Community- and landscape-level responses of reptiles and small mammals to feral-horse grazing in the Great Basin

Abstract

We investigated species- and community-level responses of squamate reptiles and granivorous small mammals to feral-horse grazing in two elevational strata across nine mountain ranges of the western Great Basin, USA. Although mammal species richness did not differ between horse-occupied and horse-removed sites, occupied sites possessed less community completeness (biotic integrity) and 1.1–7.4 times greater deer mice (Peromyscus maniculatus) than removed sites. In opposite fashion, horse-removed sites possessed greater reptile species richness and tended towards greater abundance for seven of nine species, yet unequal species pools across sites dictated that community completeness did not differ statistically between horse-removed and -occupied sites.

Introduction

As understanding of the natural environment continues to improve, ecological investigation and management move toward broader spatial scales. Although ecosystem management is rapidly becoming the dominant paradigm for investigating, managing, and monitoring ecological resources, our ability to characterize landscapes and predict landscape-level responses to single or interacting disturbances remains limited. This limitation results from a paucity of broadly collected data, an incomplete understanding of higher-order interactions, and a lack of knowledge of how diverse ecosystem components respond to disturbance (Franklin, 1993; Christensen et al., 1996).

Past studies of grazing disturbance have investigated effects of various grazing regimes on plants much more frequently than on any other ecosystem component, except perhaps soils. Studies investigating only the responses of vegetation can provide insight into the direct interaction between plant and herbivore (e.g. coevolution, effects of grazing intensity or timing on the interaction), but cannot predict responses of other taxa to grazing processes such as nutrient redistribution and soil trampling. Therefore, comprehensive understanding of ecosystem responses to large-mammal herbivory remains illusive. Although some grazing-impact studies (e.g. Bock et al., 1984; Heske and Campbell, 1991) have included larger numbers of variables, they often suffer from lack of replication and compare sites in relatively close proximity to each other (Brown and McDonald, 1995). Grazing studies such as these have provided insights into site-specific changes in various components of ecosystems, understanding as to the role of grazing intensity, duration and timing, and have investigated mechanisms of change. Though small-scale investigations allow for a high degree of experimental control, they may have questionable generality, particularly if sites are not representative of conditions or habitat types dominant on the landscape. Thus, although intensive multi-variable studies are increasing in frequency and have clarified many interactions, they may not represent spatially widespread relationships among grazers and other ecosystem components.

We investigated the grazing ecology of feral horses (Equus caballus) in the western Great Basin at a landscape scale using a diverse collection of ecosystem components. Because comparing replicates across broad spatial scales encompasses large amounts of environmental variability, we developed response variables that are less influenced by local conditions and applicable across all sites. We tested the null hypothesis that within relatively narrow elevational strata, sites experiencing horse grazing did not differ in their mammal and reptile faunas from ecologically similar sites from which horses had been removed 10–14 years earlier.

Ancestral horses had been absent from North America for at least 13,000 years (Grayson, 1989) when domestic horses (E. caballus) were brought to the southwestern United States by Spanish conquistadors near the end of the 16th century (Zarn et al., 1977; Wagner, 1983). Over time, these once-domesticated horses gave rise to populations of feral horses throughout much of western North America. In recent decades, much controversy has developed concerning the role of feral horses in ecosystems of western North America. The controversy has prompted federal legislation (the Wild Free-Roaming Horse and Burro Act of 1971) designed to protect feral horses as well as continuing calls for the complete removal of feral horses from western rangeland. At present, land managers are charged with controlling the growth of horse populations and maintaining the biotic integrity of rangeland. A crucial part of the latter mandate is to acquire knowledge of the ways in which grazing by feral horses affects the composition (e.g. species diversity) and functioning of ecosystems. Due to differences in species habitat use, mode of digestion, diet, evolutionary history, and morphology, it is inappropriate to assume that effects of grazing by feral horses are similar in nature or in magnitude to effects of grazing observed for other introduced or native species (Beever, 2003). Furthermore, although numerous studies on grazing ecology of domestic horses have been published (e.g. Reiner and Urness, 1982; Menard et al., 2002), extrapolating these results to free-roaming populations may be problematic, particularly with respect to behavior.

One important indirect effect of grazing on ecosystems may be that imposed on small-mammal communities. Mammals perform a variety of ecological functions in semi-arid ecosystems, including seed dispersal, seed predation, and herbivory, as well as soil perturbation and aeration (Brown and Harney, 1993). Consequently, changes to mammal communities may also indirectly affect the plant community. Furthermore, mammals comprise a significant portion of the vertebrate biomass in sagebrush communities and may thus exhibit the greatest response to horse grazing. Changes in mammalian communities have been interpreted to indicate deterioration or simplification in many ecosystems (Carothers et al., 1976; Reynolds and Trost, 1980; Martell, 1983; Medin and Clary, 1989). In this study, we tested for an effect of feral-horse grazing on abundance of individual mammal species as well as on evenness and completeness of small-mammal communities. Evenness may be defined as the degree to which the total number of individuals is distributed evenly among all species (as in Simpson's index), and completeness defined as the proportion of the full complement of species that might occur historically or in the absence of disturbance.

In addition to small mammals, desert squamates (lizards, snakes, and amphisbaenids) occupy a broad diversity of trophic and habitat niches. Species in the Great Basin include insectivores, carnivores, and near-omnivores. Many of these species serve as important links between lower and higher trophic levels (Stebbins, 1985). Because of these links to other ecosystem components such as plants, invertebrates, and soils, loss of squamate diversity or large decreases in abundance may further simplify ecosystems. Although domestic cattle and sheep can affect the abundances of several lizard and snake species (Busack and Bury, 1974; Szaro et al., 1985; Jones, 1988), it remains unknown how the unique grazing ecology of feral horses affects reptile abundance and diversity. Presumably, grazing by feral horses can lead to soil compaction, decreases in prey abundance, and decreases in vegetative cover, all of which may affect the abundances and diversity of reptile species. The present study represents a preliminary investigation of the relationship between feral-horse grazing and the abundances and species richness of reptiles.