Choosing what is left: the spatial structure of a wild herbivore population within a livestock-dominated landscape

Abstract

Shrublands and grasslands comprise over 30% of the land surface and are among the most exploited ecosystems for livestock production. Across natural landscapes, the distribution and abundance of wild herbivores are affected by interspecific competition for foraging resources, hunting and the development of infrastructure among other factors. In Argentine Patagonia, the abundance of domestic sheep grazing on native vegetation outnumbers the widely distributed guanaco (Lama guanicoe) and sheep ranching monopolizes the most productive lands. In this work, we aimed to assess the spatial variation in the abundance of guanacos in Península Valdés, a representative landscape of Patagonia, investigating the incidence of natural and human-related factors. We conducted ground surveys during the austral autumn in 2017 totaling 383.4 km along areas with and without sheep ranching. We built density surface models to account for the variation in guanaco abundance and obtained a map of guanaco density at a resolution of 4 km². We estimated an overall density of 11.71 guanacos.km^-2 for a prediction area of 3,196 km², although the density of guanacos tripled in areas where sheep ranching was terminated (in around 20% of the surface of Península Valdés) compared to areas with sheep. Guanacos were more abundant at lower values of primary productivity and sheep stocking rates and further from inhabited ranch buildings, suggesting competition with sheep and conflict with humans. Although guanacos selected open, grass-dominated habitats across sheep-free sites, fences dividing properties and paddocks played a significant role in the spatial structure of their population in Península Valdés affecting negatively the abundance of guanacos. Our results indicate that actions to improve habitat connectivity for guanacos, favor the coexistence among guanacos and sheep ranching, and promote responsible human activities and attitudes towards wildlife are needed.

Keywords: Anthropic factors; Distribution and abundance; Habitat selection; Lama guanicoe; Patagonia; Península Valdes; Sheep ranching; Spatial models; Wild and domestic ungulates.

Figure 1. Location of Península Valdés and distribution of the survey transects inside the study area.

Sheep ranching areas (SHEEP) are represented in dark grey while the areas where the activity has ceased (NOSHEEP) are in light grey.

Figure 2. Spatial variation in the abundance of Lama guanicoe.

Maps of population densities according to (A) the whole PV model and (B) independent models for areas with and without sheep ranching activity (SHEEP and NOSHEEP respectively). Distribution of the coefficient of variation (CV) according to (C) whole PV model and (D) models applied to SHEEP and NOSHEEP areas. Black lines show the limits of NOSHEEP areas.

Figure 3. Partial effects of the significant predictors on the abundance of Lama guanicoe according to the best-fit model for each area analyzed.

Whole Peninsula Valdés area (A–D), sheep ranching areas (SHEEP; E–H) and areas where the activity has been ceased (NOSHEEP; I and J). The solid lines represent the estimated smoothing terms (s) of each predictor and the gray shading represents 95% confidence intervals for the mean effect. The rug ticks at the bottom of the plot indicate the coverage of the range of values of each variable in the survey area. The number in brackets in each “s” gives the effective degrees of freedom (a measure of flexibility) of each term. The y-axis is on the scale of the link function.