Analysis of Megafauna Detections

Item

Title

Eng Analysis of Megafauna Detections

Date

2012

Creator

Eng Kangiser, David

Subject

Eng Environmental Studies

extracted text

Analysis of Megafauna Detections

David Kangiser

2012

© 2012 by Dave Kangiser. All rights reserved.

This Thesis for the Master of Environmental Studies Degree
by
Dave Kangiser
has been approved for
The Evergreen State College
By

________________________
Ralph Murphy, Ph.D.
Member of the Faculty

________________________
Date

4

Abstract
Animal crossing structures are increasingly being recognized as a way for
transportation agencies to reduce dangerous encounters that motorists have with
wildlife. At the Washington State Department of Transportation, this was
recognized and implemented as early as the mid 1970’s, but little monitoring
followed the installation of these structures to better understand how wildlife
interacts with the highway system if a safe crossing opportunity was present. It is
a long held theory that wildlife is most active at crepuscular periods, but does that
theory hold true for animals that use these crossing structures to access the
other side of the highway? For this thesis I monitored 6 wildlife crossing
locations on Washington State highways using motion triggered trail cameras to
better understand species composition and temporal patterns of animal
crossings. Camera images were converted to detections data that included
species, time, and ambient temperature when the animal was detected.
Detection times were included in a data matrix that related each detection to
sunrise and sunset. Chi Squared tests were used to analyze whether peak
activity was concentrated in the hour before and after sunrise and sunset.
Temperature at time of detection was analyzed to determine if animals are more
active during particular temperature ranges. Traffic volume data for the subject
stretch of roadway was also analyzed to determine if elevated volumes are a
predictor of frequency and use of crossing structures. I expect this information to
be used by the Washington State Department of Transportation to produce a
better understanding of the use of crossing structures by wildlife.

5

Acknowledgements
I would like to thank Kelly McAllister, WSDOT biologist, for his guidance
and expertise in habitat connectivity and for giving me the opportunity to be a
part of this research; Mark Bakeman, WSDOT biologist, for his help and wisdom
in statistics; Marion Carey, WSDOT manager, for supporting me as an intern in
the WSDOT Environmental Services Office; Martha Henderson, The Evergreen
State College Director of the Graduate Program on the Environment, for
sponsoring my internship; Ralph Murphy for our long bouts with metrics and Chi
Squared tests that often turned into stories of boats, sports, and our kids. I
needed that more than you know.
I would especially like to thank my wife, Stephanie, who often thought I
was speaking a different language when I tried to explain what I was doing on my
computer at 11:00 at night; and my son, Landon, for not having his dad around a
whole lot while I worked for this master’s degree. You can both have me back
now.

6

Table of Contents
Abstract….………….…………..………………………………………………….……4
Acknowledgements……………………………………………………………………5
1. Introduction.....................................................................................................7
2. Background…………………………………………………………………...…..8
2a. Road Ecology……………………..………………………….…………8
2b. Banff National Park, Canada…………….……………………….....14
2c. WSDOT Study / Internship…………………….…………………….21
2d. Site Locations………………………………….……………………...23
3. The I 90 Crossing Structures – The Whole Story…………….……………...27
4. Research Question and Hypothesis…………………………………………..32
5. Methods…………………………………………………………………………..32
6. Results……………………………………………………………………………36
7. Chi Squared Analysis…………………………………………………………...47
8. Discussion…………………………………………………...…………………..50
9. Thoughts for the Future….….……………………………...……………….….54
10. Conclusion…………………………………………………………………….….55
Appendix 1: Site Location Map.………………………….....….…57
Appendix 2: Willapa River
Species composition and frequency……………………..58
Aerial photo………………………………………………....58
Photo examples………………………………………….…59
Appendix 3: Mosquito Creek
Species composition and frequency……………………..60
Aerial photo………………………………………….……...60
Photo examples…………………………………….………61
Appendix 4: North Bend I-90 West
Species composition and frequency……………………..62
Aerial photo……………………………….………………...62
Photo examples………………………….…………………63
Appendix 5: North Bend I-90 East
Species composition and frequency……………….…….64
Aerial photo…………………………………………….…...64
Photo examples……………………………………….……65
Appendix 6: Tucker Creek
Species composition and frequency………………….….66
Aerial photo………………………………………………....66
Photo examples………………………………………….…67
Appendix 7: Deadman Creek
Species composition and frequency……………………..68

7

Aerial photo…………………………………………….…...68
Photo examples…………………………………….………69
Bibliography……………..……………………………………………………………70

1. Introduction
Human expansion is occurring at levels never experienced in history. An
expanding transportation system is inevitable with this current pace of human
expansion. With this expanding infrastructure comes a greater risk of animalvehicle collisions that further exposes flora and fauna to the risks of habitat
fragmentation and the negative consequences of island biogeography.
Transportation planning should occur with awareness of roads in the context of
surrounding habitats, ecosystems, and ecosystem processes as well as the
infrastructure network and land use. As a result of these trends, the
consideration of wildlife and their habitats in road planning and construction is
becoming increasingly common. The term for this new paradigm is road ecology.
Road Ecology by definition suggests an interdisciplinary approach to the
problems facing transportation agencies, wildlife, and other stakeholders. It is
defined as a sub-discipline of ecology that focuses on understanding interactions
between road systems and the natural environment including hydrology, wildlife
biology, plant ecology, population ecology, soil science, water chemistry, aquatic
biology, and fisheries (Lloyd, 2011 and Forman et al, 2003). Although road
ecology is still in its infancy, a piecemeal approach to this paradigm would be
useless. Therefore, theories and concepts that encompass road ecology only
work from an interdisciplinary approach.

8

Involving stakeholders and the public at all levels of the planning and
implementation of a project is essential to success. Non-governmental
Organizations (NGOs) and the public have taken advantage of their ability to
lobby government officials to persuade them to adopt measures that protect the
driving public. These measures also provide safe crossing opportunities for
animals, allowing natural gene flow, and continued use of historic ranges.
2. Background
2a. History of Road Ecology
Although the paradigm of road ecology has only recently been nationally
recognized as a sub-discipline of ecology, interactions between wildlife and
roadways date back to before European settlers arrived. Many of the main
highways in use today were once historic trails established by the Native
Americans. Take I-90 for example; the I-90 corridor through the Cascade
Mountains includes the lowest elevation pass to eastern Washington. Puget
Sound tribes, such as the Nisqually, would journey to the mountains in search of
wild game, berries, materials for clothes and baskets, and for religious purposes.
Eastern Washington tribes used that same corridor from the east to access the
mountains for their needs as well. These native peoples also travelled across
the mountains to trade and visit relatives, since inter-marrying between groups
was not uncommon. Many of the local wildlife species recognize these welltraveled routes and use them for migration from summer to winter forage areas in
the fall and reverse the migration in the spring, when the snow melts in the
highlands.

9

I-90 has evolved over the last century. Starting as a hunting and
gathering trail, it became a wagon trail accessible only when weather allowed.
Eventually it developed into a two lane mountain road, and as transportation
evolved, the route became a six lane highway, linking the Northwestern corner of
the United States to the rest of the country. Wildlife has been present during the
whole metamorphosis of the corridor and they recognize the highway is an
obstacle that cannot be easily navigated around.
The challenge facing transportation agencies today is the mandate to
provide a safe and efficient transportation system while following sound
environmental practices in the planning, design, construction, operation, and
maintenance of Washington State Department of Transportation's (WSDOT)
transportation systems and facilities (Executive order 1018.01, 2007).
Washington State is trying to improve the impacts that roadways, and especially
busy roadways, have on the environment. As a result, in July 2007, WSDOT
Secretary Douglas MacDonald signed Executive Order E 1031. This Order
recognizes the need for the Department of Transportation to realize the impacts
that the transportation system places on the environment and outlines goals for
biodiversity protection. Through the Environmental Services Office within
WSDOT, the goals set forth by the Executive Order are (Executive Order 1031,
2007):
1.
To identify affected fish and wildlife habitats as early
as possible during the planning process for projects and
programs and in preparation of regional and statewide longrange transportation plans. The planning should seek to

10

integrate state conservation and biodiversity plans and other
available natural resource information.
Transportation
planning should recognize and respond to particular
concerns and opportunities for habitat preservation and the
need for habitat connections. The earlier that habitat
concerns are taken up in project planning, the likelier that
good habitat approaches to state investment in habitat
protection and habitat connectivity can be incorporated into
projects.
2.
To locate specific opportunities to restore habitat
connectivity already damaged by human transportation
corridors.
Such opportunities should be prioritized for
maximum ecological benefit by taking account of such
factors as the multiplicity of benefits shared, as well as the
opportunity to support recovery of threatened and
endangered species, the long-term security and viability of
the habitat connection, and the cost effectiveness if
achieving connectivity gains. Such opportunities can be
located and achieved both as part of capital projects and in
ordinary maintenance activities.
3.
To cooperate and coordinate with other agencies
involved in wildlife habitat protection. This aim will ensure
compatibility of natural resource and habitat management in
adjacent areas so that wildlife connections provided at
roadways will link to functional and permanently protected
wildlife corridors. Ultimately, WSDOT and other agencies
should seek to develop a statewide habitat connectivity plan
to better integrate overall habitat management with
transportation planning.
4.
To support the use of site appropriate native plant
species in roadside landscaping and vegetation
management and to protect adjacent natural plant
communities.
5.
To develop and follow design criteria for
transportation structures that help promote fish and wildlife
movement and minimize habitat degradation. WSDOT
recognizes the Washington Department of Fish and Wildlife
manual, Design of Road Culverts for Fish Passage (2003),
as a primary source for information on fish passage designs.
Guidance, criteria, and manuals for structures affecting
terrestrial species will be developed.

11

6.
To protect and enhance important wildlife habitat
areas near highways on highway right of way in ways
compatible with highway operations, and to support efforts to
promote the traveling public's awareness and enjoyment of
wildlife in the state.
Even before WSDOT's Executive Order E 1031 was enacted, the
WSDOT's Environmental Services Office was coordinating a collaborative, multiorganization effort to address habitat connectivity. The Washington Wildlife
Habitat Connectivity Working Group, led by Washington Department of Fish and
Wildlife (WDFW) and Washington State Department of Transportation was
formed from voluntary public and private organizations. This group recognizes a
science based approach to transportation and community planning. Its function
is to recognize the role state agencies, tribes, and public stakeholders have in
conserving and connecting critical habitats. The Working Group's mission
statement sums up their objectives: "Promoting the long-term viability of wildlife
populations in Washington State through a science-based, collaborative
approach that identifies opportunities and priorities to conserve and restore
habitat connectivity" (WHCWG, 2010)
During most of the twentieth century, road construction focused on roads
integrating with the terrain and vegetation as a means of aesthetic appreciation.
It was not until the later part of the twentieth century that wildlife and natural
processes were considered as part of the planning process for road construction.
This came about as the increase in vehicles using the roadway system also saw
an increase in the number of wildlife vehicle collisions. Figure 1 shows that while

12

the number of car accidents per year has remained relatively constant, wildlife
vehicle collisions are trending upward.

Figure 1. All car vehicle collisions compared to animal vehicle collisions (Huijser and McGowen
2010)

Not only do wildlife related collisions cost the country over $8 billion a year
(USDOT, 2008), they may be significant enough in some places to affect
populations of some animals. Mortality from wildlife vehicle collisions can affect
population growth rates if mortality is greater than recruitment from immigration
and birth rates. If animals are afraid to cross roads, subpopulations could suffer
from a lack of gene flow or demographic dispersal.
There is evidence that some animals avoid roads at various distances.
For example, grizzly bears (Ursus arctos horriblis) have been documented to
avoid roadways by 500 meters (Waller and Servheen, 2005). Deer (Cervidae
family) and elk (Curvus Canadensis) studies indicate a 200 meter avoidance

13

buffer (Rost and Baily, 1979). Wolves (Canis lupus) are known to avoid high
density roadway areas as well (Thurber et al, 1994).
Animals are not only impacted by animal-vehicle collisions, but by habitat
fragmentation (Charry et al, 2009). Although direct habitat loss appears small
from an ecosystem perspective, occupying only 1% of the total land area in the
United States, the ecological impact is much greater affecting 15-20% of the
landscape (Charry et al, 2009). Large-scale edge effects can drive species that
inhabit central habitat core to local extirpation from habitat fragments and
protected areas. These species are among the most threatened species in
fragmented landscapes (Ewers et al, 2008).
Wildlife needs freedom of movement across the landscape and requires
the following (Beckmann and Hilti 2010):
1.) Large expanses of land for their daily, seasonal, or annual
ecological needs;
2.) Migratory movements between seasonal ranges for food and
breeding; and
3.) Connection between separate subpopulations for genetic
variation.
These criteria are especially true for animals that are rare, low density, or
wide-ranging (Beckmann and Hilti, 2010). Some studies have indicated that
habitat fragmentation is the number one source for diminished populations of
animals whose habitat is near a road system (Hilti et al, 2006).

14

Solutions to current habitat connectivity problems have been explored in
Banff, Canada. The park has developed a system of safe wildlife crossing
opportunities. They include a series of wildlife crossing structures that allow
opportunities for wildlife to cross the highway safely, 8' tall exclusionary game
fencing to funnel wildlife to the crossing structures, and a series of jump-outs that
allow trapped animals inside the exclusionary fencing an opportunity to escape
the roadway.
2b. Banff National Park, Canada
Since 1975, Parks Canada has made habitat connectivity a priority for the
large ungulates and carnivores in and around Banff National Park. The TransCanada Highway (TCH) bisects the park as it runs along the Bow Valley floors.
These same valley floors are used as migratory corridors for species with
migratory habits such as bighorn sheep (Ovis canadensis), elk, and deer
species. Other species with large home ranges also benefit from a series of
animal crossing structures that facilitate freedom of movement within their home
ranges.
Parks Canada has recognized the vital role of the TCH in providing a
corridor that connects eastern and western Canada. The highway, originally built
in the 1950's, started out as a scenic two lane mountain road. Today it sees
25,000 vehicles per day in the busy summertime tourist season. This is also
home to the most diverse assemblages of ungulate and large game species in
North America. A busy highway running straight through the park gave Canada

15

and its parks department an opportunity to create one of the most wildlife friendly
stretches of road in the world.
Roadways affect wildlife populations in several ways: increased mortality
due to wildlife-vehicle collisions, creating a barrier preventing animals from
moving freely, reducing habitat, and facilitating the spread of invasive species
(Foreman et al, 2003). The most obvious affect comes from wildlife vehicle
collisions. Wildlife vehicle collisions impact humans and animals alike; affecting
human safety, causing property damage, increasing insurance costs and
impeding wildlife conservation (Beckmann and Hilti, 2010). Different species
have different needs for survival and mating, therefore a broad approach to
connect habitats in Banff National Park was needed.
In 1979, The Canadian government recognized the increased traffic
volumes on the TCH and proposed a project that would double the capacity of
the highway by adding an additional lane to each lane of traffic, a process known
as "twinning" the highway (Ford et al, 2010). The project was to proceed in a
series of phases, beginning with Phase 1 in 1979, continuing through today with
phase 3B (Ford et al, 2010).
Phase 1 covered the first 13 kilometers of the TCH through Banff. The
Federal Environmental Assessment and Review process for this portion of
highway identified wildlife vehicle collisions as a major source for concern for
motorists and wildlife conservation issues. During 1978 alone there were over
110 elk vehicle collisions in these first 13 kilometers (Ford et al, 2010).
Therefore, the twinning project focused on mitigation measures to reduce wildlife

16

vehicle collisions, especially for larger animals such as ungulates and carnivores.
A wildlife exclusionary fence 2.4 meters (7.9 feet) high was constructed on both
sides of the highway to keep animals from entering the roadway. The fence
funneled animals to one of six underpasses that allowed for animal movement
under the highway. Figure 2 shows an example of exclusionary fencing.

Figure 2. Wildlife fencing installed to keep deer and elk off of the roadway (www.wsdot.wa.gov,
2012.)

Phase II of the twining project began soon after Phase I, covering the next
14 kilometers. It was competed in September 1987 with the exclusionary fencing
and four additional animal crossing underpasses. The two phases created a total
of ten animal crossing structures in 27 kilometers.

17

Next, the Canadian government submitted plans to widen the next 21
kilometers of the TCH, Phase III. This was a time of new mitigation measures for
the Canadian government. They introduced an ecological integrity based
management system within the national parks. This management system
recognized the need for large carnivores to be considered as part of mitigation
plans and should receive priority for animal crossing structure design (Parks
Canada, 1995). The fence system saw improvements in functionality. A one
meter section of chain link was added to the bottom of the fence and buried to
discourage animals from digging and passing under the fence (Ford et al, 2010).
Originally, underpasses were planned as part of mitigation, but the recent
recognition of large carnivores as a priority caused the transportation community
to look closer at what animals were using what structures. They found that some
of the large carnivores (grizzly bears and wolves primarily) preferred not to use
the confining underpasses. Consequently, two 50 meter wide overpasses were
constructed in Phase III along with ten additional underpasses. The wide open
spans proved preferable to the target species as well as several ungulate and
small mammal species. Phase III was completed in 1997 for a total of 23 wildlife
crossing structures over 45 kilometers of highway.
Monitoring the structures and fence for effectiveness fell on Parks
Canada. Park wardens, research biologists, and other trained staff drive the
improved stretch of highway on a daily basis, looking for signs of breaches in the
fencing as well as animal carcasses. This allows for complete and accurate data
regarding the effectiveness of the fence and crossing structure system as a

18

whole. The information collected is stored in a central database and includes the
date, coordinate and descriptive location, species, number of individuals, and
information obtained from necropsies (Ford et al, 2010).
At the completion of each phase, a pattern of reduced wildlife vehicle
collisions was emerging. The fence and crossing structure system reduced
wildlife vehicle collisions with ungulates by 90% and with large mammals in
general by 86% (Clevenger et al, 2002). The results are a promising sign that
the mitigation measures are working to reduce wildlife vehicle collisions.
Monitoring animal crossing structure use proved to be different from
monitoring the exclusionary fence. Researchers wanted to know more about
what species were using what structures and how often they were being used.
Before the advent of motion triggered cameras, a series of sand track pads were
deployed. The sand track pads capture the animals' tracks as they move in and
around the crossing structures. Monitoring the track pads became a task that
occurred every 2-3 days and accurately assessed what animals preferred what
structures. Monitoring in this fashion took place for 12 years. Species, direction
of movement, and number of individuals of megafauna were recorded. Figure 3
shows an example of track pads and the maintenance required for their use.
Recently, motion triggered cameras have replaced the sand track pads. Motion
triggered cameras offer greater reliability to capture animals using the crossing
structures. They also add the convenience of a longer uninterrupted monitoring
period, often going up to a month in between scheduled maintenance for
changing batteries and memory cards. The cameras also provide recordings of

19

times of animal crossings, animal behavior in response to crossing structures,
and ambient temperature during each crossing detection.

Figure 3. A biologist rakes a track pad after identifying species composition (Becker and Basting,
2010)

Armed with a better understanding of how animals respond to the different
crossing structures, the final 35 kilometers of the TCH is currently under
construction. This area consists of a different assemblage of species than the
first three phases. It is home to more large mammal species with low population
densities that are sensitive to human disturbances, such as wolverine (Gulo
gulo), grizzly bears, lynx (Lynx Canadensis) and moose (Alces alces), compared
with the typical fauna species found in the middle and lower Bow Valley, such as
cougar (Felis concolor), black bear (Usis americanus), wolves, elk and deer
(Ford et al, 2010). Consequently, larger, more open crossing structures (70

20

meters wide overpasses instead of 50 meters) at closer intervals (roughly 1.5
kilometers between structures instead of 2.7 kilometers) are planned for the final
stretch of the twinning project (Ford et al, 2010).

Figure 4. An animal crossing overpass structure in Banff National Park (ARC solutions.org, 2012)

The twinning project on the TCH in Banff National Park has produced
three main lessons for other DOTs:
1.
A long term monitoring program is needed to fully
understand animal behaviors and animal crossing structure
effectiveness (Ford et al, 2010). A minimum of three years
of data is needed to fully grasp the effectiveness of the
crossing structures. For some species, such as grizzly
bears and wolves, it may take up to five years for them to
feel comfortable enough to use a crossing structure (Parks
Canada Website www.pc.gc.ca, 2011). This allows time for
resident animals to recognize where the opportunities are for
safe crossings and what time of day to feel comfortable
enough to use the structure. That minimum length of time

21

also allows for migratory species to establish a migratory
corridor for future generations that include safe places to
cross highways provided by these structures.
2.
Use a diverse mix of crossing structures. By using a
mix of under and overpasses, culverts and expansive
bridges, and monitoring them for the minimum length of time
described above, highway planners were able to determine
the appropriate crossing structure for a particular target
species in a cost effective manner (Ford et al, 2010).
3.
Mitigation effectiveness should not be measured to
only include the reduced number of wildlife vehicle collisions,
but how effective the structures are in facilitating wildlife
gene flows, population and subpopulation dynamics, and
ecosystem functions and processes (Ford et al, 2010).
Looking at Banff National Park as a case study, research and observation
prove the animal crossing structures to work. As of July 2010, eleven different
species of large mammals have used 24 animal crossing structures more than
220,000 times since 1996 (www.pc.gc.ca, 2011). The key to making the project
an overall success is communicating to the public, especially to local
communities, that dollars used for wildlife crossing structures are justifiable.
Camera images and video footage provide the best evidence to the public that
creating wildlife crossing opportunities is a good decision for the driving public
and wildlife alike (Ford et al, 2010).
2c. WSDOT Study / Internship
Washington State Department of Transportation (WSDOT) has long
recognized the importance of keeping wildlife off its road surfaces. They are
using interns to continue the work of Patricia Cramer Ph.D. of Utah State
University and Julia Kintsch of ECO-resolutions, LLC who were contracted by the

22

WSDOT to perform a permeability study of existing structures for terrestrial
wildlife. The researchers were hired to develop a Passage Assessment System
(PAS) to help WSDOT evaluate existing transportation infrastructure for its ability
to provide terrestrial wildlife movement from one side of a roadway to another,
without the animal crossing the road at grade. The PAS system ranks
infrastructures based on Structural Functional Classes, which describes the type
of bridge or culvert. The system creates a common terminology of crossing
structures used by animals as well as assessment of the suitability for individual
animal species, classifying them into Species Movement Guilds (WHCWG 2010).
The system was then field tested along Washington roadways in linkage areas
identified in the Washington Statewide Habitat Connectivity Assessment of the
Washington Connected Landscapes Project: Statewide Analysis (Kintsch and
Cramer 2011). To validate the field tests, a series of motion triggered cameras
were installed near known animal crossing locations to assess animal use and
habits when using or approaching these structures. Seven locations were initially
selected, but due to theft and other logistical reasons, some locations were
monitored for only a couple of months.
This study is using the images captured during that initial study and
continuing monitoring efforts to assess species composition and preferred
crossing times. The sites examined in this thesis contain at least a full years'
worth of monitoring, with two exceptions. The first exception is the western
animal crossing structure on I-90 near North Bend. This site has less than a
years’ worth of data but was included due to its close proximity to the eastern I-

23

90 animal crossing structure. These structures were installed during the same
road construction upgrade of I-90 in the mid 1970's and have established
themselves as successful crossing structures for over 35 years.
The other exception is the fish passage culvert structure at Deadman
Creek on US 2 in Spokane. After camera installation in the summer of 2011, the
number of crossings by white-tailed deer (Odocoileus virginianus) quickly stood
out as being very successful with an average of over 6 (6.3194/day) deer
detections per day for the first six months of monitoring. Another way to look at it
is the structure averages one white-tailed deer detection every three hours and
20 minutes.
The other locations where crossing structures were monitored are Willapa
River at SR 6, Mosquito Creek at US 101, Tucker Creek at I-90, and the eastern
animal crossing structure at North Bend at I-90. These sites also represent the
diversity of habitats and deer species found in the state of Washington.
2d. Site Location
SR 6 at Willapa River
The Willapa River flows under State Route 6 just to the east of Raymond.
The bridge structure is an open span concrete bridge with riparian vegetation
associated with it. Managed forest land is in the vicinity as well as nearby rural
residences. This site is frequented by fishermen during the fall salmon and
winter steelhead season. It has been established for over 20 years and allows
for passage of larger animals that prefer a more open crossing opportunity.
There is no exclusionary fencing associated with this structure.

24

US 101 at Mosquito Creek
Just south of the Montesano cutoff (State Route 107) on US 101 is
Mosquito Creek, a tributary of the North River. A three sided box culvert was
installed within the last three years to aid spawning salmon and trout, replacing a
smaller corrugated pipe that spanned the width of the highway. Using current
standards for fish passable culverts, a stream simulation model was used to
determine the size of the new installation. This approach provides dry stream
banks inside the culvert for most stream flow conditions. WSDOT identified the
crossing structure as a priority for small and medium-sized animals, but
suggested that it was too small for larger animals with broader ranges, such as
elk and black bear. The structure measures 7 feet high x 15.75 feet wide x 138
feet long. At normal flows, Mosquito Creek allows small animals to utilize the
exposed banks in the box culvert, but most deer pass through the structure by
walking in the streambed. The surrounding habitat is managed Douglas fir
(Psuedotsuga menzesii) / Western hemlock (Tsuga heterophylla) forests with no
associated fencing to encourage animals to use the crossing structure. Cameras
were installed June 9, 2010 facing the openings of each side of the structure.
WSDOT estimates average traffic flow at 5,000 cars per day (Kintch and Cramer,
2011).
I-90 at Tucker Creek

25

The structure at Tucker Creek on I-90 spans a small seasonal creek. A
larger structure was installed in recent years to aid snow melt and water runoff
under I-90. A double box culvert was utilized with some exposed banks on each
side of the underpass. The structure measures 4’10” high x 9’ wide x 58’3” long.
At normal flows, Tucker Creek allows small animals to utilize the exposed banks
in the box culvert. Black-tailed (Odocoileus hemionus columbianus) and mule
deer (Odocoileus hemionus) pass through the structure using the streambed. A
railroad runs parallel to the highway along the south side of the structure. The
surrounding habitat and vegetation are composed of managed forest land with
nearby residential areas. There is a barbed wire right of way fence associated
with both sides of the openings. Traffic volumes vary seasonally, but average
25,000 to 65,000 cars/day (Kintsch and Cramer, 2011).
Deadman Creek US 2 Spokane
Just north of the city of Spokane US 2 intersects Deadman Creek. The
culvert was recently replaced to aid access to fish habitat. A large corrugated
steel pipe was used to allow natural stream flows through the structure. At
normal flows, there are exposed banks on either side of the river channel;
however the white-tailed deer that utilize the crossing structure often pass
through using the stream bed. The surrounding area is a typical riparian area
that has been enhanced through re-vegetation after the installation of the new
structure. There is no fencing associated with the highway in this area, however
the highway is built on deep fill and steep slopes above the riverbed that act as
barriers that impede animals from crossing at grade. The area in the immediate

26

vicinity of the structure is heavily frequented by humans, especially during the
warmer summer months.

I-90 North Bend West
On the west side of North Bend, I-90 is a divided highway where there are
two pairs of crossing structures: a pair referred to as I-90 West and a pair
referred to as I-90 East. Although these pairs of crossing structures are only two
miles apart, they demonstrate distinctly different characteristics in terms of
crossing structure attributes and the diversity of fauna that are utilizing them.
The crossing structures at I-90 West are open concrete bridges, but are
only about 12 feet high inside the structures. There is a small drainage ditch that
runs through the structures, but never fully inundates the substrate under the
structure, leaving ample dry ground for animals to cross underneath. The north
side of the structures is maintained for overhead power transmission lines with a
wetland just beyond the cleared transmission line right of way. The south side of
the structure is a mixed managed forest that has an abandoned access road that
parallels the highway. There is significant human activity on the access road
based on cars parked at the entrance to the access road and humans
encountered during site visits to change memory cards and batteries in the
cameras. There is an 8’ exclusionary fence associated with both sides of the
crossing structure, however numerous breaches in the fence were present until
January 31, 2012 when holes were mended and the fencing material re-attached

27

during a site visit. One location on the north side of the structure is still currently
breached due to large downed trees that have landed on the fence.

I-90 North Bend East
The other structures near North Bend are the I-90 East structures. They
are large corrugated steel bottomless arch culverts measuring about 12 feet high.
A creek runs from the south to the north, but goes sub-surface when it reaches
the entrance to the culvert. It re-emerges at the north end of the south structure,
meandering freely through the north structure. Both ends of the crossing
structure are heavily wooded by managed forests and riparian areas associated
with the creek. The vegetation type is consistent throughout the large median
area as well. An 8’ exclusionary fence is associated with both sides of the
highway, however several breaches were present until January 31, 2012 when
breaches were mended and the fencing was re-attached. There are still several
locations where the fence is damaged due to fallen trees and snags. A chainsaw
is required to fully remove the downed woody debris.
To fully understand the significance of both the I-90 wildlife structures
some historical back ground is needed.
3. The I-90 Crossing Structures: The Whole Story
In the mid 1960's the Washington State Department of Highways was
looking for ways to fix the congestion problems on US 10 that ran through the
heart of the growing city of North Bend. During this time the interstate highway

28

system was renumbered to meet the American Interstate Highway System,
creating Interstate 90. So in 1964, The Bureau of Public Lands authorized an
expansion and straightening of Interstate 90 on the west side of Snoqualmie
Pass. An Advanced Study Plan was conducted from Issaquah to North Bend by
two consulting firms, Sverdrup, Parcel & Associates and Hammond, Collier &
Associates, which concluded in July, 1966 (Final EIS 1973). The plan was
approved by the Federal Highway Administration in March 1967. A coordinating
committee was formed to include the representatives from various agencies that
had a stake in the massive project. A meeting was held on December 3, 1969 in
North Bend to discuss the route options the new I-90 would take. Several were
considered including using the existing route through the town of North Bend.
The Highways Department distributed the plans of the various route options for
comment to other agencies and local governments. North Bend citizens voiced
their concerns for having a highway run straight through their town that was
expected to balloon in traffic volumes over the next thirty years. Following the
meetings a request for a corridor was submitted to the Federal Highway
Administration and a corridor was approved for plan 'A3.' This would route I-90
to the south of North Bend at the base of Rattlesnake Mountain, allowing for
natural growth of North Bend in the future. Unfortunately, it required 2.34 miles
of Issaquah Creek to be relocated (Final EIS 1973). Issaquah Creek was a free
flowing salmon bearing stream with miles of spawning habitat available to
returning salmon. Any alteration of streams in Washington State requires a
Hydraulics Project Approval (HPA) permit to be issued by the Department of

29

Fisheries. The "Hydraulics Code" was enacted in 1943 and gives regulatory
authority to the Department of Fisheries to ensure water quality standards during
construction projects. The size of the I-90 expansion project was so massive the
agency had to share the permit issuance duties with the Department of Game.
The newly adopted National Environmental Policy Act of 1969 went into
effect on January 1, 1970. Consequently, the Federal Highway Administration
prepared an Environmental Impact Statement in May of 1970. Again, the plans
with their environmental impacts were sent to all of the government agencies
affected by the project. Most agencies, having to comment on an Environmental
Impact Statement for the first time, gave the go ahead on the project. A few
notable exceptions were the Washington Department of Ecology stating:
"…because of their ecological and esthetic values, marsh areas,
lakes and streams should essentially be kept in their natural
settings (Final EIS 1973)."
The Department of Fisheries voiced concerns over lost salmon rearing
habitat and requested other stream sections in the vicinity be enhanced to make
up for the loss of habitat. The highways department responded, saying:
"The Department of Highways does not acknowledge responsibility
for replacing lost spawning areas in other sections of the stream
outside of the project area" (Final EIS 1973).
The Department of Fisheries agreed to pursue their mitigation through issuance
of the HPA permit and ultimately signed off on the project.
The Department of Game, having leverage with the hydraulics permit,
voiced their concern that the proposed route would eliminate historic game trails

30

where deer and elk migrate between winter and summer ranges. Carl Crouse,
Director of the Department of Game, wrote in a letter, dated July 31, 1973, to the
district engineer:
"In general, you do try to adjust your plans and construction activities to
prevent or lessen losses to fish; a similar approach towards terrestrial and
avian wildlife is not apparent. This is unfortunate" (Final EIS 1973).
The Highways Department responded by saying:
"Contacts with the Regional Office for the Department of Game failed to
identify methods by which impacts to terrestrial and avian wildlife can be
mitigated…" (Final EIS 1973).
However, a letter contained in the draft EIS, dated January 22, 1971, two
and a half years before Carl Crouse’s letter, from Eugene Dziedzic, Assistant
Chief for the Department of Game, stated:
Construction should provide a tunnel approximately 10' high by 30' wide
for the game crossing portion. Lighting should be provided for peak hours
of use; the State of Colorado found 64% of their use in tunnel crossings by
wildlife occurred between 2:00 am and 5:00 am when the tunnels were
lighted. The floor should be left natural, or dirt covered, and entrance and
exits left in as natural a state as possible. Fencing should start in the
vicinity of Echo Lake and extend to the overpass at approximately mile
post 30. Fencing specifications should include use of 11' long, 7" x 9"
penta-treated posts, buried a minimum of 2 1/2'. New woven wire fencing
of 10 gauge top and bottom wires, and 12 ½ gauge filler wire should be
used. The fence should be constructed of 47" widths of such wire, giving
it a total height of approximately 7' 10" above the ground. The woven wire
should be 2 or 3" above the ground and well stretched and stapled with 1
½" staples, at a rate of 14 staples per post. The two widths should be
laced together with lacing wire of not less than 12 gauge size or hog rings
of 9 gauge size" (Final Environmental Statement 1971).
This letter was reinforced by another letter from Carl Crouse who stated:

31

"Fencing will be required to eliminate danger from wildlife crossing the
highway. These fences, along with tunnels, will minimize wildlife mortality
and also assure passage and use of habitat on either side of the highway"
(Final Environmental Statement 1971).

The Department of Highways did agree to install exclusionary game
fencing along the corridor to eliminate terrestrial wildlife entering the roadway and
to funnel game through stream crossings under the highway. However, the
Department of Highways was still trying to decide whether to install expansive
bridges or culverts. Culverts, being the cheaper option, were preferred, but
would restrict fish and wildlife from utilizing them for safe passage under the
highway. Presented with mitigation alternatives two and half years prior to the
department's response to wildlife mitigation, it appears the Department of
Highways was still unsure of the crossing structures.
The Department of Game, having the authority to issue hydraulics permits,
insisted the animal crossing structures be installed in order for the permit to go
through (DeShazo email). With recent bridge recommendations being ignored on
a different project (South King County bridge design (Bob Pfiefer email)) and
losing court cases in the recent past (Department of Game vs. Department of
Highways in Sunnyside) the Department of Game was not going to back down on
this matter.
An Environmental Report was obtained through the Washington State
Library system for the reach that contains the present day game crossing
structures under I-90 near North Bend. The report does not contain the

32

agreement that was reached; however, through tough negotiations between both
departments the animal crossing structures were installed. It's a good thing they
did. The I-90 East structure is one of the most successful black bear crossing
structures in the Western United States.
4. Research Question and Hypothesis
There has been a long held theory that wildlife is most active at dawn and
dusk. The behavior associated with twilight times is called crepuscular behavior.
This research aims to quantify crepuscular behavior for megafauna species,
therefore my research question is:
Do megafauna species exhibit crepuscular activity when offered a safe
opportunity to cross a highway?
Using the data available on the images from the motion triggered cameras, a
quantifiable relationship of peak activity time and sunrise and sunset was
developed to disprove the null hypothesis:
Ho: There is no difference in peak activity times for megafauna species,
between crepuscular periods and other periods, near our highways.
Ha: Peak activity time occurs within one hour of dawn and one hour of
dusk.
5. Methods
As mentioned earlier, this study was initially set up by Dr. Patricia
Cramer and Julia Kintsch for WSDOT. Motion triggered cameras were deployed
throughout the state at bridges and culverts to capture images of animals that
use the structures to safely pass under the roadway. Thousands of images

33

stored on the WSDOT hard drives were catalogued and, with some added
formulas for data analysis, the data matrix started to take shape. The finished
inventoried pictures describe the camera location, the date of the detection, the
time of the initial detection, the temperature, animal species, species gender if
identifiable, age class of the animal if identifiable, whether the animal used the
structure to pass safely through, and if the animal was repelled. Being repelled
refers to an animal that crossed under the structure and, for various reasons,
returned the way it came from; usually within a minute. Each animal detection
was characterized as a single species. If more than one species was observed
in the same image, they were characterized as two separate detections. If a
species was observed for a prolonged period of time, it was considered one
detection, regardless of whether the same or different individuals were involved.
If identifiable features were seen in the image, such as antlers or an ear tag, the
number of individuals was characterized accordingly. If there was an interruption
between photos of more than 30 minutes, the detection was split into two
separate events. Although the primary research considers megafauna, smaller
animals were recorded as well. Megafauna was identified as coyote size and
bigger. Considering damage caused from wildlife vehicle collisions, WSDOT was
primarily interested in megafauna.
The original intent of this study was to provide evidence for the
assumption that megafauna is most active at sunrise and sunset. To achieve
this, the date of the detection was used as a starting point. Using excel, a
numeric value was assigned to each date. Excel calculates the value by how

34

many days the detection was since January 1, 1900, giving each date and time a
unique numeric value. Using the time stamp in the photos, the time of the
detection was converted to the fraction of the day that it corresponds to. For an
example, a detection time and date from Tucker Creek where a mule deer was
detected on July 9, 2010 at 4:00 am will be used. July 9, 2010 is 40,368 days
from January 1, 1900. 4:00 am equals .16667 of the day; therefore July 9, 2010
at 4:00am is equal to the numeric value of 40,368.16667.
Sunrise and sunset were determined at midpoints of two week intervals
associated with the detections. The same formulas and calculations were used
to determine a numeric value for sunrise and sunset. The numeric values were
then compared to determine how close to sunrise or sunset a detection took
place. A formula was then written in excel to choose if the detection was closer
to sunrise or sunset (Thanks Judy Cushing), regardless of if it was closer to the
past crepuscular time period or the future crepuscular period. The output of the
formula left a comparable number that stated how far from sunrise or sunset a
detection took place. This information was put into histograms indicating how
many crossings occurred during a given time (usually seasonally) and how far
from sunrise and sunset the detections took place.
A look at raccoons at I-90 West tested the formulas and methods.
Raccoons (Procyon lotor) are nocturnal animals so they would be expected to
exhibit activity peaks before sunrise and after sunset. Sunrise and sunset are
indicated with a black line, which falls in the middle of the graph. The blue bars
show night time detections and the yellow bars (not apparent with raccoon

35

detections) show daytime detections. The red bars describe the crepuscular
period one hour before and after sunrise and sunset. As the figure 6 graphs
show, raccoon activity, indeed, peaks after sunset and before sunrise.

Figure 5. Raccoon sightings relative to sunrise at I 90 West.

sunset

36

Figure 6. Raccoon sightings relative to sunset at I 90 West.

6. Results
Willapa River
The dominant species at the Willapa River site was black-tailed deer with
101 individual detections and 23.27% of the total species composition.
Frequency of detections was .3 deer detections per day. Six coyote (Canis
latrans) detections were recorded, making up the only other megafauna species
at this site. Other species included raccoons, possums (Didelphis virginiana),
and striped skunks (Mephitis mephitis). Summer was the most active season
and winter had the lowest detections and frequencies for all species except for
possums which exhibited a spike in activity during the winter months. Deer at
this site tended to cross at a temperature range of 50-70 degrees Fahrenheit,
making up 75.36% of detections. This site lacked enough data to do a standard
chi squared test, so the G-adjusted chi squared test was done. The results show
that crepuscular activity is not significant at alpha = .05 for sunrise and sunset.
See figures 13-16 for the statistical analysis. These results are also seen in the
histograms in figure 7:

37

hrs. to sunrise

sunset

hrs. to sunset

Figure 7. Histograms characterizing detections in relation to sunrise and sunset at Willapa River

Mosquito Creek
The dominant species at Mosquito Creek was black-tailed deer with 164
detections and a frequency of .31 deer detections per day over the study period.
Black-tailed deer made up 60% of the species using this structure. Elk were the
only other megafauna species captured on camera at his location. They have

38

appeared at the west entrance, feeding on the herbaceous vegetation, but none
have attempted to enter the structure. Other species include raccoons which
make up 32% and one detection each for a migrating salmon and mink
(Neovison vison). Summer exhibited the majority of the detections and no
megafauna were detected during the winter months. The preferred temperature
range for deer crossings was between 40 and 60 degrees Fahrenheit which
accounted for 78.68% of detections.
This site lacked enough data to do a standard chi squared test, so the Gadjusted chi squared test was done. The results reflect that crepuscular activity
is not significant at alpha = .05 for sunrise and sunset. See figures 13-16 for the
statistical analysis. Histograms in figure 8 describe peak activity times may
suggest a peak evening activity period.

hrs. to sunrise

39

sunset

.Figure 8. Histograms characterizing detections in relation to sunrise and sunset at Mosquito Creek.
North Bend I-90 West
This study had originally intended the crossing structures at I-90 near
North Bend to be considered similar sites or at least similar enough to mirror
each other. The structures are approximately the same age, being built in the
mid 1970's and with similar habitat on either side of the highway. They are less
than two miles apart. The I-90 West site does not have the luxury of having more
than one season's worth of data to compare to; however there seems to be a
more diverse composition of species using the structure. Black-tailed deer,
raccoons, and black bears each make up approximately one third of the species
composition (33.95%, 28.4%, and 23.46% respectively). There were a significant
number of detections that were unclassifiable at this site due to the camera
position at the south end during low light or no light periods. A more
advantageous camera position could reveal a single dominant species.
When the study period's data was broken into seasons, a different
dominant species emerged for each season. Summer 2011 saw bears as the

40

dominant species with 30 detections and 58.82% of species composition. Blacktailed deer and raccoons were present, but at a lower composition, 17.65% and
13.73% respectively. The fall saw increased activity from black-tailed deer with
34 detections and 44.74% of composition. Raccoon activity also peaked during
fall with 23 detections worth 30.26% composition. Black bears fell to 8 detections
worth 10.53%. Bobcats (Felis rufus) appeared during the fall, although they were
rare. As stated before, a more advantageous camera position would have most
likely revealed more frequent use by bobcats. Preferred temperatures for the
dominant species crossings were in the 30-50 degree range.
This site lacked enough data to do a standard chi squared test, so
the G-adjusted chi squared test was done. The results demonstrate that
crepuscular activity is not significant at alpha = .05 for sunrise and sunset. See
figures 13-16 for the statistical analysis. Histograms in figure 9 do not reveal a
particular pattern; however, a noticeable trend emerged from this site: increased
activity of large carnivores decreases activity of prey animals and vice versa.

41

sunrise

hrs. to sunrise

hrs. to sunset

Figure 9 Histograms characterizing detections in relation to sunrise and sunset at I-90 West.

North Bend I-90 East
The site at I-90 East consisted mainly of carnivores. For the study period
from summer 2010 to spring 2011, 86.16 % of all animal detections were
carnivores. Black bears and coyotes appear to be the dominant species
consisting of 40.77% each of species composition. Black-tailed deer were
detected, but only 17 animal detections were recorded for the entire study period

42

equating to 9.34%. Bobcats and raccoons were detected as well, each
comprising 1.65% of detections. One elk was detected, and it did use the
structure.
The second half of the study period, from summer 2011 to winter 2012,
demonstrated a remarkable increase in black bear activity. While the previous
year recorded 53 bear detections for the whole season making up 40.77% of
species composition, during the 2011-2012 season 121 black bear detections
were recorded making up 57.62% of all animal detections. Frequency during the
summer months increased from .25 bears per day in 2010 to 1.2 bears per day in
2011. The reason for the increase is unclear. Camera operation times and
malfunctions could be a contributing factor, but cannot be determined. An
equally remarkable trait for this season is that only one black-tailed deer was
detected. This, along with the other study site at North Bend, begs the question:
Does increased activity in carnivores deter prey species from using the crossing
structure? Carnivorous animals make up 98.57% of all detections for the 20112012 year. Bobcats also increased in frequency with 11 animals being detected
as opposed to only 3 detections in the previous study year.
Black bears seem to prefer a temperature range similar to that of blacktailed deer. 92.85% of bear detections occurred in the 40 – 60 degree
temperature range. Naturally, peak activity occurs in the summer months.
This site lacked enough data to do a standard chi squared test, so the Gadjusted chi squared test was done. The results demonstrate that crepuscular
activity is not significant at alpha = .05 for sunrise and sunset. See figures 13-16

43

for the statistical analysis. The histograms in figure 10 reveal a preference for
crossing during the daytime.

hrs. to sunrise

hrs. to sunset

Figure 10. Histograms characterizing detections in relation to sunrise and sunset at I-90 East.

Tucker Creek I 90 mp 73
The Tucker Creek site has a mix of mule deer and black-tailed deer as the
dominant species. A summary of the study period from summer 2010 to spring

44

2011 resulted in 147 deer detections, a total of 70% of the detections. Other
species detected at the site are raccoons that make up 10%, a house cat that
made up 10.95%, squirrels (Sciurus carolinensis) made up 5.24% and rabbits
(Oryctolagus cuniculus) made up 2.86%. Summer is the most active time of year
for deer here with 99 animal detections. Frequency of use is 1.11 deer
detections per day. Fall sees a sharp decline in deer detections. Only 15 deer
detections were recorded in the fall of 2010 for a frequency of .54 deer detections
per day. Winter sees a further decline with nothing being detected in the winter
of 2011. It is unknown if this is due to the site being at an elevation where winter
time snow could be a factor, pushing animals to the valley floors or if there was a
camera malfunction during that time. This site also experiences high water flows
during the spring snow melt that could be a factor for the lack of animal
detections during that time period. The first deer detection of the 2011 year
occurred on May 15. Once deer started to be detected, they seem to increase
their frequency of use dramatically with 33 deer being detected for the remainder
of the spring season.
Deer at this site often cross in a temperature range of 40 – 60 degrees
with 74.53% of detections occurring in that range. However, recall these
detections occurred mainly in the summer months, weighting the detections to
more summerlike temperatures.
This site lacked enough data to do a standard chi squared test, so the Gadjusted chi squared test was done. The results demonstrate that crepuscular
activity is not significant at alpha = .05 for sunrise and sunset. See figures 13-16

45

for the statistical analysis. The histograms reveal a preference for nighttime
activity.

sunrise

hrs. to sunrise

hrs. to sunset

Figure 11. Histograms characterizing detections in relation to sunrise and sunset at Tucker
Creek.

Deadman Creek Spokane US 2

46

Deadman Creek animal composition consists of almost entirely whitetailed deer. A summary for the study period of summer through winter of 20112012 indicates that 99.25% of all animals using the culvert for a crossing
opportunity were white-tailed deer. The most active time of year occurred during
the fall when there were 441 white-tailed deer detections for a frequency of 4.96
animal detections per day. The winter season also displays some of the same
activity characteristics as the fall with 98.97% of species being white-tailed deer.
Frequency of use increased slightly to 5.04 deer detections per day. One moose
was detected during the summer, but was repelled from the structure. It looked
mainly to be feeding.
The summer months at the Deadman Creek site were the least used for
crossing opportunities. 248 white-tailed deer detections were recorded during
this time of year for a frequency of 2.99 deer detections per day. Temperature,
often exceeding 90 degrees in the daytime, could be a strong factor in the
decreased use during the summer. This structure was also under construction
during the summer months, affecting frequencies and detection times. Deer at
this site prefer to cross at a temperature range from the 20's to the 40's with
73.17% of detections happening in that temperature range.
Using the Chi Squared test, this site was statistically significant for whitetailed deer to exhibit detection times that were different. The histograms in figure
12 reinforce the theory that deer at this site exhibit crepuscular behavior.

47

hrs. to sunrise

sunset

hrs. to sunset

Figure 12. Histograms characterizing detections in relation to sunrise and sunset at Deadman
Creek

7. Chi Squared Analysis
A chi squared test was used to determine if the observed time of detection
was statistically significant and different from the expected time of detections,

48

which in this case is sunrise and sunset. Using excel, the calculations for the chi
squared analysis were created for the six hour time blocks of observed
detections surrounding sunrise and sunset, comparing observed times of
detections to sunrise and sunset, or what we would expect to observe. The
calculations gave chi squared statistics that were compared with a degrees of
freedom chart, ultimately determining if the observed detections were statistically
significant at the given degrees of freedom.
When a Chi Squared test is conducted, there are some assumptions to
consider. A Chi Squared test can be used when no more than 20% of the
expected counts are less than 5 and all individual counts are 1 or greater (Moore
2001). To properly conduct the analysis, the dominant species for each site was
used and in the case of North Bend west, there were multiple dominant species
to consider, consequently, multiple species of megafauna were used.
Deadman Creek in Spokane is the only site in this research to have a
sufficient sample size to do the standard Chi Squared test. Both the sunrise and
sunset Chi Squared analyses reveal a significant statistical result. See figures 13
and 14 for the statistical analysis.

49

Sunrise Chi Squared Analysis
Site Name

df

sunrise

results

Deadman Creek
Tucker Creek

10
15

35.10924
15.93506

significant @ pvalue < .001
NA

North Bend I 90 mp 29 (bears)

15

10.47743

NA

North Bend I 90 mp 27 (multi
species)
Mosquito Creek
Willapa River
all sites deer
all sites dominant species

10
15
15
20
20

12.76584
17.99794
11.15743
26.0931
28.28854

NA
NA
NA
NA
NA

Figure 13. Chi Squared analysis for sunrise

Sunset Chi Squared Analysis
Site Name
Deadman Creek

df
10

sunset
27.73685

results
significant @ pvalue < .0025

Tucker Creek
North Bend I 90 mp 29 (bears)
North Bend I 90 mp 27 (multi species)

15
15
10

17.17898
6.91093
17.40362

NA
NA
NA

Mosquito Creek
Willapa River

15
15

NA
NA

all sites deer

20

16.34875
lacks data
(n=12)
24.16685*

all sites dominant species

20

23.59207

NA

NA

Figure 14. Chi Squared analysis for sunset

When a sample size is not sufficient to perform a standard Chi Squared
analysis, a G-adjusted Chi Squared test can be used. A G-adjusted Chi Square
test accounts for a smaller sample size and is a more conservative analysis. In
the case at the Willapa River site, the number of detections relating to sunset

50

was still insufficient to properly conduct a G-adjusted Chi Squared analysis, so
the G-adjusted Chi Squared table notes the insufficient data.
Sunrise G-Adjusted Chi Squared Analysis

Site name

df

sunrise

Deadman Creek

10

35.10924

results
significant with Chi Squared
analysis

Tucker Creek

15

16.62973

not significant

North Bend I 90 mp 29 (bears)

15

not significant

North Bend I 90 mp 27

10

5.65859
15.68477

significant @ p-value = .15

Mosquito Creek

15

17.82821

not significant

Willapa River
all sites deer

15
20

8.05851
25.35177

not significant
significant at p-value = .20

all sites dominant species

20

27.70212

significant @ p-value = .15

Figure 15. G-adjusted Chi Squared analysis for sunrise

Sunset G-Adjusted Chi Squared Analysis
Site name
Deadman Creek

df
10

sunset
27.73685

results
*significant with Chi
Squared analysis

Tucker Creek

15

17.03612

not significant

North Bend I 90 mp 29 (bears)

15

8.90790

not significant

North Bend I 90 mp 27

10

16.78348

significant @ p-value = .10

Mosquito Creek

15

19.03908

significant @ p-value = .25

Willapa River

15

NA

all sites deer

20

lacks data
(n=12)
24.08331

all sites dominant species

20

26.48403

significant @ p-value = .20

significant @ p-value = .25

Figure 16. G-adjusted Chi Squared analysis for sunset

8. DiscussionThe only site with enough data to do a Chi Squared analysis was
Deadman Creek in Spokane. Both crepuscular periods were significant at alpha

51

= .05, therefore we can reject the null hypothesis that there is no difference in
peak activity times for white-tailed deer. The graphs in figures 12 point to a peak
activity period an hour before and an hour after sunrise and sunset. This would
suggest that white-tailed deer at Deadman Creek exhibit crepuscular activity.
The other study sites were not significant at alpha = .05 using the Gadjusted Chi Squared test. Two factors influenced the conclusions for the
western and central Washington sites. The first is that more data is needed. A
larger sample size would allow for a standard Chi Squared test as well as better
representation of the deer population near the study sites. The second is that
there could be other factors influencing megafauna detections near these study
sites.
The next logical variable to consider is traffic volumes. Elevated traffic
volumes have been shown to be a barrier for species dispersal and could be
affecting sites such as the I-90 sites. However if an animal was provided a safe
opportunity to cross the highway while traffic levels were elevated, would the
animal prefer to use the structures to cross the highway? This dilemma points to
confounding influences: Does elevated traffic levels deter animals from using
crossing structures because they are uncomfortable around higher traffic
volumes or are animals only using the crossing structures during higher traffic
because they offer a 100% guarantee of safe passage. I would expect this to be
a choice for individual animals to make for the following reasons: If a migrating
animal encountered I-90 and was unfamiliar with the local area, would roadway
avoidance behavior trump an opportunity to use a crossing structure to safely

52

cross the highway? Likewise, a resident animal in a similar situation,
encountering I-90 and having to decide to cross, knowing that a safe opportunity
exists in its home range; would that animal be compelled to prefer using the
structure when traffic volumes made crossing at grade virtually impossible?
Without using a locator system or genetic tests to distinguish individual animals
from each other, this type of analysis encounters too many confounding variables
to draw any conclusions from.
Other patterns that have emerged from this data and their analyses
warrant discussion. The first is the detection results at Tucker Creek. The fact
that most of the detections occurred at night point to two variables that could
explain that pattern: temperature and human activity. The location of the site, in
central Washington, and the fact that most of the detections occurred during the
summer months and the deer's preferred temperature range of 40-60 degrees, I
would expect that during the heat of the day, deer spent time bedded down
avoiding unnecessary movements.
The fact that there was a continued human presence using ATVs in the
culvert could also affect ungulate activity patterns. Research shows that human
activity around wildlife structures can shift ungulate activity to a nighttime regime
(Cramer Email and Evink 2007).
The second pattern to emerge from this data is the species composition at
the two North Bend sites. Having black bears utilize crossing structures has
proven to be difficult for some areas of the West. Consulting with Patty Cramer
revealed the frequency of bear detections here is extraordinary. In her four years

53

of monitoring 4 different states, collecting over 1.5 million photos, she has only
encountered roughly 25 successful black bear crossings (Cramer Email). The
success at both North Bend sites could be attributed to the age of the structure
since both structures have had numerous successful bear crossings and have
been established for over 35 years.
The fact that the North Bend sites see an inverse relationship between
carnivorous animals and prey animals is unique as well. It does seem logical for
a prey species to avoid a predator species, but this would again point to the age
of the structures. If carnivorous animals have established themselves as using
these structures to access resources, their scent would persist in the culverts,
ultimately discouraging use by prey species. Jon Beckman and Jodi Hilti
discovered a presence absence relationship between mother moose and her
calves and bears (Beckmann and Hilti, 2010). In some cases, mother moose
tends to give birth to her calves near roads because she knows it is a safe place
away from bears. The research suggests avoidance behavior exists in other
ungulate species and could be a factor in the species composition at the North
Bend sites. This would suggest the frequency of use by predator species is
deterring prey species from using the structure.
Adam Ford and Tony Clevenger (2010) found that there is no evidence to
suggest that predators use these wildlife funnels to capture and ambush prey, so
the frequent use does not appear to be for capturing prey species; that would
further imply that this location is not a significant migration route for migrating or

54

traveling prey species. Continued monitoring at these sites should reveal if the
study period species composition and frequencies are consistent.
The third pattern to reveal itself is the frequency of use at Deadman
Creek. The fact that the structure sees this many deer is extraordinary compared
to the other sites. What makes it even more astounding is the fact that the
structure is less than a year old.
A successful wildlife crossing structure is largely based on placement
within the landscape. The area to the north of Spokane where this structure is
located is in one of the worst hotspot in the state for deer-vehicle collisions. By
offering a safe alternative to cross the highway, the whitetail deer at this location
are now safely migrating through an otherwise hazardous area.
9. Thoughts for the Future
1. Continue collecting data at regular intervals – There is not enough data to do
most statistical analyses. Parks Canada claims that three to five years' worth of
data is needed to fully understand the relationships and patterns that develop
from wildlife using the structure (www.pc.gc.ca, 2011)
2. This analysis would prove more valuable if there was control data to compare
our findings to. Wildlife may exhibit different behavior and peak activity periods
when the influences of a roadway are not present. Woody Meyers, a WDFW
ungulate biologist, is currently conducting a study that measures activity of mule
deer on a scale of every 15 minutes.
3. Target elk as a species of interest. WSDOT thought more elk would be
detected with regularity, but they were not, and when they were, they rarely

55

crossed through the structure. Monitoring the overpass being planned for
construction east of Rock Knob near the I-90 summit should reveal if overpasses
are preferred by elk or if we need to think differently about how to keep elk off the
roads while still providing access to migration routes.
10. Conclusion
Although the study lacked sufficient data to do a standard Chi Squared
test, and the majority of the research did not support the theory that megafauna
species are most active at twilight, other patterns emerged from organizing and
cataloguing the data. Some of these patterns warrant further research. As
discussed earlier, the species composition at the North Bend sites is unique.
Further investigation at other potential sites will confirm the uniqueness of the
structure appearing to favor carnivores.
Nocturnal patterns at Tucker Creek could shed some light on the question
of how much human activity could a population tolerate without shifting activity
regimes.
The Deadman Creek area north of Spokane is a success story on how
and where to locate crossing opportunities for white-tailed deer. The fact that the
structure was used with regularity during construction points to the need for
further research for future crossing opportunities in the area.
The only conclusion the research offered was white-tailed deer exhibit
crepuscular activity at one site. Further monitoring at other research sites in
Washington State could reveal if white-tailed deer are more prone to crepuscular
activity or if other influences affect activity patterns.

56

Cataloguing and organizing the data was not just to test crepuscular
activity in megafauna species. Data collected to answer the original research
question can be applied to answer several other questions. As is common with
scientific research, this research revealed more questions than answers gained.
That's job security for the biologists and future interns at WSDOT.

57

Appendix 1
Site Locations map

58

Appendix 2
Willapa River
Species composition and Frequencies
summary without humans
species
bt deer
human
coyote
raccoon
possum
skunk
house cat
total

detections
54
0
6
59
142
6
0
267

% comp.
20.22%
0.00%
1.71%
22.10%
53.18%
1.71%
0.00%
100%

individuals % comp.
101
23.27%
0
0.00%
6
1.38%
90
20.74%
147
33.87%
6
1.38%
84
19.35%
434
100%

Frequencies
species
bt deer
human
coyote
raccoon
possum
skunk
house cat
total
total days

detections frequency/day
54
0.16
140
0.41
6
0.02
59
0.17
142
0.42
6
0.02
84
0.25
491
340

individuals
frequency/day
101
0.30
220
0.65
6
0.02
90
0.26
147
0.43
6
0.02
84
0.25
654
340

59

Willapa River camera locations

60

Willapa River East Camera

Willapa River West Camera
Appendix 3

61

Mosquito Creek
Species Composition and Frequencies
summary without humans
species
elk
bt deer
human
raccoon
mink
salmon
total

Frequencies
species
elk
bt deer
human
raccoon
mink
salmon
total days

# detections
1
108
0
63
1
1
174

# detections
1
108
0
63
1
1
534

% comp.
0.57%
62.07%
0.00%
36.21%
0.57%
0.57%
100%

frequency/day
0.0019
0.2022
0.0000
0.1180
0.0019
0.0019

# individuals
1
164
0
87
1
1
254

# individuals
1
164
0
87
1
1
534

% comp.
0.00%
57.00%
0.00%
43.00%
0.00%
0.00%
100%

frequency/day
0.0019
0.3071
0.0000
0.1629
0.0019
0.0019

Mosquito Creek camera locations

62

Mosquito Creek East Camera

Mosquito Creek West Camera

63

Appendix 4
North Bend I 90 West
Species Composition and Frequencies
summary without humans
species
bt deer
black bear
human
bobcat
coyote
rabbit
raccoon
total
Frequencies
species
bt deer
black bear
human
bobcat
coyote
rabbit
raccoon
total days

# detections
45
38
0
10
11
1
41
146

% comp.
30.82%
26.03%
0.00%
6.85%
7.53%
0.68%
28.08%
100.00%

# detections frequency/day
45
0.18
38
0.15
29
0.11
10
0.04
11
0.04
1
0.00
41
0.16
253

# individuals
55
38
0
10
12
1
46
162

% comp.
33.95%
23.46%
0.00%
6.17%
7.41%
0.62%
28.40%
100.00%

# individuals frequency/day
55
0.22
38
0.15
41
0.16
10
0.04
12
0.05
1
0.00
46
0.18
253

North Bend I 90 West camera

64

I 90 West-North Camera

I 90 West-South Camera

65

Appendix 5
North Bend I 90 East
Species Composition and Frequencies
summary without humans
species
elk
bt deer
black bear
human
bobcat
coyote
raccoon
possum
rabbit
total
Frequencies
species
elk
bt deer
black bear
human
bobcat
coyote
raccoon
possum
rabbit
total days

# detections
1
11
163
0
14
100
7
1
2
196
# detections
1
11
163
0
14
100
7
1
2
299

% comp.
0.00%
0.51%
61.73%
0.00%
5.61%
28.06%
2.55%
0.51%
1.02%
100%
frequency/day
0.0033
0.0368
0.5452
0.0000
0.0468
0.3344
0.0234
0.0033
0.0067

# individuals
1
18
174
0
14
121
9
1
2
210
# individuals
1
18
174
0
14
121
9
1
2
340

% comp.
0.00%
0.48%
57.62%
0.00%
5.24%
32.38%
2.86%
0.48%
0.95%
100%

frequency/day
0.0029
0.0529
0.5118
0.0000
0.0412
0.3559
0.0265
0.0029
0.0059

I 90 East camera locations

66

I 90 East-North Camera

I 90 East-South Camera

67

Appendix 6
Tucker Creek
Species Composition and Frequencies
summary without humans
species
deer sp (mule &bt)
human
coyote
raccoon
skunk
house cat
rabbit
squirrel
total
Frequencies
species
deer sp (mule &bt)
human
coyote
raccoon
skunk
house cat
rabbit
squirrel
total days

# detections
102
0
1
10
1
21
6
10
151

% comp.
67.55%
0.00%
0.66%
6.62%
0.66%
13.91%
3.97%
6.62%
100%

# individuals
% comp.
147
70.00%
0
0.00%
1
0.48%
21
10.00%
1
0.48%
23
10.95%
6
2.86%
11
5.24%
210
100%

# detections frequency/day
102
0.55
38
0.20
1
0.01
10
0.05
1
0.01
21
0.11
6
0.03
10
0.05
187

# individuals
frequency/day
147
0.79
64
0.34
1
0.01
21
0.11
1
0.01
23
0.12
6
0.03
11
0.06
187

Tucker Creek camera locations

68

Tucker Creek North Camera

Tucker Creek South Camera

69

Appendix 7
Deadman Creek
Species Composition and Frequencies
Species Composition Summary w/o humans
species
# detections
% comp
moose
1
0.10%
wt deer
624
99.12%
humans / dogs
0
0.00%
coyotes
2
0.20%
raccoons
6
0.59%
total megafauna
633
100%
Frequencies
species
#detections
frequency/day
moose
1
0.004
wt deer
624
2.48
humans / dogs
189
1.25
coyotes
2
0.01
raccoons
6
0.02
total days
252

# individuals
1
1072
0
2
8
1083
# individuals
1
1072
360
2
8
252

% comp.
0.07%
99.25%
0.00%
0.14%
0.55%
100%

frequency/day
0.00
4.25
1.43
0.01
0.03

Deadman Creek camera locations

70

Deadman Creek Downstream Camera

Deadman Creek Upstream Camera

71

Bibliography
Becker, Dale and Basting, Patrick. "Reconstruction of US Highway 93:
Collaboration between Three Governments." Safe Passages: Highways,
Wildlife, and Habitat Connectivity. Ed. John Beckmann. Island Press
Washington DC 2010. pp. 173-187.
Beckmann, Jon and Hilti, Jodi. "Connecting Wildlife Populations in Fractured
Landscapes." Safe Passages: Highways, Wildlife, and Habitat Connectivity. Ed.
John Beckmann. Island Press Washington DC 2010. pp. 3-16.
Charry, Barbara and Jones, Jody. Traffic Volume as a Primary Road
Characteristic Impacting Wildlife: A Tool for Land Use and Transportation
Planning. 9-13-2009.
Clevenger, A., B. Chruszcz, K. Gunson, and J. Wierzchowski. 2002. Roads and
Wildlife in the Canadian Rocky Mountain Parks – Movement, Mortality, and
Mitigation. Final Report (October 2002). Report prepared for Parks Canada,
Banff, Alberta.
Cramer, Patricia. “RE: WSDOT Crossing Structures.” Message to Dave
Kangiser. May 18, 2012. Email.
DeShazo. "RE: I-90 Game Crossings." Message to Kelly McAllister. February
4, 2012. Email.
Evink, Gary. Interactions with Roadways and Wildlife Ecology. National
Cooperative Research Program NCHRP Synthesis 305. Transportation
Research Board. Library of Congress Control No. 2002111522. Copyright 2002.
Ewers, Robert M. and Didham, Rapheal K. Pervasive Impact of Large-Scale
Edge Effects on a Beetle Community. Proceedings from the National Academy of
Science April 8, 2008 vol. 105 no. 14 5426-5429. Accessed June 8, 2012 at:
http://www.pnas.org/content/105/14/5426.short.
Ford, Adam and Clevenger, Anthony. Validity of the Prey-Trap Hypothesis for
Carnivore-ungulate Interactions at Wildlife-Crossing Structures. Conservation
Biology, vol. 24, no. 6 pp. 1679-1685. 2010.
Ford, Adam; Clevenger, Anthony; Rettie, Kathy. "The Banff Wildlife Crossing
Project: An International Public-Private Partnership." Safe Passages: Highways,
Wildlife, and Habitat Connectivity. Ed. John Beckmann. Island Press
Washington DC 2010. pp 157-172.

72

Foreman, Richard et al. Road Ecology: Science and Solutions. Island Press
Washington DC. 2003.
Hilti, Jodi, Lidicker, W. Z., Merenlender, A. M. Corridor Ecology: The Science
and Practice of Linking Landscapes for Biodiversity Conservation. Island Press,
Washington DC. 2006.
Huijser, Marcel and McGowen, Pat. "Reducing Wildlife-Vehicle Collisions." Safe
Passages: Highways, Wildlife, and Habitat Connectivity. Ed. John Beckmann.
Island Press Washington DC 2010. pp. 51-74.
Kintsch, Julia and Cramer, Patricia. Permeability of Existing Structures for
Terrestrial Wildlife: A Passage Assessment System. WSDOT Research Report.
WA-RD 777.1. Office of Research and Library Services. July 2011
Loyd, John (Lead Author); Cleveland, Cutler (Topic Editor) "Road ecology". In:
Encyclopedia of Earth. Eds. Cutler J. Cleveland (Washington, D.C.:
Environmental Information Coalition, National Council for Science and the
Environment). [First published in the Encyclopedia of Earth April 11, 2008; Last
revised Date December 24, 2011; Retrieved June 10, 2012
<http://www.eoearth.org/article/Road_ecology>
Moore, David S. Statistics: Concepts and Controversies, Fifth Edition. W.H.
Freeman and Company New York, New York 2001.
Parks Canada. 1995. Initial assessment of proposed improvements to the TransCanada Highway in Banff National Park-Phase IIIA Sunshine Interchange to
Castle Junction Interchanges. Parks Canada, Canadian Heritage, Ottowa,
Ontario.
Parks Canada website. Modified 6-21-2011. Accessed: February 12, 2012.
www.pc.gc.ca.
Pfiefer, Bob. "RE: I-90 Game Crossings." Message to Kelly McAllister.
February 5, 2012. Email.
Rost, Gregory R. and Bailey ,James A. Distribution of Mule Deer and Elk in
Relation to Roads. The Journal of Wildlife Management Vol. 43, No. 3 (Jul.,
1979), pp. 634-641
Thurber, Joanne M. et al. Gray Wolf Response to Refuge Boundaries and Roads
in Alaska. Wildlife Society Bulletin. Vol. 22, No. 1 (Spring, 1994), pp. 61-68

73

US Department of Transportation. Federal Highway Administration. WildlifeVehicle Collision Reduction Study: Report to Congress. Publication number:
FHWA-HRT-08-034. August, 2008.
Waller, John S. and Servheen, Christopher. Effects of Transportation
Infrastructure on Grizzly Bears in Northwestern Montana. Journal of Wildlife
Management 69 (3): 985-1000. 2005.
Washington State Highways Department Final Environmental Impact Statement.
SR 90 Issaquah to West Snoqualmie. Report Number FHWA-WN-EIS 73-08F.
October 1973.
Washington State Highways Department Final Environmental Statement. SR 90
West Snoqualmie to Tanner (North Bend Bypass). Report Number C.S. 1707R.
February 1971.
Washington Wildlife Habitat Connectivity Working Group (WHCWG). 2010.
Washington Connected Landscapes Project: Statewide Analysis. Washington
Departments of Fish and Wildlife, and Transportation, Olympia, WA
WSDOT Executive Order number E 1031. Protecting and Connections for High
Quality Natural Habitats. July 23, 2007.
WSDOT Executive Order number E 1018.01. Environmental Policy Statement.
April 7, 2007.
www.arcsolutions.org. copyright 2012. Website accessed June 12, 2012.
www.wsdot.wa.gov/Environment/Biologyhome. Copyright 2012. Website
accessed June 12, 2012.

74

Media

Kangiser_D-MES2012.pdf