Spatial Patterns and Population Performance of Mule Deer: Responses to Water Provisioning in Mojave National Preserve, California

There are four habitat components essential for vertebrate species: food, cover, space, and water (Mackie 1981). In areas where water is limited, but those other 3 components are readily available, the provisioning of water is expected to benefit populations of wildlife. Development of water sources in arid regions of the western United States has been a common practice used by wildlife and range managers in both state and federal agencies since the early 20th century (Broyles 1995, Rosenstock et al. 2001, Krausman et al. 2006). As evidence of the considerable popularity of this practice, water development programs were present in 10 of 11 state wildlife agencies in 1997, accounting for over 6000 developed water sources (Rosenstock et al. 1999). Despite widespread use of water developments in wildlife and range management, few empirical studies have adequately evaluated the effects of those water sources on wildlife ecology (Broyles 1995, Krausman et al. 2006, Cain et al. 2008). Broyles (1995) questioned the utility of water developments, and further suggested the necessity, benefits, and harmful side effects had not yet been evaluated. Conversely, others contend that water developments provide intrinsic benefits to wildlife populations (Rosenstock et al. 1999, Bleich 2005, Krausman et al. 2006). Although the provision of water for wildlife has developed into a controversial topic, investigators agree that there is a need for more experimental research evaluating the influence of water developments on species ecology (Broyles 1995, Broyles and Cutler 1999, Rosenstock et al. 1999, Krausman et al. 2006, Simpson et al. 2011).Over the last 2 decades, research evaluating the effects of water developments on wildlife has increased (Krausman and Etchberger 1995, Broyles and Cutler 1999, Dolan 2006, Marshal et al. 2006a, Cain et al. 2008). None of those studies, however, have conclusively determined the ecological benefits of anthropogenic water developments on wildlife. In a study of desert sheep (Ovis canadensis nelsoni) in Arizona, USA, Broyles and Cutler (1999) argued that desert sheep obtained adequate water to meet metabolic demands from available forage and, thus, were not reliant on available free-water. Nonetheless, Rosenstock et al. (2001) identified several flaws in their experimental design and interpretation of data, and ultimately determined the conclusions of Broyles and Cutler (1999) to be "useless." Cain et al. (2008b) evaluated the response of desert sheep to the removal of water developments and found little change in spatial patterns or population performance. The results of that research were, however, confounded by cool temperatures and above-average precipitation during the treatment (water-removal) phase of their experiment (Cain et al. 2008). Other published literature has been mostly subjective and the conclusions anecdotal (Dolan 2006). Based on the lack of definitive results, a long-term, experimental study was needed to account for factors such as environmental stochasticity encountered in Cain et al.'s (2008b) research or the design flaws inherent in previous research (Broyles and Cutler 1999, Rosenstock et al. 2001). The goal of my research was to identify the effects of water developments on mule deer (Odocoileus hemionus) inhabiting a Mojave Desert ecosystem. I focused my research efforts on those ungulates because many water development projects are implemented to induce changes in their spatial patterns and population performance (Broyles 1995, Krausman and Etchberger 1995, Dolan 2006, Cain et al. 2008). The results of my research could have profound implications for wildlife management in desert ecosystems, as well as for water management in the west. Mule deer are widely distributed throughout western North America and occupy a variety of habitat types, including the Canadian boreal forest, the Great Basin Desert, the Colorado Plateau, and the Mojave Desert (Wallmo 1981). The ability to adapt to extreme temperatures and precipitation gradients distinguishes mule deer from many other species of ungulate (Wallmo 1981). In areas of low habitat productivity, such as desert ecosystems, mule deer require large areas to maintain viability of populations (Marshal et al. 2006b). Nutritional quality and availability of forage (Rautenstrauch and Krausman 1989, Marshal et al. 2005), cover (Ordway and Krausman 1986), natal sites (Fox and Krausman 1994), and availability of free-standing water (Marshal et al. 2006a) are all used to assess suitability of those large areas for long-term persistence of mule deer. During times of water scarcity, mule deer are particularly reliant on the availability of free-standing water (Rautenstrauch and Krausman 1989, Rosenstock et al. 1999). In arid regions and particularly during the hot-dry season, the availability of permanent sources of water may be the most important component of habitat for mule deer. Spatial distributions of mule deer are intimately linked to availability of resources on the landscape. Seasonal changes in space use due to changes in resource availability have been documented in a variety of ungulate species including desert sheep (Cain et al. 2008) and desert mule deer (O.h. crooki; Relyea et al. 2000). During times of water scarcity, typically the hot-dry season, water content of forage is also limited. Mule deer have been reported to change distribution and home range to incorporate permanent sources of water during the hot-dry season (Rautenstrauch and Krausman 1989). Conversely, when permanent water sources are unavailable, mule deer have been shown to increase daily movements and home range size to locate sources of water (Hervert and Krausman 1986), which is likely an expensive energetic allocation. Mule deer inhabiting resource-limited environments exhibit larger home ranges and increased movements to meet their energetic demands and, alternatively, they exhibit smaller home ranges and decreased movements in environments with abundant resources (Ordway and Krausman 1986, Relyea et al. 2000, Marshal et al. 2006a, Bender et al. 2007). If water is a limiting resource in desert landscapes inhabited by mule deer, then patterns of space use would be influenced by changes in the availability of that resource.Changes in density and distribution of ungulate populations are often linked to changes to changes in body condition, productivity, and survival (McCullough 1979, Eberhardt 2002, Cook et al. 2007, Bishop et al. 2009). Moreover, populations constrained by density-dependent processes exhibit poor body condition, low productivity, and low survival when at or near ecological carrying capacity (McCullough 1979, Kie et al. 1980, Stewart et al. 2005). The underlying mechanisms of population dynamics dictating density dependence are driven primarily by the availability and quality of resources on the landscape (Kie and White 1985, Stewart et al. 2002, Bender et al. 2007, Bishop et al. 2009, Parker et al. 2009). Mule deer inhabiting desert ecosystems occur at low densities with large ranges (Marshal et al. 2006b) presumably due to limited availability of resources on the landscape (Marshal et al. 2005, Bender et al. 2007, Bleich et al. 2010). If availability of permanent sources of water is limiting in desert environments, then mule deer should reduce their effort spent acquiring free-water when sources of permanent water are provided. Mule deer could then reallocate those efforts to foraging, which could, ultimately, improve body condition, productivity, and survival. In chapter 1, I evaluated movements, distribution, and resource selection (i.e., patterns of space use) of mule deer in response to provisioning of water developments in Mojave National Preserve, California, USA from 2008-2011. I hypothesized that deer would alter movements and distribution to improve access to permanent water, particularly during times of water scarcity. I also hypothesized that proximity to permanent water would be a significant component of resource selection by mule deer, and that mule deer would exhibit selection for sites near those sources. I used general linear models with random effects for individual mule deer to evaluate the influence of water availability and season on daily movements and area of mule deer distributions. I used nonlinear mixed-effects models to evaluate the importance of water in models of resource selection by mule deer and to determine the magnitude of that water effect.In chapter 2, I evaluated responses in body condition and traits of demography of mule deer to provisioning of water developments. I hypothesized that mule deer with improved access to permanent sources of water would exhibit greater body fat, greater fetal rates, and higher survival than those in areas where water was limited. I used general linear models to evaluate the influence of water availability and climatic conditions on body condition. I then modeled the effects of water availability, body condition, and climatic conditions on fetal rates using logistic regression. Finally, I used the known fates package in Program MARK to evaluate the effects of water provisioning, climatic conditions, and individual characteristics on survival patterns of mule deer and to estimate monthly and annual probabilities of survival