Climate change is becoming an ever-increasing issue for conservation efforts, especially in dryland ecosystems where natural resources are already scarce for native species. This is increasingly true for native amphibians in the area, which are already experiencing threats to their range by human intervention, disease, and invasive species. The objectives of this study are to 1) identify how climate change impacts the distribution of native and non-native amphibian species and high priority conservation areas (HPCA) in the drylands of the Southwest United States and northern Mexico; 2) Describe the relationship between environmental variables and spatial configurations of HPCA; 3) Explore how amphibians distributions and HPCA may respond under climate change scenarios; 4) Investigate the projected change in drivers of climate change; 5) Investigate how climate change will impact the critical areas for conservation of native amphibians. Distribution maps were obtained for the 220 resident native and non-native amphibian species, and complementarity-based analysis was used to identify HPCA for amphibians. We used 34 predictor variables grouped into three categories, and ranked based on their influence in determining HPCA. Finally, Zonation, species richness, and rarity-weighted richness (RWR) were evaluated to identify complementarity to HPCA. Results show that water-related variables and -related variables such as temperature and solar radiation were the best indicators of amphibian conservation HPCA. Zonation also proved to be the best method for identifying these HPCA. This study is the first to investigate the impact of climate change on site complementarity. The results from this study will open new inquiries for biogeography and conservation biology and also have a functional use for natural resource managers in the United States and Mexico to monitor changes to these areas and plan for recovery if needed.

Riparian ecosystems comprise less than 2% of the landscape in the arid western U.S. yet provide habitat and resources to over half of arid-land wildlife species, including a broad diversity of anurans (frogs and toads). I surveyed anurans using passive acoustic monitoring to capture spring advertisement calls in wilderness area tributaries of the Verde River, Arizona, USA. In the spring and summer of 2021 and 2022, 13-29 autonomous recording units (ARUs) were deployed along perennial, intermittent, and ephemeral reaches across eight headwater streams. I characterized stream reaches based on the percent of pool, riffle, run, and side channel habitat within 100 meters of each ARU. I quantified substrate, discharge at 95% exceedance probability, flow width, and canopy cover at each site. To relate anuran occupancy and relative habitat use to environmental and hydrological variables, I evaluated acoustic data using single-species occupancy and Royle-Nichols and N-mixture (relative habitat use) models. Four species were detected in this study: canyon treefrog (Hyla arenicolor), red-spotted toad (Anaxyrus punctatus), Woodhouse’s toad (Anaxyrus woodhousii), and non-native American bullfrog (Lithobates catesbeianus), with canyon treefrog being the most ubiquitous species observed. Occupancy of canyon treefrog was greater at perennial and intermittent sites compared to ephemeral sites, and presence of pool was the most important driver of canyon treefrog occupancy and relative habitat use. Notably, this study did not detect several species with historical records in the middle Verde River watershed, including Arizona toad (Anaxyrus microscaphus) and Northern leopard frog (Lithobates pipiens). Given climate change-related flow declines and intensifying demands for water in the Southwest, maintaining stream flows that provide consistent and suitable hydroregimes for anuran breeding and larval development is of increasing importance. Determining habitat use and flow regimes necessary to support anuran populations can aid in prioritization of conservation actions related to water management and predict how changes in water availability may impact stream-breeding anurans.

Anthropogenic linear infrastructures, including roads, railways, and canals, provide important resources to humans. However, linear infrastructures can reduce landscape connectivity for many wildlife populations. To mitigate these effects, crossing structures and crossing areas can facilitate animal movement across linear infrastructures. Compared to roads, little research has evaluated the factors influencing wildlife use of crossings along major canals. The Central Arizona Project (CAP) canal is a major linear feature in Arizona, and exhibits multiple types of crossing structures and areas. In Chapter 1, the objective was to evaluate the spatial (i.e., landscape features, crossing attributes) and temporal (i.e., season, time of day) factors influencing wildlife use of overpasses (n = 43) and siphons (n = 13) along the CAP canal. Using remote wildlife cameras, 17 species were detected using overpasses and siphons along the CAP canal during one year. Animals exhibited species-specific preferences for landscape features, such as topography and vegetation, and canal crossing types, although many species decreased use of overpasses associated with human development. In Chapter 2, the objective was to evaluate the influence of human activities at overpasses on use by mule deer across multiple analytical scales. Mule deer occupancy and relative habitat use at overpasses decreased in relation to human activity, including recreation. In Chapter 3, the objective was to evaluate seasonal use of underpasses (n = 12) by mammals, reptiles, amphibians, and invertebrates. Using specialized remote wildlife cameras, 30 species were detected using underpasses along the CAP canal across three seasons, and some animals exhibited variable crossing frequencies in relation to the summer monsoon season. Overall in this project, several species of small to large-sized mammals, reptiles, amphibians, and invertebrates used a variety of crossing types, including overpasses and underpasses, along a major canal. Ultimately, this study suggests that to promote landscape connectivity for the wildlife community associated with canals and other types of linear infrastructures, it is important to provide a variety of crossing types that occur across a range of landscape characteristics.

Environmental variation impacts physiological performance in animals. As a result, many animals thermoregulate to buffer unfavorable thermal variation in their environments. Animals are only expected to thermoregulate when the benefits outweigh the costs, although both are difficult to quantify. I examined how habitats and organismal factors shape thermoregulation and physiological performance in lizards. I found that habitat structure shapes opportunities for thermoregulation in two species of Anolis lizards. In dense tropical rainforests where there is low habitat heterogeneity, the range of available microclimates is narrow. Consequently, lizards in the tropics tend to be thermal specialists – performing best over a narrow range of temperatures. This phenotype should lead to decreased performance under climate warming. I then investigated the relationship between body condition, feeding, and thermoregulation in Yarrow’s spiny lizards (Sceloporus jarrovii) using lab- and field-based experiments. In the lab experiment, when lizards were observed in an artificial thermal gradient, neither body condition nor feeding status influenced the mean body temperature. When simulated costs of thermoregulation were higher, all lizards reduced thermoregulation similarly. However, when lizards were observed in an outdoor thermal arena, individuals with lower body condition decreased thermoregulatory performance, resulting in a lower mean body temperature. Animals with poor body condition may face greater risk of predation when thermoregulating. Finally, I conducted a comparative analysis to quantify relationships between the potential for thermoregulatory performance and empirical measures of productivity (i.e., growth rates and reproductive output) in lizard populations. A model that assumes lizards are active whenever preferred temperatures were available overestimated the duration that a lizard could maintain a preferred body temperature. As such, studies equating predicted thermoregulatory performance with fitness in the context of climate change should be interpreted cautiously. Overall, environmental factors and organismal traits shape the thermoregulatory behavior of animals, ultimately affecting their physiological performance and fitness. Biologists should consider these relationships when modeling the impacts of climate change on future performance.

Understanding how wildlife interact with humans and the built environment is critical as urbanization contributes to habitat change and fragmentation globally. In urban and suburban areas, wildlife and people are often in close quarters, leading to human-wildlife interactions (HWI). In the greater Phoenix Metropolitan Area, Arizona, HWI can involve reptiles such as venomous (family Viperidae, e.g., rattlesnakes) and nonvenomous (family Colubridae, e.g., gophersnakes) snakes. Rattlesnake Solutions, LLC, a local business, removes and relocates snakes from homes and businesses in the Phoenix area and, as a collaborator, has provided records of snake removals. Using these records, I investigated taxa-specific habitat trends at two spatial scales. At the neighborhood scale (n = 60), I found that removals occurred in yards with abundant cover opportunities. At the landscape scale (n = 764), nonvenomous snakes were removed from areas of higher urbanization compared to venomous snakes. Clients of Rattlesnake Solutions, LLC, were asked to answer a short survey, designed by K. Larson and colleagues, regarding the circumstances of their snake removal event and their attitudes, perceptions, and experiences with snakes. I used responses from this survey (n =271) to investigate if prior experience with snakes influences reported attitudes towards snakes. Respondents with prior snake experiences reported more positive attitudes towards snakes and were more consistent across their responses than those without prior snake experiences. Continuing inquiry into the urban ecology of these snakes is important to fostering coexistence between snakes and people that call Phoenix home.

The Arizona toad (Anaxyrus microscaphus) is unique among bufonids because they primarily breed in streams of Arizona, New Mexico, Utah, and Nevada. Arizona toad is a species of conservation concern throughout their range. The non-native northern crayfish (Orconectes virilis) are opportunistic omnivores implicated in the declines of other native aquatic species. I wanted to determine occupancy, habitat use, and species interactions of the Arizona toad throughout its range Visual encounter surveys (VES) were completed by ASU and natural resource agency partners in the summers of 2021 and 2022 (n = 232) throughout Arizona toad range in Arizona. I used VES data and crayfish occurrence records, to determine interactions between the two species. I used broadscale environmental variables (1 km resolution) from WorldClim and EarthEnv to evaluate a relationship with Arizona toad occupancy across transects. These broadscale variables included bioclimatic variables, measures of habitat heterogeneity, measures of solar radiation, and topographic variables. In 2022 I collected fine-scale habitat data evaluating available vegetation cover and substrate composition within paired habitat plots. Fine-scale variables included canopy cover, substrate type, vegetation cover, and water depth. I applied multiple occupancy modeling approaches. Single-species model results found low toad occupancy, but high detection, as this is a rare species. Multi-species results showed no positive or negative relationship between Arizona toad and northern crayfish for both seasons. Two principal component analyses (PCA) were run on broadscale environmental variables and fine-scale habitat variables for 2021 and 2022, respectively, creating new synthetic variables for use in analysis. In 2021, the broadscale components were added to the single-species occupancy models and the top model included bioclimatic variables related to annual temperature range and precipitation. Arizona toad occupancy is lower with extreme hot temperatures and less precipitation. A logistic regression was run with the fine-scale habitat variables and the top model included PC1 and PC3. PC1 described elements related to riparian complexity, while PC3 described elements related to algae presence, including attached to cobble substrate. Arizona Toad select for certain habitats including canopy cover, shallow water, algae cover, and pebble cover. It is important to maintain riparian area habitat complexity and conserve habitat for the Arizona toad, a riparian stream specialist.

The built environment increases radiant heat exchange in urban areas by several degrees hotter compared to non-urban areas. Research has investigated how urbanization and heat affect human health; but there is scant literature on the effects of urban heat on wildlife. Animal body condition can be used to assess overall health. This parameter estimates the storage of energy-rich fat, which is important for growth, survival, and reproduction. The purpose of my research was to examine the Urban Heat Island effect on wild rodents across urban field sites spanning three strata of land surface temperature. Site level surface temperatures were measured using temperature data loggers and I captured 116 adult pocket mice (Chaetodipus spp. and Perognathus spp.) and Merriam’s kangaroo rats (Dipodomys merriami) to measure their body condition using accurate and noninvasive quantitative magnetic resonance. I used baited Sherman live traps from mid-May to early September during 2019 and 2020 in mountainous urban parks and open spaces over two summers. Rodents were captured at seven sites near the Phoenix metropolitan area; an ideal area for examining the effect of extreme heat experienced by urban wildlife. Results supported the prediction that rodent body condition was greatest in the cooler temperature stratas compared to the hottest temperature strata. I related rodent body condition to environmental predictors to dispute to environmental predictors to dispute alternative hypotheses; such as vegetation cover and degree of urbanization. Results based on measures of body fat and environmental predictors show pocket mice have more fat where vegetation is higher, nighttime temperatures are lower, surface temperatures are lower, and urbanization is greater. Kangaroo rats have more fat where surface temperature is lower. My results contribute to understanding the negative effects of extreme heat on body condition and generalized health experienced by urban wildlife because of the built environment. This research shows a need to investigate further impacts of urban heat on wildlife. Management suggestions for urban parks and open spaces include increasing vegetation cover, reducing impervious surface, and building with materials that reduce radiant heat.

Land use change driven by human population expansion continues to influence

the integrity and configuration of riparian corridors worldwide. Wildlife viability in semi-arid regions depend heavily on the connectivity of riparian corridors, since water is the primary limiting resource. The Madrean Archipelago in northern Mexico and southwestern United States (US) is a biodiversity hotspot that supports imperiled wildlife like jaguar (Panthera onca) and ocelot (Leopardus pardalis). Recent and ongoing infrastructure developments in the historically understudied US-México borderlands region, such as the border wall and expansion of Federal Highway 2, are altering wildlife movement and disconnecting essential habitat.

I used wildlife cameras to assess species occupancy, abundance, and related habitat variables affecting the use of washes as corridors for mammals in semi-arid Los Ojos (LO), a private ranch within a 530 km2 priority conservation area in Sonora, México located south of the border and Federal Highway 2. From October 2018 to April 2019, I deployed 21 wildlife cameras in five different riparian corridors within LO. I used single- season occupancy models and Royal Nichols abundance models to explore the relationship between habitat variables and use of riparian corridors by mammal communities of conservation concern within this region.

Twenty-one mammal species were recorded in the study area, including American black bear (Ursus americanus), white-tailed deer (Odocoileus virginianus) and the first sighting of jaguar (Panthera onca) in this region in 25 years. For the 11 medium- and large-bodied mammals recorded, habitat variables related to perennial river characteristics (distance to river, weekly water, and site width) and remoteness (distance from highway, elevation, and NDVI) were important for occupancy, but the direction of the relationship varied by species. For commonly observed species such as mountain lion (Puma concolor) and white-nosed coati (Nasua narica), topographic variety was highly informative for species abundance. These results highlight the importance of habitat diversity when identifying corridors for future protection to conserve wildlife communities in semi-arid regions. Additionally, this study provides robust evidence in support of mitigation measures (e.g. funnel fencing, over- or under- passes) along Federal Highway 2, and other barriers such as the border wall, to facilitate wildlife connectivity.

Human land use and land cover change alter key features of the landscape that may favor habitat selection by some species. Lizards are especially sensitive to these alterations because they rely on their external environment for regulating their body temperature. However, because of their diverse life-history traits and strategies, some are able to respond well to disturbance by using their habitat in various ways. To understand how they use their habitat and how human modifications may impact their ability to do this, biologists must identify where they occur and the habitat characteristics on which they depend. Therefore, I used species occupancy modeling to determine (1) whether disturbance predicts the presence of two sympatric congeneric (species of the same genus) lizard species Sceloporus grammicus and S. torquatus, and (2) which habitat characteristics are essential for predicting their occupancy and detection. I focused my study in central Mexico, a region of prevalent land use and land cover change. Here, I conducted visual encounter and habitat surveys at 100 1-hectare sites during the spring of 2019. I measured vegetation and ground cover, average tree diameter, and abundance of refuges. I recorded air temperature, relative humidity, and elevation. I summarized sites as either undisturbed or disturbed, based on the presence of human development. I also summarized sites by ecosystem type, desert or forest, based on vegetation composition (i.e., desert-adapted vs. non-desert-adapted plants), evidence of remnant forest, air temperature, and relative humidity. I found that S. torquatus was more likely to be present in disturbed habitat, whereas S. grammicus was more likely to be present in areas with leaf litter, tree cover, and woody debris. S. torquatus was twice as likely to be detected in forests than deserts, and S. grammicus was more likely to be detected at sites with high elevation and high relative humidity, low temperature, and herbaceous and grass cover. These results emphasize the utility of species occupancy modeling for estimating detection and occupancy in dynamic landscapes.

Riparian areas are an important resource, especially in the arid southwest, for many wildlife species. Understanding species occurrence in areas dominated by non-native vegetation is important to determine if management should be implemented. Saltcedar (Tamarix spp.) is one of the most prevalent non-native trees in riparian areas in the southwest United States and can alter vegetation structure, but little is known about how medium and large carnivores use stands of saltcedar. Three riparian forest types make up the San Pedro riparian corridor: non-native saltcedar, native mesquite (Prosopis spp.) bosque, and a mixture of native cottonwood (Populus fremontii) and willow (Salix goodingii) woodlands. My goals were to determine relative use, diversity, and occupancy of medium and large mammals across forest types to evaluate use of the non-native stands. I sampled mammals along approximately 25.7 river kilometers between July 2017 and October 2018, using 18 trail cameras (six per forest type) spaced 1km apart. I summarized environmental variables around the camera sites to relate them to species occupancy and reduced them to 4 components using a Principal Component Analysis (PCA). I observed 14 carnivore species, including bobcat (Lynx rufus), coyote (Canis latrans), and coati (Nasua narica) over 7,692 trap nights. Occupancy of some species may have been influenced by the different components, but models showed high standard errors, making it difficult to draw firm conclusions. Most mammal species used all three forest types at some level and no surveyed forest type was completely avoided or unused. Coyote tended to have greater use in the mesquite forest while canids trended toward greater use in saltcedar forest. Based on two-species occupancy models as well as activity overlap patterns, subordinate species did not appear to avoid dominant species. No species seems significantly affected by non-native saltcedar at this time.