|
nif:isString
|
-
Fieldwork was approved and sanctioned by the United States Department of Defense and the Utah Division of Wildlife Resources. Permission to access land in Great Basin was obtained from the United States Department of Defense (United States Army Dugway Proving Ground). Permission to access public land in Mojave was not required. The kit fox is a protected species and we did not handle any individuals. Thus, our study was strictly observational in nature (using infrared camera traps) and no specific permits or regulations for animal care were needed according to the Institutional Animal Care and Use Committee of Brigham Young University.
This study was conducted at sites in both the Great Basin and Mojave Deserts. The Great Basin Desert study area consisted of 915 km2 of land managed by the United States Department of Defense, United States Army Dugway Proving Ground in west-central Utah. The terrain was typical of Lake Bonneville lakebed characterized by dune systems and alkaline flats that were dominated by black greasewood (Sarcobatus vermiculatus). Where wildfires had occurred along the foothills, cheatgrass (Bromus tectorum) was common within communities of big sagebrush (Artemisia tridentata), rabbitbrush (Chrysothamnus spp. ), and juniper (Juniperus osteosperma) [15]. Elevations across the study area ranged from approximately 1300 to 1800 m. Annual weather consisted of mean air temperatures of 12.69°C (range: −20.02 to 40.58°C) and mean precipitation of 150 mm (MesoWest, Bureau of Land Management & Boise Interagency Fire Center). The US Army Dugway Proving Ground has not been grazed by domestic livestock for the last 60 years [16]. In this study area, we identified 22 permanent water sources consisting of 11 water developments for wildlife, six natural springs, and five man-made ponds. The median distance of 10,000 random points to a natural water source (e.g., springs) and any water source was 7.10 km (range: 0.04 to 19.84 km) and 3.23 km (range: 0.06 to 8.54 km), respectively. The Mojave Desert study area consisted of 1,064 km2 of public land managed by the United States Department of the Interior, Bureau of Land Management. The Mojave study area was located in extreme southwestern Utah, northwestern Arizona, and southeastern Nevada and was approximately 360 km south of Great Basin. This study area was characterized by an alternating landscape of rolling hills/ridges and dry desert washes radiating from the Beaver Dam Mountains and emptying into the Beaver Dam Wash to the southwest near the intersection of the Utah-Nevada-Arizona state borders [21]. In areas that burned within the last decade, red brome (B. rubens) was well established among surviving creosote (Larrea tridentata), Joshua-tree (Yucca brevifolia), and black-brush (Coleogyne ramosissima) communities [22]. Along the foothills, the vegetation primarily consisted of sagebrush and juniper, transitioning to pinyon pine (Pinus edulis) at higher elevations. Elevations across the Mojave study area ranged from approximately 800 to 2000 m. Annual weather consisted of mean air temperatures of 19.18°C (range: −10.04 to 41.70°C) and mean precipitation of 113 mm (MesoWest, Bureau of Land Management & Boise Interagency Fire Center). The Mojave study area was grazed by livestock from October to May. We identified 66 permanent water sources in this study area consisting of 35 water developments for wildlife, 18 water troughs/tanks for livestock, 11 natural springs, and two man-made ponds. The median distance of 10,000 random points to a natural water source and any water source was 4.75 km (range: 0.03 to 14.01 km) and 2.10 km (range: 0.02 to 8.72 km), respectively.
To verify if presence of coyotes was greater in areas with free water (hereafter wet) compared to areas without (hereafter dry), we first established wet and dry areas in both study areas. Using ArcGIS (version 10.0, Environmental Systems Research Institute, Redlands, California), we created a uniform pattern of sample points with a distance of 4 km apart for both study areas. Each of these sample points was buffered with a 2.6 km radius based on the square root of a home range for coyotes inhabiting a semi-arid environment similar to our study areas [17]. The square root of the home range is a linear measure used to approximate daily movements of mammals and birds [23], [24]. If free water was located within a buffer zone for a given sample area, we considered it a wet area. We identified water sources using databases with geospatial information for springs and water developments provided by the US Army Dugway Proving Ground and the Utah Division of Wildlife Resources. In addition, we consulted with local ranchers concerning water sources for livestock that were not in our databases. We were confident in our efforts to identify all known water sources in both study areas. We established 32 scent stations in 2011 and 39 in 2012 and monitored stations for two-week periods during July to August (hottest part of the year; Table 1). Approximately 60% of the scent stations were located in wet areas and 40% in dry areas (Table 1). Scent stations in dry areas were 2.92 km farther from a known water source compared to scent stations in wet areas (Table 1). At each scent station, we placed a scent lure (2011: fatty acid scented disc [Pocatello Supply Depot, Pocatello, Idaho, USA]; 2012: liquid scent [Murray’s Lures, Walker, West Virginia, USA]) on the ground and an infrared-triggered camera (PC 900, Reconyx©, Holmen, Wisconsin) approximately two meters from the scent either directly north or south to avoid false camera triggers by the sun.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0067800.t001 Distances from scent stations to nearest known source of free water in wet and dry areas, 2011 to 2012. To evaluate relative rates of visitation to water sources by canids, we monitored all known water sources at the Great Basin and Mojave study areas. From May to October, 2010 to 2012, we used infrared-triggered cameras to photograph canids visiting water sources. We randomly sampled water sources with cameras at both study areas for approximately two-week periods for a total of 78 weeks. To determine which water sources to sample for a given period, we generated random points within each study area using ArcMap. We then identified the nearest water source to a random point and camera-sampled as many water sources as possible (in 2010, we used six cameras in each study area for sampling compared to 15 cameras in 2011–2012). We attached cameras to metal posts and placed them approximately two meters from the edge of water where animals gained access to drink. At water sources with multiple locations of drinking access (e.g., paired tanks of water, ponds), we placed cameras at a minimum of two locations where animals could drink. We considered proximity to trails and recent sign to determine the location of cameras at ponds and large springs [25]. Our estimates of visitation by canids at large water sources were likely conservative due to the inability to monitor all potential locations where canids could access water. We assumed, however, that any potential bias was similar at large water sources from both study areas.
To test prediction one associated with the indirect effect of water hypothesis, we used z-tests for proportions [26] to compare the proportion of scent stations in wet and dry areas visited by canids at both study areas. We tested the spatial segregation component of the second prediction by comparing visits to scent stations by each canid using Kendall’s Tau-b correlation analyses. Kendall’s Tau-b correlation accounted for ties that occurred because of zero visit data due to only one species of canid primarily visiting a given station [27]. We excluded all stations that were not visited by at least one species of canid from these correlation analyses. To test the temporal segregation component of prediction two, we used time stamps available on photographs of canids at scent stations and compared these times using Mann-Whitney U-tests. To account for the transition from 23∶59 to 00∶00 h (midnight) in our analyses and thus avoid the potential difference between photographs taken immediately prior to and after midnight we marked the change between days by setting 21∶00 h (approximately sunset) as 00∶00 h. This adjustment allowed us to compare times of visitation within a continuous scale of time marked by nocturnal hours which are likely more biologically relevant to both coyotes and kit foxes [28], [29]. We used Mann-Whitney U-tests to compare the mean daily visitation rates (# of visits/# of operable camera trap days) of canids to water sources between study areas. We defined a visit as all photo captures of a species occurring within 30 min. Thus, photo captures occurring more than 30 min apart were considered independent [30]. We performed all analyses using Program R [31]. We conducted series of statistical analyses to test our hypotheses and we used a Bonferroni correction to avoid type I errors. We set the family-wise level of significance for all statistical tests at α = 0.05 (Bonferroni adjusted α = 0.01).
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}