From the Field: Capture of white-tailed deer fawns using thermal imaging technology

Transcription

1 1164 From the Field: Capture of white-tailed deer fawns using thermal imaging technology Stephen S. Ditchkoff, Joshua B. Raglin, Jordan M. Smith, and Bret A. Collier Abstract Capture of neonatal white-tailed deer (Odocoileus virginianus) often is hampered by inherent difficulties in locating study animals. A variety of techniques have been described for location and capture of fawns, including foot searches, female behavioral cues, spotlighting, and vaginal transmitter implants. However, each technique has certain limitations imposed by such factors as habitat structure or logistical difficulties. We describe a new technique for locating deer fawns in which thermal imaging technology was employed. Only 3.3 person-hours were required per fawn located and 9.4 personhours required per fawn captured. We suggest that this technique is equally or more efficient than other reported capture techniques for neonatal white-tailed deer. Key words capture, fawn, neonate, Odocoileus virginianus, thermal imager, white-tailed deer Methods for estimating demographic parameters and evaluating population dynamics commonly require the use of marked individuals (Lebreton et al. 1992). However, wildlife research involving capture of free-ranging subjects often is limited by ability to collect a suitable number of study animals, which in turn reduces predictive ability and accuracy of scientific conclusions. As a result, wildlife biologists are continually improving capture techniques, regardless of the species. In the case of white-tailed deer (Odocoileus virginianus), a variety of capture techniques for adults have been described and employed successfully, including box traps, drop-nets, etc.(see Schemnitz 1994). Neonatal fawns have been located by 1) systematic searches by foot or horseback (Bolte et al. 1970, Ballard et al. 1999), 2) observing behavioral changes in adult females as cues to the presence of a fawn (Downing and McGinnes 1969, White et al. 1972, Bartush and Lewis 1978, Huegel et al. 1985), 3) using deer vocalizations to induce females to provide behavioral cues as to the presence of a fawn (Diem 1954, Arthur et al. 1978), and 4) implanting radiotransmitters into pregnant females that will allow researchers to locate birth sites within hours of parturition (Bowman and Jacobson 1998, Carstensen et al. 2003). Each of these techniques has applicability depending upon deer density, habitat type, research objectives, and economic and logistical constraints. Here we describe a new technique for locating neonatal white-tailed deer with use of thermal imaging technology and discuss its efficiency relative to other published capture techniques. Study area This research was conducted at Brosnan Forest, a 5,830-ha tract located in the lower coastal plain of South Carolina in Dorchester County (N , W ). Brosnan Forest was owned by Wildlife Society Bulletin 2005, 33(3): Peer edited

2 From the Field Ditchkoff et al Norfolk Southern Railway and was managed as an outdoor recreation and conference facility for employees and customers of the company. This unique area had km of navigable roads. The vegetation of the area was comprised mainly of an interspersion of mature longleaf pine (Pinus palustris), bottomland hardwood drains, and mixed hardwoods. Dominant overstory species in hardwood drains were oak (Quercus spp.), sweetgum (Liquidambar styraciflua), black gum (Nyssa sylvatica), and yellow poplar (Liriodendron tulipifera), and in mixed hardwoods were sweetgum, oak, and red maple (Acer rubrum). A more complete description of vegetation on the area was reported by Sanders (2000) and Jordan (2002). In August 2003 deer density of the area was estimated at 31/km 2,and the sex ratio was estimated to be 1.4 does:buck (J. B. Raglin, Norfolk Southern Railway, unpublished data). Methods Thermal imaging system The thermal imaging camera used for locating fawns was a Raytheon PalmIR 250 Digital ( cm; Raytheon Commercial Infrared, Dallas, Tex.) with a 75-mm lens. This lens collects and focuses radiation in the 8 12-micron spectral region and has a field of view (H V) of 12 o 9 o at 1 magnification. The 1.5-kg imaging camera was powered by a nickel-metal-hydroxide rechargeable camcorder battery (6VDC) with an approximate operating time of 3.5 hours. We mounted the camera on a fully adjustable (e.g., height, aspect, and direction) tripod in the back of a pickup truck. The thermal imaging camera has a small 1.5-cm monocular viewfinder and can be used as a handheld imaging device. However, we routed the digital video output from the camera using RG-6 digital satellite cables and a 2-way splitter to both an AC/DC-powered, 23-cm television/vcr combo in the back of the truck and a 10-cm television mounted on the dashboard of the truck (Figure 1). The large television in the back of the truck was powered through the cigarette lighter in the truck and enabled multiple observers to monitor the thermal display in relative comfort, as opposed to only 1 observer being able to stare into the optical viewer of the camera. The television in the cab of the truck enabled the driver to monitor the thermal image while positioning the truck for optimal viewing of potential target animals. 10-cm Figure 1. Diagram illustrating electrical and digital connections between the Raytheon PalmIR 250 thermal imager, a 23-cm TV/VCR, and a 10-cm TV. Electrical connections are RG-6 digital satellite cables and are illustrated by curved lines. Locating fawns We conducted our searches for fawns from 22 April 2 June Deer reproductive data from the study area indicated that the median date of parturition during past years was 7 May, with 75% of the fawns being born between 10 April and 16 May (J. B. Raglin, unpublished data). Searches were conducted during periods of darkness by driving roads on the study area between 2000 and 0800 hours. By monitoring only during darkness, nighttime cooling maximized the heat differential between targets and the surrounding environment (Galligan et al. 2003), thereby increasing fawn detection probabilities. We drove the truck at ~6.5 km/hour (4 miles/hour), as this speed allowed us to maximize area covered while minimizing the chance of missing fawns. Upon location, we approached fawns using red or dimly lit headlamps and captured them by hand or with the aid of a fishing or landing net with a 4.2-m pole and 1-m-diameter net. We found that approaching fawns by a circuitous route,as opposed to a direct line,decreased the likelihood of fawns bolting before we could get close enough for capture. Once fawns were captured, we blindfolded them, took basic body measurements, and tagged them as part of a population analysis for another study. After capture and processing, we measured the direct distance from the fawn location to the vehicle. If an individual capture attempt was unsuccessful, we measured dis-

3 1166 Wildlife Society Bulletin 2005, 33(3): tance from the fawn s initial location to the vehicle. Results Figure 2. Probability density function for sighting distances of all white-tailed deer fawn locations during summer 2004 in South Carolina. Figure 3. Probability density functions for sighting distances of successful and unsuccessful white-tailed deer fawn captures during summer 2004 in South Carolina. During this pilot study, we captured a total of 26 fawns (11 male and 15 female). We also recaptured fawns on 6 occasions but did not include these data in our analysis. Mean mass of captured fawns was 4.60 kg (SE=0.29), and mean hind foot length was mm (SE = 4.3). Mean distance traveled by truck per fawn encountered was 6.06 km, and mean distance traveled for each fawn captured was km. We averaged 1.33 hours and 3.75 hours per fawn encountered and fawn captured, respectively. When adjusting these estimates for the number of observers/drivers (n = 1 4) in the truck, mean time per fawn encountered was 3.33 personhours and mean time per fawn captured was 9.38 person-hours. Fawn encounter rate required 40% fewer person-hours when only 2 workers (x - =2.5; 1 driver and 1 observer) were in the truck than when 3 workers were in the truck (x - =4.1; 1 driver and 2 observers). Similarly, 74% fewer person-hours were required when 2 workers (x - =5.7) were in the truck than when 3 workers were in the truck (x - =22.1) per fawn captured. Mean detection distance for all fawns was 38.5 m (SE = 24.4), and probability of detecting a fawn decreased as distance from the road increased (Figure 2). The greatest distance from the road at which a fawn was detected and successfully captured was 72 m. In a few cases (n=3), we detected fawns at distances greater than 75 m (n=1 at 90 m, and n=2 at 140 m), but we captured none of these deer because the fawns were older and highly mobile. Mean distance for successful captures tended to be shorter (x - =33.3; SE=15.5) than unsuccessful captures (x - =41.5;SE=28.7),although we suggest that probability of capturing a fawn was not influenced by distance from the road (Figure 3). Discussion The Raytheon PalmIR 250 detects heat differential, as opposed to commonly used low-light vision equipment that amplifies existing light so that the human eye can detect objects during low-light conditions. Because the thermal imager detects heat, it can be used during daylight hours where low-light vision equipment cannot. However, this study was conducted from 22 April 2 June 2004, when daytime ambient temperatures may approach the body temperature of a deer. During daytime hours, sun patches on the forest floor, stumps, logs, and trees create false heat signals on the digital image that can be mistaken for animals. While animals can still be detected through careful observation, ease of detection diminishes. This causes mental fatigue and likely results in potential targets being missed. The thermal imaging system was as efficient as or more efficient than other reported methods for catching neonatal white-tailed deer. Huegel et al.

4 From the Field Ditchkoff et al (1985) used doe movement data gathered with radiotelemetry as a cue that a fawn may be near and reported capture efficiencies ranging from person-hours per fawn. They suggested that efficiency improved from over 400 personhours to 14.5 due to experience of the researchers. White et al. (1972) also used doe behavior (e.g., visual observations from elevated towers in open, high-visibility habitat) as a cue to locate newborn fawns. They reported an average of 5 person-hours per successful capture. However, most regions where white-tailed deer are found would likely not be suitable for this technique because of lack of extensive open habitats. White et al. (1972) also indicated that efficiency decreased to approximately 20 person-hours per capture when vegetation was extremely thick or newborn fawn density was low (e.g.,beginning or end of the fawning period). Pusateri (2003) used ground searches to locate fawns and reported that number of person-hours required per captured fawn ranged from hours. Carstensen et al. (2003) used both doe behavior cues and vaginal implant transmitters to locate fawns. They reported that between 145 and 214 person-hours were required per captured fawn when ground-searching in areas near does that provided behavioral cues indicating a fawn was nearby. They further elaborated that between 12% and 21% of their searches were successful, and 60 personhours were required per fawn captured when using the vaginal implant transmitters. However, these data do not include the time required to capture does and implant transmitters. Vegetation and date influenced our capture efficiency. We observed that fawn detection rates using the thermal imaging system were low in areas that had thick herbaceous or shrub cover. The Raytheon PalmIR 250 thermal imaging system does not detect heat through objects, so if thick vegetation is between the target animal and the camera, heat from the animal likely will not be detected (Galligan et al. 2003). As a result, during the pilot study we confined most of our searches to pine and hardwood habitats that had minimal understory or herbaceous growth so that bedded fawns were more easily detected. We originally suspected that longleaf pine habitats with herbaceous understories would be prime habitat for locating fawns. However, only those tracts that had recently (e.g., 1 2 months previously) been burned with prescribed fire proved to be suitable. Near the end of the study, fawn capture success decreased due to the low number of newborn fawns in the area:most fawns were too old and mobile for capture. It was possible that use of a trained dog would have increased capture success for these older fawns. Downing and McGinnes (1969) found that use of a trained dog when capturing fawns increased capture success by approximately 10%. Use of loud noises or vocalizations to incite a fleeing fawn to freeze and drop (Downing and McGinnes 1969) met with little success. In our pilot study, the thermal imaging system enabled our capture efficiency of fawns to equal or exceed all other reported studies. We suspect that during subsequent field efforts,our capture success will increase due to experience with the technology, familiarity with how the system works in different habitats, and improved capture protocol. The thermal imager that we used for this study had a list price of $13,257 (U.S.). We are aware of other thermal imaging systems that range in price from approximately $6,000 to more than $40,000. While we cannot speculate on the effectiveness of other thermal systems for locating deer fawns, we do suspect that our success was not limited by the technology. On many occasions we detected small rodents (Peromyscous spp.) and lagomorphs (Sylvilagus floridanus) at distances approaching 50 m. In addition, we detected a spider at 15 m, fresh deer scat at approximately 15 m, and heat signatures (e.g., warm ground) where deer had recently (1 2 minutes earlier) been bedded at up to 40 m. We are confident that use of this technique could improve capture efficiency of wildlife in many studies and that the economic constraints of purchasing one of these systems for research purposes may well be mediated through economic savings (e.g., hourly wages and travel). In addition, the potential for use on other wildlife-related projects would further mediate cost. Prior studies have reported use of similar thermal imaging systems for surveying deer and other wildlife species, using both aerial and ground-based thermal cameras(naugle et al. 1996, Focardi et al. 2001, Haroldson et al. 2003). Other uses for this technology have yet to be reported, yet we are confident that the limits of its applicability to wildlife research will not be realized for decades. Acknowledgments. Economic and logistical support was provided by Norfolk Southern Railway and the Brosnan Forest during this research, as well as the School of Forestry and Wildlife Sciences at

5 1168 Wildlife Society Bulletin 2005, 33(3): Auburn University. We appreciate the assistance of the numerous volunteers who assisted during the study. We also would like to thank C. Ruth and the South Carolina Department of Natural Resources for approval of this project. The protocol for this study was approved by the Auburn University Institutional Animal Care and Use Committee (# ). Literature cited ARTHUR,W. J., III, G. G. HIATT,AND A.W.ALLDREDGE Response of mule deer to tape recorded fawn distress calls. Wildlife Society Bulletin 6: BALLARD, W. B., H. A. WHITLAW, S. J.YOUNG, R. A. JENKINS, AND G. J. FORBES Predation and survival of white-tailed deer fawns in northcentral New Brunswick. Journal of Wildlife Management 63: BARTUSH, W. S., AND J. C. LEWIS Behavior of whitetail does and fawns during the parturition period. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 32: BOLTE, J. R., J.A. HAIR,AND J. FLETCHER White-tailed deer mortality following tissue destruction induced by lone star ticks. Journal of Wildlife Management 34: BOWMAN, J.L.,AND H. A. JACOBSON An improved vaginalimplant transmitter for locating white-tailed deer birth sites and fawns. Wildlife Society Bulletin 26: CARSTENSEN, M., G. D. DELGIUDICE, AND B.A. SAMPSON Using doe behavior and vaginal-implant transmitters to capture neonate white-tailed deer in north-central Minnesota. Wildlife Society Bulletin 31: DIEM, K. L Use of a deer call as a means of locating deer fawns. Journal of Wildlife Management 18: DOWNING, R. L.,AND B. S. MCGINNES Capturing and marking white-tailed deer fawns. Journal of Wildlife Management 33: FOCARDI, S., A. M. DE MARINIS, M. RIZZOTTO, AND A. PUCCI Comparative evaluation of thermal infrared imaging and spotlighting to survey wildlife. Wildlife Society Bulletin 29: GALLIGAN, E.W., G. S. BAKKAN, AND S. L. LIMA Using a thermographic imager to find nests of grassland birds. Wildlife Society Bulletin 31: HAROLDSON, B. S., E. P. WIGGERS, J. BERINGER, L. P. HANSEN, AND J. B. MCANINCH Evaluation of aerial thermal imaging for detecting white-tailed deer in a deciduous forest environment. Wildlife Society Bulletin 31: HUEGEL, C. N., R. B. DAHLGREN, AND H. L. GLADFELTER Use of doe behavior to capture white-tailed deer fawns. Wildlife Society Bulletin 13: JORDAN, N. E Hatching failure and embryonic mortality in a South Carolina red-cockaded woodpecker (Picoides borealis) population. Thesis, Clemson University, Clemson, South Carolina, USA. LEBRETON, J.-D., K. BURNHAM, J. COLBERT, AND D. R.ANDERSON Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecological Monographs 62: NAUGLE, D. E., J.A. JENKS,AND B. J. KERNOHAN Use of thermal infrared sensing to estimate density of white-tailed deer. Wildlife Society Bulletin 24: PUSATERI, J. S White-tailed deer population characteristics and landscape use patterns in southwestern lower Michigan. Thesis. Michigan State University, East Lansing, USA. SANDERS, F. J Brood reduction and the insurance hypothesis as explanations for asynchronous hatching in red-cockaded woodpeckers. Thesis. Clemson University, Clemson, South Carolina, USA. SCHEMNITZ, S. D Capturing and handling wild animals. Pages in T. A. Bookhout, editor. Research and management techniques for wildlife and habitats. The Wildlife Society, Bethesda, Maryland, USA. WHITE, M., F. F. KNOWLTON, AND W. C. GLAZENER Effects of dam newborn fawn behavior on capture and mortality. Journal of Wildlife Management 39: Address for Stephen S. Ditchkoff and Jordan M. Smith: School of Forestry and Wildlife Sciences, Auburn University, Auburn, AL 36849, USA; for Ditchkoff: ditchss@auburn.edu. Address for Joshua B. Raglin: Norfolk Southern Railway, Brosnan Forest, P.O. Box 27, Dorchester, SC 29437, USA. Address for Bret A. Collier: Arkansas Cooperative Fish and Wildlife Research Unit, Department of Biological Sciences, University of Arkansas, Fayetteville, AR 72701, USA. Stephen S. (Steve) Ditchkoff is an associate professor in the School of Forestry and Wildlife Sciences at Auburn University. He received his Ph.D. in wildlife ecology from Oklahoma State University, M.S. in wildlife ecology from University of Maine, and B.S. in fisheries and wildlife from Michigan State University. His primary research interests focus on the evolutionary ecology of large mammals, with specific emphasis on white-tailed deer. Joshua B. (Josh) Raglin has been working for Norfolk Southern Railway since He is the manager of Brosnan Forest, a 7,000-ha corporate conference and outdoor recreation facility of Norfolk Southern located in Dorchester, South Carolina. He received his B.S. in wildlife management from Oklahoma State University. Jordan M. Smith is an undergraduate student at Auburn University and will receive his B.S. in wildlife science in December He has assisted with numerous research projects on white-tailed deer and elk and plans to continue his education by obtaining a M.S. degree working with large ungulates. Bret A. Collier is a post-doctoral research associate in the Department of Wildlife and Fisheries Sciences at Texas A&M University. He received an M.S. in statistics and Ph.D. in biology with the Arkansas Cooperative Fish and Wildlife Research Unit from the University of Arkansas, M.S. from Oklahoma State University in resource economics, and B.S. from Eastern Illinois University. His research interests include statistical modeling, parameter estimation, capture recapture, and Bayesian approaches to modeling wildlife populations.