Habitat use by Mountain Plovers in prairie dog colonies in northeastern New Mexico
2012; Association of Field Ornithologists; Volume: 83; Issue: 2 Linguagem: Inglês
10.1111/j.1557-9263.2012.00365.x
ISSN1557-9263
Autores Tópico(s)Animal Ecology and Behavior Studies
ResumoABSTRACT Mountain Plovers (Charadrius montanus) are grassland birds that often breed in close association with colonies of black-tailed prairie dogs (Cynomys ludovicianus). However, not all colonies provide plover nesting habitat or habitat of equal quality, and the characteristics of colonies important for plovers remain poorly understood. Over two years, I used plover distribution surveys, territory mapping, and habitat sampling to study habitat use by plovers in prairie dog colonies in shortgrass prairie in northeastern New Mexico. My objective was to document important components of plover breeding habitat in colonies by comparing characteristics of used and unused habitats at three spatial scales: colony, territory, and nest-site. I found evidence of plover breeding in 14 of 44 colonies in 2009 and 13 of 43 colonies in 2010. Based on logistic regression, the probability of a colony being occupied by plovers was positively associated with colony size, but negatively associated with mean vegetation height. Preference for larger colonies could relate to minimum habitat requirements, or a potential tendency of this species to nest in social clusters. Shorter vegetation height was strongly correlated with greater bare ground and lower forb/subshrub cover, all characteristics that may be related to plover predator avoidance and foraging microhabitat. At both the territory and nest-site scale, areas used by plovers had shorter vegetation, more bare ground, and less forb/subshrub cover than unused areas. Nest sites were also more sloped, perhaps to reduce risk of flooding, and located further away from the nearest prairie dog burrow, perhaps to reduce risk of disturbance. Overall, my results show that plover use of prairie dog colonies was influenced by landscape and habitat features of colonies, and suggest that large colonies are particularly valuable because they are most likely to contain adequate areas with preferred habitat characteristics. Los Charadrius montanus son aves de pastizal que a menudo se reproducen en estrecha asociación con colonias de los perritos de las praderas (Cynomys ludovicianus). Sin embargo, no todas las colonias proveen hábitat de anidación para C. montanus, o hábitat de la misma calidad, y las características de las colonias que son importantes para C. montanus siguen siendo poco conocidas. Durante dos años, utilicé muestreos de la distribución de C. montanus, el mapeo de territorios y el muestreo de hábitat para estudiar el uso de hábitat por C. montanus en las colonias de los peritos de la pradera de pasto corto en el noreste de Nuevo México. Mi objetivo era de documentar los componentes de hábitat importantes para la reproducción de C. montanus en las colonias de los perritos de la pradera mediante la comparación de las características de hábitat utilizados y no utilizados en tres escalas espaciales: la colonia, el territorio y el sitio de anidación. Encontré evidencia de nidificación de C. montanus en 14 de 44 colonias en 2009 y en 13 de 43 colonias en 2010. En base a la regresión logística, la probabilidad de que una colonia sea ocupada por C. montanus estaba asociada positivamente con el tamaño de la colonia, pero negativamente con la media de la altura de la vegetación. La preferencia de anidar en colonias más grandes podría estar relacionada a los requerimientos mínimos de hábitat, o a una tendencia potencial de esta especie de anidar en grupos sociales. La menor altura de la vegetación estaba fuertemente correlacionada con mayor suelo desnudo y con una menor cobertura de plantas herbáceas/subarbustos, características cuales podrían estar relacionadas con evitar la depredación, y con el microhábitat de forrajeo. Tanto a la escala de territorio como también de sitio de anidación, las áreas utilizadas por C. montanus tenían vegetación más corta, suelo más desnudo, y menos cobertura de plantas herbáceas/subarbustos que áreas no utilizadas. Sitios de anidación también tenían pendientes más inclinadas, tal vez para reducir el riesgo de inundaciones, y localizados más lejos de la madriguera mas cercana de los perritos de las praderas, tal vez para reducir el riesgo de perturbación. En general, mis resultados demuestran que el uso de las colonias de los perritos de la pradera por parte de C. montanus esta influenciada por las características del paisaje y el hábitat de las colonias, y sugieren que grandes colonias son particularmente valiosas porque tienen una mayor probabilidad de tener áreas adecuadas con características de hábitat preferidas. Mountain Plovers (Charadrius montanus) breed in disturbed prairie or semi-desert habitats, primarily in the western Great Plains (Knopf and Miller 1994, Knopf and Wunder 2006). Historically, these plovers co-existed with grazing herbivores, such as American bison (Bison bison) and black-tailed prairie dogs (Cynomys ludovicianus), over much of their breeding range, benefiting from changes in habitat structure caused by these grazers (Knopf and Wunder 2006). Mountain Plovers typically breed in habitats dominated by short vegetation (Graul 1975, Olsen and Edge 1985, Ellison-Manning and White 2001) and bare soil (Knopf and Miller 1994), both features that often result from the activities of these mammalian grazers. At present, black-tailed prairie dog colonies remain an important source of breeding habitat throughout the Mountain Plover's range. Prairie dogs influence vegetation structure on colonies by foraging, but they also clip vegetation to facilitate predator detection (Hoogland 1995). These actions not only reduce vegetation height, but also tend to suppress woody plants (Weltzin et al. 1997), reduce overall plant coverage and biomass (Whicker and Detling 1988), and expose bare soil (Archer et al. 1987). Further, the burrowing actions of prairie dogs create additional bare soil and maintain microhabitats where annual forbs can survive (Whicker and Detling 1988, Detling 2006). At some locations, particularly the taller mixed prairie of the northern Great Plains, prairie dog activity may be essential for maintaining the habitat structure that Mountain Plovers prefer. For example, in mixed prairie in Montana, plovers nest almost exclusively in colonies (Knowles et al. 1982, Knowles and Knowles 1984, Olson and Edge 1985, Dinsmore and Smith 2010). In shortgrass systems, other disturbances, such as fire, livestock grazing, or agricultural activities, can also provide conditions appropriate for plover nesting (Sager 1996, Knopf and Rupert 1999, Augustine 2011). In many cases, however, habitats generated by these disturbances, particularly livestock grazing and agriculture, appear to be of low quality for plovers (Knopf and Rupert 1999, Shackford et al. 1999, Augustine and Derner 2012), whereas current research suggests that prairie dog activities can create high quality breeding conditions for plovers in shortgrass prairie (Dreitz 2009, Augustine 2011). Despite the importance of prairie dog colonies as plover habitat, not all colonies provide appropriate plover nesting habitat or habitat of equal quality. Colony characteristics, such as size or shape (Knowles and Knowles 1984, Olson-Edge and Edge 1987, Dinsmore et al. 2005, Dinsmore and Smith 2010), age (Augustine et al. 2008), vegetation structure (Olson-Edge and Edge 1987), coverage of bare ground (Knopf and Wunder 2006), or topography (Graul 1975), have all been suggested as being potentially important. However, many aspects of plover habitat selection in prairie dog colonies remain poorly understood. I studied habitat use by breeding Mountain Plovers in black-tailed prairie dog colonies in a shortgrass prairie in northeastern New Mexico. Although Mountain Plovers are not dependent on prairie dog colonies in the southern Great Plains of New Mexico, often nesting in grasslands heavily grazed by cattle (Sager 1996), plovers were found breeding exclusively in colonies in my study area (Goguen 2012). My objective was to document important components of Mountain Plover breeding habitat in prairie dog colonies by comparing characteristics of used and unused habitats at three spatial scales: colony, territory, and nest-site. Study site. I studied Mountain Plovers from mid-May through early July 2009 and 2010 at a 16 190 ha shortgrass prairie study area located on the 233 603-ha Vermejo Park Ranch (VPR; 36° 32’ 14.0” N, 104° 44’ 17.4 W) in Colfax County, northeastern New Mexico. Although most of VPR consists of mixed-conifer forests typical of higher elevations (>2200 m), my study area was located at lower elevations (1800–2100 m) on the southeastern edge of the ranch. Climate of the region is semiarid, averaging 410 mm of precipitation annually with most falling during spring and summer. Precipitation during spring (March–June) in this region averages 110 mm, but measured 30 and 127 mm during spring 2009 and 2010, respectively (all climate data from Cimarron 4 SW station, U.S. National Oceanographic and Atmospheric Administration). The study area consisted primarily of perennial shortgrasses, particularly blue grama (Bouteloua gracilis), interspersed with other grasses, forbs, shrubs, and cacti, particularly fringed sage (Artemisia frigida), broom snakeweed (Gutierrezia sarothrae), winterfat (Eurotia lanata), prickly-pear cactus (Opuntia spp.), and tree cholla (Cylindropuntia imbricata). The study area also contained large areas dominated by woody shrubs, mostly green rabbitbrush (Chrysothamnus viscidiflorus) and four-winged saltbush (Atriplex canescens), and was transected by three stream systems supporting narrow strips of woody riparian habitat, mainly dense willow (Salix spp.). Since the late 1990s, management on VPR has focused on ecosystem restoration. American bison were reintroduced in 1997 and, during my study, between 520 and 640 adult American bison and calves were rotated among fenced allotments within the area such that some areas were actively grazed during the plover breeding season. Black-tailed prairie dog restoration efforts have also been underway since 1999, increasing the number and coverage of colonies in the study area from eight colonies totaling 15 yr, with area covered by individual colonies positively correlated with their age (Pearson r2= 0.28). Distribution of black-tailed prairie dog colonies during spring 2009 on VPR, northeastern New Mexico, showing colonies occupied by breeding Mountain Plovers. Plover distribution surveys and nest searching. To determine which prairie dog colonies were occupied by breeding plovers, I conducted morning surveys at 44 prairie dog colonies totaling 2672 ha in 2009, and 43 colonies totaling 2581 ha in 2010. One to three observers surveyed each colony on foot, walking back and forth slowly through the entire colony in routes separated by about 100 m, and stopping regularly to scan with binoculars and listen for plover vocalizations. Observers recorded plover locations on colony maps and recorded any evidence of territorial behavior (e.g., threat displays between two individuals or “falling-leaf” displays; Knopf and Wunder 2006) and breeding (e.g., head bobbing, nest scrape displays, copulation, nests with eggs, or dependent young). Each colony was surveyed at least three times each year, once during each of the following intervals: 15–31 May, 1–15 June, and 16–30 June. Many colonies were also at least partially searched opportunistically up to five additional times when colonies were visited for nest searching or habitat sampling. Most colonies were searched in both years. However, one large colony searched in 2009 was not searched in 2010 because access was restricted. At colonies where plovers were detected, additional time was spent attempting to locate nests or dependent young to confirm breeding. During these efforts, plovers were observed from >50 m using binoculars, and movements and behaviors were recorded on colony maps to assist in future nest searching and for developing territory maps to delineate habitat use on large colonies (see the section “Plover territory mapping” below). When nests were found, locations were noted with a GPS unit to allow relocation for habitat sampling after eggs hatched or nests failed. Habitat sampling and colony characteristics. Each year, I conducted extensive habitat sampling on colonies during June to allow colony- and territory-level comparisons. The basic sampling unit was a 50-m, linear transect used to designate three sampling points at 0, 25, and 50 m. Within a 5-m radius of each sampling point, I measured ground cover, vegetation height, and slope. Ground cover was visually estimated as the percentage of ground covered by grass (live, dormant, or dead), live forb/subshrub, live woody shrub, live cactus, and bare ground. I combined living, dormant, and dead grasses as one cover category because this combination best described overall importance of grass at a site regardless of the level of precipitation when sampled. The live forb/subshrub category included ground covered by herbaceous forbs as well as by two perennial subshrubs, fringed sage and broom snakeweed. Mean vegetation height was calculated by measuring the height of live or dead vegetation at 12 points distributed systematically within the 5-m radius, and slope was measured across the 5-m radius using a clinometer. Within an 11.3-m radius (0.04 ha) of the point, I estimated the density of cholla (a branching cactus that reaches >1 m), woody shrubs, and prairie dog burrows by counting individual plants or burrows. I used a larger sampling radius for these three widely dispersed habitat features to increase the probability of counting at least some individuals in each sample. Habitat-sampling transects were systematically distributed to broadly cover the entire colony, with the number of transects sampled dependent on colony size. I sampled two transects on colonies up to 10 ha in area, three transects on colonies 10 to 20 ha in area, and one additional transect for every additional 20 ha in larger colonies up to a maximum of 20 transects. At each transect starting point, I determined the orientation of the transect via a random compass bearing. Orientation of transects located near colony edges were, however, constrained so they always lay completely within the boundaries of the colony. Because landscape features could also play a role in plover habitat use, each year I also used GIS maps of colony boundaries delineated during the previous winter to determine area (ha) of each colony, and calculate the proportion of the landscape within a 2-km radius of each colony center occupied by prairie dog colonies. I chose a 2-km radius rather than 1-km because colony coverage within a 1-km radius was highly correlated (Kendall's τb > 0.50) with colony area in both years. I also measured habitat characteristics at plover nest sites and at associated non-used sites to identify features important in plover nest-site selection within a territory. Within one week after eggs hatched or a nest failed, I sampled the same habitat characteristics that were measured at each point of a habitat sampling transect (see above), centering the point on the nest cup. In addition, I measured the distance from nest cups to the nearest prairie dog burrow. I then sampled the same characteristics at a non-used site located 100 m from each nest in a direction parallel to the colony edge, but still within a plover's territory. After each breeding season, I used colony maps with behavioral observations and breeding evidence to classify each colony as either occupied by breeding plovers or unoccupied. I classified colonies as occupied if nesting was confirmed (i.e., eggs or dependent young), or if I observed plovers engaged in territorial/reproductive behaviors on >1 visit in the same general region of the colony (i.e., within 300 m). Given the cryptic nature of Mountain Plovers, some individuals may have been missed. However, given the repeated search efforts and open nature of the habitat, I believe that few, if any, colonies were misclassified as unoccupied when breeding plovers were present. Plover territory mapping. In 2010 only, I used observations of plover locations and behaviors made over at least six visits to each of seven large (104 – 431 ha) occupied colonies to develop plover territory maps (Bibby et al. 1992). I then used this information to classify habitat-sampling transects completed on these colonies as either used by plovers (i.e., encompassed by a plover territory) or unused to allow for territory-level comparisons of habitat use within these colonies. Transects that crossed a boundary between used and unused areas were excluded from these analyses. Statistical analyses. Prior to analysis, I summarized habitat data for each colony by first averaging values from all three points on each habitat sampling transect, and then averaging values from all transects in that colony. For each year, I developed a series of binary logistic regression models to determine which habitat or landscape characteristics best distinguished occupied from unoccupied prairie dog colonies. I analyzed each year separately because plovers did not occupy the same colonies each year and because habitat and landscape characteristics of colonies differed between years, particularly due to differences in precipitation. To reduce problems related to excessive variables, I first compared characteristics of occupied and unoccupied colonies for each year using univariate Mann-Whitney U-tests. Only variables with P < 0.10 in at least one year were considered for model building. However, I first calculated Kendall correlations (τb) between these remaining variables to identify redundancy. I considered two variables redundant when they showed a τb > 0.40 (P < 0.05) in both years, and retained only the one variable that showed the lowest probability from the univariate tests. I then produced logistic regression models for each year using all combinations of remaining variables, including a global model (all variables combined) and a null model (intercept only), and used Akaike's Information Criterion (AIC; Burnham and Anderson 2002) with the adjustment for small sample size (AICc) to rank each model. Models were ranked as potentially important if ΔAICc (the difference between the AICc of a given model and the lowest AICc value) was ≤2.0 AIC units (Burnham and Anderson 2002). When ranked models differed by only one parameter and had essentially the same log-likelihood values, however, I excluded the larger model to avoid including uninformative parameters (Burnham and Anderson 2002, Arnold 2010). I also calculated AIC weights (wi) as a measure of the relative support for each model, and assessed the importance of variables in each ranked model by calculating 95% confidence limits (CL) for each coefficient (β); variables were considered important if they had coefficients with 95% CLs that did not bound zero. Due to limited sample sizes at remaining spatial scales, I used only univariate tests for comparisons rather than multivariate modeling approaches. For territory-level comparisons, I summarized habitat data for used (i.e., enclosed within a plover territory) or unused portions of seven large occupied colonies in 2010 by averaging values from the appropriate colony habitat sampling transects. I then used paired-sample Wilcoxon signed rank tests (Sokal and Rohlf 1995) to compare habitat characteristics on used and unused portions of these colonies. Finally, to identify features important in plover nest-site selection within a territory, I again used paired-sample Wilcoxon signed rank tests to compare habitat characteristics of nest sites and associated non-used sites. All analyses were performed using the statistical package SPSS Statistics 18 (IBM Corporation, Somers, NY). Because most comparisons involved nonparametric tests, I reported the median and range, in addition to the mean, in most cases. I considered P≤ 0.05 as the indicator of significance for all statistical tests. Occupied versus unoccupied colonies. I searched 44 prairie dog colonies (mean area = 60.7 ha, range = 0.8–483.0 ha) an average of 3.7 times each in 2009, and 43 colonies (mean area = 60.0 ha, range = 1.2–431.4 ha) an average of 3.9 times each in 2010. In both years, ∼55% of colonies were < 25 ha, and only 10 colonies were > 90 ha. Mountain Plovers occupied (i.e., showed evidence of breeding on) 14 colonies (32%) in 2009 and 13 colonies (30%) in 2010. Over both years, 17 different colonies were occupied by plovers, with 10 occupied in both years and six in only one year. One colony occupied in 2009 was not searched in 2010. Habitat and landscape characteristics differed between occupied and unoccupied colonies in both years, but the importance of some variables differed among years (Table 1). In 2009, a dry year, colonies occupied by plovers had shorter vegetation and more bare ground than unoccupied colonies. In 2010, a wet year, occupied colonies again had shorter vegetation than unoccupied colonies, but there was no difference in the extent of bare ground. Instead, occupied colonies had more grass cover, but less forb cover than unoccupied colonies. For both years, occupied colonies were larger than unoccupied colonies and in landscapes with a greater proportion of prairie dog colony coverage within 2 km (Table 1). Over both years, six variables, including grass cover, forb/subshrub cover, bare ground cover, vegetation height, colony area, and colony cover within a 2-km radius, showed a strong enough relationship with plover occupancy (P < 0.10) to potentially be included in logistic regression model building (Table 1). In both years, however, vegetation height was negatively correlated with bare ground cover and positively correlated with forb/subshrub cover (Kendall τb > 0.45, P < 0.001). As a result, of these three variables, only vegetation height was used in the regression analyses because of its greater significance in univariate tests. The remaining four variables were used to build and compare 15 logistic regression models in each year. In 2009, only two models were competitive (ΔAIC <2.0; Table 2). One of these models, however, was excluded as it added just a single parameter to the remaining model. The best approximating model contained colony area and vegetation height, and was supported by a high relative model weight (Table 2). It predicted that the probability of a colony being occupied by breeding plovers was positively associated with colony size (β= 0.028, 95% CL: 0.008, 0.048), but negatively associated with mean vegetation height (β=–1.953, 95% CL: –3.693, –0.213). In 2010, two models were again competitive (ΔAIC <2.0; Table 2). However, one of these models (the global model) was again excluded. The best approximating model contained colony area, vegetation height, and percent grass cover and was supported by a high relative model weight (Table 2). It predicted that the probability of a colony being occupied by breeding plovers was positively associated with colony size (β= 0.079, 95% CL: 0.005, 0.153) and percent grass cover (β= 16.415, 95% CL: –2.074, 34.904), but negatively associated with mean vegetation height (β=–3.093, 95% CL: –5.894, –0.292). The variable percent grass cover, however, was not strongly supported as the CL of its coefficient bounded zero. Used versus unused portions of occupied colonies. Based on habitat sampling in seven colonies in 2010, portions of occupied colonies that contained plover territories had a lower proportion of forb/subshrub cover and shorter vegetation than portions without territories, but had a greater proportion of bare ground (Table 3). Nest sites versus non-used sites. I located 21 plover nests in nine colonies (eight nests in five colonies in 2009, and 13 nests in seven colonies in 2010). On average, plovers nested on sites with a lower proportion of forb/subshrub cover and shorter vegetation, but a greater proportion of bare ground and greater slope than unused sites (Table 4). Compared to unused sites, plover nests were also located further from the nearest prairie dog burrow (Table 4). Mountain Plovers in my study selectively used prairie dog colonies for nesting, but they also exhibited selectivity among colonies, among habitat patches within colonies, and even in sites used for nesting. At the colony level, variation in both habitat- and landscape-level characteristics influenced the suitability of individual colonies as plover breeding habitat. Logistic regression models from both years identified mean vegetation height as the most important habitat variable for distinguishing occupied from unoccupied colonies, with the probability of plover occupancy declining with increasing vegetation height. Short vegetation has previously been recognized as an important feature of plover nesting habitat (Graul 1975, Ellison-Manning and White 2001, Plumb et al. 2005), and mean vegetation height was a particularly informative measure in my study because it was correlated with two other important habitat variables: bare ground and forb/subshrub cover. Basically, mean vegetation height decreased in colonies as the proportion of bare ground increased and as coverage of taller forbs and subshrubs decreased. Based on these relationships, my results suggest that colonies with shorter vegetation and, thus, a higher probability of plover occupancy would also tend to have a higher proportion of bare ground and lower proportion of forb/subshrub cover. Bare ground has often been recognized as an important component of plover breeding habitat, with a minimum of 30% bare ground typically observed in plover breeding habitats (Knopf and Miller 1994). Plovers likely prefer sites with an abundance of bare ground and short vegetation because these characteristics help plovers avoid predators. From above, the plumage of Mountain Plovers blends in with bare ground, concealing them from avian predators, and short vegetation may aid in detection of ground predators (Augustine and Derner 2012). Abundant bare ground may also be necessary to accommodate the ground-foraging behavior of plovers (Knopf and Wunder 2006). In fact, bare ground appears to be so important that, off of prairie dog colonies, bare ground may be a more important predictor of plover habitat suitability than vegetation height. Augustine and Derner (2012) determined that intensive cattle grazing that greatly reduced vegetation height still did not substitute for prairie dog activity in terms of generating plover habitat, primarily because it did not create enough bare ground. Forb/subshrub cover had a negative effect on plover habitat use in my study area, perhaps through its effect on the amount of bare ground and vegetation height. Forbs and subshrubs often replaced bare ground as they grew, and further increased mean vegetation height because they often grew taller than the dominant short grasses. Because the prevalence of forbs and subshrubs typically increases as a colony ages (Coppock et al. 1983, Archer et al. 1987, Hartley et al. 2009), an implication of this relationship is that plover habitat suitability on colonies may decline with time. However, this was apparently not the case on my study site because most larger colonies, which were also older colonies, were occupied by plovers. These results may reflect the fact that age, or time since colonization by prairie dogs, actually varies across larger colonies, with the oldest portions generally found near the central core, and younger portions near the expanding edges (Garrett et al. 1982). Thus, habitat characteristics also vary within older colonies that are still expanding. Even though older colonies, on average, may contain considerable forb/subshrub cover, plovers may be more likely to find the habitat characteristics they require in the more recently colonized grass-dominated portions (but see Augustine et al. 2008). I did not know the time since colonization for used versus unused portions of occupied colonies in my study area. The fact that used portions contained less forb/subshrub cover than unused portions, however, suggests that plovers may have favored more recently colonized areas. Landscape features, particularly colony area, also influenced colony suitability for plovers. In both years, logistic regression models indicated that the probability of a colony being occupied by plovers increased with increasing colony area. Further, univariate tests also demonstrated that occupied colonies were, on average, in landscapes with greater coverage of colonies within 2 km. These landscape features may contribute to plover habitat suitability in at least two non-mutually exclusive ways. First, small colonies, especially those isolated from other colonies, may be avoided because Mountain Plovers require colonies of some minimum size. Estimates of home range sizes of plovers breeding in prairie dog colonies in shortgrass prairie have averaged 57 to 116 ha in area (95% minimum convex polygon; Knopf and Rupert 1996, Dreitz et al. 2005). Olson-Edge and Edge (1987), however, found that plovers nested in prairie dog colonies as small as 6 ha in mixed-grass prairie. This is similar to the minimum area of occupied colonies in my study, although the two smallest occupied colonies in my study (both ∼5 ha) were also located near (<200 m) other larger, occupied colonies, potentially allowing plovers that nested on these small colonies to also use adjacent colonies for foraging or rearing young. In Colorado, adult plovers with broods regularly moved among adjacent prairie dog colonies (Dreitz et al. 2005). Secondly, larger colonies, or colonies clustered in close proximity, may be preferred because they can support multiple plover territories. In sites with large areas of potential breeding habitat, breeding pairs sometimes seem to aggregate together, with large areas unoccupied, suggesting that plovers may prefer to settle near conspecifics (Graul 1975, Knopf and Wunder 2006). Although distribution patterns attributed to social attraction are difficult to document, and could also result from limitation of high-quality nesting habitat, this social attraction hypothesis is interesting because it provides a potential explanation for why larger colonies are preferred; larger colonies are more desirable than small because they can support clusters of breeding pairs. This hypothesis also provides a purely social explanation for why some portions of large, occupied colonies might be unused, although the fact that habitat differences did exist between used and unused portions of individual colonies suggests that the vegetation attributes of a habitat are important in territory selection. At finer spatial scales (i.e., within territories or at nests), plovers typically select nest sites with considerable bare ground (Knopf and Miller 1994, Plumb et al. 2005, Schneider et al. 2006, McConnell et al. 2009), short vegetation, typically resulting from intensive grazing or fire (Graul 1975, Olsen and Edge 1985, Parrish et al. 1993, Plumb et al. 2005), and minimal slope (Graul 1975, Knowles et al. 1982, Parrish et al. 1993, Beauvais and Smith 2003). These were all characteristics of habitats used by plovers in my study, both broadly on used portions of colonies and at sites specifically associated with nesting. At both spatial scales, used sites were relatively flat and had more bare ground and shorter vegetation than unused sites. Mean slope differed between used and unused sites only at the scale of nest sites; nests were on sites with greater slope, on average, than unused sites. Although plovers typically avoid steep slopes, sites with a slight slope may be preferred for nesting over flat areas because they allow better drainage, which is potentially important because Mountain Plover nests are vulnerable to flooding (Graul 1975, Goguen 2011). Nest sites in my study contained less forb and subshrub cover than non-unused sites. In contrast, Olson and Edge (1985) found that Mountain Plovers nesting in prairie dog colonies in mixed prairie in Montana preferentially selected areas with a greater density of forbs and fringed sage, a subshrub, possibly because they made incubating birds less conspicuous. In fact, Olson and Edge (1985) hypothesized that mid-sized colonies (6–50 ha) supported optimal plover densities in their study area because these colonies maintained adequate forb and subshrub cover, whereas larger colonies were often denuded of all vegetative cover. These results suggest that forbs and subshrubs may affect plover habitat quality differently in mixed prairie than shortgrass prairie. Alternatively, the importance of these plants may vary depending on the particular species involved. Mountain Plovers in my study tended to avoid nesting close to prairie dog burrows, with plover nests typically located further from the nearest burrows than non-used sites. Using artificial nests, Baker et al. (2000) found that nest predation rates were higher on prairie dog colonies than in adjacent uncolonized habitats. Although prairie dogs were not nest predators, they may have disturbed eggs out of curiosity. Given that plover nests are typically tended by a single adult (Graul 1973, Dinsmore and Knopf 2005), and must at times be left unattended during foraging, perhaps placing nests further from prairie dog burrows minimizes such incidental disturbance. Regardless of this potentially negative interaction, prairie dog colonies are selectively used by nesting plovers throughout their range, and appear to provide productive habitats for both nesting and brood-rearing (Dinsmore et al. 2003, Dreitz 2009). My results support the general recommendation of Dinsmore (2003) that efforts to increase prairie dog populations will also increase the abundance and distribution of Mountain Plovers. However, my results also suggest that efforts should focus on creating larger colonies, or at least clusters of colonies in close proximity, to ensure that an adequate area of preferred habitat (i.e., short vegetation, abundant bare ground, and few forbs and subshrubs) is present. I thank the management of Vermejo Park Ranch for allowing me access to the property, and thank L. Goguen, L. Klomps, and R. Ward for assistance in field work. D. Long deserves special thanks for sharing his prairie dog colony data, and for creating the prairie dog colony map used in this paper. This manuscript benefited greatly from comments by D. Augustine, M. Wunder, and an anonymous reviewer, and from discussion provided by several colleagues as part of a Faculty Works-In-Progress session at the Pennsylvania State University, Hazleton campus. Funding and support for this research was provided by Faculty Research Development Grants from the Pennsylvania State University, Hazleton. All methods were approved by the Pennsylvania State University Institutional Animal Care and Use Committee.
Referência(s)