WOLF (Canis lycaon) PUP SURVIVAL, DISPERSAL, AND MOVEMENTS IN ALGONQUIN PROVINCIAL PARK, ONTARIO

A Thesis Submitted to the Committee on Graduate Studies in Partial Fulfillment of the Requirements for the Degree of Master of Science with a Major in Ecology and Ecosystem Processes

TRENT UNIVERSITY Peterborough, Ontario, Canada Copyright by Kenneth J. Mills 2006 Watershed Ecosystems Graduate Program June 2006

ABSTRACT Wolf (Canis lycaon) pup survival, dispersal, and movements in Algonquin Provincial Park, Ontario, Canada Kenneth J. Mills

I measured monthly survival and dispersal rates for 73 juvenile (age 3.5 31 weeks) wolves (Canis lycaon) in Algonquin Provincial Park using intraperitoneal VHF transmitters, radio collars and VHF ear tag transmitters. Intraperitoneal radio transmitters were surgically implanted in 53 wolf pups at den sites using inhalation anesthesia with sevoflurane. Mean parturition date was April 29 and mean age at capture was 5.5 weeks. All surgeries were completed without complications, and post-operative infections or complications were identified. Pup survival was high (0.954 per month) and constant during the study period and most mortality was from natural causes. Pups dispersed as early as August and dispersal rates were higher than those inferred previously for juvenile wolves (0.033 per month). Pup movements and home site use reflected increased independence as summer progressed.

Key words: Algonquin Provincial Park, Canis lycaon, dispersal, home site, intraperitoneal transmitter, juvenile, mortality, movements, pup, sevoflurane, surgery, survival, wolves

ii

AUTHORIZATION TO SUBMIT THESIS

Co-Supervisor

Date

Co-Supervisor

Date

Committee Member

Date

Internal Examiner

Date

External Examiner

Date

iii

ACKNOWLEDGEMENTS I am indebted to many who assisted me throughout the course of my graduate research. I am most grateful to my wife, Jaime, for the patience and encouragement she provided during our tenure at Trent University and Algonquin Provincial Park. I thank Brent Patterson and Dennis Murray, my primary advisors, whose support and input were essential and greatly appreciated throughout the course of the research. Each also contributed to written portions of the manuscript via revisions and suggestions on analyses. My parents, Jim and Diane, in-laws Tom and Janeen, and my brothers and sisters were very supportive throughout my education. I thank Paul and Suzy Shalla and Richard and Patsy Shalla for their consistent friendship, advice, expertise and willingness to always help in any situation. I am also indebted to Amelia Argue and Jessica Condon, who dedicated much of their free time during two summers to capturing and monitoring pups. Karen Loveless was also instrumental in pup captures. Joe Marquardson, Derek Meier, Kevin Middel, Marie-Helene OuelletDAmours, Lauren Patterson, Derek Potter, Brad Steinberg, Sue Tully and Dan Votour also provided field assistance. Special thanks go to Mark Mills for traveling to Algonquin to assist in finding dens and capturing pups. This research would not have been possible without the assistance of many veterinarians who donated their time, talents, and resources for surgical procedures. Dr. Graham Crawshaw of the Toronto Zoo assisted in developing the procedure and donated many weeks of his free time conducting surgical procedures in Algonquin. Dr. Crawshaw also provided much of the material and expertise included in Chapter

iv

2, including the written surgical protocol and the evaluation of the surgery and its effectiveness. Dr. Craig Mosley of the University of Guelph and Oregon State University designed and built the anesthesia machine used in all surgeries, volunteered much of his time in developing surgical protocols and traveled from Oregon to assist in the field portions of the research. Dr. Mosley also contributed to the implant manuscript by adding constructive revisions and expertise. I also thank Rhonda Benke, and Drs. Amanda Dame, David McLelland, and Rolf Olberg for donating their time to travel to Algonquin to perform some of the implantations. I thank Dr. Jason McCloud and the staff of the Algonquin Animal Hospital for preparation and sterilization of the equipment. Dr. Dale Smith of the University of Guelph also assisted in developing the surgical protocols. I also thank Sandpiper Industries for a timely equipment grant and Scent-Lok for donating field clothing used during trapping efforts for wolf adults and pups. The staff at Algonquin Provincial Park, the Friends of Algonquin Park, and Harkness Laboratory of Fisheries Research provided timely assistance, housing and resources throughout the study. Thanks go to Doug Brown, Kevin Clute, Peg Darraugh, Kathy Pordonick, Gary Ridout, Richard Szczygiel, Marie Shalla, Rick Stronks, John Swick, John Winters, and Glen Yanke. This research was supported by grants from Algonquin Provincial Park, Algonquin Forest Authority, Friends of Algonquin Park and the Ontario Ministry of Natural Resources Wildlife Research and Development Section.

vi

TABLE OF CONTENTS Page TITLE PAGE ABSTRACT AUTHORIZATION TO SUBMIT THESIS ACKNOWLEDGEMENTS LIST OF FIGURES i ii iii iv x

CHAPTER 1: Introduction

CHAPTER 2: Field implantation of intraperitoneal radio transmitters for Eastern wolf (Canis lycaon) pups using inhalation anesthesia with sevoflurane Abstract Introduction Methods Results Discussion

5 6 8 11 13

CHAPTER 3: Wolf pup survival, dispersal, and movements in Algonquin Provincial Park, Ontario, Canada Abstract Introduction

17

18 19

vii

Methods Study area Pup capture Survival analysis Dispersal analysis Movement and home site use Results Pup capture Survival patterns Dispersal patterns Movement and home site use Discussion

20 20 20 22 23 23 24 24 25 26 26 27

CHAPTER 4: Conclusion

34

LITERATURE CITED

37

APPENDIX: Effects of GPS collar sampling frequency on estimated movement distance and home range size of wolves Abstract Introduction Methods Study area

44

45 46 48 48

viii

GPS transmitter efficiency Wolf movements Territory delineation Results GPS transmitter efficiency Wolf movements Territory delineation Discussion Literature Cited

49 49 51 52 52 53 53 54 58

ix

LIST OF FIGURES CHAPTER 2 Figure 2.1 15

A detailed view of an intraperitoneal VHF radio transmitter (7.9 cm length x 1.7 cm diameter; ATS, Isanti, MN, U.S.A.) prior to insertion into the peritoneal cavity of a 5week old wolf pup in Algonquin Provincial Park. The implant consists of a lithium battery (right), circuit board (middle) and a coiled antenna (left) packaged in a biologically inert resin. Figure 2.2 Surgical setup for field implantation of intraperitoneal VHF transmitters into wolf pups using sevoflurane at den sites in Algonquin Provincial Park. The setup includes an oxygen cylinder with a dial up combination regulator/flow meter attached to the table leg and a sevoflurane vaporizer placed on the table in a Bains anesthetic circuit with the respective hosing and tight-fitting anesthetic face mask. 16

CHAPTER 3 Figure 3.1 Nelson-Aalen cumulative hazard function ( 95% CI) for wolf pups in Algonquin Provincial Park (2002 05). Cumulative hazard was calculated using all observed mortalities excluding two pups that died from capture related disturbance. Figure 3.2 Relationship between pup age separated into four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) and mean ( SE) 31 30

duration of home site occupancy for wolf pups implanted in Algonquin Provincial Park. Figure 3.3 Relationship between pup age separated into four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) and mean ( SE) distance moved between home sites or relocations for wolf pups implanted in Algonquin Provincial Park. Figure 3.4 Relationship between pup age separated into four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) and litter cohesion expressed as the percentage of the litter that was together, for wolf pups implanted in Algonquin Provincial Park. 33 32

APPENDIX Figure A.1 Estimated transmitter battery duration (mean SE) and the number of successful locations (mean SE) obtainable for Lotek model 4000 GPS collars using fix intervals of increasing length (n = 3 collars). Figure A.2 Logarithmic model for the relationship between fix interval (FI) and distance ratio (DR = distance measured for the x-hr FI subsample / distance measured for the 5-min FI subsample) for wolf movements measured using GPS transmitters on 5-min FIs for 64 63

xi

24 hour sampling periods (n = 13 movement paths). The logarithmic fit was y = 0.672+ -0.127 ln(x). Figure A.3 65

Relationship between distance ratio and the fractal dimension (D) for wolf movement paths measured using GPS collars in Algonquin Provincial Park (n = 33 movement paths). The fractal dimension is calculated as: D = log(n) / (log(n) + log(d/L)), where n = no. of steps in the movement path, d = maximum diameter of the movement path, and L = measured length of the movement path. Figure A.4 66

Logarithmic model for to the relationship between fix interval and area ratio (AR = area measured for the x-day FI / area measured for the .25-day FI) for (a) 100% Minimum Convex Polygon (n = 8 wolf territories) and (b) adaptive kernel territory size (n = 8 wolf territories) measured using GPS collars in Algonquin Provincial Park. The logarithmic fit for the fix interval and MCP area ratio was y = 0.800 + 0.175 ln(x). Figure A.5 Relationship between number of GPS telemetry locations (n = 309 locations) and percent total area of a minimum convex polygon territory for a collared wolf (W129) in Algonquin Provincial Park. 67

xii

CHAPTER 1 Introduction Juvenile recruitment, as determined by the production and survival of offspring, is an important component of the dynamics of most vertebrate populations (Peterson 1977, DelGiudice et al. 1991, Fuller et al. 2003). Therefore, the ability to accurately measure juvenile demography is essential to understanding overall population change and assessing population condition (Van Ballenberghe and Mech 1975, Mech 1977, Kerley et al. 2003, Owen-Smith et al. 2005). However, demographic features of juvenile carnivores, specifically wolves, are difficult to measure, primarily because of concerns regarding disturbance at home sites and potential abandonment of handled juveniles (Chapman 1977, Mech 1977, Laurenson and Caro 1994, Mowat et al. 1996, Fernandez et al. 2002, Echols et al. 2004). Monitoring techniques that are applicable for juvenile carnivores are also lacking, primarily because current techniques require external radiocollars that are excessively large and do not allow for subsequent growth. Despite intense research focused on wolves, previous methods developed for assessing pup survival are generally qualitative and potentially biased (Rausch 1967, Van Ballenberghe and Mech 1975, Mech 1977, Harrington et al. 1983, Fuller 1989a, Theberge and Theberge 2004). These qualitative techniques also fail to cover the early juvenile period from birth to 5 months (Mech 1977, Fuller et al. 2003). Mortality factors for pups during this period are essentially unknown, which is concerning since pups of this age range appear particularly susceptible to natural mortality and disease (Mech 1977, Johnson et al. 1994, Mech and Goyal 1993, Mech

and Goyal 1995, Fuller et al. 2003). Therefore, research focused on wolf pup demography from birth to 5 months should be considered a priority. In Algonquin Provincial Park (APP), research conducted from 1987 to 1999 suggested that pup survival was low and that recruitment was not adequate to maintain wolf numbers within APP (Theberge and Theberge 2004). This, in addition to high levels of human-caused mortality outside the Park, prompted a wolf harvest moratorium in all townships surrounding the APP which was implemented in December 2001. The initiation of this moratorium required intense monitoring of wolf dynamics within APP in order to determine the condition and viability of the wolf population and the effectiveness of the moratorium. To meet these requirements, it was determined that a focus must be directly placed on measuring wolf pup production, survival and recruitment in addition to adult demography. Because of this research focus and the lack of data regarding juvenile survival, I developed, with assistance of zoo and wildlife veterinarians, a technique that could be used for monitoring individual juvenile carnivores such as wolf pups. This technique involved the implantation of an intraperitoneal very high frequency (VHF) transmitter using gas anesthetics (CHAPTER 2). The a-priori requirements of the protocol were as follows, 1) it must be safe for the pups and incur minimal short and long-term pain and discomfort, 2) it must be applicable for 4 6-week-old wolf pups, 3) it must have a minimal effect on subsequent pup survival and 4) it must be performed in the field in order to minimize handling time. I describe the surgical procedures and the specific equipment that was required to perform the implant

surgeries in the field. I also report the viability of the surgical protocol and the results of the surgical procedures in relation to these requirements. The development of a successful radio transmitter implanting technique allowed for the close monitoring of individual wolf pups in APP and the quantitative evaluation of pup survival, dispersal and movements during the period from birth through autumn, something that had not previously been possible. I report on the survival, dispersal and cause-specific mortality rates of wolf pups from June, when pups were 3.5 8 weeks old, to the end of November and examine the effects of social and ecological factors (e.g., litter size, pack size, and sex) that may affect pup survival and dispersal (CHAPTER 3). I also measured the duration of home site use, movement distance between home sites and litter cohesion, which could not be directly assessed with previous methods. In addition, I examined social and ecological factors and their effects on these aspects of home site use and pup movements.

CHAPTER 2

FIELD IMPLANTATION OF INTRAPERITONEAL RADIO TRANSMITTERS FOR EASTERN WOLF (Canis lycaon) PUPS USING INHALATION ANESTHESIA WITH SEVOFLURANE

Abstract In a study of wolf pup mortality, intraperitoneal radio transmitters were surgically implanted in 53 (27 male and 26 female) 3.5 to 8-wk-old Eastern wolf (Canis lycaon) pups at den sites in Algonquin Provincial Park, Ontario, Canada over two breeding seasons (2004 and 2005). Pups were manually removed from dens and initially injected with butorphanol at a dosage of 0.2 mg/kg for sedation and interoperative analgesia. Anesthesia was induced and maintained with 3% sevoflurane in oxygen via a face mask. Meloxicam (0.3 mg/kg IM) was given to provide additional analgesia. All surgeries were completed without complications and pups were readily accepted back into the packs. No post-operative complications were identified but 2 pups from a single litter drowned as a result of being moved by pack members to a flooded den following the surgery. For 6 pups necropsied following natural deaths, transmitters were found lying free within the abdominal cavity, and in one case the transmitter was found in the pelvic cavity, with no evidence of blockage or inflammation within the cavity or at the surgical site. Inhalation anesthesia provided extremely rapid induction (1 min) and recovery (<3 min) and was completely controllable with no residual anesthetic effects. The equipment for inhalation anesthesia was readily portable in field packs, and has considerable advantages over injectable drugs for small and very young animals such as wolf pups. The utility of the procedure is demonstrated by the minimal effect it had on subsequent pup survival, the rapid recovery of pups following surgery, and the lack of long term complications as determined by subsequent necropsies of pups following natural deaths.

Introduction There have been few studies into denning ecology and pup survival of wolves (Pimlott et al. 1969, Fritts and Mech 1981, Ballard and Dau 1983, Fuller 1989b, Ciucci and Mech 1992, Boyd et al. 1993). This likely reflects concern over the negative effects of den disturbance and the difficulty in marking and monitoring juvenile carnivores (Chapman 1977, Mech 1977, Laurenson and Caro 1994, Mowat et al. 1996, Fernandez et al. 2002, Echols et al. 2004). Investigations of wolf pup (Canis lycaon) survival in Algonquin Provincial Park (APP), Ontario were initiated in 2004 as part of a broader study examining the population dynamics of Park wolves (Patterson et al. 2004). The study required that pups be individually marked with very high frequency (VHF) transmitters in early summer to allow for monitoring. In previous years, wolf pups in APP had been fit with padded VHF radiocollars or VHF ear tag transmitters, but these tended to fail prematurely or were removed by adults (B.R. Patterson et al., unpublished data). Early monitoring is essential for determining specific survival rates and causespecific mortality during the summer. This is particularly important for populations that are prone to diseases, such as canine parvovirus, that can be a significant mortality factor for wolf pups less than three months of age (Meunier et al. 1981, Johnson et al. 1994, Mech and Goyal 1993, Mech and Goyal 1995). To effectively monitor pups during this early juvenile period, anesthetic and surgical techniques were developed to implant VHF transmitters in pups during the time that they could still be easily captured in their dens.

The use of implantable VHF transmitters in wildlife is increasing and has been employed principally in species where behavior and/or body morphology restrict the use of external transmitters (e.g., Green et al. 1985, Rosatte and Kelly-Ward 1988, Echols et al. 2004), but may be particularly applicable for juveniles where external attachment of transmitters is problematic due to body growth. However, only a few studies highlight the use of implantable transmitters in either canids or juvenile animals, all of which were conducted in lab settings which required that individuals be captured and transported to a surgical facility (Green et al. 1983, Rosatte and Kelly-Ward 1988, Way 2000). These techniques are not acceptable for wolf pups because it requires them to be separated from the pack for extensive periods, which might cause abandonment and subsequent pup mortality (Chapman 1977, Mech 1977, Laurenson and Caro 1994, Mowat et al. 1996, Fernandez et al. 2002). Most surgical protocols in wildlife utilize injectable anesthetics with slow induction and recovery periods which may be unacceptable for use in juveniles because of lengthened disturbance time at the den site, and increased risk of rejection. Because wolf pups are usually moved to new den or rendezvous sites immediately following human disturbance (Joslin 1967, Chapman 1977; see Chapter 3), it is important for pups to be fully recovered from the surgical procedure before relocation. Inhalant anesthetics provide smooth and rapid induction and recovery (Johnson et al. 1998) but the standard equipment is generally not conducive for field studies, especially where manual transportation to a remote surgical site is required. Herein, I report an effective surgical protocol for implanting peritoneal VHF

transmitters in wolf pups that can be easily performed in the field using inhalation anesthesia.

Methods Wolf den sites in APP (45oN, 78oW) were located via monitoring of radiocollared adults in spring 2004 and 2005. Wolves began showing signs of affinity for den locations during the last week of April and first week of May in each year. Dens were located by searching the area of localization on foot using 1 to 8 field personnel between May 28th and June 16th in each year, roughly five weeks after animals appeared to localize at a suspected den site. Once the den was found, all pups were captured manually and placed individually in breathable cloth bags in which they were weighed and restrained until surgery. All pups were handled with vinyl gloves to reduce residual scent transmission throughout the handling procedures. Transmitters (M1235, Advanced Telemetry Systems, Isanti, Minnesota, U.S.A.: 7.9 cm length x 1.7 cm diameter; Figure 2.1; or model IMP/300M, Telonics, Inc., Mesa, Arizona, U.S.A.: 8.1 cm length x 2.3 cm diameter) were individually packaged and previously sterilized with vaporized hydrogen peroxide (Sterrad, Advanced Sterilization Products, Irvine, California, U.S.A.). Multiple sets of basic surgical instrument and drapes were packaged individually and sterilized by autoclaving. Surgery was performed at the den site on a lightweight portable camping table, and the surgical suite and anesthetic apparatus were enclosed in a mosquito net suspended from a suitable tree (Figure 2.2). The oxygen cylinder was affixed to the leg of the table and the vaporizer was placed on the table.

Up to 30 minutes prior to the surgical procedure, pups were removed individually from the bags and given an intramuscular (IM) injection of butorphanol (Torbugesic, Wyeth Animal Health, Guelph, Ontario, Canada) at 0.1 mg/kg for sedation and intraoperative analgesia. Anesthesia was then induced with 8% sevoflurane (Sevoflurane, Abbott Laboratories, St. Laurent, Quebec, Canada) in oxygen with a flow rate of 2-3 L/min, via a tight-fitting anesthetic face mask. The anesthetic apparatus consisted of a precision sevoflurane vaporizer (Penlon Sigma Delta. Penlon Ltd., Abingdon, Oxon, UK), a D or E size aluminum oxygen cylinder, a dial-up combination regulator/flowmeter (Model 3108-L, Inovo Inc, Naples, Florida, U.S.A.), a Bains anesthetic circuit, and a 2 m hose for waste gas. Total weight of the apparatus including the aluminum oxygen cylinder was 11.4 kg with the E cylinder, and 9.7 kg with the smaller D size cylinder. Once the pups were relaxed and at a surgical plane of anesthesia, the sevoflurane concentration was reduced to 2-3% and the oxygen flow to 1-2 L/min. Anesthetic depth was further adjusted based on the assessment of reflex responses, heart rate and respiratory rate which were monitored continuously throughout the anesthesia. The ventral abdomen was clipped with battery operated clippers and prepared for aseptic surgery with 4% chlorhexidine surgical scrub (Vet Solutions Surgical Scrub and Handwash, Vet Solutions, Fort Worth, Texas, U.S.A.) followed isopropyl alcohol, which was repeated once, and finally with the application of povidone iodine (Betadine, Purdue Pharma, Pickering, Ontario, Canada). A surgical drape was then applied. A 2 cm incision was made through the skin and ventral abdominal wall approximately 1 cm caudal to the umbilicus, and the

transmitter was placed loosely into the abdominal cavity. The abdominal wall was sutured in a simple interrupted pattern with 4/0 polydioxanone (PDS, Ethicon, Johnson and Johnson, Piscataway, New Jersey, U.S.A.) followed by subcuticular suture with the same material, and then by the external application of nbutylcyanoacrylate surgical adhesive (VetBond, 3M, London, Ontario, Canada) to the skin wound. Upon completion of the surgery, the vaporizer was turned off, the mask removed after one minute, and the pups were held during the recovery period. A transponder (PIT tag; Avid Canada, Calgary, Alberta, Canada) was inserted subcutaneously (SC) on the dorsal midline between the scapulas, and the wound sealed with surgical glue. Each pup received an injection of meloxicam (Metacam, Boehringer, Burlington, Ontario, Canada) at 0.2-0.3 mg/kg SC for longer-lasting post-surgical analgesia, and the antibiotic cefazolin (Cefazolin, Novopharm, Toronto, Ontario, Canada) at 25 mg/kg SC for the prevention of post-operative infection. A blood sample was taken from each pup for complete blood count and pathogen serology, and hair samples were collected for genetic and nutritional studies. Implanted transmitters were not removed at the end of the study as the stress induced by capture and surgical removal of the transmitters would be excessive. In previous studies, transmitters were left in place without any detectable complications (Davis et al. 1984, Reid et al. 1986). All protocols for both field seasons were approved by the Ontario Ministry of Natural Resources Wildlife Animal Care Committee and the Trent University Animal Care Committee.

10

Results Twenty pups from five different litters were captured and implanted in 2004, and 33 pups from eight litters were implanted in 2005. Mean age at capture for both years was 5.5 0.4 weeks (SE; range: 3.5 8 weeks) and mean weight was 2.9 0.1 kg (range: 1.45 5.1 kg). All pups were easily removed from the den by hand and were calm enough to handle manually without additional restraint. Actual surgical time averaged 15 minutes, and total handling time including surgical preparation, PIT tag injection, treatment, blood and hair sampling and recovery averaged 35 minutes. Induction of anesthesia to a surgical level with sevoflurane averaged just 1.4 min (range: 1-2 min). No pups showed signs of physical stress or had difficulty breathing throughout the procedure. All pups were awake within four minutes of removal of the face mask and were ambulatory and able to walk back into the den within 10 minutes following surgery without apparent sedative effect or ataxia. Due to initial technical problems with the oxygen cylinder, the first pup handled was anesthetized with ketamine (Vetalar, Bioniche Animal Health, Belleville, Ontario, Canada) at 25 mg/kg combined with midazolam (Versed, Hoffman-La Roche, Mississauga, Ontario, Canada) at 0.25 mg/kg IM. Relaxation and anesthesia occurred within 3 minutes. However, the addition of butorphanol (0.1 mg/kg IM) was required to reduce response to the surgical procedure and the pup also received some sevoflurane by mask during the procedure. The other fifty-two pups received butorphanol and sevoflurane only for the anesthetic protocol. There were no post-operative complications or mortalities associated with surgical procedures. However, two 3.5-week old pups from the same litter died

11

following capture after they were relocated by pack members to a flooded den and subsequently drowned. Post-mortem necropsies confirmed that there was no infection, inflammation or peritonitis associated with either the surgical incision or the implanted transmitter for these pups. There was also no sign of complications related to the implant among four additional pups necropsied after naturally-caused deaths occurring 50 - 134 days after surgery. Transmitters were found floating freely in the abdominal cavity in three of these pups and floating freely in the pelvic cavity of the fourth. No pups were abandoned after handling and all packs returned to the den site or relocated the pups to new dens or rendezvous sites within a few hours of the team leaving the site. Most pups (~75%) were known to be alive and the transmitters were still functioning in November, approximately six months after deployment. This excludes nine transmitters implanted in pups that died prior to the end of November and three ATS transmitters that malfunctioned and became inoperable prematurely (28 October, 3 November, and 23 November). After February signals tended to weaken and contact was lost with many pups. Effective reception range for aerial telemetry was 1.5 - 10.3 km depending on flight altitude and a combination of factors such as the general orientation to the fixed H-antenna mounted on the aircrafts wing strut and the topographic features surrounding the pups location. The range for ground telemetry was approximately 0.5 - 1.1 km and was strongly affected by the topography between the transmitter and receiver.

12

Discussion This protocol is the first to describe a surgical procedure using inhalation anesthesia equipment that was designed to be used in remote locations. The advantages of inhalation anesthesia are the rapid and smooth induction and recovery, the high degree of controllability without the need for additional injectable anesthetic if the procedure is prolonged, and minimal residual drug effects once the anesthetic is discontinued. Inhalants produce classical anesthesia with dose-dependent unconsciousness and good muscle relaxation compared with the injectable dissociative anesthetics, such as tiletamine, ketamine, or ketamine combinations. Ketamine alone provides inadequate relaxation for surgery and anesthesia is shortlived. The addition of alpha-2 agonists, such as xylazine and medetomidine, to ketamine increase the depressant effects on cardiovascular function, and cause potentially longer and rougher induction and recovery periods. As a general rule, the administration of alpha-2 agonists to pediatric animals is not advised (Lemke 2004). Rapid return of the pups normal behavior without sedation or other side-effects is particularly important to reduce the risk of maternal rejection or aggression, and acceptance by siblings. Overall, this protocol was designed to minimize handling times and stress while providing the highest level of anesthetic care and a surgical technique that was possible in a field situation. The safety of the protocol is demonstrated by the fact that there were no postoperative mortalities due directly to surgical procedures and no pups abandoned by adults following capture. The two pups that died soon after capture drowned after the parents moved them to a flooded den. The age of these pups at the time of capture

13

suggest that captures should be focused on pups >3.5 weeks old when they would be more mobile and independent (Packard 2003). No other pup mortalities occurred within 30 days after surgery in either year. All pups necropsied lacked inflammation or infection in the abdominal cavity or other adverse effects related to surgical procedures or the implanted transmitter. The low effective range of the transmitters hindered my ability to relocate individuals that ranged widely (Koehler et al. 2001), such as pups that had dispersed, but was highly effective when pups were located within their natal territory. Despite the range limitations, implanted transmitters are the only described method that can be used to effectively monitor wolves during the early juvenile period. If further monitoring is required, implanted pups could be recaptured and fitted with VHF collars or ear tags in autumn to monitor survival and dispersal during the remainder of the year. Overall, I found the anesthetic and surgical techniques to be safe and effective for VHF transmitter deployment in juvenile wolves in a field setting. This protocol may also be appropriate for implanting other mammals with transmitters where injectable techniques and/or external transmitters are either undesirable or ineffective. The required surgical equipment was extremely effective for the required procedures, was easy to use and could be carried in backpacks into the den site or other remote locations by two or more field personnel.

14

Figure 2.1. A detailed view of an intraperitoneal VHF radio transmitter (7.9 cm length x 1.7 cm diameter; ATS, Isanti, MN, U.S.A.) prior to insertion into the peritoneal cavity of a 5-week old wolf pup in Algonquin Provincial Park. The implant consisted of a lithium battery (right), circuit board (middle) and a coiled antenna (left) packaged in a biologically inert resin.

15

Figure 2.2. Surgical setup for field implantation of intraperitoneal VHF transmitters into wolf pups using sevoflurane at den sites in Algonquin Provincial Park. The setup included an oxygen cylinder with a dial up combination regulator/flow meter attached to the table leg and a sevoflurane vaporizer placed on the table in a Bains anesthetic circuit with the respective hosing and tight-fitting anesthetic face mask.

16

CHAPTER 3

WOLF PUP SURVIVAL, DISPERSAL, AND MOVEMENTS IN ALGONQUIN PROVINCIAL PARK

17

Abstract Accurate determination of juvenile survival and recruitment rates is essential for assessment of status and viability of animal populations. Currently, the demographic attributes of juvenile carnivores, specifically wolves, are poorly known but of considerable conservation interest. I measured survival and dispersal rates for 73 juvenile (age 3.5 31 weeks) wolves (Canis lycaon) in Algonquin Provincial Park using implantable VHF transmitters, padded VHF radiocollars, and VHF ear tag radio transmitters. Pups moved longer distances between home sites, spent less time at each home site, and spent more time apart, as they aged. Monthly pup survival was relatively high (0.954; 95% CI: 0.931, 0.978) and constant during the June November period. In APP, most pup mortality was from natural causes (n = 9), such as accidents, conspecific strife, or predation by black bears (Ursus americanus), with few deaths (n = 2) from anthropogenic causes like shooting and poisoning. Some pups dispersed as early as August (age 15 weeks) and monthly dispersal rates were surprisingly high for young pups (0.033; 95% CI: 0.013, 0.053). The natal pack to which pups belonged affected the probability of survival but I failed to detect any influence of pack size or litter size on pup survival or dispersal, although our sample size precluded robust analysis.

18

Introduction Basic demographic attributes such as survival and dispersal are difficult to measure in juvenile mammals (Echols et al. 2004), and therefore in many cases their contribution to population dynamics remains unclear. Yet, in theory juvenile survival and dispersal should play an important role in determining population trends and viability (DelGiudice et al. 1991, Fuller et al. 2003). Theory also suggests that carnivore populations should have high juvenile survival, particularly among social species where helping from non-breeders may be demographically beneficial (Moehlman 1979, Harrington et al. 1983, Jennions and Macdonald 1994, Russell et al. 2002). However, these conclusions are largely speculative because little quantitative data exist regarding juvenile survival and dispersal patterns prior to the age where individuals can safely be marked with radio-collars (Fuller et al. 2003). This shortcoming emphasizes the need for more focused research on the demographic attributes of juvenile carnivores. Because wolves are social carnivores, survival rates of juveniles should be relatively high when resources are not limiting (Harrington et al. 1983, Heard and Williams 1992, Fuller et al. 2003). Most studies of pup survival are based on passive monitoring techniques including litter counts across time, monitoring radio-collared animals from late summer on, or conducting howling surveys (e.g., Rausch 1967, Van Ballenberghe and Mech 1975, Mech 1977, Harrington and Mech 1982, Fuller 1989a, Theberge and Theberge 2004). These methods only provide estimates of apparent survival rates, which are potentially incomplete or biased because timelines are truncated and individuals can easily be lost to follow-up. It follows that accurate

19

mortality and dispersal rate assessment for juvenile wolves requires a more focused and comprehensive approach. I investigated survival and dispersal among juvenile wolves (Canis lycaon) in APP, using intraperitoneal radio-transmitters surgically-implanted in 4 6-week-old wolf pups. I predicted that pre-winter survival rates would be low based on previous surveys conducted in APP (Theberge and Theberge 2004). I also predicted that dispersal rates would be low and pup movements would increase in length and frequency as summer progressed.

Methods Study Area Algonquin Provincial Park lies at the southern edge of the Canadian Shield in south-central Ontario, Canada (45N, 78W; 7571 km2; Patterson et al. 2004). Densities and basic dynamics of wolves within the park have been relatively consistent at least since 2004 when we initiated an intensive study of the adult wolf population (B. Patterson et al., unpubl. data). Primary prey species in the park included moose (Alces alces), white-tailed deer (Odocoileus virginianus), and beaver (Castor canadensis; Pimlott et al. 1969). Den and rendezvous sites used by wolves were generally located in conifer-dominated habitats near permanent water sources (Joslin 1967, Pimlott et al. 1969, Norris et al. 2002). Pup Capture I identified potential wolf den sites in spring via aerial and ground tracking of radio-collared adults (2004; n = 11 packs, 2005; n = 15 packs; Patterson et al. 2004).

20

In order to schedule wolf pup captures appropriately, we visited den sites 4 6 weeks after the date that collared adult wolves first localized at potential dens. These potential den sites were searched on foot by 1 8 field personnel and pups were captured manually from the den. Captured pups were weighed, sexed, and implanted with a VHF radio-transmitter (2 x 8 cm, Advanced Telemetry Systems, Isanti, MN or Telonics, Inc., Mesa, AZ) in the peritoneal cavity (Crawshaw et al., in prep). Each transmitter was equipped with either a motion-sensitive mortality switch with 4-hour delay (n = 38), or a temperature-sensitive mortality switch which was triggered if body temperature fell below 32C (n = 15). I estimated the age of captured pups using: 1) original localization dates of collared pack members, and 2) pup size and degree of tooth eruption at the time of capture (Rutter and Pimlott 1968, Mech 1970, Van Ballenberghe and Mech 1975). Final age of pups was determined using the latter criterion, but both measures were found to provide comparable estimates of parturition date (paired-t: t12 = 0.73, P = 0.48), with the mean difference between both metrics being 0.8 0.2 weeks (SE; range: 0.1 2.0 weeks). Pups were returned to the den when they became ambulatory following surgery, and in the hours after all pups were processed and the field team had left the site den sites were monitored to ensure that pups were not abandoned by adults. I also captured pups opportunistically from mid-July through late October 2002 2005 using Victor Softcatch #3 and #1.5 foothold traps placed in or near rendezvous sites or along logging roads. Each captured pup was sexed, weighed, and radio-tagged with a VHF ear tag transmitter (Advanced Telemetry Systems, Isanti,

21

MN) or padded VHF collar (Model LMRT3, Lotek Engineering, Inc., Newmarket, ON). I monitored transmittered pups for survival via ground or aerial telemetry every 2 3 days after capture until the end of August, and weekly thereafter. The only exceptions to this monitoring protocol occurred when packs shifted home sites or when telemetry flights were canceled due to poor flying conditions. This occurred rarely in summer but was more common after September when weather conditions were variable. Pup cause of death was determined by evidence at the mortality site, and necropsies were conducted by personnel from the Canadian Cooperative Wildlife Health Centre (Guelph, ON). Survival Analysis Survival rate estimates were obtained using the program Micromort using the sum of all radio days for all pups (Heisey and Fuller 1985). Monthly survival rates were calculated through November 30 of each year, with all pups known to be alive on that date being right-censored. I also fit a parametric Weibull survival analysis to test the hypothesis of constant survival rate during the study period. If the dimensionless shape parameter (p) generated by the Weibull model did not differ from 1.0, the risk of mortality was considered to be constant (McCallum 2000). I evaluated the determinants of pup survival using Cox Proportional Hazards regression models. Variables under consideration included the natal pack of each pup, sex of the pup, and the number of adult members in the natal pack (i.e., the maximum number of individuals verified to be with the pack on > 1 occasion during the previous winter). I also included litter size as a time-dependent covariate in

22

survival analyses. Because pup survival within litters might not be independent, we used clustering to identify littermates in all Cox Proportional Hazards analyses in order to achieve robust variance estimation (Lin and Wei 1989). Dispersal Analysis I defined dispersal as emigration from the natal territory without returning for the remainder of the study period (Gese and Mech 1991). I used the program Micromort to estimate residency rates, and dispersal rates were then calculated as 1 residency rates (Heisey and Fuller 1985, Patterson and Messier 2001). In some cases pups disappeared from their natal territory before the anticipated expiry of transmitter batteries, and were not subsequently relocated following extensive searching both on the ground and from the air to a radius exceeding 20 km from the natal territory; I considered such individuals as probable dispersers. CPH models with clustered littermates were used with natal pack, sex, litter size, and pack size as covariates. Movement and Home Site Use I recorded the succession of home sites (i.e. den and rendezvous sites collectively; Pimlott et al. 1969) that each monitored litter used throughout the summer as well as the number of days spent at each home site and the linear distance moved between successive home sites. I also recorded the date that each litter abandoned rendezvous sites and began moving with the pack. Movement characteristics of ageing pups were described by dividing the study period into four intervals: 1) denning (age < 6 weeks), 2) early rendezvous (age 6 11.9 weeks), 3) late rendezvous (age 12 17.9 weeks) and 4) rendezvous abandonment (age 18 weeks). Each interval was then used in a repeated-measures

23

ANOVA to examine the effects of age class, natal pack, litter size, and pack size on the described movement metrics. I quantified litter cohesion, defined as the proportion of the litter that was located together at a given time, via repeated-measures ANOVA using the same age classes and dependent variables as in the movement analyses. I also examined the effects of the same variables on the age that pups abandoned rendezvous sites using simple multiple regression. All movement and home site occupancy data were log transformed prior to analysis to promote normality.

Results Pup Captures I estimated the average birth date of pups in APP to be April 29 1.9 days for 2004 and 2005 (SE; range: April 18 May 11). However, the estimated parturition date for pups born in 2005 (April 27 2.8 days) was earlier than in 2004 (May 4 2.2 days; t11 = 2.37, P = 0.04). I conducted pup capture and implanting from May 29 June 16 during both years. Fifty-three pups from 13 litters (2004: n = 20 pups from 5 litters; 2005: n = 33 pups from 8 litters) were implanted with transmitters. I believe that all pups were implanted for 10 litters, whereas only a subsample of the remaining 3 litters was radio-marked. Mean age of captured pups was estimated as 5.5 0.4 weeks (range: 3.5 8 weeks), mean litter size was 4.9 0.5 pups (range: 2 7 pups), and the sex ratio of marked pups was close to parity (27M:26F). I determined that adults returned to each den site shortly after our departure and relocated the pups to a new den site, usually 0.4 2.0 km away. I found no evidence of pup abandonment

24

following transmitter implant, and only two cases (3.8% of pups handled) where my handling procedure may have resulted in negative impacts to pup survival. These two littermates were the youngest pups captured during the study and drowned after they had been moved by pack-members from the capture site to a flooded den. These two mortality events were censored from further analyses. During trapping efforts in 2002 05, 15 additional pups were radio-collared and 5 were fitted with ear tag radio transmitters in 2003 (fitting date: July 12 October 27). Thus, my total sample of radio-marked pups was 73 individuals from 28 different litters and 16 different packs. Survival Patterns Monthly survival of wolf pups during the study period (all years combined) was 0.954 (95% CI: 0.931, 0.999; 0.715 for the entire 31-week interval). I recorded 14 pup mortalities during the study period; 9 were proximally due to natural causes, 2 were human-caused, and 3 were due to unknown causes. The main causes of natural mortality included conspecific strife (n = 2), drowning (n = 2), black bear predation (n =2), and other causes such as injuries inflicted by ungulates. Human-caused mortalities included poisoning and shooting. Monthly cause-specific mortality rates were 0.030 (95% CI: 0.010, 0.049) for natural deaths, 0.007 (95% CI: 0.000, 0.016) for human-caused deaths, and 0.010 (95% CI: 0.00, 0.021) for unknown causes. The Weibull survival analysis revealed that mortality risk was constant throughout the study period (p = 0.96, 95% CI: 0.59, 1.56, z1 = -0.16, P = 0.871). However, examination of the Nelson-Aalen cumulative hazard function suggests qualitative temporal differences in risk (Figure 3.1). Litter size, pack size, and sex

25

did not significantly affect pup mortality risk (all P 0.11). However, mortality risk differed between packs (z = 0.25, P = 0.03), with packs experiencing pup mortality ranging from 0 50% of the total litter and 4 of 15 litters (with >1 transmittered pup) having multiple pups deaths. I was unable to identify specific factors predisposing pups to higher mortality risk within given packs. Dispersal Patterns A total of 10 pups dispersed during the study period; 4 were confirmed and the remainder were suspected. The estimated monthly dispersal rate was 0.033 (95% CI: 0.013, 0.053) during the study period (0.213 [95% CI: 0.087, 0.322] for the entire 31-week study period). The youngest age that any pup was verified to have dispersed was 15 weeks. Natal pack, sex, litter size, and pack size did not affect dispersal tendencies for pups in APP (all P 0.18). None of the known dispersers were verified to have formed or joined an existing pack prior to November 30. In fact, one disperser was killed by conspecifics in a territory adjacent to its natal territory within days following dispersal, another spent much of the autumn localized near a municipal garbage dump, and the 2 remaining pups were known to have survived past November 30. Movement and Home Site Use The duration of home site occupation decreased, and number of shifts between home sites increased, as pups aged (F3,36 = 163.9, P < 0.01; Figure 3.2). Dens and early rendezvous sites were occupied much longer than rendezvous sites used at the end of summer. Pack size was correlated to the number of days that litters spent at individual home sites (F3,36 = 3.0, P = 0.03) but the effect size was small and

26

probably insignificant (increase of 0.21 days per adult individual in the pack). Natal pack and litter size were not important determinants of home site occupancy (all P 0.19). The distance that litters moved increased as pups aged (F3,36 = 55.2, P < 0.01; Figure 3.3). Litters were moved very short distances, if at all, in their first 6 weeks but moved progressively farther as they grew older, up to a mean of 3.5 km per move when pups were in the late rendezvous age class. The natal pack, litter size, and pack size did not affect movement distance (all P 0.64). Litters were located apart more often during and after abandoning rendezvous sites than when they were younger (F2,9 = 5.7, P = 0.03; Figure 3.4). Litter cohesion was not affected by the natal pack, litter size, or pack size (P 0.38). The 2005 litter cohort abandoned rendezvous sites at a mean age of 17.3 0.7 weeks, which was 3.4 weeks before the 2004 cohort (20.7 0.8 weeks; t7 = -3.46, P = 0.01). The natal pack of the litter and pack size did not affect the age at which litters abandoned rendezvous sites (all P 0.45) but litter size showed a marginally significant inverse relationship with the age when rendezvous sites were abandoned (F1,12 = 4.62, P = 0.07).

Discussion This study is the first to monitor wolf pups continuously from weaning to the beginning of their first winter and thus presents the only direct and unbiased estimates of pup survival and dispersal during the summer and fall periods. Summer pup survival in APP was relatively high and most causes of death were due to natural causes. The only covariate that affected mortality risk was the natal pack of the litter,

27

indicating that packs vary in their ability to raise pups to independence. For example, one pack was verified to produce litters in all summers from 2003 2005 but succeeded in raising only 2 pups to independence during these years, whereas another pack produced pups in all summers during the same period and raised at least 10 pups to independence. The differential survival of pups in different packs might reflect variations in parental care or prey availability within the respective territory (Harrington et al. 1983, Boertje and Stephenson 1992). I detected pups dispersing from natal packs at an earlier age than those recorded previously for wolves (Mech and Boitani 2003). I did not document any significant social, physical or environmental factors that could explain the mechanism behind early juvenile wolf dispersal. Therefore, the detection of early dispersal for pups in APP may merely reflect the benefit of monitoring individual juvenile pups as opposed to the indirect means generally used to assess pup demography in the past. Conversely, given the genetic distinctiveness of eastern wolves and gray wolves (Wilson et al. 2000, Grewal et al. 2004) and the evolutionary relatedness of eastern wolves and coyotes (Wilson et al. 2000), early dispersal among eastern wolves in APP may simply reflect differences in life history. It is understood that juvenile dispersers generally have reduced survival and production (Gese and Mech 1991, Schauster et al. 2002), which is likely the case for dispersing pups in our study. However, additional research is needed to determine the mechanisms promoting early dispersal of wolf pups in APP and the survival and success of pups following dispersal.

28

My monitoring protocol also revealed interesting patterns related to pup movements from the natal denning period through to the abandonment of home sites. As pups aged, I found that the distance moved between home sites increased and the duration of stay at home sites and the cohesiveness of litters decreased, reflecting the increasing mobility and independence of pups. Varying social factors such as pack size, litter size and the natal pack of the litter generally had little effect on pup movements. While these results are generally comparable to several previous studies (Joslin 1967, Pimlott et al. 1969, Fuller 1989b), mine is the first to measure wolf pup movement patterns directly through continuous monitoring. I present the first detailed analyses of wolf pup survival, dispersal and movements from the time of weaning through the beginning of winter. The demonstrated safety of the capture and implanting protocols should alleviate concerns related to den disturbance, and thus hold promise for use in other carnivore populations. Because data regarding juvenile survival and dispersal are essential to accurately describe population dynamics and assess viability (Peterson 1977, DelGiudice et al. 1991, Fuller et al. 2003), these demographic rates are likely to be a necessity in many carnivore conservation and management programs. With added emphasis on the estimation of demographic rates for all cohorts in a population, it will be easier to determine the status and needs of many carnivore populations.

29

Cumulative Hazard

0.00
0

0.05

0.10

0.15

0.20

0.25

50

100

150

200

Pup Age (days)

Figure 3.1. Nelson-Aalen cumulative hazard function ( 95% CI) for wolf pups in Algonquin Provincial Park, Ontario (2002 05). Cumulative hazard was calculated using all observed mortalities excluding two pups that died from capture related disturbance.

30

Figure 3.2. Relationship between mean ( SE) duration of home site occupancy and pup age expressed as four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) for wolf pups in Algonquin Provincial Park.

31

Figure 3.3. Relationship between mean ( SE) distance moved between home sites or relocations and pup age expressed as four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) for wolf pups in Algonquin Provincial Park.

32

Figure 3.4. Relationship between litter cohesion expressed as the percentage of the litter that was together, and pup age expressed as four increasing age classes (denning, early rendezvous, late rendezvous and rendezvous abandonment) for wolf pups in Algonquin Provincial Park.

33

CHAPTER 4 Conclusion The surgical protocol developed for implanting peritoneal VHF transmitters was safe for use in wolf pups, required minimal handling times, and did not cause any cases of abandonment by adults or apparently compromise survival through complications related to surgeries or the implanted transmitters. The equipment required to perform the surgeries could easily be carried to remote locations by field personnel and all surgeries were successfully performed at the den site. The safety and effectiveness of this protocol suggest it could be used to study juvenile specific demography for other carnivore populations to assess more fully overall population dynamics and viability. Pups were born during the last week of April and the first week of May in APP and most pups were captured at 5.5 weeks of age. Production was at a normal level for wolves, with a mean litter size of 4.9 (Pimlott et al. 1969, Mech 1970, Fuller et al. 2003). Movements increased in distance and frequency and litter cohesion decreased as pups grew older, representing increased independence and mobility. Factors such as pack size, litter size, and the natal pack of the litter had little or no obvious effect on pup movement characteristics. Survival was relatively high for transmittered wolf pups, but was within the range of apparent survival rates reported for wolf pups in other populations (Peterson et al. 1984, Fuller 1989a, Pletscher et al. 1997, Mech et al. 1998). Mortality rates were constant during the study period and most mortality was due to natural causes. Pups began dispersing from their natal packs at much younger ages than has been previously recorded (Gese and Mech 1991,

34

Fuller et al. 2003). This was likely, at least in part, a function of methodology and the ability to monitor young dispersers with the use of implanted transmitters. The knowledge of these demographic rates for wolf pups was important in the assessment of overall demography and viability of wolves within APP. I was able to determine that pup production, survival and recruitment were relatively high and that most mortality factors were natural in origin. Despite the success of these monitoring techniques and the full understanding of population dynamics for wolves in APP, there is a need for further development of techniques and research. The VHF transmitter implants used for this research were the largest that could be used in pups as young as 4 weeks but still lacked signal strength and battery longevity. These weaknesses made it difficult to determine if missing pups had dispersed or if their transmitters had failed. Dispersing pups were rarely relocated due to the weak signal strength of implanted transmitters. However, it is possible that some transmitters failed before November, yet we considered this unlikely because transmitter batteries were calculated to last > 10 months. Therefore, most missing pups were considered probable dispersers but their inferred status could have added a larger margin of error to dispersal estimates than if the status of all transmittered pups was precisely known. More intensive monitoring using aerial telemetry would have increased our ability to track dispersing pups because they generally made predispersal movements (i.e., movements within the natal territory independent of other pack members) prior to leaving natal territories that were difficult to detect using ground telemetry alone. The short battery life of implanted transmitters also precluded annual measures for recruitment rates as most transmitters had failed before the pups were 1 year old.

35

More powerful transmitters would alleviate both these shortcomings as it would allow dispersing pups to be located more easily and would allow for continual monitoring of the entire first year of life for pups. More research is also required to determine the ecological factors that affect survival and early dispersal of pups in APP. The focus of future research should also include the assessment of prey densities within APP, which has not yet been determined, because prey availability may significantly affect all aspects of wolf demography (Mech 1977, Harrington et al. 1983, Fuller et al. 2003). This research was the first to monitor young pups from weaning to the end of November and is, therefore, the first to document unbiased survival and dispersal rates for this period. These techniques also allowed the determination of causespecific mortality for individual pups, which aids in the understanding of selective pressures for juvenile wolves. The knowledge of these demographic features, in addition to the assessment of adult cohorts, is essential for the determination of population conditions and viability. The techniques described should be considered for use in other carnivore species and are especially applicable for monitoring juvenile demography characteristics of species of concern.

36

LITERATURE CITED Ballard, W. B., and J. R. Dau. 1983. Characteristics of gray wolf, Canis lupus, den and rendezvous sites in southcentral Alaska. Canadian Field-Naturalist 97: 299-302. Boertje, R. D. and R. O. Stephenson. 1992. Effects of ungulate availability on wolf reproductive potential in Alaska. Canadian Journal of Zoology 70:2441-2443. Boyd, D. K., R. R. Ream, D. H. Pletscher, and M. W. Fairchild. 1993. Variation in denning and parturition dates of a wild gray wolf, Canis lupus, in the Rocky Mountains. Canadian Field-Naturalist. 107: 359-360. Chapman, R. C. 1977. The effects of human disturbance on wolves (Canis lupus l.). MSc. Thesis, University of Alaska. Ciucci, P., and L. D. Mech. 1992. Selection of wolf dens in relation to winter territories in northeastern Minnesota. Journal of Mammalogy 73: 899-905. Davis, J. R., A. F. Von Recum, D. D. Smith, and D.C. Guynn Jr. 1984. Implantable telemetry in beaver. Wildlife Society Bulletin 12:322-324. DelGiudice, G. D., L. D. Mech, and U. S. Seal. 1991. Gray wolf density and its association with weights and hematology of pups from 1970 to 1988. Journal of Wildlife Diseases 27:630-636. Echols, K. M., M. R, Vaughan, and H. D. Moll. 2004. Evaluation of subcutaneous implants for monitoring American black bear cub survival. Ursus 15:172-180. Fernndez, N., F. Palomares, and M. Delibes. 2002. The use of breeding dens and kitten development in the Iberian lynx (Lynx pardinus). Journal of Zoology 258:1-5.

37

Fritts, S. H., and L. D. Mech. 1981. Dynamics, movements, and feeding ecology of a newly protected wolf population in northwestern Minnesota. Wildlife Monographs 80: 79 pp. Fuller, T. K. 1989a. Population dynamics of wolves in north-central Minnesota. Wildlife Monographs 105, 41 pp. Fuller, T.K. 1989b. Denning behavior of wolves in North-central Minnesota. The American Midland Naturalist 121:184-188. Fuller, T. K., L. D. Mech, and J. F. Cochrane. 2003. Wolf population dynamics. Pages 161191 in L. D. Mech and L. Boitani, editors. Wolves: behavior, conservation, and ecology. University of Chicago Press, Chicago, Illinois, USA. Gese, E.M., and L.D. Mech. 1991. Dispersal of wolves (Canis lupus) in northeastern Minnesota, 1969-1989. Canadian Journal of Zoology 69:2946-2955. Green, J.S., R. T. Golightly, S. Lyndaker Lindsey, and B. R. Leamaster. 1985. Use of radio transmitter implants in wild canids. Great Basin Naturalist 45: 567-570. Grewal, S. K., P. J. Wilson, T. K. Kung, K. Shami, M. T. Theberge, J. B. Theberge and B. N. White. 2004. A genetic assessment of the Eastern wolf (Canis lycaon) in Algonquin Provincial Park. Journal of Mammalogy 85:625-632. Harrington, F. H. and L. D. Mech. 1982. Fall and winter homesite use by wolves in northeastern Minnesota. Canadian Field-Naturalist 96:79-84. Harrington, F. H., L. D. Mech, and S. H. Fritts. 1983. Pack size and wolf pup survival: their relationship under varying ecological conditions. Behavioral Ecology and Sociobiology 13:19-26.

38

Heard, D. C. and T. M. Williams. 1992. Distribution of wolf dens on migratory caribou ranges in the Northwest Territories, Canada. Canadian Journal of Zoology 70:1504-1510. Heisey, D. M. and T. K. Fuller. 1985. Evaluation of survival and cause-specific mortality rates using telemetry data. Journal of Wildlife Management 49:668674. Jennions, M. D. and D. W. Macdonald. 1994. Cooperative breeding in mammals. Trends in Ecology and Evolution 9:89-93. Johnson, M. R., D. K. Boyd, and D. H. Pletscher. 1994. Serologic investigations of canine parvovirus and canine distemper in relation to wolf (Canis lupus) pup mortalities. Journal of Wildlife Diseases 30: 270-273. Johnson, R., E. Striler, D. Sawyer, and D. Brunson. 1998. Comparison of isoflurane with sevoflurane for anesthesia induction and recovery in adult dogs. American Journal of Veterinary Research 59: 487-491. Johnson, S. P., T. Gellatt, R. B. Heath, and W. Taylor. 2004. Field inhalation anesthesia in free- ranging juvenile Stellers sea lions (Eumetopias jubatus). In Proceedings of the American Association of Zoo Veterinarians and Wildlife Disease Association Joint Conference, San Diego, pp. 506-508. Joslin, P. W. B. 1967. Movements and home sites of timber wolves in Algonquin Park. American Zoologist 7:279288. Kerley, L. L., J. M. Goodrich, D. G. Miquelle, E. N. Smirnov, H. B. Quigley, and M. G. Hornocker. 2003. Reproductive parameters of wild female Amur (Siberian) tigers (Pantera tigris altaica). Journal of Mammalogy 84:288-298.

39

Koehler, G. M., P. B. Griggs Hall, M. H. Norton, and D. J. Pierce. 2001. Implantversus collar-transmitter use on black bears. Wildlife Society Bulletin 29:600605. Laurenson, M. K. and T. M. Caro. 1994. Monitoring the effects of non-trivial handling in free-living cheetahs. Animal Behaviour 47:547-557. Lemke, K. A. 2004. Perioperative use of selective alpha-2 agonists and antagonists in small animals. Canadian Veterinary Journal. 45: 475-479. Lin, D. Y. and L. J. Wei. 1989. The robust inference for the Cox proportional hazards model. Journal of the American Statistical Association 84:10741079. McCallum, H. 2000. Population parameters: Estimation for ecological models. Blackwell-Science, Oxford University Press, New York, New York, USA. Mech, L. D. 1970. The wolf: the ecology and behavior of an endangered species. University of Minnesota Press, Minneapolis, Minnesota, U.S.A. Mech. L. D. 1977. Productivity, mortality, and population trends of wolves in northeastern Minnesota. Journal of Mammalogy 58:559-574. Mech, L. D., and S. M. Goyal. 1993. Canine parvovirus effect on wolf population change and pup survival. Journal of Wildlife Diseases 29: 330-333. Mech, L. D., and S. M. Goyal. 1995. Effects of canine parvovirus on gray wolves in Minnesota. Journal of Wildlife Management 59: 565-570. Mech, L. D., L. G. Adams, T. J. Meier, J. W. Burch, and B. W. Dale. 1998. The wolves of Denali. University of Minnesota Press, Minneapolis, Minnesota, USA.

40

Mech, L. D. and Boitani. 2003. Wolf social ecology. Pages 134 in L. D. Mech and L. Boitani, editors. Wolves: behavior, conservation, and ecology. University of Chicago Press, Chicago, Illinois, USA. Meunier, P. C., L. T. Glickman, M. J. Appel, and S. J. Shin. 1995. Canine parvovirus in a commercial kennel: epidemiologic and pathologic findings. Cornell Veterinarian 71: 96-110. Moehlman, P. D. 1979. Jackal helpers and pup survival. Nature 277:382-383. Mowat, G., B. G. Slough, and S. Boutin. 1996. Lynx recruitment during a snowshoe hare population peak and decline in southwest Yukon. Journal of Wildlife Management 60:441-452. Norris, D. R., M. T. Theberge, and J. B. Theberge. 2002. Forest composition around wolf (Canis lupus) dens in eastern Algonquin Provincial Park, Ontario. Canadian Journal of Zoology 80:866872. Owen-Smith, N., and D. R. Mason. 2005. Comparative changes in adult vs. juvenile survival affecting population trends of African ungulates. Journal of Animal Ecology 74:762-773. Packard, J. M. 2003. Wolf behavior: reproductive, social, and intelligent. Pages 35-65 in L. D. Mech and L. Boitani, editors. Wolves: behavior, conservation, and ecology. University of Chicago Press, Chicago, Illinois, USA. Patterson, B. R. and F. Messier. 2001. Social organization and space use of coyotes in eastern Canada relative to prey distribution and abundance. Journal of Mammalogy 82:463-477. Patterson, B. R., N. W. S. Quinn, E. F. Becker, and D. B. Meier. 2004. Estimating

41

wolf densities in forested areas using network sampling of tracks in snow. Wildlife Society Bulletin 32:938947. Peterson, R. O. 1977. Wolf ecology and prey relationships on Isle Royale. U. S. National Park Service Scientific Monograph Series, no. 11. Washington D.C., U.S.A. Pimlott, D. H., J. A. Shannon, and G. B. Kolenosky. 1969. The ecology of the timber wolf in Algonquin Provincial Park. Ontario Department of Lands and Forests, Fish and Wildlife Research report No. 87. Pletscher, D. H., R. R. Ream, D. K. Boyd, M. W. Fairchild, and K. E. Kunkel. 1997. Population dynamics of a recolonizing wolf population. Journal of Wildlife Management 61:459-465. Rausch, R. A. 1967. Some aspects of the population ecology of wolves, Alaska. American Zoologist 7:253-265. Reid, D. G., W. E. Melquist, J. D. Woolington, and J. M. Noll. 1986. Reproductive effects of intraperitoneal transmitter implants in river otters. Journal of Wildlife Management 50:92-94. Rosatte, R. C., and P. M. Kelly-Ward. 1988. A surgical procedure for implanting radio transmitters in striped skunks, Mephitis mephitis. The Canadian FieldNaturalist 102: 713-715. Russell, A. F., T. H. Clutton-Brock, P. N. M. Brotherton, L. L. Sharpe, G. M. McIlrath, F. D. Dalerum, E. Z. Cameron, and J. A. Barnard. 2002. Factors affecting pup growth and survival in co-operatively breeding meerkats Suricata suricata. Journal of Animal Ecology 71:700-709.

42

Rutter, R. J. and D. H. Pimlott. 1968. The world of the wolf. J. B. Lippincott Company, Philadelphia, Pennsylvania, U.S.A. Schauster, E. R., E. M. Gese, and A. M. Kitchen. 2002. Population ecology of swift foxes (Vulpes velox) in southeastern Colorado. Canadian Journal of Zoology 80:307-319. Theberge, J. B. and M. T. Theberge. 2004. The wolves of Algonquin Park: a 12 year ecological study. University of Waterloo, Department of Geography publication series; #56. Van Ballenberghe, V., and L.D. Mech. 1975. Weights, growth, and survival of timber wolf pups in Minnesota. Journal of Mammalogy 56:44-63. Way, J.G. 2000. Ecology of Cape Code coyotes. Msc. Thesis, Univ. of Connecticut, 107 pp. Wilson, P. J., S. Grewal, I. D. Lawford, J. N. M. Heal, A. G. Granacki, D. Pennock, J. B. Theberge, M. T. Theberge, D. R. Voigt, W. Waddell, R. E. Chambers, P. C. Paquet, G. Goulet, D. Cluff, and B. N. White. 2000. DNA profiles of the Eastern Canadian wolf and the red wolf provide evidence for a common evolutionary history independent of the gray wolf. Canadian Journal of Zoology 78:2156-2166.

43

APPENDIX

EFFECTS OF GPS COLLAR SAMPLING FREQUENCY ON COLLAR EFFICIENCY AND ESTIMATED MOVEMENT DISTANCE AND HOME RANGE SIZE OF WOLVES

44

Abstract Global Positioning System (GPS) technology has greatly improved the ability to investigate animal ecology at fine spatial and temporal scales, but many GPS telemetry-based investigations have underutilized the technical capabilities of modern transmitters. We estimated wolf (Canis lycaon) movement distance and territory size using variable GPS transmitter sampling frequencies, to evaluate the importance of programming details on estimated movement patterns, territory size, and transmitter performance. Estimated movement distance decreased exponentially as sampling frequency was reduced from one location every 5 min to 24 hrs, implying that intense sampling frequencies were required for accurate measurement of fine-scale wolf movements. Wolf territory size estimates using 95% minimum convex polygons declined with decreasing sampling frequency, whereas estimates using adaptive kernel methods were comparable across a range of sampling intensities. Estimated transmitter battery longevity increased as the interval between fixes was lengthened, but transmitter efficiency, defined as the number of successful fixes obtained during battery lifespan of the transmitter, was highest with fix intervals of intermediate length. We conclude that GPS transmitters are effective for documenting fine-scale movements of animals, but their performance is strongly dependent upon transmitter programming and scale of analysis. Effective use of GPS transmitters will require study-specific assessment of the trade-offs between sampling needs versus transmitter efficiency and longevity.

45

Introduction Integration of Global Positioning System (GPS) technology into wildlife telemetry has greatly improved the potential for investigating animal ecology at fine temporal and spatial scales (Rodgers and Anson 1994, Girard et al. 2002). The resolution of this technology has been tested relative to landscape topography (Gamo et al. 1999, DEon et al. 2002, Frair et al. 2004, Cain et al. 2005), vegetation characteristics (Rempel et al. 1995, Moen et al. 1996, Frair et al. 2004, DeCesare et al. 2005), and animal behavior patterns (Moen et al. 1996, Bowman et al. 2000, DEon and Delparte 2005). Transmitter performance, longevity, and efficiency with respect to GPS location fix interval (i.e., the amount of time between location fixes) have received less attention (Girard et al. 2002, Ppin et al. 2004, Cain et al. 2005). Further investigations into these aspects of GPS telemetry are necessary to better understand transmitter performance under a variety of settings, so that ecological questions regarding home range and space use requirements can be addressed more effectively. Wolves (Canis lycaon) are an ideal species for testing the efficacy of GPS transmitters because they are highly mobile, maintain specific territories, and have a large body size that can handle a relatively large transmitter package. This permits testing the effects of varying transmitter sampling frequencies on estimates of movement distance and territory size, and also allows for assessing sampling optimization from the perspective of transmitter efficiency and longevity. GPS transmitter efficiency (i.e., total no. of successful fixes recorded during the lifespan of a single battery pack) and battery longevity have not been studied empirically in a

46

natural field setting under varying fix intervals. Most tests of transmitter performance have been conducted using stationary radio-collars often deployed in open environments, or else have used inconsistent equipment and environments (e.g., DEon et al. 2002, Cain et al. 2005). Rodgers (2001) suggested that GPS-fix intervals 4 hours are more efficient than longer fix intervals because the GPS receiver does not have to reacquire satellite ephemeris data and can therefore generate a position in less time. Also, although software programs typically provided with GPS transmitters estimate battery longevity according to fix intervals, these programs do not estimate expected acquisition times and fix success rates across a range of fix intervals. Given the lack of independence between location acquisition times and fix interval, existing programs thereby fail to fully assess the efficiency of GPS transmitters in relation to sampling frequency (i.e., fix interval). We predicted that our own assessment of GPS transmitter performance would reveal that transmitter longevity increases with increasing fix interval length, but that fix intervals of intermediate length actually are advantageous from the perspective of transmitter efficiency if the programmed fix acquisition time is an important criterion for determining relocation success. Most recent studies on wolf movements have relied on very high frequency (VHF) telemetry (e.g., Merrill and Mech 2000, Jedrzejewski et al. 2001). One drawback of such technology for wolves and other species is that measured movements are relatively inaccurate (Musiani et al. 1998, Rooney et al. 1998, Kauhala and Tiilikainen 2002, Merrill and Mech 2003). GPS transmitters can be programmed to capture animal locations with sampling frequencies that that are short

47

( 5 min) and accurate enough (error < 15 m/location) to estimate relatively precise movement distances (D'Eon et al. 2002). Previous research suggests that data on movement distance and home range size estimation is lost when using sampling frequencies of increasing length (Rooney et al. 1998, De Solla et al. 1999, Girard et al. 2002, Ppin et al. 2004). Nonetheless, we are aware of no published tests of GPS systems using fix intervals (FI) of < 15 min to assess fine-scale animal movements. There is also little data available to assess techniques and sampling frequencies that optimize home range estimators using GPS technology (Girard et al. 2002). Such work is required to determine the necessary programming schedule to accurately measure wolf movement patterns.

Methods Study Area Algonquin Provincial Park (APP), Ontario, Canada (45N, 78W) lies on the southern edge of the Canadian Shield and covers 7571 km2 (Pimlott et al. 1969, Patterson et al. 2004). Situated at the transition between boreal and mixed-deciduous forest, the landscape is composed mostly of pines (Pinus strobus, P. resinosa, P. banksiana) and hardwoods and other deciduous species (Betula papyrifera, Populus tremuloides, Quercus rubra, Acer saccharum, A. rubrum). Elevation ranges from 180-580 m (Cook et al. 1999). The topography is characterized by rocky hills and ridges in the west and flatter and less rocky terrain in the east (Joslin 1967).

48

GPS transmitter efficiency In summer 2003, 3 adult wolves were fitted with GPS transmitters (Model 4000; Lotek Engineering, Inc., Newmarket, Ontario). Each GPS transmitter was scheduled to obtain one location every 0.25, 1.5, 2, 6 and 12 hr for a total of 576, 4718, 1149, 312 and 230 fix attempts. We calculated the mean fix acquisition time (transmitters terminated unsuccessful attempts at 180 sec) and mean fix success rate (no. successful fixes / no. of attempts) for each transmitter at each FI. We then estimated transmitter battery duration at each FI using the respective mean acquisition time and computer software provided by the manufacturer (GPS Plus, Lotek Engineering, Inc., Newmarket, Ontario). We evaluated GPS transmitter efficiency at each FI by determining the mean number of successful locations that would be generated over the lifespan of the battery. Wolf movements In autumn 2004, we fitted two additional adult wolves with GPS transmitters (Model 4400M; Lotek Engineering, Inc., Newmarket, Ontario) that were scheduled to obtain location fixes at 5-min intervals during 24-hr periods (n = 14 days). Sampling was not conducted on consecutive days in order to ensure sample independence (e.g., wolves moved little for several days after an ungulate had been killed and thus locations over several days at a kill site were correlated). Location fixes were attempted only once at each scheduled time. Data were downloaded from GPS transmitters to a handheld command unit (GPS 4X00 compatible Handheld Command Unit, Lotek Engineering, Inc.) during regular telemetry flights. All data from each sample day was then subsampled at FIs

49

of increasing length (FIs ranging from one location every 10 min to 24 hr). Total daily movement distance was estimated by summing the straight line distance between successive points, and was mapped and measured for each FI subsample set in ArcView GIS 3.2a (Environmental Systems Research Institute, Inc., Redlands, California). In the case of unsuccessful location fixes, the next successful fix was used in its place. We excluded from the analysis one sample day with <75% location success (7% of sample days). We included periods where wolves were apparently inactive because of the subjectivity involved in differentiating fine scale movements from actual resting periods. Any movements measured during these inactive periods attributable to location error were typically much less than the distance between fixes obtained while wolves were moving and were expected to have minimal effects on observed patterns. We calculated a distance ratio by dividing the estimated distance traveled using progressively longer FIs by the distance estimated using the respective 5-min FI for that movement path (i.e., distance ratio = [meters moved at the subsampled FI] / [meters moved at the 5-min FI]). We fit standard logarithmic decay models to scatter plots of FI vs. the distance ratio and determined the relative fit of each model. Because the complexity of movement paths may affect distance ratios (i.e., linear paths have distance ratios that remain close to 1.00), the fractal dimension, D, was calculated for each movement path following methods described in Bascompte and Vil (1997). The fractal dimension quantifies the convolution present in a search path using the number of steps, the length, and the widest diameter of the path. The value of D is equal to 1.0 when the length and maximum diameter of the path are

50

equal and increases as the complexity of the path increases (Katz and George 1985). We compared estimates of D calculated using movement paths generated from 15min FIs (total n = 33 movement paths) obtained from 5 adult wolves, to the respective 24-hr distance ratio using simple linear regression. We predicted an inverse relationship between these variables because complex movement paths should generate smaller distance ratios relative to more linear movement paths. Territory delineation Data from 8 transmittered adult wolves were used to assess territory delineation using locations obtained from GPS transmitters using variable sampling frequencies (Models 4000, 4400S, and 4400M; Lotek Engineering, Inc., Newmarket, Ontario). Locations collected between 1 October and 30 April were used to define winter territories when wolves are not restricted in their movements due to pup rearing activities. The duration of data collection for each GPS transmitter used in this analysis averaged 117 20 days (SE; range: 76211 days). GPS location sets from each transmittered wolf were subsampled by taking one location every 0.25, 0.5, 1, 3.5, and 7 days. Territory size was measured for each FI subsample set using a 95% fixed mean minimum convex polygon (MCP) as well as kernel-based territory size estimate. Adaptive kernels were calculated using increasing proportions of the generated smoothing parameter (starting at 0.1 h_ref then increased at 0.1 increments) until a single polygon territory was obtained (A. Rodgers, Ontario Ministry of Natural Resources and J. Kie, United States Forest Service, personal communication). This technique was used to avoid analytical problems often

51

encountered when using least squares cross-validation with large data sets (Hemson et al. 2005). For both estimators the change in territory size estimated using progressively longer sampling frequencies was quantified by dividing estimated territory area at each sampling frequency by the respective territory size estimate from the 0.25-day FI (i.e., the area ratio). We then determined the best fit model for the relationship between 95% MCP territory size and sampling frequency using information-theoretic methods (Anderson et al. 2000). We conducted MCP territory size bootstraps for each territory (50 replicates using the 0.25-day FI) and determined the mean number of locations required to obtain a 95% MCP. All spatial measurements were performed using the Home Range Extension (A. R. Rodgers and A. P. Carr, Ontario Ministry of Natural Resources) in ArcView 3.2a. We ignored autocorrelation within the data because the data continued to exhibit a high degree of dependence even when using extended fix intervals (24 hr) (e.g., Reynolds and Laudr 1990, Rooney et al. 1998, De Solla et al. 1999). Data for analyses of transmitter efficiency, movements and territory size were collected during autumn and winter (23 September 30 April) to minimize the effect of vegetation and canopy cover on GPS transmitter performance (Rempel et al. 1995, Moen et al. 1996, Frair et al. 2004, DeCesare et al. 2005).

Results GPS transmitter efficiency In general, mean fix success rates decreased and location acquisition times increased with increasing FI (fix success: 95 3%, 93 3%, 83 5%, 75 5, 79

52

7%; location acquisition time: 52 10 sec, 72 8 sec, 99 10 sec, 119 8 sec, 100 13 sec for 0.25, 1.5, 2, 6, and 12-hr FIs, respectively). Transmitters set to obtain locations at 1.5-hr FIs were estimated to be the most efficient. In fact, the shorter fix acquisition time for the 1.5-hr FI should allow a transmitter programmed at this sampling frequency to collect approximately 3,150 more successful locations than if a 2-hr FI was used. Transmitter efficiency was low for the 0.25-hr FI, despite high fix success and short acquisition times, due to the high sampling frequency. Transmitter efficiency was also low for the 6 and 12-hr FIs as a result of long fix acquisition times and low fix success rates (Figure A.1). Wolf movements Mean daily movement distance was estimated as 21.4 1.5 km (SE; range: 11.231.6 km), and the respective mean hourly movement rate was 0.83 0.25 km/hr (range: 0.471.31 km/hr) for sample days using 5-min FIs (n = 13 days). The distance ratio for this dataset followed closely a logarithmic decay (logarithmic: AICc = 5.26, RMSE: 0.104; Figure A.2). This indicated a substantive reduction in estimated movement distance for any FIs exceeding 5 min. The mean fractal dimension for all movement paths was 1.25 0.03 (range: 0.96-1.86; n = 33 paths) and was inversely related to the distance ratio measured for the 24-hr FI subsamples (t31 = -6.46, P < 0.001, Figure A.3). Territory delineation Mean 95% MCP territory size was 228 67 km2 (range: 50620; n = 8 territories) measured using 0.25-day FIs. All territory size estimates decreased with increasing FI (Figure A.4a). The model explaining the rate of reduction in estimated

53

territory size due to decreasing sampling frequency followed a logarithmic decay (logarithmic: AICc = 0.38, RMSE: 0.092; Weibull RMSE: 0.082; Figure A.4a). A mean of 127 24.5 locations, comprising 34.1 5.5% of all collected locations, was required to define territory sizes equal in area to the respective 95% MCPs measured using all available fixes for each territory (Figure A.5). Mean kernel territory size was 198 57 km2 (range: 39515; n = 8 territories) measured using 0.25-day FIs. There was little change in estimated territory size as FI increased beyond 0.25 hr (Figure A.4b). In fact, adaptive kernel home ranges were much more robust to variable sampling frequency, with a mean deviation from the 0.25-day FI kernel size being only 5.7 3.8% whereas the equivalent mean deviation for the MCP subsamples was 30.0 4.0%.

Discussion GPS transmitter efficiency, movement distance, and territory size estimates are scale-dependent and are greatly affected by variable FIs (Schaefer and Mahoney 2003). Fix intervals of intermediate length (e.g., 1.5 hr) had the highest efficiency because they balanced the time required for location acquisition with the number of daily location attempts. They also provided transmitter operational lifespan that would be adequate for most long-term studies (~1.75 years), especially for those where the redeployment of collars was planned. Transmitters set to FIs of intermediate length would also have the potential to generate more successful fixes throughout their operational lives than those set to obtain fixes at less intense sampling frequencies. Therefore, researchers should be aware that, while long FIs

54

(e.g., > 2 hr) can extend monitoring periods, the acquisition of locations at consistent time intervals may be compromised due to unsuccessful location attempts. This problem could bias assessment of fine-scale movement patterns in many species (e.g., Frair et al. 2004). Fix intervals <1.5 hr were also less efficient, despite short acquisition times and high fix success rates, due to the high number of daily fix attempts that accelerated battery power depletion. Our estimated movement distance followed a logarithmic decay with increasing fix interval that was more extreme at short fix intervals. This model was sufficiently precise to be used to reliably estimate actual distance traveled from data sets where longer FIs (e.g., 15 min) are used. The logarithmic decay models we generated are consistent with previous observations from studies using coarser sampling frequencies applied to less mobile species (e.g., Reynolds and Laundr 1990, De Solla et al. 1999, Ppin et al 2004). However, the rapid deviation from the actual distance traveled as FI increases has not been explored fully in previous tests of GPS telemetry, and highlights the necessity for fine-scale measurements and appropriate GPS transmitter programming schedules when studying subtle animal movements. This is illustrated by the fact that movement distances estimated using 15-min FI subsamples underestimated movements measured using 5-min FIs by a mean of 15.2%. The degree by which actual movements are underestimated with increasing FIs is also dependent on the complexity of the movement path of the animal being monitored. This relationship deserves attention when studying animal movements, especially when movements are measured using straight line distance between

55

successive relocations. When movement paths are tortuous, the use of long FIs will increasingly underestimate actual movement distance, thereby providing another justification for the use of intensive sampling frequencies with GPS telemetry. The sharp reduction in estimated MCP territory size with decreasing sampling frequency brings into question the validity of territory size measurements obtained using typical VHF telemetry monitoring protocols. In large studies where many animals are monitored simultaneously it will be difficult to obtain the number of locations required to adequately assess territory size from VHF telemetry alone. Even if locations were collected daily, most territory sizes would be underestimated. In fact, in the present study we continued to underestimate wolf territory sizes by a mean of 16.1% when using locations collected daily for 2.5 7 months. Reviews comparing home range size or movement distances between different animal populations (e.g., Fuller and Murray 1998, Girard et al. 2002, Fuller et al. 2003) may be flawed if sampling protocols are inconsistent between studies. We determined that 95% MCP home ranges were highly sensitive to sampling intensity, which is alarming given that this metric is often used in studies of wolf territory sizes (e.g., Messier 1985, Kernohan et al. 2001). In contrast, adaptive kernel home ranges were more robust to variable sampling frequencies and should be the preferred method for measuring animal home ranges. An average of 127 locations were required to obtain accurate 95% MCPs for the eight wolf territories that we measured, whereas kernels estimated relatively accurate territory sizes using locations collected only weekly (range: 1232 locations). Thus, adaptive kernel home range estimates are clearly more efficient.

56

To summarize, we suggest using the shortest FI possible when the primary question concerns fine-scale animal movements. This reduces the operational life of the transmitter battery, but provides a better representation of fine-scale movement patterns. Longer FIs (e.g., 12 hr) are adequate for coarser data needs such as home range analysis. However, there exists a trade-off between transmitter battery duration and efficiency. To maximize efficiency when using long FIs, satellite availability in the particular study area should serve to determine specific transmitter scheduling patterns (Johnson and Barton 2004). Our results provide relevant insight regarding GPS telemetry study design. The pattern of data loss with increasing FI length has important implications and should not be ignored. When deploying GPS transmitters, researchers need to assess the costs and benefits of different sampling frequencies relative to research objectives and transmitter performance. In cases where GPS transmitters can be retrieved and re-deployed at relatively low cost compared to their initial purchase costs (e.g., Merrill et al. 1998), designing field studies to ensure that data quality is consistent with needs should take precedence over increasing transmitter longevity.

57

LITERATURE CITED Anderson, D. R., K. P. Burnham, and W. L. Thompson. 2000. Null hypothesis testing: problems, prevalence, and an alternative. Journal of Wildlife Management 64:912923. Bowman, J. L., C. O. Kochanny, S. Demarais, and B. D. Leopold. 2000. Evaluation of a GPS collar for white-tailed deer. Wildlife Society Bulletin 28:141145. Cain, J. W. III, P. R. Krausman, B. D. Jansen, and J. R. Morgart. 2005. Influence of topography and GPS fix interval on GPS collar performance. Wildlife Society Bulletin 33:926-934. Cook, S. J., D. R. Norris, and J. B. Theberge. 1999. Spatial dynamics of a migratory wolf population in winter, south-central Ontario (19901995). Canadian Journal of Zoology 77:17401750. DeCesare, N. J., J. R. Squires, and J. A. Kolbe. 2005. Effect of forest canopy on GPSbased movement data. Wildlife Society Bulletin 33(3):926-934. DEon, R. G., R. Serrouya, G. Smith, and C. O. Kochanny. 2002. GPS radiotelemetry error and bias in mountainous terrain. Wildlife Society Bulletin 30(2):430 439. DEon, R. G., and D. Delparte. 2005. Effects of radio-collar position and orientation on GPS radio-collar performance, and the implications of PDOP in data screening. Journal of Applied Ecology 42(2):383388. De Solla, S. R., R. Bonduriansky, and R. J. Brooks. 1999. Eliminating autocorrelation reduces biological relevance of home range estimates. Journal of Animal Ecology 68:221234.

58

Frair, J. L., S. E. Nielsen, E. H. Merrill, S. R. Lele, M. S. Boyce, R. H. M. Munro, G. B. Stenhouse, and H. L. Beyer. 2004. Removing GPS collar bias in habitat selection studies. Journal of Animal Ecology 41:201212. Fuller, T. K., and D. L. Murray. 1998. Biological and logistical explanations of variation in wolf population density. Animal Conservation 1:153157. Fuller, T. K., L. D. Mech, and J. F. Cochrane. 2003. Wolf population dynamics. Pages 161191 in L. D. Mech and L. Boitani, editors. Wolves: behavior, conservation, and ecology. University of Chicago Press, Chicago, Illinois, USA. Gamo, R. S., M. A. Rumble, R. Lindzey, and M. Stefanich. 1999. GPS radio collar 3D performance as influenced by forest structure and topography. Pages 464 473 in J. H. Eiler, D. J. Alcorn, and M. R. Neuman, editors. Proceedings of the 15th International Symposium on Biotelemetry. Juneau, Alaska, USA. Girard, I., J. Ouellet, R. Courtois, C. Dussault, and L. Breton. 2002. Effects of sampling effort based on GPS telemetry on home-range size estimations. Journal of Wildlife Management 66(4):12901300. Hemson, G., P. Johnson, A. South, R. Kenward, R. Ripley, and D. MacDonald. 2005. Are kernels the mustard? Data from global positioning systems (GPS) collars suggests problems for kernel home-range analyses with least-squares crossvalidation. Journal of Animal Ecology 74:455-463. Jedrzejewski, W., K. Schmidt, J. Theuerkauf, B. Jedrzejewska, and H. Okarma. 2001.

59

Daily movements and territory use by radio-collared wolves (Canis lupus) in Bialowieza Primeval Forest in Poland. Canadian Journal of Zoology 79:1993 2004. Johnson, C. E., and C. C. Barton. 2004. Where in the world are my field plots? Using GPS effectively in environmental field studies. Frontiers in Ecology and the Environment 2(9):475482. Joslin, P. W. B. 1967. Movements and home sites of timber wolves in Algonquin Park. American Zoologist 7:279288. Katz, M. J. and E. B. George. 1985. Fractals and the analysis of growth paths. Bulletin of Mathematical Biology 47(2):273286. Kauhala, K., and T. Tiilikainen. 2002. Radio location error and the estimates of home-range size, movements, and habitat use: a simple field test. Annales Zoologici Fennici 39:317324. Kernohan, B. J., R. A. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and movements. Pages 125166 in J. J. Millspaugh and J. M. Marzluff, editors. Radio tracking and animal populations. Academic Press, San Diego, California, USA. Merrill, S. B., L. G. Adams, M. E. Nelson, and L. D. Mech. 1998. Testing releasable GPS radiocollars on wolves and white-tailed deer. Wildlife Society Bulletin 26(4):830835. Merrill, S. B., and L. D. Mech. 2000. Details of extensive movements by Minnesota wolves (Canis lupus). American Midland Naturalist 144:428433. Merrill, S. B., and L. D. Mech. 2003. The usefulness of GPS telemetry to study wolf

60

circadian and social activity. Wildlife Society Bulletin 31(4):947960. Messier, F. 1985. Social organization, spatial distribution and population density of wolves in relation to moose density. Canadian Journal of Zoology 63:1068 1077. Moen, R., J. Pastor, Y. Cohen, and C. C. Schwartz. 1996. Effects of moose movement and habitat use on GPS collar performance. Journal of Wildlife Management 60(3):659668. Musiani, M., H. Okarma, and W. Jedrzejewski. 1998. Speed and actual distance traveled by radiocollared wolves in Bialowieza Primeval Forest (Poland). Acta Theriologica 43(4):409416. Patterson, B. R., N. W. S. Quinn, E. F. Becker, and D. B. Meier. 2004. Estimating wolf densities in forested areas using network sampling of tracks in snow. Wildlife Society Bulletin 32(3):938947. Ppin, D., C. Adrados, C. Mann, and G. Janeau. 2004. Assessing real daily distance traveled by ungulates using differential GPS locations. Journal of Mammalogy 85(4):774780. Pimlott, D. H., J. A. Shannon, and G. B. Kolenosky. 1969. The ecology of the timber wolf in Algonquin Provincial Park. Ontario Department of Lands and Forests, Fish and Wildlife Research report No. 87. Rempel, R. S., A. R. Rodgers, and K. F. Abraham. 1995. Performance of a GPS animal location system under boreal forest canopy. Journal of Wildlife Management 59(3):543551. Reynolds, T. D., and J. W. Laundr. 1990. Time intervals for estimating pronghorn

61

and coyote home ranges and daily movements. Journal of Wildlife Management 54(2):316322. Rodgers, A. R., and P. Anson. 1994. Animal-borne GPS: tracking the habitat. GPS World. July 1994. Rodgers, A. R. 2001. Tracking animals with GPS: the first 10 years. Pages 110 in A. M. Sibbald and I. J. Gordon, editors. Tracking animals with GPS. The Macaulay Institute, Aberdeen, Scotland. Rooney, S. M., A. Wolfe, and T. J. Hayden. 1998. Autocorellated data in telemetry studies:time to independence and the problem of behavioural effects. Mammal Review 28:8998. Schaefer, J. A., and S. P. Mahoney. 2003. Spatial and temporal scaling of population density and animal movement: a power law approach. Ecosience 10:496-501.

62

Figure A.1. Estimated transmitter battery duration (mean SE) and the number of successful locations (mean SE) obtainable for Lotek model 4000 GPS collars using fix intervals of increasing length (n = 3 collars).

63

Figure A.2. Logarithmic model for the relationship between fix interval (FI) and distance ratio (DR = distance measured for the x-hr FI subsample / distance measured for the 5-min FI subsample) for wolf movements measured using GPS transmitters on 5-min FIs for 24 hour sampling periods (n = 13 movement paths). The logarithmic fit was y = 0.672+ -0.127 ln(x).

64

Figure A.3. Relationship between distance ratio and the fractal dimension (D) for wolf movement paths measured using GPS collars in Algonquin Provincial Park (n = 33 movement paths). The fractal dimension is calculated as: D = log(n) / (log(n) + log(d/L)), where n = no. of steps in the movement path, d = maximum diameter of the movement path, and L = measured length of the movement path.

65

b Figure A.4. Logarithmic model for to the relationship between fix interval and area ratio (AR = area measured for the x-day FI / area measured for the .25-day FI) for (a) 100% Minimum Convex Polygon (n = 8 wolf territories) and (b) adaptive kernel territory size (n = 8 wolf territories) measured using GPS collars in Algonquin Provincial Park. The logarithmic fit for the fix interval and MCP area ratio was y = 0.800 + -0.175 ln(x).

66

Figure A.5. Relationship between number of GPS telemetry locations (n = 309 locations) and percent total area of a minimum convex polygon territory for a collared wolf (W129) in Algonquin Provincial Park.

67