Wildlife accident reduction study and monitoring : Arizona State Route 64

              Wildlife Accident Reduction
Study and Monitoring:
Arizona State Route 64
Final Report 626
November 2012
Arizona Department of Transporta􀆟􉽯on
Research Center
Wildlife Accident Reduction Study
and Monitoring:
Arizona State Route 64
Final Report 626
November 2012
Prepared by:
Norris L. Dodd, Jeffrey W. Gagnon, Scott Sprague,
Susan Boe, and Raymond E. Schweinsburg
Arizona Game and Fish Department
5000 W. Carefree Hwy.
Phoenix, AZ 85068
Prepared for:
Arizona Department of Transportation
in cooperation with
U.S. Department of Transportation
Federal Highway Administration
This report was funded in part through grants from the Federal Highway Administration,
U.S. Department of Transportation. The contents of this report reflect the views of the
authors, who are responsible for the facts and the accuracy of the data, and for the use
or adaptation of previously published material, presented herein. The contents do not
necessarily reflect the official views or policies of the Arizona Department of
Transportation or the Federal Highway Administration, U.S. Department of
Transportation. This report does not constitute a standard, specification, or regulation.
Trade or manufacturers’ names that may appear herein are cited only because they are
considered essential to the objectives of the report. The U.S. government and the State
of Arizona do not endorse products or manufacturers.
Technical Report Documentation Page
1. Report No.
FHWA-AZ-12-626
2. Government Accession No.
3. Recipient’s Catalog No.
4. Title and Subtitle
Wildlife Accident Reduction Study and Monitoring:
Arizona State Route 64
5. Report Date
November 2012
6. Performing Organization Code
7. Authors
Norris L. Dodd, Jeffrey W. Gagnon, Scott C. Sprague, Susan Boe,
and Raymond E. Schweinsburg
8. Performing Organization Report No.
9. Performing Organization Name and Address
Arizona Game and Fish Department
Research Branch
5000 W. Carefree Hwy.
Phoenix, AZ 85068
10. Work Unit No.
11. Contract or Grant No.
SPR-000-1(171)626
12. Sponsoring Agency Name and Address
Arizona Department of Transportation
Research Center, 206 S. 17th Ave., MD075R
Phoenix, AZ 85007
ADOT Project Manager: Estomih Kombe, Ph.D., PE
13.Type of Report & Period Covered
Final Report
14. Sponsoring Agency Code
15. Supplementary Notes
Prepared in cooperation with the U.S. Department of Transportation Federal Highway Administration
16. Abstract
The research team assessed elk (Cervus elaphus), mule deer (Odocoileus hemionus), and pronghorn (Antilocapra
americana) movements and vehicle collision patterns from 2007 through 2009 along a 57 mi stretch of State Route
(SR) 64 to develop strategies to improve highway safety and wildlife permeability. This study followed the SR 64 2006
Final Wildlife Accident Reduction Study that recommended nine wildlife passage structures and further monitoring to
determine the best locations for passage structures and fencing. Research objectives were to:
 Assess wildlife movements, highway crossing patterns, and permeability across SR 64.
 Assess relationships of wildlife crossings and distribution to vehicular traffic volume.
 Investigate wildlife-vehicle collision spatial and temporal incidence and patterns.
 Determine use of Cataract Canyon Bridge by wildlife for below-grade passage.
 Develop recommendations to enhance highway safety and wildlife permeability.
The team tracked 23 elk, 11 deer, and 15 pronghorn with Global Positioning System (GPS) receiver collars, yielding
mean passage rates of 0.44, 0.54, and 0.004 crossings/approach, respectively. In total, 167 wildlife-vehicle collisions
were analyzed. Traffic volume influenced permeability and wildlife-vehicle collision patterns. The team recommended
11 passage structures, including Cataract Canyon Bridge, which had modest current wildlife use, along with wildlife
fencing to reduce collisions and promote permeability for elk, deer, and pronghorn.
17. Key Words
Elk, GPS telemetry, fencing, highway impact, mule deer,
permeability, pronghorn, traffic volume, wildlife passage
structures, wildlife accident reduction, wildlife-vehicle
collisions
18. Distribution Statement
This document is available to the U.S.
public through the National Technical
Information Service, Springfield,
Virginia 22161
23. Registrant’s Seal
19. Security Classification
Unclassified
20. Security Classification
Unclassified
21. No. of Pages
111
22. Price
TABLE OF CONTENTS
1.0 INTRODUCTION ................................................................................................... 3
2.0 LITERATURE REVIEW ........................................................................................ 9
2.1 Background .....................................................................................................9
2.2 Research Justification ....................................................................................12
2.3 Research Objectives ......................................................................................16
3.0 Study Area ............................................................................................................. 17
3.1 Natural Setting ...............................................................................................17
3.1.1 Climate ............................................................................................ 17
3.1.2 Vegetation ....................................................................................... 19
3.1.3 Wildlife Species .............................................................................. 19
3.1.4 Cataract Canyon Bridge .................................................................. 21
3.2 Traffic Volume ..............................................................................................22
4.0 Methods.................................................................................................................. 25
4.1 Wildlife Capture, GPS Telemetry, and Data Analysis ..................................25
4.1.1 Elk Capture ..................................................................................... 25
4.1.2 Mule Deer Capture .......................................................................... 25
4.1.3 Pronghorn Capture .......................................................................... 25
4.1.4 GPS Analysis of Animal Movements ............................................. 27
4.1.5 Calculation of Crossing and Passage Rates .................................... 27
4.1.6 Calculation of Pronghorn Approaches ............................................ 29
4.1.7 Calculation of Weighted Crossings and Approaches...................... 30
4.2 Traffic Volume and Animal Distribution Relationships ...............................30
4.3 Wildlife-Vehicle Collision Relationships .....................................................31
4.4 Wildlife Use of Cataract Canyon Bridge ......................................................31
4.5 Identification of Passage Structure Sites .......................................................33
5.0 Results .................................................................................................................... 37
5.1 Wildlife Capture, GPS Telemetry, and Data Analysis ..................................37
5.1.1 Elk Capture, Movements, and Highway Permeability .................... 37
5.1.2 Mule Deer Capture, Movements, and Highway Permeability ........ 41
5.1.3 Pronghorn Capture, Movements, and Highway Approaches .......... 42
5.2 Traffic Relationships .....................................................................................46
5.2.1 Elk-Traffic Relationships ................................................................ 46
5.2.2. Mule Deer–Traffic Relationships ................................................... 46
5.2.3. Pronghorn-Traffic Relationships .................................................... 51
5.3. Wildlife-Vehicle Collision Relationships .....................................................51
5.4 Wildlife Use of Cataract Canyon Bridge ......................................................58
5.5 Identification of Passage Structure Sites .......................................................59
5.5.1 Passage Structure Recommendations by Highway Section ............ 64
6.0 Discussion .............................................................................................................. 69
6.1 Wildlife Permeability ....................................................................................69
6.2 Wildlife Distribution and Traffic Relationships ............................................71
6.3 Wildlife-Vehicle Collision Relationships .....................................................72
6.4 Cataract Canyon Bridge Wildlife Use ...........................................................74
6.5 Identification of Passage Structure Sites .......................................................75
6.5.1 Passage Structure Design Considerations ....................................... 75
6.5.2 Role of Passage Structure Spacing ................................................. 79
6.5.3 Role of Fencing ............................................................................... 79
7.0 Conclusions and Recommendations ...................................................................... 83
7.1 Final Wildlife Accident Reduction Study Role .............................................83
7.2 Wildlife Permeability and Passage Structures ..............................................84
7.2.1 Elk and Mule Deer Permeability ..................................................... 84
7.2.2 Pronghorn Permeability .................................................................. 85
7.3 Impact of Traffic and Noise ..........................................................................86
7.4 Passage Structure Design and Placement ......................................................86
7.4.1 Role of Passage Structure Spacing ................................................. 87
7.4.2 Role of Fencing ............................................................................... 87
7.5 Highway Safety and Wildlife-Vehicle Collisions .........................................88
7.6 Monitoring .....................................................................................................89
References ......................................................................................................................... 91
LIST OF FIGURES
Figure 1. Landownership, Mileposts, 0.1 mi Segments, Highway Sections
A–E, and Preliminary Wildlife Passage Structures in the SR 64
Study Area Identified by the Final Wildlife Accident Reduction
Study (ADOT 2006). ................................................................................ 15
Figure 2. Great Basin Conifer Woodland Adjacent to SR 64 (Top) with
Open to Dense Stands of Pinyon and Juniper and Cliffrose, Apache
Plume, and Other Shrubs, and Plains and Great Basin Grasslands
(Bottom) Dominated by Blue and Black Grama, Galleta, and
Needle-and-Thread Grasses. ..................................................................... 18
Figure 3. Density Distributions for the Three Target Species of Research
along SR 64: Elk (Left), Mule Deer (Center), and Pronghorn
(Right). ...................................................................................................... 20
Figure 4. Cataract Canyon Bridge on SR 64. ........................................................... 21
Figure 5. Hourly Traffic Volume (Vehicles per Hour) along SR 64, Arizona,
from 2007 through 2009. Note the Low Volume of Traffic during
Nighttime Hours (00:00–04:00). ............................................................... 23
Figure 6. Photographs of Capture Techniques Used for Elk, Mule Deer and
Pronghorn along SR 64 and GPS-Collared Animals: Elk Captured
with Net-Covered Clover Trap (Top), Darting to Immobilize Mule
Deer (Middle), and Net-Gunning of Pronghorn from a Helicopter
(Bottom). ................................................................................................... 26
Figure 7. GPS Locations and Lines between Successive Fixes to Determine
Highway Approaches and Crossings in 0.10 mi Segments. The
Expanded Section Shows GPS Locations and Lines between
Successive Fixes to Determine Approaches to the Highway
(Shaded Band) and Crossings. Example A Denotes an Approach
and Crossing; Example B Denotes an Approach without a
Crossing. ................................................................................................... 29
Figure 8. Reconyx™ Camera Mounted on a Wood Strip Glued to the
Concrete Surface of Each SR 64 Cataract Canyon Bridge Culvert
Cell to Monitor Wildlife Use. ................................................................... 32
Figure 9. Images of a Mule Deer Doe (Left) and Spike Bull Elk (Right)
Recorded by Reconyx™ Cameras Mounted in the SR 64 Cataract
Canyon Bridge Culvert Cells. ................................................................... 32
Figure 10. SR 64 Crossings by GPS-Collared Elk along the Entire Study Area
(Top) and Sections A through E of the 2006 Final Wildlife
Accident Reduction Study and Enlarged to Show Crossings along
Sections D and E (Bottom). ...................................................................... 39
Figure 11. SR 64 Weighted Crossings by GPS-Collared Elk along the Entire
Study Area (Top) and Sections A through E of the 2006 Final
Wildlife Accident Reduction Study and Enlarged to Show Crossings
along Sections D and E (Bottom). ............................................................ 40
Figure 12. Mule Deer GPS Fixes along the SR 64 Study Area, as well as
Fixes for Two Deer Captured North of Flagstaff (Numbers 43 and
172). .......................................................................................................... 42
Figure 13. SR 64 Highway Crossings (Top) and Weighted Crossings
(Bottom) by GPS-Collared Mule Deer along Highway Section E
by 0.1 mi Segment. ................................................................................... 43
Figure 14. Highway Approaches (Top) and Weighted Approaches (Bottom)
Made to within 0.3 mi of SR 64 by GPS-Collared Pronghorn and
Sections A through E of the 2006 Final Wildlife Accident
Reduction Study. ....................................................................................... 45
Figure 15. Mean Probability That GPS-Collared Elk Occurred within 330 ft
Distance Bands along SR 64 at Varying Traffic Volumes. ...................... 48
Figure 16. Mean SR 64 Passage Rates by Two-Hour Time Blocks (Reflected
by the Midpoint of the Blocks) and Corresponding Mean Traffic
Volumes during Each Time Block for Elk (Bottom) and Mule Deer
(Top). ........................................................................................................ 49
Figure 17. Mean Probability That GPS-Collared Mule Deer Occurred within
330 ft Distance Bands along SR 64 at Varying Traffic Volumes. ............ 50
Figure 18. Mean Probability That GPS-Collared Pronghorn Occurred within
330 ft Distance Bands along SR 64 at Varying Traffic Volumes. ............ 52
Figure 19. Frequency of Elk and Mule Deer Collisions with Vehicles by
SR 64 Milepost from 2007 through 2009. ................................................ 53
Figure 20. Proportion of SR 64 Single-Vehicle Accidents by Milepost from
1998 through 2008 that Involved Wildlife. ............................................... 54
Figure 21. SR 64 Elk and Mule Deer Collisions with Vehicles by Time of
Day and Associated Traffic Volume. ........................................................ 55
Figure 22. SR 64 Elk and Mule Deer Collisions with Vehicles by Day and
Associated Traffic Volume. ...................................................................... 56
Figure 23. SR 64 Elk and Mule Deer Collisions with Vehicles by Month and
the Mean Traffic Volume. ......................................................................... 58
Figure 24. Ratings for 95 SR 64 0.6 mi Segments Using Wildlife Movement,
WVC Data, and Other Criteria to Determine the Location of
Potential Wildlife Passage Structures. Red Bars Denote Segments
Where Underpasses Were Recommended in the 2006 Final
Wildlife Accident Reduction Study and Orange Where Overpasses
Were Recommended (Table 9). The Green Bars Represent
Segments Where Additional Structures Are Recommended as a
Result of This Study.................................................................................. 61
Figure 25. Recommendations for SR 64 Wildlife Passage Structures and
Wildlife Fencing for Highway Section A (Left) and Section B
(Right). ...................................................................................................... 65
Figure 26. Recommendations for SR 64 Wildlife Passage Structures and
Wildlife Fencing for Highway Sections D and E. .................................... 66
Figure 27. Various Wildlife Passage Structure Options for SR 64, Including
CON/SPAN® Pre-Cast Concrete Arches for Overpasses, with a
Rendering of a Pronghorn Overpass on US 89 Integrated into Cut
Slopes (Top Left) and a Stand Alone Overpass on US 93 in
Montana (Top Right), Single-Span Bridged Underpasses Similar to
Those Used on SR 260 (Center), and Corrugated Multi-Plate Arch
Underpasses Used along US 93 in Montana (Bottom). ............................ 78
Figure 28. An Electrified Barrier Installed in the Pavement to Prevent
Wildlife from Breaching the Fenced Corridor at the Fencing
Terminus. This Mat was Installed on I-40 Off-Ramps in New
Mexico. ..................................................................................................... 81
LIST OF TABLES
Table 1. Vehicle Accidents Involving Collisions with Elk and Mule Deer
along SR 64 from 1991 through 2003, Including the Mean Number
of Collisions (per Year and per Mile). ...................................................... 13
Table 2. SR 64 Sections and Mileposts with Proposed Wildlife Mitigation
Measures for Focal Wildlife Species Identified in the 2006 Final
Wildlife Accident Reduction Study. ........................................................... 14
Table 3. Comparative Mean Values for GPS-Collared Animals by Species
Determined from GPS Telemetry along SR 64. ....................................... 38
Table 4. Mean Probabilities that GPS-Collared Elk, Mule Deer, and
Pronghorn Occurred within Distance Bands from SR 64 at Varying
Traffic Volumes. Documented from 2007 through 2009. ........................ 47
Table 5. WVCs Involving Elk and Mule Deer on SR 64 Sections from 2007
through 2009, including the Total Number and Mean Collisions
(per Mile). ................................................................................................. 51
Table 6. Frequency of Elk and Deer Collisions with Vehicles along SR 64
by Time Period. ......................................................................................... 57
Table 7. Frequency of Elk and Deer Collisions with Vehicles along SR 64
by Season. ................................................................................................. 57
Table 8. Number of Animals by Species that Entered and Successfully
Crossed through Cataract Canyon Bridge on SR 64, and Success
Rates. ......................................................................................................... 59
Table 9. Wildlife Passage Structure Locations along SR 64 by Milepost and
Highway Section and Types Recommended in the Various 2006
Final Wildlife Accident Reduction Study Alternatives and Those
Recommended as a Result of the Current Wildlife Movements
Study. ........................................................................................................ 63
ACRONYMS AND ABBREVIATIONS
AADT average annual daily traffic
ADOT Arizona Department of Transportation
AGFD Arizona Game and Fish Department
ATR automatic traffic recorder
BLM Bureau of Land Management
DPS Department of Public Service
ft foot or feet
GCNP Grand Canyon National Park
GMU Game Management Unit
GPS Global Positioning System
hr hour(s)
I-40 Interstate 40
MCP minimum convex polygon
mi mile(s)
MP milepost
mph miles per hour
NF National Forest
ROW right(s)-of-way
SDI Shannon diversity index
SE standard error
SR State Route
U.S. United States
US 89 U.S. Route 89
US 93 U.S. Route 93
US 180 U.S. Route 180
VHF very high frequency
WVC wildlife-vehicle collision
LIST OF SPECIES
Animals
Badger Taxidea taxus
Black bear Ursus americanus
Black-tailed jackrabbit Lepus californicus
Caribou Rangifer tarandus
Coyote Canis latrans
Desert cottontail rabbit Sylvilagus audubonii
Elk Cervus elaphus
Gray fox Urocyon cinereoargenteus
Grizzly bear Ursus arctos horribilis
Moose Alces alces
Mountain lion Puma concolor
Mule deer Odocoileus hemionus
Pronghorn Antilocapra americana
Raccoon Procyon lotor
Squirrel Spermophilus variegatus
Striped skunk Mephitis mephitis
White-tailed deer Odocoileus virginianus couesi
Wolf Canis lupus
Plants
Apache plume Fallugia paradoxa
Big sagebrush Artemisia spp.
Black grama Bouteloua eriopoda
Blue grama Bouteloua gracilis
Cliffrose Cowania mexicana
Galleta Pleuraphis jamesii
Gambel oak Quercus gambelii
Needle-and-thread grass Hesperostipa comata
One-seed juniper Juniperus monosperma
Pinyon Pinus edulis
Ponderosa pine Pinus ponderosa
Rabbitbrush Ericameria nauseosa
Winterfat Ceratoides lanata
ACKNOWLEDGMENTS
This project was funded by the Arizona Department of Transportation (ADOT) Research
Center and the Federal Aid Wildlife in Restoration Act, Project W-78-R, supporting
Arizona Game and Fish Department (AGFD) research. The research team commends
ADOT for its proactive commitment to promoting wildlife connectivity. The support of
the Federal Highway Administration, including former Environmental Project Manager
Steve Thomas, was instrumental to the funding and conduct of the project.
Many individuals at ADOT provided support and guidance in this project. The research
team commends Roadway Predesign Section for its commitment to developing
preliminary strategies for resolution of wildlife-highway conflicts. Estomih Kombe of the
ADOT Research Center provided project oversight and coordination. The team thanks
John Harper and Chuck Howe of the Flagstaff District for their tremendous support and
innovative management. Doug Eberline and Jennifer Toth of the Multimodal Planning
Division provided traffic data support. The team also thanks Todd Williams and Justin
White, of the Office of Environmental Services as well as Bruce Eilerts and Siobhan
Nordhaugen, formerly with the same office, for their commitment to the project and
overall efforts to address wildlife permeability.
AGFD Flagstaff Region personnel played a crucial role in supporting the project,
including Ron Sieg, Tom McCall, Carl Lutch, and David Rigo. The outstanding capture
support provided by Larry Phoenix was vital to the success of pronghorn capture and the
project. The research team also is indebted to the capable pilots of Papillion Helicopters.
Kari Ogren, Rob Nelson, and Chad Loberger of the AGFD Research Branch provided
invaluable field support, data collection, and analysis.
Highway patrol officers with the Arizona Department of Public Safety (DPS) Flagstaff
District made an outstanding effort to record all wildlife-vehicle collisions along State
Route (SR) 64. They collected and recorded information that was important to the
research project beyond what was required on accident reports. The research team is
particularly grateful to Matt Bratz for his coordination of accurate accident data
collection by DPS and his valuable input into the 2006 Wildlife Accident Reduction
Study.
The Kaibab National Forest (NF) provided invaluable logistical support during the
project. In particular, Jeffrey Waters provided project guidance and coordination, as well
as assistance with pronghorn capture.
The Arizona Antelope Foundation, the Rocky Mountain Elk Foundation, the Arizona Elk
Society, the Arizona Deer Association, and the Mule Deer Foundation were crucial in
helping to meet matching funding requirements for the project. Their interest and
commitment to efforts to promote wildlife permeability are sincerely appreciated.
The Technical Advisory Committee (TAC) provided many suggestions toward improving
the project’s effectiveness and applicability. Its tremendous support, oversight, and
commitment throughout the project are appreciated.
1
EXECUTIVE SUMMARY
The ADOT Research Center funded this study through a funding allocation from the
Federal Highway Administration (FHWA) State Planning and Research Program (SPR).
The Arizona Game and Fish Department was assigned the lead role in the execution of
the study and making recommendations based on the results. This partnership was made
possible with a joint project agreement (JPA) between the two state departments. The
study would concentrate on a 57 mile stretch of SR 64 beginning at the southern end,
which is the junction with Interstate 40. The focus would be a thorough evaluation of the
movement of elk, mule deer, and pronghorn in relation to highway and habitat
characteristics, traffic volumes, wildlife related accidents, and existing highway assets
like bridges.
The incidence of wildlife vehicle collisions along State Route 64 (SR 64) in Arizona has
been on the rise and thus a growing safety issue. Data collected over a ten year period
ending in 2008 showed 42 percent of single vehicle accidents in the study area involved
wildlife. The national average for wildlife related accidents is only five percent. In
addition, on a five mile stretch of highway at the north end of the study area wildlife
related accidents accounted for 75 percent of all single vehicle accidents.
Apart from the safety issue, good wildlife management means that we need to pay
attention to whether highway infrastructure may be creating a barrier to essential wildlife
movement within its habitat. In the long term, for wildlife to flourish, it is important that
man made barriers do not create scattered ‘islands’ of smaller and smaller animal
populations. Such an unintended segregation of wildlife populations has the potential to
result in diminished genetic strength and other weaknesses related to small numbers that
ultimately leads to a slow death for the affected species. It is therefore important that
efforts to address one issue (like wildlife vehicle collisions) are not done in a manner that
worsens the situation with respect to another important consideration. Solutions need to
be developed that strike a good balance between these needs.
What the Data Shows
For purposed of data collection and analysis, the designated length of highway for the
study was divided into small sections one tenth of a mile long. This would enable
researchers to clearly identify and chart out where within the proximity of the highway
animals were located or seen to make successful crossings. Monitoring of animal
movements was made possible by the use of Global Positioning System (GPS) collars on
animals. Successful crossings by an animal were identified when two consecutive
location coordinates for a collared animal matched locations on opposite sides of the
highway.
In addition to the use of GPS collars to monitor animal movement and highway
crossings, wildlife-vehicle collision data and traffic volume data were collected during
the study as well as from other sources. Relevant evaluations by other researchers in
prior years were also reviewed to see whether they were in agreement with or
2
contradicted any of our results. The evaluation attempted to establish relationships
between wildlife crossing levels and corresponding wildlife vehicle accidents and traffic
levels for different highway segments.
Analysis of the data established a positive relationship between wildlife crossing figures
and vehicle collisions for both elk and mule deer. No pronghorn crossings or
accidents/collisions were documented. It is thought that the highway may constitute
enough of a barrier that pronghorn will not venture to approach it. In comparison to
highway approach and crossing data seen for elk and deer in other locations like State
Route 260, the approach and crossing levels documented for SR 64 are considerably
lower. This is thought to be explained in part by the absence of attractive wet
meadow/riparian foraging habitat areas. Overall, high traffic volumes were associated
with lower wildlife (in this case elk and mule deer) approaches to and crossings of
highways. Where these high traffic volumes lasted only short durations, and thus could
be considered temporary, animals could be expected to return to habitat close to the
highway when the period of high traffic volume ended.
Summary of Conclusions and Recommendations
 For some wildlife (like elk and mule deer), wildlife vehicle accident data can be
used reliably in the identification of locations where wildlife crossing assets can
make a big impact. When possible, GPS tracking is useful for supplemental data.
 For animals like pronghorn which have a strong tendency to keep their distance
from busy highways, GPS tracking studies are crucial for collecting the data
necessary to identify potential solutions.
 Measures to reduce noise and negative visual impacts near wildlife crossing
infrastructure have potential to enhance the effectiveness of these assets.
 Passage structure designs should consider some important characteristics, some
specific to the main animal species of concern, noted in the report to maximize the
use of these assets.
 Based on the full set of data collected and analyzed as part of this study, the
research team identified a total of eleven potential wildlife passage locations. As
opportunities arise, consideration should be made towards the implementation of
some underpasses/overpasses or the retrofitting of bridges as the case may be. The
specific details are provided for each of the potential locations.
 For existing or future wildlife passage structures, having a length of appropriate
fencing to ‘channel’ animals to the point of crossing is an important part of
achieving maximum benefits from what is typically a sizeable infrastructure
investment.
3
1.0 INTRODUCTION
The research team assessed wildlife-highway relationships from 2007 through 2009 along
a 57 mile stretch of State Route (SR) 64, the highway linking Interstate 40 (I-40) and
Grand Canyon National Park (GCNP) in north-central Arizona. The incidence of
wildlife-vehicle collisions (WVCs) involving elk and mule deer along this stretch of
highway is a significant and growing concern, as is the ability of wildlife to cross the
highway corridor, or permeability. This predominately two-lane highway will be
reconstructed in the future to a four-lane divided highway to address growing traffic
volume and the incidence of WVCs. The average annual traffic volume on SR 64 was
4275 vehicles per day during the study period, but traffic levels at night were low,
averaging less than 10 vehicles per hour for a 4 hour period.
In a Final Wildlife Accident Reduction Study (1991–2003), the Arizona Department of
Transportation (ADOT) commissioned the development of a proactive assessment of
WVCs and potential mitigation measures to reduce the incidence of WVCs along
approximately 50 miles of SR 64 (185.5–235.4). This assessment (ADOT 2006)
recommended that nine passage structures be integrated into future highway
reconstruction of SR 64. It also recognized the need to conduct further field evaluation
and monitoring to determine the best locations for wildlife passage structures and the
extent of fencing needed to funnel animals to the structures.
The study called for assessing wildlife use of Cataract Canyon Bridge to determine
whether its design is conducive to wildlife passage. The assessment addressed the
potential barrier effect on pronghorn and recommended that this issue also be addressed
with monitoring. As a result of these recommendations, this research project was initiated
in 2007, with the following objectives:
 Assess elk (June 2007 through October 2009), mule deer (April 2008 through
October 2009), and pronghorn (January 2008 through January 2009) movements,
highway crossing patterns, and distribution relative to SR 64 and determine
permeability across the highway corridor.
 Investigate the relationships of elk, mule deer, and pronghorn highway crossing and
distribution patterns to SR 64 vehicular traffic volume (2007 through 2009).
 Investigate WVC patterns and relationships to elk, mule deer, and pronghorn
movement and highway crossing patterns in relation to SR 64 (2007 through 2009).
 Assess the degree to which Cataract Canyon Bridge is used by wildlife for below-grade
passage (July 2008 through December 2009).
 Develop recommendations to enhance elk, mule deer, and pronghorn highway
permeability along SR 64 through the application of wildlife passage structures and
ungulate-proof fencing.
4
MOVEMENTS AND PERMEABILITY
The research team determined highway crossings and calculated the crossing and passage
rates for elk, mule deer, and pronghorn using Global Positioning System (GPS)
telemetry. Passage rates served as the team’s relative measure of highway permeability,
calculated as the number of times animals crossed SR 64 in proportion to the number of
times animals approached to within 0.15 mi. The research team tracked 23 elk fitted with
GPS collars and accrued 107,055 GPS relocations. Elk crossed the highway 843 times, an
average of 0.12 times per day, with the highest proportion of crossings (60 percent)
occurring during the driest season (April–July).
Travel to limited water sources likely influenced movement and crossing patterns. The
elk passage rate averaged 0.44 crossings per approach, 52 percent lower than the rate
found during previous research on SR 260 sections with similar highway standards (Dodd
et al. “Evaluation of Measures,” 2007). The elk crossing distribution was not random and
exhibited several peak crossing zones, especially at the north end of the study area.
The research team tracked 11 mule deer fitted with collars and accrued 29,944 GPS fixes.
Deer crossed SR 64 550 times, an average of 0.26 times per day—twice as frequently as
elk. Seasonal deer crossings were more consistent than seasonal elk crossings, though
46 percent of crossings occurred during late summer and fall (August–November). The
average deer passage rate was 0.54 crossings per approach, which was higher than the
rate for elk. The mule deer crossing distribution did not occur in a random fashion. It
exhibited two peak crossing zones at the north end of the study area, with 92 percent of
the crossings occurring along a 3.2 mile stretch between Grand Canyon Airport and the
GCNP, in the vicinity of Tusayan.
The research team tracked 15 pronghorn with GPS collars that amassed 56,433 GPS
fixes. Only a single GPS-collared pronghorn crossed SR 64 (three times), for a crossing
rate average of 0.001 crossings per day. The mean pronghorn passage rate was a
negligible 0.004 crossings per approach, indicating that SR 64 is a near total barrier to
pronghorn passage. Pronghorn approached the highway 4269 times, and the distribution
was not random. The approach distribution exhibited three peaks along SR 64, with the
largest peak near the south boundary of the Kaibab National Forest, north of Valle.
TRAFFIC RELATIONSHIPS
In cooperation with ADOT, the research team measured traffic volume using a permanent
automatic traffic recorder. The pattern of elk and mule deer distribution with fluctuating
traffic was consistent with published models that indicated reduced “habitat
effectiveness” near the highway.
The use of habitat within 990 ft of the highway, as measured by probability of presence
of all three species, was clearly reduced at higher traffic volumes. The mean proportion
for the three species occurring within 990 ft of SR 64 dropped nearly in half, from 0.34 at
less than 100 vehicles per hr to 0.19 at 200 to 300 vehicles per hour. However, elk and
deer returned to areas within 330 ft of the highway in proportions greater than 0.12 when
5
traffic volumes were low. The impact to habitat effectiveness for these two species thus
was temporary.
The highest levels of permeability for elk and deer (passage rates greater than
0.70 crossings per approach) occurred at night when traffic was lowest. Pronghorn, on the
other hand, are diurnal and are active when traffic is heaviest. Along SR 64, pronghorn
uniformly avoided habitats adjacent to the highway (within 330 ft), thus reflecting a
permanent loss in habitat effectiveness.
Peak daytime traffic volumes along SR 64 approach 10,000 vehicles per day, a volume
at which highways become strong barriers to wildlife passage. Pronghorn appeared more
sensitive to traffic volume impact than elk and deer, and their avoidance of the area
adjacent to the highway is problematic in terms of implementing effective passage
structures to promote permeability.
WILDLIFE-VEHICLE COLLISION RELATIONSHIPS
The incidence of WVCs along SR 64 is a growing highway safety issue, with an increase
in collisions from that documented in the 2006 Final Wildlife Accident Reduction Study
report (36.7 per year) to 52.0 per year during this study (ADOT 2006). The research team
recorded 167 WVCs, with elk accounting for 59 percent of the accidents and mule deer
accounting for 35 percent. SR 64 sections on Kaibab National Forest lands at the north
and south ends of the study area had the highest incidence of elk and deer collisions,
though the collision rate on the north end was more than twice the rate on the south end
near I-40. No WVCs involving pronghorn were recorded during the study.
The spatial association between WVCs and GPS-determined crossings at the 1.0 mi scale
was significant for elk and mule deer. From 1998 through 2008, 42 percent of all single-vehicle
accidents in the study area involved wildlife, compared with the national average
of just 5 percent. On the five miles at the north end of the study area, wildlife-related
accidents accounted for more than 75 percent of all single-vehicle accidents.
The observed frequency of elk-vehicle collisions by time of day was different from our
expectations, with the highest proportion of elk collisions (50 percent) recorded during
evening hours. There was a negative association between elk-vehicle collisions and
traffic volume by hour. Deer-vehicle collisions also varied by time of day, with 49
percent of accidents recorded during the evening. Accidents during the morning and
midday, when traffic volume was highest, accounted for 43 percent of deer-vehicle
collisions.
There was a significant difference in the observed versus expected frequency of elk-vehicle
collisions by season. The driest season of the year, early spring–summer (April–
July), accounted for 43 percent of all elk-vehicle collisions; late summer–fall accounted
for another 38 percent. The association between elk collisions and mean monthly traffic
volume was significant, which was not the case for deer. For mule deer, the incidence of
collisions was relatively constant through much of the year, except for the late summer–
fall season (August–November), when nearly half of all collisions occurred. The
6
association between highway crossings and collisions by month was significant for elk
and mule deer.
Using nationally accepted cost estimates associated with elk and mule deer collisions,
and based on 2007–2009 WVCs, the annual cost associated with SR 64 vehicle collisions
is estimated to be $612,513 for elk and $162,168 for deer, or a total of $774,681 per year;
over 20 years, the total cost from WVCs would exceed $15.5 million (Huijser et al.
2007).
CATARACT CANYON BRIDGE WILDLIFE USE
To quantify wildlife use of Cataract Canyon Bridge, the research team employed single-frame
cameras in each of the four box-culvert cells; these self-triggering cameras
provided infrared nighttime illumination to record animals crossing through the bridge at
night. In total, 126 wildlife images were recorded by cameras, including 13 elk and 37
mule deer. In addition to wildlife, substantial human presence was documented at the
bridge, with a total of 191 humans and 29 all-terrain vehicles passing under the structure.
Of the limited number of elk and mule deer that approached the bridge, 92 percent and
89 percent of these species, respectively, crossed through the bridge cells. The majority
(89 percent) of deer use occurred from August through October. Elk use occurred only in
October and December, with no approaches the rest of the year. Of all deer and elk bridge
crossings, 64 percent occurred in the 4 hr period between 11:00 p.m. and 3:00 a.m..
Though the documented wildlife use of the bridge was nominal, the research team’s
expectation for significant use was also low because wildlife fencing to limit at-grade
crossings and funnel animals to the bridge could not be accomplished as hoped. Despite
the limited wildlife use recorded on the cameras, the research team nonetheless believes
that the bridge has the potential to be a highly effective retrofitted wildlife passage
structure due to the comparatively high rates of mule deer and elk that crossed through
with minimal behavioral resistance.
The bridge exceeds all recommended structural and placement guidelines for effective
elk and mule deer passage structures. The high level of human use should not
significantly limit effective wildlife use of the structure because wildlife use occurs in the
evening and nighttime hours; human use occurs during daylight hours.
IDENTIFICATION OF PASSAGE STRUCTURE SITES
The research team used elk and mule deer highway crossings, WVCs, pronghorn
approaches, and the proportions of animals crossing or approaching within each segment,
among other criteria, to rate 95 0.6 mile segments for suitability as potential passage
structure locations, this 57 mile area extends into GCNP and included all areas where elk,
mule deer and pronghorn approached the highway and ranged outside of the area defined
by ADOT (2006). Additional criteria included land ownership and topography that
would support passage structure construction. The ratings ranged from 1 to 33 points on a
40-point scale and averaged 10.0 points per segment. The research team’s ratings
7
identified 11 priority wildlife passage structure locations; the 0.6 mile segments with
these structures averaged a 20.7 rating.
Six sites were conducive to underpasses, and five were at sites where the terrain was
conducive to overpasses and would promote pronghorn permeability. Of the nine wildlife
underpass locations identified in the 2006 Final Wildlife Accident Reduction Study report,
the research team’s rating of potential passage structure sites corroborated that eight were
warranted.
In addition to the passage structure sites recommended in the 2006 Final Wildlife
Accident Reduction Study report, which were based largely on WVC records and sites
where the topography could support a structure, the team identified three additional
passage structures. One of these was an underpass at the Kaibab National Forest–GCNP
boundary, and the other two structures are overpasses recommended for pronghorn
passage in an area where no WVC was recorded during this study or documented in the
2006 Final Wildlife Accident Reduction Study report.
A variety of passage structure types can be considered for use along SR 64, including the
single-span bridges used effectively along SR 260, cost-effective multi-plate arch
underpasses, and pre-cast concrete arches. The 11 structures recommended by the team
are spaced 1.5 to 2.3 mi apart, with this spacing generally consistent with guidelines for
elk and pronghorn.
The situation for pronghorn is very different from that for elk and deer. For pronghorn,
fencing in association with passage structures is not needed to preclude at-grade
pronghorn crossings, but it is important in providing a visual cue as to a path across the
highway barrier, provided no fencing is used at the mouth of the passages. For
pronghorn, minimizing the impact of high daytime traffic may be more critical than
fencing, especially given pronghorn avoidance of the habitats adjacent to SR 64. A
comprehensive set of measures to reduce traffic-associated impact could create “quiet
zones” along the highway corresponding to passage structures and could facilitate
pronghorn permeability.
Wildlife fencing plays an integral role with passage structures in achieving objectives for
reducing WVCs, promoting highway safety, and improving wildlife permeability,
especially for elk and deer. Failure to erect adequate fencing in association with passage
structures, even when spaced adequately, has been found to substantially reduce their
effectiveness. The research team identified a 14.2 mile section of the highway where
fencing would be needed to meet WVC reduction and permeability objectives.
8
9
2.0 LITERATURE REVIEW
2.1 BACKGROUND
Direct and indirect highway impacts have been characterized as some of the most
prevalent and widespread forces altering ecosystems in the United States (Noss and
Cooperrider 1994, Trombulak and Frissell 2000, Farrell et al. 2002). Forman and
Alexander (1998) estimated that highways have affected more than 20 percent of the
nation’s land area through habitat loss and degradation.
It is estimated that as many as 1.5 million collisions involving deer occur annually in the
United States (Conover 1997). Wildlife-vehicle collisions (WVCs) cause human injuries,
deaths, and tremendous property loss (Reed et al. 1982, Schwabe and Schuhmann 2002).
More than 38,000 human deaths attributable to WVCs occurred in the United States from
2001 through 2005, and the economic impact exceeds $8 billion a year (Huijser et al.
2007). The most pervasive impacts of highways on wildlife, however, are the barrier and
fragmentation effects resulting in diminished habitat connectivity (Noss and Cooperrider
1994, Forman and Alexander 1998, Forman 2000).
Highways block animal movements between seasonal ranges or other vital habitats.
This barrier effect fragments habitats and populations, reduces genetic interchange
(Gerlach and Musolf 2000, Epps et al. 2005, Riley et al. 2006), and limits dispersal of
young (Beier 1995), all disrupting viable wildlife population processes. Long-term
fragmentation and isolation renders populations more vulnerable to the influences of
catastrophic events and may lead to extinctions (Hanski and Gilpin 1997). Fencing that
blocks wildlife and livestock access across highways without provisions for adequate
passage may exacerbate barrier effects.
Though numerous studies have alluded to highway barrier effects on wildlife
(e.g., Forman et al. 2003), relatively few have provided quantitative data relative to
animal passage rates. Most studies have focused on the efficacy of passage structures in
maintaining wildlife permeability, the ability of animals to pass across highways
(Clevenger and Waltho 2003, Ng et al. 2004). Assessments of highway fragmentation
effects on relatively small, less mobile mammals (Swihart and Slade 1984, Conrey and
Mills 2001, McGregor et al. 2003) have proved easier to accomplish than assessments for
larger, more mobile species that are limited by cost-effective techniques to measure
permeability.
Paquet and Callaghan (1996) used winter track counts adjacent to highways and other
barriers to determine passage rates by wolves. Very high frequency (VHF) radio
telemetry has also been used to assess wildlife movements and responses to highways,
often pointing to avoidance of highways and roads (Brody and Pelton 1989, Rowland
et al. 2000).
Only a limited number of studies have addressed permeability in an experimental
(e.g., before and after construction) context with research controls (Hardy et al. 2003;
Roedenbeck et al. 2007; Olsson 2007; Dodd et al., “Evaluation of Measures,” 2007).
10
Olsson (2007) documented an 89 percent decrease in the mean moose-crossing rate
between before- and after-reconstruction levels along a highway in Sweden. Dyer et al.
(2002) compared actual road to simulated road network crossing rates, where caribou
crossed actual roads less than 20 percent as frequently as simulated networks.
Dodd et al. (“Assessment of elk,” 2007) stressed the value of a quantifiable and
comparable metric of permeability. They calculated elk highway passage rates from
Global Positioning System (GPS) telemetry to conduct before-after-control
reconstruction comparisons along SR 260. Dodd et al. (“Effectiveness of Wildlife,” in
review) reported that overall elk (n = 100) passage rates averaged 0.50 crossings per
approach. Among reconstruction classes, the mean elk passage rate for the before-reconstruction
control class (0.67) was 39 percent higher than the mean after-reconstruction
passage rate (0.41). They also calculated white-tailed deer passage rates
along SR 260; the rates averaged only 0.03 crossings per approach on control sections.
On reconstructed sections with passage structures, the passage rate was significantly
higher (0.16 crossings per approach).
Along United States Route 89 (US 89), Dodd et al. (“Effectiveness of Wildlife,” in
review) used the same consistent methodology and found the mean pronghorn (n = 31)
passage rate to be negligible—0.006 crossings per approach. US 89 constitutes a near-total
barrier to pronghorn passage.
In addition to the permeability insights gained from the previously discussed GPS
telemetry studies, the SR 260 and US 89 studies furthered the understanding that traffic
volume plays in the highway barrier effect. Theoretical models (Mueller and Berthoud
1997) suggest that highways averaging 4000 to 10,000 vehicles per day present strong
barriers to wildlife and would repel animals from the highway. Gagnon et al. (“Traffic
volume alters,” 2007) found that increasing vehicular traffic volume decreased the
probability of at-grade crossings by elk, which shifted their distribution away from the
highway with increasing traffic volume, consistent with Mueller and Berthoud (1997) and
Jaeger et al. (2005).
For white-tailed deer, Dodd and Gagnon (2011) found that at-grade SR 260 passage rates
were consistently low (fewer than 0.1 crossing per approach) across all traffic volumes.
Pronghorn also were consistently negatively impacted by traffic volume, even at low
levels, and distribution remained constant among all distances from US 89 and across all
traffic volumes up to 500 vehicles per hr (Dodd et al. “Effectiveness of Wildlife,” in
review). Whereas elk and deer highway crossings occur at night when traffic volume is
lowest, pronghorn are diurnal and active when traffic volumes are typically at their
highest (Gagnon et al. “Traffic volume alters,”2007) contributing to their low
permeability.
Collectively, these Arizona studies using consistent, comparable methodologies and
metrics have added substantially to the understanding of highway impact to wildlife
permeability and traffic volume relationships for multiple species and highways
exhibiting different traffic patterns. This understanding will further benefit from
continued studies that assess permeability for additional species and on highways that
11
expand the range of experimental conditions under which permeability is assessed
(Jaeger et al. 2005).
Numerous assessments of WVC patterns have been conducted, most focusing on deer
(Reed and Woodward 1981, Bashore et al. 1985, Romin and Bissonette 1996, Hubbard et
al. 2000). Only recently have WVC assessments specifically addressed elk-vehicle
collision patterns (Gunson and Clevenger 2003, Biggs et al. 2004, Dodd et al.
“Effectiveness of Wildlife,” in review, Gagnon et al. 2010). Insights gained from such
assessments have been instrumental in developing strategies to reduce WVCs
(Romin and Bissonette 1996, Farrell et al. 2002), including planning passage structures
to reduce at-grade crossings and to maintain permeability (Clevenger et al. 2002).
Consistent tracking of WVCs constitutes a valuable tool to assess the impact of highway
construction (Romin and Bissonette 1996) and the efficacy of passage structures and
other measures (e.g., fencing) in reducing WVCs (Reed and Woodard 1981, Ward 1982,
Clevenger et al. 2001, Dodd et al., “Evaluation of Measures,” 2007).
Increasingly, structures designed to promote wildlife passage across highways are being
implemented throughout North America, especially large bridges (e.g., underpasses or
overpasses) designed specifically for large animal passage (Clevenger and Waltho 2000,
Bissonette and Cramer 2008). Whereas early passage structures were typically
approached as single-species mitigation measures to address WVCs (Reed et al. 1975),
the focus today is more on preserving ecosystem integrity and landscape connectivity
benefiting multiple species (Clevenger and Waltho 2000).
Transportation agencies are increasingly receptive to integrating passage structures into
highways to address safety and ecological needs (Farrell et al. 2002). At the same time,
there is increasing expectation that such structures will benefit multiple species and
enhance connectivity (Clevenger and Waltho 2000), and that scientifically sound
monitoring of wildlife response to these measures will occur to improve effectiveness
(Clevenger and Waltho 2003, Hardy et al. 2003). Corlatti et al. (2009) argued for long-term
monitoring of wildlife passages to evaluate their effectiveness in maintaining
connectivity and promoting population and genetic viability, thus justifying their
relatively high cost.
Wildlife passage structures have indeed shown benefit in promoting wildlife passage for
a variety of wildlife species (Farrell et al. 2002; Clevenger and Waltho 2003; Dodd et al.,
“Assessment of elk,” 2007; Gagnon et al. 2011). Dodd et al. (“Evaluation of Measures,”
2007) found that elk passage rates along one section of SR 260 increased 52 percent to
0.81 crossings per approach once reconstruction was completed and ungulate-proof
fencing linking passage structures was erected. This pointed to the efficacy of passage
structures and fencing in promoting permeability, as well as achieving an 85 percent
reduction in elk-vehicle collisions (Dodd et al., “Evaluation of Measures,” 2007).
Gagnon et al. (“Effects of traffic,” 2007) found that traffic levels did not influence elk
passage rates during below-grade underpass crossings. This finding shows the benefit
of underpasses and fencing in promoting permeability by funneling elk to underpasses
12
where traffic has minimal effect compared with crossing at-grade during high traffic
volumes (Gagnon et al. “Traffic volume alters,”2007). The fivefold higher white-tailed
deer permeability along SR 260 after reconstruction with passage structures compared
with controls suggests the efficacy of passage structures; like elk, deer passage rates were
minimally affected by traffic on sections where passage structures facilitated below-grade
passage (Dodd and Gagnon 2011).
Structural characteristics and placement of wildlife crossing structures are important to
maximizing wildlife use (Reed et al. 1975; Foster and Humphrey 1995; Clevenger and
Waltho 2000, 2003; Dodd et al., “Evaluation of Measures,” 2007; Gagnon et al. 2011).
Prior studies modeled structural factors accounting for differences in wildlife use
(Clevenger and Waltho 2000, 2005; Ng et al. 2004). Gagnon et al. (2011) assessed five
factors, of which structural design and placement characteristics had the greatest
influence on elk use of SR 260 underpasses. However, given sufficient time for
habituation, most structures became equally effective for elk, even in spite of structural
or placement limitations.
The spacing between passage structures is also an important consideration (Bissonette
and Adair 2008). Dodd et al. (“Effectiveness of Wildlife,” in review) and Gagnon et al.
(2010) found considerable variation in mean elk passage rates (ranging from 0.09 to
0.81 crossings per approach) on three reconstructed SR 260 sections, likely reflecting the
corresponding variation in passage structure spacing ranging from 1.5 to 0.6 miles, with a
strong negative association with increased distance between structures (r = -0.847;
Dodd et al. “Effectiveness of Wildlife,” in review). Bissonette and Adair (2008) assessed
recommended passage structure spacing for several species tied to isometric scaling of
home ranges to define appropriate passage structure spacing distance. They
hypothesized that when used with other criteria this approach will help maintain
landscape permeability for a range of species.
Most assessments of wildlife passage structure use have been for newly constructed
structures implemented as part of major highway reconstruction projects (Clevenger and
Waltho 2000, 2005; Gagnon et al. 2011). However, some assessments have been of
primarily drainage structures retrofitted to serve as wildlife passage structures with the
erection of fencing to limit at-grade crossings and funnel animals to structures
(Gordon and Anderson 2003, Ng et al. 2004).
In Arizona, such retrofitting has considerable promise as a cost-effective approach to
minimizing WVCs and promoting permeability (Gagnon et al. 2010), particularly
compared with costly highway reconstruction that may not occur on some highways for
decades. As such, there is a need to better understand the potential effectiveness of
existing structures for retrofitting applications, including structural design characteristics
that may limit effectiveness.
2.2 RESEARCH JUSTIFICATION
The incidence of WVCs along SR 64 between I-40 and GCNP is a significant and
growing concern. In the future, this predominantly two-lane highway will be
13
reconstructed to a four-lane divided highway to address growing traffic volume and the
incidence of WVCs.
To help address the WVC issue, ADOT commissioned the development of a proactive
assessment of WVCs and potential mitigation measures to reduce their incidence along
SR 64. In ADOT (2006) it was reported that 48 percent of 475 accidents recorded along
SR 64 in the five-year period from October 1998 through September 2003 involved
collisions with wildlife, primarily elk and mule deer (Table 1). This study developed and
evaluated alternatives and associated mitigation measures for consideration in the
planned feasibility study for the eventual reconstruction of SR 64.
Table 1. Vehicle Accidents Involving Collisions with Elk and Mule Deer
along SR 64 from 1991 through 2003, Including the Mean Number
of Collisions (per Year and per Mile).
SR 64
section
Elk-vehicle accidents Mule deer-vehicle
accidents
Mileposts Total
Mean
(per
year)
Mean
(per
mile)
Total
Mean
(per
year)
Mean
(per
mile)
A 185.5–204.7
(19.2 mi) 58 4.5 3.0 79 6.1 4.1
B
204.7–212.5
(7.8 mi)
2 0.2 0.3 2 0.2 0.3
C
212.5–214.3
(1.8 mi)
0 0.0 0.0 1 0.1 0.6
D 214.3–223.4
(9.1 mi) 6 0.5 0.7 3 0.2 0.3
E 223.4–235.4
(12.0 mi) 97 7.5 8.1 238 18.3 19.8
All 185.5–235.4
(49.9 mi) 163 12.5 3.3 315 24.2 6.3
Source: Final Wildlife Accident Reduction Study (ADOT 2006)
The earlier study report (ADOT 2006) delineated five SR 64 sections (A–E) based on
land ownership and habitat (Figure 1). This study developed two mitigation alternatives
for three of the sections (Table 2) to address the past incidences of WVCs, including the
construction of as many as seven wildlife underpasses and three overpasses, depending
on the selected alternatives (Table 2; Figure 1). None of the passage structures were
recommended along highway sections where American pronghorn were a focus (Sections
B and D), partly because no WVC involving this species was recorded from 1991
through 2003. However, SR 64 likely constitutes a significant barrier to pronghorn
passage similar to US 89 to the east, where no pronghorn-vehicle collisions were
recorded either (Dodd et al. 2011).
14
Table 2. SR 64 Sections and Mileposts with Proposed Wildlife Mitigation Measures for
Focal Wildlife Species Identified in the 2006 Final Wildlife Accident Reduction Study.
SR 64
section Mileposts
Mitigation
alternative
Proposed wildlife
mitigation measures Focal
wildlife
Underpass Overpass Fencing species
A 185.5–204.7
(19.2 mi)
A(W)-1
A(W)-2
2a
1a
0
1
Yes
Yes
Elk, mule
deer
B 204.7–212.5
(7.8 mi)
None 0 0 No Pronghorn
C 212.5–214.3
(1.8 mi)
None 0 0 No
None;
human
develop-ment
D 214.3–223.4
(9.1 mi)
D(W)-1
D(W)-2
1
0
0
1
Yes
Yes
Elk,
mule deer,
pronghorn
E 223.4–235.4
(12.0 mi)
E(W)-1
E(W)-2
4
4
0
1
Yes
Yes
Elk,
mule deer
a Includes Cataract Canyon Bridge, which will be used as a passage structure
The same study (ADOT 2006) identified the need to conduct further field evaluation and
monitoring to determine the best locations for wildlife passage structures and the extent
of ungulate-proof fencing needed to funnel animals to passage structures. The report
indicated that the focus of such monitoring should be from mile post (MP) 222.0 to MP
235.4, where the highest incidence of WVCs involving elk and mule deer has occurred in
the past.
The report called for the monitoring of current and potential (e.g., with added funnel
fencing) wildlife use of Cataract Canyon Bridge at MP 187.3; Section A to determine
whether this multiple box culvert design is conducive to wildlife passage.
The report also addressed the potential barrier effect to pronghorn (especially along
Section B) and recommended that this issue also be further evaluated with monitoring. In
that report it was also recommended that a cooperative research project between ADOT
and the AGFD be initiated in advance of the feasibility study and final design for
highway reconstruction such that refined, site-specific information can be incorporated
into the final reconstruction plans.
15
Figure 1. Landownership, Mileposts, 0.1 mi Segments, Highway Sections A–E,
and Preliminary Wildlife Passage Structures in the SR 64 Study Area Identified by the
Final Wildlife Accident Reduction Study (ADOT 2006).
16
In 2007, an interagency agreement between ADOT and the AGFD was executed for the
SR 64 research project (Project JPA07-026T), with funding provided by the ADOT
Research Center. This research project is significant from several perspectives.
The cited study (ADOT 2006) conducted for SR 64 represents the first assessment of its
type in Arizona, forming the proactive basis from which to develop strategies to mitigate
WVCs and obtain refined information with further monitoring and research.
The project also reflects the incremental process in addressing wildlife connectivity and
permeability needs embodied in Arizona’s Wildlife Linkages Assessment (Arizona
Wildlife Linkages Workgroup 2006). General connectivity needs identified in the
assessment (e.g., Linkage No. 12; Coconino Plateau–Kaibab National Forest) were also
addressed in the 2006 Final Wildlife Accident Reduction Study, which called for further
monitoring to assess site-specific needs and refined strategies for promoting permeability.
2.3 RESEARCH OBJECTIVES
Pursued largely as a result of the 2006 Final Wildlife Accident Reduction Study, this
research project will add considerably to the understanding of wildlife movements in
relation to highways and provide information to support data-driven design planning for
the planned reconstruction of SR 64. Focusing on elk, mule deer, and pronghorn, this
research project complements previous research on wildlife-highway permeability, traffic
volume, and WVC relationships (Dodd et al. “Effectiveness of Wildlife,” in review,
2011; Gagnon et al. “Traffic volume alters,” 2007). The specific research objectives of
this research project were to:
 Assess elk (June 2007 through October 2009), mule deer (April 2008 through
October 2009), and pronghorn (January 2008 through January 2009) movements,
highway crossing patterns, and distribution relative to SR 64 and determine
permeability across the highway corridor.
 Investigate the relationships of elk, mule deer, and pronghorn highway crossing and
distribution patterns to SR 64 vehicular traffic volume (2007 through 2009).
 Investigate WVC patterns and relationships to elk, mule deer, and pronghorn
movement and highway crossing patterns in relation to SR 64 (2007 through 2009).
 Assess the degree to which Cataract Canyon Bridge is used by wildlife for below-grade
passage (July 2008 through December 2009).
 Develop recommendations to enhance elk, mule deer, and pronghorn highway
permeability along SR 64 through the application of wildlife passage structures and
ungulate-proof fencing.
17
3.0 STUDY AREA
SR 64 is the highway connecting I-40 to Grand Canyon National Park (GCNP). It is
classified as a rural principal arterial highway. The focus of this research project was a
57 mile stretch of highway starting at I-40, approximately 2 miles east of Williams (MP
185.5), and ending at the GCNP boundary (MP 237.0) just north of the community of
Tusayan, Coconino County, Arizona (latitude 35º25'–35º99'N, longitude 112º12'–
112º15'W; Figure 1). SR 64 runs north–south and intersects US 180 at Valle (MP 213.5);
US 180 links SR 64 to Flagstaff, 40 miles to the southeast. The majority of SR 64 now is
a two-lane highway, with occasional passing lanes.
3.1 NATURAL SETTING
The study area is at the southwest extent of the Colorado Plateau physiographic province.
The south half of the study corridor lies within the San Francisco Peaks Volcanic Field
(Hansen et al. 2004). The study corridor adjacent to SR 64 varies in elevation from
6000 ft between Red Lake and Valle to 6930 ft at the south end of the study area near
Kaibab Lake, and 6600 ft elevation at Grand Canyon Airport in Tusayan. The topography
is a mix of mesas, cinder cones, and broken terrain with rolling hills, ridges, and valleys
interspersed with large, relatively flat grassland areas (Figures 1 and 2).
Land ownership adjacent to the highway includes U.S. Forest Service (Kaibab NF) lands
(35 percent of the corridor), including 5 miles at the south end and 13 miles at the north
end of the study area (Figure 1). In between, land ownership is a mix of interspersed
Arizona State Trust (25 percent) and private lands (40 percent), with much of the private
land subdivided for development (Figure 1). Existing development is concentrated near
the communities of Red Lake, Valle, and Tusayan (Figure 1).
3.1.1 Climate
Generally, the climate is characterized as semiarid, dominated by hot summers and cool
winters. At the south and north ends of the study area, near Williams and the Grand
Canyon, respectively, the average maximum temperature is 64 °F, with July being the
warmest month (mean = 84 °F); highs can approach 95 °F. Winter daily low temperatures
average 35 °F at Williams and 32 °F at the Grand Canyon, with January being the coolest
month (mean = 19 °F); winter lows at the Grand Canyon often dip below 0 °F.
Precipitation varies considerably along the length of the study area, with Williams
averaging 21.6 inches annually, including an average snowfall accumulation of
69.3 inches. Precipitation drops off to the north along SR 64 at Valle, where it averages
only 9.4 inches annually, with 4.8 inches of annual snow accumulation. Precipitation at
Tusayan (Grand Canyon Airport) is greater than at Valle but is still less than the south
portion of SR 64; annual precipitation averages 13.7 inches, with annual snowfall of
44.3 inches. Precipitation occurs primarily during intense and localized thunderstorms
associated with the summer monsoon and more widespread frontal storms that pass
through the state in the winter.
18
Figure 2. Great Basin Conifer Woodland Adjacent to SR 64 (Top) with Open to Dense
Stands of Pinyon and Juniper and Cliffrose, Apache Plume, and Other Shrubs, and Plains
and Great Basin Grasslands (Bottom) Dominated by Blue and Black Grama, Galleta, and
Needle-and-Thread Grasses.
19
3.1.2 Vegetation
Vegetation is diverse and exhibits characteristics of the Petran Montane Conifer Forest,
the Great Basin Conifer Woodland, and the Plains and Great Basin Grassland biotic
communities (Brown 1994). Ponderosa pine dominates montane coniferous forests at the
southernmost and northernmost portions of the study areas. Gambel oak occurs in the
overstory. Big sagebrush, rabbitbrush, and cliffrose dominate the understory. Sagebrush
is particularly prevalent at the drier north extent of the study area.
The ponderosa pine–dominated forest adjacent to SR 64 is interspersed with small
openings in draws and flats vegetated by sagebrush, blue grama, and other grasses.
Though these sites are dry, they nonetheless may correspond to WVC incidence hotspots
as observed for wet meadow habitats adjacent to SR 260 by Dodd et al. (“Evaluation of
Measures,” 2007) and Manzo (2006). Sparse to dense pinyon and one-seed juniper
dominate the overstory of extensive Great Basin conifer woodlands, with sagebrush,
cliffrose, Apache plume, and other shrubs in the understory, along with blue grama and
other grasses in openings (Figure 2). Conifer woodlands transition to Plains and Great
Basin grasslands dominated by blue and black grama, galleta, and needle-and-thread
grasses, with winterfat and sage interspersed with sparse junipers (Figure 2).
3.1.3 Wildlife Species
The focal species of this study were elk, mule deer, and pronghorn. SR 64 separates
Game Management Units (GMUs) 7 and 10 south of Valle and bisects GMU 9 between
Valle and Tusayan. Elk are found in moderate densities at the south end of the study area
and at high densities at the north end of the study area, with low densities in between
(Figure 3). During the project (2007 through 2009), the AGFD surveyed an average of
424 elk in GMU 9, with a robust ratio of bulls in relation to cows and calves (39 bulls :
100 cows : 32 calves). In GMU 7, an average of 355 elk were surveyed annually
(22 bulls: 100 cows : 39 calves).
At the north extent of the study area, mule deer are commonly seen along the highway
corridor and occur in high densities; they occur in moderate densities at the south extent
and in low densities in between (Figure 3). During the study, an average of 303 mule deer
were surveyed by the AGFD each year (17 bucks : 100 does : 42 fawns). In GMU 7,
extending far east of SR 64, 151 deer were classified in surveys each year (25 bucks : 100
does : 37 fawns).
The pronghorn population levels in grassland and open woodland areas of GMUs 9 and 7
are considered average relative to other northern Arizona populations. However, pockets
of habitat east of SR 64 in GMUs 7 and 9 hold high densities of pronghorn (Figure 3),
and the population on the east side is larger than the population on the west. In GMU 9,
an average of 119 animals were surveyed each year during this project (47 bucks : 100
does : 26 fawns). For GMU 7, an average of 247 pronghorn were surveyed each year
(31 bucks : 100 does : 25 fawns).
20
Figure 3. Density Distributions for the Three Target Species of Research along SR 64:
Elk (Left), Mule Deer (Center), and Pronghorn (Right).
21
3.1.4 Cataract Canyon Bridge
Reflective of the generally rolling terrain along the highway corridor, Cataract Canyon
Bridge at MP 187.3 near Kaibab Lake is one of the most substantial bridge structures
along the 57 mi study area (Figure 4). This 44 ft wide reinforced concrete box-culvert
bridge was constructed in 2001, with four 26 ft spans, for a total length of 104 ft. The
bridge cells have a 16 ft vertical clearance. Due to the moderate to high elk numbers and
moderate mule deer density in the vicinity of this bridge, there was an opportunity to
evaluate existing wildlife use to document and better understand the efficacy of this
structure type to serve as an effective wildlife crossing structures without fencing.
Dodd et al. (“Evaluation of Measures,” 2007) stressed the need for ungulate-proof funnel
fencing to guide animals toward passage structures to achieve desired wildlife use.
Without fencing, animals continue to cross the highway at grade. Though ADOT was
amenable to such fencing near Cataract Canyon Bridge and conducted a formal analysis,
there were too many concerns to move forward with fencing to coincide with this
research project. One of the foremost concerns related to addressing potential “end run”
or forcing of elk to another location along SR 64 at the termini of the fencing, including
immediately adjacent to I-40. As such, no fencing was erected in association with this
bridge during the project.
Figure 4. Cataract Canyon Bridge on SR 64.
22
3.2 TRAFFIC VOLUME
Average annual daily traffic (AADT) volume on SR 64 was measured at 4343 vehicles
per day in 2008 and 4208 vehicles per day in 2009, or an average of 4275 vehicles per
day. Since late 2007, traffic volume has been continuously measured by a permanent
automatic traffic recorder (ATR) installed near Valle. Traffic volumes were highest
during daytime hours (Figure 5).
Compared with other study areas in central and northern Arizona, including SR 260
(Dodd et al. “Effectiveness of Wildlife,” in review), US 89 (Dodd et al. 2011), and
Interstate 17 (Gagnon et al. “Elk Movements Associated,�� in review), SR 64 is unique in
that traffic is virtually nonexistent during the late nighttime hours, averaging less than
10 vehicles per hr for a 4 hr period (Figure 5). This reflects the predominant tourist
destination nature of motorists traveling SR 64 to and from GCNP and not using this
route for regional or interstate travel beyond GCNP at night. This is also reflected in the
relatively small proportion of commercial vehicles traveling SR 64—only 0.07 percent
during the day and 0.05 percent at night—compared with SR 260, where commercial
traffic exceeded 40 percent at night (Dodd et al. “Effectiveness of Wildlife,” in review),
and US 89, where commercial vehicles made up a third of the nighttime traffic (Dodd et
al. 2011).
Peak annual traffic volume along SR 64 between 10:00 and 17:00 averaged 355 vehicles
per hr, equivalent to an AADT volume of 8800 vehicles per day, though in summer it was
considerably higher. Mean monthly traffic volume was highest during summer (June–
August; 5710 AADT), when it was three times higher than volume during the lowest
traffic months of December–February (1775 AADT). Traffic volume was highest on
Saturdays—nearly 20 percent higher than during the rest of the week.
23
Figure 5. Hourly Traffic Volume (Vehicles per Hour) along SR 64, Arizona, from 2007
through 2009. Note the Low Volume of Traffic during Nighttime Hours (00:00–04:00).
24
25
4.0 METHODS
4.1 WILDLIFE CAPTURE, GPS TELEMETRY, AND DATA ANALYSIS
4.1.1 Elk Capture
The research team captured elk at 14 sites adjacent to SR 64, along Highway Sections A
and E, at the north and south extremes of the study area; 11 sites were concentrated
adjacent to Section E. Elk were trapped primarily in net-covered Clover traps (Clover
1954) baited with salt and alfalfa hay; all traps were within 0.5 mi of the highway
corridor and near permanent water sources (Figure 6).
Elk were also captured by nighttime darting from a vehicle along the highway aided with
a spotlight. The low volume of traffic late at night made such capture possible. Trapped
animals were physically restrained, and all animals were blindfolded, ear-tagged, and
fitted with global positioning system (GPS) receiver collars (Figure 6). Darted elk were
administered a reversal drug when handling was complete. Elk were instrumented with
Telonics, Inc., Model TGW-3600 store-on-board GPS collars programmed to receive a
GPS relocation fix every two hours. All collars had very high frequency (VHF) beacons,
mortality sensors, and programmed release mechanisms to allow recovery. Battery life
for the GPS units was approximately two years.
4.1.2 Mule Deer Capture
The research team attempted to trap mule deer in small Clover traps baited with sweet
feed and salt at five sites adjacent to Section E but had no success. Therefore, the team
shifted to nighttime darting from a vehicle along or immediately adjacent to SR 64, aided
by a spotlight. Again, the low volume of traffic late at night made such capture feasible.
Deer were blindfolded, ear-tagged, and fitted with GPS receiver collars (Figure 6), and
then administered a reversal drug. Most deer were fitted with Telonics, Inc. Model TGW-
3500 GPS receiver collars (Figure 6) programmed to receive a fix every two hours, with a
battery life of 11 months. Four deer were fitted with Telonics Generation IV GPS
receiver collars that had a one-year battery life. All collars had VHF mortality sensors
and programmed release mechanisms for recovery.
4.1.3 Pronghorn Capture
The research team captured pronghorn using a net gun fired from a helicopter (Firchow et
al. 1986, Ockenfels et al. 1994, Dodd et al. 2011; Figure 6). A fixed-wing aircraft and
numerous ground spotters using optics equipment were employed to search for pronghorn
during capture to minimize helicopter searching. Pronghorn were captured during winter
to minimize heat-related stress on animals as well as deleterious effects on females that
could occur if captured later in their pregnancies.
26
Figure 6. Photographs of Capture Techniques Used for Elk, Mule Deer and Pronghorn
along SR 64 and GPS-Collared Animals: Elk Captured with Net-Covered Clover Trap
(Top), Darting to Immobilize Mule Deer (Middle), and Net-Gunning of Pronghorn from a
Helicopter (Bottom).
27
The team’s capture objectives were to:
 Instrument as nearly an equal number of pronghorn on each side of SR 64 as possible
because the research team suspected that SR 64 would be a barrier similar to US 89
(Dodd et al. 2011).
 Spread the collars among as many different herds as possible along the length of the
study area.
 Capture animals within 2 mi of SR 64.
Upon capture, pronghorn were immediately blindfolded and untangled from the capture
net. Animals were fitted with a GPS collar and marked with a numbered, colored ear tag
(Figure 6). Tissue samples were taken from animals’ ears with a paper punch and
preserved for future genetic analysis. The research team instrumented pronghorn with
Telonics, Inc. Model TGW-3500 store-on-board GPS receiver collars programmed to
receive 12 GPS fixes per day, with one fix every 90 minutes between 04:00 and 22:00.
GPS units had a battery life of 11 months. All collars had VHF beacons, mortality
sensors, and programmed release mechanisms to allow recovery.
4.1.4 GPS Analysis of Animal Movements
Once GPS collars were recovered and data downloaded, the research team employed
ArcGIS® Version 8.3 Geographic Information System software (ESRI®, Redlands,
California) to analyze GPS data similar to analyses done for elk by Dodd et al.
(“Assessment of elk,” 2007), white-tailed deer (Dodd and Gagnon 2011), and pronghorn
(Dodd et al. 2011). The team calculated individual minimum convex polygon (MCP;
connecting the outermost fixes) home ranges composed of all GPS fixes (White and
Garrott 1990). Differences in means were assessed by analysis of variance, and means
were reported with ±1 standard error (SE).
Crossings were compared among the following seasons:
 Late spring–summer (April–July).
 Late summer–fall (August–November).
 Winter–early spring (December–March).
4.1.5 Calculation of Crossing and Passage Rates
The team divided the entire length of SR 64 from I-40 to the Grand Canyon village into
570 sequentially numbered 0.1 mi segments corresponding to the units used by ADOT
for tracking WVCs and highway maintenance, and identical to the approach used by
Dodd et al. (“Evaluation of Measures,” 2007, Figure 1). The number and proportion of
GPS fixes within 0.15, 0.30, and 0.60 mi of SR 64 were calculated for each animal.
28
To determine highway crossings, the team drew lines connecting all consecutive GPS
fixes. Highway crossings were inferred where lines between fixes crossed the highway
through a given segment (Dodd et al. “Evaluation of Measures,” 2007, Figure 7). Animal
Movement ArcView Extension Version 1.1 software (Hooge and Eichenlaub 1997) was
used to assist in animal crossing determination. The research team compiled crossings by
individual animal by highway segment, date and time, and calculated crossing rates for
individual elk, mule deer, and pronghorn by dividing the number of crossings by the days
a collar was worn.
The research team calculated passage rates for collared animals, which served as its
relative measure of highway permeability (Dodd et al., “Evaluation of Measures,” 2007).
An approach was considered to have occurred when an animal traveled from a point
outside the 0.15 mi buffer zone to a point within 0.15 mi of SR 64, determined by
successive GPS fixes (Figure 7). The approach zone corresponded to the road-effect zone
associated with traffic-related disturbance (Rost and Bailey 1979, Forman et al. 2003)
previously used for elk and white-tailed deer by Dodd et al. (“Evaluation of Measures,”
2007, Dodd and Gagnon 2011). Animals that directly crossed SR 64 from a point beyond
0.15 mi were counted as an approach and a crossing.
The research team calculated passage rates as the ratio of recorded highway crossings to
approaches. The research team tested the hypothesis that the observed spatial crossing
distribution among 0.10 mi segments did not differ from a discrete randomly generated
distribution using a Kolmogorov-Smirnov test (Clevenger et al. 2001; Dodd et al.,
“Assessment of elk,” 2007), a test that is sensitive to the difference in ranks and shape of
the distributions. The team used linear regression to assess the strength of associations
between passage rates and traffic volume, as well as between the frequency of highway
crossings and WVCs at the MP (1.0 mi) scale using WVC records from 1991 through
2009 on those highway stretches with elk or deer crossings.
29
Figure 7. GPS Locations and Lines between Successive Fixes to Determine Highway
Approaches and Crossings in 0.10 mi Segments. The Expanded Section Shows GPS
Locations and Lines between Successive Fixes to Determine Approaches to the Highway
(Shaded Band) and Crossings. Example A Denotes an Approach and Crossing; Example
B Denotes an Approach without a Crossing.
4.1.6 Calculation of Pronghorn Approaches
Based on previous northern Arizona pronghorn highway telemetry research (Ockenfels et
al. 1997, van Riper and Ockenfels 1998, Bright and van Riper 2000, Dodd et al. 2011),
the research team anticipated few pronghorn crossings or approaches to within 0.15 mi
associated with crossings. As such, the team used the number of approaches by
pronghorn to within 0.30 mi to determine the distribution of animals adjacent to SR 64
for the purposes of assessing the need for, and potential location(s) of, passage structures,
as was done for the US 89 study (Dodd et al. 2011). Use of this greater approach distance
also was deemed appropriate given the relatively open nature of pronghorn habitat,
pronghorn reliance on visual stimuli in risk avoidance (Gavin and Komers 2006), and
pronghorn mobility over long distances (Yoakum and O’Gara 2000) compared with other
ungulates.
Pronghorn highway approaches were determined for animals approaching from each side
of SR 64 and both sides combined. The research team tested the hypothesis that the
A
B
30
observed spatial approach distribution among 0.10 mi segments did not differ from a
discrete randomly generated approach distribution using a Kolmogorov-Smirnov test
(Dodd et al. 2011).
4.1.7 Calculation of Weighted Crossings and Approaches
To account for the number of individual elk, deer, and pronghorn that crossed (and
approached, in the case of pronghorn) each highway segment adjacent to SR 64, as well
as evenness in crossing frequency among animals, the research team calculated Shannon
diversity indexes (SDIs; Shannon and Weaver 1949) for each segment using this formula:
Thus, to calculate SDI (or H  ) for each highway segment, the researchers calculated and
summed all the -(pi ln pi) for each animal that had approaches in the segment, where each
pi is defined as the number of individual collared elk and deer crossings and approaches
for pronghorn within each segment divided by the total number of respective crossings or
approaches in the segment. SDI were used to calculate weighted crossing or approach
frequency estimates for each segment, multiplying uncorrected crossings or approach
frequency by SDI. Weighted highway crossings and approaches better reflected the
number of crossing and approaching animals and the equity in distribution among elk,
deer, and approaching pronghorn (Dodd et al., “Evaluation of Measures,” 2007).
4.2 TRAFFIC VOLUME AND ANIMAL DISTRIBUTION RELATIONSHIPS
The research team measured traffic volume using a permanent automatic traffic recorder
(ATR) programmed to record hourly traffic volumes. ADOT’s Data Team provided data
collected from the ATR immediately north of Valle.
The research team examined how the proportion of elk, deer, and pronghorn GPS
relocations at different distances from the highway varied with traffic volume by
calculating the proportion of fixes in each 330 ft distance band, out to a maximum of
3300 ft. As done for elk (Gagnon et al. “Effects of traffic,” 2007), white-tailed deer
(Dodd and Gagnon 2011), and pronghorn (Dodd et al. 2011), the research team combined
traffic and GPS data by assigning traffic volumes for the previous hour to each GPS
location using ArcGIS Version 9.1 and Microsoft® Excel® software.1 This allowed the
team to correlate the traffic volume each animal experienced in the hour prior to
movement to a particular point, regardless of distance traveled.
To avoid bias due to differences in the number of fixes for individual animals, the
proportion of fixes occurring in each distance band for each animal was used as the
sample unit, rather than total fixes. The research team calculated a mean proportion of
animals and fixes for all animals within each 330 ft distance band at varying traffic
1 Microsoft and Excel are either registered trademarks or trademarks of Microsoft Corporation in the
United States and/or other countries.
31
volumes: less than 100, 101–200, 201–300, 301–400, 401–500, and 501–600 vehicles per
hr (Gagnon et al. “Effects of traffic,” 2007).
The team compared elk, mule deer, and pronghorn distribution and highway impact along
SR 64, and compared the species-specific distributions to those for elk (Gagnon et al.
“Effects of traffic,” 2007) and white-tailed deer on SR 260 (Dodd and Gagnon 2011) and
pronghorn on US 89 (Dodd et al. 2011). The team also assessed and compared species-specific
highway passage rates by time of day and associated traffic volume and used
linear regression to assess the strength of associations between WVCs and traffic volume.
4.3 WILDLIFE-VEHICLE COLLISION RELATIONSHIPS
The research team documented the incidence of WVCs along all SR 64 sections using
two methods. First, the research team relied on the submission of forms by agency
personnel, primarily DPS highway patrol officers, to determine the incidence of WVCs
during the study. DPS patrol officers made a concerted effort to record the species and
sex of animals involved in WVCs where such could be determined. These records were
augmented by regular searches of the highway corridor for evidence of WVCs by
research personnel.
The database compiled from the consolidated (non-duplicate) records included the date,
time, and location (to the nearest 0.1 mi) of the WVCs, the species involved, and the
reporting agency. WVC records were compiled and summarized by highway section by
year. Where duplicate reports of WVCs were made by DPS and research team searches,
the locations were compared to determine their accuracy (Barnum 2003, Gunson and
Clevenger 2003).
The research team used a database compiled by the ADOT Traffic Records Section
which includes DPS accident reports to determine the proportion of single-motor vehicle
accidents that involved wildlife along the respective highway sections. Huijser et al.
(2007) reported that nearly all WVCs are single-vehicle crashes. The research team
compared WVC incidence by season (using the same seasons for highway crossings),
month, day, and time (2 hr intervals), and used chi-square tests to compare observed
versus expected WVC frequencies.
4.4 WILDLIFE USE OF CATARACT CANYON BRIDGE
To quantify wildlife use of Cataract Canyon Bridge, the research team employed
Reconyx™ professional model single-frame cameras installed within each of the four
box-culvert cells (Figure 8). The encased cameras were mounted on wood strips attached
to the culvert walls with glue, thus avoiding the need to make modifications to the walls
that might impact their integrity. These self-triggering cameras provided infrared
illumination to record animals crossing through the bridge at night. The cameras were
programmed to record up to five frames per second, providing near-video-like tracking of
animals as they approached and crossed through the bridge cells. Images were date, time,
and temperature stamped (Figure 9), digitally stored, and periodically downloaded for
analysis.
32
Figure 8. Reconyx™ Camera Mounted on a Wood Strip Glued to the
Concrete Surface of Each SR 64 Cataract Canyon Bridge Culvert Cell to Monitor
Wildlife Use.
Figure 9. Images of a Mule Deer Doe (Left) and Spike Bull Elk (Right)
Recorded by Reconyx™ Cameras Mounted
in the SR 64 Cataract Canyon Bridge Culvert Cells.
The research team analyzed camera data to determine the frequency of occurrence of
animals by species (and people) that entered and then passed through the bridge cells.
Though not able to determine underpass passage rates as done by Dodd et al. (“Video
surveillance to assess,” 2007) using video camera systems, the proportion of animals that
entered the culvert cells and ultimately passed through provided some indication of the
relative acceptance by animals to cross SR 64 below grade via Cataract Canyon Bridge.
33
4.5 IDENTIFICATION OF PASSAGE STRUCTURE SITES
Sawyer and Rudd (2005) identified several important considerations for locating the most
suitable sites in which to place passage structures, primarily for pronghorn, though these
criteria are applicable for other species. In this assessment of potential passage structure
sites and to validate the preliminary findings in ADOT (2006), the research team
considered each of the criteria identified by Sawyer and Rudd (2005) but recognized that
the 0.1 mile segment length used was too small and cumbersome to discern and analyze
differences among segments.
Dodd et al. (“Evaluation of Measures,” 2007) reported that the optimum scale to address
management recommendations for accommodating wildlife passage needs using GPS
telemetry or WVC data was at the 0.6 mi scale. Making recommendations at this scale
allows ADOT engineers latitude to determine the best technical location for passage
structures along the segment. Thus, the team aggregated the 570 0.1 mi segments from
MP 185.5 to MP 235.4 into 95 0.6 mi segments for analysis. The research team addressed
passage structure needs for the entire highway study area as well as each individual
highway section.
Sawyer and Rudd (2005) identified animal abundance as a primary criterion for the
consideration of passage structure sites. The research team focused this criterion on the
larger population levels adjacent to the entire study area versus by segment. For
pronghorn, Sawyer and Rudd (2005) stressed that passage structures were more
appropriate in linking populations with “abundant numbers (hundreds)” than small
isolated populations that may not benefit to the same degree and exhibit a high likelihood
of encountering passage structures. The pronghorn, elk, and mule deer populations
adjacent to SR 64 indeed number well into the hundreds, with the herds for all three on
both sides of the highway still viable and reproducing.
The team used the other segment-specific criteria identified by Sawyer and Rudd (2005)
with some modifications to rate each of the 95 0.6 mi segments, considering GPS
telemetry findings with other pertinent factors, as done for US 89 by Dodd et al. (2011).
Because passage structures that have the potential to benefit permeability for multiple
species are preferred (Clevenger and Waltho 2000), some ratings for elk, deer, and
pronghorn were additive, thus weighting those sites that may yield benefit to multiple
species. However, because pronghorn range largely did not overlap the higher-density
portions of elk and mule deer ranges (Figure 3) and because few, if any, WVCs involving
pronghorn were anticipated, the team made separate passage structure recommendations
for this species.
34
The team’s rating criteria/categories were as follows:
Elk highway crossings – Due to the anticipated availability of highway crossing
data for elk, this rating was based on the proportion of SR 64 crossings made by
GPS-collared elk within each aggregated 0.6 mi segment across the entire study
area. The ratings for elk crossings were additive to mule deer crossings and
pronghorn approaches. Categories used include:
0 No crossings
1 1–2% of total elk crossings
2 3–4% of total elk crossings
3 5–6 % of total elk crossings
4 7–8% total elk crossings
5 >8% of total elk crossings
Mule deer highway crossings – This rating was also based on the proportion of
SR 64 crossings made by GPS-collared mule deer within each aggregated 0.6 mi
segment. However, because deer were only captured adjacent to SR 64 Section E,
rather than adjacent to a greater length of SR 64, the ratings reflect higher
proportions of crossings. The ratings for mule deer crossings were additive to
those for elk crossings and pronghorn approaches. Categories used include:
0 No crossings
1 1–2% of total crossings for the species
2 3–4% of crossings for the species
3 5–6 % of crossings for the species
4 7–8% of crossings for the species
5 >8% of total crossings for the species
Pronghorn approaches – This criterion was considered indicative of where
animals potentially would approach and cross SR 64 via a passage structure and
was based on the proportion of approaches to within 0.3 mi on both sides of the
highway for aggregated 0.6 mi segments. This rating was additive with elk and
mule deer crossing ratings where GPS-collared animals overlapped. Categories
used include:
0 No approaches
1 1–3% of total pronghorn approaches
2 3–5% of total pronghorn approaches
3 5–15% of total pronghorn approaches
4 15–25% of total pronghorn approaches
5 >25% of total pronghorn approaches
35
Elk, mule deer, and pronghorn distribution – This rating was based on the number
of different GPS-collared animals that crossed SR 64 for elk and deer and were
relocated within the 0.3 mi approach zone for pronghorn. This rating was additive
for each of the three species where data overlapped. Categories used include:
0 No animals crossing or approaching
1 1–2% of all animals crossing or approaching
2 2–4% of all animals crossing or approaching
3 4–6% of all animals crossing or approaching
4 6–8% of all animals crossing or approaching
5 >8% of all animals crossing or approaching
Wildlife-vehicle collisions – The number of non-duplicate WVCs recorded by
0.6 mi segment during the project (2007–2009) for elk, mule deer, pronghorn, and
other large mammals such as mountain lion, black bear, badger, etc. categories
used include:
0 No WVC
1 1–2 total WVCs
2 3–4 total WVCs
3 5–6 total WVCs
4 7–8 total WVCs
5 >8 total WVCs
Land status – This criterion reflected the ability to conduct construction activities
on and outside the ADOT right-of-way (ROW), such as creating approaches with
fill material for overpasses. Categories used include:
0 Private
1 State Trust
3 National Park Service – GCNP (preservation and natural process
focus)
5 Federal – U.S. Forest Service (multiple-use focus)
Human activity – Ideally, no human activity should occur within the vicinity of a
passage structure; however, road access, businesses, visitor pullouts, and other
activities occur adjacent to US 89. Categories used include:
0 Significant human activity (business, housing, etc.)
1 Moderate human activity (access road, visitor pullout)
3 Limited human activity
5 No human activity
36
Topography – The ability to situate overpasses oriented along existing ridgelines
that pronghorn, elk, or deer can traverse, or locate underpasses in association with
wide gentle drainages is desirable. Categories used include:
0 Terrain not suited for a passage structure (steep, broken)
1 Topography marginal for a passage structure (flat)
3 Topography could accommodate a passage structure (small
drainage)
5 Topography ideally suited for passage structure (large drainage for
underpass or ridgeline for overpass)
In addition to the above criteria, the research team also considered other factors in its
identification of potential passage structure sites. These factors included whether the
0.6 mile segments coincided with the preliminary sites recommended in ADOT (2006), if
the types of structures were suited for the site, and how the priority segments from this
study relate to the minimum recommended passage structure spacing determined by
Bissonette and Adair (2008).
37
5.0 RESULTS
5.1 WILDLIFE CAPTURE, GPS TELEMETRY, AND DATA ANALYSIS
5.1.1 Elk Capture, Movements, and Highway Permeability
From June 2007 through October 2009, the research team tracked and recovered data
from 23 elk (13 females, 10 males) instrumented with GPS receiver collars; 17 elk were
trapped in Clover traps and six were captured by darting with immobilizing drugs. Only
three elk were captured at the far south end of the study area adjacent to Section A, while
the remainder were captured at the north end adjacent to Section E (Figure 1).
GPS collars were affixed to elk for an average of 302.9 days (33.4 SE), during which
time the collars accrued 107,055 GPS fixes for a mean of 4654.6 fixes per elk (322.6).
Of the GPS fixes, 12,483 (11.6 percent) were recorded within 0.6 mi of SR 64, and 3796
(3.5 percent) of the fixes were made within 0.15 mi; the proportion of fixes near SR 64
were considerably lower than those for SR 260, where 48.5 percent occurred within
0.6 mi of the highway (Dodd et al. “Effectiveness of Wildlife,” in review). Elk traveled
an average of 1082.1 ft (105.3) between GPS fixes. Males traveled slightly farther
between crossings (1107.1 ft 57.5) than females (1051.2 ft 142.1). Elk minimum
convex polygon (MCP) home ranges averaged 284.8 mi2 (56.3), and male home ranges
(479.5 mi2 165.5) which were significantly larger (t21 = 2.45, P < 0.001) than female
home ranges (199.6 mi2 17.8).
GPS-collared elk crossed SR 64 843 times, for a mean of 40.1 crossings per elk (11.2).
On average, elk crossed SR 64 0.12 time per day (0.03), and ranged from 1 to
200 crossings per elk, though four elk never crossed SR 64. The highest proportion of
crossings occurred in late spring–summer (April–July; 60 percent), followed by late
summer–fall (August–November; 27 percent), and only 13 percent during winter–early
spring (December–March). The overall elk passage rate averaged 0.44 crossing per
approach (0.07; Table 3). There was no significant statistical difference between mean
passage rates for female and males elk (0.46 0.08 vs. 0.43 0.12 respectively).
The crossing distribution by elk among SR 64 0.1 mi segments was not random and
exhibited several peak crossing zones (Figure 10), especially at the north end of the study
area. The observed crossing distribution differed significantly from a random distribution
(Kolmogorov-Smirnov d = 0.309, P < 0.001). The limited number of crossings at the
southern area (Section A) reflects that only three elk were captured in this area and that
these elk crossed SR 64 only an average of three times versus the study area average of
40.1 crossings per elk.
A total of 48 crossings occurred along Section D, with one apparent peak crossing zone,
and 786 occurred along Section E, where several crossing peaks were registered by elk
(Figure 10). The highest concentration of different crossing elk occurred between
Segments 470 and 500, with seven animals (30.4 percent of all elk) and a mean of 3.2
different animals per 0.1 mi segment.
38
Weighted highway crossings reflect the frequency of elk crossings, the number of
individual elk that crossed at each segment, and the evenness in crossing frequency
among all collared animals. The weighted elk crossing pattern (Figure 11) was noticeably
different than the uncorrected crossing distribution. Crossing peaks on Sections A and D
were absent in the weighted crossing distribution due to the relatively small number of
elk that crossed at these peak zones (Figure 10).
Table 3. Comparative Mean Values for GPS-Collared Animals by Species Determined
from GPS Telemetry along SR 64.
Parameter
Mean value per GPS-collared animal by species (±SE)
Elk (n = 23) Mule deer (n = 11) Pronghorn (n = 15)
No. of highway crossings
40.1
(11.2)
55.0
(16.4)
0.2
(0.2)
Highway crossings per day
0.12
(0.03)
0.26
(0.06)
0.001
(0.001)
GPS fixes ≤0.6 mi of
highway per year
632.2
(164.4)
1921.4
(637.7)
791.0
(184.4)
GPS fixes ≤0.15 mi of
highway per year
346.6
(97.7)
1054.9
(354.2)
370.1
(84.1)
GPS fixes ≤0.06 mi of
highway per year
198.8
(63.2)
522.4
(167.9)
116.7
(21.4)
Highway approaches per day
0.21
(0.04)
0.58
(0.12)
0.27
(0.16)
Passage rate
(crossings per approach)
0.44
(0.07)
0.54
(0.08)
0.004
(0.002)
MCP home ranges (mi2)
284.8
(56.3)
141.1
(48.3)
85.8
(21.1)
Distance traveled between
GPS fixes (ft)
1107.1
(57.5)
942.2
(232.9)
845.2
(46.6)
39
Figure 10. SR 64 Crossings by GPS-Collared Elk along the Entire Study Area (Top)
and Sections A through E of the 2006 Final Wildlife Accident Reduction Study and
Enlarged to Show Crossings along Sections D and E (Bottom).
Sections: A B C D E
Section D Section E
40
Figure 11. SR 64 Weighted Crossings by GPS-Collared Elk along the Entire Study
Area (Top) and Sections A through E of the 2006 Final Wildlife Accident Reduction
Study and Enlarged to Show Crossings along Sections D and E (Bottom).
Sections: A B C D E
Section D Section E
41
5.1.2 Mule Deer Capture, Movements, and Highway Permeability
From April 2008 through October 2009, the research team tracked and recovered data
from 11 mule deer (8 females, 3 males) instrumented with GPS receiver collars. Deer
were captured adjacent only to Section E at the far north extent of the study area.
GPS collars were affixed to deer for an average of 207.9 days (38.7), during which time
they accrued 29,944 GPS fixes, for a mean of 5988.5 fixes per deer (128.0). Deer were
relocated near SR 64 more than elk, with 12,047 (57.2 percent) fixes recorded within
0.6 mi of SR 64, and 3796 (15.6 percent) of the fixes made within 0.15 mi of the
highway. Mule deer traveled an average of 942.2 ft (232.9) between GPS fixes. Males
traveled slightly farther between crossings (1003.7 ft 323.5) than females (820.9 ft
155.2). Home ranges averaged 141.1 mi2 (48.3); male home ranges (189.8 mi2 184.9)
were not significantly different (P = 0.343) from female ranges (132.3 mi2 51.4).
Five mule deer (two males, three females) captured along SR 64 exhibited extreme long-distance
(more than 100 mi) movements away from SR 64 to the south, most
independently of each other. All five followed the same travel corridor to an area
northwest of Flagstaff and west of the San Francisco Peaks (Figure 12). The mean home
range of these five deer (342.1 mi2 41.9) was 22 times greater than those that did not
make such movements (15.5 mi2 3.6; t11 = 10.0, P < 0.001). The factors contributing to
such movement patterns is being addressed in another study by the research team, but
such movements point to the need to maintain landscape connectivity for far-ranging
species.
Collared deer crossed SR 64 550 times, for a mean of 55.0 crossings per deer (16.4).
On average, deer crossed SR 64 twice as frequently as elk, or 0.26 times per day (0.06).
Deer crossings ranged from 2 to 147 crossings per deer, and one deer never crossed SR
64. Seasonal deer crossings were more consistent than elk, though 46 percent occurred
during late summer–fall (August–November), followed by 28 percent in late spring–
summer (April–July), and 26 percent in winter–early spring (December–March). The
overall deer passage rate was higher than that for elk and averaged 0.54 crossing per
approach (0.08; Table 3). The mean passage rate for female deer (0.59 crossing per
approach; 0.08) was higher than that for male deer (0.34 0.34).
The deer crossing distribution by 0.1 mi segments did not occur randomly and exhibited
two peak crossing zones; all crossings occurred along Section E beyond Segment 420
(Figure 13). The observed mule deer crossing distribution differed significantly from a
random distribution (Kolmogorov-Smirnov d = 0.281, P < 0.001). The two crossing
peaks occurred along a 3.2 mi stretch (Segments 480–512) between the entrances to
Grand Canyon Airport and GCNP (Figure 13); 505 crossings (92 percent of total)
occurred along this stretch of highway, though deer were captured along the length of
Section E.
42
Figure 12. Mule Deer GPS Fixes along the SR 64 Study Area, as well as Fixes for Two
Deer Captured North of Flagstaff (Numbers 43 and 172).
When considering the number of crossing mule deer in calculating weighted crossings,
the same two peaks in the crossing distribution were even more apparent. Deer crossings
between Segments 450 and 470 largely disappeared and were restricted to two segments
when weighted crossings were calculated (Figure 13).
5.1.3 Pronghorn Capture, Movements, and Highway Approaches
The research team instrumented and tracked 15 pronghorn (10 females, 5 males) with
GPS receiver collars from January 2008 through January 2009. Due to disparity in the
distribution of pronghorn herds (Figure 3) adjacent to SR 64, coupled with the prevalence
of closed private lands across much of the pronghorn range, the team was not able to
achieve its objective of collaring an equal number of animals on each side of the
highway; 10 were captured on the east side and five on the west.
GPS collars were affixed to pronghorn an average of 298.1 days (29.6), during which
time the collars accrued 56,433 GPS fixes for a mean of 3762.2 fixes per pronghorn
(339.0).
43
Figure 13. SR 64 Highway Crossings (Top) and Weighted Crossings (Bottom) by GPS-Collared
Mule Deer along Highway Section E by 0.1 mi Segment.
44
Of the GPS fixes accrued for pronghorn, 1426 (3 percent) occurred within 0.15 mi of SR
64, or an average of 95.1 (26.1) fixes per animal; all but one pronghorn approached the
highway to within 0.15 mi. All 15 pronghorn approached to within 0.60 mi of the
highway, accruing 9729 GPS fixes (17 percent of all fixes), with a mean of 648.6
(151.2) fixes per animal. During the duration of GPS tracking, pronghorn traveled an
average of 845.2 ft (46.6) between fixes (1.5 hr). Females traveled farther between fixes
(905.5 ft 44.6) than males (724.1 ft 72.2). Pronghorn home ranges averaged 85.8 mi2
(21.1), and there was no difference between male (88.6 mi2 29.6) and female (84.4 mi2
34.5) home ranges (P = 0.469).
Only a single GPS-collared pronghorn crossed SR 64 during tracking—a female that
crossed three times; none of the other 14 collared pronghorn crossed the highway. The
pronghorn crossing rate averaged 0.001 crossings per day. The mean pronghorn passage
rate was a negligible 0.004 crossings per approach (Table 3).
The frequency of approaches by pronghorn to within 0.30 miles of SR 64 yielded
considerably more information than crossings for the determination of potential passage
structure locations. Pronghorn approached the highway to within 0.30 miles 4269 times
(Figure 14), for a mean of 284.6 (±69.0) approaches per animal and a range of 2 to 907
approaches.
The observed approach distribution did not occur in a random distribution (Kolmogorov-
Smirnov d = 0.883, P < 0.001). Partly owing to the disparity in the number of collared
animals on the east and west sides of SR 64, it is not unexpected that 3623 approaches
were from the east and only 465 from the west. However, the approaches per animal were
also dramatically different; 362.2 approaches per animal (±74.4) on the east side versus
91.2 approaches per animal (±121.3) from the west. All but two approaches to SR 64
between Segments 1 and 220 were made by pronghorn approaching from the east, though
a limited number of animals were captured on the west side along this stretch.
Shannon diversity index (SDI)-weighted pronghorn approaches totaled 2756.7, and the
distribution pattern changed considerably from the uncorrected approach distribution.
The peak in crossings at the south end of the study area disappeared, owing to there being
only a single male that approached here. The weighted distribution of approaches also
showed an increased peak in approaches at the north extent of pronghorn range, between
Segments 310 and 390. Between Segments 340 and 370, 11 different collared animals
(73.3 percent of total) approached SR 64, with a mean of 6.9 different animals per 0.1
mile segment. The peak in approaches between Segments 180 and 220 at the center of the
study area remained prevalent even after SDI-weighted approaches were calculated
(Figure 14), though approaches here were attributable to just two pronghorn.
45
Figure 14. Highway Approaches (Top) and Weighted Approaches (Bottom) Made to
within 0.3 mi of SR 64 by GPS-Collared Pronghorn and Sections A through E of the
2006 Final Wildlife Accident Reduction Study.
Sections: A B C D E
Sections: A B C D E
46
5.2 TRAFFIC RELATIONSHIPS
5.2.1 Elk-Traffic Relationships
The research team’s elk distribution analysis was based on 12,483 GPS fixes recorded
within 3300 ft of the highway. Frequency distributions of mean probabilities showed a
shift in distribution away from SR 64 with increasing traffic volume (Table 4; Figure 15).
The shift away from the highway occurred even at relatively low traffic volume
(Figure 15), with a 64 percent decrease in probability of elk occurring within 660 ft from
traffic volume less than 100 vehicles per hr (0.28 probability) to 200 to 300 vehicles per
hr (0.10). The mean probability of elk occurring within 660 ft of SR 64 remained
constant at 0.08 from 200 to 600 vehicles per hr. The mean probability of elk occurring
farther away from the highway (1650 and 1980 ft distance bands) increased 65 percent
from traffic less than 100 vehicles per hr (0.17) to 500 to 600 vehicles per hr (0.28).
Elk passage rates by 2 hr time blocks ranged from 0.01 (18:00–20:00) to 0.72 crossings
per approach (04:00–06:00), with the passage rate between midnight and 04:00 when
traffic was nearly absent along the highway (Figure 5) averaging 0.63 crossings per
approach (0.05; n = 45). This nighttime rate was more than three times higher than the
mean passage rate during the rest of the day, averaging 0.19 crossings per approach
(0.04; n = 72; Figure 16). There was a significant negative association between the elk
passage rate by 2 hour blocks and increasing traffic volume (r = -0.660, P = 0.022). The
passage rate by day averaged 0.48 crossings per approach (±0.16) and was relatively
constant for most days (0.42–0.49) except for Tuesday (0.64).
5.2.2. Mule Deer–Traffic Relationships
The research team’s mule deer distribution analysis was based on 12,047 GPS fixes
recorded within 3300 ft of the highway. Mule deer frequency distributions of combined
mean probabilities showed shifts in distribution away from the highway with increasing
traffic volume, though not as dramatic as for elk (Table 4; Figure 17). At low traffic
volumes less than 200 vehicles per hr, probabilities for deer occurring within 660 ft of the
highway averaged 0.21 but dropped when traffic was more than 200 vehicles per hr and
remained static out to 1980 ft, averaging 0.11.
Mean deer probabilities of occurring within the 1650–1980 ft distance band largely
remained unchanged across traffic volume classes (mean = 0.21) up to 500 vehicles per
hr but dropped to a mean probability of 0.14 at more than 500 vehicles per hr. The most
dramatic shift in deer distribution occurred in the intermediate distance bands, 990–1320
ft from SR 64, with the probability of deer occurring here doubling from 0.12 at less than
100 vehicles per hr to 0.24 at just 100 to 200 vehicles per hr; the probability of deer
occurring at this distance remained static up to 500 vehicles per hr and averaged 0.21.
Deer passage rates by 2 hour time blocks ranged from 0.03 (19:00–21:00) to 0.78
crossings per approach (03:00–05:00 a.m.), with the passage rate between midnight and
04:00 a.m. when traffic was absent, averaging 0.58 crossings per approach (0.03). This
nighttime rate was more than two times the mean passage rate during the rest of the day,
47
averaging 0.28 crossings per approach (0.02; Figure 16). Unlike elk, the deer passage
rate remained relatively high (0.61 crossings per approach [0.02]) well into the morning
hours up until the 09:00–11:00 time block (Figure 16).
Due to the passage rates remaining high into the morning hours, the negative association
between the deer passage rate by 2 hr blocks and increasing traffic volume was not
significant (r = -0.07, P = 0.831). The passage rate by day of the week averaged 0.47
crossings per approach (±0.10), which was relatively constant for most days, and ranged
from 0.42 on weekend days to 0.50 the remainder of the week.
Table 4. Mean Probabilities that GPS-Collared Elk, Mule Deer,
and Pronghorn Occurred within Distance Bands from SR 64 at Varying Traffic Volumes.
Documented from 2007 through 2009.
Distance
from
highway (ft)
by species
Probability of occurring in distance band by traffic volume
(vehicles per hour)
<100 100–200 200–300 300–400 400–500 500–600
0–990
Elk
Mule deer
Pronghorn
All
0.36
0.38
0.23
0.34
0.21
0.32
0.18
0.22
0.18
0.23
0.16
0.19
0.18
0.17
0.15
0.17
0.15
0.18
0.12
0.14
0.15
0.19
0.11
0.14
990–1980
Elk
Mule deer
Pronghorn
All
0.26
0.27
0.30
0.28
0.26
0.34
0.33
0.30
0.28
0.33
0.32
0.30
0.31
0.31
0.34
0.32
0.31
0.31
0.31
0.31
0.40
0.21
0.35
0.33
1980–2970
Elk
Mule deer
Pronghorn
All
0.28
0.29
0.37
0.30
0.38
0.26
0.38
0.36
0.27
0.34
0.43
0.40
0.38
0.39
0.38
0.38
0.43
0.37
0.47
0.43
0.33
0.43
0.45
0.40
48
Figure 15. Mean Probability That GPS-Collared Elk Occurred within 330 ft Distance
Bands along SR 64 at Varying Traffic Volumes.
0
0.1
0.2
330 660 990 132016501980
Mean probability
Distance from highway (ft)
500-600 vehicles/hr
49
Figure 16. Mean SR 64 Passage Rates by Two-Hour Time Blocks (Reflected by the
Midpoint of the Blocks) and Corresponding Mean Traffic Volumes during Each Time
Block for Elk (Bottom) and Mule Deer (Top).
50
Figure 17. Mean Probability That GPS-Collared Mule Deer Occurred within 330 ft
Distance Bands along SR 64 at Varying Traffic Volumes.
51
5.2.3. Pronghorn-Traffic Relationships
The team’s distribution analysis was based on 9729 pronghorn GPS fixes recorded within
3300 ft of SR 64. Regardless of traffic volume, even at the lowest levels, pronghorn
distribution within 660 ft of the highway was low, with all combined probabilities less
than 0.07 (Figure 18). In the 990–1320 ft distance bands, the mean combined probability
of occurrence dropped from 0.28 at less than 100 vehicles per hr but stayed relatively
constant thereafter (0.20–0.21) up to 400 vehicles per hr, with a slight drop to 0.13 at
400–500 vehicles per hr. In the 1980 ft distance band, the mean probability of pronghorn
occurring here increased from 0.07 at less than 100 vehicles per hr to 0.13 at just 100–
200 vehicles per hr, and up to 0.16 at 500–600 vehicles per hr (Figure 18).
The proportion of pronghorn GPS approaches made to within 0.15 mi of SR 64 (n = 951)
varied throughout the day. Nearly half the approaches (48 percent) occurred between
16:30 and 19:00, when traffic was at its highest level during the day. During the morning
hours (05:30–10:00 hr), 18 percent of the approaches occurred, as did an equal proportion
of approaches made during midday hours (10:00–16:30).
5.3. WILDLIFE-VEHICLE COLLISION RELATIONSHIPS
From 2007 through 2009, DPS highway patrol officers and research team members
recorded 157 WVCs involving elk and mule deer along the SR 64 study area (Table 5).
Elk accounted for 63 percent (n = 99) of these WVCs, followed by mule deer, which
accounted for 35 percent (n = 58). In addition, three coyotes, three rabbits and one
mountain lion, black bear, and badger each were involved in WVCs during the study
period. No collisions involving pronghorn were recorded. In total, 77 WVCs were
recorded in 2007 (46 elk, 28 deer, 4 other), 51 in 2008 (33 elk, 16 deer, 2 other), and
40 in 2009 (20 elk, 14 deer, 3 other). DPS highway patrol accident reports indicated that
27 human injuries occurred in WVCs during the study.
Table 5. WVCs Involving Elk and Mule Deer on SR 64 Sections from 2007 through
2009, including the Total Number and Mean Collisions (per Mile).
SR 64
section
Elk collisions Deer collisions All collisions
Total Mean
(per mile) Total Mean
(per mile) Total Mean
(per mile)
A 26 1.3 30 1.6 56 2.9
B 4 0.5 0 – 4 0.5
C 0 – 0 – 0 –
D 13 1.4 3 0.3 16 1.8
E 56 4.7 25 2.1 81 6.8
All 99 1.9 58 1.2 157 3.1
52
Figure 18. Mean Probability That GPS-Collared Pronghorn Occurred within 330 ft
Distance Bands along SR 64 at Varying Traffic Volumes.
53
Section E (MP 223.4 to MP 235.4) had the highest incidence of elk and deer collisions, as
well as collisions per mile and more than twice the collisions per mile than Section A
(MP 185.5 to MP 204.7; Figure 19); these sections account for the highest density elk and
deer range along SR 64 (Figure 3). The spatial association between elk-vehicle collisions
and crossings at the 1.0 mi scale was significant (r = 0.811, n = 27, P < 0.001), as was the
association for deer (r = 0.705, n = 10, P = 0.022).
Figure 19. Frequency of Elk and Mule Deer Collisions with Vehicles by SR 64
Milepost from 2007 through 2009.
From 1998 through 2008, 41.7 percent of all single-vehicle accidents recorded along
SR 64 involved wildlife, compared with the national average of just 4.6 percent (Huijser
et al. 2007; Figure 20). The proportion of accidents involving wildlife (recorded by MP)
was as high as 87 percent (MP 233), with wildlife-related accidents accounting for more
than 75 percent of all single-vehicle accidents along five mileposts (all from MP 229 to
MP 234; Figure 20).
For accidents where time was reported by DPS, the incidence of elk and deer collisions
varied considerably among time periods (Table 6; Figure 21). The highest proportion of
elk collisions (50 percent) occurred from 5:00 p.m. to 10:00 p.m., followed by 39 percent
from 11:00 p.m. to 04:00 a.m. Only 11 percent of elk accidents occurred from 05:00 to
10:00 a.m., and none were recorded from 11:00 a.m. to 4:00 p.m. The observed
frequency of elk-vehicle collisions by time period differed from expected values (2 =
62.5, df = 3, P < 0.001).
54
Figure 20. Proportion of SR 64 Single-Vehicle Accidents by Milepost from 1998
through 2008 that Involved Wildlife.
The negative association between the occurrence of elk-vehicle collisions and traffic
volume by hour was significant (r = -0.723, P = 0.001) in spite of the disproportionately
low incidence of collisions that occurred in the morning when traffic volume was low
(Figure 21).
The timing of deer-vehicle collisions was more variable than those for elk (Table 6;
Figure 21), though the observed frequency differed significantly from the expected by
time period (2 = 26.8, df = 3, P < 0.001). Though 49 percent of accidents involving deer
occurred during the evening, only 8 percent occurred at night when traffic volume was
lowest. Conversely, during the times of the day when traffic volume was at its highest,
late morning and midday, a combined 43 percent of deer-vehicle collisions occurred
(Figure 20), partly accounting for the poor association between deer collisions and traffic
volume (r = 0.016, P = 0.941).
The incidence of elk collisions by day of the week did not vary significantly (P = 0.800),
though there were fewer collisions on Thursday than other days (Figure 22). For deer,
however, the incidence of collisions on Monday was more than double that of the other
six days, and the observed frequency of collisions by day was marginally different from
55
what was expected (2 = 11.5, df = 6, P = 0.075). Neither association between both elk
and deer collisions versus mean daily traffic volume was significant (P = 0.879 and P =
0.562, respectively).
Figure 21. SR 64 Elk and Mule Deer Collisions with Vehicles by Time of Day and
Associated Traffic Volume.
56
Figure 22. SR 64 Elk and Mule Deer Collisions with Vehicles by Day
and Associated Traffic Volume.
There was a significant difference in the observed versus expected frequency of elk-vehicle
collisions by season (Table 7; 2 = 17.4, df = 2, P < 0.001). The driest season,
early spring–summer (April–July), accounted for 43 percent of all elk-vehicle collisions
along SR 64, while late summer–fall (August–November) accounted for another 38
percent (Table 7; Figure 23).
The association between elk collisions and mean monthly traffic volume was significant
(r = 0.789, P = 0.002). For mule deer, the incidence of collisions was relatively constant
through much of the year, except the late summer–fall season when nearly half of all
collisions occurred (Table 7; Figure 23). The association between deer-vehicle collisions
and traffic volume was not significant (P = 0.210). The association between elk crossings
and collisions by month was significant (r = 0.583, P = 0.047), as was the association for
deer (r = 0.686, P = 0.014).
57
Table 6. Frequency of Elk and Deer Collisions with Vehicles
along SR 64 by Time Period.
Time period Hours
Frequency of WVCs (%)
Elk Mule deer
Evening 17:00–22:00 47
(50.0%)
32
(49.2%)
Nighttime 23:00–04:00 37
(39.4%)
5
(7.7%)
Morning 05:00–10:00 10
(10.6)
19
(29.2%)
Midday 10:00–16:00 0
(–)
9
(13.8%)
Table 7. Frequency of Elk and Deer Collisions with Vehicles
along SR 64 by Season.
Season Months
Frequency of wildlife vehicle collisions (%)
Elk Mule deer
Winter–early spring Dec–Mar 18
(18.2%)
12
(20.3%)
Late spring–summer Apr–Jul 43
(43.4%)
18
(30.5%)
Late summer–fall Aug–Nov 38
(38.4%)
29
(49.1%)
58
Figure 23. SR 64 Elk and Mule Deer Collisions with Vehicles by Month
and the Mean Traffic Volume.
5.4 WILDLIFE USE OF CATARACT CANYON BRIDGE
Camera monitoring of Cataract Canyon Bridge was conducted from July 2008 through
December 2009. A total of 126 wildlife images, including 13 elk and 37 mule deer, were
recorded by the four cameras in the bridge cells (Table 8).
Of the limited number of elk that approached the bridge during the study, 92 percent
successfully crossed through the bridge cells, while the remaining 8 percent turned back.
For mule deer, for which a greater number of successful crossings were recorded (n = 37)
than elk, 89 percent of crossings were successful (Table 8). For smaller mammal species,
including gray fox, raccoon, skunk, and various squirrel species, only 6 percent of these
animals went all the way through the structure, while 94 percent turned back.
The relatively low mobility of some of these species (e.g., squirrels) may have limited
their potential for crossing through the bridge. The vast majority of deer underpass use
occurred from August through October, with 89 percent of the entries into the bridge in
these three months. Elk use occurred only in October and December, with no entries the
rest of the year. Of all deer and elk bridge crossings, 64 percent occurred in the 3 hr
period from 10:00 p.m. to 01:00 a.m.
59
Table 8. Number of Animals by Species that Entered and Successfully Crossed through
Cataract Canyon Bridge on SR 64, and Success Rates.
Species
Animals entering bridge Animals crossing through bridge
Success
No. Proportion No. Proportion rate
Elk 13 0.10 12 0.19 0.92
Mule deer 37 0.29 33 0.53 0.89
Gray fox 5 0.04 2 0.48 0.40
Raccoon 10 0.08 2 0.03 0.20
Skunk 7 0.06 1 0.02 0.14
Squirrel 54 0.43 0 – –
In addition to wildlife use of Cataract Canyon Bridge, substantial presence by people was
documented. In total, 191 humans were recorded, averaging 15.9 people per month;
29 all-terrain vehicles were recorded at the bridge. Human use of the bridge was largely
restricted to daytime hours from 10:00 a.m. to 5:00 p.m., when 75 percent of the use
occurred.
5.5 IDENTIFICATION OF PASSAGE STRUCTURE SITES
The research team used elk and mule deer highway crossings, WVCs, pronghorn
approaches, and the proportions of animals crossing or approaching within each segment,
among other criteria, to rate 95 0.6 mi segments for suitability as potential passage
structure locations. Additional criteria included land ownership and topography that
would support passage structure construction. Ratings of the 94 0.6 mi segments from
MP 185.5 to MP 235.4 for their suitability for potential passage structures ranged from
1 to 33 points (mean = 10 points) of a possible 40 points (Figure 24). The highest-rated
(33 points) 0.6 mi segment (Segment 82, MP 234.5 to MP 235.0) was on the Kaibab NF
just south of the south entrance to Grand Canyon Airport, which corresponded to the
stretch of highway with the highest proportion of elk crossings (14.3 percent;
121 crossings) and mule deer crossings (42.3 percent; 227 crossings), as well the
highest incidence of WVCs (n = 10).
The land ownership and terrain at this segment further make this site suited for a passage
structure. ADOT (2006) identified this site as warranting an overpass, though it was not
included in the preferred alternative. The next two highest-rated 0.6 mi segments scored
24 points each; one (Segment 81) was just south of the highest-rated segment, further
po