Evaluation of measures to minimize wildlife-vehicle collisions and maintain wildlife permeability across highways

              EVALUATION OF MEASURES TO

MINIMIZE WILDLIFE- VEHICLE

COLLISIONS AND MAINTAIN

PERMEABILITY ACROSS HIGHWAYS:

Arizona Route 260

Final Report 540

Prepared by:

Norris L. Dodd, Jeffrey W. Gagnon, Susan Boe,

Amanda Manzo, and Raymond E. Schweinsburg

Arizona Game and Fish Department

Research Branch

2221 West Greenway Road

Phoenix, AZ 85023

August 2007

Prepared for:

Arizona Department of Transportation

206 South 17th Avenue

Phoenix, Arizona 85007

in cooperation with

U. S. Department of Transportation

Federal Highway Administration

The contents of this report reflect the views of the authors who are responsible for the facts and

accuracy of the data presented herein. The contents do not necessarily reflect official views or

policies of the Arizona Department of Transportation or the Federal Highway Administration. The

report does not constitute a standard, specification, or regulation. Trade or manufacturers’

names that 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- 07- 540

2. Government Accession No.

3. Recipient's Catalog No.

5. Report Date

AUGUST 2007

4. Title and Subtitle

EVALUATION OF MEASURES TO MINIMIZE WILDLIFE-VEHICLE

COLLISIONS AND MAINTAIN WILDLIFE

PERMEABILITY ACROSS HIGHWAYS: Arizona Route 260

6. Performing Organization Code

7. Au thors

Norris L. Dodd, Jeffrey W. Gagnon, Susan Boe, Amanda Manzo,

and Raymond E. Schweinsburg

8. Performing Organization Report No.

10. Work Unit No.

9. Performing Organization Name and Address

Arizona Game and Fish Department

Research Branch

2221 West Greenway Road

Phoenix, Arizona 85023

11. Contract or Grant No.

ECS File No. JPA 01- 152

JPA 04- 024T

13. Type of Report & Period Covered

FINAL REPORT

January ‘ 02 – December ‘ 06

12. Sponsoring Agency Name and Address

Arizona Department of Transportation

206 S. 17th Avenue

Phoenix, AZ 85007

ATRC Project Manager: Estomih Kombe

14. Sponsoring Agency Code

15. Supplementary Notes

Prepared in cooperation with the U. S. Department of Transportation, Federal Highway Administration

16. Abstract

We conducted wildlife- highway relationships research from 2002– 2006 along a 17- mile stretch of State

Route 260 in Arizona which is being reconstructed in five phases with 11 wildlife underpasses and 6

bridges. Reconstruction phasing allowed us to use a before- after- control experimental approach in our

research. The objectives of our research were;

1) Assess and compare wildlife use of underpasses. 2) Evaluate highway permeability and wildlife

movements among reconstruction classes. 3) Characterize wildlife- vehicle collision patterns and changes

with reconstruction. 4) Assess relationships among highway traffic volume and wildlife vehicle collisions,

elk crossing patterns, and wildlife use of underpasses. 5) Assess the role that ungulate- proof fencing plays in

wildlife vehicle collisions, wildlife use of underpasses, and wildlife permeability. 6) Provide ongoing

highway reconstruction implementation guidance.

We used video surveillance to assess and compare wildlife use of five underpasses at which we recorded

8,455 animals and 11 different species; 5,560 of these animals ( 65.8%) crossed through the underpass. We

employed Global Positioning System telemetry to assess highway permeability across SR 260, with 65 elk

fitted with receiver collars. Elk crossed State Route 260 5,749 times. Elk permeability on reconstructed

highway ( 0.43 crossings/ approach) was half that of control sections. Permeability increased 60% after

ungulate- proof fencing was erected on a reconstructed section. Effective monitoring and adaptive

management yielded benefits to highway safety and wildlife permeability alike.

17. Key Words

Cervus elaphus, deer, elk, GPS telemetry, fencing,

highway crossings, highway impact, Odocoileus spp.,

permeability, traffic volume, video surveillance, wildlife

underpasses, wildlife- vehicle collisions.

18. Distribution Statement

No restriction. Document is

available to the U. S. public through

the National Technical Information

Service, Springfield, Virginia 22161

19. Security Classification

Unclassified

20. Security Classification

Unclassified

21. No. of Pages

185

22. Price

23.

Registrant’s

Seal

SI* ( MODERN METRIC) CONVERSION FACTORS

APPROXIMATE CONVERSIONS TO SI UNITS APPROXIMATE CONVERSIONS FROM SI UNITS

Symbol When You

Know Multiply By To Find Symbol Symbol When You Know Multiply By To Find Symbol

LENGTH LENGTH

in inches 25.4 millimeters mm mm millimeters 0.039 inches in

ft feet 0.305 meters m m meters 3.28 feet ft

yd yards 0.914 meters m m meters 1.09 yards yd

mi miles 1.61 kilometers km km kilometers 0.621 miles mi

AREA AREA

in2 Square inches 645.2 square millimeters mm2 mm2 square millimeters 0.0016 square inches in2

ft2 square feet 0.093 square meters m2 m2 square meters 10.764 square feet ft2

yd2 square yards 0.836 square meters m2 m2 square meters 1.195 square yards yd2

ac acres 0.405 hectares ha ha hectares 2.47 acres ac

mi2 square miles 2.59 square kilometers km2 km2 square kilometers 0.386 square miles mi2

VOLUME

VOLUME

fl oz fluid ounces 29.57 milliliters mL mL milliliters 0.034 fluid ounces fl oz

gal gallons 3.785 liters L L liters 0.264 gallons gal

ft3 Cubic feet 0.028 cubic meters m3 m3 cubic meters 35.315 cubic feet ft3

yd3 Cubic yards 0.765 cubic meters m3 m3 cubic meters 1.308 cubic yards yd3

NOTE: Volumes greater than 1000L shall be shown in m3.

MASS MASS

oz ounces 28.35 grams g g grams 0.035 ounces oz

lb pounds 0.454 kilograms kg kg kilograms 2.205 pounds lb

T short tons

( 2000lb) 0.907 megagrams

( or “ metric ton”)

mg

( or “ t”)

mg

( or “ t”)

megagrams

( or “ metric ton”) 1.102 short tons

( 2000lb) T

TEMPERATURE ( exact)

TEMPERATURE ( exact)

º F Fahrenheit

temperature

5( F- 32)/ 9

or ( F- 32)/ 1.8 Celsius temperature º C º C Celsius temperature 1.8C + 32 Fahrenheit

temperature

º F

ILLUMINATION

ILLUMINATION

fc foot- candles 10.76 lux lx lx lux 0.0929 foot- candles fc

fl foot- Lamberts 3.426 candela/ m2 cd/ m2 cd/ m2 candela/ m2 0.2919 foot- Lamberts fl

FORCE AND PRESSURE OR STRESS FORCE AND PRESSURE OR STRESS

lbf poundforce 4.45 Newtons N N Newtons 0.225 poundforce lbf

lbf/ in2 poundforce per

square inch 6.89 kilopascals kPa kPa kilopascals 0.145 poundforce per

square inch lbf/ in2

SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380

TABLE OF CONTENTS

1.0 EXECUTIVE SUMMARY ......................................................................................... 1

1.1 Video surveillance to assess highway underpass use by elk.............................. 2

1.2 Effects of traffic volume on elk use of wildlife highway underpasses .............. 2

1.3 Assessment of elk highway permeability using Global Positioning System

telemetry ............................................................................................................ 3

1.4 Role of fencing in promoting wildlife underpass use and highway

permeability ....................................................................................................... 4

1.5 Influence of fluctuating traffic volume on elk distribution and highway

crossings............................................................................................................. 4

1.6 Characteristics of elk- vehicle collisions and comparison to highway crossing

patterns............................................................................................................... 5

1.7 Influence of environmental factors on elk highway crossings........................... 6

1.8 Preliminary assessment of factors influencing wildlife use of highway

underpasses ........................................................................................................ 6

1.9 Overall conclusions and recommendations ....................................................... 7

2.0 INTRODUCTION...................................................................................................... 11

2.1 BACKGROUND ............................................................................................. 11

2.2 EXPERIMENTAL APPROACH..................................................................... 12

2.2.1 Phased Construction and Adaptive Management ................................ 13

2.2.2 Experimental Approach ....................................................................... 13

2.3 RESEARCH OBJECTIVES............................................................................ 14

2.4 REPORT ORGANIZATION........................................................................... 18

3.0 STUDY AREA............................................................................................................ 19

4.0 VIDEO SURVEILLANCE TO ASSESS HIGHWAY UNDERPASS USE BY

ELK ............................................................................................................................ 25

4.1 INTRODUCTION ........................................................................................... 25

4.2 STUDY AREA ................................................................................................ 26

4.3 METHODS...................................................................................................... 28

4.3.1 Video Surveillance System Components and Layout.......................... 28

4.3.2 Video Data Analysis ............................................................................ 28

4.3.3 Time- Lapse Validation ........................................................................ 30

4.3.4 Statistical Analysis............................................................................... 30

4.4 RESULTS ........................................................................................................ 32

4.5 DISCUSSION.................................................................................................. 36

5.0 EFFECTS OF TRAFFIC VOLUME ON ELK USE OF WILDLIFE

HIGHWAY UNDERPASSES .................................................................................. 41

5.1 INTRODUCTION ........................................................................................... 41

5.2 MATERIALS AND METHODS..................................................................... 42

5.3 RESULTS ........................................................................................................ 43

5.4 DISCUSSION.................................................................................................. 45

6.0 ASSESSMENT OF ELK HIGHWAY PERMEABILITY USING GLOBAL

POSITIONING SYSTEM TELEMETRY .............................................................. 51

6.1 INTRODUCTION ........................................................................................... 51

6.2 METHODS...................................................................................................... 52

6.2.1 Elk Capture and GPS Collars............................................................... 52

6.2.2 GPS Accuracy Validation .................................................................... 52

6.2.3 GPS Data Analysis of Elk Movements................................................ 53

6.2.4 Statistical Analysis............................................................................... 55

6.3 RESULTS ........................................................................................................ 57

6.3.1 GPS Fix Accuracy................................................................................ 57

6.3.2 Elk Highway Movements .................................................................... 57

6.3.3 Habitat and Elk Movements................................................................. 60

6.4 DISCUSSION.................................................................................................. 62

7.0 ROLE OF FENCING IN PROMOTING WILDLIFE UNDERPASS USE AND

HIGHWAY PERMEABILITY ................................................................................ 67

7.1 INTRODUCTION ........................................................................................... 67

7.2 STUDY AREA ................................................................................................ 69

7.3 METHODS...................................................................................................... 71

7.3.1 Comparison of elk crossing patterns and permeability........................ 71

7.3.2 Comparison of elk- vehicle collision patterns ...................................... 72

7.3.3 Wildlife Underpass Use Comparison................................................... 73

7.4 RESULTS ........................................................................................................ 74

7.4.1 Comparison of elk crossing patterns and permeability........................ 74

7.4.2 Comparison of elk- vehicle collision patterns ...................................... 76

7.4.3 Wildlife Underpass Use Comparison................................................... 79

7.5 DISCUSSION.................................................................................................. 79

8.0 INFLUENCE OF FLUCTUATING TRAFFIC VOLUME ON ELK

DISTRIBUTION AND HIGHWAY CROSSINGS................................................. 83

8.1 INTRODUCTION ........................................................................................... 83

8.2 METHODS...................................................................................................... 84

8.3 RESULTS ........................................................................................................ 86

8.4 DISCUSSION.................................................................................................. 88

8.4.1 Traffic volume and elk distribution ..................................................... 88

8.4.2 Traffic volume, probability of highway crossing and highway

permeability ..................................................................................................... 90

8.4.3 Elk distribution and collisions ............................................................. 90

9.0 CHARACTERISTICS OF ELK- VEHICLE COLLISIONS AND

COMPARISON TO GLOBAL POSITIONING SYSTEM- DETERMINED

HIGHWAY CROSSING PATTERNS..................................................................... 93

9.1 INTRODUCTION ........................................................................................... 93

9.2 METHODS...................................................................................................... 94

9.2.1 Wildlife- vehicle collision tracking ...................................................... 94

9.2.2 Elk- vehicle collision relationships to AADT and elk population

estimates........................................................................................................... 95

9.2.3 Comparison of elk- vehicle collision by highway section and

construction classes.......................................................................................... 96

9.2.4 GPS telemetry assessment of elk highway crossings .......................... 96

9.2.5 Comparison of elk- vehicle collision and elk highway crossing

patterns............................................................................................................. 96

9.3 RESULTS ........................................................................................................ 98

9.3.1 Elk- vehicle collision relationships to AADT and elk population

estimates........................................................................................................... 98

9.3.2 Comparison of elk- vehicle collisions by highway section and

construction class........................................................................................... 101

9.3.3 Cost- benefit of measures to reduce elk- vehicle collisions................. 103

9.3.4 Comparison of elk- vehicle collision and elk highway crossing

patterns........................................................................................................... 104

9.3.5 Spatial relationships between elk- vehicle collision and crossing

patterns........................................................................................................... 110

9.3.6 Temporal relationships between elk- vehicle collision and crossing

patterns........................................................................................................... 110

9.4 DISCUSSION................................................................................................ 113

9.4.1 Elk- vehicle collision relationships to AADT and elk population

estimates......................................................................................................... 115

9.4.2 Comparison of elk- vehicle collisions by highway section and

construction classes........................................................................................ 117

9.4.3 Cost- benefit of measures to reduce elk- vehicle collisions................. 118

9.4.4 Comparison of elk- vehicle collisions and elk highway crossings ..... 118

10.0 INFLUENCE OF ENVIRONMENTAL FACTORS ON ELK HIGHWAY

CROSSINGS ............................................................................................................ 123

10.1 INTRODUCTION ......................................................................................... 123

10.2 METHODS .................................................................................................... 125

10.2.1 Determination of elk highway crossing patterns ............................... 125

10.2.2 Habitat assessment ............................................................................. 125

10.2.3 Statistical analysis.............................................................................. 126

10.3 RESULTS ...................................................................................................... 127

10.4 DISCUSSION................................................................................................ 129

11.0 PRELIMINARY ASSESSMENT OF FACTORS INFLUENCING WILDLIFE

USE OF HIGHWAY UNDERPASSES.................................................................. 133

11.1 INTRODUCTION ......................................................................................... 133

11.2 STUDY AREA .............................................................................................. 133

11.3 METHODS .................................................................................................... 135

11.3.1 Video surveillance systems................................................................ 135

11.3.2 Assessment of wildlife use of underpasses........................................ 135

11.3.3 Modeling factors influencing wildlife underpass use ........................ 136

11.4 RESULTS ...................................................................................................... 137

11.4.1 Wildlife underpass use and passage rates .......................................... 137

11.4.2 Modeling of factors influencing wildlife underpass use.................... 138

11.4.3 Influence of underpass structure and placement................................ 144

11.4.4 Influence of video surveillance monitoring length ............................ 145

11.4.5 Influence of season ............................................................................ 146

11.4.6 Influence of time of day..................................................................... 147

11.5 DISCUSSION................................................................................................ 149

12.0 CONCLUSIONS AND RECOMMENDATIONS................................................. 153

12.1 Highway planning and monitoring ................................................................ 153

12.2 Role of riparian and meadow habitats ........................................................... 153

12.3 Wildlife underpasses...................................................................................... 154

12.4 Influence of traffic volume on wildlife.......................................................... 154

12.5 Wildlife permeability relationships................................................................ 155

12.6 Highway safety/ wildlife- vehicle collisions.................................................... 155

12.7 Role of ungulate- proof fencing...................................................................... 155

12.8 Future State Route 260 reconstruction sections............................................. 156

REFERENCES..................................................................................................................... 157

LIST OF FIGURES

Figure 3.1. Location of our study area and the five highway sections where phased highway

reconstruction has been ongoing since 2000, and the location of wildlife underpasses and

bridges........................................................................................................................ ................... 20

Figure 3.2. The existing narrow two- lane roadway ( left; Doubtful Canyon section) is being

reconstructed to a four- lane divided highway ( right; Preacher Canyon section), State Route 260,

Arizona........................................................................................................................ .................. 21

Figure 3.3. State Route 260 study area ( at the pedestrian/ wildlife underpass on the Christopher

Creek Section), Arizona................................................................................................................. 22

Figure 3.4. Aerial view of Little Green Valley riparian- meadow complex adjacent to the

Preacher Canyon section................................................................................................................ 22

Figure 3.5. Average annual daily traffic for State Route 260, Arizona. ( ADOT Control Road

monitoring station) for the period 1990– 2005. .............................................................................. 23

Figure 4.1. Little Green Valley riparian- meadow complex ( center photo) adjacent to State

Route 260 in Arizona, into which the west ( top photo) and east ( bottom photo) wildlife

underpasses open. .......................................................................................................................... 27

Figure 4.2. Layout ( top) of video surveillance system components at the west and east Little

Green Valley wildlife underpasses and the location of elk- proof and highway right- of- way

fencing, State Route 260, Arizona. ................................................................................................ 29

Figure 4.3. Frequency of elk crossing through both wildlife underpasses by time, crossing

toward and returning from the Little Green Valley meadow– riparian complex, State Route 260,

Arizona, determined from video surveillance from September 2002– September 2005................ 35

Figure 4.4. Frequency of elk observed crossing at the two Little Green Valley wildlife

underpasses by day determined by video surveillance from September 2002– September 2005,

and average daily traffic volume determined from a traffic counter along State Route 260,

Arizona, for the period December 2003- December 2004.............................................................. 36

Figure 4.5. Combined mean passage rate ( number of crossing elk/ number of approaching elk)

by month at the two Little Green Valley wildlife underpasses, State Route 260, Arizona,

determined by video surveillance conducted January 2003- September 2005. ............................ 37

Figure 4.6. Indian Gardens wildlife underpass on the Kohl’s Ranch section of State Route 260,

Arizona, completed in March 2006. .............................................................................................. 38

Figure 5.1. Proportion (± SE) of individual elk exhibiting: a) alert- hesitation, b) flight- retreat,

and c) feeding behavior when observed approaching five underpasses along State Route 260,

Arizona, at varying traffic levels. .................................................................................................. 45

Figure 5.2. Mean passage rates for elk during winter and summer ( solid line) and fall and

spring migration period ( dotted line) through five wildlife underpasses at varying traffic levels

along State Route 260, Arizona, 2003- 2005.................................................................................. 46

Figure 6.1. Cow elk caught in a Clover trap ( left) and being fitted with a GPS receiver collar,

along State Route 260, Arizona. .................................................................................................... 52

Figure 6.2. Highway segments ( 0.10 mi) delineated along State Route 260, Arizona, used to

compile highway crossings by elk, and the 0.15- mi distance buffer in which approaches to the

highway were determined. ............................................................................................................. 54

Figure 6.3. Minimum convex polygon ( MCP) home range ( White and Garrott 1990) adjacent

to the study area for bull elk no. 5 in which we generated the same number of random points

( dots) as successful Global Positioning System ( GPS) fixes ( n = 3,815) to compare distribution. 56

Figure 6.4. Mean frequency of observed Global Positioning System ( GPS) fixes and random

points generated for 33 elk fitted with GPS receiver collars that occurred in buffer zones within

0.6 mi of State Route 260, Arizona................................................................................................ 60

Figure 6.5. Frequency distribution of elk highway crossings by 0.1- mi segment, highway

section, and reconstruction class along State Route 260, Arizona, determined from 33 elk fitted

with Global Positioning System ( GPS) receiver collars. ............................................................... 61

Figure 6.6. Frequency distribution of elk highway crossings by 0.1- mi segment for Christopher

Creek section, State Route 260, Arizona, and projected proportions of total crossings

intercepted with planned and proposed elk- proof fence. ............................................................... 66

Figure 7.1. Location of wildlife underpasses and bridges along the Christopher Creek section

of State Route 260, Arizona, and delineation of different treatments to deter wildlife passage

onto the highway and funnel animals toward passage structures. ................................................. 70

Figure 7.2. Alternatives to fencing to deter at- grade wildlife crossings along the Christopher

Creek section of State Route 260, Arizona.................................................................................... 71

Figure 7.3. Number of elk- vehicle collisions recorded along the Christopher Creek section of

State Route 260, Arizona, in 2004 ( top; total 51 collisions) before ungulate- proof fencing was

erected and 2005 ( bottom; 12 collisions) after fencing was erected.............................................. 77

Figure 7.4. At- grade and below- grade ( through 6 wildlife underpasses) elk passage rates at

varying traffic volume levels along State Route 260, Arizona, ( figure from Gagnon et al.

2007c). ............................................................................................................................... ........... 80

Figure 8.1. Permanent traffic counting station installed along the Little Green Valley section of

State Route 260, Arizona, in December 2003................................................................................ 85

Figure 8.2. Mean probability that Global Positioning System ( GPS)- collared elk ( n = 44)

occurred within each 330- ft distance band from State Route ( SR) 260, Arizona, at varying

traffic volumes: a) < 100, b) 100- 200, c) 200- 300, d) 300- 400, e) 400- 500, f) > 600 vehicles/ hr. .87

Figure 8.3. Probabilities of elk crossing State Route ( SR) 260, Arizona, at varying traffic

volumes under different scenarios, derived from the best possible model selected by Akaike’s

Information Criterion. .................................................................................................................... 89

Figure 9.1. Bull elk killed in elk- vehicle collision along the Christopher Creek section of State

Route 260, Arizona, ( left) and a vehicle that struck an elk and sustained substantial property

damage and injured the driver........................................................................................................ 93

Figure 9.2. Number of elk- vehicle collisions ( 1994– 2006) and weighted elk highway crossings

for 33 elk fitted with Global Positioning System collars 2002– 2004 by 0.1- mi segments and

sections along State Route 260, Arizona. .................................................................................... 105

Figure 9.3. Number of elk- vehicle collisions recorded 2001– 2006 on the Preacher Canyon

( bottom), Christopher Creek ( middle), and control sections ( Little Green Valley and Doubtful

Canyon) of State Route 260, Arizona. ......................................................................................... 106

Figure 9.4. Actual elk- vehicle collisions recorded along State Route 260, Arizona, from 2001–

2006, compared to levels predicted from our multiple regression equation using average annual

daily traffic volume and elk population data. .............................................................................. 108

Figure 9.5. Number of elk- vehicle collisions along State Route 260, Arizona, by AADT levels

with and without measures to reduce collision incidence of including underpasses and fencing

assuming a 60% reduction in elk- vehicle collisions with measures, and the economic benefit

associated with the difference in the number of elk- vehicle collisions at varying AADT .......... 109

Figure 9.6. Coefficients of determination ( r2) for linear regression comparisons of elk- vehicle

collisions to elk crossings and weighted elk crossings conducted at various scales along State

Route 260............................................................................................................................ ........ 112

Figure 9.7. Frequency of elk- vehicle collisions and weighted elk crossings determined from 33

elk fitted with GPS collars 2002– 2004, by 0.6- mi sections along State Route 260. ................... 112

Figure 9.8. Proportions of elk- vehicle collisions ( solid line) and elk highway crossings ( dashed

line) by month along State Route 260, Arizona........................................................................... 114

Figure 9.9. Proportions of elk- vehicle collisions ( solid line) and elk highway crossings ( dashed

line) for bull elk by month along State Route 260, Arizona........................................................ 114

Figure 9.10. Elk- vehicle collision frequency by day and as corrected with daily AADT factors

accounting for differential daily traffic volume........................................................................... 115

Figure 9.11. Proportions of elk- vehicle collisions ( bars) and elk highway crossings ( dashed

line) by 2- hour time interval along State Route 260, Arizona..................................................... 116

Figure 9.12. Absolute departure ( by 0.5 hour increments) from sunrise or sunset for elk- vehicle

collisions along State Route 260.................................................................................................. 116

Figure 10.1. Schematic representation of the sampling block used to assess slope, aspect, and

canopy cover at 0.2- mi high and low frequency elk highway crossing sites along State Route

260, Arizona........................................................................................................................ ........ 127

Figure 10.2. Mean distance to permanent water sources and meadow habitats ( and SE bars)

from the center fn 0.2- mi elk highway crossing segments, comparing high and low frequency

crossing segments along State Route 260, Arizona. .................................................................... 128

Figure 10.3. Relationships between the mean frequency of weighted elk crossings at 0.2- mi

crossing segments along State Route 260, Arizona, and the distance to the nearest permanent

water sources ( dashed line) and meadow habitat ( solid line) ...................................................... 129

Figure 10.4. Classification and Regression Tree ( CART) modeling decision tree for variables

that described high and low frequency elk highway crossing sites, including proximity to

nearest permanent water, proximity to meadow habitat less than or greater than 680 ft, and the

proximity to meadow habitat less than or greater than 3,100 ft from 0.2- mi crossing segments

along State Route 260, Arizona ................................................................................................... 132

Figure 11.1. Aerial ( left) and entry ( right) views of the West Little Green Valley underpass

looking north.......................................................................................................................... ..... 139

Figure 11.2. Aerial ( left) and entry ( right) views of the East Little Green Valley underpass

looking north.......................................................................................................................... ..... 140

Figure 11.3. Aerial ( left) and entry ( right) views of the Pedestrian- Wildlife underpass looking

north. ............................................................................................................................... ............ 141

Figure 11.4. Aerial ( left) and entry ( right) views of the Wildlife 2 underpass looking north. ... 142

Figure 11.5. Aerial ( left) and entry ( right) views of the Wildlife 3 underpass looking north. ... 143

Figure 11.6. Number of elk groups that approached the five wildlife underpasses ( left) and

mean elk passage rate for the underpasses ( right) over three years of video surveillance

monitoring, State Route 260, Arizona. ........................................................................................ 146

Figure 11.7. Mean monthly elk passage rates for five underpasses at which video surveillance

monitoring has been ongoing for four years ( 2002– 2006) along State Route 260, Arizona. ...... 147

Figure 11.8. Number of elk underpass crossings ( left) and mean passage rate ( right) at five

underpasses along State Route 260, Arizona, determined from video surveillance 2002– 2006.148

Figure 11.9. Number of elk crossings ( left) and passage rate ( right) by time of day for five

underpasses along State Route 260, Arizona, determined from video surveillance 2002– 2006. 1148

Figure 11.10. The Indian Gardens wildlife underpass on the Kohl’s Ranch section ( looking

north from the east- bound lanes bridge), State Route 260, Arizona, where nearly all originally-planned

concrete walls were removed upon construction, widening the underpass floor for

wildlife passage, retaining native vegetation, and improving openness...................................... 151

LIST OF TABLES

Table 2.1. Dates that highway reconstruction was initiated and completed for the five

reconstruction sections on State Route 260, Arizona, and years of research accomplished

under various construction classes as part of research conducted 2002– 2006. ....................... 14

Table 3.1. State Route 260 reconstruction sections, reconstruction status, mileposts and

length, and the number of wildlife passage structures planned or built as part of the

reconstruction................................................................................................................. ......... 19

Table 4.1. Physical characteristics associated with the two wildlife underpasses ( UP) at which

we conducted video monitoring focusing on elk from September 2002– September 2005, State

Route 260, Arizona. ................................................................................................................. 26

Table 4.2. Results obtained from the Bradley- Terry logit model for paired preferences to

assess elk preference in selection of two wildlife underpasses ( UP) on State Route 260,

Arizona, for approaching and crossing. ................................................................................... 33

Table 4.3. Probabilities ( pe) of use of two wildlife underpasses ( UP) by elk groups and 95%

confidence intervals ( CI) obtained from multiple logistic regression given the combined

effects of underpass ( east versus west) and season ( summer [ April– September] versus

winter [ October– March]), and elk passage rates and two- proportion 95% CI for the same

underpass use categories. ......................................................................................................... 34

Table 4.4. Proportion of individual elk that displayed various behaviors near the two

wildlife underpasses ( UP) while approaching and crossing, and the associated 95% CI for

differences in the proportions ( Agresti 1996).......................................................................... 34

Table 5.1. Number of successful and unsuccessful elk crossings and passage rates ( no.

crossing/ no. approaching) by elk groups observed by video surveillance at five wildlife

underpasses along State Route 360, Arizona, at varying traffic levels.................................... 44

Table 5.2. Number of individual elk exhibiting flight responses while passenger vehicles

and semis passed overhead at low and high traffic levels during attempted crossings at five

wildlife underpasses along State Route 260 in central Arizona, 2003 – 2005. ....................... 45

Table 6.1. Highway crossings by section along State Route 260, Arizona, of 33 elk fitted

with Global Positioning System ( GPS) telemetry collars, including the length and

construction status of each section, number of elk that crossed the highway within each

section, and mean passage rate. ............................................................................................... 58

Table 6.2. Mean values we calculated by elk class for highway crossing, approach, and

passage rate parameters, and minimum convex polygon ( MCP) home ranges determined

from Global Positioning System ( GPS) telemetry along State Route 260, Arizona; and

results of ANOVA ( all df = 1, 31) tests of differences in means between cows and bulls. .. 59

Table 6.3. Mean crossings/ day and passage rates ( crossings/ approach) calculated by

highway reconstruction class along State Route 260, Arizona, from 33 elk fitted with Global

Positioning System ( GPS) telemetry collars............................................................................ 61

Table 6.4. Mean proportions of vegetation types comprising minimum convex polygon

( MCP home ranges, proportion of vegetation types within 0.6 mi of State Route 260,

Arizona, mean proportion of elk Global Positioning System ( GPS) fixes by vegetation type,

and chi- squared ( χ2) comparison of observed versus expected proportions of each

vegetation type used by 33 elk fitted with GPS collars. .......................................................... 63

Table 7.1. Physical characteristics associated with wildlife underpasses ( UP) and bridges

on the Christopher Creek section of State Route 260, Arizona, and whether video

surveillance of wildlife use was conducted at the passage structure. ...................................... 69

Table 7.2. Mean elk crossings/ day and passage rate ( crossings/ approach) for elk fitted with

Global Positioning System ( GPS) telemetry collars by highway reconstruction treatment,

Christopher Creek section, State Route 260, Arizona. ............................................................ 76

Table 7.3. Mean proportion of elk highway crossings along the Christopher Creek section,

State Route 260, before and after ungulate- proof fencing was erected and the mean

proportion of change ( Δ) with fencing..................................................................................... 78

Table 7.4. Mean collisions/ season ( 2002– 2006) by season and highway reconstruction class

along the Christopher Creek section, State Route 260, Arizona.............................................. 81

Table 8.1. Parameters for the only supported model ( best model) of 22 possible for the

probability of 40 elk crossing State Route ( SR) 260, Arizona, using Akaike’s Information

Criterion ( AIC), and its comparison to individual factors and the null model, - 2 log-likelihoods,

number of parameters ( k), AIC adjusted for small sample sizes ( AICc), AICc

difference ( ΔAICc), and Akaike weights ( wi).......................................................................... 88

Table 9.1. Frequency of wildlife- vehicle collisions by species and average annual daily

traffic ( AADT) volume for State Route 260, Arizona, and elk population estimates for

management units adjacent to SR 260, for the period 1994– 2006. ......................................... 99

Table 9.2. Number of animals killed in wildlife- vehicle collisions along State Route 260,

Arizona, by species documented by DPS and AGFD between 2001– 2006, with age and sex

of classified animals and proportion of classified animals. ................................................... 100

Table 9.3. Frequency of elk- vehicle collisions by State Route 260 highway section,

Arizona, recorded for the period 2001– 2006 by DPS and AFGD, and a comparison of the

total elk- vehicle collisions to the total in the ADOT database ( see Table 9.1) for the same

period ............................................................................................................................... ..... 101

Table 9.4. Proportion of single- vehicle accidents involving wildlife by State Route 260

highway section, Arizona, 1994– 2005................................................................................... 102

Table 9.5. Number of elk fitted with Global Positioning System ( GPS) collars versus those

that were killed in elk- vehicle collisions along State Route 260, Arizona, by highway

crossing frequency class. ....................................................................................................... 104

Table 9.6. Number of elk- vehicle collisions by State Route 260 highway section, Arizona,

1994– 2006, and mean collisions/ mi/ year (± SE) for each section.......................................... 107

Table 9.7. Dates of construction initiation and completion for SR 260 highway sections,

Arizona, and mean number of elk- vehicle collisions ( EVC) from 1994– 2006 (± SE) by

highway construction classes ( before, during, and after reconstruction)............................... 108

Table 9.8. Summary of elk crossings, Shannon Diversity Index ( SDI), and weighted

crossings by highway section along State Route 260, Arizona, determined from 33 elk fitted

with GPS telemetry collars, May 2002– April 2004............................................................... 109

Table 9.9. Elk- vehicle collision ( EVC) relationships between highway crossings and

weighted crossings by 33 Global Positioning System- collared elk at various scales along

State Route 260, including correlation coefficients ( r) and coefficients of determination ( r2). 111

Table 10.1. G2- statistics associated with five variables used in the Classification and

Regression Tree modeling of high and low frequency elk crossing segments along State

Route 260, Arizona. ............................................................................................................... 131

Table 11.1. Structural characteristics associated with wildlife underpasses on State Route

260, Arizona, at which video camera surveillance was conducted from 2004– 2006 to assess

wildlife use, and the year in which underpass construction was completed and monitoring

initiated. ............................................................................................................................... . 135

Table 11.2. Number of animals by species recorded by video cameras at the West Little

Green Valley underpass, number crossing through the underpass, and the passage rate ( no.

crossing/ no. approaching). ..................................................................................................... 139

Table 11.3. Number of animals by species recorded with video surveillance at the East

Little Green Valley underpass, number crossing through the underpass, and the passage rate

( no. crossing/ no. approaching)............................................................................................... 140

Table 11.4. Number of animals by species recorded with video surveillance at the

Pedestrian- Wildlife underpass, number crossing through the underpass, and the passage rate

( no. crossing/ no. approaching)............................................................................................... 141

Table 11.5. Number of animals by species recorded with video surveillance at the Wildlife

2 underpass, number crossing through the underpass, and the passage rate ( no. crossing/ no.

approaching). ......................................................................................................................... 142

Table 11.6. Number of animals by species recorded with video surveillance at the Wildlife

3 underpass, number crossing through the underpass, and the passage rate ( no. crossing/ no.

approaching). ......................................................................................................................... 143

Table 11.7. Likelihood ratio test results for five factors modeled by multiple logistic

regression analysis for successful elk crossing at five underpasses along State Route 260,

Arizona, assessed by video camera surveillance conducted 2004- 2005, including logistic

regression chi- square ( χ 2) statistic, probability, and degrees of freedom ( DF)..................... 144

Table 11.8. Probability of a successful elk crossing at five wildlife underpasses along State

Route 260, Arizona, determined by logistic regression......................................................... 145

Table 11.9. Comparison of odds of a successful crossing at 5 wildlife underpasses along SR

260 in central Arizona. The number on the left side of each ratio is associated with the

structures listed in column one .............................................................................................. 145

Table 11.10. Probability of a successful elk crossing by year at five wildlife underpasses

along State Route 260, Arizona, determined by logistic regression ...................................... 146

Table 11.11. Probability of a successful elk crossing by season at five wildlife underpasses

along State Route 260, Arizona, USA, determined by logistic regression ............................ 147

Table 11.12. Probability of a successful elk crossing by time class at five wildlife

underpasses along State Route 260, Arizona, determined by logistic regression.................. 148

TABLE OF SPECIES

Animals

Black bear Ursus americanus

Caribou Rangifer tarandus

Elk Cervus elaphus

Grizzly bear Ursus arctos

Javelina Tayassu tajacu

Moose Alces alces

Mountain goats Oreamnos americanus

Mountain lion Puma concolor

Mule deer Odocoileus hemionus

Pronghorn Antilocapra americana

Rocky Mountain elk Cervus elaphus nelsoni

White- tailed deer Odocoileus virginianus cousei

Wolf Canis lupus

Plants

Douglas fir Pseudotsuga menzeisii

Gambel oak Quercus gambelii

Juniper Juniperus spp.

Manzanita Arctostaphalos pungens

Pinyon Pinus edulis

Ponderosa pine Pinus ponderosa

White fir Abies concolor

ACKNOWLEDGMENTS

This study was funded by the Arizona Department of Transportation’s ( ADOT) Arizona

Transportation Research Center ( ATRC), and the Federal Aid Wildlife in Restoration Act,

Project W- 78- R supporting Arizona Game and Fish Department ( AGFD) research. The

Tonto National Forest ( TNF) and Federal Highway Administration ( FHWA) provided

additional funding to the project that made our application of Global Positioning System

telemetry possible. We thank Terry Brennan, Robert Ingram, and Duke Klein of the TNF,

and Paul Garrett and Steve Thomas of FHWA for their commitment to making this project

possible.

Many individuals at ADOT provided endless support and guidance in this project and were

instrumental to its success, especially Estomih Kombe, Bruce Eilerts, Siobhan Nordhaugen,

and Doug Brown ( now at the Arizona Department of Administration). Mark Catchpole,

Doug Eberline, Jami Rae Garrison, and Dale Buskirk of the Transportation Planning

Division provided invaluable traffic data and support. We sincerely thank Tom Foster,

Myron Robison, David Gerlach, William Pearson, James Laird, Tom Goodman, Jack

Tagler, and Dallas Hammit of the Prescott District for their commitment to adaptive

management and willingness to respond to our data and recommendations. Though often at

a cost of valuable time and limited funds, their commitment maximized the effectiveness of

wildlife structures and highway safety.

The cooperation of John Anderson, Walt Cline, Bob Ochoa ( Boy Scouts of America),

Mikey Marazza, and Tom Dunney ( Arizona State University), allowing us to trap elk on

private lands, contributed greatly the success of the Global Positioning System telemetry

portion of our study.

The AGFD’s Mesa Region played a crucial role in the project, especially key personnel:

Tim Holt, Henry Apfel, John Dickson, Craig McMullen, and Jon Hanna. Research Branch

personnel Kari Ogren, Fenner Yarborough, Scott Sprague, and Tim Rogers assisted with the

laborious task of viewing and analyzing videotape and keeping our video camera

surveillance systems “ up and running.” The logistical support provided by Tonto Creek

Fish Hatchery personnel was invaluable to our project, particularly the hospitality and

assistance provided by Larry Peterson to whom we dedicate this report, John Diehl, Larry

Duhamell, Mike Weisser, and Trevor Nelson.

We offer a special thanks to the Arizona Department of Public Safety Highway Patrolmen

in the Payson District whose efforts to document wildlife- vehicle collisions were

instrumental not only to the success of our project, but invaluable in helping resolve

wildlife- vehicle conflicts across Arizona, making its highways safer.

The ATRC’s Technical Advisory Committee for this project provided many suggestions

toward improving its effectiveness and applicability. Their tremendous support, oversight,

and commitment throughout the duration of the project were greatly appreciated.

1

1.0 EXECUTIVE SUMMARY

We studied wildlife- highway relationships from 2002– 2006 along a 17- mi stretch of

State Route ( SR) 260, in central Arizona, USA. This stretch is being reconstructed from

a two- lane to a four- lane divided highway in five phases incorporating 11 wildlife

underpasses and six bridges. Phased reconstruction allowed us to use a before- after-control

experimental approach in our research. The objectives of our research were to:

• Assess and compare wildlife use of underpasses.

• Evaluate highway permeability and wildlife movements among reconstruction

classes.

• Characterize wildlife- vehicle collision patterns and changes with reconstruction.

• Assess relationships among highway traffic volume and wildlife- vehicle

collisions, elk crossing patterns, and wildlife use of underpasses.

• Assess the role that ungulate- proof fencing plays in wildlife- vehicle collisions,

wildlife use of underpasses, and highway permeability to wildlife.

• Provide ongoing highway reconstruction implementation guidance.

We used video camera surveillance to assess and compare wildlife use of five underpasses.

We recorded 8,455 animals and 11 different species; 5,560 of these animals ( 65.8%)

crossed through the underpasses. Our underpass passage rates ranged from 0.10– 0.68

crossing/ approach, and underpass structure and placement was the most important factor of

five modeled influencing the probability of successful underpass crossing by elk ( Cervus

elaphus nelsoni). We used Global Positioning System ( GPS) telemetry data collected from

65 elk fitted with GPS receiver collars to assess movement patterns and highway

permeability. Our elk crossed SR 260 5,749 times. Elk permeability on reconstructed

highway ( 0.43 crossings/ approach) was half that of the control sections. Permeability

increased 60% after ungulate- proof fencing was erected on a reconstructed section, linking

underpasses. Fencing also resulted in > 80% reduction in elk- vehicle collisions and

improved underpass effectiveness as elk and deer ( Odocoileus spp.) underpass passage rates

increased from 0.12 to 0.56 crossings/ approach after fencing. While we found that traffic

volume did affect elk highway crossing and distribution patterns at highway grade, it had

little or no impact on elk crossing below grade through underpasses. We assessed spatial

and temporal patterns of elk- vehicle collisions ( n = 571). Annual elk- vehicle collisions

were related to traffic volume and elk population levels. Mean elk- vehicle collisions during

reconstruction ( 11.6/ year) was higher than before ( 4.4/ year) and after reconstruction

( 6.5/ year) for each section. The benefit of reduced elk- vehicle collisions from underpasses

and fencing was projected to approach $ 1 million/ year. We compared elk- vehicle collisions

and elk GPS crossings at five scales and found that the strength of the relationship and

management utility were optimized at the 0.6- mi scale. The proximity of riparian- meadow

habitats and permanent water along SR 260 influenced the pattern of elk GPS crossings and

2

elk- vehicle collisions. Together, effective monitoring and adaptive management improved

both highway safety and highway permeability to wildlife.

We report our research findings in eight separate chapters or volumes addressing various

aspects of our research. A summary for each volume follows below, as well as an overall

summary of project conclusions and recommendations.

1.1 VIDEO SURVEILLANCE TO ASSESS HIGHWAY UNDERPASS USE BY

ELK

We used integrated video systems to compare wildlife use of two bridged wildlife

underpasses on the Preacher Canyon section from September 2002– September 2005.

Both underpasses opened into the same riparian- meadow complex, were situated < 850 ft

apart, and had different below- span characteristics and dimensions. Our objectives were

to compare elk response to the underpasses and test hypotheses that passage rate,

probability of use, and behavioral response at the two underpasses did not differ. We

related differences in elk use and response to underpass design characteristics. Elk

accounted for > 90% of the animals we recorded on videotape, with 3,708 elk in 1,266

groups recorded at the two underpasses. We used multiple logistic regression to predict

the probability of underpass use by elk incorporating the combined effects of underpass,

season, and year. Season had the greatest effect on underpass use, with the probability of

underpass use in summer ( 0.81) higher than winter ( 0.58) when migratory elk less

habituated to the underpasses were present. A pattern of high summer (> 0.80) and low

winter passage rates (< 0.40), regardless of underpass, existed in all three years of video

surveillance. Underpass design characteristics also had an effect on the probability of elk

crossing the underpass; the probability of use of the underpass with two times the

openness ratio, half the length for elk to traverse, and sloped earthen sides ( 0.75) was

higher than the neighboring underpass with concrete walls ( 0.66). Proportions of elk

displaying behaviors indicative of resistance to crossing were dependent on underpass

and were higher at the underpass with concrete walls. In all cases, elk preferred the more

open underpass with natural earthen sides. We believe that differences in underpass

length and the concrete walls contributed to differences in elk use and behavioral

response. Continued video surveillance of these and other underpasses will allow us to

evaluate their efficacy in promoting wildlife permeability.

1.2 EFFECTS OF TRAFFIC VOLUME ON ELK USE OF WILDLIFE

HIGHWAY UNDERPASSES

Structures that allow wildlife to cross the highway corridor are increasingly used to

mitigate potential negative impacts of roadways, but little is known about how varying

traffic levels may limit their effectiveness, either by reducing wildlife passage rates or by

causing animals to cross highways at other locations where they could potentially cause

collisions. We monitored five wildlife crossings SR 260 using video surveillance to

determine if traffic levels or traffic types ( semi- trailer truck versus automobile) affected

elk use of wildlife underpasses. We examined elk crossing behavior at wildlife

underpasses at two critical points during the crossing period: 1) when elk initially

3

approached the underpass, and 2) after elk entered the underpass. Passage rates at low,

intermittent traffic volume ( 0- 2 vehicles/ min = 0.59, 2- 4 vehicles/ min = 0.75) and at

higher traffic levels ( 4- 6 vehicles/ min = 0.73, > 6 vehicles/ min = 0.71) were not markedly

reduced compared to passage rates when no vehicles were present ( 0.65). Passage rates

varied seasonally, likely due to the presence of migratory elk unused to underpasses

during part of the year, but even during migratory periods traffic level had minimal effect

on the passage rate. Thus, increasing traffic did not substantially reduce the effectiveness

of wildlife underpasses as viable means of mitigating wildlife population fragmentation,

at least at the traffic levels we studied. Semis were five times more likely than were

passenger vehicles to cause a flight behavior when traffic levels were intermittent versus

when traffic was continuous, possibly due to the sudden increase in sound and vibration.

If flight away from underpasses causes animals to cross the highway at other points and

thereby increase the potential for elk- vehicle collisions, measures that reduce traffic noise

and visual stimuli caused by passing vehicles at underpasses may be warranted.

1.3 ASSESSMENT OF ELK HIGHWAY PERMEABILITY USING GLOBAL

POSITIONING SYSTEM TELEMETRY

Highways have significant direct and indirect impact to natural ecosystems, including

wildlife barrier and fragmentation effects, resulting in diminished habitat connectivity

and highway permeability. We used GPS telemetry to assess elk movement patterns and

permeability across SR 260. The highway was reconstructed in phases, allowing for

comparison of highway crossing and passage rates during various stages of

reconstruction. We instrumented 33 elk ( 25 female, 8 male) with GPS receiver collars

May 2002- April 2004. Our collars accrued 101,506 GPS fixes with 45% occurring

within 0.6 mi of the highway. Nearly two times the proportion of locations occurred

within 0.6 mi of the highway compared to randomly generated locations. We believe elk

were attracted to the highway corridor by riparian- meadow foraging habitats that were

seven times more concentrated within the 0.6- mi zone around the highway compared to

the mean proportion within elk use areas encompassing all GPS fixes. Elk crossed the

highway 3,057 times; crossing frequency and distribution along the highway was

aggregated compared to random. Crossing frequency within 0.1- mi highway segments

was negatively associated with the distance to riparian- meadow habitats. Mean observed

crossing frequency ( 92.6 crossings/ elk) was lower than random ( 149.6 crossings/ elk).

Cows crossed 4.5× as frequently as bulls. Highway permeability among reconstruction

classes was assessed using passage rates ( ratio of highway crossings to approaches); our

overall mean passage rate was 0.67 crossings/ approach. The mean passage rate for elk

crossing the highway section where reconstruction was completed ( 0.43

crossings/ approach) was half that of sections under reconstruction and control sections

combined ( 0.86 crossings/ approach). Permeability was jointly influenced by the size of

the widened highway and associated vehicular traffic on all lanes. Crossing frequency

was used to delineate where ungulate- proof fencing yielded maximum benefit in

intercepting and funneling crossing elk toward underpasses and reducing elk- vehicle

collisions. Use of passage rates provides a quantitative measure to assess highway

permeability, conduct future before and after construction comparisons, and develop

mitigation strategies to minimize the impact of highways on wildlife.

4

1.4 ROLE OF FENCING IN PROMOTING WILDLIFE UNDERPASS USE

AND HIGHWAY PERMEABILITY

Ungulate- proof fencing has been used successfully to mitigate the incidence of wildlife-vehicle

collisions on highways throughout North America. While fencing is often

regarded as an integral component of effective wildlife passage structures, limited

information or guidelines exist for the application of fencing in conjunction with wildlife

passages. Fencing itself may limit wildlife permeability across highways and exacerbate

the barrier effect of highways on wildlife populations. The five- mile reconstructed

Christopher Creek section was opened to traffic six months before ungulate- proof fencing

was erected, linking four wildlife underpasses and three bridges. To assess the role of

strategically placed fencing along 49% of the section, we compared before and after

fencing elk- vehicle collision incidence, wildlife use of underpasses, and elk highway

permeability. From 2002– 2006, we documented 110 elk- vehicle collisions. The

incidence of collisions increased over three fold after highway reconstruction was

completed but before fencing was erected. After fencing, the incidence of elk collisions

declined 87%. We employed video camera surveillance systems at two underpasses to

compare wildlife use for nine months before and 11 months after fencing was erected.

Before fencing, we recorded 500 elk and deer at the underpasses, of which only 12%

successfully passed through the underpasses; 81% of animals continued to cross the

highway at grade. After fencing, of 595 elk and deer recorded, 56% crossed successfully

and no animals crossed the highway at grade. The probability of an approaching animal

crossing through an underpass increased from 0.09 to 0.56 with fencing, and the

combined odds of a crossing through the underpass after fencing was 13.6: 1 compared to

before fencing. We used GPS telemetry to assess highway permeability and crossing

patterns. We instrumented 22 elk ( 16 female, 6 male) with GPS receiver collars April

2004– October 2005, during which time our collars accrued 87,745 GPS fixes. The elk

highway passage rate after the highway was opened to traffic but before fencing was

erected ( 0.54 crossings/ approach) was 32% lower than the level determined from a

previous study for the section during reconstruction ( 0.79 crossings/ approach). Once

fencing was erected, the passage rate increased 52% to 0.82 crossings/ approach. The

proportion of elk crossings that occurred along fenced highway stretches declined 50%

while the proportion of crossings along unfenced highway increased 40%. Fencing plays

an important role in reducing the incidence of wildlife- vehicle collisions and increasing

the effectiveness of wildlife passage structures. Furthermore, fencing in combination

with a relatively high density of passages ( one structure/ 0.7 mi) promoted elk highway

permeability by funneling animals toward the underpass where resistance to crossing was

lower than that associated with crossings at grade.

1.5 INFLUENCE OF FLUCTUATING TRAFFIC VOLUME ON ELK

DISTRIBUTION AND HIGHWAY CROSSINGS

We linked 38,709 GPS locations collected from December 2003– June 2006 from 44 elk

fitted with GPS collars and hourly SR 260 traffic data, totaling more than 6,470,000

vehicles to determine how elk distribution varied with traffic volume, and how traffic

5

volume related to highway crossings. The probability of elk occurring near the highway

decreased with increasing traffic volume, indicating that habitat near the highway was

used by elk, but primarily when traffic volumes were relatively low (< 100 vehicles/ hr).

We used multiple logistic regression combined with Akaike’s Information Criteria to

identify factors potentially important in influencing probability of elk crossing the

highway. We found that increasing traffic volume reduced the overall probability of

highway crossing, but this effect depended on both season and the proximity of riparian-meadow

habitat, with elk crossing highways at higher traffic volumes during spring and

fall migratory periods and when accessing these riparian- meadow foraging areas.

Overall, our results indicate that: 1) fluctuations in traffic volume should be considered in

models of habitat effectiveness for elk, 2) the effect of traffic volume on probability of

highway crossing, and therefore highway permeability, will depend on how a highway

affects access to daily and seasonally important resources, and 3) increased traffic

volume alone will not prevent elk from crossing the roadway and therefore development

of effective wildlife passages or motorist warning signs could reduce the probability of

elk- vehicle collisions, especially during migratory periods if placed near riparian-meadow

habitat or other areas with preferred resources.

1.6 CHARACTERISTICS OF ELK- VEHICLE COLLISIONS AND

COMPARISON TO HIGHWAY CROSSING PATTERNS

We assessed spatial and temporal patterns of elk- vehicle collisions from 1994- 2006 ( n =

571) along SR 260. We used GPS telemetry to assess spatial and temporal patterns of elk

highway crossings and compare to elk- vehicle collisions patterns. Annual elk- vehicle

collisions were related to traffic volume and elk population levels. elk- vehicle collisions

occurred in a non- random pattern. With three of the sections completed, mean elk-vehicle

collisions while they were under reconstruction ( up until ungulate- proof fencing

was erected; 11.6/ year) was higher than mean before- construction elk- vehicle collisions

( 4.4/ year) and after reconstruction ( 6.5/ year) elk- vehicle collisions for each section. On

the first section completed in 2001 with limited fencing ( 13%), elk- vehicle collisions did

not differ among before, during, and after construction classes, even though mean traffic

volume increased 67% from before- to after- construction levels, pointing to the benefit of

three passage structures and fencing. On another section completed in 2004, elk- vehicle

collisions increased > 2.5× when opened to traffic but before fencing was erected; elk-vehicle

collisions dropped > 70% once fencing was installed. The benefit associated with

reduced elk- vehicle collisions from underpass and fencing was projected to approach $ 1

million/ year. We accrued 101,506 fixes from 33 elk ( 25 females, 8 males) fitted with

GPS collars 2002- 2004. Elk crossed the highway 3,057 times in a non- random pattern.

We compared elk- vehicle collisions and crossings at five scales; the strongest

relationship was at the highway section scale. Strength of the relationship and

management utility were optimized at the 0.6- mi scale. Elk- vehicle collision frequency

was associated with proximity to riparian- meadow habitats adjacent to the highway at the

section and 0.6- mi scales. Though both fall elk- vehicle collisions and crossings exceeded

expected levels, the proportion of elk- vehicle collisions in September- November ( 49%)

exceeded the proportion of crossings and coincided with the breeding season, migration

of elk from summer, and high use of riparian- meadow habitats adjacent to the highway.

6

There was no difference in the proportion of elk- vehicle collisions and crossings by day,

as both reflected avoidance of crossing the highway during periods of highest traffic

volume. Though traffic volume was highest from Thursday- Saturday, the proportion of

elk- vehicle collisions was lower than expected. A higher proportion of elk- vehicle

collisions ( 59%) occurred relative to crossings ( 33%) in the evening ( 1700- 2300 hr); 34%

of elk- vehicle collisions occurred within a one- hour departure of sunset, and 55.5%

within a two- hour departure. Elk- vehicle collision data are valuable in developing

strategies to maintain permeability and increase highway safety including selecting

locations of passage structures.

1.7 INFLUENCE OF ENVIRONMENTAL FACTORS ON ELK HIGHWAY

CROSSINGS

Vehicle collisions with ungulates are recognized as a serious problem because they

threaten human safety and cause tremendous property damage, as well as impact wildlife

populations. Such risks to humans and wildlife make it important to understand how

environmental factors influence ungulates highway crossing patterns. Several studies

have described site- specific variables at ungulate- vehicle accident sites, but none have

used highway crossing data. Also, previous studies lack information on how riparian-meadow

habitats influence road crossings. We used GPS data from 33 elk collared from

2002– 2004 to determine where they crossed SR 260. Our GPS collars yielded more than

101,000 GPS fixes from which we determined 3,057 crossings of the highway. We

delineated 90 0.2- mi segments along a 17- mi stretch of the highway and calculated

weighted elk crossings associated with each segment. We selected the 20 0.2- mi

segments that exhibited the highest and the 20 with the lowest weighted elk crossing

frequencies, and measured various habitat factors associated with these segments. To

assess the influence of habitat factors on elk highway crossing patterns, we conducted

field validations and Geographic Information System analysis to model five

environmental parameters: 1) proximity to nearest meadow, 2) proximity to nearest

permanent water source, 3) forest canopy, 4) slope, and 5) aspect. We employed a

Classification and Regression Tree ( CART) approach to determine the hierarchical order

of importance for each variable tested for both our high and low frequency elk highway

crossing segments. Proximity to water and meadow were tied for most influential factors

associated with high frequency weighted crossing sites. The CART- derived

classification tree indicated that water occurred less than 2,500 ft from all high crossing

segments, and meadow occurred less than 3,100 ft from 85% of the high crossing sites.

The results from this study provide important information that can be used to mitigate the

incidence of wildlife- vehicle collisions and design safe and ecologically sensitive

highways.

1.8 PRELIMINARY ASSESSMENT OF FACTORS INFLUENCING

WILDLIFE USE OF HIGHWAY UNDERPASSES

We assessed and compared wildlife use of five underpasses on the Preacher Canyon and

Christopher Creek sections of SR 260, using data from video camera surveillance

conducted 2002– 2006. Our video surveillance systems were designed to capture animals

approaching and crossing though the underpasses, allowing us to measure passage rates

7

( crossings/ approach). We recorded 8,455 animals and 11 different species on 1,100

hours of videotape. Overall, 5,560 of these animals, or 65.8%, crossed through the

underpasses. Elk accounted for the majority of the animals documented at the

underpasses ( 73.8%), while white- tailed deer and mule deer accounted for 10.9% and

7.4% of the total, respectively. Our mean elk underpass passage rates ranged from 0.10–

0.68 crossing/ approach for the five monitored underpasses. We used multiple logistic

regression to select factors important in predicting probability of a successful crossing

through the underpasses by elk; we modeled the influence of underpass ( structure and

placement), season, length of monitoring, time of day, and day of the week. We used a

general linear model with a logit link to determine probabilities of a successful crossing

for each of the factors selected. We found that four factors were important in predicting

the probability of a successful crossing once elk approached the underpass. Our

underpass factor was the most important one, suggesting that underpass structure and

placement was of primary importance in predicting the probability of successful elk

passage at the underpass, with probabilities of successful elk crossing ranging from 0.09–

0.77. The length of time an underpass was monitored was the second most important

factor selected in our logistic regression modeling ( with the probability of successful

crossing increasing from 0.52 in the first year of monitoring to 0.69– 0.71 in the

subsequent four years), followed closely by season and time of day. The probability of

successful elk crossing during fall, when migratory elk not habituated to our underpasses

were present along SR 260 was 0.59 compared to 0.71 during summer. Day of the week,

our surrogate factor for traffic volume did not have a significant influence on crossing

probabilities at our below grade passage structures.

1.9 OVERALL CONCLUSIONS AND RECOMMENDATIONS

Our research underscored the ability to integrate transportation and ecological objectives

into highway construction activities, yielding tangible benefits to both highway safety

and wildlife permeability. The combination of phased construction, adaptive

management, and effective monitoring of measures to reduce wildlife- vehicle collisions

and promote permeability were instrumental to achieving transportation and ecological

objectives. We recommend that such an approach to highway construction and

monitoring be pursued elsewhere in the state whenever possible. In the instance of SR

260, ADOT prioritized the reconstruction of the five sections on the historic incidence of

wildlife- vehicle collisions. Our research validated this prioritization; the strong

association between elk- vehicle collisions and highway crossings underscored the utility

and value of wildlife- vehicle collision data in planning wildlife mitigation measures

ranging from passage structures to ungulate- proof fencing. We strongly recommend that

all work units within ADOT and other agencies make a concerted effort to collect and

archive wildlife- vehicle collision data throughout Arizona, utilizing the standardized

interagency collision report form.

We found that the presence of riparian and wet meadow habitats constituted the “ engine”

that drove conflicts between the highway and wildlife along SR 260. Elk- vehicle

collision and elk highway crossing patterns were closely associated with proximity to

riparian- meadow habitats. Future highway construction activities should avoid such

8

limited, valuable habitats where possible. Wildlife underpasses located adjacent to

riparian- meadow habitats received high levels of use by wildlife due to their movement

toward these preferred foraging areas, as well as animal propensity to travel along

drainages. Where highway alignments near riparian- meadow habitat are unavoidable,

such sites are excellent locations to consider wildlife passage structures.

Wildlife underpasses were highly effective in promoting below- grade wildlife crossings,

with two- thirds of over 8,500 animals recorded during video surveillance having crossed

through an underpass. These underpasses were instrumental to improving highway safety

through reduction of wildlife- vehicle collisions and promoting wildlife permeability.

Structural design characteristics and placement of underpasses are important

considerations to maximizing their efficacy in promoting wildlife passage, and structural

characteristics were the most important factor in determining the probability of successful

crossings by wildlife. Underpass openness is crucial to achieving high probability of

successful underpass use. The distance that animals must travel through an underpass is

an especially important factor in maximizing efficacy, and should be minimized in

underpass design. Elk avoided an underpass where concrete mechanically stabilized

earth walls were erected for soil stabilization, compared to a neighboring underpass with

more natural 2: 1 sloped earthen sides. We recommend that the application of concrete

walls be avoided in wildlife underpasses. Visibility through underpasses should be

maximized during design and implementation. Where underpasses occur on divided

highways we recommend that the bridges be placed in line where possible to maximize

visibility by animals through the structures. Wildlife underpass placement should avoid

areas of high human activity or congregation that occur outside daytime hours.

We documented a recurring seasonal pattern where elk underpass passage rates dropped

from summer levels > 0.90 crossings/ approach to below 0.40 during the fall when

migratory elk moved through the SR 260 corridor. Migratory elk do not appear to exhibit

the same propensity for habituation to underpasses as resident elk. Additional ungulate-proof

fencing may be needed to address this seasonal drop in underpass passage rates.

Long term monitoring will provide valuable insights on changes in wildlife use patterns.

Ungulate- proof fencing in conjunction with underpasses will expedite the wildlife

learning process.

Traffic levels fluctuated greatly on an hourly, daily and seasonal basis through our study

area, averaging between 7,000– 8,500 average annual daily traffic ( AADT). We found

that traffic volume influenced elk crossing patterns and distribution at highway grade.

With increasing traffic levels, we found reduced probability of successful elk highway

crossings at grade, crossings occurred later in the evening when volume levels abated,

and elk moved away from the highway as volumes increased. Unsuccessful attempts to

cross SR 260, or “ repels” typically coincided with high traffic volume. Conversely, at

our monitored wildlife underpasses, traffic volume on SR 260 overhead did not have an

effect on elk approaching and successfully crossing through the underpasses below grade.

This finding was of paramount importance to understanding the efficacy of underpasses

in promoting wildlife permeability.

9

GPS telemetry afforded us an unprecedented opportunity to assess and compare wildlife

permeability among highway reconstruction classes, as well as assess permeability before

and after the erection of ungulate- proof fencing. Reconstruction from a two- lane to four-lane

divided highway reduced wildlife permeability by half compared to that of our

control sections. On one section, the during reconstruction passage rate dropped 34%

after reconstruction but before fencing was erected. Yet the elk passage rate increased

54% after half of the section was strategically fenced with ungulate- proof fencing. Thus,

fencing in conjunction with underpasses promoted wildlife permeability as animals were

funneled toward underpasses and bridges where they crossed below grade with minimal

impact from traffic passing above. In addition to playing an instrumental role in

promoting permeability, ungulate- proof fencing was crucial to achieving effective use of

underpasses, especially those not located in proximity to meadow habitats. Without

fencing, elk and deer continued to cross SR 260 at grade immediately adjacent to

underpasses. The 50% of the section that was fenced was projected to intercept 89% of

elk crossings determined from GPS telemetry, and yielded an 83% reduction in elk-vehicle

collisions in the year after fencing. Fencing is an integral component of wildlife

mitigation measures in reducing elk- vehicle collisions and promoting wildlife

permeability.

With two of the five SR 260 sections reconstructed to date integrating underpass and

ungulate- proof fencing, 2006 was the first year that the incidence of actual elk- vehicle

collisions dropped below the level predicted from modeling based on average annual

daily traffic ( AADT) volume and elk population levels. Our model predicted even

greater benefit as AADT increases. Thus, the complement of measures implemented to

date has achieved its objective in mitigating the impact of highway reconstruction and

increasing traffic volume, and the benefit is expected to grow now that the third section is

complete ( Kohl’s Ranch) and the entire Preacher Canyon section is being fenced under

an enhancement grant project. With only a modest increase in AADT, we estimated an

annual benefit from reduced elk- vehicle collisions of nearly $ 1 million/ year.

Compared to the first three reconstructed sections, the remaining two exhibited relatively

few wildlife- vehicle collisions or collared elk crossings. The exception is the limited

areas where riparian- meadow habitat is located in close proximity to the highway.

10

11

2.0 INTRODUCTION

2.1 BACKGROUND

Highways directly and indirectly create some of the most prevalent and widespread

changes to the ecosystem in the United States ( Noss and Cooperrider 1994, Trombulak

and Frissell 2000, Farrell et al. 2002). The estimated 500,000 to 700,000 deer killed each

year in collisions on U. S. highways directly affect the ecosystem ( Romin and Bissonette

1996a, Schwabe and Schuhmann 2002). Collisions also cause human injuries, deaths,

and tremendous property loss ( Reed et al. 1982, Schwabe and Schuhmann 2002), and

disproportionately affect threatened or endangered species ( Foster and Humphrey 1996).

Highways indirectly impact ecosystems by causing habitat loss and blocking animal

movements. Forman and Alexander ( 1998) estimated that highways have caused habitat

loss and degradation in more than 20% of the U. S. Blocking of animal movements

between seasonal ranges or other vital habitats is perhaps highways’ most pervasive

environmental impact. ( Noss and Cooperrider 1994, Forman and Alexander 1998,

Forman 2000). Their fragmentation of habitats and populations reduces genetic

interchange ( Gerlach and Musolf 2002) and limits dispersal of young ( Beier 1995); all

serving to disrupt viable wildlife population processes. Long- term fragmentation and

isolation renders populations more vulnerable to catastrophic events and may lead to

extinctions ( Hanski and Gilpin 1997). Fencing to prevent wildlife and livestock access to

highways may exacerbate the barrier effect unless provision is made for passage.

Though numerous studies have alluded to highways’ barrier effects on wildlife ( e. g., see

Forman et al. 2003), few have yielded quantitative data on animal passage rates,

particularly in an experimental context ( e. g., pre- and post- highway construction). Many

studies have focused on the efficacy of passage structures at allowing wildlife to avoid at-grade

crossings ( Clevenger and Waltho 2003, Ng et al. 2004) or have relied on modeling

to assess highways' passability, or permeability to wildlife ( Singleton et al. 2002).

Assessments of the habitat fragmentation highways cause for relatively low- mobility

small mammals have yielded quantifiable results from mark- recapture trapping, but

assessments for larger, far- ranging species have been limited by cost ( Swihart and Slade

1984, Conrey and Mills 2001, McGregor et al. 2003). Paquet and Callaghan ( 1996) used

winter track counts adjacent to highways and other barriers to determine passage rates by

wolves ( Canis lupus), something few other studies have reported. 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), but seldom directly addressing permeability as Gibeau et al. ( 2001) did for grizzly

bears ( Ursus arctos).

Numerous studies have been conducted on the spatial and temporal patterns of wildlife-vehicle

collisions, most focusing on deer ( Reed and Woodard 1981, Bashore et al. 1985,

Romin and Bissonette 1996b, Hubbard 2000). Only recently have researchers

specifically addressed patterns of collisions with elk ( Cervus elaphus) ( Gunson and

Clevenger 2003, Biggs et al. 2004). Insights gained from such studies have been

12

instrumental in developing strategies to reduce collisions with wildlife ( Romin and

Bissonette 1996a, Farrell et al. 2002), including planning passage structures to reduce at-grade

crossings and maintain passage ( Clevenger et al. 2002).

Consistent tracking of wildlife- vehicle collisions is a valuable tool to assess the impact of

highway construction on wildlife ( Romin and Bissonette 1996b) and the efficacy of

passage structures and other measures ( e. g., fencing) in reducing wildlife- vehicle

collisions ( Reed and Woodard 1981, Ward 1982, Clevenger et al. 2001a). Though data on

wildlife- vehicle collisions is valuable, no study has investigated or validated the

relationships between these collisions and the spatial and temporal crossing patterns of the

wildlife involved. In fact, Barnum ( 2003) reported that these data were not useful in

identifying crossing zones, largely due to inaccurate reporting.

Underpasses, overpasses and other structures designed to promote safe passage of large

animals across highways are being built more frequently throughout North America

( Clevenger and Waltho 2000). Whereas early passage structures were typically designed

to mitigate the impact on a single- species ( Reed et al. 1975), the focus today is more on

preserving ecosystem integrity and landscape continuity to benefit multiple species

( Clevenger and Waltho 2000). Transportation agencies increasingly are receptive to

integrating passage structures into highways to address both safety and ecological needs

( Farrell et al. 2002). However, they increasingly expect that such structures will benefit

multiple species and enhance access to habitat ( Clevenger and Waltho 2000); and that

scientifically sound monitoring and evaluation of wildlife response will be done to

improve future effectiveness ( Clevenger and Waltho 2003, Hardy et al. 2003).

Just as varied approaches have been used to assess wildlife passage, a multitude of

methods measures have been used to measure wildlife use of passage structures. Most

studies’ data have come from underpass track counts ( Clevenger and Waltho 2000,

Gloyne and Clevenger 2001), event recorders ( Foster and Humphrey 1995), or single-frame

camera images ( Ng et al. 2004). Using frequency- of- use data to compare passage

structure use has a potential bias due to heterogeneous animal distribution or the

differential funneling caused by varying amounts of wildlife- proof fencing; and fails to

account for animals not using passage structures or that are resistant to crossing. To

address such biases, Clevenger et al. ( 2001b) estimated expected passage frequencies

derived from track assessments of relative abundance, and Clevenger and Waltho ( 2003)

calculated species performance ratios from radio telemetry, pellet transects, and habitat

suitability indices. Reed et al. ( 1975) compared animal evidences at the entrance and

exits of an underpass to calculate activity indices, while Gordon and Anderson ( 2003)

used behavioral quantification as a measure of wildlife response.

2.2 EXPERIMENTAL APPROACH

The reconstruction of State Route ( SR) 260 is one of the most comprehensive projects of

its type in North America: eleven large wildlife underpasses1 and six bridges ( 1 passage

1 Each underpass but one actually consists of two structures with an atrium between them. They will be

referred to as single structures herein.

13

structure/ mi) are being built to allow wildlife passage and improve highway safety, This

project rivals the landmark efforts to improve wildlife passage and reduce losses from

collisions with vehicles in Banff National Park, Alberta, Canada, which has 24 passage

structures in 28 mi ( 0.86 structures/ mi; Clevenger and Waltho 2003), as well as those

planned for the U. S. Highway 93 reconstruction in Montana, which has 42 passage

structures in 56 mi ( 0.75 structures/ mi; Western Transportation Institute 2005).

2.2.1 Phased Construction and Adaptive Management

In addition to its scope in addressing conflicts with wildlife, the SR 260 upgrade is

noteworthy for two other reasons: it is being done in phases and information gained in

early phases is being used to improve work in later ones. The section of SR 260 to be

upgraded was divided into five parts; each part is being reconstructed according to a

priority set by ADOT. These parts are identified by a settlement or prominent feature:

Preacher Canyon, Christopher Creek, Kohl’s Ranch, Little Green Valley, and Doubtful

Canyon. The phasing of reconstruction has facilitated effective construction oversight by

ADOT and allowed the sections with higher priority to be done first under limited

funding. The incidence of wildlife- vehicle collisions was a key factor used in setting

priority for upgrade ( Route 260- Payson to Heber EIS, ADOT Environmental Planning

Section, Phoenix, AZ).

Doing the reconstruction in phases has also facilitated sharing of our preliminary findings

with ADOT project managers for their use in addressing wildlife- related issues.

Preliminary insights from studies done in the early phases have been used to improve

wildlife passage structures, identify appropriate stretches for ungulate- proof fencing to

maximize underpass effectiveness and minimize wildlife- vehicle collisions, and select

appropriate sites for other measures ( e. g., wildlife escape jumps and gates) in sections

whose work is still underway or is yet in the planning stage Though such an adaptive

management approach can yield continuous improvement to the quality of highway

construction, especially relating to highway safety, it does come at a potential cost when

construction delays and increased project budget expenditures occur.

2.2.2 Experimental Approach

The reconstruction of SR 260 in phases afforded us the opportunity to assess the impact

on wildlife of highway reconstruction at various stages. Hardy et al. ( 2003) and

Roedenbeck et al. ( 2007) stressed the value of conducting “ before- after, control- impact”

( BACI; Underwood 1994) assessments to determine the effects on wildlife of highway

construction and the efficacy of measures to reduce wildlife- vehicle collisions and

promote passage. The phased reconstruction of SR 260 and the presence of experimental

controls gave us the opportunity to conduct such an assessment. During our research, we

assessed wildlife relationships and response to one section that was reconstructed prior to

the initiation of research, two where construction was initiated during our project,

yielding before-, during- and after- reconstruction data, and two sections that served as

research controls ( Table 2.1).

14

Our research focused on evaluating the effectiveness of measures designed to minimize

wildlife- vehicle collisions, especially those involving elk, and to maintain wildlife

passage across the highway. The first phase was initiated under Joint Project Agreement

01- 152, which was executed with ADOT in January 2002. . It focused on the Preacher

Canyon section, which was the first to be reconstructed. Work on this section was

completed in 2001 ( Table 2.1). Research under this phase served as a “ pilot study” for

the development and evaluation of various techniques for gathering data to use in

assessing the effectiveness of the various measures to minimize wildlife- vehicle

collisions and facilitate wildlife passage across the highway corridor. Phase II of our

project continued through July 2006 under Joint Project Agreement 04- 024T, which was

executed with ADOT in December 2003. This phase focused on the Christopher Creek

section, which in late 2004 became the second section completed ( Table 2.1). We also

continued monitoring the Preacher Canyon section in this phase. In November 2005,

Joint Project Agreement 06- 004T was finalized with ADOT, which authorized Phase III

of our research. This phase focused on the Kohl’s Ranch section, which was completed in

early 2006 ( Table 2.1); research under this phase will continue through June 2008.

Table 2.1. Dates that highway reconstruction was initiated and completed for the five

reconstruction sections on State Route 260, Arizona, and years of research accomplished

under various construction classes as part of research conducted 2002– 2006.

Construction upgrade Years of study by construction class

Highway section Begun Completed Before During After

Preacher Canyon

1999

2001

0

0

5

Christopher Creek 2002 2004 1 2 2

Kohl’s Ranch 2003 2006 2 2 1

Little Green Valley Control 5 0 0

Doubtful Canyon Control 5 0 0

2.3 RESEARCH OBJECTIVES

Our research addressed six objectives:

Objective 1. Assess and compare wildlife use of wildlife underpasses constructed

along State Route 260, and evaluate the efficacy of video surveillance as a means of

assessing wildlife use of underpasses.

Researchers have used various methods to gather data to assess wildlife response to

passage structures ( Hardy et al. 2003), including track counts ( Rodríguez et al. 1996;

Clevenger et al 2001b; Clevenger and Waltho 2000, 2003), event recorders ( Reed et al.

15

1975, Foster and Humphrey 1995), and infrared motion or heat sensor single- frame

cameras ( Servheen et al. 2003, Brudin 2003, Ng et al. 2004). Video cameras have had

only limited use in the past for assessing passage structure use ( Reed et al. 1975, Gordon

and Anderson 2003, Plumb et al. 2003). Video surveillance has an advantage over other

techniques because animal behavior can be assessed, especially when the animal resists

or fails to cross ( Hardy et al. 2003). Video surveillance also allows identification and

classification ( e. g., sex, age) of individual animals, which track counts do not ( Hardy et

al. 2003). Although video camera surveillance has been minimally used to assess use of

passage structures, such monitoring has nonetheless provided insights that were not

obtained from other methods ( Reed et al. 1975, Gordon and Anderson 2003).

To meet this objective, we evaluated the use of video surveillance to assess and compare

wildlife response to underpasses constructed during the reconstruction of SR 260 in the

first phase of our research. Focusing on the first two completed wildlife underpasses ( in

the Preacher Canyon section), we tested the hypothesis that wildlife frequency of use,

passage rates, and behavioral response did not differ at these underpasses and we

evaluated the efficacy of using passage rate and behavioral response measures to compare

wildlife use of passage structures. We explored seasonal wildlife use and response at the

two underpasses, related differences in response to underpass characteristics where

possible, and considered relationships with highway traffic volume.

Under Phase II, we expanded monitoring and assessment to the underpasses constructed

on the Christopher Creek section with three additional video surveillance systems.

Information from these underpasses is still preliminary, especially compared to the

relatively long term monitoring conducted at the two underpasses in the Preacher Canyon

section as part of Phase I. Our sixth video camera surveillance system was installed at an

underpass on the Kohl’s Ranch section as part of Phase III of our project; this data is also

preliminary.

Our findings for this objective are reported in Sections 4 and 11.

Objective 2. Evaluate wildlife movements across SR 260 before, during, after

reconstruction using Global Positioning System ( GPS) telemetry.

The use of GPS telemetry in wildlife movement studies has become increasingly popular,

cost- effective, and reliable ( Rodgers et al. 1996). With continuous automated tracking at

set time intervals, reduced observer bias ( compared to VHF telemetry), and potential to

collect large datasets, GPS telemetry has revolutionized wildlife movement studies. GPS

telemetry is increasingly used to address previously- difficult questions ( e. g., Anderson

and Lindzey 2003), and holds tremendous potential to facilitate highway passage

assessment and determine spatial and temporal highway crossing patterns of wildlife.

Under this objective, we used GPS telemetry to investigate elk passage across SR 260,

comparing their approach, crossing, and passage rates by sections before, after and at

intermediate stages of construction. We evaluated quantitative measures of elk highway

passage using GPS telemetry; and assessed spatial and temporal influences on elk

16

movements. We conducted separate GPS telemetry assessments under Phases I and II of

our research, with a third assessment ongoing in conjunction with research Phase III.

Our findings for this objective are reported in Sections 6 and 7.

Objective 3. Characterize the temporal and spatial patterns of wildlife- vehicle

collisions and changes associated with highway reconstruction ( before, during, after

reconstruction), and compare wildlife- vehicle collisions to GPS- determined crossing

patterns.

Under this objective, we characterized the nature of elk- vehicle collision patterns along

SR 260, and compared collision incidence associated with the highway before, during,

and after reconstruction. We sought to validate the priority for reconstruction set for the

highway sections based on wildlife- vehicle collisions. We compared spatial and

temporal patterns of elk- vehicle collisions to elk- highway crossings determined by GPS

telemetry as a means to validate the management utility of elk- vehicle collision data in

developing strategies to reduce collisions and promote passage. Overall, this objective

focused on evaluating the ultimate effectiveness in reducing elk- vehicle collisions of the

full complement of measures ( e. g., wildlife underpasses and ungulate- proof fencing)

implemented along SR 260, as well as the benefit/ cost relationships of such measures.

We reported the results of our research for this objective in Section 9.

Objective 4. Evaluate the relationships among highway traffic volume and wildlife-vehicle

collisions, elk crossing patterns, and wildlife use of underpasses.

Although researchers disagree about whether increasing traffic volume is the primary

reason for increasing ungulate- vehicle collisions ( McCaffery 1973; Reilly and Green

1974; Allen and McCullough 1976; Case 1978; Romin and Bissonette 1996), many

recognize that traffic volume is an important factor, among others such as wildlife

population fluctuations, wildlife behavior, driver behavior, and temporal and spatial

environmental factors ( Carbaugh et al. 1975, Bashore et al. 1985, Groot Bruinderink and

Hazebroek 1996, Haikonen and Summala 2001, Seiler 2004, Gunson and Clevenger

2003, Manzo 2006).

Traffic may serve as a “ moving fence” that can render highways impassable to wildlife

( Bellis and Graves 1978). One theoretical model ( Iuell et al. 2003) predicted that

highways become impassable barriers to most wildlife at 10,000 vehicles/ day, potentially

leading to fragmentation and rapid genetic isolation like that documented for bighorn

sheep ( Epps et al. 2005). Alternatively, because traffic varies seasonally, weekly and by

time of day, some animals may be able to cross highways with high traffic volume during

periods when traffic volume is relatively low.

Average annual daily traffic ( AADT) along SR 260 is high and is increasing due to the

tourist, recreational, and commercial traffic that travels this highway. SR 260 links

Phoenix to White Mountain communities ( e. g., Show Low, Pinetop- Lakeside,

Springerville- Eagar) and high- mountain recreation areas ( e. g., White Mountain Apache

17

Reservation, Apache- Sitgreaves National Forest), as well as Interstate 40. We used GPS

telemetry data to assess relationships of AADT to elk distribution and highway approach

and crossing patterns so as to assess the impact of traffic volume at highway grade. At

wildlife underpasses, we assessed the relationships of traffic volume on wildlife crossing

below grade.

Our findings from traffic volume- related research are reported in Sections 5 and 8.

Objective 5. Assess the role that ungulate- proof fencing plays in the incidence of

wildlife- vehicle collisions, wildlife use of underpasses, and wildlife permeability

across the highway.

Research has shown that ungulate- proof fencing effectively reduces wildlife- vehicle

collisions, especially when used in conjunction with passage structures ( Ward 1982,

Lavsund and Sandegren 1991, Romin and Bissonette 1996, Clevenger et al. 2001,

Forman et al. 2003). Though fencing is generally regarded as effective in reducing

collisions with wildlife, mixed results have been reported ( Falk et al. 1978), especially

where animals cross at the ends of fencing resulting in zones of increased collisions

( Feldhamer et al. 1986, Woods 1990, Clevenger et al. 2001). Furthermore, fencing is

costly and requires substantial maintenance ( Forman et al. 2003), potentially contributing

to transportation managers’ reluctance to fence extensive stretches of highways. While

fencing is often regarded as an integral component of effective passage structures ( Romin

and Bissonette 1996, Forman et al. 2003), limited information or guidelines exist for the

use of fencing with wildlife passage structures. As fences themselves constitute effective

barriers to ungulate passage across highways ( Falk et al. 1978), fencing may exacerbate

the barrier effect associated with highways alone ( see Section 6), particularly where

effective measures to accommodate animal passage are lacking.

During the reconstruction of SR 260, ADOT’s practice of integrating 8- ft ungulate- proof

fencing with underpasses and bridges has been to erect limited wing fences ( fewer than

300 ft) outward from bridge abutments to funnel animals toward the structures. As part

of our research, we addressed the efficacy of this approach to fencing and used the

adaptive management approach during reconstruction to recommend and evaluate the

strategic placement of fencing to intercept crossing wildlife, reduce wildlife- vehicle

collisions, promote effective use of underpasses by wildlife, and maintain highway

permeability.

The results of our research related to the role of ungulate- proof fencing are found in

Section 7.

Objective 6. Provide ongoing, highway construction and maintenance guidance

throughout all construction phases.

As our research was part of an ongoing adaptive management approach to the highway

reconstruction project; we provided guidelines for maintaining wildlife permeability,

minimizing wildlife- vehicle collisions, improving wildlife underpass design to maximize

18

the likelihood of high acceptance and use by wildlife, and the strategic placement of

ungulate- proof fencing.

We report the results and applications of the adaptive management part of our project in

Sections 4, 6, 7, and 11.

2.4 REPORT ORGANIZATION

This report has our findings from Phases I and II, and to a limited degree Phase III. First,

we describe the study area to set the context for our research. The research conducted to

meet our objectives is reported in the following sections. In the Conclusion and

Recommendations section, we tie together the information from the previous sections.

19

3.0 STUDY AREA

We conducted this study along a 17- mi stretch of SR 260, beginning 9 mi east of Payson, and

extending to the base of the Mogollon Rim in central Arizona ( lat 34° 15’– 34° 18’ N, long

110° 15’– 111° 13’ W; Figure 3.1). The existing two- lane highway is being upgraded to a four-lane

divided highway ( Figure 3.1). In places, the footprint of the upgraded highway exceeds

0.3 mi in width ( Figure 3.2). When completed, the highway will have 11 wildlife

underpasses specifically intended to reduce at- grade elk crossings and elk- vehicle collisions,

as well as six bridges over large canyons and streams that will accommodate wildlife use

( Figure 3.1, Table 3.2). All but one of the underpasses consists of two structures, one for

each roadway, with an atrium between them. The highway reconstruction is being done in

five phases, each phase focusing on a single section ( Figure 3.1, Table 3.2). Reconstruction

of three sections is now complete.

The Preacher Canyon section was the first completed; all lanes were opened to traffic in

November 2001. This section has two bridged underpasses and a large bridge over Preacher

Canyon ( Figure 3.1); 0.4 mi ( 13%) of the section was fenced with 8- ft ungulate- proof fencing

associated with the two underpasses near Little Green Valley. The Christopher Creek section

was completed in December 2004; it has had four wildlife underpasses and three bridges in

place since 2003. All lanes in the Christopher Creek section were opened to traffic in July

2004 before all fencing associated with the underpasses was completed. Here, fencing and

alternatives to fencing ( e. g., swaths of large rock rip- rap) were implemented along half the

section in association with passage structures. The Kohl’s Ranch section, the most recently

reconstructed, was completed in March 2006; this section has one wildlife underpass and 1.5

bridges ( only one bridge span was built over Thompson Draw, with the other to be done

under the Little Green Valley section). Reconstruction of the Little Green Valley and

Doubtful Canyon sections will be in or after 2008.

Table 3.1. State Route 260 reconstruction sections, reconstruction status, mileposts and

length, and the number of wildlife passage structures planned or built as part of the

reconstruction.

Reconstruction Highway Length Wildlife passages

Highway section status mileposts ( mi) Underpasses Bridges

Preacher Canyon Completed 2001 260.0– 263.0 3.0 2 1

Little Green Valley Control 263.1– 265.5 2.5 1 0.5

Kohl’s Ranch Completed 2006 265.6– 269.5 4.0 1 1.5

Doubtful Canyon Control 269.6– 272.5 3.0 3 0

Christopher Creek Completed 2004 272.6- 277.0 4.5 4 3

All 260.0- 277.0 17.0 11 6

20

Mogollon

Rim

Preacher

Canyon

Kohl’s

Ranch

Little

Green

Valley

Doubtful

Canyon

Christopher

Creek

Figure 3.1. Location of our study area and the five highway sections where phased

highway reconstruction has been ongoing since 2000, and the location of wildlife

underpasses and bridges. The shaded areas correspond to riparian- meadow habitats

located adjacent to the highway. Topographic relief reveals the study area’s proximity to

the Mogollon Rim escarpment, the dominant physiographic feature within the study area.

21

Figure 3.2. The existing narrow two- lane roadway ( left; Doubtful Canyon section) is

being reconstructed to a four- lane divided highway ( right; Preacher Canyon section),

State Route 260, Arizona.

Our study area lies within the ponderosa pine ( Pinus ponderosa) association of the montane

coniferous forest community ( Brown 1994a). Elevations along SR 260 range from 5,220–

6,560 feet ( ft.) The Mogollon Rim escarpment to the north is the dominant landform, rising

precipitously to 7,860 ft ( Figures 3.1 and 3.3). Vegetation adjacent to the highway grades

from mixed forest of ponderosa, pinyon ( P. edulis), juniper ( Juniperus spp.), and live oak

( Quercus spp.) on the lower elevation Preacher Canyon and Little Green Valley sections, to

forests predominated by ponderosa with interspersed Gambel oak ( Q. gambelii) at higher

elevations to the east ( Christopher Creek section). Chaparral ( e. g., manzanita;

Arctostaphalos pungens) with sparse pinyon, live oak, and ponderosa pine is prevalent on the

drier south- facing slopes. In canyons emanating from the Mogollon Rim within our study

area, mixed- conifer forest of ponderosa pine, Douglas fir ( Pseudotsuga menzeisii), white fir

( Abies concolor) and Gambel oak are found. Numerous riparian and wet meadow habitats

occur at several locations along the highway corridor ( Figure 3.1); some meadows are more

than 60 acres ( Figure 3.4). Several perennial streams flow adjacent to portions of the

highway, including Little Green Valley Creek ( Preacher Canyon, Little Green Valley

sections), Tonto Creek ( Kohl’s Ranch section), Christopher Creek ( Doubtful Canyon,

Christopher Creek sections), Hunter Creek ( Christopher Creek section), and Sharp Creek

( Christopher Creek section) ( Figure 3.1).

Climatic conditions within the study area are mild, with a mean maximum monthly

temperature ( July) for Payson of 90.3 º F, and mean minimum monthly temperature ( January)

of 19.6 º F. Annual precipitation averages 20.7 inches ( in.), with a mean of 21.3 in. of

snowfall in winter; precipitation has averaged ⅔ of normal since 2002.

Average annual daily traffic ( AADT) on this portion of SR 260 ( ADOT Control Road traffic

monitoring station) doubled in 10 years from 3,100 in 1994 to nearly 6,300 in 2002, and

increased to 8,700 (+ 38%) in 2003 ( Figure 3.5; ADOT Data Management Section). Over the

same period, annual wildlife- vehicle collisions involving ungulates and large carnivores on

this stretch of SR 260 increased from 28 to 44, with a mean of 35.9 (± 2.5 SE; Dodd et al.

2006).

22

Figure 3.3. State Route 260 study area ( at the pedestrian/ wildlife underpass on the

Christopher Creek Section), Arizona. The Mogollon Rim escarpment rises in the

distance above ponderosa pine forest adjacent to the highway corridor. The solar panels

power our video camera surveillance system to monitor wildlife use of the underpass.

Figure 3.4. Aerial view of Little Green Valley riparian- meadow complex adjacent to the

Preacher Canyon section. Such habitats are very important to wildlife for food and water,

especially in proximity to forest cover.

23

Rocky Mountain elk ( Cervus elaphus nelsoni) were a focus of our research for several

reasons. First, elk accounted for more than 80% of all collisions between vehicles and

wildlife ( Dodd et al. 2006) and the vast majority of property loss and human injuries

associated with these collisions. Elk are large animals that can readily support our GPS

telemetry collars, yielding substantial long- term data on movements in relation to the

highway corridor, and were relatively easy to trap.

Our study area has both resident and migratory elk herds. Resident elk were common,

especially near meadow and riparian habitats. Migratory elk come off the Mogollon Rim

with the first snowfall of more than 12 in., typically in late October ( Brown 1990,

1994b). Brown ( 1990) reported that 85% of the elk residing within his Mogollon Rim

herd unit migrated to an area below but within six mi of the base of the Mogollon Rim,

which encompasses our study area. Elk return to summer range with forage green- up at

higher elevations ( Brown 1990). The Arizona Game and Fish Department estimated the

resident elk population in the game management units encompassing our study area at

1,500- 1,600 ( Arizona Game and Fish Department, Game Management Branch,

unpublished data), though not all elk resided in proximity to SR 260. White- tailed deer

( Odocoileus virginianus cousei) were frequently seen in our study area, while mule deer

( O. hemionus) were less common and more localized on the Christopher Creek section.

Figure 3.5. Average annual daily traffic for State Route 260, Arizona. ( ADOT Control

Road monitoring station) for the period 1990– 2005.

Year

Average annual daily traffic volume

2000

3000

4000

5000

6000

7000

8000

9000

1990 1992 1994 1996 1998 2000 2002 2004

24

25

4.0 VIDEO SURVEILLANCE TO ASSESS HIGHWAY

UNDERPASS USE BY ELK 2

4.1 INTRODUCTION

Recognition of highways’ impact on wildlife has increased dramatically in the past decade

( Forman et al. 2003). In addition to direct habitat loss ( Forman 2000), mortality from vehicle

collisions has been recognized as a serious and growing problem for wildlife and motorists

( Reed et al. 1982, Farrell et al. 2002). Annual vehicle collisions with deer alone in the U. S.

exceed 1.5 million ( Conover 1997). Highways play a pervasive role as barriers to free

movement of wildlife, fragmenting and isolating habitats, reducing genetic interchange ( Epps

et al. 2005), and increasing population susceptibility to catastrophic events ( Forman and

Alexander 1998, Trombulak and Frissell 2003).

Structures designed to promote wildlife passage across highways are increasingly being built,

particularly large bridges designed specifically for large animal passage ( Clevenger and

Waltho 2000). Whereas in the past managers typically built early passage structures as

single- species mitigations ( Reed et al. 1975), managers today have directed their focus

toward preserving habitat continuity to benefit multiple species ( Clevenger and Waltho

2000). Transportation agencies are increasingly receptive to building passage structures to

meet safety and ecological needs ( Farrell et al. 2002), and there is increasing expectation that

they will indeed yield desired benefits ( Clevenger and Waltho 2000). Scientifically sound

monitoring of wildlife response to passage structures is crucial to improving future

effectiveness ( Clevenger and Waltho 2003, Hardy et al. 2003).

Researchers have used various techniques to measure wildlife use of passage structures

( Hardy et al. 2003), including track counts ( Rodríguez et al. 1997; Clevenger et al 2001a;

Clevenger and Waltho 2000, 2003), triggered event recorders or counters ( Reed et al. 1975,

Foster and Humphrey 1995), and infrared motion or heat sensor single- frame cameras

( Brudin 2003, Servheen et al. 2003, Ng et al. 2004). Video cameras have had only limited

use ( Reed et al. 1975, Sips et al. 2002, Gordon and Anderson 2003, Plumb et al. 2003).

Several measures have been used to describe wildlife use of passage structures. Most studies

have enumerated frequency of use ( Clevenger and Waltho 2000, Gloyne and Clevenger

2001, Sips et al. 2002, Ng et al. 2004). However, frequency of use may be a biased index of

passage structure efficacy. It is subject to differential funneling of animals by varying

amounts of fencing and heterogeneous animal distribution, and does not account for non- use

attributable to structure characteristics or alternative crossing locations, as addressed by Reed

et al. ( 1975), Clevenger et al. ( 2001a), and Clevenger and Waltho ( 2003, 2005).

Video surveillance has advantages over other techniques for assessing passage because

managers can evaluate animal behavior, especially when the animals resist - or don’t

2 An early version of this chapter was published in the Journal of Wildlife Management ( see Dodd et al.

2007a)

26

complete - crossing ( Hardy et al. 2003, Gordon and Anderson 2003). Though few studies

have used video surveillance to assess use of passage structures, such monitoring has

provided insights that could not be obtained from other methods ( Reed et al. 1975, Gordon

and Anderson 2003, Plumb et al. 2003).

Our objective was to evaluate video surveillance for assessing and comparing wildlife

response to underpasses. We examined Rocky Mountain elk use of the first two underpasses

completed as part of the SR 260 reconstruction. We tested hypotheses that elk passage rate

( crossing frequency/ approach frequency), probability of use, and behavioral response did not

differ at the two underpasses. We monitored seasonal elk use of the underpasses to test the

hypothesis that passage rates and probability of use did not differ by season. We related

differences in elk use to underpass design and provided guidelines for future design to

maximize the likelihood of use by elk and other wildlife.

4.2 STUDY AREA

We conducted our study at two bridged underpasses constructed specifically for wildlife

passage along the Preacher Canyon section of SR260 ( Figure 4.1). Both opened to the south

into Little Green Valley, a relatively lush riparian- meadow foraging area contrasted by dense

forest cover on the north side of the highway ( Figure 4.1). The two underpasses were less

than 850 ft. apart ( Figure 4.1). Though both were of similar open- span bridge construction

and length ( 135 ft), the below- span characteristics and dimensions were markedly different

( Figure 4.1, Table 4.1). The east underpass had vegetated earthen sides that made it more

open and natural compared to the west underpass, which had concrete, mechanically

stabilized earth ( MSE) walls ( Figure 4.1). ADOT installed 8 ft high ungulate- proof fencing

along 0.4 mi of the highway to funnel animals to the two underpasses ( Figure 4.2).

Table 4.1. Physical characteristics associated with the two wildlife underpasses ( UP) at

which we conducted video monitoring focusing on elk from September 2002– September

2005, State Route 260, Arizona. ( see Figure 4.1).

Characteristic

East UP

West UP

Construction type Open I- beam span Open I- beam span

Bridge span distance 135 ft 135 ft

Maximum height above floor ( H) 22 ft 38 ft

Atrium ( between bridges) a 36 ft 36 ft

Width at floor ( W) 32 ft 52 ft

Lengthb ( L) 175 ft 365 ft

Side construction 2: 1 sloped earth/ vegetation MSEc concrete walls to 20 ft

Openness ratiod 12.3 5.5

a - Atrium = width of opening between eastbound and westbound bridge spans at each underpass

b - Length = distance for animals to fully negotiate underpass, including fill at mouth of underpass

c - MSE = mechanically stabilized earth

d - Openness = ( W × H) / L ( Reed et al. 1979)

27

Figure 4.1. Little Green Valley riparian- meadow complex ( center photo) adjacent to State

Route 260 in Arizona, into which the west ( top photo) and east ( bottom photo) wildlife

underpasses open. Note their proximity and the different soil stabilization features, the west

with concrete walls, and the east with 2: 1 sloped earthen sides.

28

4.3 METHODS

4.3.1 Video Surveillance System Components and Layout

At each underpass we installed video surveillance systems comprised of four low- lux, high-resolution

black and white ( B& W) video cameras linked to a 12- volt videocassette recorder

( VCR) with alarm input and a B& W quad- screen splitter. To illuminate the area covered by

our cameras, we installed infrared ( IR) 60 LED illuminators ( 9 at the east underpass, 7 at the

west). We used five IR photo- beam triggers at each underpass to detect approaching and

crossing animals. We operated both systems on 120- volt AC power converted to 12- volt DC

power for distribution to all equipment via buried wiring. We operated the camera systems at

the east underpass from September 2002– September 2005; the camera system at the west

underpass we operated from November 2002– September 2005.

At each we oriented our video systems to record animals approaching from the north side

only ( Figure 4.2), recording the elk as they traveled from forest cover into Little Green

Valley. We believe that elk that approached from the north had a greater degree of discretion

in use of the underpasses or alternate crossing locations compared to elk already in Little

Green Valley that had to return to cover via an underpass. Nonetheless, we recorded animals

crossing from both the north and south. We installed two cameras approximately 100– 115 ft

from each underpass ( Figure 4.2) to record animals approaching to within approximately 200

ft. of the underpass along drainages leading to it. We mounted a camera atop a 15- ft pole in

each underpass to record animals entering and crossing. We oriented a camera toward the

highway to record traffic, while other cameras simultaneously recorded approaching or

crossing wildlife ( Figure 4.2); we reported results of this monitoring in Gagnon ( 2006).

We placed IR photo- beam triggers approximately 1.5 ft above ground oriented such that

animals could not approach the underpass without tripping a trigger. To avoid recording

delays, we operated all components continuously so that VCRs immediately began recording

when triggered, with all cameras recording simultaneously. We programmed our VCR alarm

to record for two minutes each time an animal successively tripped a trigger. Twelve- volt

DC blowers and heaters ensured continuous operation during heat and cold.

4.3.2 Video Data Analysis

We extracted the following data from the video tape: date, time of day, total time animals

spent in the area, species, sex, age, number of animals, number of animals approaching and

crossing through the underpass, direction of travel, and various behaviors.

We calculated monthly elk passage rates as the proportion of elk groups that passed through

the underpass from the north relative to the frequency of groups that approached from the

same direction. We counted as an approach when animals crossed over the 3.5 ft right- of-way

( ROW) fence approximately 130– 160 ft from the mouth of the underpass ( Figure 4.2).

ADOT did not remove this fence during underpass construction due to presence of livestock;

instead, at each underpass it threaded the top two stands of wire through 20- ft lengths of PVC

pipe to create elk jumps. We counted it as a group crossing when half or more the elk in a

group passed through an underpass.

29

Figure 4.2. Layout ( top) of video surveillance system components at the west and east

Little Green Valley wildlife underpasses and the location of elk- proof and highway right- of-way

fencing, State Route 260, Arizona. We oriented video cameras to record wildlife

approaching each underpass from the north ( 2 cameras), animals crossing through the

underpass from both the north and south ( 1 camera), and simultaneous traffic on the

highway while animals approach and pass through the underpass ( 1 camera). The bottom

photo shows a group of elk passing through the west underpass, though the lead cow is

showing resistance. Note the illumination provided by infrared lights to observe animals at

nighttime.

Cameras

Photo- beam triggers

Ungulate- proof fence

Right- of- way fence

N

30

We classified behavioral responses of individual elk into five approach and three crossing

categories to quantify acceptance or resistance to using the underpasses, similar to

Gordon and Anderson ( 2003). For approaching elk, we assigned frequencies to 1 or more

of these categories:

• Would not cross – elk left without crossing an underpass.

• Enter underpass and retreat – elk entered an underpass but retreated outside it.

• Alarmed flight – elk that approached or entered an underpass, but rapidly departed

in an alarmed manner.

• Feeding in area – elk that fed in the area between the ROW fence and the center

of an underpass.

• Standing or milling about – elk that stood or milled about in the area between the

ROW fence and center of an underpass.

For elk that crossed through either underpass from the north, we classified the degree to

which they exhibited hesitation or paused in an alert posture ( excluding feeding

behavior). We quantified delay by the time it took the elk to move from the mouth to

beyond the center of the underpass:

• No delay – elk crossed with less than 10 seconds combined hesitation.

• Minor delay – elk crossed with 11- 30 seconds combined hesitation.

• Obvious delay – elk crossed after a combined delay more than 30 seconds. We

classified elk that retreated or fled in alarm from the underpass before finally

crossing as exhibiting obvious delay.

4.3.3 Time- Lapse Validation

We conducted 24- hour time- lapse taping on five occasions at both underpasses to

compare the number of elk groups and individual animals recorded by VCRs in time-lapse

mode to the number recorded when the VCR alarm log reflected that animals had

activated a photo- beam trigger. We relied on our VCR internal alarm counters while

viewing the time- lapse video recordings to determine what proportion of approaching and

crossing animals our photo- beam triggers detected.

4.3.4 Statistical Analysis

Clevenger and Waltho ( 2000) stressed the benefit of multi- species assessments of

passage structure use, and Little et al. ( 2002) raised issues regarding predator- prey

interactions at passage structures. Though we pursued a multi- species assessment ( see

Gagnon et al. 2006), our observations for most species were relatively small compared to

31

those for elk, which accounted for more than 90% of the animals recorded on our

videotapes. This limited our ability to make statistical inferences regarding underpass

use for species other than elk.

We used elk group observations to assess underpass use to address poten