconsistent’ here; there’s nothing inconsistent about finding different habitat relationships for different species in a different ecological contexts.

Line 376: More important than what? I’m not picking up on your meaning here.

Line 380: This sentence also reads like a holdover from your last version.

Line 386: Clarify that you mean individual differences in resource selection. As I read your suggestion for future research here, it seems even more critical that you show readers that these differences in selection coefficients are not simply functional responses to availability. If they are, then your suggestion that they represent alternative strategies is much weaker.

Line 401: It’s up to you how much you want to synthesize the results of your various analyses for the readers of this paper, but I can’t help but think you’re leaving out some critical knowledge from your last paper in this conclusion. Just reading this, I’m left with the impression that grazing systems are irrelevant to sharp-tailed grouse in this ecosystem. But what about the fact that you showed nest success was significantly lower in these rotational grazing systems? If they’re not avoiding them, but they’re achieving lower reproductive success in them, then RGS aren’t just neutral or not quite as good as we’d hoped, they’re a (potentially) bad management practice for sharp-tailed grouse!

Again, this is largely a matter of personal preference, but I would personally love to see some synthesis of what you’re finding out there. Don’t assume people will search out all your research and put the puzzle together themselves.

Table S2: This table title should probably include more detail so it can stand alone. Something like: “Support for models predicting the spatial grain of each landcover variable that best predicts third-order female sharp-tailed grouse habitat selection during the breeding season, 2016–2018, based on deviance information criteria (DIC) values.

Table S3: Same comment as above.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Joseph Smith

Attachment

Submitted filename: milligan sharptail R1 review.pdf

Click here for additional data file.^{(111.4KB, pdf)}

PLoS One. 2020 Jun 4;15(6):e0233756. doi: 10.1371/journal.pone.0233756.r004

Author response to Decision Letter 1

5 May 2020

Additional Editor Comments (if provided):

Line 200: I’d like to see more detail on your methods for the second-order resource selection

analysis, rather than just referring the reader to a couple citations. Which variables were

considered, how was model fitting and model selection accomplished? There’s no way I could

re-create your analysis with the details you give here.

Author response: Thank you for these comments. We believe that it is important to include both second and third order selection because these analyses evaluate habitat selection at different spatial scales, but we have followed the first reviewer’s recommendation and changed our second order analysis so that it uses resource selection functions similar to the third order analysis rather than compositional analysis. This has allowed us to include additional continuous variables, such as distance to oil pad, which could not be accommodated with our previous analysis approach. We have outlined our approach in detail and believe that it adds important information to the manuscript.

Reviewer #1

Lines 32-33: Very minor point of subject-verb agreement: “…strategies…are important for the conservation of sharp-tailed grouse. “

Author response: We have corrected this sentence (Line 33).

Author response: This citation refers to a thesis on sharp-tailed grouse in Alberta that provided the best empirical evidence of which we are aware regarding sharp-tailed grouse as indicator species (Line 69). We agree that there are other plans that cite sharp-tailed grouse as indicator species, but they did not evaluate this concept, which is why we chose this citation. We can provide additional citations if the editor feels that is necessary.

Author response: Following this recommendation, we have redone our second order analysis and now use resource selection functions which evaluate all the variables included in the within-home range analysis.

Author response: We can defer to the editor on this.

Author response: Our second order analysis now includes these continuous variables, including roads and oil pads, so we have updated our discussion to reflect this.

Author response: We have revised our analysis, so that it now includes additional variables which altered our discussion, so we have removed this sentence.

Line-by-line comments on PONE-D-19-33753R1:

Line 190: This is a little too vague to reproduce. Is it edge length divided by home range size?

Author response: We have clarified that edge density was edge length divided by home range size (Lines 196-198).

Author response: This is a good point and we clarified that the study area was defined by a utilization distribution, which is a more appropriate term than home range (Lines 205-206).

Author response: We have updated our second order analysis and we only included one home range from each individual, which we clarified in the methods section (Lines 202-203).

Figure 1: Point taken re: interpretability, and I’ll defer to the AE’s judgement on this, too, but it’s awfully hard to see any differences among the means on this scale.

Author response: We can defer to the editor on this, but we found no evidence for a difference in home range size among grazing systems, which is what is depicted in this graph.

Figure 2: Need units on the x-axis label (m/km2, or whatever these are).

Author response: We have included units on the x-axis.

Line 308: Strike ‘the variable of’.

Author response: We have removed this phrase.

Figure 3: The y-axis label could be a lot more informative than ‘effect size.’ Maybe, ‘population level selection coefficient,’ or ‘global mean selection coefficient.’

Author response: We have revised the y-axis label following this recommendation.

Figure 4: Comparing this figure to Figure 3, it’s apparent that these are not the individual-level slopes per se, but the individual-level adjustments to the population-level means shown in Figure 3. This should either be explained more clearly in the figure caption, or (and I’d prefer this) you could show boxplots of the actual individual-level slope estimates. This would show the location and scale of these slope parameters in one place.

Author response: We have fixed this figure so that it now shows boxplots of the actual individual-level slope estimates.

Author response: This is a good point and we have included an additional figure (Figure 6) to allow readers to better evaluate our conclusions and to demonstrate that we are seeing highly flexible habitat selection, not a functional response.

[see pdf version of comments!]

Author response: Thank you for the useful suggestion. We have included a figure (Figure 6) like the one you suggested that plots resource use against availability and combined it with the previous Figure 5 to show the number of individuals selecting for and against specific habitat variables. This figure includes important information about the variation in individual selection and is similar to previously published figures evaluating functional responses.

Author response: We have revised our discussion and removed this sentence.

Line 361: Reconsider the word ‘inconsistent’ here; there’s nothing inconsistent about finding different habitat relationships for different species in a different ecological contexts.

Author response: Given our updated results, we have revised this sentence (Line 368).

Line 376: More important than what? I’m not picking up on your meaning here.

Author response: We have revised our discussion and removed this sentence.

Line 380: This sentence also reads like a holdover from your last version.

Author response: We have removed this sentence.

Author response: We have clarified that we mean individual differences in resource selection (Line 399).

Author response: This is a good suggestion and we have incorporated findings from previous research to better synthesize our results (Lines 370-373 and 404-406).

Author response: We have altered the title to be more informative following this suggestion.

Table S3: Same comment as above.

Author response: We have altered the title to be more informative following this suggestion.

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(20.9KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0233756.r005

Decision Letter 2

W David Walter

13 May 2020

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

PONE-D-19-33753R2

Dear Dr. Milligan,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

W. David Walter, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

I appreciate the authors edits to the previous reviews and their detailed responses to those reviews. Specifically, updating your compositional analysis to second-order selection with resource selection functions was much appreciated. I found a few items that I would appreciate the authors paying attention to in the type setting stage. As the authors are the experts on these topics, I just want to be sure they are presented properly in the final version of the manuscript.

Line 150: Should “lessees” be “leassees” or changed to “leaseholders” for clarity? Current spelling seems odd but if that is correct then please ignore my comment.

Lines 151: AUM? Spell out first use unless the authors believe this is known and understood worldwide?

Table 3: Remove the Cumulative AIC weight column as previously requested by a reviewer and I believe was confirmed removed by the authors in their rebuttal?

Line 309: Please change “Within the home range” to “At the third order” to be consistent and clear that this represents the start of the section on third-order selection of habitat.

Table S2 and S3: Spell out “Ag”to be clear it is agriculture. Also, Table S3 has S6 in the document?

Please be sure to go through all Tables throughout manuscript to check for inclusion and citing. Do the same for Supplemental Figures and Tables as well.

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0233756.r006

Acceptance letter

W David Walter

21 May 2020

PONE-D-19-33753R2

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

Dear Dr. Milligan:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. W. David Walter

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Appendix. Example code for Bayesian logistic regression model evaluating third order habitat selection using the R package R-INLA.

(DOCX)

Click here for additional data file.^{(16.2KB, docx)}

S1 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of grassland measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S2 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of wooded draws measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S3 Fig. Simulation results evaluating the number of available points necessary for convergence of the proportion of row crop agriculture measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S4 Fig. Simulation results evaluating the number of available points necessary for convergence of edge density measured at different buffer distances.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

S5 Fig. Simulation results evaluating the number of available points necessary for convergence of variables measured at a single scale.

(TIF)

Click here for additional data file.^{(3.4MB, tif)}

The number of parameters (K), AIC_c values, AIC_c values, model weights (w_i), and log-likelihoods are reported.

(DOCX)

Click here for additional data file.^{(16.6KB, docx)}

The number of parameters (K), AIC_c values, AIC_c values, model weights (w_i), and log-likelihoods are reported.

(DOCX)

Click here for additional data file.^{(15.4KB, docx)}

(DOCX)

Click here for additional data file.^{(17.5KB, docx)}

(DOCX)

Click here for additional data file.^{(17.4KB, docx)}

S1 Data

(ZIP)

Click here for additional data file.^{(10.5MB, zip)}

Attachment

Submitted filename: milligan sharptail nest selection review.pdf

Click here for additional data file.^{(121KB, pdf)}

Attachment

Submitted filename: Response to Reviewers.pdf

Click here for additional data file.^{(163.1KB, pdf)}

Attachment

Submitted filename: milligan sharptail R1 review.pdf

Click here for additional data file.^{(111.4KB, pdf)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(20.9KB, docx)}

Data Availability Statement

All relevant data are within the paper and its Supporting Information files.

[pone.0233756.ref001] 1.Johnson MD. Measuring habitat quality: a review. The Condor. 2007;109(3):489–504. [Google Scholar]

[pone.0233756.ref002] 2.Jones J. Habitat selection studies in avian ecology: a critical review. The Auk. 2001;118(2):557–62. [Google Scholar]

[pone.0233756.ref003] 3.Boyce MS, Johnson CJ, Merrill EH, Nielsen SE, Solberg EJ, Van Moorter B. Can habitat selection predict abundance? J Anim Ecol. 2016;85(1):11–20. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref004] 4.Beyer HL, Haydon DT, Morales JM, Frair JL, Hebblewhite M, Mitchell M, et al. The interpretation of habitat preference metrics under use–availability designs. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010;365(1550):2245–54. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref005] 5.McLoughlin PD, Morris DW, Fortin D, Vander Wal E, Contasti AL. Considering ecological dynamics in resource selection functions. J Anim Ecol. 2010;79(1):4–12. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref006] 6.Hebblewhite M, Merrill EH. Trade-offs between predation risk and forage differ between migrant strategies in a migratory ungulate. Ecology. 2009;90(12):3445–54. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref007] 7.Wiens JA, Milne BT. Scaling of ‘landscapes’ in landscape ecology, or, landscape ecology from a beetle’s perspective. Landscape Ecol. 1989;3(2):87–96. [Google Scholar]

[pone.0233756.ref008] 8.Hagen CA, Sandercock BK, Pitman JC, Robel RJ, Applegate RD. Spatial variation in lesser prairie-chicken demography: a sensitivity analysis of population dynamics and management alternatives. J Wildl Manage. 2009;73(8):1325–32. [Google Scholar]

[pone.0233756.ref009] 9.McNew LB, Gregory AJ, Sandercock BK. Spatial heterogeneity in habitat selection: Nest site selection by greater prairie-chickens. J Wildl Manage. 2013;77(4):791–801. [Google Scholar]

[pone.0233756.ref010] 10.Allen AM, Månsson J, Jarnemo A, Bunnefeld N. The impacts of landscape structure on the winter movements and habitat selection of female red deer. European Journal of Wildlife Research. 2014;60(3):411–21. [Google Scholar]

[pone.0233756.ref011] 11.Krausman PR, Naugle DE, Frisina MR, Northrup R, Bleich VC, Block WM, et al. Livestock grazing, wildlife habitat, and rangeland values. Rangelands. 2009;31(5):15–9. [Google Scholar]

[pone.0233756.ref012] 12.Alkemade R, Reid RS, van den Berg M, de Leeuw J, Jeuken M. Assessing the impacts of livestock production on biodiversity in rangeland ecosystems. Proc Natl Acad Sci. 2013;110(52):20900–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref013] 13.Fleischner TL. Ecological costs of livestock grazing in western North America. Conserv Biol. 1994;8(3):629–44. [Google Scholar]

[pone.0233756.ref014] 14.Boyd CS, Beck JL, Tanaka JA. Livestock grazing and sage-grouse habitat: impacts and opportunities. Journal of Rangeland Applications. 2014;1:58–77. [Google Scholar]

[pone.0233756.ref015] 15.Fuhlendorf SD, Engle DM. Restoring heterogeneity on rangelands: Ecosystem management based on evolutionary grazing patterns. Bioscience. 2001;51(8):625–32. [Google Scholar]

[pone.0233756.ref016] 16.Coppedge BR, Fuhlendorf SD, Harrell WC, Engle DM. Avian community response to vegetation and structural features in grasslands managed with fire and grazing. Biol Conserv. 2008;141(5):1196–203. [Google Scholar]

[pone.0233756.ref017] 17.McNew LB, Winder VL, Pitman JC, Sandercock BK. Alternative rangeland management strategies and the nesting ecology of greater prairie-chickens. Rangeland Ecol Manage. 2015;68(3):298–304. [Google Scholar]

[pone.0233756.ref018] 18.Winder VL, McNew LB, Pitman JC, Sandercock BK. Effects of rangeland management on survival of female greater prairie-chickens. J Wildl Manage. 2018;82(1):113–22. [Google Scholar]

[pone.0233756.ref019] 19.Engle DM, Fuhlendorf SD, Roper A, Leslie DM Jr. Invertebrate community response to a shifting mosaic of habitat. Rangeland Ecol Manage. 2008;61(1):55–62. [Google Scholar]

[pone.0233756.ref020] 20.Limb RF, Fuhlendorf SD, Engle DM, Weir JR, Elmore RD, Bidwell TG. Pyric–herbivory and cattle performance in grassland ecosystems. Rangeland Ecol Manage. 2011;64(6):659–63. [Google Scholar]

[pone.0233756.ref021] 21.Fuhlendorf SD, Townsend DE II, Elmore RD, Engle DM. Pyric-herbivory to promote rangeland heterogeneity: evidence from small mammal communities. Rangeland Ecol Manage. 2010;63(6):670–8. [Google Scholar]

[pone.0233756.ref022] 22.Lipsey MK, Naugle DE. Precipitation and soil productivity explain effects of grazing on grassland songbirds. Rangeland Ecol Manage. 2017;70(3):331–40. [Google Scholar]

[pone.0233756.ref023] 23.Holechek JL, Gomez H, Molinar F, Galt D. Grazing studies: what we’ve learned. Rangelands. 1999;21:12–6. [Google Scholar]

[pone.0233756.ref024] 24.Schieltz JM, Rubenstein DI. Evidence based review: positive versus negative effects of livestock grazing on wildlife. What do we really know? Environmental Research Letters. 2016;11(11):113003. [Google Scholar]

[pone.0233756.ref025] 25.Briske DD, Derner J, Brown J, Fuhlendorf SD, Teague W, Havstad K, et al. Rotational grazing on rangelands: reconciliation of perception and experimental evidence. Rangeland Ecol Manage. 2008;61(1):3–17. [Google Scholar]

[pone.0233756.ref026] 26.Frisina MR. Grazing private and public land to improve the Fleecer Elk Winter Range. Rangelands. 1991;13(6):291–4. [Google Scholar]

[pone.0233756.ref027] 27.Montana Department of Fish Wildlife and Parks. Upland game bird enhancement program strategic plan. Helena, MT, USA: 2011. [Google Scholar]

[pone.0233756.ref028] 28.Sliwinski M, Burbach M, Powell L, Schacht W. Factors influencing ranchers’ intentions to manage for vegetation heterogeneity and promote cross-boundary management in the northern Great Plains. Ecol Soc. 2018;23(4):45–62. [Google Scholar]

[pone.0233756.ref029] 29.Hormay AL. Principles of rest-rotation grazing and multiple-use land management: US Department of the Interior, Bureau of Land Management; 1970. [Google Scholar]

[pone.0233756.ref030] 30.Hormay AL, Evanko AB. Rest-rotation grazing: a management system for bunchgrass ranges: California Forest and Range Experiment Station; 1958. [Google Scholar]

[pone.0233756.ref031] 31.Kraft JD. Vegetation characteristics and lesser prairie chicken responses to land cover types and grazing management in western Kansas [Thesis]. Manhattan, KS, USA: Kansas State University; 2016.

[pone.0233756.ref032] 32.Roersma SJ. Nesting and brood rearing ecology of plains sharp-tailed grouse (Tympanuchus phasianellus jarnesi) in a mixed-grass/fescue ecoregion of southem Alberta [Thesis]. Winnipeg, MB, Canada: University of Manitoba; 2001.

[pone.0233756.ref033] 33.Hillman CN, Jackson WW. The sharp-tailed grouse in South Dakota: South Dakota Department of Game, Fish and Parks; 1973. [Google Scholar]

[pone.0233756.ref034] 34.Poiani KA, Merrill MD, Chapman KA. Identifying conservation-priority areas in a fragmented Minnesota landscape based on the umbrella species concept and selection of large patches of natural vegetation. Conserv Biol. 2001;15(2):513–22. [Google Scholar]

[pone.0233756.ref035] 35.Bergerud AT, Gratson MW. Survival and breeding strategies of grouse In: Bergerud AT, Gratson MW, editors. Adaptive strategies and population ecology of northern grouse. Minneapolis, MN, USA: University of Minnesota Press; 1988. p. 473–577. [Google Scholar]

[pone.0233756.ref036] 36.Winder VL, McNew LB, Pitman JC, Sandercock BK. Space use of female greater prairie-chickens in response to fire and grazing interactions. Rangeland Ecol Manage. 2017;70(2):165–74. [Google Scholar]

[pone.0233756.ref037] 37.Lautenbach JM, Plumb RT, Robinson SG, Hagen CA, Haukos DA, Pitman JC. Lesser prairie-chicken avoidance of trees in a grassland landscape. Rangeland Ecol Manage. 2017;70(1):78–86. [Google Scholar]

[pone.0233756.ref038] 38.Dettenmaier SJ, Messmer TA, Hovick TJ, Dahlgren DK. Effects of livestock grazing on rangeland biodiversity: A meta-analysis of grouse populations. Ecology and Evolution. 2017; 10.1002/ece3.3287 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref039] 39.Goddard AD, Dawson RD, Gillingham MP. Habitat selection by nesting and brood-rearing sharp-tailed grouse. Canadian Journal of Zoology. 2009;87(4):326–36. [Google Scholar]

[pone.0233756.ref040] 40.Marks JS, Marks VS. Habitat selection by Columbian sharp-tailed grouse in west-central Idaho: USDI Bureau of Land Management; 1987. [Google Scholar]

[pone.0233756.ref041] 41.Saab VA, Marks JS. Summer habitat use by Columbian sharp-tailed grouse in western Idaho. Great Basin Nat. 1992;52(2):166–73. [Google Scholar]

[pone.0233756.ref042] 42.Ryan MR, Burger LW, Jones DP, Wywialowski AP. Breeding ecology of greater prairie-chickens (Tympanuchus cupido) in relation to prairie landscape configuration. Am Midl Nat. 1998;140:111–22. [Google Scholar]

[pone.0233756.ref043] 43.Winder VL, Carrlson KM, Gregory AJ, Hagen CA, Haukos DA, Kesler DC, et al. Factors affecting female space use in ten populations of prairie chickens. Ecosphere. 2015;6(9):1–17. [Google Scholar]

[pone.0233756.ref044] 44.Niemuth ND. Identifying landscapes for greater prairie chicken translocation using habitat models and GIS: a case study. Wildl Soc Bull. 2003;31:145–55. [Google Scholar]

[pone.0233756.ref045] 45.Manzer DL. Sharp-tailed grouse breeding success, survival, and site selection in relation to habitat measured at multiple scales [Dissertation]. Edmonton, AB, Canada: University of Alberta; 2004.

[pone.0233756.ref046] 46.Cope MG. Distribution, habitat selection and survival of transplanted Columbian sharp-tailed grouse (Tympanuchus phasianellus columbianus) in the Tobacco Valley, Montana [Thesis]. Bozeman, MT, USA: Montana State University; 1992.

[pone.0233756.ref047] 47.Winder VL, McNew LB, Gregory AJ, Hunt LM, Wisely SM, Sandercock BK. Space use by female greater prairie-chickens in response to wind energy development. Ecosphere. 2014;5(1):1–17. [Google Scholar]

[pone.0233756.ref048] 48.Stonehouse KF, Shipley LA, Lowe J, Atamian MT, Swanson ME, Schroeder MA. Habitat selection and use by sympatric, translocated greater sage-grouse and Columbian sharp-tailed grouse. J Wildl Manage. 2015;79(8):1308–26. [Google Scholar]

[pone.0233756.ref049] 49.Dinkins JB, Conover MR, Kirol CP, Beck JL, Frey SN. Greater Sage-Grouse (Centrocercus urophasianus) select habitat based on avian predators, landscape composition, and anthropogenic features. The Condor: Ornithological Applications. 2014;116(4):629–42. [Google Scholar]

[pone.0233756.ref050] 50.Hagen CA, Pitman JC, Loughin TM, Sandercock BK, Robel RJ, Applegate RD. Impacts of anthropogenic features on habitat use by Lesser Prairie-Chickens. Studies in Avian Biology. 2011;39:63–75. [Google Scholar]

[pone.0233756.ref051] 51.Mysterud A, Ims RA. Functional responses in habitat use: Availability influences relative use in trade-off situations. Ecology. 1998;79(4):1435–41. [Google Scholar]

[pone.0233756.ref052] 52.Johnson DH. The comparison of usage and availability measurements for evaluating resource preference. Ecology. 1980;61(1):65–71. [Google Scholar]

[pone.0233756.ref053] 53.McDonald L, Erickson W, Boyce M, Alldredge J. Modeling vertebrate use of terrestrial resources In: Silvy NJ, editor. The Wildlife Techniques Manual Volume 1: Research. Baltimore, MD, USA: Johns Hopkins University Press; 2012. p. 410–29. [Google Scholar]

[pone.0233756.ref054] 54.Curtin CG, Sayre NF, Lane BD. Transformations of the Chihuahuan Borderlands: grazing, fragmentation, and biodiversity conservation in desert grasslands. Environmental Science & Policy. 2002;5(1):55–68. [Google Scholar]

[pone.0233756.ref055] 55.LANDFIRE. LANDFIRE Existing Vegetation Type layer. U S Department of Interior, Geological Survey http://wwwlandfiregov/indexphp [2016, November 18]. 2013.

[pone.0233756.ref056] 56.Milligan MC, Berkeley L, McNew LB. Effects of rangeland management on the nesting ecology of sharp-tailed grouse. Rangeland Ecol Manage. 2020;73:128–37. [Google Scholar]

[pone.0233756.ref057] 57.Seaman DE, Millspaugh JJ, Kernohan BJ, Brundige GC, Raedeke KJ, Gitzen RA. Effects of sample size on kernel home range estimates. J Wildl Manage. 1999;63:739–47. [Google Scholar]

[pone.0233756.ref058] 58.Worton BJ. Kernel methods for estimating the utilization distribution in home-range studies. Ecology. 1989;70(1):164–8. [Google Scholar]

[pone.0233756.ref059] 59.Manly B, McDonald L, Thomas D, McDonald T, Erickson W. Resource selection by animals: statistical analysis and design for field studies. Kluwer, Nordrecht, The Netherlands: 2002. [Google Scholar]

[pone.0233756.ref060] 60.Northrup JM, Hooten MB, Anderson CR Jr, Wittemyer G. Practical guidance on characterizing availability in resource selection functions under a use–availability design. Ecology. 2013;94(7):1456–63. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref061] 61.Burnham KP, Anderson DR. A practical information-theoretic approach. New York, NY, USA: Springer; 2002. [Google Scholar]

[pone.0233756.ref062] 62.Arnold TW. Uninformative parameters and model selection using Akaike’s Information Criterion. J Wildl Manage. 2010;74(6):1175–8. [Google Scholar]

[pone.0233756.ref063] 63.McGarigal K, Cushman SA, Ene E. FRAGSTATS v4: Spatial pattern analysis program for categorical and continuous maps. University of Massachusetts, Amherst: 2012;http://www.umass.edu/landeco/research/fragstats/fragstats.html. [Google Scholar]

[pone.0233756.ref064] 64.Laforge MP, Vander Wal E, Brook RK, Bayne EM, McLoughlin PD. Process-focused, multi-grain resource selection functions. Ecol Model. 2015;305:10–21. [Google Scholar]

[pone.0233756.ref065] 65.Gillies CS, Hebblewhite M, Nielsen SE, Krawchuk MA, Aldridge CL, Frair JL, et al. Application of random effects to the study of resource selection by animals. J Anim Ecol. 2006;75(4):887–98. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref066] 66.Thomas DL, Johnson D, Griffith B. A Bayesian random effects discrete-choice model for resource selection: Population-level selection inference. J Wildl Manage. 2006;70(2):404–12. [Google Scholar]

[pone.0233756.ref067] 67.Laforge MP, Brook RK, van Beest FM, Bayne EM, McLoughlin PD. Grain-dependent functional responses in habitat selection. Landscape Ecol. 2016;31(4):855–63. [Google Scholar]

[pone.0233756.ref068] 68.Holbrook JD, Squires JR, Olson LE, DeCesare NJ, Lawrence RL. Understanding and predicting habitat for wildlife conservation: the case of Canada lynx at the range periphery. Ecosphere. 2017;8(9):e01939. [Google Scholar]

[pone.0233756.ref069] 69.Muff S, Signer J, Fieberg J. Accounting for individual-specific variation in habitat-selection studies: Efficient estimation of mixed-effects models using Bayesian or frequentist computation. J Anim Ecol. 2019;89(1):80–92. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref070] 70.Fong Y, Rue H, Wakefield J. Bayesian inference for generalized linear mixed models. Biostatistics. 2010;11(3):397–412. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref071] 71.Wang X, Ryan YY, Faraway JJ. Bayesian regression modeling with INLA. Boca Raton, LA, USA: Chapman and Hall/CRC; 2018. [Google Scholar]

[pone.0233756.ref072] 72.Boisvert JH, Hoffman RW, Reese KP. Home range and seasonal movements of Columbian sharp-tailed grouse associated with conservation reserve program and mine reclamation. West N Am Nat. 2005;65:36–44. [Google Scholar]

[pone.0233756.ref073] 73.Christenson CD. Nesting and brooding characteristics of sharp-tailed grouse (Pedioecetes phasianellus jamesi) in southwestern North Dakota [Thesis]. Grand Forks, ND, USA: University of North Dakota; 1970.

[pone.0233756.ref074] 74.Patten MA, Pruett CL, Wolfe DH. Home range size and movements of greater prairie-chickens. Studies in Avian Biology. 2011;39:51–62. [Google Scholar]

[pone.0233756.ref075] 75.Milligan MC, Berkeley LI, McNew LB. Effects of rangeland management on the survival of adult sharp-tailed grouse. J Wildl Manage. In review. [Google Scholar]

[pone.0233756.ref076] 76.Holloran MJ, Kaiser RC, Hubert WA. Yearling greater sage-grouse response to energy development in Wyoming. J Wildl Manage. 2010;74(1):65–72. [Google Scholar]

[pone.0233756.ref077] 77.Milligan MC. Effects of grazing management on sharp-tailed grouse ecology in mixed-grass prairies [Dissertation]. Bozeman, MT, USA: Montana State University; 2019.

[pone.0233756.ref078] 78.Montana UNRCS-. Technical Guide—R058AE001MT. 1983.

[pone.0233756.ref079] 79.Matthews TW, Tyre AJ, Taylor JS, Lusk JJ, Powell LA. Habitat selection and brood survival of greater prairie-chickens. Studies in Avian Biology. 2011;39:179–91. [Google Scholar]

[pone.0233756.ref080] 80.Sih A, Bell AM, Johnson JC, Ziemba RE. Behavioral syndromes: an integrative overview. The Quarterly Review of Biology. 2004;79(3):241–77. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref081] 81.McLoughlin PD, Boyce MS, Coulson T, Clutton-Brock T. Lifetime reproductive success and density-dependent, multi-variable resource selection. Proceedings of the Royal Society B: Biological Sciences. 2006;273(1593):1449–54. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref082] 82.Leclerc M, Dussault C, St-Laurent M-H. Behavioural strategies towards human disturbances explain individual performance in woodland caribou. Oecologia. 2014;176(1):297–306. [DOI] [PubMed] [Google Scholar]

[pone.0233756.ref083] 83.Réale D, Dingemanse NJ, Kazem AJ, Wright J. Evolutionary and ecological approaches to the study of personality. Philosophical Transactions of the Royal Society B. 2010;365:3937–46. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0233756.ref084] 84.Sih A, Cote J, Evans M, Fogarty S, Pruitt J. Ecological implications of behavioural syndromes. Ecol Lett. 2012;15(3):278–89. [DOI] [PubMed] [Google Scholar]

PERMALINK

Habitat selection of female sharp-tailed grouse in grasslands managed for livestock production

Megan C Milligan

Lorelle I Berkeley

Lance B McNew

Roles

Abstract

Introduction

Study area

Methods

Results

Table 1. Home range size (95% volume contour) for radio-marked female sharp-tailed grouse monitored in the 3 grazing systems during the breeding seasons of 2016–2018.

Fig 1. Female sharp-tailed grouse breeding season home range size (± SE) by grazing system.

Table 2. Support for candidate models predicting the home range size of female sharp-tailed grouse during the breeding seasons of 2016–2018.

Fig 2. Relationship (± 85% confidence intervals) between the density of edge habitat (total landcover edge length / polygon area) and breeding season home range size for female sharp-tailed grouse.

Table 3. Support for candidate models predicting second order selection, or home range selection, of female sharp-tailed grouse during the breeding seasons of 2016–2018.

Fig 3. Relationship (± 85% confidence intervals) between the proportion grassland (A), the proportion wooded draws (B), distance to oil pad (C), and the proportion rest-rotation (D) and the relative probability of breeding season home range selection for female sharp-tailed grouse.

Fig 4. Fixed effects representing population-level habitat selection of sharp-tailed grouse within their home ranges during the breeding season, with error bars representing 95% credible intervals.

Fig 5. Variation in individual-specific slopes for each variable evaluated in the third order habitat selection analysis.

Discussion

Conclusions

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

W David Walter

Roles

Author response to Decision Letter 0

Decision Letter 1

W David Walter

Roles

Author response to Decision Letter 1

Decision Letter 2

W David Walter

Roles

Acceptance letter

W David Walter

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases