https://cpw.cvlcollections.org/files/original/b42989b0e9a9d9d6cb765ccb81bc2fdb.pdf ce4bd240e2989e5a99c861152d62e96b PDF Text Text The research in this publication was partially or fully funded by Colorado Parks and Wildlife. Dan Prenzlow, Director, Colorado Parks and Wildlife • Parks and Wildlife Commission: Marvin McDaniel, Chair • Carrie Besnette Hauser, Vice-Chair Marie Haskett, Secretary • Taishya Adams • Betsy Blecha • Charles Garcia • Dallas May • Duke Phillips, IV • Luke B. Schafer • James Jay Tutchton • Eden Vardy �Wildlife Society Bulletin 43(2):222–230; 2019; DOI: 10.1002/wsb.971 Original Article Less Invasive Monitoring of Cougars in Colorado’s Front Range MAT W. ALLDREDGE,1 Colorado Parks and Wildlife, 317 W Prospect Road, Fort Collins, CO 80526, USA TASHA BLECHA, Colorado Parks and Wildlife, 317 W Prospect Road, Fort Collins, CO 80526, USA JONATHAN H. LEWIS, Colorado Parks and Wildlife, 317 W Prospect Road, Fort Collins, CO 80526, USA ABSTRACT From 2014 to 2016, in the Front Range of Colorado, USA, we assessed noninvasive approaches to sampling cougar (Puma concolor) populations in an attempt to provide a new method that would be less field intensive, less expensive, and could be applied over large spatial extents compared with current methods. We assessed the use of predator calls to lure cougars to a site with remote camera traps for detection and also evaluated hair snags at sites to noninvasively identify individual animals. Predator calls effectively attracted cougars to specific sites with an average of 82 unique photographic detections of cougars per survey year (0.03 detections/trap‐night). However, obtaining hair samples from these animals was less effective because animals did not always pass through hair snags and ability to uniquely identify individuals by genotype was poor. We evaluated different approaches to estimating cougar density and found mark–resight to be a viable option in our study system. Mark–resight density estimate after correcting for partial use of the sampling area by cougars was 4.1 cougars/100 km2 (95% CI = 2.4, 5.8). Our results indicate that combining methods of noninvasive genetic sampling and auditory calls to monitor cougar populations can provide reliable density estimates over large geographic areas and areas with significant amounts of inaccessible private lands. © 2019 The Wildlife Society. KEY WORDS abundance, auditory, cougar, mark–resight, noninvasive, predator calls Puma concolor. Reliable estimates of population size are beneficial for setting harvest quotas, evaluating management practices, and understanding dynamics of predator–prey systems. Unfortunately, with many carnivore species, it can be difficult and expensive to obtain these estimates. This is especially true with cougars (Puma concolor) because of their low densities, secretive nature, and unpredictable response to lures. Most reliable estimates of population size for cougars have come from intensive capture‐and‐monitoring studies, which are expensive and time‐consuming (Logan 1983, Lindzey et al. 1994, Murphy 1998, Logan and Sweanor 2001). Across western North America, obtaining population estimate for cougars is a priority for 86% of natural resource agencies (Beausoleil et al. 2008). Additionally, Beausoleil et al. (2008) reported that agencies commonly extrapolated cougar population size from hunter harvest data (mandatory check), habitat assessments, and published estimates. These approaches provide some information to manage cougar populations, but they are limited. Extrapolation from hunter harvest is primarily reactionary because this approach relies on composition of the harvest to suggest population status (Anderson Received: 23 April 2018; Accepted: 11 March 2019 1 E‐mail: mat.alldredge@state.co.us 222 and Lindzey 2005). Extrapolation from published estimates requires strong assumptions about similarities among areas and is not robust to subtle differences in environmental, demographic, or management differences among areas. Currently there are no reliable estimates of cougar density for Colorado, USA, and this is especially true within the urban–wildland interface of Colorado’s Front Range. This is problematic because Colorado Parks and Wildlife (CPW) managers require more accurate estimates of cougar populations to develop better management strategies to achieve healthy, self‐sustaining populations and help mitigate human–wildlife conflicts. The CPW manages cougar populations based on large (several 1,000 km2) Data Analysis Units (DAUs) with specific cougar population objectives. These objectives are designed to maintain healthy cougar populations across the state in conjunction with maintaining healthy ungulate populations and minimizing human– cougar conflict. To better manage cougars within the state and address public concerns about cougar numbers, it is important to develop techniques and obtain reliable estimates of cougar density that are cost‐effective and cover large areas that can be applied at the scale of DAUs. Exurban landscapes provide unique challenges to sampling because they are dominated by private lands with limited access, which makes current sampling methodologies logistically infeasible and ineffective. Wildlife Society Bulletin • 43(2) �Typical assessments of cougar density rely on extensive telemetry studies to effectively mark all the animals in the population or capture–recapture approaches (Anderson and Lindzey 2005, Cougar Management Guidelines Working Group [CMGWG] 2005). Many cougar researchers believe that complete enumeration of the cougar population through long‐term studies attempting to mark all animals in the population is the most reliable method to obtain population estimates (Cougar Management Guidelines Working Group [CMGWG] 2005). An alternative approach that has been used to estimate cougar population size is the 2‐sample Lincoln–Petersen estimator in a mark–recapture framework (Anderson and Lindzey 2005). This method also requires a marked population and is subject to all the Lincoln–Petersen model assumptions, which include constant probability of capture among all individuals and time periods, and population closure (no changes in the population from emigration–immigration or births–deaths; Williams et al. 2002). Neither of these approaches are realistic at a management scale across large areas typical of Colorado’s DAUs or in areas dominated by private property. There is difficulty and expense associated with physical mark–recapture techniques for estimating cougar abundance and they can be applied only at small scales; therefore, alternate techniques have been developed, such as track surveys, the use of scat detection dogs, and bio‐darting. However, track surveys were reported to be an ineffective method for estimating abundance of cougars in some areas, but effective in other areas (Sargeant et al. 1998, Choate et al. 2006, Gompper et al. 2006, Sawaya et al. 2011). Track surveys have been used to assess cougar population trends (Smallwood and Fitzhugh 1991, 1995; Smallwood 1994; Cunningham et al. 1995), but actual relationships to population size are generally weak (Van Dyke et al. 1986, Van Sickle and Lindzey 1992, Cunningham et al. 1995). The use of scat‐detection dogs was demonstrated to be effective for estimating cougar density, but this approach can be expensive and limited in spatial extent (Davidson et al. 2014). Scat collections can either be done by searching transects with human observers or with trained dogs (Harrison et al. 2004, Smith et al. 2005, Davidson et al. 2014). Although use of scats for noninvasive genetic sampling (NGS) may sound appealing, the actual encounter rate of scats may be prohibitively low and the ability to cover large areas logistically infeasible to make this a viable management option. Therefore, this technique would have limited applicability at the management scale used in Colorado and would be logistically infeasible within the Front Range because of private property access. Two studies effectively estimated cougar density in Montana, USA, using bio‐darting to obtain tissue samples from cougars treed by hounds, and backtracking cougars during winter to obtain hair samples (Russell et al. 2012, Proffitt et al. 2015). Additionally, Beausoleil et al. (2016) developed an effective sampling approach using a citizen scientist approach and bio‐darting to estimate cougar density at a game management unit scale in northeastern Washington, USA. This approach does account for detection probability, but is still field intensive. Bio‐darting also is based on finding tracks and being able to tree an individual to get a sample, which may not effectively sample all portions of the population, especially where property access can be limited. Furthermore, DNA‐based estimation methods, including bio‐darting and snow‐tracking, may be biased high because of inflated minimum counts caused from genotyping errors (Sawaya et al. 2017). To address limitations of other techniques relative to spatial extent and inaccessible areas due to private lands, an alternative approach would be to collect hair or tissue from cougars that are lured into a site. Lures and hair snags in conjunction with remote camera traps have been used to successfully survey carnivores (Sargeant et al. 1998, Long et al. 2003, Choate et al. 2006, Crooks et al. 2008, Sawaya et al. 2011), lynx (Lynx canadensis; McDaniel et al. 2000, Schmidt and Kowalczyk 2006), bobcats (L. rufus; Harrison 2006), ocelots (Leopardus pardalis; Weaver et al. 2005), other felids (Harrison 1997, Downey et al. 2007), and cougars (Long et al. 2003). Although the use of lures with hair snags has proven effective on many species, such as bears (Ursus spp.) and some felids, baits and scents have been found relatively ineffective at luring cougars to a specific site, even when cougars are known to be in close proximity (Long et al. 2003, Choate et al. 2006, Yeager 2016). Few surveys have incorporated auditory calls despite the fact that felids may exhibit a greater response to auditory cues than to olfactory stimuli (Chamberlain et al. 1999). To attract cougars to sampling locations, Yeager (2016) reported that auditory calls were more effective than bait or scent lures. For example, they were able to detect the majority of radiocollared cougars (19 of 20) on the Colorado Front Range using auditory calls placed in cubbies. Furthermore, all 7 radiomonitored cougars on the Uncompahgre Plateau in Colorado were similarly detected (Yeager 2016). However, application of this approach to assess density estimation has not been implemented. Local environments can also dictate which sampling approaches are used because access to private lands, vegetation structure, and climate can impede study designs. For example, although Sawaya et al. (2011) reported snow‐ tracking as an effective method to estimate cougar density in their study area, Alldredge (2012) reported that varying snow conditions and limited access to private lands made snow‐tracking logistically infeasible in the Front Range of Colorado. Similarly, Beausoleil et al. (2016) presented an effective strategy to use citizen scientists and bio‐darting, but given the extent of private property on the Front Range of Colorado, and presumably other areas with a wildland– urban interface, this would not be practical. Study area size may also dictate sampling approaches because many (mark– recapture, snow‐tracking, scat detection) would be impractical at large scales. For management purposes, areas of interest tend to be quite large; for example, the typical DAU in Colorado generally exceeds 5,000 km2. We implemented an NGS survey to test the efficacy of this approach in conjunction with an ongoing cougar Alldredge et al. • Less Invasive Monitoring of Cougars in Colorado’s Front Range 223 �research project (2006–2016) where cougars were already collared with a Global Positioning System (GPS). The NGS sampling sites were established using a 25‐km2 grid across the study area using predator calls, cameras, and hair collection devices. We assessed the reliability of this approach to detect cougars relative to the marked sample based on photographic evidence at sites. We also assessed the reliability of obtaining hair samples and genotypes from those hair samples relative to photographic evidence of the presence of cougars at sites. We also evaluated the feasibility of using capture–recapture and mark–resight models to estimate cougar densities STUDY AREA We conducted the study along Colorado’s Front Range, in Boulder, Jefferson, Gilpin, and Larimer counties from 2014 to 2016. The study area was defined by Hwy 36, Hwy 72, Hwy 93, and I‐70, with the actual grid area of 800 km2 (Fig. 1). The area was a foothill‐montane system with elevations ranging between 1,595 and 3,173 m. Dominant vegetation throughout the area was ponderosa pine (Pinus ponderosa), Douglas fir (Pseudotsuga menziesii), Rocky Mountain juniper (Juniperus scopulorum), mountain mahogany (Cercocarpus montanus), Gambel’s oak (Quercus gambelii), serviceberry (Amelanchier alnifolia), and bitterbrush (Purshia tridentata). The study area was quality cougar habitat with mule deer (Odocoileus hemionus) and elk (Cervus canadensis) as primary prey, as well as smaller alternate prey species (Blecha et al. 2017). Human development and use of the area was intensive, with land use consisting of 3% urban, 27.3% exurban, and 69.7% rural areas. The patch‐work of private and public lands throughout the study area limited recreational hunting of cougars. Cougar mortality was primarily human‐caused, including vehicle collisions and management‐related removals (Alldredge 2016). METHODS During 2014–2016, we used a systematic layout to ensure sampling stations were distributed across the study area. We used a grid‐cell size equal to one‐quarter of a female cougar home range size, which is typically recommended to ensure a nonzero probability of detection for all animals within the survey area (Otis et al. 1978, White et al. 1982, Williams et al. 2002). The average home‐range size for female cougars on the Front Range was approximately 100 km2 (Alldredge 2016), so we used a 5 × 5‐km grid cell size as our primary grid. A secondary 1 × 1‐km grid was overlaid on the primary grid. We randomly selected a cell from the secondary grid within each primary grid cell, omitting all cells on the edges. We also selected the 4 adjacent diagonal cells, for 165 total sites (Fig. 2). Within each selected cell, we selected specific sites to build a cubby based on likely areas to attract a cougar, property access, and logistics. We had 2 active sites and 3 passive sites during each sampling period within each primary grid. To attract 224 cougars, active sites employed a call, camera, scent, visual lure, and 1–2 hair‐snaring devices, whereas passive sites lacked calls and cameras because of logistical and budgetary constraints. We randomly selected active sites without replacement from the 5 available sites in each grid at time 0, 4, and 8 weeks and rotated them accordingly, such that all 5 sites were active sites once during the season and one was an active site twice. We checked all sites at approximately weekly intervals for signs of visitation and hair, and checked batteries in cameras and calls. Sixty‐six active sites and 99 passive sites were active throughout the study. All active sites were similar in design, containing the same elements. The main attractant was a predator call (Wasatch Wildlife Products® Custom FurFindR®, Magna, UT, USA) programmed to play a 5‐second distressed fawn sound on 30‐second intervals. These calls were also equipped with light sensors rendering them inactive during daylight hours. We attached calls with a cable approximately 1 m up from the base of a tree. We then built a cubby (brush and logs built up around the tree to direct the approach to the call) around the call leaving an obvious entry or exit. We constructed a perimeter hair snag using 18‐gauge, high‐tensile barbed wire with 4 prongs/barb (7‐cm spacing) around cubbies to snag hair as cougars approach the site. We stretched wire around available trees, forming a polygon around the cubby of approximately 10‐m radius. We also installed strands of barbed wire at cubby entrances to snag hair. During 2014, we placed barbed wire horizontally and terrain features affected the height of wires, consequently affecting whether a site was designed for a cougar to step over, under, or through 2 strands of barbed wire. During 2015–2016, we designed all cubbies with vertical wires in a v‐shape (narrower at the bottom), so that cougars had to walk between to snag hair. We also placed 2 sticky rollers on the barbed wire as a secondary hair snag at each site. A sticky roller was a 15‐cm‐long, 2‐cm‐diameter piece of polyvinyl chloride pipe coated with a sticky substance (Tree Tranglefoot®, Grand Rapids, MI, USA), with the wire run through it to snag hair. At each site, we positioned an infrared motion‐sensor camera (Reconyx®, Holmen, WI, USA; PC85 Rapidfire® or PC800 Hyperfire®) set to rapidly take 5 photos when triggered. To minimize the possibility of sample contamination (multiple animals leaving hair) and degradation, we checked the sites for activity every week. We considered hair on a single barb as one sample and denoted quantity with a score of 1–3 (<5 hairs, 6–15 hairs, and >15 hairs), following Yeager (2016). We removed hair using sterile tweezers and resterilized the barb by passing a flame under it (Kendall et al. 2008, Settlage et al. 2008). We placed hair in a small paper envelope, and placed envelopes in a plastic bag with a desiccant and stored them at room temperature (Taberlet and Luikart 1999). If hair was on the sticky rollers, we collected the entire roller, wrapped it in wax paper, and placed it in a plastic bag. Hair samples were processed at the Molecular Ecology Lab at the U.S. Geological Survey Fort Collins Science Wildlife Society Bulletin • 43(2) �Figure 1. Study area for population sampling of cougars during 2014 to 2016, located on the Front Range of Colorado, USA. Center. When possible, we extracted DNA from 10 hairs, following suggestions to achieve correct genotypes at 99% confidence (Taberlet et al. 1996, Goossens et al. 1998, Boersen et al. 2003) using Qiagen DNeasy® Tissue Kits (Qiagen Inc., Valencia, CA, USA). We genotyped samples using 10 microsatellite loci (FCA096, FCA035, FCA043, FCA149, FCA026, FCA057, FCA090, FCA132, FCA254, and B207‐2) that have been shown to have high variability Alldredge et al. • Less Invasive Monitoring of Cougars in Colorado’s Front Range 225 �Figure 2. Study area boundary and grid layout for noninvasive genetic sampling for cougars on the Front Range of Colorado, USA, during 2014 to 2016. Larger squares represent the 25‐km2 grid. White 1‐km2 cells represent the randomly selected secondary cells where actual lure sites were subjectively placed. in cougars (Menotti‐Raymond and O’Brien 1995, Ernest et al. 2000, Sinclair et al. 2001, Anderson et al. 2004). We amplified the DNA by polymerase chain reaction (PCR) using individually dye‐labeled primers specific to each loci primer following recommendations of Yeager (2016). We analyzed each locus via GeneMapper® (Thermo Fisher Scientific, Waltham, MA, USA). To assess error, we compared results from hair genotyping with archived blood and tissue samples collected during capture. When possible, we reprocessed hair samples shown to contain error at ≥1 allele. 226 We recorded camera detections (photographs of cougars) once (i.e., as a single capture event) for each night, regardless of the number of photographs. Although cougars do not have clear natural markings, we assumed that multiple detections of an unmarked cougar at a site during any night were of the same animal because of the strong solitary nature of cougars. Following the robust design framework (Kendall 1999), we created detection histories using detection of independent individuals over 6 primary 2‐week sampling periods across each year. Detection histories for each primary period were divided into 2 secondary 1‐week Wildlife Society Bulletin • 43(2) �periods. Dependent kittens and subadults photographed with their mothers were not counted in the detection history. This approach allows for multiple detections of the same individual within any sampling period (e.g., sampling with replacement; McClintock et al. 2009). We used GPS data from monitored cougars to confirm identity of radiocollared cougars. We used mark–resight analyses to estimate cougar density because we annually monitored 4–19 cougars fitted with GPS radiocollars in our study area during 2014–2016 (McClintock and White 2009, McClintock et al. 2009). The small sample size of uniquely identified individuals based on genetic samples precluded us from using capture– recapture models to estimate population size or density (Kendall 1999, Williams et al. 2002). We used the Poisson‐ log‐normal mark–resight model implemented in Program MARK, to obtain a spatially uncorrected estimate of population size and density as cougars per 100 km2 (White and Burnham 1999, McClintock et al. 2009). We conducted preliminary analyses using all 3 years of data, but estimates for 2015 and 2016 did not produce biologically reasonable estimates because of a limited number of marked individuals during those years. Therefore, we only present analyses and results for 2014. We fixed transition probabilities to 0 in these models because we accounted for time on and off grid separately based on GPS data of collared cougars. The apparent survival parameter was fixed to 1 because the time period was short and no mortality was observed for collared cougars or expected during this time period. The remaining model parameters were the number of unmarked individuals in the population during interval j (Uj) and intercept for mean resighting rate during interval j (αj), which were either modeled as constants across all time periods or allowed to vary across time periods. To explore heterogeneity in detection probabilities, we examined models with the individual heterogeneity (mixture model) and without heterogeneity. Model selection was based on Akaike’s Information Criterion that has been corrected for small sample sizes (AICc) to determine which model was best supported by these data (Burnham and Anderson 2002). We did not test models to address differences between males and females because sex could not be determined with certainty for unmarked individuals in all photographs. We then corrected detection probability for spatial use of the sampled area as the time spent on grid using GPS location data from the collared cougars over the duration of the sampling period (White and Shenk 2001). We programmed GPS collars to obtain 7 locations/day, which resulted in up to 588 locations/cougar (if all fixes are successful) during the sampling period. We then calculated time on grid as the number of locations on grid divided by the total number of locations for each collared cougar. The mean and variance were then calculated across collared cougars to give an estimate of the probability ( p̄ ) of a cougar being on the grid during the sampling period (White and Shenk 2001). From this, the corrected or actual density of cougars (D̂a ) per 100 km2 could be estimated as Dˆa = Nˆ p ̅ A 100, where A is the sampled area. The associated variance is then ˆ (Dˆa ) = Var 2 ˆ ˆ (Nˆ ) Nˆ Var (p ̅ ) + p ̅ 2 Var (A / 100) 2 , (White and Shenk 2001). RESULTS During January–March of 2014–2016, surveys photographed 42–118 unique cougars at cubbies and all detections occurred at active sites with calls (Table 1). Approximately 61% of all cubbies visited by cougars resulted in successful collection of hair samples. Of those hair samples, 33.3% were successfully genotyped and represented 20.3% of cougars photographed at cubbies. Comparing genotypes from samples to known individuals, proper identification of genotypes occurred for 87% (13% error rate) with an estimated 17.7% (0.61 × 0.333 × 0.87) chance of successfully genotyping an individual that visited a cubby. During 2014, 19 GPS‐collared cougars available for detection used the sampling grid at least once during the survey period. Of these, 15 cougars were detected at least once on camera. The number of uncollared cougars detected was 5, 20, 6, 3, 4, and 10 across the 6 primary sampling periods, respectively. When collared cougars were within the 25‐km2 grid during a sampling period, the detection probability was 0.042 (SE = 0.0001). However, when a collared cougar was within the 1‐km2 grid of a primary site during a sampling period, detection probability was 0.37 (SE = 0.001). The best model for mark–resight analysis was constant detection probability across all time periods, constant number of unmarked individuals, and no individual heterogeneity. Support for the other models was minimal (ΔAICc > 8.20; Table 2). Based on the top‐ranked model, the estimated detection probability was 0.23 (SE = 0.034) for all time periods with an estimated population size of 46 (SE = 7.94) cougars. This uncorrected estimate would result in a density of 5.8 cougars/100 km2 (95% CI = 4.2–8.1) Based on GPS locations of collared cougars during the sampling time frame, the average time on grid was 0.72 (SE = 0.176), ranging between 0.20 and 1.0. Based on this, the corrected density estimate was 4.1 cougars/100 km2 (95% CI = 2.4–5.8). Table 1. Number of photographs, hair samples, and successful genotypes by year from the Colorado, USA, Front Range cougar population survey during 2013–2016. 2014 2015 2016 Alldredge et al. • Less Invasive Monitoring of Cougars in Colorado’s Front Range No. pictures Hair samples Genotypes 86 42 118 55 (64.0%) 32 (76.2%) 51 (43.2%) 20 (36.4%) 11 (34.4%) 15 (29.4%) 227 �Table 2. Model selection results based on Akaike’s Information Criterion that has been corrected for small sample sizes (AICc) for 2014 mark– resight analysis for cougars in the Front Range of Colorado, USA. Model selection was based on models with either constant or time‐varying (j) parameters for the number of unmarked individuals (Uj), mean resighting rate (αj), and presence of individual heterogeneity (σ2j). Number of model parameters (K), AICc weights (wi), and deviance are also reported. Uj Constant Constant Constant Constant Time Time Time Time αj σ 2j K ΔAICc wi Deviance Constant Time Constant Time Constant Constant Time Time No No Yes Yes No Yes No Yes 2 7 3 8 7 8 12 13 0.00 8.23 9.11 10.24 17.03 19.24 23.68 25.17 0.97 0.02 0.01 <0.01 <0.01 <0.01 <0.01 <0.01 178.7 179.2 187.5 177.3 179.8 177.8 169.8 168.2 DISCUSSION To our knowledge, our study is one of the first to use auditory calls to estimate cougar density and provide an estimate of cougar density within a wildland–urban interface. Colorado has relied on cougar population projections based on habitat assessments and published density estimates for management purposes. This is a common approach used by most management agencies for estimating cougar numbers, but our approach provides an alternative that should produce realistic population estimates at the scale of typical management units. In addition to this, we provide the first known estimate of cougar density in the urban–wildland interface of the Front Range of Colorado. Fragmented landscapes that are dominated by private property are very restrictive on techniques that can be used to sample cougars. Our approach has the potential to be applied across the urban–wildland interface in Colorado and in other exurban areas throughout the west where estimates of cougar densities are needed. We had limited success extracting hair samples and genotypes from cougars. We have visual evidence of cougars entering cubbies with strands of barbed wire rubbing on the animal that nevertheless resulted in no hair samples. Although failure to collect hair could be attributed to barbs failing to pull hair from cougars or by hair being windblown from barbs, low success genotyping collected hair was likely due to the collection of shed hair lacking roots rather than pulled hair that typically contain roots. Russell et al. (2012) reported similar results from hair samples collected from back‐tracking cougars, with only 23% of back‐tracks and 13% of 165 hair samples resulting in DNA of sufficient quality for individual identification. False assignments to known individuals were also an issue (13%) in our study. Low‐quality DNA can partially explain this because some level of genotyping error is common (allelic dropout and false alleles), but sample contamination with other cougars is also driving false assignments. Multiple cougars (females with dependent young) entering a site during a sampling period likely leave samples from all of these individuals and if they cross the snag in the same location these samples will be 228 mixed. Sawaya et al. (2017) report genotyping errors that were sufficient to inflate minimum counts from 20 to 36 for cougars (overestimate of 44%) and from 49 to 64 for wolverines (Gulo gulo; overestimate of 23%), well above the 13% error we report. All cougar detections occurred at cubbies with predator calls, but proper site selection that improves the effectiveness of calls is important. We recommend the random selection of 1‐km2 grids with the selection of the “best” location within that by human judgment and experience (i.e., areas where cougars typically travel). Based on our assessment of collared cougar visits to cubbies with predator calls, we believe cougars are more likely influenced by calls when they are within 1 km of a call. Yeager (2016) tested the effective distance of the call using a decibel meter and examining GPS locations of collared cougars and found that 550 m was probably the maximum effective distance for these calls to elicit a response from a cougar. Certainly, increasing the density of calls would also increase the detection probability, but this often is not realistic because of time and money, so we emphasize the importance of site selection. Our mark–resight model allowed for detection with replacement, and was ideal for camera and call data because it accounted for detections of the same individual cougar within a sampling period (McClintock and White 2009). However, these detections should still remain independent from each other, which can be a problem if a cougar has a kill near the call and revisits frequently. Examination of GPS data in situations where a collared cougar was detected multiple times in the same night revealed that all of these cases (n = 3) were associated with a kill within 400 m of the call and camera. Therefore, we suggest that, when building the detection history, multiple observations of an individual at the same site in the same night be counted as a single detection. Both spatially corrected and uncorrected estimates of cougar density we obtained were greater than most other published estimates of cougar density. Cougar densities (not corrected for space use) reported from other studies typically ranged between 1.2 and 4.7 cougars/100 km2 (Logan et al. 1986, Ross and Jalkotzy 1992, Choate et al. 2006). Cougar densities corrected for space use were reported at 4.5 and 5.2 cougars/100 km2 in the Bitterroot Mountains of Montana, a lightly hunted population (Proffitt et al. 2015). Greater estimates in our study could be related to methodological differences, because most other reported densities are based on intensive marking studies and likely represent minimum densities, as also noted by Proffitt et al. (2015). It is likely that not all portions of a cougar population are sampled with intensive marking studies because some animals are harder to detect or occur in remote areas, which can be accounted for using calls and cameras. Additionally, high cougar densities observed on the Front Range may have been influenced by low hunting pressure in our study area due to limited access to private lands, as well as abundant prey. However, Choate et al. (2006) reported a cougar density of 2.2 cougars/100 km2 in an unhunted population. Wildlife Society Bulletin • 43(2) �Correcting cougar population densities for space use is very important; failing to correct for this space use will result in overestimates of density because the area being used by animals is actually larger than what is assumed by the sampling grid (White and Shenk 2001). Our estimated density after correcting for the amount of time on and off grid for cougars was at the lower bound of the 95% confidence interval of the uncorrected estimate. Accounting for this is especially important for cougars because they are highly mobile and utilize large home ranges (Dickson and Beier 2002; females ~100 km2 with some males >750 km2, unpublished data from this study). With the exception of Beausoleil et al. (2016) and Proffitt et al. (2015), densities reported above were not corrected for spatial extent and therefore, could be biased high. Spatially explicit capture–recapture models, could also be used in a similar fashion to estimate density with our data, especially if telemetered animals were not available (Efford et al. 2004, Efford and Fewster 2013). We were mostly unsuccessful in genotyping all individuals that visited a site, but continuing to develop the NGS approach to reliably obtain quality genetic samples from cougars is worthwhile. This technique would provide population estimates at a lower cost than mark–resight and would remove all need to capture and mark individuals. This approach would also provide additional information beneficial to managers. Sex structure of the population could be estimated from these data. In addition, an assessment of population‐level diet composition could be assessed from hair samples using stable isotope analysis (Moss et al. 2016). MANAGEMENT IMPLICATIONS Colorado, and other western states, currently base cougar population estimates primarily on habitat assessments and published density estimates. We have presented an alternative approach that has the potential to give robust density estimates of cougars across large spatial scales that are equivalent to scales at which cougars are managed. Additionally, this approach has the potential to be applied in logistically difficult areas that may preclude other approaches, such as the exurban Front Range of Colorado where private property limits access. Accounting for how mobility of cougars affects population estimates is also important. Typical harvest management strategies for cougars are quota systems and failure to correct for space use in estimates could result in an overharvest. Based on the correction factor estimated in this study, a 15% harvest rate of the uncorrected population estimate may actually be a 21% harvest rate based on the corrected density estimate. This difference could certainly make a difference in whether prescribed harvest rates are maintaining stable populations or causing declines. ACKNOWLEDGMENTS This project was funded by Colorado Division of Wildlife Federal Aid in Wildlife Restoration Project W‐153‐R and Colorado Division of Wildlife game cash funds. We appreciate all of the field efforts of the technicians on this study. We also appreciate the numerous land owners, including county and city properties that allowed us access. The efforts of those that reviewed this manuscript were greatly appreciated. Many thanks to the Associate Editor and reviewers that commented on this manuscript and their insightful suggestions. LITERATURE CITED Alldredge, M. W. 2012. Cougar demographics and human interactions along the urban–exurban Front Range of Colorado. W‐204‐R1 Annual Report, Colorado Parks and Wildlife, Fort Collins, USA. Alldredge, M. W. 2016. Cougar demographics and human interactions along the urban–exurban Front Range of Colorado. W‐204‐R1 Annual Report, Colorado Parks and Wildlife, Fort Collins, USA. Anderson, C. R. Jr., and F. G. Lindzey. 2005. Experimental evaluation of population trend and harvest composition in a Wyoming cougar population. Wildlife Society Bulletin 33:179–188. Anderson, C. R., F. G. Lindzey, and D. B. McDonald. 2004. Genetic structure of cougar populations across the Wyoming Basin: metapopulation or megapopulation. Journal of Mammalogy 85:1207–1214. Beausoleil, R. A., D. Dawn, D. A. Martorello, and C. P. Morgan. 2008. Cougar management protocols: a survey of wildlife management agencies in North America. Pages 205–241 in D. E. Toweill, S. Nadeau, and D. Smith, editors. Proceedings of the Ninth Mountain Lion Workshop. Idaho Department of Fish and Game, Sun Valley, USA. Beausoleil, R. A., J. D. Clark, and B. T. Maletzke. 2016. A long‐term evaluation of biopsy dart and DNA to estimate cougar density: an agency–citizen science collaboration. Wildlife Society Bulletin 40:583–592. Blecha, K. A., R. B. Boone, and M. W. Alldredge. 2017. Hunger mediates apex predator’s risk avoidance response in wildland–urban interface. Journal of Animal Ecology 87:609–622. Boersen, M. R., J. D. Clark, and T. L. King. 2003. Estimating black bear population density and genetic diversity at Tensas River, Louisiana using microsatellite DNA markers. Wildlife Society Bulletin 31:197–207. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multi‐ model inference: a practical information theoretic approach. Second edition. Springer‐Verlag, New York, New York, USA. Chamberlain, M. J., J. W. Mangrum, B. D. Leopold, and E. P. Hill. 1999. A comparison of attractants used for carnivore track surveys. Proceedings of the Southeastern Association of Fish and Wildlife Agencies. Southeastern Association Fish & Wildlife Agencies 53:296–304. Choate, D. M., M. L. Wolfe, and D. C. Stoner. 2006. Evaluation of cougar population estimators in Utah. Wildlife Society Bulletin 34:782–798. Cougar Management Guidelines Working Group [CMGWG]. 2005. Cougar management guidelines. First edition. WildFutures, Bainbridge Island, Washington, USA. Crooks, K. R., M. Grigione, A. Scoville, and G. Scoville. 2008. Exploratory use of track and camera surveys of mammalian carnivores in the Peloncillo and Chiricahua Mountains of southeastern Arizona. Southwestern Naturalist 53:510–517. Cunningham, S. C., L. A. Haynes, C. Gustavson, and D. D. Haywood. 1995. Evaluation of the interaction between mountain lions and cattle in the Aravaipa–Klondyke area of southeast Arizona. Arizona Game and Fish Department Technical Report No. 17, Phoenix, USA. Davidson, G. A., D. A. Clark, B. K. Johnson, L. P. Waits, and J. R. Adams. 2014. Estimating cougar densities in northeast Oregon using conservation detection dogs. Journal of Wildlife Management 78:1104–1114. Dickson, B. G., and P. Beier. 2002. Home‐range and habitat selection by adult cougars in southern California. Journal of Wildlife Management 66:1235–1245. Downey, P. J., E. C. Hellgren, A. Caso, S. Carvajal, and K. Frangioso. 2007. Hair snares for noninvasive sampling of felids in North America: do gray foxes affect success? Journal of Wildlife Management 71:2090–2094. Efford, M., D. Dawson, and C. Robbins. 2004. DENSITY: software for analyzing capture–recapture data from passive detector arrays. Animal Biodiversity and Conservation 27:217–228. Alldredge et al. • Less Invasive Monitoring of Cougars in Colorado’s Front Range 229 �Efford, M. G., and R. M. Fewster. 2013. Estimating population size by spatially explicit capture–recapture. Oikos 122:918–928. Ernest, H. B., M. C. T. Penedo, B. P. May, M. Syvanen, and W. M. Boyce. 2000. Molecular tracking of mountain lions in the Yosemite Valley region in California: genetic analysis using microsatellites and faecal DNA. Molecular Ecology 9:433–441. Gompper, M. E., R. W. Kays, J. C. Ray, S. D. Lapoint, D. A. Bogan, and J. R. Cryan. 2006. A comparison of noninvasive techniques to survey carnivore communities in northeastern North America. Wildlife Society Bulletin 34:1142–1151. Goossens, B., L. P. Waits, and P. Taberlet. 1998. Plucked hair samples as a source of DNA: reliability of dinucleotide microsatellite genotyping. Molecular Ecology 7:1237–1241. Harrison, R. L. 1997. Chemical attractants for Central American felids. Wildlife Society Bulletin 25:93–97. Harrison, R. L. 2006. A comparison of survey methods for detecting bobcats. Wildlife Society Bulletin 34:548–552. Harrison, R. L., P. B. S. Clarke, and C. M. Clarke. 2004. Indexing swift fox populations in New Mexico using scats. American Midland Naturalist 151:42–49. Kendall, W. L. 1999. Robustness of closed capture‐recapture methods to biolations of the closure assumption. Ecology 80:2517–2525. Kendall, K. C., J. B. Stetz, D. A. Roon, L. P. Waits, J. B. Boulanger, and D. Paetkau. 2008. Grizzly bear density in Glacier National Park, Montana. Journal of Wildlife Management 72:1693–1705. Lindzey, F. G., W. D. Van Sickle, B. B. Ackerman, D. Barnhurst, T. P. Hemker, and S. P. Laing. 1994. Cougar population dynamics in southern Utah. Journal of Wildlife Management 58:619–624. Logan, K. A. 1983. Mountain lion population and habitat characteristics in the Big Horn Mountains of Wyoming, Thesis, University of Wyoming, Laramie, USA. Logan, K. A., L. Irwin, and R. Skinner. 1986. Characteristics of a hunted mountain lion population in Wyoming. Journal of Wildlife Management 50:648–654. Logan, K. A., and L. L., Sweanor. 2001. Desert puma: evolutionary ecology and conservation of an enduring carnivore, Island Press, Washington, D.C., USA. Long, E. S., D. M. Fecske, R. A. Sweitzer, J. A. Jenks, B. M. Pierce, and V. C. Bleich. 2003. Efficacy of photographic scent stations to detect mountain lions. Western North American Naturalist 63:529–532. McClintock, B. T., and G. C. White. 2009. A less field‐intensive robust design for estimating demographic parameters with mark–resight data. Ecology 90:313–320. McClintock, B. T., G. C. White, M. F. Antolin, and D. W. Tripp. 2009. Estimating abundance using mark–resight when sampling is with replacement or the number of marked individuals is unknown. Biometrics 65:237–246. McDaniel, G. W., K. S. McKelvey, J. R. Squires, and L. F. Ruggiero. 2000. Efficacy of lures and hair snares to detect lynx. Wildlife Society Bulletin 28:119–123. Menotti‐Raymond, M. A., and S. J. O’Brien. 1995. Evolutionary conservation of ten microsatellite loci in four species of Felidae. Journal of Heredity 86:319–322. Moss, W. E., M. W. Alldredge, and J. N. Pauli. 2016. Quantifying risk and resource use for a large carnivore in an expanding urban–wildland interface. Journal of Applied Ecology 53:371–378. Murphy, K. M. 1998. The ecology of the cougar (Puma concolor) in the northern Yellowstone ecosystem: interactions with prey, bears, and humans, Dissertation, University of Idaho, Moscow, USA. Otis, D. L., K. P. Burnham, G. C. White, and D. R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildlife Monographs 62. Proffitt, K. M., J. F. Goldberg, M. Hebblewhite, R. Russell, B. S. Jimenez, H. S. Robinson, K. Pilgrim, and M. K. Schwartz. 2015. Integrating resource selection into spatial capture–recapture models for large carnivores. Ecosphere 6(11):e239. Ross, P. I., and M. G. Jalkotzy. 1992. Characteristics of a hunted population of cougars in southwestern Alberta. Journal of Wildlife Management 56:417–426. Russell, R. E., J. A. Royle, R. Desimone, M. K. Schwartz, V. L. Edwards, K. P. Pilgrim, and K. S. McKelvey. 2012. Estimating abundance of 230 mountain lions from unstructured spatial sampling. Journal of Wildlife Management 76:1551–1561. Sargeant, G. A., D. H. Johnson, and W. E. Berg. 1998. Interpreting carnivore scent‐station surveys. Journal of Wildlife Management 62:1235–1245. Sawaya, M. A., C. B. Anton, M. Barrueto, A. P. Clevenger, H. B. Quigley, T. K. Ruth, D. R. Stahler, and C. C. Wilmers. 2017. Quality control measures reveal substantial effects of genotyping errors on DNA‐based mark–recapture results. Pages 180–181 in C. R. McLaughlin and M. Vieira, editors. Proceedings of the 12th mountain lion workshop, May 15– 18, 2017, Estes Park, Colorado, USA. Sawaya, M. A., T. K. Ruth, S. Creel, J. J. Rotella, J. B. Stetz, H. B. Quigley, and S. T. Kalinowski. 2011. Evaluation of noninvasive genetic sampling methods for cougars in Yellowstone National Park. Journal of Wildlife Management 75:612–622. Schmidt, K., and R. Kowalczyk. 2006. Using scent marking stations to collect hair samples to monitor Eurasian lynx populations. Wildlife Society Bulletin 32:462–466. Settlage, K. E., F. T. Van Manen, J. D. Clark, and T. L. King. 2008. Challenges of DNA‐based mark–recapture studies of American black bears. Journal of Wildlife Management 72:1035–1042. Sinclair, E. A., E. L. Swenson, M. L. Wolfe, D. C. Choate, B. Bates, and K. A. Crandall. 2001. Gene flow estimates in Utah’s cougars imply management beyond Utah. Animal Conservation 4:257–264. Smallwood, K. S. 1994. Trends in California mountain lion populations. Southwestern Naturalist 39:67–72. Smallwood, K. S., and E. L. Fitzhugh. 1991. The use of track counts for mountain lion population census. Pages 59–67 in C. E. Braun, editor. Mountain Lion–Human Interactions Symposium and Workshop. Colorado Division of Wildlife, Denver, USA. Smallwood, K. S., and E. L. Fitzhugh. 1995. A track count for estimating mountain lion Felis concolor californica population trend. Biological Conservation 71:251–259. Smith, D. A., K. Ralls, B. L. Cypher, and J. E. Maldonado. 2005. Assessment of scat‐detection dog surveys to determine kit fox distribution. Wildlife Society Bulletin 33:897–904. Taberlet, P., S. Griffin, B. Goossens, S. Questiau, V. Manceau, N. Escaravage, L. P. Waits, and J. Bouvet. 1996. Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Research 24:3189–3194. Taberlet, P., and G. Luikart. 1999. Non‐invasive genetic sampling and individual identification. Biological Journal of the Linnean Society 68:41–55. Van Dyke, F. G., R. H. Brocke, and H. G. Shaw. 1986. Use of road track counts as indices of mountain lion presence. Journal of Wildlife Management 50:102–109. Van Sickle, W. D., and F. G. Lindzey. 1992. Evaluation of road track surveys for cougars (Felis concolor). Great Basin Naturalist 52:232–236. Weaver, J. L., P. Wood, D. Paetkau, and L. L. Laack. 2005. Use of scented hair snares to detect ocelots. Wildlife Society Bulletin 33:1384–1391. White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capture–recapture removal methods for sampling closed populations, Los Alamos National Laboratory Publication LA‐8787‐NERP, Los Alamos, New Mexico, USA. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46: 120–139. White, G. C., and M. T. Shenk. 2001. Population estimation with radio‐ marked individuals. Pages 329–350 in J. Millspaugh, and J. M. Marzluff, editors. Radio tracking and animal populations. Academic Press, San Diego, California, USA. Williams, B. K., J. D. Nichols, and M. J. Conroy. 2002. Analysis and management of animal populations. Academic Press, San Diego, California, USA. Yeager, K. L. 2016. A noninvasive method using auditory predator calls and hair snares to detect and genetically sample cougars (Puma concolor), Thesis, Colorado State University, Fort Collins, USA. Associate Editor: Hinton. Wildlife Society Bulletin • 43(2) � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Journal Articles Description An account of the resource CPW peer-reviewed journal publications Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Less invasive monitoring of cougars in Colorado's Front Range Description An account of the resource <span>From 2014 to 2016, in the Front Range of Colorado, USA, we assessed noninvasive approaches to sampling cougar (</span><i>Puma concolor</i><span>) populations in an attempt to provide a new method that would be less field intensive, less expensive, and could be applied over large spatial extents compared with current methods. We assessed the use of predator calls to lure cougars to a site with remote camera traps for detection and also evaluated hair snags at sites to noninvasively identify individual animals. Predator calls effectively attracted cougars to specific sites with an average of 82 unique photographic detections of cougars per survey year (0.03 detections/trap-night). However, obtaining hair samples from these animals was less effective because animals did not always pass through hair snags and ability to uniquely identify individuals by genotype was poor. We evaluated different approaches to estimating cougar density and found mark–resight to be a viable option in our study system. Mark–resight density estimate after correcting for partial use of the sampling area by cougars was 4.1 cougars/100 km</span>²<span> (95% CI = 2.4, 5.8). Our results indicate that combining methods of noninvasive genetic sampling and auditory calls to monitor cougar populations can provide reliable density estimates over large geographic areas and areas with significant amounts of inaccessible private lands.</span> Bibliographic Citation A bibliographic reference for the resource. Recommended practice is to include sufficient bibliographic detail to identify the resource as unambiguously as possible. <p>Alldredge, M. W., T. Blecha, and J. H. Lewis. 2019. Less invasive monitoring of cougars in Colorado's Front Range. Wildlife Society Bulletin 43:222–230. <a href="https://doi.org/10.1002/wsb.971" target="_blank" rel="noreferrer noopener">https://doi.org/10.1002/wsb.971</a></p> Creator An entity primarily responsible for making the resource Alldredge, Mathew W. Blecha, Tasha Lewis, Jonathan H. Subject The topic of the resource Abundance Auditory Cougar Mark–resight Noninvasive Predator calls <em>Puma concolor</em> Extent The size or duration of the resource. 9 pages Date Created Date of creation of the resource. 2019-07-05 Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/InC-NC/1.0/" target="_blank" rel="noreferrer noopener">In Copyright - Non-Commercial Use Permitted</a> Format The file format, physical medium, or dimensions of the resource application/pdf Language A language of the resource English Is Part Of A related resource in which the described resource is physically or logically included. Wildlife Society Bulletin Type The nature or genre of the resource Article