Trade-offs between food availability and predation risk in desert environments: The case of polygynous monomorphic guanaco (Lama guanicoe)

Journal of Arid Environments 97 (2013) 136e142

Contents lists available at SciVerse ScienceDirect

Journal of Arid Environments journal homepage: www.elsevier.com/locate/jaridenv

Trade-offs between food availability and predation risk in desert environments: The case of polygynous monomorphic guanaco (Lama guanicoe) Pablo Acebes*, Juan E. Malo, Juan Traba Terrestrial Ecology Group-TEG, Departamento de Ecología, Facultad de Ciencias, Universidad Autónoma de Madrid, C/. Darwin, 2, E-28049 Madrid, Spain

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 June 2012 Received in revised form 21 December 2012 Accepted 31 May 2013 Available online

Habitat selection by ungulates is determined by the quantity, quality and distribution of trophic resources as well as by predation risk. It may also vary in relation to species-specific reproductive strategies and social organization. The guanaco (Lama guanicoe), a highly social and sexually size-monomorphic wild camelid typical of arid lands, is ideal for evaluating behavioural responses of this type, since most studies have done on dimorphic ungulates in temperate environments, where trophic resources are abundant. We recorded the group size and social structure of guanaco during both dry and wet seasons of 2005e2007 in an Argentinean desert, where pumas (Puma concolor) are the sole predators. Remote sensing data were used to calculate five variables that reflected trophic availability and terrain morphology for each guanaco group and for an equivalent number of random controls. Habitat use did not differ between types of social groups but differed between seasons. Guanacos used less productive and less steep areas during the breeding season, irrespective of juvenile:adult ratios in the family groups, and larger groups occupied flatter areas. Overall, guanaco habitat selection prioritizes reducing predation risk to the extent that animals occupy areas offering the minimum productivity capable of meeting their energy requirements. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Habitat selection Mating system NDVI Social organization Ungulate

1. Introduction Populations of large herbivores are generally regulated via tope down mechanisms such as predation (Hopcraft et al., 2010; Sinclair et al., 2003) and by bottomeup constraints in primary production (Hopcraft et al., 2010; McNaughton et al., 1989). Large herbivores may respond to changes in resource availability or predation with behavioural adjustments (Kie, 1999; Sinclair and Arcese, 1995), which may affect demographic parameters (Creel et al., 2007). Such mechanisms operate differently within species according to their type of social organization (Jarman, 1974). Predation risk effects arise when a prey alters its behaviour in response to predators, and these behavioural responses carry costs. Indeed, risk effects can be larger than direct effects of predation, and can be influential even when the direct rate of predation is zero (Creel and Christianson, 2008). It is generally accepted that selection will favour individuals that optimally balance the benefits of risk reduction against its costs (Lima, 2002). Behavioural responses to reduce predation risk include changes in habitat use (Creel et al.,

* Corresponding author. Tel.: þ34 914978916; fax: þ34 914978001. E-mail address: [email protected] (P. Acebes). 0140-1963/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jaridenv.2013.05.017

2005) or in group size (Creel and Winnie, 2005). The response may also differ between sexes: males usually seek habitats which offer high trophic availability, whereas females with offspring select habitats that firstly offer security against predators and secondly provide abundant forage (Main et al., 1996; McCullough, 1999). Thus, habitat selection to maximize reproductive fitness represents a trade-off between maximizing foraging opportunities and minimizing predation risk (Kie, 1999; McCullough, 1999). The guanaco (Lama guanicoe Müller), a highly social mediumsized South American ungulate, is sexually monomorphic in body size and exhibits a resource defence polygyny mating system (Franklin, 1983). Three types of groups may be encountered during the breeding season (González et al., 2006): family groups (i.e. a territorial male with adult females and their offspring), groups of non-territorial males, and solitary territorial males that are seeking or defending a territory without females. Outside the breeding season, guanaco group composition varies according to environmental conditions. Sedentary populations are observed when weather and forage supply is stable, allowing populations to maintain territories all year round (Franklin, 1983; González et al., 2006). However, in areas with particularly snowy winters with a drastic reduction of food availability, guanacos may move to more sheltered areas, losing their territoriality and forming large mixed

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

herds composed of adults of both sexes, juveniles and newborns (Franklin, 1983; González et al., 2006). Most studies on ungulates have done in temperate ecosystems (Creel et al., 2005; Kunkel and Pletscher, 2000; Pierce et al., 2004; Theuerkauf and Rouys, 2008), but little information about desert ecosystems is available, where resources are extremely scarce and prey may accept higher predation risk. Predators establish a ‘landscape of fear’ (Laundré et al., 2001) whose topography is determined by the level of predation risk that prey face in different habitat types. When foraging in this landscape, prey will often shift their use from riskier to safer areas which may represent a change to poorer quality habitat, resulting in a decreased diet quality (Hernández and Laundré, 2005). However this behavioural decision might not be faced when safer habitats are of very low abundant forage. On the other hand, most hypotheses in this field have been tested on dimorphic ungulates, which generally show marked sexual segregation (Main et al., 1996; Ruckstuhl and Neuhaus, 2000), but more studies on monomorphic ungulates are needed. As Ruckstuhl and Neuhaus (2000) emphasized, non-dimorphic species may be ideal to explain sexual differences in habitat use, forage selection, predator avoidance or activity budgets, because body size effects are absent. The guanaco thus becomes an interesting species for analysing the trade-off between resource availability and predation risk as a function of its social organization. The puma, the natural predator of guanacos, is a solitary ambush and stealthy carnivore that relies on vegetation cover and terrain features to approach close enough to their prey before attacking, hiding on bushy vegetation, in steeper gradients or rocky terrain (Bank and Franklin, 1998). As pumas do not chase their prey through long distances, early detection by guanacos can be especially advantageous to increase their likelihood to escape (Marino and Baldi, 2008). It is the sole predator of guanacos in the Northwestern Argentina, where this study was conducted. The area is dominated by sparse shrubland of less than 20% of plant cover, with a very scant and seasonal herbaceous layer growing in the wet season, which together with the greening of the shrub vegetation represents a slight increase in forage availability during this season. The minimum plant cover (i.e. lowest forage abundance) of the area corresponds to habitats located on level flat ground of fine-textured substrates, while steep hard terrain contains taller and denser vegetation cover (Acebes et al., 2010b). We here use field data and satellite-based methods: (i) to firstly determine the social organization of a small and sedentary guanaco population (Acebes et al., 2010a) in both the breeding and nonbreeding seasons over three years; (ii) to evaluate whether social units differ in habitat selection as a function of trophic availability and potential predation risk; and (iii) to determine whether any patterns depend on group size, the number of offspring and the yearling/adult ratio. In this ecological context we expect to find that: (a) there will be a different habitat selection among types of guanaco social organization: male groups and solitary males will occupy more productive zones than family groups, accepting a higher predation risk given that they do not have calves (more vulnerable to predation) in an attempt to maximize their body condition in order to gain access to reproduction (Main et al., 1996) and; (b) family groups will occupy areas of low predation risk, even if they have a lower productivity, and this tendency will be more pronounced for groups with higher number of offspring and/or higher yearling/adult ratio.

137

of the IschigualastoeTalampaya World Heritage Site together with Talampaya National Park (La Rioja province). The mean altitude is 1300 m a.s.l. The park comprises an area of 60 369 ha. The nearest village is 10 km from the park limit. The climate is dry desert with a mean annual temperature below 18  C (range 10 e45  C) and a mean temperature in the hottest month above 22  C. Mean annual precipitation ranges from 80 to 140 mm, concentrated in summer (NovembereFebruary). Monte Desert is the dominant biome (Márquez et al., 2005). The predominant vegetation is sparse shrubland, dominated by species of Zygophyllaceae (Zuccagnia punctata, Larrea spp. and Bulnesia retama), Fabaceae (Prosopis spp., Geoffroea decorticans, Cercidium praecox) and Chenopodiacae (Atriplex spp. and Suaeda divaricata). Cacti (Tephrocactus spp., Opuntia sulphurea and Echinopsis spp.) and Bromeliads (Tillandsia spp. and Deuterocohnia longipetala) are also frequent but to a lesser extent (Acebes et al., 2010b). The puma is the unique predator of guanacos in the area. Throughout its geographic range the puma is commonly associated with forested areas or in dryer more-open regions, generally occurring in habitats with dense understory vegetation and with increased topographic relief (Franklin et al., 1999). During the studied period (2005e2007; 107 field days), evidence from predation of two guanaco calves, some guanaco carcasses and several recent puma tracks and faeces were recorded, and two pumas were sighted by park guards, always in habitats of steep terrain and dense shrub cover, but no puma tracks on open-flat habitats with scant or no vegetation were detected. 2.2. Guanaco data The field work was conducted in FebruaryeMarch 2005, 2006 and 2007 (wet season), when guanacos breed, and in August 2005 and 2006 (dry season), corresponding to the maximum and minimum abundance of trophic resources respectively, although these differences are not marked (Acebes, unpubl. data). Two researchers with binoculars (10  42) surveyed the area by vehicle through line e transects surveys from all the roads and tracks, covering 6700 km in total. Surveys were conducted during day-light hours, travelling at 10e40 km/h. Roads and tracks crossed all habitat types (gradient from bare ground to densest plant cover) and topography (i.e. open-flat, steep or rugged terrain), ensuring that landscape heterogeneity was properly examined. The following were recorded when a solitary guanaco or a herd was sighted: (1) exact geographical location using GPS, a laser range-finder and a precision compass; (2) group size; (3) social and age structure: the number of adult males and females, juveniles (yearlings >1 and <1.5 years old) and yearlings (up to one year old) and; (4) social unit: family group, male group, solitary male or mixed herds. Because of their sociability, large size, bright brown colour and low vegetation, guanacos were easily observed. When animals were too distant for individuals to be identified they were approached on foot until this was possible. Guanaco flight events occurred at a mean distance of 141 m (Malo et al., 2011). The local guanaco population is sedentary and comprised of fewer than 400 individuals (0.38 individuals/km2, Acebes et al., 2010a), so there was a chance that the same individual or group could be recorded more than once in the same season. 2.3. Habitat characteristics

2. Materials and methods 2.1. Study area The study was carried out in Ischigualasto Provincial Park, in San Juan province, Northwestern Argentina (29 550 S, 68 050 W), part

Habitat variables that could be important to guanaco were obtained from a Landsat 7 ETM þ image, acquired on February 26 2002, and from an ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) digital elevation model (DEM). Both images had a resolution of 30  30 m. The Normalised Difference

138

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

mean and unit variance), which is necessary when variables are measured in very different units and the differences between variances are not the focus of the analysis (Quinn and Keough, 2002). Multivariate analyses of variance (MANOVA) were employed to evaluate whether habitat use by different social units of guanaco varied as a function of season or year, using the first two components obtained from the PCA as response variables and the type of guanaco social unit, year and season as predictor variables. A MANCOVA was also undertaken to determine whether habitat use by family groups differed as a function of their group size and the number of calves, with year and season as categorical predictors, and group size and the yearling/adult ratio as continuous predictors. We tried substituting the yearling/adult ratio for the number of yearlings or the number of offspring (yearlings þ juveniles) and results were unaltered. Tukey’s honestly significant difference test for unequal samples sizes was used a posteriori to compare pairs of means. Group size was square root-transformed and gradient log-transformed. Analysis of habitat use with respect to availability was done in two stages, as outlined above. The first PCA employed values of the five variables for guanaco observations (N ¼ 114) and control sites (N ¼ 114). Differences between use and availability were analysed by a MANOVA, in which the first two components of the PCA were introduced as response variables and the guanaco observation factor versus control point was the categorical predictor. The second analysis repeated the process but this time using only availability data from within the protected area. The Varimax rotation was used to allow a better interpretation of the components. A previous analysis comparing 350 to 114 random points to define availability was performed, and results were unaltered, therefore we decided to balance the statistical design with the number of guanaco observations. All analyses were performed with STATISTICA 8.0 (StatSoft, 2007).

Vegetation Index (NDVI) was calculated from the red: near-infrared reflectance ratio (NDVI ¼ [NIR  RED]/[NIR þ RED]). NDVI is especially useful due to the strong correlation between aboveground net primary productivity and absorbed photosynthetically active radiation (Kerr and Ostrovsky, 2003). We also used information derived from band 1 (0.45e0.52 mm) and band 5 (1.55e 1.75 mm) of the same image. Band 1 discriminates between bare ground and vegetation (Kerr and Ostrovsky, 2003): high values represent a high proportion of bare ground, giving additional information on forage abundance where ecosystem productivity is very low (Pettorelli et al., 2005). Band 5 is associated with soil moisture and vegetation (Kerr and Ostrovsky, 2003): low values indicate high moisture, which may influence the quantity of the plant forage available to herbivores (Olff et al., 2002). Using these bands, together with NDVI, we ensured that all vegetation types were registered. Terrain gradient was obtained, and its ruggedness estimated, from the DEM. Ruggedness was determined by means of the ruggedness measure vector proposed by Sappington et al. (2007). This is a more appropriate estimate than other indices since it is derived from aspect and not slope, allowing both to be independent variables in the analyses. The process employs a 3  3 pixel window moving over the terrain model. Index values range from 0 to 1, representing very smooth and very rough zones respectively. See Hansen et al. (2009) for a similar approach. We therefore used remote sensing data to properly describe the habitat structure in terms of vegetation cover and morphology terrain, assuming that vegetation cover together with steep terrain are perceived as an obstructive cover by guanacos, as pumas can conceal and approach to prey without being detected (Marino and Baldi, 2008). Habitat availability was analysed in two steps. In the first step, as many random points (i.e. pseudo-absences) as there were guanaco sightings were established on the viewshed model obtained by the DEM from the same surveyed roads and tracks, in a buffer of 500 km around them, considering the suitable range distance to detect animals according to habitats structure. In the second, given that 98% of guanaco observations fell within the protected zone, a new set of random points were established solely for the area that fell within the park. Habitat selection by guanaco is thus explained firstly in terms of availability within the entire prospected area and secondly within the park. The rationale is that the species’ absence outside the protected area may be due to poaching and not to habitat characteristics. Five environmental variables (NDVI, band 1, band 5, slope and ruggedness) were determined in both estimates for each random point and guanaco observation. Spatial and remote sensing analyses employed ArcGIS 9.3 (ESRI, USA) and ERDAS IMAGINE 9.1 (Leica Geosystems) software.

The major social unit was the family group (63.2%), followed by solitary males (22.8%) and male-only groups (14%) (Table 1). Family group size did not differ either between seasons (ManneWhitney U-test: U ¼ 597.49, P ¼ 0.995) or years (KruskaleWallis test: H ¼ 2.63, P ¼ 0.268). There were similarly no differences in the sizes of male-only groups between seasons (MeW test: U ¼ 28.50, P ¼ 0.713) and years (KeW test: H ¼ 1.09, P ¼ 0.581) (Table 1). Yearling/adult ratio varied between seasons (ANOVA, F1,67 ¼ 6.62, P ¼ 0.012) and years (F2,67 ¼ 3.79, P ¼ 0.027), being higher in the breeding season (mean  SD; 0.40  0.26) than in the non-breeding season (0.22  0.24) (Table 1).

2.4. Data analysis

3.2. Habitat use by guanaco social units

The average value of each variable was calculated for all observations of the same individual or herd in the same season (min ¼ 1, max ¼ 5). The sample size of the 324 field sightings was thus reduced to 114, so avoiding pseudoreplication. Group size differences between seasons (breeding/non-breeding) and years (2005/ 2006/2007) were analysed via non parametric ManneWhitney U and KruskaleWallis tests, respectively. Differences in the yearling/ adult ratio between family groups were analysed by means of an ANOVA with season and year as factors. A Principal Components Analysis (PCA) with the five environmental variables was carried out to obtain orthogonal variables (i.e. food availability and potential predation risk) that represent the bidimensional space in which guanaco habitat use occurs. A correlation matrix was chosen (based on variables standardised to zero

The first two components of the PCA accounted for 73.7% of the variance. The first component (PCA 1) represents a plant productivity gradient: zones with a high proportion of bare ground and low soil moisture () versus zones with plant cover (þ) (Table 2). The second component (PCA 2) comprises a topographic gradient: rugged, level zones (þ) versus smooth, steep slopes (). Habitat use by guanacos as defined by these two PCA components revealed no differences between types of social unit (MANOVA, F4,214 ¼ 1.01, P ¼ 0.405) (Fig. 1a) nor between years (F4,214 ¼ 1.12, P ¼ 0.349), but there were significant differences between seasons (F2,107 ¼ 3.95, P ¼ 0.022) (Fig. 1b). The ruggedness variable was the only one to give significant results in post hoc tests. During the breeding season, guanacos occupied level areas of lower plant productivity, the converse being the case during the non-breeding season (Fig. 1b).

3. Results 3.1. Demography and social organization

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

139

Table 1 Guanaco social groups recorded during the breeding (BS) and non-breeding seasons (NBS) during the three study years. Mean group size is given, with percentage of occurrence of each social group per season in parentheses. For family groups mean number of yearlings and juveniles and the yearling/adult ratio (Y/A) are given. Social unit

Family groups Male groups Solitary males Total Yearlings Juveniles Y/A Ratio

2005

2006

2007

BS

NBS

BS

NBS

BS

6.29 (68.8) 4.25 (12.5) 1 (18.8) 4.89 (100) 1.82 2.64 0.56

7.14 (58.3) 3.17 (25) 1 (16.7) 5.12 (100) 1.00 2.07 0.25

8.27 (76.5) 2.00 (5.9) 1 (17.6) 6.62 (100) 1.62 2.15 0.24

8.18 (57.1) 2.25 (9.5) 1 (33.3) 5.22 (100) 1.50 2.08 0.17

8.21 (61.1) 3.60 (13.9) 1 (25) 5.77 (100) 2.32 3.14 0.43

The analysis centred on family groups showed the same pattern as for all guanacos combined, with inter-seasonal differences in habitat use (MANCOVA, F2,63 ¼ 4.38, P ¼ 0.017) and as a function of group size (F2,63 ¼ 3.94, P ¼ 0.024). The plant productivity variable was the only one to give significant results in a posteriori tests on the effect of season, whereas larger groups occupied more level areas. There were no significant results for the effects of year (F4,126 ¼ 0.83, P ¼ 0.510), yearling/adult ratio (F2,63 ¼ 1.43, P ¼ 0.247), number of yearlings (F2,63 ¼ 1.21, P ¼ 0.310) or the total number of offspring (F2,63 ¼ 0.42, P ¼ 0.660).

Total

7.72 (63.2) 3.36 (14) 1 (22.8) 5.54 (100) 1.72 2.5 0.34

proportion of family groups throughout the year (63%, Table 1), compared to other guanaco populations (31%, Puig and Videla, 1995). The detected stability in the number and structure of guanaco social organization throughout the study period may be influenced by the monomorphic nature of the species, along with the scarce forage availability. In general, polygynous ungulates display marked sexual dimorphism, and during the breeding season males have temporary territories or harems, and less commonly form leks, but they do not defend their females from rivals outside the reproductive period (Emlen and Oring, 1977),

3.3. Habitat selection The first two components of the PCA for all guanaco observations and the random control points accounted for 71.3% of the variance. These two components were interpreted as proxies of trophic availability and predation risk. Habitat selection by guanacos was non-random (MANOVA, F2,225 ¼ 11.92, P < 0.001) (Fig. 2a). Post hoc tests showed significant differences only for plant productivity, this being greater in control points (Table 3). The variance absorbed by the second PCA, with the control points restricted to the protected zone, was 74.9%, with the same component configuration. The MANOVA once again showed differences (F2,225 ¼ 3.73, P ¼ 0.026) (Fig. 2b), which were significant in post hoc tests, in this case between guanaco observation points and control points for the topography component. 4. Discussion Our results show an apparent year-long stability in types of guanaco social organization as well as that the different social units do not differ in habitat use in this hyperarid ecosystem. Habitat use by guanacos nevertheless changed between seasons, the larger groups selecting areas of low forage abundance but with level terrain during the breeding season. 4.1. Demography and social organization Our results point to the stability of different social units, showing no mixed groups at any time of year, and the relative high Table 2 Factor loadings of the first two components of the PCA obtained from NDVI, Band 1, Band 5, slope and ruggedness (VRM). Variable

PCA 1

PCA 2

NDVI Band 1 (barren ground) Band 5 (dryness) Logslope VRM (ruggedness) Explained variance Proportion of total (%)

0.875* 0.904* 0.855* 0.193 0.064 2.353 47.1

0.019 0.268 0.019 0.730* 0.851* 1.329 26.6

Asterisks indicates significant values (p < 0.05). Variable slope was log-transformed.

Fig. 1. Means and standard deviations of factor scores of guanaco observations in the two resultant PCA axes (a) as a function of social group type, (b) for the breeding (BS) and non-breeding seasons (NBS). The continuous line represents axis 1: Vegetational gradient (high values) vs bare ground (low values). The broken line represents axis 2: level but rugged areas (high values) vs steep but smooth areas.

140

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

The apparent year-long stability of social organization together with the scarcity of trophic resources of the area (NDVI z 0.1, Table 3) not spatially clumped (Acebes et al., 2010b) suggests that resource defence polygyny mating system described for this species (Franklin, 1983) may be modified into a female (harem) defence mating system. Previous studies in Patagonia have shown that breeding males monopolize territories around ephemeral wetlands which offer abundant forage and excellent visibility during the breeding season (Bank et al., 2003). However, where resources are scant and not patchily distributed it is unenergetically defendable for a male which has to expend most of its time budget on such defence (Emlen and Oring, 1977). Although not previously described in the guanaco, shifts in types of mating systems as a function of population density, availability of resources and females (Emlen and Oring, 1977) have been documented in several ungulates, including elk Cervus elaphus (Carranza et al., 1996), fallow deer Dama dama (Apollonio et al., 1992), topi Damaliscus lunatus (Gosling, 1991) and wild ass Equus africanus (Moehlman, 1998). It would thus be interesting to collect data from different regions to test whether these shifts really occur and under which conditions such changes take place in guanacos. Despite the harsh environment, the yearling/adult ratio in the breeding season and the occurrence of small family groups suggest that all females succeed in getting pregnant. The yearling/adult ratio in the breeding season estimated in the present study (0.41) is relatively high for a low density population when compared to other guanaco populations in Chilean Patagonia (0.32; own unpublished data) where the population density is higher (43 individuals/km2, Sarno and Franklin, 1999). This may indicate that adopting female defense polygyny is an effective strategy. Nevertheless, environmental rigour, in the forms of trophic scarcity and extreme climatic conditions, along with predation, may account for the sharp decline in this ratio during the non-breeding season (0.21) (Bank et al., 2002; Sarno and Franklin, 1999; Sarno et al., 1999) as well as for its interannual variation. 4.2. Habitat selection Fig. 2. Means and standard deviations of factor scores for guanaco observation sites and for control areas in the two resulting axes of the PCA (a) in the total area prospected and (b) within the protected area only. The continuous line represents axis 1: Vegetational gradient (high values) vs bare ground (low values). The broken line represents axis 2: level but rugged areas (high values) vs steep but smooth areas.

when instead two sexes segregate (Main et al., 1996; Ruckstuhl and Neuhaus, 2000). In guanacos both sexes have the same risk of predation as well as similar trophic requirements (Ruckstuhl and Neuhaus, 2000). The low density and the sedentary nature of this population (Acebes et al., 2010a) may also influence the stability of guanaco social organization. Indeed guanaco social behaviour in this harsh environment is similar to some equid species (Equus burchelli, Equus zebra, Equus przewalskii, Equus caballus), nonruminant ungulates that maintain large harems year-round (Linklater, 2000; Neuhaus and Ruckstuhl, 2002). Table 3 Values (mean  SD) of the five environmental variables calculated from the untransformed remote sensing data. Values are given for guanaco locations, control locations within the protected area and for the entire prospected zone (N ¼ 114 in all cases). Variable

Guanacos

Control (park)

Control (total)

NDVI Band 1 (barren ground) Band 5 (dryness) VRM (ruggedness) Slope

0.098  0.023 113.118  21.937

0.105  0.027 112.294  22.966

0.128  0.032 103.339  21.135

135.782  29.290 0.497  0.164 2.146  2.317

135.913  36.254 0.516  0.192 3.539  3.372

123.839  33.523 0.515  0.176 4.085  5.195

Foraging patterns, as defined by the quantity, quality and availability of food (Hansen et al., 2009), are limited by the conditions within the protected area, where forage is less abundant than in the surrounding areas (Table 3). Bearing in mind that desert ecosystems offer sparse plant cover of poor nutritional value (NoyMeir, 1973), bottomeup processes led by forage scarcity tend to limit herbivore abundance (McNaughton et al., 1989) rather than topedown constraints (Sinclair et al., 2003). Thus, the low guanaco density would account for the low density of its predator (Hopcraft et al., 2010). Nevertheless, although predation may have a small regulatory effect on prey at a population level, behavioural responses of the latter when confronting predation risk may be significant (Creel and Christianson, 2008; Lima, 2002). Regarding our expectation that habitat selection differs between guanaco social units, male-only groups and solitary males selecting habitats which offer high quality forage irrespective of a high predation risk (Bank et al., 2003; Main et al., 1996) and family groups occupying habitats that minimize predation risk (Bank et al., 2003), is not confirmed by our results. The explanation may rest in the scant variation in foraging niches in desert areas: resources are not clumped into high-quality patches but rather distributed along gradients of limited plant productivity, as commented above (Acebes et al., 2010b). Our results nevertheless suggest that, when compromising between maximising forage benefits and reducing predation risk, guanacos incline towards reducing the latter by occupying minimally productive areas that can meet their energy requirements although possibly at significant cost (Creel and

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

Christianson, 2008). Seasonality was in fact one of the most important factors shaping guanaco habitat selection. Thus, during the wet-breeding period, guanacos, especially the larger groups, occur in level areas of low NDVI, i.e. with low forage abundance and a high proportion of bare ground, but high visibility to properly detect pumas. Indeed, in these areas no pumas where detected at all. With this decision, guanacos prevent or minimize the ambush hunting strategy of pumas in two ways: (i) pumas do not find the required stalking cover (i.e. dense vegetation cover or steep terrains) to hide and approach close enough to prey before attacking (Laundré, 2010) and (ii) guanacos early detect pumas increasing their likelihood to escape, as pumas do not chase their prey through long distances. In other words, guanacos select open, low-risk “refuge” habitats where puma lethality is low. On the contrary, guanacos in a poorer physical condition or guanacos without calves might spend more time in riskier habitat with higher food availability. Laundré (2010) has recently shown that deer spend most of their time in low-risk open areas (poorer in food resources), while pumas spend most of their time in forest patches, that is, riskier habitats for deer (but rich in food resources), increasing pumas success rate. Guanacos of Chilean Patagonia (Bank et al., 2003) and guanacos and vicuñas (Vicugna vicugna) of the Argentine Puna (Cajal, 1989) occur in open level areas with abundant grazing, and avoid areas associated with higher predator density, such as those with dense vegetation, steeper gradients or rocky terrain. Alternatively, puma predation on guanacos in Patagonia was higher than expected in tree and shrub habitats than in open-grassland ones, providing such habitats more successful hunting opportunities (Bank et al., 2002). However, our results show that the trade-off between low forage abundance but high visibility (i.e. refuge habitats) cannot be maintained in the dry season, when trophic resources are extremely scarce (Sinclair and Arcese, 1995). As forage abundance decreases in this season, guanacos spend more time searching for food to fulfil their energy requirements, resulting in less time allocation to antipredator vigilance (Marino and Baldi, 2008). Thus animals move to areas of steep terrain with greater plant cover despite the greater predation risk there, that is, areas where pumas were detected. It is likely that other factors, and not only predation risk, might account for these seasonal habitat shifts, such as the need for particular nutrients found in the growing forbs during the wet season. Dietary studies are needed to clarify this question, but germination of annual plants in the wet season, although very scant, was more relevant in steep terrain with denser vegetation cover, particularly under shrubs (Acebes, unpub. data), thanks to the ‘nurse plant syndrome’, i.e. higher nutrient content, shade and lower soil water evaporation under shrubs which may further facilitate germination of seeds and growth of seedlings (Armas and Pugnaire, 2005). Results of family groups reinforce the general pattern obtained from all guanaco observations, the larger groups occupying the more level areas. An alternative explanation to just predation risk effects might be related to the density of conspecifics i.e. a density-dependent effect: smaller groups are displaced towards steeper areas by intraspecific competition with larger groups. Nevertheless this argument does not seem consistent since guanaco density in the area is low and territory defense is weak, or even non-existent under context of potential female mating system (Emlen and Oring, 1977). On the other hand, we found no differences with respect to yearling/adult ratio, yearling numbers, or total offspring numbers (yearlings þ juveniles). Marino and Baldi (2008) have shown that the larger the guanaco family group the less time invested in vigilance and the longer spent foraging. Larger groups may thus occupy less productive habitats in the breeding season while minimizing predation risk, as commented above. Other ungulates, such as the elk (C. elaphus), alter their habitat

141

selection from forest areas to less favourable grasslands as a form of antipredator behaviour (Creel et al., 2005; Laundré, 2010). Finally, further studies should focus on other sexually sizemonomorphic ungulates in desert regions, to improve the knowledge of the factors that shape their habitat selection when trophic resources are very scarce and predation risk exists.

Acknowledgements This research was funded by a Biological Conservation project from the BBVA Foundation (INTERMARG Project). Partial support for UAM researchers was provided by the REMEDINAL-2 research network (S-2009/AMB/1783). The authors would further like to thank the Ischigualasto Park staff for their help.

References Acebes, P., Traba, J., Malo, J.E., Ovejero, R., Borghi, C.E., 2010a. Density and habitat use at different spatial scales of a guanaco population (Lama guanicoe) in the Monte desert of Argentina. Mammalia 74, 57e62. Acebes, P., Traba, J., Peco, B., Reus, L., Giannoni, S.M., Malo, J.E., 2010b. Abiotic gradients drive floristic composition and structure of plant communities in the Monte Desert. Revista Chilena de Historia Natural 83, 395e407. Apollonio, M., Festa-Bianchet, M., Mari, F., Mattioli, S., Sarno, B., 1992. To lek or not to lek: mating strategies of male fallow deer. Behavioral Ecology 3, 25e31. Armas, C., Pugnaire, F.I., 2005. Plant interactions govern population dynamics in a semi-arid plant community. Journal of Ecology 93, 978e989. Bank, M.S., Franklin, W.L., 1998. Mountain lion (Puma concolor patagonica) feeding observations and attacks on guanacos (Lama guanicoe). Mammalia 62, 599e605. Bank, M.S., Sarno, R.J., Campbell, N.K., Franklin, W.L., 2002. Predation of guanacos (Lama guanicoe) by southernmost mountain lions (Puma concolor) during historically severe winter in Torres del Paine National Park, Chile. Journal of Zoology 258, 215e222. Bank, M.S., Sarno, R.J., Franklin, W.L., 2003. Spatial distribution of guanaco mating sites in southern Chile: conservation implications. Biological Conservation 112, 427e434. Cajal, J.L., 1989. Uso de hábitat por vicuñas y guanacos en la reserva San Guillermo, Argentina. Vida Silvestre Neotropical 2, 21e31. Carranza, J., Fernandez-Llario, P., Gomendio, M., 1996. Territoriality in rutting red deer. Ethology 102, 793e805. Creel, S., Winnie, J.A., 2005. Responses of elk herd size to fine-scale spatial and temporal variation in the risk of predation by wolves. Animal Behaviour 69, 1181e1189. Creel, S., Winnie, J., Maxwell, B., Hamlin, K., Creel, M., 2005. Elk alter habitat selection as an antipredator response to wolves. Ecology 86, 3387e3397. Creel, S., Christianson, D., Liley, S., Winnie, J.A., 2007. Predation risk affects reproductive physiology and demography of elk. Science 315, 960. Creel, S., Christianson, D., 2008. Relationships between direct predation and risk effects. Trends in Ecology and Evolution 23, 194e201. Emlen, S.T., Oring, L.W., 1977. Ecology, sexual selection and the evolution of mating systems. Science 197, 215e223. Franklin, W.L., 1983. Contrasting socioecologies of South American’s wild camelids: the vicuña and guanaco. In: Eisenberg, J.F., Kleiman, D.G. (Eds.), Advances in the Study of Mammalian Behavior, Special Publication No. 7, American Society of Mammalogists. Allen Press, Kansas, pp. 573e629. Franklin, W.L., Johnson, W.E., Sarno, R.J., Iriarte, J.A., 1999. Ecology of the Patagonia puma Felis concolor patagonica in southern Chile. Biological Conservation 90, 33e40. González, B.A., Palma, R.E., Zapata, B., Marín, J.C., 2006. Taxonomic and biogeographic status of guanaco Lama guanicoe (Artiodactyla, Camelidae). Mammal Review 36, 157e178. Gosling, L.M., 1991. The alternative mating strategies of male topi, Damaliscus lunatus. Applied Animal Behaviour Science 29, 107e119. Hansen, B.B., Herfindal, I., Aanes, R., Saether, B.-E., Henriksen, S., 2009. Functional response in habitat selection and the trade-offs between foraging niche components in a large herbivore. Oikos 118, 859e872. Hernández, L., Laundré, J.W., 2005. Foraging in the ‘landscape of fear’ and its implications for habitat use and diet quality of elk Cervus elaphus and bison Bison bison. Wildlife Biology 11, 215e220. Hopcraft, J.G.C., Olff, H., Sinclair, A.R.E., 2010. Herbivores, resources and risks: alternating regulation along primary environmental gradients in savannas. Trends in Ecology and Evolution 25, 119e128. Jarman, P.J., 1974. The social organisation of antelope in relation to their ecology. Behaviour 48, 215e266. Kerr, J.T., Ostrovsky, M., 2003. From space to species: ecological applications for remote sensing. Trends in Ecology and Evolution 18, 299e305. Kie, J.G., 1999. Optimal foraging risk of predation: effects on behavior and social structure in ungulates. Journal of Mammalogy 80, 1114e1129.

142

P. Acebes et al. / Journal of Arid Environments 97 (2013) 136e142

Kunkel, K.E., Pletscher, D.H., 2000. Habitat factors affecting vulnerability of moose to predation by wolves in southeastern British Columbia. Canadian Journal of Zoology 78, 150e157. Laundré, J.W., Hernández, L., Altendorf, K.B., 2001. Wolves, elk, and bison: reestablishing the “landscape of fear” in Yellowstone National Park, USA. Canadian Journal of Zoology 79, 1401e1409. Laundré, J.W., 2010. Behavioral response races, predatoreprey shell games, ecology of fear, and patch use of pumas and their ungulate prey. Ecology 91, 2995e3007. Lima, S.L., 2002. Putting predators back into behavioral predatoreprey interactions. Trends in Ecology and Evolution 17, 70e75. Linklater, W.L., 2000. Adaptive explanation in socio-ecology: lessons from the Equidae. Biological Reviews of the Cambridge Philosophical 75, 1e20. Main, M.B., Weckerly, F.W., Bleich, V.C., 1996. Sexual segregation in ungulates: new directions for research. Journal of Mammalogy 77, 449e461. Malo, J.E., Acebes, P., Traba, J., 2011. Measuring ungulate tolerance to human with flight distance: a reliable visitor management tool? Biodiversity and Conservation 20, 3477e3488. Márquez, J., Martínez-Carretero, E., Dalmasso, A., Pastrán, G., Ortiz, S., 2005. Las áreas protegidas de la provincia de San Juan II. La vegetación del Parque Provincial Ischigualasto. Multequina 14, 1e27. Marino, A., Baldi, R., 2008. Vigilance patterns of territorial guanacos (Lama guanicoe): the role of reproductive interests and predation risk. Ethology 114, 413e423. McCullough, D.R., 1999. Density dependence and life-history strategies of ungulates. Journal of Mammalogy 80, 1130e1146. McNaughton, S.J., Oesterheld, M., Frank, D.A., Williams, K.J., 1989. Ecosystem-level patterns of primary productivity and herbivory in terrestrial habitats. Nature 341, 142e144. Moehlman, P.D., 1998. Feral asses Equus africanus: intraspecific variation in social organization in arid and mesic habitats. Applied Animal Behaviour Science 60, 171e195. Neuhaus, P., Ruckstuhl, K.E., 2002. The link between sexual dimorphism, activity budgets, and group cohesion: the case of plain zebra (Equus burchelli). Canadian Journal of Zoology 80, 1437e1441.

Noy-Meir, I., 1973. Desert ecosystems: environments and producers. Annual Review of Ecology and Systematics 4, 25e51. Olff, H., Ritchie, M.E., Prins, H.H.T., 2002. Global environmental controls of diversity in large herbivores. Nature 415, 901e904. Pettorelli, N., Vik, J.O., Mysterud, A., Gaillard, J.-M., Tucker, C.J., Stenseth, N.C., 2005. Using the satellite-derived NDVI to assess ecological responses to environmental change. Trends in Ecology and Evolution 20, 503e510. Pierce, B.M., Bowyer, R.T., Bleich, V.C., 2004. Habitat selection by mule deer: forage benefits or risk of predation? Journal of Wildlife Management 68, 533e541. Puig, S., Videla, F., 1995. Comportamiento y organización social del guanaco. In: Puig, S. (Ed.), Técnicas para el manejo del guanaco. IUCN, Buenos Aires, Argentina, pp. 97e118. Quinn, G.P., Keough, M.J., 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, Cambridge. Ruckstuhl, K.E., Neuhaus, P., 2000. Sexual segregation in ungulates: a new approach. Behaviour 137, 361e377. Sappington, J.M., Longshore, K.M., Thompson, D.B., 2007. Quantifying landscape ruggedness for animal habitat analysis: a case study using bighorn sheep in the Mojave Desert. Journal of Wildlife Management 71, 1419e1426. Sarno, R.J., Franklin, W.L., 1999. Population density and annual variation in birth mass of guanaco in southern Chile. Journal of Mammalogy 80, 1158e1162. Sarno, R.J., Clark, W.R., Bank, M.S., Prex, W.S., Behl, M.J., Johnson, W.E., Franklin, W.L., 1999. Juvenile guanaco survival: management and conservation implications. Journal of Applied Ecology 36, 937e945. Sinclair, A.R.E., Arcese, P., 1995. Population consequences of predation sensitive foraging: the Serengeti wildebeest. Ecology 76, 882e891. Sinclair, A.R.E., Mduma, S., Brashares, J.S., 2003. Patterns of predation in a diverse predator prey system. Nature 425, 288e290. StatSoft, Inc., 2007. STATISTICA (Data Analysis Software System), Version 8.0. www. statsoft.com. Theuerkauf, J., Rouys, S., 2008. Habitat selection by ungulates in relation to predation risk by wolves and humans in the Bia1owiez_ a Forest, Poland. Forest Ecology and Management 256, 1325e1332.