Wildlife Population Management I

Contributed Oral

Desert Bighorn Sheep Population Dynamics and Resource Selection in the San Andres Mountains, New Mexico
Miranda Strasburg, Lindsay Smythe

The San Andres Mountains in south-central New Mexico represent one of the largest, relatively undisturbed Chihuahuan Desert landmasses in the US and serve as crucial habitat for desert bighorn sheep (Ovis canadensis). However, their vast range creates a challenge to managers monitoring bighorn sheep abundance. To inform aerial survey routes, in 2017-2018, GPS collars were placed on 13 bighorn sheep to identify their daily locations. We used these GPS locations to determine how habitat use varies between sexes and to develop a resource selection function (RSF) model to determine areas preferentially used by bighorn sheep. Overall, these bighorn sheep use similar habitat regardless of sex, though ewes use steeper habitats closer to grasslands than rams. The results of our RSF model suggest that topography plays a greater role in habitat selection than resource availability, though further analysis is needed to verify if this is the case. We also constructed a stage-based projection matrix for female bighorn sheep with three life-history stages — lambs, yearlings, and adults— using population estimates from aerial surveys and vital rate estimates from the collars to estimate population dynamics. Under a low reproduction scenario, we observed dramatic decreases in the population over time with this effect being mitigated when the percent of ewes having female lambs was increased. Elasticity analyses revealed that ewe survival has the greatest impact on population growth. Indeed, if ewe survival is decreased by only 10%, the population is expected to decline, while large changes in lamb survival have little influence on population growth. Given that ewes play the largest role in population growth and that both rams and ewes tend to use habitat with similar characteristics, future monitoring efforts in this region should focus on areas most utilized by ewes and management strategies that maximize ewe survival. 

G.U.T. Check – GPS, Ultrasound, and Transmitters to Properly Evaluate Reproduction in Elk
Avery Corondi, Jeremiah Banfield, W. David Walter, Justin Brown

Understanding reproduction and recruitment of ungulate populations is necessary for managing a sustainable population and creating comprehensive management plans. Investigating natality and subsequent juvenile survival is costly, labor intensive, and difficult but is often necessary to thoroughly assess reproductive output in large mammals. Advances in technology have allowed us to examine these relationships more efficiently. In the winters of 2020 and 2021, vaginal implant transmitters (VITs) paired with GPS-capable collars were deployed in pregnant Rocky Mountain cow elk (Cervus canadensis). Elk were chemically immobilized with one of two drug combinations, Butorphanol-Azaperone-Medetomidine (BAM) or Nalbuphine-Medetomidine-Azaperone (NalMed-A) that were administered with remote drug delivery systems either through free-darting or after capture in Clover traps. Pregnancy was assessed in the field using ultrasound and later confirmed by pregnancy-specific protein B (PSPB) analysis of serum. VITs provided notification of parturition events and calves were located and equipped with GPS collars within 24 hours of birth. We intensively monitored calf movements and assessed sources of mortality. Overall, both drug combinations provided adequate sedation for processing the cow elk; however, partial or full redosing was occasionally necessary with NalMed-A to reach a suitable level of sedation. One hundred VITs were deployed during 2020-2021. Of the 50 VITs deployed in 2020, 47 were expelled and resulted in 45 calves located. Ultrasound pregnancy results were 99% accurate relative to PSPB serum results, with one cow diagnosed as open via ultrasound and pregnant based on PSPB. Sources of calf mortality include dystocia, predation, disease, starvation, trauma, malnutrition, and harvest. Calves equipped with GPS collars will provide additional information to mortality sources as well as maternal investment and how it influences calf survival. This is the first study to document details on pregnancy rates, calf movements, and mortalities using a variety of recent technological advances.

Population Dynamics of a Declining White-Tailed Deer Population in Northern Georgia
Adam Edge, Jacalyn Rosenberger, Charlie Killmaster, Christina Johannsen, David Osborn, Karl Miller, Gino D’Angelo

Throughout the Chattahoochee National Forest (CNF) in the Appalachians of northern Georgia, white-tailed deer (Odocoileus virginianus) populations have been in decline for decades. Based on hunter harvest data from 8 wildlife management areas (WMAs) within the CNF, buck harvest success rates (i.e. bucks harvested /hunter/days available) declined 64% from 1979-2018. Georgia Department of Natural Resources instituted increasingly restrictive harvest regulations for antlerless deer, but populations failed to recover. In an associated study from 2018-2020 on Blue Ridge and Coopers Creek WMAs within the CNF, annual fawn survival was estimated to be 0.168 (95% CI=0.098-0.288) with predation as the leading cause of mortality. Thus, low fawn recruitment is likely a primary cause of population decline. We radio-collared 14 yearling and 45 adult deer in concurrence with fawns to estimate field-based vital rates (e.g., survival and fecundity) to parameterize stage-structured, female-only population models with 3 life-stages. We developed 5 scenarios that incorporated antlerless harvest restrictions and varying fawn survival rates: 1) observed survival rates, 2) no antlerless harvest, 3) 5% antlerless harvest, 4) observed survival rates + moderate fawn survival (0.270), and 5) observed survival rates + high fawn survival (0.430). Our objective was to compare annual population growth rates (λ) of each scenario projected 10 years into the future for management insight. We hypothesized that antlerless harvest restrictions and increased fawn survival would be needed to achieve positive growth. Our results suggest restricting antlerless harvest alone is insufficient to recover populations (λ = 0.965 – 0.982). Scenarios that included antlerless harvest restrictions in addition to moderate to high fawn survival were the only scenarios resulting in positive population growth (λ = 1.017 – 1.084). Therefore, management strategies to improve fawn survival in coordination with antlerless harvest restrictions are likely necessary for deer population sustainability in the southern Appalachian region.

Assessing Impacts of Common Raven Density on Great Sage-Grouse to Develop Science-Driven Adaptive Management Strategies
Sarah Webster, Peter Coates, Shawn O’Neil, Brianne Brussee, Seth Dettenmaier, Cali Roth, John Tull, Mark Ricca, Pat Jackson, Jonathon Dinkins, Ann Moser, Lee Foster, David Delehanty

The common raven (Corvus corax; hereafter raven) is a behaviorally flexible predator with drastically increasing populations that negatively impact sensitive prey species, including greater sage-grouse (Centrocercus urophasianus; hereafter sage-grouse). Accurate estimates of raven density remain difficult to obtain, and effective, adaptive management protocols are needed to mitigate negative impacts of increasing raven populations. We mapped raven density across the Great Basin, USA, and evaluated effects of density on sage-grouse nest survival in order to estimate a predator-prey conflict threshold for sage-grouse. We found increasing raven density adversely impacted sage-grouse nest survival, especially above a threshold of ~0.40 ravens km2. Importantly, average raven density across our study extent was 0.54 ravens km2 (95% CI = 0.42–0.70). We used underlying data to demonstrate a science-based adaptive approach to inform management of ravens in western landscapes with emphasis on sage-grouse habitats. We also developed a map delineating areas used by breeding versus non-breeding ravens (i.e. resident or non-resident, respectively) as local management options could vary based on raven breeding status. Our science application is amenable to different management objectives and is a valuable resource for managers wanting to ameliorate impacts of ravens on sage-grouse and other sensitive species. Findings are preliminary and provided for best timely science.

Changes in Spatial Distribution Decoupled from Abundance for Greater Sage-Grouse in Bi-State Distinct Population Segment
Megan Milligan, Peter Coates, Mark Ricca, Brian Prochazka, Shawn O’Neil, John P. Severson, Steven Mathews, Shawn Espinosa, Scott Gardner, Sherri Lisius, David Delehanty

Changes in distribution or abundance are frequently used independently to evaluate trends and status of wildlife populations. It is often assumed, although not explicitly stated, that these measures are correlated. However, if population distribution and abundance become disconnected, such as when changes in a subpopulation drive overall population trends, then focusing on a single indicator, such as abundance, can mask important losses in distribution. We evaluated changes in both population abundance and distribution from 1995 to 2021 for greater sage-grouse (Centrocercus urophasianus) in the Bi-State Distinct Population Segment, a genetically distinct and isolated population straddling the border of Nevada and California on the edge of the sage-grouse geographic range. First, we used an integrated population model incorporating lek counts and demographic data to predict annual population abundance across all sub-populations (i.e., lek complexes). Second, we used telemetry data to develop phenological and reproductive life-stage resource selection functions to map predicted habitat. Finally, we used predicted abundances to estimate a weighted utilization distribution of space use that, with the predicted habitat maps, allowed us to evaluate changes in population distribution over time. Although overall population abundance remained stable over both the short- ( λ  = 0.99, 95% CRI = 0.70-1.30) and long-term ( λ  = 1.02, 95% CRI = 0.74-1.42), the distribution of occupied habitat declined. This was due to significant losses among 3 sub-populations, while one larger subpopulation expanded, translating to a loss of total area over time and a 77% probability that the range of any given subpopulation contracted. The contractions in distribution combined with stable population trends suggests long-term patterns in redistribution of sage-grouse among subpopulations, with peripheral subpopulations declining while the largest core population increased. This decoupling between trends of abundance and distribution could have implications for metapopulation persistence as peripheral populations become more vulnerable

Machine Learning and Computer Vision for Wildlife Population Management
Mikey Tabak, John Lombardi

The ability to observe wildlife remotely is central to population management. Camera traps, acoustic detectors, and flights of crewless aircraft (drones and fixed-wing aircrafts) are often employed in remote sensing projects. These remote sensing operations collect large amounts of images or sound files that must be processed and analyzed before the data can be used to inform management decisions. Deep learning provides a tool to automatically and rapidly process these data. We used computer vision and deep learning to build models that automatically process images and acoustic recordings of wildlife. Specifically, we evaluated the effectives of using image classification, object detection, and object segmentation computer vision models to analyze remote sensing data. We found that object detection models are able to rapidly classify, detect, and count animal species in camera trap images, as well as images from flights of drones and fixed wing aircraft, with accuracies of 90-98%. We also found that image classification models were able to classify bat species from acoustic recordings with 93% accuracy. We found that object detection and object segmentation models are more effective than classification models at removing empty images from datasets, and they have the advantage of automatically counting and locating animals within images. These models can rapidly (100-1,000 images per minute) process data on laptop computers.

The ability to automatically process images and acoustic recordings allows researchers to expand and more rapidly obtain data for wildlife monitoring programs. Furthermore, these models can be deployed for continuous monitoring and allow for data acquisition in “real time.” In this presentation, I will introduce these different types of deep learning computer vision methods, provide results from our research, and describe applications in wildlife management.

Testing a New Passive Acoustic Recording Unit to Monitor Wolves
Shannon Barber-Meyer, Vicente Palacios, Barbara Marti Domken, Lori Schmidt

As part of a broader trial of noninvasive methods to research wild wolves (Canis lupus) in Minnesota, USA, we explored whether wolves could be remotely monitored using a new, inexpensive, remotely-deployable, noninvasive, passive acoustic recording device, the AudioMoth. We tested the efficacy of AudioMoths in detecting wolf howls and factors influencing detection by placing them at set distances from a captive wolf pack and compared those recordings with real-time, on-site howling data between May 22 – June 17, 2019. We identified 1531 vocalizations that we grouped into 428 vocal events (236 solo howl series and 192 chorus howls). The on-site AudioMoth correctly recorded 100% of chorus and solo howls that were also documented in real-time. The remote array detected 49.5% of chorus and 11.9% of solo howls (at least one unit detected the event). The closest remote AudioMoth (0.54 km, 1/3 mi) detected 37% of choruses and 8.9% of solo howls. Chorus howls (9.4%) were detected at even the farthest unit (3.2 km, 2 mi). Favorable wind (carrying source howls to the remote units) and calm (no wind) conditions increased detectability and detection distance of chorus howls. Temperature was inversely related to detection. Given the detection distances we observed, AudioMoths are probably most useful in studying specific sites during periods when wolves move less frequently (e.g., during late spring and summer at homesites or potentially during winter at kill sites of very large prey). AudioMoths would also be useful in a passive sampling array (e.g., occupancy studies), especially when in concert with other methods such as camera trapping (Garland et al. 2020). Additional research should be conducted in areas with different environmental variables (e.g., wind, temperature, habitat, topography) to determine performance under varying conditions and also when fitted with a parabolic dish.

Demographics of Gray Wolf Packs Recolonizing Variable Habitat Types in Central Wisconsin
Theresa Simpson

Most gray wolves in the United States live in a series of disjunct populations.  Management of these isolated populations is important in sustaining the species.  The Central Forest Region (CFR) of Wisconsin has been home to one such disjunct population since the early 1990’s.  My objective was to determine if individual gray wolf pack demographics were impacted by either time period of recolonization or pack territory habitat quality. Using 18 years of historic CFR wolf monitoring data, I divided this gray wolf recolonization into three distinct time periods: Early (1994-1999), Mid (2000-2005), and Late (2006-2012).  I used GIS to define habitat classes of individual pack territories as Optimal, Mixed, and Marginal, based on features known to influence wolf habitat selection or avoidance. These were: (1) percent public land, (2) percent agriculture, and (3) road density (km/ km2).  I analyzed the influence of time periods and habitat classes on pack territory size, mid-winter pack size, pup production, wolf-human conflicts, human-caused wolf mortalities, territory persistence, and sustained reproductive performance.  I found pack demographics were similar across time periods, except for pup production that was slightly lower during the Mid Time Period. Wolf-human conflicts increased significantly over time, but when wolf population size was accounted for, the time effect disappeared.  In contrast to the weak effects of time period, packs existing in Marginal Habitat had smaller mid-winter pack sizes and lower reproductive performance, and experienced greater conflicts with humans, and six-time greater human-caused mortalities than the other habitat classes.  This study provides possible habitat parameter thresholds of wolf pack tolerance to human-altered environments and identifies which wolf pack demographic parameters most impact wolves in human-dominated environments.  This method of wolf territory assessment can be used to direct wolf hunts to Marginal Habitat, while limiting wolf hunts in Optimal and Mixed Habitat.  

Contributed Oral
Location: Virtual Date: November 2, 2021 Time: 3:00 pm - 4:00 pm