Examining the Relationship Between Landscape Connectivity and the Breeding Effort of the Red-Legged Frog (Rana Aurora) in Western Washington Wetlands

Item

Title

Eng Examining the Relationship Between Landscape Connectivity and the Breeding Effort of the Red-Legged Frog (Rana Aurora) in Western Washington Wetlands

Date

2012

Creator

Eng Holcomb, Chris

Subject

Eng Environmental Studies

extracted text

EXAMINING THE RELATIONSHIP BETWEEN LANDSCAPE
CONNECTIVITY AND THE BREEDING EFFORT OF THE REDLEGGED FROG (Rana aurora) IN WESTERN WASHINGTON
WETLANDS

by
Chris Holcomb

A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2012

©2012 by Chris Holcomb. All rights reserved.

This Thesis for the Master of Environmental Studies Degree
by
Chris Holcomb

has been approved for
The Evergreen State College
By

________________________
Gerardo Chin-Leo, Ph.D.
Member of the Faculty

________________________
Date

ABSTRACT
Examining the Relationship Between Landscape Connectivity and the Breeding Effort of
the Red-Legged Frog (Rana aurora) in Western Washington Wetlands
Chris Holcomb
Amphibians provide valuable ecosystem services in many environments. However, over
the last 30 years, populations of many amphibian species have been declining, largely
due to habitat destruction and fragmentation. The Red-Legged Frog, Rana aurora, favors
mature forests for the non-breeding portion of the year and utilizes forests at relatively far
distances from the wetlands and ponds in which it breeds. Without careful planning and
landscape stewardship, the expected levels of human development may cause significant
declines of R. aurora in the Puget Sound lowlands. An estimate of the level of upland
habitat loss and fragmentation that R. aurora can tolerate is an important area for
research. This study contributes to the understanding of the effects of habitat
fragmentation on R. aurora. I analyzed 14 sites, each of which included a wetland with
habitat considered to be ideal for R. aurora breeding: physical characteristics of these
wetlands included seasonal or semi-permanent hydrology and dominance by emergent
vegetation or partial dominance by small shrubs. The sites varied from each other with
respect to upland connectivity characteristics when land covers within 2 km of each
wetland in the sample were considered. Using R. aurora egg mass counts in each wetland
as an index for the breeding population size, I found a positive relationship between
breeding effort and more extensive, well-connected habitats on all sides of the study
wetland. There was a strong correlation (r2=.79) between egg mass quantities and the size
of the forest patch that was physically connected to each study wetland in each site. In
addition, I observed a significant difference in the average quantity of egg masses in sites
that were near a road and those that were farther away. Sites that were located within
.25km of a road averaged 60 egg masses while those that were farther away from roads
averaged 268 egg masses (p<0.05). Other connectivity factors were analyzed
qualitatively; higher traffic levels on nearby roads coincided with lower population size.
Easier access to secondary forest patches coincided with higher population size. Higher
population numbers coincided with landscapes devoted to wilderness preservation,
second growth forest preservation, and timber production while urban landscapes and
those featuring mixtures of forestry, rural residential development, agriculture and
highways coincided with smaller populations. Suggestions for further research include
increasing the sample size and analyzing the connectivity that surrounds each wetland
with a least cost analysis in GIS. Least cost analysis assigns numbers that represent
energy expenditure and risk of death to various land covers in a landscape and models a
species success at crossing such landscapes.

Table of Contents
Chapter 1: INTRODUCTION.......................................................................................................... 1
Overview ...................................................................................................................................... 1
Observed Declines and Official Listings ..................................................................................... 2
The Ecological Significance of Amphibians and R. aurora ........................................................ 3
Human Impacts to Rana aurora .................................................................................................. 4
Habitat and Connectivity ......................................................................................................... 8
Justification for this Study ........................................................................................................... 9
Research Questions ................................................................................................................ 10
Chapter 2 LITERATURE REVIEW .............................................................................................. 11
The Red Legged Frog: Summary of Biology and Ecology........................................................ 11
The Importance of Upland Habitat ........................................................................................ 14
Connectivity Definition and Overview ...................................................................................... 19
The Importance of Habitat Connectivity to Amphibians ........................................................... 20
Assessment of Connectivity: Formal Approaches ..................................................................... 23
Assessment of Connectivity: More Recent Approaches ............................................................ 23
Lowland Forest Connectivity in Western Washington .............................................................. 27
Landscape History: Pre-European and American Influences ............................................... 27
American Influences: 1850 to Present ................................................................................... 29
Chapter 3 Methods ......................................................................................................................... 33
Wetland Selection ...................................................................................................................... 33
Study Site Selection Criteria ...................................................................................................... 34
Final Selection of Sites .............................................................................................................. 37
Data Collection Methods ........................................................................................................... 42
Qualitative Landscape Characterization .................................................................................... 43
Quantitative Analysis ................................................................................................................. 44
Chapter 4 RESULTS...................................................................................................................... 45
Primary Forest Patch Size within 2 Km and Breeding Effort .................................................... 59
The Effect of Roads and Traffic on Breeding ............................................................................ 60
Chapter 5 DISCUSSION AND RECOMMENDATIONS ............................................................ 63
v

Part 1: Scientific Conclusions .................................................................................................... 63
Summary .................................................................................................................................... 63
Forest Patch Size and Breeding Effort ...................................................................................... 64
Breeding Effort and Nearby Busy Roads ................................................................................... 64
Access to Secondary Forest Patches.......................................................................................... 65
Edge Effects within Forest Patches ........................................................................................... 66
Position of the Wetland within the Primary Forest Patch ......................................................... 67
General Land Use Objectives of the Area Surrounding the Site ............................................... 67
Part 1 Conclusion ....................................................................................................................... 70
Part 2: Recommendations for Further Research ........................................................................ 71
Habitat Use ................................................................................................................................ 71
Conduct a Similar Study in the Cascades and Olympic Ranges ................................................ 72
Conduct Genetic Research to Determine Possible .................................................................... 72
Conduct Research on Road Impacts .......................................................................................... 72
Use Radio Telemetry to Research Habitat Preferences and Migration Distances .................... 73
Use GIS to Inventory Forest Patches in the Puget Sound Lowlands ......................................... 73
Part III: Management Recommendations .................................................................................. 73
Background ................................................................................................................................ 73
Conservation Recommendations ................................................................................................ 75
Literature Cited .......................................................................................................................... 81

v

List of Figures
Figure 1: Red legged frog (Rana aurora) identification. ...................................................................... 12
Figure 2 A breeding wetland in mid March, 2011................................................................................ 13
Figure 3 The same breeding wetland in late June, 2011. ..................................................................... 13
Figure 4 Rana Aurora egg mass .............................................................................................................. 14
Figure 5 Functional connectivity diagram ............................................................................................... 24
Figure 6 Locations of Selected Study Sites ........................................................................................... 39
Figure 7 Study Site 1, Eastern Thurston County ................................................................................. 46
Figure 8 Sites 2 and 3, in the Town of Rainier ..................................................................................... 47
Figure 9 Site 4, Northwest Whatcom County ....................................................................................... 48
Figure 10 Site 5, in the City of Puyallup ............................................................................................... 49
Figure 11 Site 6, Between Kent and Covington, King County ............................................................. 50
Figure 12 Site 7 North of Tenino, Thurston County ............................................................................ 51
Figure 13 Site 8 Northeast of Eatonville, Pierce County ...................................................................... 52
Figure 14 Site 9, Baker River area, North Cascades National Park .................................................... 53
Figure 15 Site 10 Rainier Training Area, Joint Base Lewis-McChord ................................................ 54
Figure 16 Site 11 Northwest of Eatonville, Pierce County ................................................................... 55
Figure 17 Site 12 Skagit River Valley, Skagit County .......................................................................... 56
Figure 18 Site 13, Rainier Training Area, Joint Base Lewis-McChord ............................................... 57
Figure 19 Site 14 Cedar River Watershed, central King County......................................................... 58
Figure 20 Rana Aurora Breeding Effort as a Function of Primary Forest Patch Size ......................... 60
Figure 21 Effect of Nearby Roads on Rana Aurora Breeding Effort ......................................................... 61

vi

List of Tables
Table 1 Rana aurora Predators....................................................................................................................4
Table 2 Alternate Hypothesis ..................................................................................................................... 10
Table 3 Descriptions of Selected Sites ........................................................................................................ 40
Table 4 Descriptions of Rejected Sites ....................................................................................................... 41
Table 5 Quantitative Results ...................................................................................................................... 59
Table 6 Effect of Traffic Levels on Breeding Effort: Qualitative Analysis ............................................ 62
Table 7 Rana aurora Friction Values for Different Land Covers ........................................................... 90

vii

Acknowledgements

I would like to thank and acknowledge my parents Robin and Annette Holcomb for their
patience and support and my reviewer, Dr. Gerardo Chin-Leo, for his patience and
optimistic encouragement. I would also like to thank Dr. Marc Hayes, for serving as an
invaluable resource on amphibian ecology. I would like to thank James Lynch and Scott
Richardson (both of Joint Base Lewis McChord) and Steve Walker of the Whatcom
County Land Trust for their suggestions on study sites. I would like to thank Ron Tressler
(Seattle City Light), Heidy Barnett and Sally Nickelson (both of Seattle Public Utilities)
for sharing data. I would like to thank Karin Grimland (Whatcom County Land Trust)
and Ronald Gay (USFS) for assisting with fieldwork. I would like to thank all of the
property owners who gave me permission to enter their land and look for frog eggs and
they include: the Anderson Family Trust, Kathleen O’Farrell, Gary Shroeder, Richard
Fischer, John and Olive Vincent, Mark and Twila Miller, the Kehoe family and the
Vekved family. I would like to thank taxpayers and administrators of public lands
including those of North Cascades National Park and the Washington Department of
Natural Resources. I would like to thank the Thurston County Land Trust and the
Whatcom County Land Trust, Seattle City Light, Joint Base Lewis-McChord, Wilcox
Farms and Manke Timber Corporation, all of whom provided me with access to study
sites.

viii

Chapter 1: INTRODUCTION
Overview
Amphibians occupy valuable niches in aquatic and terrestrial environments but are
decreasing in abundance throughout the world. According to Wells (2007), amphibians’ highly
permeable skin, small size, ectothermic metabolism, dependence on aquatic habitats and
dependence on interconnected habitats make them particularly susceptible to a variety of human
impacts. Some research has examined the impacts to many amphibian species over broad
landscapes (e.g. Rubbo and Kiesecker 2005; Skids et al.2007; Egan and Paton, 2008) while other
research has focused on very specific impacts to one or a few species (e.g. Chan-McLeod 2003;
Schuytema and Nebeker, 1999, Deguise and Richardson 2009). This research indicates that some
threats figure more prominently in the lives of each species, genus or order than other threats.
The Red Legged Frog (Rana aurora) formally the Northern Red Legged Frog (Rana
aurora aurora) is a medium sized frog that favors coniferous or mixed coniferous / deciduous
forests ranging from southwestern British Columbia to coastal Northern California (Jones et al
2005). R. aurora appears to be less tolerant of heavily urbanized areas than other endemic
amphibian species, most notably the Pacific Treefrog (Pseudacris regilla) (Nussbaum, 1983;
Richter et al. 2008). This study will examine how a group of closely related habitat connectivity
characteristics relate to populations of R. aurora. Such as the case with many frogs, Rana aurora
is much more adept at crossing less than ideal habitat areas than salamanders. Although it
migrates similar distances from its breeding grounds as the Western Toad (Anaxarus boreas) it
appears to be tolerating human land use changes better than this species (Adams et al. 1998).
Also, R. aurora is persisting much better in western Washington than its closely related cousin,
the Oregon Spotted Frog (Rana pretiosa) (Adams et al. 1999). This species was once wide1

spread in western Washington but currently is documented in only several breeding sites in the
state. Although R. aurora appears to be coping with human impacts better than these other native
amphibians, it is important to better understand its landscape habitat requirements and to consider
future impacts to this species in the face of the anticipated increases in human population and
development in western Washington.

Observed Declines and Official Listings
Researchers started realizing that R. aurora was not present in landscapes that are heavily
urbanized or devoted to agriculture in the early 1980’s. Allan D. St John conducted a series of
amphibian and reptile surveys throughout Oregon in the 1980s and observed that R. aurora were
not present in urban areas or expansive areas devoted to agriculture, even if wetlands were
present (St. John, 1982, 1984, 1985, 1987). In their 1983 field guide, Nussbaum and others stated
that the R. aurora ‘seem[ed] to be less common than it once was’ in Oregon’s Willamette Valley.
This area has been heavily devoted to agriculture, is occupied by Interstate 5 and has been
steadily increasing in human population for much of Oregon’s history (Bury, 2008). More recent
surveys have documented similar R. aurora declines (Blaustein and Wake, 1990; Jennings and
Hayes, 1994; COSEWIC 1 2006, 2012).
Because of observed declines, R. aurora has been regarded by 4 of the 6 governments
that are responsible for managing it as being comparatively abundant but necessary to monitor.
In California, it is considered a species of Special Concern (DFG, 2011). The Oregon
Department of Fish and Wildlife places R. aurora in its least concern category which is ‘SV’ for
‘sensitive vulnerable’ in the Willamette Valley area (ODFW, 2011). The Washington
Department of Fish and Wildlife does not grant Rana aurora a designation (WDFW, 2012). The

1

COSEWIC means ‘Committee on the Status of Endangered Wildlife in Canada’. This body publishes
reports assessing status on many species.
2

British Columbia Ministry of the Environment includes Rana aurora on its ‘Blue list’ 2 (BC
Ministry of the Environment, 2011). On the United States federal level, the US Fish and Wildlife
Service does not list Rana aurora as being ‘Endangered’, ‘Threatened’, Sensitive’ or ‘Candidate’
(USFWS, 2012). The Council on Sensitive and Endangered Wildlife in Canada (COSEWIC) has
designated it as a ‘Species of Special Concern’ (2006, 2012).

The Ecological Significance of Amphibians and R. aurora
Lentic breeding amphibians are important components of aquatic and terrestrial
ecosystems. Being generally high-fecundity animals, amphibians in isolated wetlands have been
shown to produce 1400kg of amphibian biomass in a breeding season (Gibbons, 2006). Tadpoles
significantly control algae and periphyton (Mallory and Richardson, 2005).Without tadpoles,
extreme algae growth can cause eutrophication, which reduces biodiversity (Bedford et al. 2001).
Tadpoles also serve as a food source for native fishes, other amphibians and certain insects
(Calef, 1973; Licht 1974) (See Table 1: Rana Aurora Predators). Once metamorphs develop into
frogs and leave the wetland, they transfer energy and nutrients from the aquatic habitat to the
terrestrial habitat (Register et al., 2005). Adult amphibians mainly feed on detritivorous insects
on the forest floor and therefore slow down rates vegetative decomposition (Davic and Welsh,
2004). Finally, amphibians are colorful and cryptic providing an aesthetic value and encouraging
people to connect with the natural environment.
Because of their life history, Rana aurora offer these services to a specific part of the
ecosystem at a specific time. R. aurora has one of the highest fecundities among local
amphibians, with each egg mass containing between 750- 2000 eggs (Jones et al., 2005). This
results in a large supply of tadpoles in the early spring which consume algae and periphyton and
constitute a significant food source for predators. Since R. aurora migrate comparatively far

2

A ‘Blue list’ species is defined as ‘at risk but not extirpated, endangered or threatened’
3

distances from the breeding area , they bring their ecosystem services (transfer of aquatic area
nutrients, food for larger animals and consumption of insects) to forests located far from aquatic
areas. While Pseudacris regilla can also be found on the forest floor, R. aurora are markedly
larger and therefore consume more insects and different species of insects.
Common Name

Scientific Name

Red Legged Frog Life Stage that it
Preys On

Northwestern Salamander

Ambystoma gracile

Tadpole

Bullfrog *

Rana catesbeiana

Adult

Giant waterbug

Belostomatidae spp.

Tadpole

Laval diving beetle

Dytiscidae spp.

Tadpole

Dragon and damselfly larvae

Odanata spp.

Tadpole

Giant diving beetle

Lethocerus americanus

Tadpole

Cutthroat trout

Salmo clarkia

tadpole, adult

Rainbow trout

Salmo gairdneri

tadpole, adult

Bluegill*

Lepomis macrochirus

Tadpole

Western Gartersnake

Thamnophis sirtalis

adult, eggs

Belted Kingfisher

Megaceryle alcyon

Adult

Raccoon

Procyon lotor

adult, eggs

Great Blue Heron

Ardea Herodias

Adult

Table 1 Rana aurora Predators

Human Impacts to Rana aurora
Studies done in the field and lab have shown that a variety of human activities impact R.
aurora. This is due to vulnerable amphibian physiology and its dependence on both aquatic and
adjoining upland habitats. Impacts can therefore be grouped into toxics, hydrological impacts,
disease, parasites, introduced species and habitat loss. Some factors may take a toll on a
population over time while others such as road building and land clearing carry immediate
impacts. Finally, some impacts are facilitated by others. Habitat fragmentation, for example, not

4

only renders habitat less accessible but facilitates the spread of introduced species that compete
and depredate the species in question.
Many studies on amphibian landscape impacts work with the concept of urbanization
which encapsulates several impacts. ‘Urbanization’ is an imprecise term but Marzluff (2008)
defines it as an increase in ‘cities, suburbs and their surrounding built areas’ and McDonnell and
Picket (1990) define ‘urban’ as an area with ‘high human population density coupled with
increased energy use and extensive alteration of the landscape’. In this thesis I will consider
urbanization to be a land development trend that includes both of these definitions. Landscapes
dedicated to agriculture and timber production will not be considered ‘urban’ while landscapes
devoted to other forms of commerce as well as housing and transportation will be considered
urban. Urbanization generally results in habitat loss, habitat fragmentation, hydrological impacts
to aquatic areas, the increased presence of toxics, increased noise and light pollution and the
spread of alien species (Mitchell and Brown, 2008).
Loss of aquatic habitat for breeding has occurred as Washington became industrialized
but this trend has been significantly slowed in the past 20 years. Lane and Taylor (1996) have
estimated that by 1988, 39% of Washington State’s wetland area had been eliminated. This trend
slowed around 1990 with the ‘No Net Loss’ doctrine which enforced the sections 301 (a) and 404
of the Clean Water Act more vigorously. Although a court battle eliminated hydrologically
isolated wetlands from the Clean Water Act protections, growth management regulations in
Washington and California in the early 1990s have been helpful in protecting isolated wetlands
(WDOE 2001; CSWRCB 2005) many of which are ideal R. aurora habitat. Isolated wetlands in
Oregon remain less protected since state growth management regulations were essentially
overturned. Nonetheless, R. aurora in western Washington continue to suffer from the legacy of
wetland loss in many areas that were developed first in the state. These areas generally include
river valleys, deltas, the eastern margin of Puget Sound and areas that were first settled and
5

dedicated to commerce and industry. Many of these areas are now intensively urbanized or
devoted to agriculture (van Stavaren et al 2006).
An increase in impermeable surfaces over the landscape has been shown to affect
hydrology in ways that are adverse to R. aurora. Increased impermeable surface area in the
surrounding landscape leads to more pronounced changes in water levels and more permanent
inundation of aquatic areas (Holland et al.,1995; Thom et al., 2001;Azous and Horner, 2001;
Kentula et al. 2004). Rapid decreases in water level have been shown to strand R. aurora and
Ambystoma gracile (Northwest Salamander) egg masses above the water level (Klaus Richter,
pers. observation). This stranding can expose egg masses to freezing or desiccation.
Additionally, permanent inundation facilitates predatory fish and introduced frog (American
Bullfrog, Lithobates catesbeinus, and Green Frog, Lithobates clamatans) populations (Adams
1999; Ostergaard, 2001). It also facilitates the development of shrub-dominated communities that
are not as amenable to R. aurora breeding as wetlands that are dominated by emergent vegetation
or shallow open water (Reinelt et al. 1998). For these reasons, urbanization in an area has the
potential make the area’s emergent wetlands less suitable even adjacent forests are also preserved.
Since amphibians have semipermeable skin and are associated with low-lying aquatic
habitats that drain wide areas, they are impacted by toxic substances that are applied over the
surrounding landscape. Substances originating from a host of sources have been proven to impact
R. aurora or its close cousin the California red legged frog (Rana draytonii). It is important to
remember that fertilizers and biocides are used for commercial and residential landscaping in
addition to agriculture and timber production, thus making these substances wide-spread
throughout the R. aurora range. Laboratory experiments have revealed that R. aurora embryos
can be negatively affected by even small amounts of ammonium sulfate (NH3SO4) and
ammonium nitrate (NH3NO3), which are common components of fertilizers (Schuytema and
Nebecker 1999, 2000). They are sensitive to doses that are much lower than are commonly
6

applied (Marco et al., 1990). Few field studies have analyzed the affects of biocides 3 but Hayes
and others (2008) contend that these substances could pose a problem given their ubiquity in
many parts of the R. aurora range. Various biocides have been implicated as factors endangering
Rana draytonii (Davidson et al. 2001) and this could be a harbinger for R. aurora.
A wide variety of industrial and consumer products contain endocrine disrupting
compounds which cause male frogs to develop female characteristics. These compounds are a
particularly large concern because even small dosages of them can adversely affect amphibians.
Bettaso and others (2002) documented the presence of a biomarker in male R. aurora at several
northwestern California sites that indicated that they had been exposed to endocrine disrupters.
This finding suggests that populations in even rural areas throughout the range are being exposed
to endocrine disrupters.
Scientists have long suspected that expanding Lithobates catesbeinus populations have
been a factor in native amphibian declines but more study is required and to date no actual
evidence for this has been documented (Hayes et al. 2008). Part of this is due to the fact that it is
difficult to select sites to experiment with bullfrogs since other habitat-related factors come into
play (Hayes et al., 2008). By conducting field experiments, Kieseker and Blaustein (1998, 1999)
have found that when both R. aurora and Lithobates catesbeinus occupy the same habitat, R.
aurora are seemingly forced into deeper habitat that is less optimal for them. Cook and Jennings
(2007) point out that Rana draytonii breeds about 2.5 months earlier than Lithobates catesbeinus
so presence of large adult populations or developing larvae do not overlap. Since R. aurora
breeds at the same time of year, these results could plausibly be extended to it. This collective
research suggests that adults of these two species may compete for resources to R. aurora’s
detriment but that more research is required to determine if Lithobates catesbeinus is significantly
impacting R. aurora populations.
3

The term ‘biocides’ includes insecticides, fungicides and herbicides.
7

Fish have been shown to impact R. aurora populations by predation and or working in
concert with bullfrogs. Kiesecker and Blaustein (1998) demonstrated that if smallmouth bass
(Microperus dolomieui) are present in water bodies with Lithobates catesbeinus larvae R. aurora
growth and survivorship was negatively affected, possibly because R. aurora are forced into
deeper water with more fish. Trout (Oncorehynchus spp.) have been shown to prey on native
amphibians (McGarvie-Hirner and Cox, 2007). Bluegill (Leponis macrochirus) encourage
Lithobates catesbeinus survival (Adams et al., 2003). While introduced fish negatively affect
native amphibians, it should be noted that fish depend on areas with permanent inundation which
are only one type of aquatic area that R. aurora utilize for breeding. At present, introduced warm
water fish appear to be a greater threat to R. aurora than bullfrogs and greenfrogs. This threat is
more significant in areas undergoing increased urbanization since urbanization leads to more
permanent hydroperiods (Holland et al, 1995; Thom et al, 2001).
Habitat and Connectivity

In their assessment of all of the threats confronting R. aurora, Hayes and others (2008)
contend that loss and fragmentation of terrestrial habitat may be the greatest threat to the species.
When they are away from breeding habitat, R. auroras utilize forest landscapes almost
exclusively (Haggard, 2000; Chan-McLeod, 2003; Jones et al., 2005). In addition, R. aurora has
been shown to migrate as far as 4.8 km from breeding areas (Hayes, 2004), meaning that the
species may require extensive connectivity more than other native amphibians. Some studies
have found a positive relationship between R. aurora abundance in aquatic areas and the amount
of forest cover within 1 or 2km (Richter and Azous, 2001; Ostergaard 2001; Ostergaard et al.
2008). While patches of forests may exist near aquatic areas, they are of little use to amphibians
if they cannot be reached or are particularly difficult to reach by ranid frogs 4 (Fahrig 1997;

4

‘Ranid frogs’ are all frogs belonging to the genus Rana, or ‘true frogs’. This is a world wide genus of frogs
and conclusions about the biology and ecology of many of them can be extended to Rana aurora.
8

Stevens and Baguette, 2008). Roads are particularly treacherous barriers and their capacity to
fragment habitat increases with the level of traffic on them (Gibbs 1998; Cushman 2006;
Eigenbrod et al., 2007). They have also been shown to kill large numbers of migrating R. aurora
in one study (Beasely, 2002). The concepts of ‘functional connectivity’ and ‘landscape
complementation’ deal with the degree to which a given species can utilize the broader landscape,
when the species’ habitat requirements and the landscape’s fragmentation are considered (Crooks,
2007). Mathias (2008) used GIS friction analysis of land cover maps to assess functional
connectivity for R. aurora in King County, Washington. Although she did not incorporate field
data, she assessed the landscape based on R. aurora’s ability and the risk the species incurred to
cross most land covers. She found that the more urbanized western part of the county was less
connected than the central part of the county. Her research however, did not take into account
actual abundance from field data.

Justification for this Study
Although R. aurora populations can be observed in suburban and exurban areas, the fact
that it cannot be found in more intensely urban areas suggest that this species has limits as to how
much human development it can tolerate. The species has been clearly decimated in the urban
core of not only large cities but smaller towns. Pseudacris regilla, by contrast, is frequently
observed in such areas. R. aurora is closely associated with forest habitats but the expansive
forests that have covered the species’ habitat for most of its history are no longer extant. R.
aurora is now living in landscapes that are covered with a patchwork of forest (of varying age
classes), pastures, clear cuts, residential development, business districts and roads. Forest patches
vary with respect to size, connectivity to breeding habitat and connectivity to other forest patches.
Despite its relatively high ability to cross sub-optimal habitat and utilize forests patches
across the landscape, local scientists have expressed concern that R. aurora will decline as

9

western Washington increases in population over the upcoming decades (Shuett-Hames et al,
2007; Hayes et al. 2008). Hayes and others (2008) have expressed a need to examine the
importance of habitat connectivity for R. aurora. Using breeding effort as an index of population
size, this research analyzes how the size of the forest patch adjacent to the breeding area and the
proximity of busy roads to the breeding area affect R. aurora breeding effort. Egg mass censuses
were taken on thirty wetlands but many of these were thrown out of the study because it was
believed that other factors besides surrounding connectivity were affecting populations. This
resulted in a sample size of 14 selected wetlands that reflected a range of connectivity to forest
within 2 kilometers. The study was guided by 4 research questions. From these questions I have
developed a series of alternate hypothesis (Table 2).
Research Questions

1). How does the size of the immediate forest patch affect breeding effort?
2). How does the presence of roads within .25 km of the wetland affect breeding effort?
3). How does the level of traffic on nearby roads affect breeding effort?
4). How does connectivity between the immediate patch and the neighboring patches affect
breeding effort?
Alternate Hypotheses
Breeding effort will be positively related to the size of the immediate patch because
larger patches represent a larger degree of continuous ideal habitat.
Breeding effort will be negatively related to the presence of roads within .25 km of the
wetland edge.
Breeding effort will be inversely related to traffic intensity on neighboring roads.
Breeding effort will be positively related to the ease at which frogs can travel between
the immediate patch and neighboring forest patches.
Table 2 Alternate Hypothesis

10

Chapter 2 LITERATURE REVIEW
In order to learn about the accrued scientific knowledge on R. aurora life history, ecology
and the threats relating to the species, I conducted searches for peer-reviewed journal articles on
data bases specializing in ecology, zoology and biology. I used key words such as ‘red legged
frog’, ‘Rana aurora’, ‘amphibians, ‘habitat connectivity’, ‘urbanization’ and ‘dispersal’ and
‘migration’ to select articles that had these topics in their abstracts. I reviewed recent books on
amphibians. I surveyed thesis work that had been done at universities in the R. aurora range. I
surveyed unpublished US Forest Service research on R. aurora and amphibians. Finally, I
surveyed the bibliographies of some of these written works to gather other relevant sources.

The Red Legged Frog: Summary of Biology and Ecology
R. aurora is a medium-sized (50-100 SVL 5), lentic (still water) breeding frog. Most
individuals have a dark patch around an eye with brown irises and a red groin patch. The back
can be tan, brown or reddish-brown and spots are usually present (Fig. 1)(Jones et al., 2005). The
R. aurora range extends from Mendocino County, California in the south northward through all
of Vancouver Island to the Margaret Bay area of British Columbia, Canada. In California, the
range is close to the coast but in Oregon and Washington it extends to mid elevations (up to 365
m or 1200 ft) of the Cascades. In mainland British Columbia it extends inland from the Straights
of Georgia roughly 200 miles in the south to about 100 miles in the Margaret Bay area. The
species is absent from the higher elevations of the Olympic Mountains (Jones et al., 2005; Pearl,
2005). Because of urbanization, it is absent from heavily urbanized areas on the east shore of the
Puget Sound, the Portland, Oregon metropolitan area and the metropolitan areas in southwest
British Columbia (St John 1982, 1984, 1985, 1987; Nussbaum et al. 1983; Jennings and Hayes
1994; COSIWIC 2011).
5

SVL means ‘snout vent length’. It is the length from the animals snout to its vent (anus).
11

Figure 1: Red legged frog (Rana aurora) identification.
The coloring is variable throughout the range. Reddish markings on the ventral side of the legs are
usually diagnostic. LEFT: specimen from Multnomah Co. Oregon, CENTER: Humboldt Co.
California, RIGHT: ventral view of adult. All photos by Gary Nafis, californiaherps.org.

Breeding and Larval Development
Rana aurora generally gather en masse at the same aquatic area to breed every winter,
commonly in February (Licht 1969). Small wetlands with semi-permanent inundation and ample
amounts of emergent or aquatic-bed vegetation constitute ideal breeding habitat (Pearl et al.,
2005) but Rana aurora also breed in lakes and slow-flowing water (under 5cm/second) (Klaus
Richter, personal obs.). Prolific breeding does not occur in forested or shrub dominated wetlands,
most likely due to lack of sunlight and nutrients (Shelley 2002; Shelley and Golon 2003). Figures
2 and 3 show a wetland that typifies ideal breeding habitat.

12

Figure 2 A breeding wetland in mid March,

Figure 3 The same breeding wetland in late June,
2011.

2011

Photos above are the same view of Site 11. Figure 2 shows the area inundated to 60 cm and Figure 3
shows the same area inundated to 38cm several months later. This is ideal R. Aurora breeding
habitat; it is interspersed with common cattail (Typha latifolia) in the foreground and background,
slough sedge (Carex obnupta) in the foreground, and Douglas spiraea (Spiraea douglasii) which is in
the background and reddish-colored in the spring. The National Wetland Inventory classifies this as
a palustrine emergent wetland with seasonal inundation (PEMC). 278 egg masses were observed
here. Photo by Chris Holcomb

Male frogs arrive at breeding areas first. Populations at lower elevations and at lower
latitudes tend to breed earlier, probably due to temperature. Storm (1960) observed that frogs in
the Corvalis, Oregon area arrived at breeding sites on December 8 while Licht (1969) observed
that frogs in British Columbia did not arrive at breeding sites until February or March when air
temperatures reached 10°C. In conducting fieldwork for this thesis, I observed that oviposition
had started earlier at sites in west Pierce County, Washington than at higher elevation sites in the

13

central part of the county. Once couples form, the
male and female undergo amplexus and the female
then deposits a globular mass of 530-830 eggs. The
capsules around each egg quickly absorb water
causing the mass to have a jelly-like consistency
and to grow to the size of a large cantelope (Figure

4). Egg masses are attached to aquatic vegetation,

Figure 4 Rana Aurora egg mass

usually in water that is 48 to 70 cm deep (Storm,
1960; Licht, 1969; Calef 1973). I observed that masses are generally in the upper 36 cm of the
water column. Breeding activity is often concentrated in the northern part of t he aquatic area,
probably because this area has the most sunlight exposure.
Embryos develop over the course of 10-30 days. As the embryonic stage progresses, the
egg mass becomes less spherical and becomes laced with algae and sediment (Figure 4, photo on
right). Once the tadpoles hatch they tend to stay on or near the egg mass for a short time.
Tadpoles reach metamorphosis 11-14 weeks after hatching. Juvenile frogs tend to stay in the
wetland anywhere between 2 weeks and 2 months after they reach metamorphosis (Storm 1960;
Licht 1974; Brown 1975).
The Importance of Upland Habitat

Adult R. aurora usually leave the breeding area in spring and spend a solitary life in
uplands. They have been shown to select forested areas when leaving breeding ponds
(Rothermel, 2004) and are found in greater abundance in forests (Haggard 2000; Aubry 2000;
Chan-McCleod 2003). However, they are capable of crossing more open habitats like clear cuts
(Chan-Mcleod 2003; Chan-McCleod and Moy, 2006) and roads (Beasely, 2002). They have also
been seen in low-density residential areas (Holcomb, personal obs.) if such areas are small
14

enough and include such features as wetlands, ditches, and forest patches. Studies in different
parts of the range suggest a variety of seasonal travel distances. Hayes (2007) has demonstrated
that R. aurora can travel up to 4.8 km from the breeding wetland in the central Oregon Cascades,
which is comparatively far for many local amphibians. Haggard (2000) found that frogs only
moved 80 m at her study site on the northern California coast. Most researchers feel that between
2 and 3 kilometers is an average one-way migration distance for R. aurora (Mathias, 2008).
Semlitsch (2008) has stated particularly far movements (such as Hayes’s 4.8 km observation)
likely represent extreme distances that are undertaken by very few individuals. Semlitsch (2008)
has stated that for all lentic-breeding amphibians, dispersal and migration are different processes
that are done at different times of the life history. He defines migration as seasonal movements
generally by adults from breeding areas into adjacent upland which are followed by returns to the
same aquatic area to breed. Dispersal is often undertaken by juvenile frogs and constitutes
movements from their breeding area over the upland to new breeding areas (Rothermel, 2004).
These dispersal movements may be done over the course of two or three years until the animal is
sexually mature. Since radio telemetry techniques can only be used on adult frogs, there is sparse
information on R. aurora juvenile movements in uplands, but they are presumed to serve a large
role in dispersal, as Rothermel (2004) has described.
Due to the expense and challenges of radio telemetry work on small animals, movement
data is sparse. However, radio telemetry research has been undertaken in a wide range of habitats
and places within the R.aurora range and it is probable that average dispersing and migrating
distances vary with habitat and location. Haggard (2001) analyzed movement from breeding
ponds on the northern California coast. Hayes et al (2001, 2007) analyzed the Umpqua Basin of
Oregon. Serra-Shean (2001) analyzed movement out of a large wetland in western Washington.
Semlitsch (2008) has theorized that ranid frogs migrate by making sustained trips, triggered by
nocturnal rainfall, before they stop in an area and remain comparatively sedentary for long

15

periods. Several bodies of research suggest that frogs move under 10m a day at times (Haggard
2000; Ritson and Hayes 2000; Schuett-Hames 2004). These shorter movements may take place
after periods of far sustained movement when frogs have found a good place to forage. SchuettHames (2004) employed video recording to document frog behavior and determined that adults
spend long periods of time under complex understory feeding. Such behavior likely enables them
to stay concealed from predators, conserve energy and water and build up energy reserves for
later travel. Shuett-Hames’s observations possibly describe Semlitsch’s idea of frogs remaining
relatively stationary for periods lasting months after periods of sustained travel. How R. aurora
overwinter is one of the least understood aspects of their life history (Hayes et al. 2008). Post
metamorphic individuals have been observed to spend the winter in breeding ponds (Ritson and
Hayes 2000) but it is believed that the majority of adults overwinter in uplands.
The collective research over the past 35 years has given us a moderately-clear picture of
the types of forest habitats and features favorable to R. aurora. There is little information on
amphibian use of the extensive old-growth forests that predated American influence (Mathias,
2008) but there has been some research on old growth patches that currently exist (eg. Gilbert and
Allawine, 1991). R. aurora favor mature forests that have at least some understory and are either
dominated by conifers, deciduous trees or are mixed (Aubry, 2000; Haggard 2000; SchuettHames, 2004; Gomez and Anthony, 1996).
The US Forest Service examined Pacific Northwest native forest amphibian communities
in the 1980s over three different study areas of western Washington and Oregon. This research by
Gillbert and Allawine, 1991, Aubry and Hall, 1991, Bury et al 1991 stated the importance of large
woody debris, recognizing unique and botanically diverse microhabitats, and the proximity of
aquatic areas. This research however, did not consider clear cuts or aspects related to
connectivity; it only studied habitat aspects in unmanaged Douglas fir forests.

16

Keith Aubry conducted research on amphibian presence in managed forests in the early
1990’s and considered additional aspects of forest structure. The study took place on private
forest lands southeast of Eatonville, Washington in Central Pierce County, near the edge of the R.
aurora range. Unlike the aforementioned USFS work, this research included an analysis of
amphibian use of clear cuts. Aubry also analyzed second growth forests that were dominated by
Douglas fir but included other conifers and broadleaf trees. R. aurora were most abundant in the
oldest age class in which they comprised 5.3% of all amphibian captures. Far fewer R. aurora
were captured in the clear cut plots and the pre canopy plots and none were caught in the closed
canopy plots. In addition, the Aubry study concluded that R. aurora abundance was positively
associated with leaf litter depth and the abundance of shrubbery and negatively associated with
elevation and cover of exposed rock. The elevation relationship is not surprising given that R.
aurora do not inhabit areas above 1200 m (Pearl 2005) and parts of the study area were close to
that elevation.
Martin and McComb (2003) analyzed amphibian associations in second growth
patchwork landscapes that typify commercial logging areas in Oregon’s Coast Range. The study
area was influenced by an expansive wildfire in the mid 1800’s but had been used for commercial
timber production for 40 years prior to the study. The forests in the study were dominated by
Douglas fir but also included Sitka spruce (Picea sitchensis), western hemlock (Tsuga
heterophyla), red alder (Alnus rubra) and big leaf maple (Acer macrophyllum). They delimited
13 forest types based on tree age, level of tree type dominance (deciduous or conifer) and amount
of canopy closure and concluded that R. aurora prefer ‘mixed, large sawtimber’. This forest type
is defined as being: ‘<70% conifer or hardwood composition, > 20% cover, > 53.3 cm dbh’ 6
(Martin and McComb 2003). Gomez and Anthony (1996) conducted a similar study but with
only 5 forest types in Oregon and concluded that R. aurora were more abundant in deciduous

6

‘dbh’ means diameter at breast height
17

forests. This is probably due to the fact that red alder forests produce substrates with more
nutrient levels and hence more invertebrate prey species for amphibians (Shirley, 2004).
The most recent research recognizes how far R.aurora travel and the realities of the
silviculture landscape. This leads to the question, under what conditions will R. aurora cross
clear cuts and can small residual forest patches facilitate migrations? Chan-McCleod has
researched R. aurora movement in fragmented forest landscapes in British Columbia and came to
similar conclusions that Schuett-Hames (2004) did: frogs do travel through open habitat but seem
to prefer forests. She concluded that clear cuts under 12 years old pose significant barriers to
R.aurora movement but those frogs will be more likely to enter and move through them under
certain conditions. The study was undertaken from August through October when both rain and
high temperatures are extant. Compared with forest habitats, frogs permeated clear cuts at a rate
of 16.7% when rain was absent but temperatures and humidity measurements were at their
average level during the trial period. However, under the maximum observed noon temperature,
the rate of entry into clear cuts dropped to 2.3%. Additionally, streams 3 m wide seemed to
encourage entry into clear cuts while streams under 1.5 meters did not significantly affect this
(Chan-McCleod, 2003). In another study that evaluated R. aurora use of residual tree patches left
in clear cuts, Chan-McCleod and Moy (2006) determined that: 1) when travelling through clear
cuts, frogs intercepted patches largely by chance and were not likely to gravitate toward such
patches unless they were 5-20 meters from them and 2) frogs tended to select patches that were
over .8 ha in area.
To conclude, in the second growth forest landscape, R. aurora appear to select mature
forests that are either dominated by deciduous trees or are a mixture of deciduous trees and
conifers with complex understory. They will move through clear cuts but this behavior is
facilitated by rainfall and cooler temperatures and is more often undertaken by larger individuals
that can more easily withstand environmental pressures. R. aurora will also utilize small patches
18

of forest but do not appear to seek out such areas unless they are within 20 meter of them and
they are larger- over .8 ha in area. Given their physiological constraints, clear cuts and open areas
pose significant challenges to R. aurora and highlight problems with habitat fragmentation.

Connectivity Definition and Overview
‘Connectivity’pertains to the geographic size of habitats and the magnitude and nature to
which they are linked to other habitats (Groom, 2008). Sanjayan (2007) states that connectivity is
related to the degree of movement of organisms and processes. Talley and others (2007) also
provide a broad definition, stating that connectivity is not just about animals going across the
landscape spreading genes, it relates to material and energy moving across landscapes.
Adriaensen and others (2003) state that the inverse of habitat connectivity is ‘landscape
resistance’ or ‘isolation’. ‘Fragmentation’ is the process of separating contiguous expanses of
habitats into disparate parts. Fragmentation makes it more difficult for organisms to utilize the
entire habitat that was originally available to them. Fahrig (2003) emphasizes that the concepts of
fragmentation and habitat loss should be separated. She states that while fragmentation in and of
itself results of loss in habitat, it also renders existing habitat blocks less accessible to organisms
and thus has unique effects on the species in question.
While some organisms may persist in these patches of habitat soon after they are
fragmented, they may suffer ill effects over time. Some reasons for this are that other prefragmentation components of the ecosystem may disappear while native organisms that are better
adapted for the new landscape may proliferate. Invasive species may enter the system. If native
organisms are not as adept at leaving the patch, they may suffer the effects of a limited gene pool.
With respect to lentic breeding amphibians, two types of connectivity are important; the
first type of connectivity is landscape complementation, or the arrangement of two important but
different habitat types, specifically breeding habitat and upland habitat (Dunning et al., 1992).
19

This link is important because both habitat types are essential to their life cycle and animals
transfer materials between the two habitats and (Talley et al. 2006; Kupferberg, 1997; Anderson
et al, 1991). If the landscape is too fragmented, this transfer cannot take place. The second type
of connectivity relates to connections with other breeding areas and amounts of contiguous
upland habitat for wide ranging animals to utilize. This thesis generally deals with this second
type of connectivity. The almost universal assumption underlying discussions of connectivity is
that habitats were well connected prior to the very recent influences related to human agriculture,
industrialization and urbanization.

The Importance of Habitat Connectivity to Amphibians
Connectivity, including links from the aquatic area to key upland habitats as well as
linkages between such habitats enables amphibians to utilize upland habitat. This enables them to
take advantage of food and cover resources that uplands have an abundance of. In turn
connectivity for amphibians makes it possible for upland habitats and human communities to be
shaped by amphibian ecosystem services. Amphibians are a significant consumer of forest floor
invertebrates as well as being a significant food source for larger carnivores (Wells, 2005).
Amphibian populations in individual aquatic areas periodically crash and are dependent
on being ‘rescued’ by colonization from the broader ‘metapopulation’. Such crashes occur due to
insufficient reproduction and immigration, habitat succession, the proliferation of a predator and
long term drought. Metapopulations are comprised of many separate populations, each breeding
in its own pond year after year, but that are each close enough to be contacted by individuals from
other populations (Marsh and Trenham, 2001). Metapopulations are identified by genetic
analysis (Marc Hayes, personal communication). Metapopulation theory originated from the
assumption that all lentic breeding amphibian populations were highly philopatric and did not
disperse far. Smith and Green (2005) questioned this, noting that many species are less

20

philopatric than previously thought, can travel further than previously thought and that females
may be more selective in choosing breeding grounds.
Aquatic areas need to be close enough to be occasionally reached by dispersing juveniles
to be part of metapopulations. In addition, sufficient levels of habitat connectivity need to be in
place to make colonization possible and for this reason researchers concerned with amphibian
conservation have devoted time to metapopulation studies in recent years (Trenham and Shaffer
2005, Trenham et al. 2003, Trenham, 1998, Skelly and Meir 1997, Driscoll 1997).
Metapopulation studies on amphibians consistently suggest that the more common and welldistributed breeding wetlands are throughout the landscape, the higher the probability that
turnovers can be prevented. Constant colonization and population can be restored in the event of
a population die off in any given wetland (Trenham et al. 2003).
Whether or not R. aurora have a metapopulation structure is unclear (Hayes et al., 2008).
If breeding areas are within 500 m of each other, they may have a more patchy population
structure as described by Petranka and Hollbrook (2006). In addition, Hayes and others (2008)
suggest that the species may be able to survive population crashes because it has relatively high
fecundity and is long lived (8 to 12 years). This means that populations at breeding sites could
eventually make up for bad reproductive years. However, Hayes and others also speculate that R.
aurora populations may function as metapopulations due to their far migration tendencies.
Whether or not R. aurora have a metapopulation structure, more of a patchy population structure
in certain areas, or can overcome occasional population crashes due to lifespan and fecundity is
unknown. Nonetheless, in management efforts it is probably wise to consider how well members
of one population can contact those of another population.
In contrast with more sedentary lungless salamander species, ranid frogs tend to be more
susceptible to negative genetic effects if different populations cannot occasionally exchange

21

genetic material (Wells, 2005). In addressing genetic issues for all animals, Frankam (2006)
listed seven factors that determine the susceptibility that a population of one species may have to
negative genetic effects. These include 1) the number of population fragments 7, 2) the
geographic distribution of the population fragments, 3) the dispersal ability of the species, 4)
migration rates between fragments, 5) degree of connectivity between fragments, 6) the time (in
generations of the species) that the fragmentation took place and 6) the susceptibility of the
species to inbreeding depression. Many of these points are relevant to ranid frogs since they are
small, slow moving animals with narrow habitat requirements.
No research has investigated how R.aurora genetics has been affected by human caused
habitat fragmentation but the species is persisting in many areas that have had low surrounding
functional connectivity for decades. This suggests that a species’ ability to cross adverse land
covers does not particularly give it an advantage at exchanging genes with other populations.
Studies on other ranid frog species suggest that human disturbance may already be affecting these
isolated R. aurora populations. Reh and Seitz (1990) and Hitchings and Beebee (1997) found
significant differences in genetic differentiation with increasing pond distance, suggesting that
isolation will not only prevent a rescue of a crashed population but eventual inbreeding
depression. The Reh and Seitz study as well as 2 other studies in Europe on ranid frog species
found that roads, railways and urbanization caused increased genetic difference or distinct genetic
groups (Vos et al. 2001; Sefner et al. 2011). Metapopulation studies at sites with low human
fragmentation and habitat destruction between other had little genetic difference between sites (eg
Gill 1978; Berven 1995; Trenham 1998; Seppa and Laurila 1999; Skelly et al. 1999). These
results suggest roads, railways and cities can result in greater genetic difference among their
populations of Rana aurora and possibly cause inbreeding depression within these populations.

7

In the case of lentic breeding amphibians, a ‘population fragment’ would mean an aquatic area with a
population of a species. Large aquatic areas with separated habitat areas may have a distinct population
fragment in each habitat area.
22

Assessment of Connectivity: Formal Approaches
In many amphibian environments, including R. aurora’s Puget Sound basin, vast
contiguous forest habitat areas are no longer in existence and animals are often left to travel
between patches of suitable habitat via territory that is less amenable to amphibian’s survival.
Originally, connectivity was assessed based on the size, shape and arrangement of patches of
ideal habitat. Since ecologists had some idea of the distances that different species traveled every
year, studies evaluated the size and distances of patches of ideal habitat in the landscape that fell
under a given distance from breading ponds (e.g. 1km, 2km, 3km).
While large expanses of ideal habitat were naturally deemed the most optimal levels of
connectivity, the size of patches and their arrangement were also considered. Generally this was
done by incorporating circular buffer functions in GIS programs with the breeding pond at the
center of the circle. Additionally ‘Nearest neighbor’ functions in GIS programs considered the
distance of patches of ideal habitat from each other. Prugh (2009) determined that nearest
neighbor patches were particularly poor predictors of abundance and occupancy by a target
species and that buffer functions were not much better. This is due to the fact that land covers
between patches were not considered. In recent years, connectivity studies have been divided into
two categories. Physical or structural connectivity considers and arrangement of habitat and all
other land cover types. Functional connectivity considers how a species behaves in all land
covers in addition to its ideal habitat.

Assessment of Connectivity: More Recent Approaches
Recently more connectivity studies have adapted new approaches that address other land
covers in the species migration and dispersal zone in addition to ideal habitat. Functional
connectivity is the degree to which a landscape can be crossed by an individual of a given species
and is based on behavior. It is largely based on the behavior that animals exhibit on different land
23

covers (Stevens et al, 2006, With et al. 1999, Goodwin and Fahrig, 2002) and is comprised of 2
components. Patch resistance is the level of difficulty that an area poses for a given species to
cross. Boundary permeability is the degree to which one habitat type can be crossed by a given
species (Stamps et al. 1987, Wiens et al. 1997). The functional connectivity approach has a
distinct advantage over earlier approaches to assessing connectivity in the sense that it more
accurately embraces the realities of human-impacted landscapes. Even frogs -- small, slowmoving ectotherms-- will cross adverse landscapes to utilize more ideal habitat types. Figure 4
illustrates an example. If a relatively far-dispersing animal like R. aurora wants to reach other
patches of forest beyond the one that surrounds its breeding pond, a road or heavily urbanized
area separating the ‘initial patch’ from, say, ‘patch A’ will be more consequential than ruralresidential land with lightly traveled roads, even if Patch B on the other side of such land is
farther than Patch A.

Patch A

Town

Small City

Patch B
Initial Patch

Harder Travel

Easier Travel

Key
Cropland

Urban Area

Highway

Smaller Road

Figure 5 Functional connectivity diagram

24

Functional connectivity of a landscape for a given species is assessed by creating ‘permeability
models’ using GIS applications. Ray and others (2002) employed this method to assess habitat
connectivity for two lentic breeding amphibians in Switzerland. This research comprehensively
describes the process, has served as a model for subsequent work and is the source for
information in this paragraph. ESRI’s software (eg. ArcMap, ArcView, ArcInfo) is the most
commonly used GIS tool. In this software, aerial photographs are used to construct a grid-based
landscape layer which codes different land covers. Grids are constructed in GIS using the raster
format which characterizes landscapes in square-shaped cells instead of lines and polygons. The
resolution (cell size) of raster-based values’ are then assigned to each cell based on land cover
type. The friction values indicate both the risk of mortality and energy expenditure that a given
species incurs for crossing a particular land cover and are based on previous studies of the species
and similar species as well as professional judgment. For forest amphibians, such as R. aurora,
more open artificial habitats lead to water loss or changes in optimal body temperature and may
involve more dangers such as maps can be increased or decreased depending on objectives.
Numbers or ‘friction cars or a higher chance of predation due to lower cover. Within these areas,
animals may move more quickly in order to reach a more amenable land cover. Frogs can often
be observed quickly crossing roads by hopping. These physiological effects and behavioral
responses to increased threats in areas devoid of forest cover translate into higher energy
expenditures. Once land cover and friction layers are created, the ‘cost distance function’ (in
ArcMap) is employed to calculate the ‘maximum cost of migration’ (MCM) for crossing each cell
in the model. This is done by multiplying the friction value of a land cover by the ‘maximum
distance of migration’ (MDM). The MDM is considered to be the distance in meters that the
species generally travels in ideal circumstances without habitat fragmentation and thus
characterizes the species inherent migration or dispersal tendencies. Here is a summary for
determining the MCM:

25

MCM = MDM * (friction value of land cover type)
For studies of lentic-breeding amphibians, the cost distance function can be centered
around each breeding pond in the landscape and set to calculate the MCM of a species moving
away from the breeding pond in all directions, thus simulating dispersal or seasonal migration.
As a virtual animal moves over cells within this GIS model, the friction value assigned to each
cell is subtracted from the animals MCM. The migration ends when the virtual animal has lost all
its energy equal to its MCM. The collective MCM for each pond is then averaged: higher
percentages of ideal habitat connected to the pond would therefore result in greater average MCM
for a species leaving the pond. Ray and others then used ‘generalized additive models’ to
measure relationships between each species to land covers. These models were used because of
their ability to handle non-linear relationships between dependent and independent variables.
Mathias (2008) applied the methods outlined by Ray and others (2002) to model
functional connectivity for R. aurora in King County, Washington. The Mathias study is
particularly valuable not only because it was the first functional connectivity study for R. aurora
but because central and western King County is one of the most urbanized areas in the state and
this development trend is expected to continue. Consulting regional amphibian experts, namely
Marc Hayes, Joanne Schuett-Hames, Klaus Richter and Ken Jacobsen, Mathias developed friction
values for the western Washington landscape. These are given in Appendix A.
Mathias produced a landscape map for central and western King County at 30 m
resolution (one cell representing 90 sq meters of land). Using this landscape map, she
incorporated the least cost function to create friction maps that assumed a 1000m MDM and a
3000m MDM. This was done because research has shown that R. aurora generally travel
between 1000 m and 3000m. Mathias also created 2m (one cell representing 4 sq meters of
land) resolution landscape and friction maps for the Bear Creek basin in northwest king county,

26

also illustrating functional connectivity under 1000m and 3000m scenarios. She did not
incorporate data from real frog movements but general patterns on land cover and connectivity
were obvious.
Mathias’ work elucidates not only on R. aurora habitat connectivity but also on least cost
modeling. First, she found that western King County—characterized by Seattle, Bellevue,
Renton and Interstate 5-- is significantly less connected than central King County which is more
rural. Secondly, connectivity correlated strongly with mature forests and lower road density.
Finally, the landscape was significantly more connected for R. aurora if the 3000 m MDM is
assumed for both the broad part of the county and the Bear Creek Basin. In other words, if we
assume that R. aurora typically migrates 3000 m one way each year, it is better able to handle the
adverse affects of urbanization because it is more likely to encounter new breeding ponds and
utilize other forest patches. The finer resolution of the Bear Creek Basin maps revealed less
connectivity because it is able to pick up roads which are significant barriers. However, the lower
resolution maps of broader geographical areas are valuable because they reveal connectivity
patterns over broad regions. Additionally, a portion of animals do cross roads so these lower
resolution maps evaluate broader connectivity for this segment of the population.

Lowland Forest Connectivity in Western Washington
Landscape History: Pre-European and American Influences

Since the latest retreat of the Puget Lobe of the Vashon Glacier 13 thousand years ago
climate changes and human influence have shaped the ecosystem. Originally, much of the Puget
Sound Lowlands were dominated by prairies and oak (Quercus spp.) woodlands. Climate change
about 8000 years ago started a trend toward temperate coniferous forests. Native Americans
preserved prairies and oak woodlands in many places. Recently American and European land use
practices have profoundly altered the biome.

27

Prairies and oak woodlands took a foothold due to a 5000-year period of warmer weather
that followed the glacial retreat (Bowcutt, 2009). When the climate began to cool about 8000
years ago, Douglas fir (Pseudotsuga menziezii) colonized much of the prairies. Native
Americans, however, controlled this succession in some areas in order to conduct agriculture,
hunt and gather acorns. Crawford and Hall (1997) estimate that at the time of European and
American contact, prairies and oak savannah covered 150,000 acres on areas abutting the
southern Puget Sound and scattered portions of the Chehalis River basin. Currently, prairies and
oak woodlands persist most notably in what is now the low-lying areas of Lewis, Thurston and
Pierce Counties (Kruckeberg, 1991; Duer 1999).
Much of the Puget Sound lowlands became dominated by coniferous forests. This is
largely due to a cooling climate 8000 years ago. Kruckeberg (1999) mentions that both Captain
Vancouver and Malaspina, sailing for Britain and Spain respectively in the late 1700’s, observed
mature conifer forests growing to the Puget Sound’s shore in many areas. This provides a picture
of the Puget Sound Lowland landscape prior to European and American contact. This forest type
typically involves a succession of red alder (Alnus rubra) colonizing barren or recently-burned
areas, Douglas fir (Pseudostuga menziezii) and finally western hemlock (Tsuga heterophylla)
dominating at the end of the succession. Western red cedar (Thuja plicata) grows in wetter or
more shaded areas often forming groves. Bigleaf maple (Acer macrophylum) is often
interspersed within conifer forests. Sitka spruce (Picea sitchensis) favors coastal areas, valleys
with much precipitation or wetter areas. A plethora of additional tree species are present in the
region, each favoring specific moisture, shade, altitude and soil characteristics (Kruckeberg,
1991). The succession cycle is generally restarted by wildfires started by lightning (Garman et al.
1990).

28

American Influences: 1850 to Present

European and American settlement generally began in the 1850s and steadily expanded.
American settlers from the east mainly settled along the Puget Sound, in prairies and along rivers.
Native Americans had originally established permanent settlements along the coasts and in river
valleys but shifted to permanent settlements along the coast as American settlement expanded.
Originally, settlers came with the intention of practicing agriculture on prairies and converting
Native Americans to Christianity. Settlements soon became more sophisticated and timber
extraction was gradually expanded(Cox, 1999).
Timber production expanded as the market for lumber expanded and technological
innovations came into play. Not only was the wood used for northwest towns but it was exported
to create urban centers in California during the gold rush. Wood was also used for steam energy
to drive ships. Up until the early 1900’s, logging was restricted to taking place along rivers that
could be used to transport the timber to Puget Sound. In addition, the market and available
technology resulted in selective logging. Douglas Fir (Pseudostuga menziezii) was the only
species targeted and larger firs, along with all the other species, were left since the saw blades in
lumber mills were too small to process them. Railroads and the ‘Steam Donkey’, a machine for
yarding timber, enabled timber extractors to efficiently clear land well beyond rivers. Western
Red Cedar (Thuja plicata) also became a valuable commodity. Aided by a network of railroads,
the extraction of old growth timber continued but at farther and farther distances from the Puget
Sound and large rivers. Trucks and more advanced yarding machines started to be widely utilized
in the 1940’s – indicative that old growth timber was being extracted at more rugged areas
beyond the reach of railroads (Cox, 1999).
Since the time of American settlement, timber extraction, agriculture and urbanization
eliminated forest and altered the forest structure of the Puget Sound Lowlands and lower

29

elevations of the Cascade and Olympic Ranges- the R. aurora range. Much of the Puget Sound
Lowlands was logged by the 1920s and by the mid 1930’s most of the lowlands of western
Washington had been logged at least once (Andrews and Cowlin, 1940). Agriculture was carried
out in river valleys and deltas and small towns developed along the coast and large rivers and
established trade routes. Up until the 1930s, artificial replanting was generally not practiced and
clear cut areas were left to reforest naturally. As a result, Red Alder became more widespread;
bog soil samples from the mid 20th century reflect a greater percentage of red alder pollen than
that of other species (Cox, 1999).
The conservation ethic grew and in the 1960s and North Cascades National Park and
wilderness areas were created, thus preserving some old growth forests 8 in the Cascades and
Olympic Ranges (Franklin, 2007). National parks often preserved old growth in valleys in
addition to alpine and subalpine areas while wilderness areas were generally established at higher
elevations. Some river valleys in the three national parks protected lowland old growth forests
but this was limited. Since many lower elevations in the Cascade and Olympic Ranges were
reserved for silviculture, connections between old growth patches in protected areas was
dependent on the configuration of mature conifer or deciduous forests. In addition, commercial
forestry led to biologically simple forests. Patches of replanted single-species Douglas fir became
increasingly common since it was the most valuable species for lumber, pulp and paper (Franklin,
2007).
In addition to timber extraction, American land use beliefs and practices changed the fire
regime. The mid to late 1800’s were characterized by expansive forest fires, usually set by

8

The term ‘old growth forests’ is has many different meanings but in this paper it is used to describe
Pacific Northwest temperate rainforests that feature a variety of tree ages, average age of dominant
species approaching half the maximum longevity (about 150 years for shade tolerant species), some old
trees with ages 300 years or older, natural regeneration of dominant tree species within canopy gaps and
presence of standing dead or dying trees. This definition was devised by Moseler, Thompson and Pendrel
in 2003.
30

settlers (Garman 1990). The rapid rate of logging in the early 1900s also increased the amount of
‘slash’ or unwanted wood on the ground which facilitated powerful fires. As the 20th century
progressed, the US forest service and other agencies carried out a policy of attempting to prevent
and fight any fire. This led to further fuel accumulations which resulted in powerful fires that left
much more flora, fauna and soil micro organisms in the landscape dead than the less intense fires
of an earlier age. Since the 1990s, forest managers have attempted to ameliorate decades of fire
suppression by controlled burning (Puettman et al 2008).
Starting in the 1940s, construction of state and federal highways and a gradual influx of
people accelerated forest habitat loss and fragmentation. As environmental sentiments started to
grow in the 1960s, old growth forests were preserved. Western Washington, for example
typically has relatively intact old growth forest habitat patches in the larger valleys of National
Parks and wilderness areas, the I-90 corridor and Department of Natural Resources (DNR) land
on the western Olympic Peninsula. These old growth fragments are largely isolated from one
another and border lands covered by a mosaics of different age classes of timber as well as
agricultural land and land that is steadily becoming more urbanized in certain areas (Cox, 1999).
In the 1991, timber extraction slowed somewhat in the national forest lands of the Cascades and
Olympics when the Northern Spotted Owl (Strix occidentalis) and the Marbled Murrelet
(Brachyramphus mamarotus) were both listed as threatened under the Endangered Species Act
(USFWS, 2004; 2009). As a result of this ruling, logging on National Forest Service lands has
been greatly reduced.
Urban development has been the principle habitat fragmentation force in the Puget Sound
Lowlands over the past 40 years. This is mainly true in areas nearest the Puget Sound and federal
and state highways. Agriculture persists in the river valleys and industry tends to stay in
established urban areas. The region continues to attract people from around the country and is
expected to increase in population by 1.5 million people by 2020. In 1990, the state legislature
31

passed the Growth Management (GMA) which incorporates various mechanisms to discourage
housing sprawl. Nonetheless, landowners often have some economic incentive for converting
forest and agricultural areas to residential housing, especially if such areas are within driving
distance of an urban or suburban center.
Even though data is limited on the spatial patterns of forest clearing, it is possible that
much of the lowlands were adequately well connected for R. aurora in the midst of timber
extraction. This can be attributed to two factors. First, people generally cleared areas that were
small enough to allow for frog movement. Second, the species’ has shown an ability to utilize
stands of deciduous trees, which naturally took root after humans clear cut land.

32

Chapter 3 Methods

Methodology was based on the literature review, which provided valuable information
on the biology and ecology of R. aurora as well as on current accepted sampling and analysis
procedures.

Wetland Selection
In selecting aquatic areas to study, I considered wetland type, presence of forest buffers
around wetlands and levels of surrounding human impact to the landscape. Ultimately, property
owner approval played a role in determining what wetlands were studied. I refer to the final 14
wetlands that I chose to compare as ‘Study Wetlands’. While selection commenced in December
2010 and lasted through March of 2011, I continued to search for more sites until the end of my
study. Initially, I attempted to focus on wetlands in Thurston and Pierce Counties but as the study
progressed I obtained some in King, Skagit and Whatcom Counties. Mainly, I selected wetlands
by utilizing Geographical Information Systems (GIS) software and publicly-available GIS
websites but I also relied on the advice of regional experts and land managers.
Many wetlands were selected using Arc Map 9.3.1 software (Environmental Systems
Research Institute, Redlands, California) or county GIS systems available online. I created 4
separate maps for Thurston, Pierce, Skagit and Whatcom Counties. In each map, I first added a
2009 aerial photo that was created from the US Department of Agriculture (USDA) National
Agricultural Inventory Program (NAIP). Specifically, I visited the USDA Geospatial Data
Gateway website (http://datagateway.nrcs.usda.gov/ ) and downloaded compressed NAIP county
mosaics for the appropriate Washington county. This data had a coordinate system of Universal
Transverse Mercator (UTM) and was projected in the North American Datum (NAD) of 1983.
Since it was raster data, this coordinate and projection, UTM NAD 83, was adapted for each
33

wetland selection map. I then overlaid the most current National Wetland Inventory (NWI)
shapefile from the NWI website (http://www.fws.gov/wetlands/ Data/Data Download.html). It is
widely known that NWI data is incomplete and inaccurate (Gale and Kudray, 2000; Johnston and
Maysembourg, 2002) but it is generally effective at showing the inundated and non-forested
wetlands that are the most ideal breeding habitat for R. aurora. Additionally, NWI classifies
wetlands by the Cowardin system (Cowardin et al., 1979) which gives substantial information on
vegetation and water regime and this enabled me to identify appropriate wetlands.
I then downloaded road shapefiles from county GIS agency websites and acquired parcel
ownership data from county assessor offices. I utilized 2006 Thurston County Parcel data owned
by the Evergreen State College. For Pierce County, I purchased 2011 parcel data shapefile from
the Pierce County Assessor’s office. In the case of King County, I relied exclusively on the online
King County IMap service to select wetlands and determine ownership
(http://www.kingcounty.gov/operations /gis/Maps/iMAP.aspx). This tool included orthophotos,
links to assessor data and information on King County wetlands in addition to those recognized
by NWI.

Study Site Selection Criteria
Random sampling was not employed to arrive at the final selection of study wetlands.
This was due to the fact that, considering my budget and time limitations, it was not realistic to
accumulate a large number of wetlands to randomly select from. Instead, I relied on selecting
wetlands that were similar with respect to several physical parameters but differed with respect to
the dependent variables which were the size of the initial forest patch and the distance to the
nearest paved road.
Initially, there were 211 wetlands that I was interested in including in the study based on
plant communities, water regime, water quality and size. I then sought permission to enter as
34

many of these as I could. In order to qualify as being good R. aurora breeding habitat, the water
needed to have sufficient light exposure to allow embryonic development and food resources
(Storm, 1960; Licht, 1969; Calef 1973; Brown, 1975), it needed to have relatively thin-stemmed
vegetation for egg masses to adhere to (Storm, 1960; Licht, 1969; Calef 1973), and it needed to
have a hydroperiod of at least 6 months which is sufficient to allow animals to complete
metamorphosis (Richter and Azous, 1995; Hayes et al., 2008). Considering these requirements, I
used GIS applications to find palustrine emergent wetlands with seasonal inundation (PEMC)
and palustrine emergent wetlands with semipermanent flooding (PEMF). These are categories of
freshwater wetlands that are generally under 6 feet in depth and include non-woody vegetation
protruding from the water surface (Cowardin et. al. 1979). I included wetlands with lower-case
modifiers; a common example was ‘PEMFb’ wetlands which are influenced by American beavers
(Castor Canadensis). I also included PEMC or PEMF portions of larger wetland complexes.
Although wetlands like this can be found along the margins of lakes and ponds, I avoided
deepwater habitats because I wanted to avoid the confounding variables of predatory fish and
bullfrogs (Rana catesbiana) that are associated with more permanent water bodies. In addition to
selecting wetlands with these parameters, some wetlands on Joint Base Lewis McChord were
recommended to me by biologists experienced with the area.
In addition to water regime and plant community, I based my selection on wetland size
and distance to other wetlands that would be appropriate breeding areas. I selected wetlands
between ½ acre and 10 acres in area. Marc Hayes of WDFW recommended this range because
wetlands smaller than ½ acre may not adequately reflect surrounding habitat characteristics and
wetlands over 10 acres would be too difficult to census in a reasonable time frame. Marc Hayes
also suggested that wetlands should be separated from other potential breeding sites by at least
400 meters so that they can represent distinct R. aurora populations. Popescu and Gibbs (2010)
and Petranka and others (2004) also stipulate this.

35

In an attempt to limit the variable of poor water quality, I attempted to select wetlands
that were surrounded by forests or grassy areas that were not lawns or cultivation. This was the
case in all but three study wetlands. Study wetland 5 in Puyallup was bordered by an office
property and a road. Study wetland 6 in King County was partially bordered by a lawn. Study
wetland 12 in Skagit County was partially bordered by a hay field. Despite the fact that these
land features were adjacent to these wetlands, none of them demonstrated high levels of algae
growth and they all had pH values from the 6 to 7 range. I therefore included these three sites
because they were representative of a certain level of surrounding development.
Once I obtained ownership information for a wetland that I was interested in analyzing, I
sent a letter describing my study and requesting entry to the landowner. With the input of Dr.
Martha Henderson of the Evergreen State College, I developed a letter template which I used for
a while and then shortened. Both letter templates are included in Appendix B. I sent out a total
of 236 letters to landowners in Thurston, Pierce and King Counties. I ultimately got permission
to ender 29 properties, some of which were owned by the same land owner. Fifteen people
contacted me to decline my request and 14 letters were returned by the Postal Service. Earlier in
the study, I attempted to call about 15 people who had not responded but either could not find
phone numbers for them or got very negative responses.
Several professionals gave me suggestions on appropriate wetlands to survey and
permission to do so. Joint Base Lewis McChord (JBLM), owned by the US Department of
Defense, straddles Thurston and Pierce Counties and holds large expanses of relatively
undisturbed Puget Sound Lowland habitat (Adams and Bury, 1998). JBLM biologists Jim Lynch
and John Richardson suggested what areas to survey on the base and gave me permission to enter.
I selected wetlands in Whatcom and Skagit Counties toward the end of the Rana aurora
embryonic phase and time was therefore short. For this reason I almost exclusively surveyed
wetlands on either public land or land trust land and relied on the guidance of others. The
36

Washington Department of Natural Resources (DNR) granted me permission to enter their lands
throughout the state. US Forest Service (USFS) biologist Ron Gay facilitated surveys in North
Cascades National Park and adjacent Mount Baker-Snoqualmie National Forest lands. Steve
Walker and Karen Grimland of the Whatcom Land Trust gave me permission and assistance in
surveying their lands. Jennifer Bohanon of WDFW also gave me site suggestions and landowner
contacts in Whatcom County.
For two of the study site wetlands, I analyzed fieldwork results that were obtained by
other people. Egg mass count data for site 12 was done by Ron Tressler of Seattle City Light.
Egg mass count data for Site 14 was conducted by Cedar River Watershed biologists Heidy
Barnett and Shelly Nickelson. The site 14 data reflects a five year average for Deep Lake
(Barnett and Nickelson, 2008).

Final Selection of Sites
Even after carefully pre-selecting sites, it is necessary to visit them in order to determine
if they can be considered appropriate R.aurora breeding areas. Ultimately, I surveyed 27 sites but
rejected 14 because I felt that other factors at them besides those relating to connectivity affected
R.aurora breeding. In some cases, I felt that I was surveying the area too late and that egg masses
had hatched and disintegrated. In other cases, I arrived at a site to find that the area appropriate
for breeding was smaller than .5 acres. Additionally, I omitted other sites because they were
excessively shaded , were impacted by water quality issues, were dominated by Ambystoma
gracile, or I suspected that it had an excessive fish or bullfrog presence. I finally chose to use 14
sites that were all of sufficient size, had the appropriate plant community characteristics and
lacked influence by other species.
Ultimately my sites were spread from Thurston County to the Canadian border (Fig 6).
The majority of sites were located in the lowlands of Thurston and Pierce Counties and this is
37

probably representative since these areas have a large number of palustrine wetlands. Property
owner permission strongly determined what sites were available to me and for this reason there
was likely a bias to my selection. Most of the owners that I contacted were residential property
owners and 12 of them gave me permission to enter. People who granted me permission did not
seem to be politically opposed to what I was doing and were even interested in my results. I also
contacted corporate owners although this was a very small subset of the total people I contacted.
The majority of them, including timber companies, gave me permission. One timber company
denied my request on the basis that I did not have adequate insurance. A few agribusinesses cited
food security concerns and denied my request. Land trusts and city utility companies allowed me
to enter. I entered public land without seeking permission.

38

Figure 6 Locations of Selected Study Sites

39

Number, Name

1. Bald Hill Road

General Location

East Thurston Co.

UTM

535121.14mE

Tax parcel # or
Ownership
22603310000

Size (ha)

1.73

North of the town of
Rainier, Thurston Co. ,
East of Hubbard street

523792.79mE

-most of wetland was about 50 cm
-Dominated by Phalaris arundinacea; patches of Typha
latifolia, Carex obnupta, Scirpus microcarpus

5193718.70mN
2. Rainier #1

Characteristcs

21604310000

.52

-about 48cm
-East section; Spiraea douglassii dominates on the northern
edge, ---sedge dominate central portion,

5193923.11mN

-West section; Spirea douglassii around edges, interior
comprised of Scirpus microcarpus, Carex obnupta, Juncus
effusus
3. Rainier #2

North of the Town of
Rainier, Thurston Co.

524065.28mE

63550006800
(Vincent)

.17

-Spiraea grows around t he

.19

-Wetland generally has about 45 cm of inundation

5194600.56mN
63550007000
(Miller)
4. Veckved

NW Whatcom Co.

528938.81mE

-Alnus Rubra around edges, Phalaris arundinacea
dominates, also Lysichiton americanum and Carex obnupta

5427509.27mN
5. Puyallup

Puyallup, directly East
of Pierce Co. Airport

554794.83mE

0419275011

.05

-Spiraea douglasii around edges and in a strip in the middle,
Labrador tea, Scirpus subterminalis in the open water areas,
also Juncus effusus

5217256.77mN

6. Lake Youngs
area

-About 80 cm of inundation

SE King County,
unincorporatded area
between Kent and
Covington

567032.06mE

7. Taylor family
LP

Central Thurston Co.,
north of Tenino
adjacent to private
timber land

511922.99mE
5192765.28mN

11608230000
(Taylor Family
LP)

.75

-70 cm water depth; this is a beaver pond with a dam at the
southern end and another in the middle. Southern section
avg. 70 cm depth
-wide variety of emergent plants

8. Clear Lake

Western Pierce Co.,
NW of Eatonville

545313.71mE

Manke Timber
Co.

.24

-About 80 cm deep in February

1222059015

.64

-Dominated by Phalaris arundinacea. Spiraea douglasii and
Salix spp. along edge

5250383.71mN

-East 2/3 of wetland is emergent and dominated by Phalaris
arundinacea

5193018.20mN

9. Baker River

10. Pipeline

11. Stringtown Rd

12. Harrison
Slough

13. No Name
Lake

North Cascades
National Park, on the
Baker River trail about
2 miles from end of
road at Baker Lake

606962.83mE

West side of the Rainier
Training Area, JBLM,
SW of Rainier Road

517378.66mE
5197618.83mN

US Army

Rapjohn Lake area, NW
of Eatonville, N of
Stringtown Rd

551502.24mE

Manke Timber
Co.

Skagit River Valley,
east of Rockport,
between the river and

608261.13mE

Central Rainier
Training Area, JBLM,
east of Rainier Rd.

520951.60mE

5403125.41mN

-Averages 70 cm

1.69
National Park
Service

-About 65 cm in early April 2011
-Plants not carefully recorded, relatively open water,
sedges, rushes

2.99

-About 5 feet of water in February 2011
-The north half of the lake is open water and dominated by
emergent plants,while the southern and southwest portions
are dominated by Salix spp. The northern portion was
considered habitat. Much of this area is dominated by
Typha latifolia but there are some patches of Spiraea
douglasii

.38

-About 30-40 inches of water During the winter of 2011
- The eastern half of the wetland is appropriate for RAAU
breeding while the western half is scrub-shrub or forested.
This is a very scenic wetland with a variety of plant
communities.

5193018.20mN

Seattle City
Light

.63

-Oxbow pond with up to 10 feet of inundation
- Pond is dominated by Nuphar luteum and Potomegeton
natans. Typha latifolia and Spiraea douglasii on edges.

5371407.76mN
US Army

1.31

-Circular depression that gradually increases in depth to
about 80 cm in the center.

5198344.71mN
-Completely dominated by Phalaris arundinacea

14. Deep Lake

Part of ’14 Lakes
cluster; Western portion
of Cedar River
Watershed, south King
County

583756.20mE
5249761.29mN

Seattle Public
Utilities

1.2

-Lake was relatively devoid of emergent vegetation due to
recently increased water levels; most of the egg masses
were found attached to the branches of dead, submerged
trees that had been installed to increase amphibian breeding

Table 3 Descriptions of Selected Sites

40

Site Name

Location

Ballard

Central Pierce
Co

UTM
550252.72mE

Tax Parcel #

RAAU Egg
Mass Count

Reason for Rejection

0418314039

9

-Under .167 ha

5205588.49mN

-Water quality issues from being in a cow pasture
-Part of area is shaded by trees

Park 1

Baker River
Road, Skagit
Co

593282.19mE

40

-Directly bordered by paved road, which would constitute a unique
disturbance not represented by the other Study Wetlands.

5388356.98mN
-Other large, unsurveyed breeding ponds in area so the role of this wetland
in RAAU breeding was unknown

Sumas
Mountain
Beaver Pond

NW
Whatcom Co,

Wickerhsam
1

SW Whatcom
County off
SR 9

558703.61mE

SW Whatcom
County, off
SR9

558324.04mE

North
Whatcom Co.

576336.62mE

Wickersham
2
Potters Pond

559469.86mE

19

-This is mainly an AMGR pond. 95 AMGR egg masses were found.

13

Forested wetland, not ideal

10

-Actual area where RAAU were breeding was under .0836

5419226.76mN

5389293.49mN

5388720.75mN

-45 AMGR masses in surrounding areas
0

Was visited in April, season likely passed

0

Good habitat. Possibly visited too late

5

-This is mainly an AMGR breeding area.

12

-area under .167 ha

5416169.83mN
Ranger Lake

JBLM1, west
RTA2

516429.36mE
5197752.44mN

Springer Lake
Beaver pond

Central
Thurston Co.

509705.87mE
5198601.88mN
509693.35mE

Springer Lake
Stormwater
pond,
adjacent
beaver pond

Central
Thurston Co.

Pond, NE of
No Name
Lake

JBLM,
Central RTA

Paine Jr
(Trustee)

SW of Yelm,
Thurston Co.

- This was mainly research on natural wetlands, not stormwater ponds,
hence this site would have been unique.

5199172.15mN

-The beaver pond was unique in other ways; it was heavily shaded
521125.67mE

521180.76mE

12

-area under .167 ha

5198897.35mN

5201099.67mN

532468.61mE

22605140000

0

-The owners informed me that this was once used to raise bullfrogs

12618220000

0

-pond as deep as 6 ft (according to the owner)

-Heavily shaded; in the same forest patch as site 13.

5194618.08mN
Studabaker

Kehoe
Clearcut

West
Thurston Co.,
off of 140th
Ave, SW

500649.54mE

Clearcut
Parcel

512398.15mE

-water was not clear and it was stormy. Owner informed me that he had
stocked it with trout, much algae- probably from a Christmas tree farm on
opposite shore

5191803.90mN

33

-Wetland was in the middle of a clearcut, so it was different than all of the
other sites. It was surrounded by the same forest patch as site # 7 so I
chose to use that site instead.

10

Mainly a AMGR breeding area: 45 egg masses counted

5193047.50mN
Kehoe Beaver
Pond 1

Clearcut
parcel

512486.28mE
5193120.34mN

PB Lumber 1

In forest, east
of Kehoe
clearcut

513020.80mE

0

-Pretty good habitat; large, shallow pond with islands. Possibly surveyed
too late

PB Lumber 2

Large pond N
of PB
Lumber 1

513441.37mE

0

-Good habitat around edges; beaver pond. Possible fish presence: I heard a
loud splash. Possibly surveyed too late;

East of ‘PB
lumber2’

513441.37mE

0

-pretty good habitat; beaver pond. Deep but with many islands of sedges

PB Lumber 3

5193161.11mN

5193226.10mN

-Possibly surveyed too late

Table 4 Descriptions of Rejected Sites

41

Data Collection Methods
Egg mass counting was employed because it indicated the level of breeding effort at a
given wetland. Breeding effort can be used as an index to estimate population levels of adult
frogs breeding at the wetland and gives an idea of the amount of frogs that are able to utilize the
surrounding landscape. Egg mass counts provide a more accurate estimate of population trends
than other methods (Patton and Harris, 2010). Per information by Patton and Harris (2010),
R.aurora is well suited for this method because the species is a relatively explosive breeder,
individual masses are clearly separated, egg masses persist for about 5 weeks, masses are large
and tend to be found in predictable areas (Jones et al, 2005).
The 2011 breeding season was unique but I attempted to obtain a sufficient sample size
of wetlands by expanding my surveys beyond Thurston and Pierce Counties and by utilizing data
from other researchers. I visited each wetland once from February 10 to April 5, 2011. R.aurora
throughout the region had bred particularly early in 2011, probably due to a period of warmer
weather in January (Jennifer Bohanon, WDFW, personal communication). As a result, I believe
that I surveyed some sites too late to obtain an accurate count of R. aurora breeding activity. I
omitted these sites from the selection. I therefore visited sites in King County in mid March and
sites in Skagit and Whatcom Counties in early April. I incorporated 3 sites from this part of the
state; all seemed to reflect later breeding, probably due to a later date at which the water reached
high enough temperatures. I also included 2 sites that had been surveyed by other investigators.
I conducted a census of egg masses for each wetland. I chose a census over sampling
because I learned that R. aurora often deposit egg masses unevenly throughout wetlands, even if
conditions throughout the entire wetland are equally favorable with respect to sunlight exposure,
emergent vegetation and water depth. I attempted to visually search the entire wetland, by
viewing it from the shore, walking on logs that extended into the water and by wading through it

42

while wearing chest waders. I used an inflatable raft to observe some sites. I wore polarizing
sunglasses in order to see through the Sun’s glare on the water surface. I tallied each egg mass in
a field notebook. I tallied Northwestern Salamander (Ambystoma gracile) egg masses in order to
ascertain if this species was dominating the wetland as a breeding site. Ambystoma gracile and R.
aurora favor similar wetlands. Ambystoma gracile prey on Rana aurora tadpoles so I considered
a large proportion of Ambystoma gracile masses to be a confounding variable and did not include
wetlands with this characteristic in the selection.

Qualitative Landscape Characterization
I arranged the sites from lowest egg mass counts to highest and then used aerial photo
layers on Arc Map 10 to ascertain patterns of connectivity and fragmentation based on
information in the literature. I considered the arrangement, quantity and size of 1) the primary
forest patch touching the wetland, 2) nearby forest patches, 3) roads, 4) neighborhoods, 5)
pastures, 6) clear cuts and 7) business areas. I observed a general relationship involving the size
of the primary forest patch and roads and decided to study this quantitatively. I also observed
relationships between egg mass counts and 1) the degree of edge effects in the primary forest
patch, 2) the size and accessibility of secondary forest patches and habitat and 3) the broader land
use zoning in the broader landscape.
After a quick assessment of the sites, it was apparent to me that sites that were at the edge
of forest patches had lower egg mass counts, regardless of how extensive the primary forest patch
was. I chose to examine roads that lay within .25 km of the site because previous studies have
shown that roads or deforestation at roughly this distance have the strongest effect on
populations. Eigenbrod and others (2008) concluded that the strongest effects on anuran
abundance and biodiversity in wetlands occurred when roads or deforestation was within 500 m

43

of the wetland edge. Semlitsch and Bodie (2003) found that 95% of the population of a given
species utilizes upland habitat within 159 – 290 m from the wetland.

Quantitative Analysis
I used Arc Map 10 to measure the forest patch sizes and degree of road disturbance.
First, I created maps of each county that included NAIP 2009 aerial photos, 2010 NWI data and
layers for roads. For each wetland, I created a 2km wide buffer around the wetland because
Hayes and others (2008) state that seasonal movements exceeding 1 km may be typical (Mathias,
2008). I then used the area tool to measure the size of the primary forest patch and the area of the
buffer. In order to find the percentage of the buffer that was covered by the primary forest patch,
I divided the later by the former. I plotted ‘Egg mass counts’ against ‘percentage of buffer
covered by primary forest patch’ in Excel. I tested the relationship for significance by importing
the graph into the program JMP (Statistical Analysis Systems, Cary, North Carolina) and used a
chi squared test to test for significance.
Based on my observations of sites at the edge of forest patches and on the literature, I
used Excel to compare sites that had at least one road within .25 km of the wetland edge. Since
this was count data and non-parametric, I used the JMP program to subject the data to the
Wilcoxon / Kruskal-Wallis test s (rank sums) to test for significance. The literature states that the
traffic level of nearby roads is a significant factor in amphibian diversity and abundance in
aquatic areas (eg. Fahrig et al. 1995; Eigenbrod et al. 2008; Mazerolle 2004). To address traffic
intensity, I obtained as much traffic count data as I could for roads that were near each study site
from county and Washington Department of Transportation reports, web tools and conversations
with officials. Based on the literature, I classified roads as having low, moderate, or heavy
traffic and speculated on their affects at the sites.

44

Chapter 4 RESULTS

I numbered the sites in ascending order based on egg mass counts (Site 1 having the
lowest count and Site 14 having the highest count) and, after looking at maps surrounding sites,
observed landscape connectivity patterns. I observed that egg mass counts seemed to increase as
the size of the primary forest patch increased. Many of the sites in the middle of this selection
were on the edge of the primary forest patch and within .25 km of a road. I analyzed the sites
quantitatively by calculating the percentage of the area within the 2 km buffer that was covered
by the primary forest patch and tallying up the sites that had a busy road within .25 km of the
study wetland. I compared both of these quantities with egg mass counts.

45

Figure 7 Study Site 1, Eastern Thurston County

46

Figure 8 Sites 2 and 3, in the Town of Rainier

47

Figure 9 Site 4, Northwest Whatcom County

48

Figure 10 Site 5, in the City of Puyallup

49

Figure 11 Site 6, Between Kent and Covington, King County

50

Figure 12 Site 7 North of Tenino, Thurston County

51

Figure 13 Site 8 Northeast of Eatonville, Pierce County

52

Figure 14 Site 9, Baker River area, North Cascades National Park

53

Figure 15 Site 10 Rainier Training Area, Joint Base Lewis-McChord

54

Figure 16 Site 11 Northwest of Eatonville, Pierce County

55

Figure 17 Site 12 Skagit River Valley, Skagit County

56

Figure 18 Site 13, Rainier Training Area, Joint Base Lewis-McChord

57

Figure 19 Site 14 Cedar River Watershed, central King County

58

Primary Forest Patch Size within 2 Km and Breeding Effort
I calculated the percentage of the area within 2 km that was covered by the primary forest
(the contiguous forest patch touching each study wetland) (Table 5). I then plotted breeding
effort (egg mass counts) against the size of the initial patch (Figure 7). There was a significant
(p=0.0001) and strong positive relationship: as the primary forest patch increased in size within 2
km, breeding effort increased. TheR2 value of the regression analysis was .79, indicating that 79%
of the variation in breeding effort could be predicted by patch size.
Site
Egg Mass
Number Count

Primary Forest
Patch Size,
Including Area
Beyond 2km
radius Buffer

Primary Patch
size within 2km
radius buffer

Percentage of
Buffer filled
by Primary
Forest Patch

Busy Paved
Roads
within
.25km of
wetland

1

0

6.88

6.88

5.25

Yes

2

4

5.25

5.25

0.37

Yes

3

7

14.65

14.65

0.01

Yes

4

12

234.8

191.2

14.1

Yes

5

17

71.6

71.6

5.4

Yes

6

19

220

293.8

21.9

Yes

7

100

561

488.9

35.2

Yes

8

120

538

489.5

37.3

Yes

9

154

863.8

333.1

25.3

No

10

198

1725

1116

68.9

No

11

265

431

409

30.6

Yes

12

300

2000

645

64.7

No

13

305

4683.49

1195.5

89.1

No

14

387.7

48227

1020.5

77.4

No

Table 5 Quantitative Results
59

Red Legged Frog Breeding Effort as a
Function of Primary Forest Patch Size
Red Legged Frog Egg Mass Count

450
400

R² = 0.792

350
300
250
200

Egg Mass Count

150

Linear (Egg Mass Count)

100
50

y = 4.013x - 1.4006

0
-50 0

20
40
60
80
% of Buffer Covered by Primary Forest Patch

100

Figure 20 Rana Aurora Breeding Effort as a Function of Primary Forest Patch Size

The Effect of Roads and Traffic on Breeding
Road effects were analyzed quantitatively and qualitatively. Eggs mass counts was
significantly less in areas near busy roads (<.25 km). This is shown graphically (Figure 19). Egg
mass counts in areas < 0.25 km from road were also compared to those >.25 km from a road
using the Wilcoxon/Krustal Wallis rank sums test. This test showed that there was a significant
difference in egg mass counts between those sites (p<0.01). The effect of traffic levels was
analyzed more qualitatively (Table 6). Sites 1 through 7 were moderately or heavily impacted by
traffic levels and Sites 8 through 14 were either not impacted by traffic levels or incurred very
low impacts from traffic levels. This reflected a general relationship between traffic levels and
breeding effort.

60

Figure 21 Effect of Nearby Roads on Rana Aurora Breeding Effort

61

Site

Nearby Roads (road
closest to site is given
first)

1

-Bald Hill Rd
-128th
-138th
-118th
-127th
133

Average
Daily
Traffic
9
(ADT)
4976
192
135
756
549
1376

Medium
Low
Low
Low
Low
Low

2005
2005
1994
2001

4

-‘H’ Street
-Delta Line Rd

803
242

Low
Low

2007
2001

5

-SR 161
-110th Ave E
-152nd St S.
-122nd Ave E.
-224th

4500
4675
7600
8150
5162

Medium
Medium
High
High
Medium

2004
2010
2010
2010
2011

-Old Highway 99
-Offut Lake Road
-22nd Ave
-416th Ave
-Dean Kreger Road

4450
1157
<200
200
425

Medium
Low
Low
Low
Low

2004
1999
estimated
2004
2006

2
and
3

6

7
8

Traffic
Intensity
10
Rating

Year of
Traffic
Count

9

Comments

Moderately impacted by roads and traffic. There are few roads in the area but Bald
Hill road is directly adjacent to the site and separates it from a large forest patch.
Other forest patches to the east are smaller and more distant
Heavily impacted by roads and traffic. Many small residential roads in the area for
which data isn’t available. Traffic intensity for them are probably ‘low’ as they are
similar to 118th and 127th. 133rd Avenue lies between the site and the vast forest areas
of JBLM.
Moderately impacted by roads and traffic. Few roads in the area. The ‘0 road’ which
is in British Columbia and runs along the border appears to be moderately busy but no
traffic count data is available for it.
Heavily impacted by roads and traffic. The primary forest patch is surrounded on all
sides by either dense urban development or moderately to heavily traveled roads.
Beyond these land covers are additional busy roads and development.
Heavily impacted by roads and traffic. A contiguous 3/5 of the buffer area is
characterized by residential area with feeder streets. 224th is immediately south of the
site. Traffic levels are likely the same on other similarly sized roads but traffic count
data is sparse.
Moderately impacted by roads and traffic. Highway 99 is within .25 km to the west of
the site. Much of the land east of the site is not bounded by paved roads.
Low Impact by Roads and traffic. The roads nearest the site have low traffic counts.
Much of the primary forest area is not bounded by roads.

No road or traffic impacts. The entire buffer area is in North Cascades National Park.

10

-Rainier Road

5142

Medium

2008

Low to Medium impact. Rainier road lies close to the site but virtually no other roads
are in the buffer area.

11

-Stringtown Road
-Eatonville Cutoff
-SR 7
-SR 161

450
3800
2000
8400

Low
Medium
Low
High

2001

Low to medium impact. Stringtown road is within 200 meters of the wetland but is
not highly traveled. The other highways are more distant.

12

-Rockport-Cascade
-Martin Ranch Rd.

203
50

Low
Low

2009

Low Impact. These two roads are lightly travelled and do not isolate populations
from forested areas to the south

13

No Road Impacts. There are no paved roads in the buffer area

14

No Road Impacts. There are no paved roads in the buffer area.

Table 6 Effect of Traffic Levels on Breeding Effort: Qualitative Analysis

9

Traffic counts are given for the portions of the road nearest the site.
Traffic Intensity Ratings are taken loosly from those devised by Fahrig et al. (1995): Low: 500-3500, Med: 5000-6000 and High:8500-13,000. I am considering
low to mean 0-3500, Med: 3500-8500 and high: >8500.

10

62

Chapter 5 DISCUSSION AND RECOMMENDATIONS
Part 1: Scientific Conclusions
Summary
This study quantifies the population of R. aurora, (as estimated by eggs counts) that
breed in a selection of wetlands and attempts to explain these quantities based on the surrounding
landscape characteristics of each wetland. An underlying concept in many connectivity and
habitat principles around wetlands is that lentic breeding amphibians that favor forest habitats use
visual or other sensory clues to gravitate toward forest when leaving the wetland (Semlitsch,
1998; Walston and Mullin, 2008). Animals will therefore move toward any area around the
wetland that is forested and then continue to exploit appropriate habitat or migrate from there.
They will even attempt to cross roads, some of the most adverse land covers. This study has
found 1) significant positive correlations with the breeding population represented in study
wetlands and the size of the primary forest patch adjacent to the wetland and significant negative
correlations with breeding population size and the presence of roads within a quarter mile of
some of the study wetlands.
In order to fully characterize the surrounding landscape more exact quantitative methods
should be applied to landscapes that are comprised of a mosaic of land covers and that lie within
the 2km radius of the wetland. This study quantifies 1) the area of the primary forest and 2) the
affect of roads within .25 km of the breeding site. Within the 2km buffer, this study addresses
landscape characteristics beyond the impacting road in a way that is more qualitative and
incompletely quantitative. Beyond the initial impacting road, which R. aurora attempt to cross,
lie secondary forest patches, pastures and other roads. The collective impacts of all these land
covers need to be assessed. The shape of primary forest patches also needs to be assessed with

63

more quantitative methodology to determine the degree of edge effects. The most current
developments in amphibian landscape ecology suggest that conducting a GIS least cost analysis
for the area within 3km of each study site would accomplish this. Additionally, the sample size
of 14 selected sites is small; a minimum of 20 sites is recommended to draw stronger conclusions
(Hayes, personal communication). As a result of these factors, this study is preliminary and only
suggests ideas for future investigations and provides justification for more exact methodology.

Forest Patch Size and Breeding Effort
As evidenced by Fig 6, breeding effort was positively correlated with the size of the
primary forest patch. In the sample of 14 sites, primary forest patch sizes ranged in size from
covering .01% of the 2km-wide buffer (study wetland 3) to covering 89.1% of the buffer (study
wetland 13). Generally, the sites reflected an increase in breeding effort as the primary forest
patch increased in size. Forest habitats in western Washington preserve moisture, maintain a
more constant microclimate, protect animals from radiation and provide invertebrates for food.
These results are consistent with conclusions reached by Aubry, (2001), Chan-McCloed (2003),
Chan-McCleod and Moy (2006), Haggard (2000 ), and Schuett-Hames (2004) that Rana aurora
favor precanopy and mature forests. On an intuitive level, larger forest patches come closer to
resembling the large expanses of old growth forests that Rana aurora evolved in.

Breeding Effort and Nearby Busy Roads
As evidenced by Figure 7, the 4 sites with the most egg masses did not have a busy road
within .25 km while all the others did. In addition to resulting in habitat loss, creating edge
effects and altering local hydrology, roads are significant barriers to dispersal and migration
(Eigenbrod et al., 2008; Fahrig et al. 1995; Beasely 2000) This study did not analyze land cover
mosaics in the more disturbed areas around the sites but in many cases a busy paved road lies
within .25 km of the site. For sites 1, 2, 3, 4, 5, 6 and 7 (Figures 8-13) the initial road is merely
64

the first obstacle for reaching other smaller habitat areas. For Sites 8, 9, 10 and 11 (Figures 1417) this road is all that separates animals on study sites from secondary forest patches and other
aquatic areas. In the case of study wetland 12 (Figure 18), a paved road, Martin Ranch Road,
does exist south of the site but this road is probably not sufficiently busy (50 ADT) to create a
sizeable barrier to forest patches south of it.

Access to Secondary Forest Patches
In this study, ‘secondary forest patches’ are defined as other forest patches within 2 km
from the study wetlands that are not contiguous with the primary forest patch that is adjacent to
the study wetland. Although primary patches provide the best habitat and the best corridors
radiating out from breeding sites, R. aurora do attempt to cross deforested areas (Haggard, 2001;
Chan-McCleod 2004) so considering other forest patches is important. This study suggested that
some breeding populations have easier access to secondary forest patches based on the type of the
intervening landscape cover and the size of such covers.
Trends were observed over the study sites that suggest a positive relationship between
breeding effort and the population’s access to secondary forest patches. Study wetlands 2 and 3
are within 600 m of each other, have very small primary forest patches and are similar with
respect to size and characteristics (Figure 8). Study wetland 3 however, is separated from a forest
and wetland area by about 500 m of pasture. Site 2 is surrounded on three sides by rural
residential development. Factors that separate wetlands from secondary forest patches may also
play a role in the increasing egg mass numbers in sites with larger numbers and larger primary
forest patch sizes.
Many of the sites in the selection that had middling values for breeding effort were
located at the edge of large forest patches, thus rendering a large part of their 2 km buffers to be
substandard habitat in the form of mosaics of roads, pastures, residential development and small
65

patches of forest. Animals in site 7 (100 egg masses) could exploit wetland, streams and forest
patches to the west but a residential road (Chein Hill Road), Old Highway 99, and railroad tracks
separate Site 7 from these areas (Figure 12). Study wetland 8 (120 egg masses) lies to the west of
smaller forest patches and aquatic areas that lie within pastures (Figure 14). Only one road
separates these areas. Site 11 (265 egg masses) is similarly separated by a single road
(Stringtown Road) from upland forests and aquatic areas associated with Ohop Creek to the south
(Figure 16).

Edge Effects within Forest Patches
Edge effects may have played a role in reducing habitat at some sites. Where the forest
patch meets a pasture, clear cut, residential lawn or road, the forest habitat along this edge is
degraded. Edges result in an increase in sunlight penetrating the forest which leads to denser
understory vegetation and a more variable microclimate (Weyrauch and Grubb, 2004). Edge
effects probably play a significant role in the sites with the smallest primary forest patches -- Sites
1, 2 and 3 – because the forest patches associated with these sites are so small and separated that
light easily penetrates them (Figures 8 and 9). Sites 4 and 5 had markedly larger primary forest
patches than the first 3 sites but the egg mass counts (12 and 17 masses, respectively) did not
correspond as tightly with this additional habitat (Figure 10 and 11). Site 4’s primary forest patch
is interspersed by 7 gravel driveways and rural residences and 3 clear cuts. It also has a
comparatively circuitous shape. Site 5’s primary forest patch is circuitous in shape and the
southern section is interspersed by 4 large dense scrub wetlands. These areas are less than ideal
habitat and also create edges in the adjacent forests. Edge effects may affect population sizes on
these sites because they reduce habitat quality. Sites with greater egg mass counts simply have a
larger primary patch size which mitigates edge effects since the ratio of area to circumference is
larger. Additionally, some of the larger sites have primary forest patches that are more circular
and not as circuitous which also results in a greater circumference to area ratio.
66

Position of the Wetland within the Primary Forest Patch
Sites that were positioned more in the middle of the primary forest patch generally had
higher egg mass counts than those that were positioned more toward the edge. Although it was
constrained by natural features of steep gradients and the Baker River, Site 9 (154 egg masses)
was located in the middle of a long forest patch (Figure 14). Site 10 (198 egg masses) was also in
the middle of a large forest patch but the fact that a prairie lied to the north and that a deforested
70 m-wide gas line easement / dirt road to the northeast bisected this patch may serve to lessen
the Rana aurora population associated with Site 10 (Figure 15). Site 13 is almost completely
surrounded by contiguous mature forests within 2 km from the breeding area and this area is only
broken up by narrow dirt roads (Figure 18). Site 14 (387 egg masses) is surrounded by at least 1
km of mature forests on all sides. The Cedar River forms a barrier about 1.25 km to the south,
and a utility easement is located about 1km east of it but mature forests exist beyond the easement
(Figure 19).

General Land Use Objectives of the Area Surrounding the Site
This study suggested the broader land use objective of the landscape surrounding the site
is a factor in R.aurora population levels. Land use is largely governed by economics,
government objectives (e.g. national defense) geography and cultural values and is generally
organized by ownership, regulations and zoning. Broad land use objectives and zoning can shape
habitat use over a large scale which can have a unique affect on habitat, depending on the
objective. Some objectives may include factors already discussed such as ‘Edge effects’ and
‘Accessibility of Secondary Forest Patches’. The 14 sites fall into 5 land use objectives; urban,
rural mixed use, timberland, second growth forest preservation and wilderness protection
1. Urban

67

Site 5 (17 egg masses) exists in a relatively large forest patch but this patch is within the
City of Puyallup, where the landscape is dedicated to dense residential development, commerce,
and the transportation infrastructure that accompanies these activities. The amount of pavement,
buildings and traffic severely isolate Site 5 from other habitat beyond 2km (Figure 10). Sites 2
and 3 lie within a small town, Rainier, and are surrounded by residential development. Site 6,
(Figure 11) is impacted by suburban development, even though it is in unincorporated King
County. Over half of its 2 km buffer area is covered by suburban residential areas and, probably
more importantly, many moderately travelled roads that separate small forest patches. While this
area is in unincorporated King County, it
2. Rural Mixed Use
Four of the sites exist in unincorporated, flat parts of the Puget Sound Lowlands with
mixed land use objectives. Sites 1, 4, 13, and 14 (Figures 7, 9, 18 and 14 respectively) fall into
this category. They have a mosaic of commercial forestry holdings; pasture land, rural residential
areas and rural roads that are traveled to varying degrees. R. aurora populations across all these
sites reflect the size of the primary forest patch and the ease to which R. aurora can move to
secondary patches.
3. Protected Wilderness
Site 9 (154 egg masses) is within North Cascades National Park, an area that is
maintained for biodiversity and other environmental values as well as recreation (Figure 14).
Although the area is highly constrained by steep topography and the Baker River and much of the
buffer area is comprised of alpine meadows, rock and glaciers, the preserved status has prevented
many of the human disturbances existent at the other sites. This site is unique in that it is
comprised of old growth forests. Old growth forests may have a higher carrying capacity than
mature second growth, which would allow a greater density of frogs.
68

4. Timber Production
Site 12 (300 egg masses) exists at the edge of an area almost exclusively devoted to
timber production in the Cascade Mountains (Figure 17). Some pastures intersperse the primary
forest patch and the Skagit River, a barrier, lies 1km to the north. The Washington Department of
Natural Resources manages lands south of the site and this area is different than timber lands in
the ‘Rural Mixed Use’ category. These timberlands are much more expansive and are a mosaic of
clearcuts, old growth, mature and other intermediary forest age classes. Site 12 is bordered on the
north by pasture and the primary forest patch is smaller than other primary forest patches in the
selection but the high egg mass count suggests that this landscape is more conducive to R. aurora
populations than the ‘Rural Mixed Use’ landscape R. aurora may be able to tolerate these
landscapes better than ‘Rural Mixed Use’ landscapes because of the lack of moderately or
heavily-travelled roads and the fact that even very young forests provide better habitat than
pastures and residential development. Although 2 roads are present south of the site, they are
comparatively lightly travelled, most likely due to few residences and businesses to the east.
Martin Ranch Road has 50 ADT and Rockport Cascade Road has 203 ADT. Since these roads
are lightly travelled, frogs can probably reach forested areas south of them relatively easily.
5. Second Growth Forests Preservation
Two of the highest egg mass counts in the selection Site 13 (305 egg masses) (Figure 18)
and Site 14 (387 egg masses) (Figure 19) exist in large patches of mature second growth
coniferous forests that are preserved or very lightly logged. This land use type reflects territory
that, like much of western Washington, was initially logged but subsequently reforested and then
preserved under other objectives besides timber production. JBLM maintains forests for military
training purposes. JBLM occasionally logs areas but timber sales are limited in size and
subjected to more stringent environmental standards than many timber harvesting operations

69

(McAllister, 2001). JBLM managers are actually attempting to replicate some old growth forest
characteristics throughout the base (Adams 2000). Site 10 is also on JBLM but has less
connectivity than Site 13, owing, in part to other land use objectives beyond those of the base.
Site 10 is bisected by a moderately-busy road (Rainier Road) and a gas line easement (Figure 15).
The City of Seattle maintains the Cedar River Watershed for high quality drinking water and
therefore does not log it. While these forests have lower carrying capacities than the old growth
forests at Site 9, the sheer size of the forest patches are conducive to relatively large Rana aurora
populations.

Part 1 Conclusion
From 32 surveyed sites, this study selected 14 that were similar to each other with respect
to being appropriate R. aurora breeding sites. These 14 sites reflected a range of conditions with
respect to connectivity in upland forest habitats and this appeared to be reflected in the R. aurora
breeding effort, represented by egg mass counts. Breeding effort is proportional to the population
and the population levels reflect accessibility to appropriate habitat. Egg mass counts can
therefore be used to gauge the levels of upland habitat connectivity.
This study found statistically significant relationships between the size of the primary
forest patch and the presence of busy roads within .25km of the breeding wetland. The larger the
forest patch is and the more it surrounds the site, the more amenable upland habitat R. auroa
populations have at their disposal that is directly connected to the places that they breed in.
Crossing roads pose a substantial risk to R. aurora and other anurans so the closer a road is to a
breeding site, the more it fragments the landscape for R. aurora breeding in the wetland.
This study also observed relationships with other connectivity measures although it only
analyzed them qualitatively. 1) The higher the proportion of edge on the forest patch, the lower
the level of habitat quality within forest patch. Small primary forest patches or those with very
circuitous shapes had lower egg mass counts. 2) Non-forested areas such as utility easements and
70

maintained prairies adjacent to the breeding site may limit R. aurora breeding effort, even if the
breeding site is in the middle of a broad forest area. 3) The broader regional land use objective
appears to have an effect on R. aurora populations. Urban areas have lower populations, even if
the forest patch is comparatively large. Higher populations are associated with: 1) protected
wilderness, 2) large areas of preserved second growth and 3) landscapes that are strictly devoted
to timber production. These three relationships often overlapped with other relationships but I
speculated what factors were influencing R. aurora populations given the knowledge reflected in
the literature and what I observed within the selection. For example, Site 11 was located at the
edge of a landscape devoted to timber production and had one of the highest egg mass counts in
the selection. It also was bordered on the north by a pasture and was connected to a primary
forest patch that was much smaller than sites within the selection that had between 300-387 egg
masses, two factors that are inconsistent with such a high egg mass count. This led me to
speculate that the sparsely-populated landscape in the Cascades reserved primarily for timber
production was more conducive to R. aurora populations than rural mixed use landscapes with a
greater human presence.

Part 2: Recommendations for Further Research
Habitat Use
A study with similar objectives to this one should be undertaken with a greater sample
size, more narrow selection parameters and analyzed with GIS methods. Marc Hayes, amphibian
expert with WDFW has recommended that a selection of 30 sites, selected from an original set of
50 wetlands, would be ideal but that ’20 would work’ (Hayes, personal communication). A larger
sample size would likely result in more robust results and enable us to make better
generalizations about Rana aurora habitat requirements. I would also recommend that the
selection focus on the Puget Sound Lowlands and not include sites in the National Parks and
areas devoted to logging in the Cascades. These landscapes each have variables that the Puget

71

Sound lowlands lack. National Parks have old growth forests and natural barriers. Logging
landscapes are characterized by vast areas that are uninhabited by people, only have a few paved
roads and are covered in a patchwork of different age classes of trees.
A ‘least cost’ analyses with GIS should be used to analyze the results. Such an analysis
would account for all of the land covers and the different metrics associated with connectivity
such as edge effects, and summarize all of them into how appropriate the area is for a given
species. Molly Mathias-Levitt has already conducted a least cost analysis for King County, and
the Bear Creek Basin in particular (Mathias, 2006) but without field data. It would be valuable to
obtain breeding effort data for a selection within King County and observe how the results
correspond with Mathias least cost map.

Conduct a Similar Study in the Cascades and Olympic Ranges
R. aurora range also includes lower elevations of the Cascades and Olympics as well as
the lowlands between Olympia and Portland and lower mountains such as the Willapa Hills. A
study similar to this one in these areas would provide data on how R. aurora is doing in
landscapes that are relatively uninhabited and devoted to timber extraction and wilderness
protection. Such a study could provide insight on forest management and possible perspective on
how the species is doing in the more populated Puget Sound Lowlands.

Conduct Genetic Research to Determine Possible
R. aurora populations may be impacted over the long term if the forest patch that they
occupy is too isolated to allow for genetic exchange with other populations outside the patch.
Because of the embeddednes of human settlement and infrastructure in the Puget Sound
Lowlands, R. aurora populations depend on habitat islands of varying sizes and varying levels of
connectivity with other habitat islands. Research on. The level of habitat fragmentation that leads
to inbreeding would long term conservation strategies for R. aurora.

Conduct Research on Road Impacts
72

While it is generally assumed that roads cause a sizeable impact on amphibian
populations, only one study, Beasley (2002), that directly addresses road impacts has been
conducted in the Pacific Northwest. Much of the work has been done in southern Ontario,
Canada by Lenore Fahrig and fellow researchers. A greater understanding on how R. aurora are
able to negotiate a variety of roads would provide more information on how they can survive in
heavily impacted landscapes. Site 11 would be a good subject for this study since it is a strong
breeding area and Stringtown Road (450 ADT) is within 200 meters of much of its southern end.

Use Radio Telemetry to Research Habitat Preferences and Migration Distances
Although radio telemetry research is expensive, a clearer picture of how far R. aurora
typically travel is still needed. Such information would better inform researchers and policy
makers on what scale R. aurora travel. It would be especially valuable to gain more information
migration in ‘rural mixed use areas’. This land cover type is at the fringe of exurban
development. Up to this point, radio telemetry work has been done on animals in timber
production landscapes.

Use GIS to Inventory Forest Patches in the Puget Sound Lowlands
This research will be valuable for forecasting habitat levels in the future. Such an
analysis should take into account whether or not areas are on private timber land: it is possible
that market conditions will lead to logging on these lands at roughly the same time and therefore
markedly reducing patches of mature forests. This analysis should also take into account whether
or not breeding wetlands are adjacent to patches of mature forest. This question would suggest
areas for wetland creation. A third aspect of this analysis would be to determine if appropriate
Rana aurora breeding habitat exists within or adjacent to these forest patches. This information
would be helpful for selecting wetland mitigation and restoration sites.

Part III: Management Recommendations
Background
73

As stated in Chapter 1, R. aurora occupies a range extending from the northern California
Coast to in British Columbia. In western Oregon and Washington, they inhabit areas from sea
level to 1100m (3400 ft) elevation. Although populations should be monitored, R. aurora may be
more impacted by urbanization and fragmentation within the parts of its range that are closer to
cities than areas that are dedicated to timber extraction and wilderness protection. Within this
range, the Portland-Eugene metropolitan area, Olympia to Everett metropolitan area and
southwest British Columbia are currently highly populated and will continue to grow in
population. Rural areas near these population centers will also become more populated.
The Puget Sound Lowlands are expected to increase by 1.5 million people by 2020 and
this poses a significant threat to the abundance of Rana aurora within the exurban fringe (ShuettHames, 2006). Forest patches of varying sizes dot the Puget Sound Lowland landscape. Some of
these forest patches are preserved for the sake of watershed protection, public recreational land,
conservation easements and conservation trust lands. Others are on private forest land and land
that could be converted to housing developments. The continued existence of these forests is
therefore more tenuous.
The state of Washington has enacted the Growth management Act (GMA) to a) enhance
established cities and towns and b) to facilitate the rural industries such as logging and agriculture
and c) to preserve important natural resource areas. The GMA mandates that local jurisdictions
establish codes to promote these three objectives. GMA codes that protect aquatic areas by
mandating that they have buffers and that impacts to them be mitigated will continue to
ameliorate impacts to R. aurora. Additionally, GMA laws that encourage forestry and seek to
concentrate additional residential and commercial development in urban growth areas will benefit
Rana aurora populations in the Puget Sound Lowlands. Transfer of development rights programs
allow rural land owners to sell their development rights to developers of urban areas are currently

74

being practiced in King County and to a lesser extent Pierce and Snohomish Counties because the
demand is the strongest in those areas. TDR programs will also benefit R. aurora.
While GMA legislation is beneficial to R. aurora population persistence, we must realize
that it is not enough. Private land owners still have the opportunity to develop forest lands to
some extent in unincorporated areas under GMA so more deforestation will continue to occur.
Additionally, although GMA mandates the preservation of lands for forestry market conditions
may result in harvests in roughly similar time frames. More research is required on private forest
lands in the Puget Sound Lowlands and the expected timeframes that they will be harvested.
Growing cities will result in increased transportation between cities and this will result in
increased traffic volumes, require the construction of more roads and the expansion of some
existing roads. Road impacts are therefore likely to increase.
In addition to concerns about the future availability of forest lands and increased road
impacts, the fact that R. aurora will be subjected to surviving on an array of habitat islands in the
Puget Sound Lowlands may have long term ill effects. The isolation could result in inbreeding.
In bread individuals have been shown to be smaller, more lethargic, less resistant to many threats,
and less able to reproduce. The risk of inbreeding would increase with the degree of isolation.

Conservation Recommendations
Recommendations for R. aurora management are multidisciplinary; take into account the
most prescient needs and the realities of an increasing human population and changing landscape.
Many of these recommendations require further research and are thus linked to the previous
section.
Continue to Conserve Aquatic Habitat

75

The GMA’s mandate to conserve and protect critical areas is important for one aspect of
R. aurora’s life cycle and should thus be continued. Through buffers and mitigation, Critical
Areas Laws help to maintain water quality and regulate flows. Wetland typing --commonly
employed in consulting for determining wetland regulations-- enables the land owner and
municipal employees to realize how surrounding landscapes would be amendable to wildlife
passage.
Promote Compact Human Communities and Minimize Road Development
Laws designed under this paradigm including GMA measures and TDR programs -- will
encourage the preservation of larger tracks of landscapes that are relatively amenable to R.
aurora. This thesis and other research suggest that larger patches of upland habitat contribute to
larger populations. The Growth Management Act seeks to concentrate development in or just
outside of established cities and towns. Urban areas are some of the most hostile landcovers to
Rana auora and other amphibians so it is better to concentrate human housing and commerce in
these areas and stem the tide of development in more rural areas. This would have the added
benefit of slowing the construction of roads throughout the countryside and stemming the
increase in traffic flow on at least some existing roads.
The GMA will also help to preserve wetland habitat, forests that surround wetlands and
‘greenbelts’ within urban areas. Protecting the amphibian aquatic and upland habitat within
urban areas would contribute to the gene pool of these species throughout their range and enable
urban amphibians to provide urban wetlands and forests with their ecological services as well as
educational opportunities for urban residents. Such populations may run the risk of becoming
geographically isolated, however, so artificially supplementing them with animals or egg masses
from neighboring areas may be required to maintain genetically healthy populations.

76

Transfer of development right (TDR) programs should continue to further channel
development into established urban areas while financially benefiting rural landowners. These
programs have been established in 5 of the Puget Sound’s most populous counties. The King
County TDR program is currently the most active, due to King County having the most demand
for housing. Hopefully the continued influx of new people will bring TDR programs into wider
use.
Creation and Restoration of Aquatic Habitat
Freshwater wetlands with appropriate amphibian habitat attributes should be established
in appropriate places. Many of these wetlands were lost in the decades that preceded the CWA.
Some were lost due to the chanellization of rivers while others were drained and filled to
facilitate agriculture and urban development in river valleys and deltas. The Puget Sound
lowlands features large blocks of forest bordered by urbanized areas that had wetlands at one
time, a prime example being Tiger and Cougar Mountain areas adjacent to Bellevue, Newcastle
and Issaquah. Wetlands with seasonal and semipermanent hydroperiods and emergent vegetation
should be established near these large forest blocks so that amphibian populations can take
advantage of these forest reserves and shape their ecology with their ecosystem services. Large
forests patches with only one or a few appropriate ponds would also be good locales for
additional ponds so that metapopulation structure can be enhanced.
Rannap and others (2009) recommend establishing new wetlands in clusters and adding
wetlands to areas where there are just a few natural wetlands. Fostering hydroperiods that are
short enough to discourage fish presence yet long enough to allow for amphibian larval
development is currently a challenging goal due to limited scientific results. Creating clusters of
wetlands with each wetland in a cluster being designed to have a different hydroperiod and plant
community, has increased chances of the probability of the ‘right’ hydropriod and plant

77

community being present within the cluster for a given species. If a variety of hydroperiod and
plant community conditions are made available in one area, each of the native amphibian species
can utilize the breeding habitat that it is best suited for (Rannap et al. 2009).
‘Wetland mitigation’ is a well established societal endeavor for creating, enhancing and
restoring wetlands and natural areas adjacent to them, but the success rate of this process needs to
be raised substantially. It is generally carried out (or supposed to be carried out) by developers
who are proposing impacts to wetlands or their buffers and need to fulfill permit requirements for
doing so (Hough and Robertson, 2007). The Clean Water Act, GMA and shoreline management
(SMA) all require wetland mitigation if wetland functions and values are impacted by a project.
The three types of compensatory wetland mitigation are 1) on site projects, 2) in lieu fee
programs and 3) wetland mitigation banking. Since the 1970s, on site mitigation has been the
most common strategy, but this has resulted in failure over half the time (Murphy et al. 2009;
Johnson et al, 2004). In lieu fee programs allow the developer to compensate for proposed
impacts by paying a fee to a government agency which in turn puts these funds toward mitigation
projects (Ecology, 2006). Under the approval of government agencies, Wetland mitigation banks
sell credits to developers applying to do things that impact wetlands so that they can fulfill permit
requirements. The banks can sell additional credits as their projects gain success. Washington
State has established WAC 173 700 to standardize mitigation bank creation. At the time of
writing, Washington State has twelve banks operating and five under review by the Department
of Ecology (DOE, 2012).
Both in lieu fee programs and banks allow regional planners and ecologists to coordinate
mitigation efforts that compensate for wetland losses most appropriately and address regional
needs (such as amphibian breeding habitat). Wetland mitigation banks have the added benefit of
using market forces to encourage mitigation success. The more traditional on site approach is
still widely used. On site projects are successful when sufficient knowledge about site hydrology
78

guides design, when goals and objectives are realistic, and when sufficient monitoring and
maintenance is undertaken (Ecology, 2006). Hopefully wetland mitigation can be harnessed to
establish new breeding sites in appropriate places.
All land development codes mandate the establishment of stormwater ponds and
amphibians have been shown to utilize them for breeding (Ostergaard et al. 2008). It is important
that such ponds have vegetation structures that facilitate egg mass establishment,that they have
hydroperiods that allow for larval development yet inhibit occupancy by fish and that surrounding
areas have sufficient levels of forest habitat and functional connectivity. It would be beneficial
for ecologists to identify existing stormwater ponds that, because of surrounding landscape, serve
as good breeding sites. Once this is known, they could be modified and managed. It would also
be beneficial for ecologists to work with transportation agencies to identify proposed stormwater
ponds that have potential for complementing large forests blocks as breeding sites. Such sites
could therefore be designed to encourage amphibian use.
Establish and Protect Upland Habitat
In landscapes that are near urban areas and that are subjected to conversion to suburban
or urban land use, cluster zoning should be enacted to consolidate housing. While cluster
developments are a good tool, it is best to leave conservation areas physically connected to other
natural areas as opposed to establishing ‘habitat islands’ in the middle of developed areas
(Abercrombie, 2004). Establishing conservation easements is an even better strategy because
planners can identify broad areas that are particularly good habitat attributes and designate them
as protection zones.
R. aurora are capable of crossing many land covers but the most preferable are either
mature or old growth forests (Haggard, 2000; Chan-McCleod, 2004). National forests and
national parks occupy interior areas of the Cascade and Olympic Ranges and this includes areas
79

that are low enough to be considered R. aurora habitat. Blocks of state forest lands are distributed
in lower elevations of the major mountain ranges.
Encouraging large interconnected blocks of mature forests, leaving islands of trees in
clear cuts, and attempting to keep clearcuts smaller and maintaining connectivity between forest
and appropriate breeding wetlands would all greatly improve habitat for R. aurora on timber
lands. Forest lands in the lower elevations of the Cascade Ranges, Olympic Range and most of
the Willipa Hills and Black Hills are generally distributed between large timber producers and
small forest land owners. There is little that can be done to encourage or mandate R. aurora
habitat conservation measures among private foresters, largely because R. aurora is not
threatened or endangered under the Endangered Species Act.

80

Literature Cited
Adams, M. J. (1999). Correlated factors in amphibian decline: exotic species and habitat
change in western Washington. Journal of Wildlife Management, 63, 1162-1171.
Adams, M. J., Bury, R. B., & Swarts, S. A. (1998). Amphibians of the Fort Lewis
Military Reservation, Wasshington; sampling techniques and community patterns.
Northwestern Naturalist, 79, 12-18.
Adams, M. J., West, S. D., & Kalmach, L. (1999). Amphibian and reptile surveys of the
U.S. Navy lands on the Kitsap and Toandos Peninsulas, Washington.
Northwestern Naturalist, 80, 1-7.
Adriaensen, F., Chardon, J. P., De Blust, G., Swinnon, E., Villalba, S., Gulinch, H., &
Matthysen, E. (2003
The application of 'least-cost' modelling as a functional landscape model. Landscape and
Urban Planning, 64, 14.
Alberti, M., Weeks, R., & Coe, S. (2004). Urban land-cover change analysis in central
Puget Sound. Photogrammetric Engineering & Remote Sensing, 70(9), 10431052.
Aubry, K. B. (2000). Amphibians in managed, second-growth Douglas-fir forests. The
Journal of Wildlife Management, 64(4), 1014-1052.
Aubry, K. B., & Hall, P. (1991). Terrestrial amphibian communities in the southern
Washington Cascade Range. (PNW-GTR-285).
Baldwin, R. F., Calhoun, A. J. K., & de Maynadier, P. G. (2006). Conservation planning
for amphibian species with complex habitat requirments: a case study using
movements and habitat selection of the wood frog Rana sylvatica. Journal of
Herpetology, 40(4), 442-453.
Baldwin, R. F., & de Maynadier, P. G. (2009). Assessing threats to pool-breeding
amphibian habitat in an urbanizing landscape. Biological Conservation, 142,
1628-1638.
Barnett, H. and S. Nickelson (2008) 14 Lakes 2008 Egg Mass Surveys. Seattle Public
Utilities. Available online at: J:\SSW\WS541\Public\IDTeams\Monitoring\HCP Effectiveness
Monitoring\Other Species\Amphibs\14 Lakes Egg Mass Survey Report.doc

Blaustein, A. R., Romansic, J. M., Scheessele, E. A., Han, B. A., Pessier, A. P., &
Longcore, J. E. (2005). Interspecific variation in susceptibility of frog tadbpoles
to pathogenic fungus Batrachochytrium dendrobatidis. Conservation Biology,
81

19(5), 1460-1468.
Brown, C. J., Blossey, B., Maerz, J. C., & Joule, S. J. (2006). Invassive plant and
experimental venue affect tadpole performance. Biological Invassions, 8, 327338.
Brown, H. (1975). Reproduction and development of the red-legged frog, Rana aurora, in
Northwestern Washington Northwest Science, 49(4), 241-252.
Bury, R. B., Aubry, K. B., & Corn, P. S. (1991). Regional patterns of terrestrial
amphibian communities in Oregon and Washington. (PNW-GTR-285).
Calef, G. W. (1973). Spatial distribution and effective breeding population of red legged
frogs in Marion Lake, British Columbia. Canadian Field Naturalist, 87, 279-284.
Chan-McCleod, A. C. A. (2003). Factors affecting the permeability of clearcuts to redlegged frogs. Journal of Wildlife Management, 67(4), 663-670.
Chan-McCleod, A. C. A., & Moy, A. (2006). Evaluating residual tree patches as stepping
stones and short-term refugia for red-legged frogs. The Journal of Wildlife
Management, 71(6), 1836-1844
Colburn, T. D., Dumanoski, D., & Myers, J. P. (1996). Our Stolen Future: Are We
Threatening Our Fertility, Intelligence and Survival? New York, NY: Penguin
Cook, D. G., & Jennings, M. R. (2007). Microhabitat use of the California red-legged
frog and introduced bullfrog in a seasonal marsh. Herpetologica, 63(4), 430-440.
Corn, P. S., & Bury, R. B. (1991). Terrestrial amphibian communities in the Oregon
Coast Range. (PNW-GTR-285).
Crooks, K. R., & Sanjayan, M. (Eds.). (2006). Connectivity conservation: maintaining
connections for nature. Cambridge, England: Cambridge University Press.
Croteau, M. C., Gibson, J., Gallant, N., & Lean, D. (2008). Toxicological threats to
herpetofauna in urban and suburban environments. In J. C. Mitchell & R. E. Jung
Brown (Eds.), Urban Herpetology (Vol. 3). Salt Lake City, UT: Society for the
Study of Amphibians and Reptiles.
Crump, M. L. (2010). Amphibian diversity and life history. In C. K. J. Dodd (Ed.),
Amphibian Ecology and Conservation: A Handbook of Techniques (pp. 3-17).
Oxford, England: Oxford University Press.
Cushman, S. A. (2006). Effects of habitat loss and fragmentation on amphibians: a
review and prospectus. Biological Conservation, 128, 231-240.

82

Davic, R. D., & Welsh, H. H., Jr. (2004). On the ecological roles of salamanders
Annual Review of Ecology, Evoluton and Systematics 35, 405-435.
de Solla, S. R., Pettit, K. E., Bishop, C. E., Cheng, K. M., & Elliott, J. E. (2002). Effects
of agricultural runoff on native amphibians in the Lower Fraser River Valley,
British Columbia, Canada. Environmental Toxicology and Chemistry, 21(2), 353360.
Dodd, C. K. J. (Ed.). (2010). Conservation and management. New York: Oxford
University Press.
Dodd, C. K. J., & Cade, B. S. (1998). Movement patterns and the conservation of
amphibians breeding in small, temporary wetlands. Conservation Biology, 12(2),
331-339.
Dumas, P. C. (1966). Studies of the Rana species complex in the pacific northwest.
Copeia, 1966(1), 60-74.
F.C., P., & Miller, J. R. (2008). Habitat and landscape characteristics underlying anuran
community structure along an urban-rural gradient. Ecological Applications,
18(2), 1107-1118.
Fagan, W. F., & Calabrese, J. M. (Eds.). (2006). Quantifying connectivity: balancing
metric performance with data requirements. Cambridgel, England: Cambridge
University Press.
Fahrig, L. (2003). Effects of Habitat Fragmentation on Biodiversity. Annual Review of
Ecological and Evolutionary Systematics, 34, 487-515.
Fahrig, L., Pedlar, J. H., Shealagh, E. P., Taylor, P. D., & Wegner, J. F. (1995). Effect of
road traffic on amphibian density. Biological Conservation, 73, 177-182.
Frankham, R. (Ed.). (2006). Genetics and landscape connectivity. Cambridge, England:
Cambridge University Press.
Garman, S.L., Swanson, F.J. & Spies, T.A. Past, present and future landscape patterns in
the Douglas Fir region of the Pacific Northwest. In Forest Fragmentation:
Wildlife and management Implications. Eds. J. A. Rochelle, Lehmann, L,
Wisniewski, J., 1999. Brill Academic Publishers, Leiden, Netherlands.
Gibbs, J. P. (1998). Distribution of woodland amphibians along a forest fragmentation
gradient. Landscape Ecology, 13, 263-268.
Gibbs, J. P., & Shriver, W. G. (2005). Can road mortality limit populations of poolbreeding amphibians? Wetlands Ecology and Management 13, 281-289.
Gilbert, F. F., & Allwine, R. (1991). Terrestrial amphibians of the Oregon Cascade
83

Range. (PNW-GTR-285).
Goodman, B.L. & Fahrig, L. (20020 How does landscape structure influence landscape
connectivity? Oikos, 99, 552-570.
Groom, M. J., Meffe, G. K., & Carroll, C. R. (2006). Principles of conservation biology
(Vol. 3rd). Sunderland, MA: Sinauer Associates, Incorporated.
Hagar, J. C. (2007). Wildlife species associated with non-coniferous vegetation in the
Pacific Northwest conifer forests: a review Forest Ecology and Management
246, 108-122.
Halliday, T. (Ed.). (2005). Diverse pheonomena influencing amphibian population
declines. London, England: University of California Press.
Hamer, A. J., & McDonnell, M. J. (2010). The response of herpetofauna to urbanization:
infering patterns of persistence from wildlife databases. Austral Ecology, 35, 568580.
Hanski, I., Mononen, T., & Ovaskainen, O. (2011). Eco-evolutionary metapopulation
dynamics and the spatial scale of adaptation. American Naturalist, 177(1), 29-43.
Harper, E. B., Rittenhouse, T. A. G., & Semlitsch, R. D. (2008). Demographic
consequenses for terrestrial habitat loss for pool-breeding amphibians: predicting
extinction risks associatated with inadequate size of buffer zones. Conservation
Biology, 22(5), 105-1215.
Hayes, M. P., Quinn, T., Richter, K. O., Schuett-Hames, J. P., & Serra Shean, J. T. (Eds.).
(2008). Maintaining lentic-breeding amphibians in urbanizing landscapes: the
case study of the northern red-legged frog (Rana aurora): Society for the
Conservation of Amphibians and Reptiles.
Hayes, M.P., Pearl, C.A. & Rombough, C.J (2001). Rana aurora aurora movement.
Herpetological Review 32, 1, 35-36.
Hayes, T., Haston, K., Tsui, M., Hoang, A., Haeffele, C., & Vonk, A. (2002).
Feminization of male frogs in the wild. Nature, 419 895-896.
Hayes, T. B., Collins, A., Lee, M., Mendoza, M., Noriega, N., Stuart, A., & Vonk, A.
(2002a). Hermaphoditic, demasculinized frogs after exposure to the herbicide
atrazine at low ecologically relevant doses. Procedings of the National Academy
of Sciences, 99(5467-5480).
Hayes, T. B., Haston, K., Tsui, M., Hoang, A., Haeffele, C., & Vonk, A. (2002b).
Femination of male frogs in the wild. Nature, 419, 895-896

84

Hels, T., & Buchwald, E. (2001). The effect of road kills on amphibian populations.
Biological Conservation, 99, 331-340.
Hitchings, S.P. & Beebee, T.J.C (1997). Genetic substructuring as a result of barriers to
gene flow in urban Rana temporaria (common frog) populations: implications for
biodiversity conservation. Heredity, 79 117-127.
Homan, R. N., B.S. Windmiller, B. S., & Reed, J. M. (2004). Critical thresholds
associated with habitat loss for two vernal pool-breeding amphibians. Ecological
Applications, 14(5), 1547-1553.
Joly, P., Miaud, C., Lehmann, A., & Grolet, O. (2001). Habitat matrix effects on pond
occupancy in newts. Conservation Biology, 15(1), 239-248.
K.O. Richter and Azous, A. (1995). Amphibian occurence and wetland characteristics in
the Puget Sound Basin. Wetlands, 15(3), 7
Kiesecker, J. M., & Blaustein, A. R. (1998). Effects of introduced bullfrogs and
smallmouth bass on microhabitat use, growth, and survival of native red-legged
frogs (Rana aurora). Conservation Biology, 12(4), 776-787.
Kruckeberg, A. R. (1991). The Natural History of Puget Sound Country. Seattle and
London: University of Washington Press.
Lannoo, M. (Ed.). (2005). Amphibian Declines: The Conservation Status in the United
States. London, England: University of California Press.
Lehtinen, R. M., Galatowitsch, S. M., & Tester, J. R. (1999). Consequences of habitat
loss and fragmentation for wetland amphibian assemblages. Wetlands, 19(1), 112.
Luck, G. W., Daily, G. C., & Erlich, P. R. (2003). Population diversity and ecosystem
services. Trends in Ecology and Evolution, 18, 331-336.
Marsh, D. M., & Trenham, P. C. (2001). Metapopulation dynamics and amphibian
conservation Conservation Biology, 15(1), 40-49.
Martin, K. J., & McComb, B. C. (2003). Amphibian habitat associations at patch and
landscape scales in the central Oregon Coast Range. Journal of Wildlife
Management 67(4), 672-683.
Mathias, M. (2008). Assessing functional connectivity for the northern red-legged frog
(Rana aurora) in an urbanizing landscape. Master's Thesis, University of
Washington, Seattle, Washington.
Moilanen, A., & Hanski, I. (Eds.). (2006). Connectivity and metapopulation dynamics in
highly fragmented landscapes. Cambridge, England: Cambridge University Press.
85

Nussbaum, R. A., Brodie, E. D., & Storm, R. M. (1983). Amphibians and reptiles of the
Pacific Northwest. Moscow, ID: University of Idaho Press.
Ostergaard, E. C., Richter, K. O., & West, S. D. (Eds.). (2008). Amphibian use of
stormwater ponds in the Puget Sound Lowlands (Vol. 3). Salt Lake City, Utah:
Society for the Study of Amphibians and Reptiles.
Paton, P. W. C., & Harris, R. N. (2010). Egg mass and nest counts. In C. K. J. Dodd
(Ed.), Amphibian Ecology and Conservation (pp. 143-166). Oxford, England:
Oxford University Press.
Pearl, C. A. (Ed.). (2005). Rana aurora (Baird and Girard, 1852[b]) northern red-legged
frog. London, England: University of California Press.
Petranka, J.W. and Holbrook, C.T. (2006). Wetland restoration for amphibians: should
local sites be designed to support metapopulations or patchy populations?
Restoration Ecology, 14, 3, 404-411.
Popescu, V. D., & Gibbs, J. P. (Eds.). (2010). Landscape ecology and GIS methods. New
York: Oxford University Press.
Pough, F. H., Andrews, R. M., Cadle, J. E., Crump, M. L., Savitzky, A. H., & Wells, K.
D. (2004). Herpetology (3rd ed.). Upper Saddle River, New Jersey: Prentice Hall.
Ray, N., Lehmann, A., & Joly, P. (2002). Modelling spatial distribution of amphibian
populations: a GIS approach based on habitat matrix permeability. Biodiversity
and Conservation, 11, 2143-2165.
Reh, W., & Seitz, A. (1990). The influence of land use on the genetic structure of
populations of the common frog Rana temporaria. Biological Conservation, 54,
239-249.
Reinelt, L., Horner, R., & Azous, A. (1998). Impacts of urbanization on paulustrine
(depressional freshwater) wetlands-research and management in the Puget Sound
region. Urban Ecosystems, 2, 219-236.
Richter, K. O., & Azous, A. (1995). Amphibian occurence and wetland characteristics in
the Puget Sound Basin. Wetlands, 15, 306-312.
Richter, K. O., Kerr, D. O., & Blessing, B. J. (Eds.). (2008). Buffer-only wetland
protection: implications for pond-breeding amphibians
(Vol. 3). Salt Lake City, UT: Society for the Study of Amphibian and Reptiles.
Ritson, P. I., & Hayes, M. P. (2000). Late season activity and overwintering in the
northern red-legged frog (Rana aurora aurora). Final report to the US Fish and
86

Wildlife Service, . Portland,OR.
Rubbo, M. J., & Kiesecker, J. M. (2005). Amphibian breeding distribution in an
urbanized landscape. Conservation Biology, 19(2), 504-511.
Schuett-Hames, J. P. (2004). Northern red-legged frog (Rana aurora aurora) terrestrial
habitat use in the Puget Lowlands of Washington. Master's Degree, The
Evergreen State College, Olympia, Washington.
Schuett-Hames, J. P., Neatherlinn, E. A., Azerrad, J. M., Hicks, T. L., Jacobson, J. E.,
Tirhi, M. J., & Wilhere, G. F. (2006). Lanscape planning for Washington State's
wildlife: managing for biodiversity in developing areas. Northwestern Naturalist,
87(2), 186.
Semlitsch, R. D. (2008). Differentiating migration and dispersal processes for pondbreeding amphibians. Journal of Wildlife Management, 72(1), 260-267.
Semlitsch, R. D. (2000). Principles for management of aquatic-breeding amphibians.
Journal of Wildlife Management, 64(3), 615-631.

Shaffer, T. P. C. a. H. B. (2005). Amphibian upland habitat use and its consequences for
population viability. Ecological Applications, 15(4), 10.
Shean, J. T. S. (2002). Post-Breeding Movements and Habitat Movements by the
Northern Red-legged Frog, Rana aurora aurora, at Dempsey Creek, Thurston
County, Washington. Master's Thesis, The Evergreen State College, Olympia,
Washington.
Shuytema, G. S., & Nebeker, A. V. (1999). Effects of ammonium nitrate, sodium nitrate
and urea on red-legged frogs, Pacific treefrogs and African clawed frogs. Bulletin
of Environmental Contamination and Toxicology, 63, 357-364.
Skelly, D.K. & Golon, J. (2003) Assimilation of natural benthic substrates by 2 species
of tadpoles. Herpetelogica 59 (1) 37-42.
Skidds, D. E., Goulet, F. C., Paton, P. W., & Mitchell, J. C. (2007). Habitat correlates of
reproductive effort in wood frogs and spotted salamanders in an urbanizing
watershed. Journal of Herpetology, 14(3), 439-450.
Sparling, D. W., Richter, K. O., Calhoun, A., & Micacchion, M. (2001). Methods for
evaluating wetland condition: using amphibians in bioassessments of wetlands.
(EPA 822-R-01-0071). Washington D.C.
Spyreas, G. W., B.W., Plocher, A.E., Ketzer, D.M., Mathews, J., Ellis, J., Heske, E.
(2010). Biological consequenseces of invasion by reed canarygrass (Phalaris
87

arundinacea). Biological Invassions, 12, 1253-1267.
Stevens, V., É., L., Wesselingh, R. A., & Baguette, M. (2006). Quantifying functional
connectivity: experimental assessment of boundary permeability for the natterjack
toad (Bufo calamita). Oecologia, 150, 161-171.
Storm, R. (1960). Notes on the breeding biology of the red-legged frog (Rana aurora
aurora). Herpetologica, 16(4), 251-259.
Swartwout, D. J. (1982). An evaluation of the National Wetlands Inventory in
Massachusetts. Master of Science, University of Massachusetts, Amherst,
Massachusetts.
Talley, D. M., Huxel, G. R., & Holyoak, M. (Eds.). (2006). Connectivity of the landwater interface. Cambridge, England: Cambridge University Press.
Tavernia, B. G., & Reed, J. M. (2009). Spatial extent and habitat context influences the
nature and strength of relationships between urbanization measures. Landscape
and Urban Planning, 92, 47-52.
Taylor, P. D., Fahrig, L., & With, K. A. (Eds.). (2006). Landscape connectivity: a return
to the basics. Cambridge, England: Cambridge University Press.
Theobold, D. M., Hobbs, N. T., Bearly, T., Zack, J. A., Shenk, T., & Riebsame, W. E.
(2000). Incorporating biological information in local land-use decision making:
designing a system for conservation planning. Landscape Ecology, 15, 35-45.
Theobold, D. M., T., S., Kline, J., Maxwell, B., Hobbs, N. T., & Dale, V. H. (2005)
Ecological support for rural land-use planning. Ecological Applications, 5(6),
1906-1914.
Thom, R. M., Borde, A. B., Richter, K. O., & Hibler, L. F. (Eds.). (2001). Land Use and
Watersheds: Human Influence on Hydrology and Geomorphology in Urban and
Forest Areas (Vol. 2). Washington D.C.: American Geophysical Union.
Todd, B. D., Luhring, T. M., Rothermel, B. B., & Gibbons, J. W. (2009). Effects of forest
removal on amphibian migrations: implications for habitat and landscape
connectivity. Journal of Applied Ecology 46, 554-561.
Trenham, P. C., Koenig, W. D., Mossman, M. J., Stark, S. L., & Jagger, L. A. (2003).
Regional dynamics of wetland-breeding frogs and toads: turnover and synchrony.
Ecological Applications, 13(6), 10.
Trenham, P. C., & Shaffer, H. B. (2005). Amphibian upland habitat use and its
consequenses for population viability. Ecological Applications, 15(4), 1158-1168.

88

USFWS. (2012). Species Profile- Northern Red-Legged Frog (Rana aurora aurora)
Retrieved May 11, 2012, 2012, from
http://ecos.fws.gov/speciesProfile/profile/speciesProfile.action?spcode=D02I
Vos, C.C.Antonisse-DeJong, A,G. Goedarts, P.W, & Smulders, M.J.M. (2001).Genetic
similarity as a measure for connectivity between fragmented populations of the
moor frog, (Rana arvalis). Heredity, 86 598-608.

Safner, T., Miaud, C., Gaggioti, O., Decout, S. Rioux, D., Zundel, S. & Munel, S. (2001)
Combining demography and genetic analysis to assess the population structure of
amphibians in a human dominated landscape. Conservation Genetics. 12, 161173.
Stevens, V. & Baguette, M. (2008). Importance of habitat quality and landscape
connectivity for the persistence of endangered natterjack toads. Conservation
Biology, 22(5), 1194-1204.
Walston, L.J. & Mullin, S.J. (2008). Variation in amount of surrounding forest habitat
influences the initial orientation of junile amphibians emigrating from breeding
ponds. Canadian Journal of Zoology, 86, 141-146.
Werner, E. E., Yurewicz, K. L., Skelly, D. K., & Relyea, R. A. (2007). Turnover in an
amphibian metacommunity: the role of local and regional factors. Oikos, 116,
1713-1725.
White, R. The Altered Landscape: Social Change and the Land in the Pacific Northwest.
in Regionalism and the Pacific Northwest, eds. William G. Robbins, Robert J.
Frank, and Richard E. Ross (Corvalis Oregon State University Press, 1983), p.
111.
Weyrauch, S. L., & Grubb, T. C. J. (2004). Patch and landscape characteristics associated
with the distribution of woodland amphibians in an agricultural fragmented
landscape: an information-theoretic approach. Biological Conservation, 115, 443450.
Wiens, J.A., Schooley, R.I. & Weeks, R.D.Jr.. (1997). Patchy landscapes and animal
movements: do beetles percolate?.. Oikos, 78, 257-264.
With, K.E., Gardner, R.L. & Turner, M.G. (1997). Landscape connectivity and
population distributions in heterogeneous environments.. Oikos, 78, 151-169.

89

Appendix A
Land cover Class

Description

Friction Cost

Non-forested wetland

Wetlands associated with open water

5

Deciduous & Mixed Forest

>80% Deciduous Trees, or 10-90% each Deciduous and Coniferous Trees

5

Coniferous Forest

>80% Coniferous Trees

5

Small Open Water

Small lakes, small reservoirs, small streams

20

Regenerating Forest

Forest replanted after logging

20

Grass

Developed Grass and Grasslands

40

Clearcut Forest

Cleared forest without significant regrowth and very dry grass

40

Agriculture

Row Crops, Pastures

50

Cleared for Development

Cleared Land

50

Shoreline

Marine Shoreline

50

Light Intensity Urban

20-50% Impervious Area

60

Local Roads

Not designated as arterials

60

Medium Intensity Urban

50-80% Impervious Area

80

Collector Arterials

Collectors that serve very little through traffic and

80

serve a high proportion of the local traffic
Heavy Intensity Urban

>80% Impervious Area

Barrier:
Infinite

Large Open Water

Large lakes, large rivers

Barrier:
Infinite

High Traffic Roads

Freeways, Principal arterials, Minor arterials. Serve “ through traffic” and are busy
roads

Barrier:
Infinite

Table 7 Rana aurora Friction Values for Different Land Covers

90

Appendix B
Landowner Contact Letter Template
Chris Holcomb
-AddressOlympia, Washington 98502
December 28, 2010
[Name}
[Address}
Dear _______:
I am writing you because I would like to ask your permission to briefly enter your
property (on 123rd Avenue SE, Yelm) on a few days this winter and spring for the
purposes of ecological research. I am a student in the Masters of Environmental Studies
program at the Evergreen State College and am doing research on two of our native
amphibians, the Red Legged Frog and the Northwestern Salamander. I am studying the
breeding activity of these two animals on 30 wetlands throughout Thurston County so
that I can get some idea of how the level of development around the wetland affects
breeding. After having had worked in the wetland consulting field for years, I can attest
that if I were allowed on your land to conduct this research, it would not result in any
additional constraints on what you can do on your property.
Why your property? I am interested in including a wetland on your property in my
research because it is the appropriate size and type. I learned about your wetland by
studying the National Wetland Inventory website’s ‘Wetland Mapper’ feature. This
information was generated from satellite infrared photography. The Thurston County
Geodata website provided me with property ownership information.
What is involved in me coming on your property to look at your wetland? First, I
would like to visit the area to make sure that it is good habitat and appropriate to include
in my study. I was hoping to do a visit at whatever time is permissible for you from
January 22 through the 30th. If it seems like good habitat, I would like to visit again
anytime from March 26th to April 3rd to make sure that these two animals are actually
breeding in the wetland. I would ascertain this by wading through the wetland and
looking for their egg masses. It is possible that I or someone else would be interested in
visiting your property a few times in 2012 and actually counting the egg masses, but this
is uncertain at this point. Of course, I would be more than happy to make arrangements
with you on when would be a good times to visit. If these dates are not good, I can come
91

on other dates. I would be glad to meet you or if you would rather I can just come and do
what I have to do without bothering you. I can follow any important instructions that you
may have such as shutting livestock fences. If you are renting the land out, I would be
more than happy to communicate with tenants.
Who will get this information? The results of this study will be contained in my
master’s thesis and possibly a scientific journal article. Neither document will include
property ownership information, wetland categories, ratings, or buffer widths. My
research would not provide any additional information to government agencies about
your property: a master’s thesis is not a valid document for permit applications.
Amphibian activity does not affect wetland buffer widths. Finally, everyone already has
access to the websites that I mentioned, so the wetland is already known to the world.
I hope that you will grant me access to your property a few times this winter and spring.
Again, I would be glad to meet you, notify you in advance of my visit and follow any
special instructions. I would also be interested in any information that you have on
seasonal water levels in the wetland, land use history or amphibians that you have
observed. I can be reached by mail at the above address, by phone at [phone number] or
by e-mail at [email address]. I would really appreciate knowing your thoughts on this by
January 10th 2011, but feel free to contact me at your convenience.
Thank you very much for your consideration.
Sincerely,

Chris Holcomb
Evergreen MES student

92

93

Media

Holcomb_C-thesis2012.pdf