-
extracted text (extracttext:extracted_text)
-
THE OREGON SPOTTED FROG (RANA PRETIOSA) IN LOWLAND WESTERN
WASHINGTON, USA: A POPULATION, PARENTAGE, & NON-BREEDING
HABITAT ANALYSIS
by
Chelsea D. Waddell
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2015
©2015 by Chelsea D. Waddell. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Chelsea D. Waddell
has been approved for
The Evergreen State College
by
________________________
Dina Roberts, Ph.D.
Member of the Faculty
________________________
Marc P. Hayes, Ph.D.
Senior Research Scientist
Habitat Program
Washington Department of Fish & Wildlife
________________________
Date
ABSTRACT
The Oregon spotted frog (Rana pretiosa) in lowland western Washington, USA:
A population, parentage, & non-breeding habitat analysis
Chelsea D. Waddell
The at-risk Oregon spotted frog (OSF, Rana pretiosa) has disappeared from much of
its geographic range. It is endangered in Washington State, and considered threatened
under the US Endangered Species Act. Much research has been devoted to improving
OSF breeding habitat management. Despite these important efforts, adult non-breeding
habitat utilization in western Washington remains poorly known, and this is a significant
gap in our understanding of this species. In western Washington, many OSF populations
are small, genetically isolated, and embedded in a rapidly urbanizing matrix. Given these
habitat limitations, determining the total habitat footprint of each OSF population is
critical to their conservation. This study investigated the spatial relationship between
breeding and non-breeding habitat utilization patterns of adult OSFs by using genetic
sampling for one small population. This effort exploited the fact that a large fraction of
egg masses (n=109) laid in 2014 (February-March) at the target study site, West Rocky
Prairie (WRP), had already been genetically sampled. This effort sampled adult OSFs
genetically in their non-breeding active-season habitat (July- September), and linked
those adults (n=56) to breeding locations based on parentage of egg masses using
CERVUS 3.0.7. Straight-line distance measurements of parent:offspring pairs (n=12)
revealed that parents (n=2) traveled >2km and (n=1) >1km between breeding and nonbreeding habitat. Based on microsatellites (n=12), AR=3.833, and COLONY analysis,
N e =25 (CI95: 15 to 43), 54% of sampled adults had ≥1 sibling within the sampled
population, suggesting a recent bottlenecking for OSF at WRP. Most (83%) of all
captured frogs were found in a small (10×6m) pond at WRP, indicating that non-breeding
habitat may be limited. Management of OSF should consider all habitats that may
contribute to its vulnerability. This study provides critical information about OSFs at
WRP, and a basis of what to expect for non-breeding active-season habitat in other OSF
populations.
TABLE OF CONTENTS
LIST OF FIGURES .......................................................................................................... vii
LIST OF TABLES ............................................................................................................ ix
ACKNOWLEDGEMENTS ................................................................................................x
CHAPTER 1: INTRODUCTION THE OREGON SPOTTED FROG ..............................1
CURRENT STATUS OF THE OREGON SPOTTED FROG .......................................1
LISTING STATUS .................................................................................................2
HABITAT REQUIREMENTS & LIFE HISTORY ........................................................3
HABITAT THREATS & MANAGEMENT .................................................................11
STUDY RATIONALE ..................................................................................................12
CHAPTER 2: OREGON SPOTTED FROG HABITAT UTILIZATION .......................14
METHODS ....................................................................................................................14
FIELD METHODS ...............................................................................................14
WEST ROCKY PRAIRIE STUDY AREA ......................................................14
BREEDING HABITAT SURVEYS .................................................................18
NON-BREEDING HABITAT SURVEYS .......................................................19
Visual Encounter Surveys .........................................................................20
Animal Processing ....................................................................................24
Field Sampling for Genetic Material ........................................................29
Habitat Characteristic Measurements .......................................................30
Survey Area by Region & Organisms Observed ......................................30
ANALYSIS OF EFFORT METHODS ................................................................31
SPATIAL ANALYSIS METHODS .....................................................................32
RESULTS ......................................................................................................................34
EGG-MASS SURVEYS .......................................................................................34
NON-BREEDING HABITAT SURVEY AREA .................................................35
EFFORT ................................................................................................................37
ADULT LOCATIONS BY LAND COVER TYPE .............................................39
LAND COVER .....................................................................................................42
iv
ANIMAL PRESENCE ......................................................................................42
SURVEY AREAS .............................................................................................43
WEST SIDE SURVEY AREA .........................................................................45
POND SOUTH OF TILLEY POND .................................................................53
EAST SIDE SURVEY AREA ..........................................................................55
DORSAL PATTERN RECOGNITION ...............................................................57
CHAPTER 3: OREGON SPOTTED FROG GENETICS ................................................58
INTRODUCTION .........................................................................................................58
OVERVIEW OF POPULATION GENETICS & AMPHIBIAN DECLINE .......58
OREGON SPOTTED FROG GENETICS: CURRENT KNOWLEDGE ............60
APPLICATIONS OF POPULATION GENETICS ..............................................65
PARENTAGE ANALYSIS OVERVIEW ............................................................65
METHODS ....................................................................................................................67
FIELD METHODS ...............................................................................................67
SOURCES OF DNA .............................................................................................68
MICROSATELLITES (MICROSATELLITE LOCI) ......................................69
LABORATORY METHODS: MICROSATELLITES ........................................70
PARENTAGE ANALYSIS: METHODS USED .................................................75
EXCLUSION ....................................................................................................76
LIKELIHOOD ..................................................................................................77
CATEGORICAL ALLOCATION & CERVUS 3.0.7. ANALYSIS ................78
FRANz ANALYSIS & PARENTAL RECONSTRUCTION ............................82
SIBLING RELATIONSHIP (SIBSHIP) RECONSTRUCTION
& COLONY ANALYSIS ..........................................................................82
SPATIAL ANALYSIS .........................................................................................84
RESULTS ......................................................................................................................85
SAMPLED ADULTS & OFFSPRING: A COMPARISON OF DORSAL
PATTERN & MICROSATELLITE TOOLKIT .......................................85
EFFECTIVE POPULATION SIZE (N e ) ANALYSIS .........................................88
ALLELIC RICHNESS ASSESSMENT ...............................................................89
ADULT RELATEDNESS ....................................................................................92
v
PARENTAGE ASSINGMENT RESULTS ..........................................................93
SPATIAL ANALYSIS OF PARENT & OFFSPRING LOCATIONS ................96
CHAPTER 4: DISCUSSION ............................................................................................99
PARENT OFFSPRING LOCATIONS .........................................................................99
POPULATION STRUCTURE ...................................................................................101
EFFECTIVE POPULATION, ALLELIC RICHNESS & RELATABILITY ....101
HABITAT AVAILABILITY & POPULATION SIZE ......................................102
GENETIC MARKERS .......................................................................................104
LAND COVER CHARACTERISTICS ......................................................................105
LOCATIONS OF ADULTS .......................................................................................106
HABITAT VARIABLES & THEIR EFFECTS ON OREGON
SPOTTED FROG CAPTURE ............................................................................107
JUVENILE OREGON SPOTTED FROGS ................................................................109
DORSAL PATTERN RECOGNITION METHODS .................................................110
MANAGEMENT IMPLICATIONS & FUTURE RESEARCH ................................111
SPECIFIC APPLICATIONS TO THE OREGON SPOTTED
FROG RANGE-WIDE .......................................................................................113
LITERATURE CITED ...................................................................................................115
APPENDIX A: Standard English and Scientific Names for flora and
fauna found in the WRP Survey Area. ................................................................120
APPENDIX B: Field Survey Protocol for Buccal Swab Sampling ................................121
APPENDIX C: Mitochondrial DNA & Single Nucleotide Polymorphisms ..................122
APPENDIX D: Parentage Analyses: Unreported Methods Used
& Other Methods ................................................................................................124
vi
LIST OF FIGURES
FIGURE 1.1. Map: Current Distribution of Oregon spotted frog in Washington State ......2
FIGURE 1.2. Communal Egg Mass ...................................................................................5
FIGURE 1.3. Oviposition Habitat ......................................................................................6
FIGURE 2.1.1. Map: Washington State: West Rocky Prairie ..........................................15
FIGURE 2.1.2. Map: West Rocky Prairie Land Ownership..............................................16
FIGURE 2.1.3. Map: West Side Survey Area: Key Locations..........................................17
FIGURE 2.1.4. Walking Visual Encounter Surveys .........................................................20
FIGURE 2.1.5. Floating Visual Encounter Surveys .........................................................21
FIGURE 2.1.6. Floating Minnow Traps ...........................................................................22
FIGURE 2.1.7. Incidental Species: Minnow Traps ...........................................................24
FIGURE 2.1.8. Snout to Vent Measurements....................................................................25
FIGURE 2.1.9. Shank Measurements ...............................................................................26
FIGURE 2.1.10. Mass Measurements ..............................................................................26
FIGURE 2.1.11. Male & Female Identification ...............................................................27
FIGURE 2.1.12. Dorsal Pattern ........................................................................................28
FIGURE 2.1.13. Study Specific Photo Book ....................................................................29
FIGURE 2.2.1. Map: West Rocky Prairie Egg Mass Locations 2014 ...............................35
FIGURE 2.2.2. Map: Non-Breeding Habitat Survey Area West Rocky Prairie 2014.......36
FIGURE 2.2.3. Map: West & East Side Survey Area ......................................................37
FIGURE 2.2.4. Small Pond: West Side Survey Area .......................................................39
FIGURE 2.2.5. Map: Locations of Observed and Captured Adults .................................41
vii
FIGURE 2.2.6. Map: West Side Survey Area: Land Cover Type ....................................46
FIGURE 2.2.7. Water Level Changes in the Small Pond ..................................................47
FIGURE 2.2.8. East Channel ............................................................................................48
FIGURE 2.2.9. West Channel ...........................................................................................49
FIGURE 2.2.10. Clearing North of North Channel ..........................................................50
FIGURE 2.2.11. North Channel .......................................................................................51
FIGURE 2.2.12. Area West of West Channel ..................................................................52
FIGURE 2.2.13. Tilley Pond ............................................................................................53
FIGURE 2.2.14. Map: Land Cover Type: Pond South of Tilley Pond .............................54
FIGURE 2.2.15. Pond South of Tilley Pond ....................................................................55
FIGURE 2.2.16. Map: Land Cover Type: East Side Survey Area ...................................56
FIGURE 2.2.17. East Side Survey Area ...........................................................................57
FIGURE 3.1.1. Map: OSF geographic range & clades .....................................................61
FIGURE 3.1.2. Principle Coordinates Analysis: OSF Clades ..........................................62
FIGURE 3.3.1. Map: Locations of OSF Parents and their Assigned Offspring ...............98
viii
LIST OF TABLES
TABLE 1.1. Annual Variation in Oviposition Start & End Times .....................................4
TABLE 1.2. Frequency of OSF Breeding Habitat Locations .............................................7
TABLE 2.2.1. Summary of Effort per Adult Captured ....................................................38
TABLE 2.2.2. Summary of Effort per Adult Partitioned by Two-Hour
Time Intervals ...............................................................................................................38
TABLE 2.2.3. Land Cover Type Area & Number of Captured Adults Within Each .......43
TABLE 2.2.4. Survey Regions & Species Observed ........................................................44
TABLE 3.2.1. Microsatellite Markers Used .....................................................................72
TABLE 3.3.1. Matching Adults Based on Dorsal Pattern & Genotype ...........................87
TABLE 3.3.2. Effective Population Size of Sampled Population at WRP .......................88
TABLE 3.3.3. Number of Individuals, Loci Tested, Average Number of Alleles,
& Mean Expected He ....................................................................................................89
TABLE 3.3.4. Allelic Richness: Summary Table From CERVUS
Allelic Richness Output ................................................................................................91
TABLE 3.3.5. Sibling Relationships ................................................................................92
TABLE 3.3.6. Parentage Assignments .............................................................................94
TABLE 3.3.7. Straight-Line Distance Between Parents and Offspring ............................97
ix
ACKNOWLEDGEMENTS
I would like to begin by saying that this was a tremendous collaborative effort and in no
way could I have achieved this work on my own. These collaborations included the
Habitat program at Washington Departments of Fish & Wildlife, the Molecular Genetics
Laboratory at WDFW, faculty at The Evergreen State College, and many wonderful
volunteers.
I am extremely grateful for the opportunities these collaborations provided, as they
helped make this project a fantastic learning experience and a significant contribution to
the conservation of the Oregon spotted frog.
Thanks to Marc Hayes for his advisement, support, field assistance, and expertise.
Without his tremendous contributions, and guidance, this project would not have been
possible.
Thanks to my reader, Dina Roberts, who encouraged me to forge these collaborations and
guided me through the thesis writing and analysis process.
Thanks to everyone in the Molecular Genetics Laboratory at WDFW. In particular, I’d
like to thank the MGL Director, Ken Warheit, for allowing me to utilize the resources in
the MGL and for assisting me with the analysis. I’d also like to thank Maureen Small for
helping me with the analysis. Additionally, I’d like to thank Cherril Bowman for her
assistance and guidance in the laboratory.
Thanks to Julie Tyson for teaching me many of the field methods required for this study,
and for your wisdom and guidance throughout this project.
Many thanks to the incredible volunteers who came out and helped with the field
component of this study. If it wasn’t for all of you, I would not have been able to do this
project. In particular, I’d like to thank my brother, Cameron Smith, for spending weeks in
the field with me catching frogs; you really are the best field assistant. Many thanks to
Molly Stover Cox, Bonnie Blessing, Sierra Blakely, Sarah Davis, Alex Uhrich, Lynda
Manden, Nikhil Narkhede, and David Snyder for your time, and effort as volunteers.
To my loving husband, Alex Waddell, thank you for all of your support and
encouragement throughout this process. From my first class, early mornings in the field,
to my completion of this program, you have been by my side. Thank you.
Finally, with tremendous gratitude, I’d like to thank the Dr. Holly Reed Conservation
Fund of Point Defiance Zoo & Aquarium for your funding contributions to this work.
x
CHAPTER 1
INTRODUCTION
CURRENT STATUS OF THE OREGON SPOTTED FROG
The Oregon spotted frog (OSF, Rana pretiosa) is a highly aquatic ranid species,
endemic to the Pacific Northwest (USFWS, 2015). The OSF has a historical range from
Southwestern British Columbia to northern California (Hallock, 2013). However, due to
impacts from human development, the species distribution has declined precipitously.
The current distribution is much reduced, and historically widespread populations in the
Willamette Valley, Oregon, and areas in California have been extirpated. However,
isolated populations still exist in Oregon and Washington, and extend into southern
British Columbia (Cushman & Pearl, 2007). Based on conservative historical distribution
estimates, 79% of OSF populations have been lost; however, losses may actually reach
90% (Hayes, 1997). Although reduced, Washington has remaining populations located in
North Puget Sound, South Puget Sound, and western Klickitat County (Hallock, 2013)
(Figure 1.1).
1
Figure 1.1. Current distribution of Oregon spotted frog populations in Washington State.
Triangles do not represent single populations, but general areas of populations across the
OSF’s range in Washington State. Figure adapted from Hallock (2013).
LISTING STATUS
As of August 29th, 2014, the US Fish and Wildlife Service (USFWS) formally
listed the OSF (Rana pretiosa) as a Threatened Species (USFWS, 2015). However, the
OSF has been considered a State Endangered Species in Washington since 1997
(Hallock, 2013), and Endangered in Canada since 2000 (Haycock, 2000). These listings
limit activities that could be deemed harmful to this species (Hallock, 2013). The OSF is
classified in Oregon as a Critically Sensitive Species, and in California as a Species of
Special Concern.
2
HABITAT REQUIREMENTS, LIFE HISTORY, & MANAGEMENT
Dependent on the stage of its life cycle, the OSF has varying seasonal hydrologic
habitat requirements, which are partitioned into three temporal categories (Watson et al.,
2003). Breeding season occurs between February and early April, the non-breeding
season occurs between April and early October, and the overwintering season occurs
between October and February of each year (Watson et al., 2003; M. Hayes, unpublished
data). This life-history necessitates an understanding of what defines suitable habitat for
each life stage as the season progresses. During the late non-breeding season in
Washington, water availability may be limited due to lower amounts of precipitation
during that time of year. Given the fact that this species is fully aquatic during all phases
of its life cycle, multiple challenges may arise regarding its management if we do not
fully understand its requirements for suitable habitat.
BREEDING SEASON
During the breeding season, between February and early April, adult males and
females congregate in shallow, seasonal pools created by the seasonal expansion of their
permanent water habitat from rain and snowmelt (McAllister & Leonard, 1997). Oregon
spotted frogs require water depths less than 30 cm (average 18.5 cm) (Pearl & Adams,
2009) because these shallow waters warm quickly, which is important for embryonic
development (Licht, 1971). OSF also prefers shallow waters with emergent and
submerged “vegetation types (which) provide feeding areas, refuge from predators, and
warmer water” (Pearl & Adams, 2009). The breeding season lasts approximately four
weeks and the seasonal shallow habitat serves as oviposition (egg laying) sites
3
(McAllister & Leonard, 1997). Table 1.1 highlights the annual variation in oviposition
times of OSF populations in British Columbia and Washington State across multiple
years.
Table 1.1. Annual variation in oviposition start and end dates times, which has been
linked to water temperature. 1 Depths & 2 Temperatures measured within 48 hours of
egg-laying. _3 Information from Licht (1971) where egg laying began at 6°C at the center
of breeding ponds and 20.7°C adjacent to egg masses.
Table adapted from McAllister & Leonard (1997).
Breeding females in lowland sites breed every year (Licht, 1974), they lay a
single egg clutch (mass) per year, and a single male generally fertilizes a single clutch
(Phillipsen et al., 2009). In McAllister & Leonard (1997) adult females laid egg masses
with an average of 643 eggs per mass in communal clusters of 10-75 masses, although
higher numbers of masses (>100) have been reported (Tyson & Hayes, 2014). Figure 1.2
illustrates a communal egg mass.
4
Figure 1.2. Oregon spotted frogs communal egg mass. This mass includes >50 Oregon
spotted frog egg clusters (blue circle) with a single clutch (red circle). Figure adapted
from Kapust et al. (2012).
Oregon spotted frogs frequently breed in the same geographic locations each year,
and depending on topography and seasonal water variation, will sometimes use the same
oviposition site each year (Kapust et al., 2012; Watson et al., 2003). Figure 1.3 shows
breeding habitat for the OSF.
5
Figure 1.3. High quality oviposition habitat for the OSF includes a shallow, seasonally
flooded wetland, where Reed Canary Grass is infrequent. Figure adapted from Hallock
(2013).
Oregon spotted frog eggs develop between 14 and 30 days, the eggs hatch and
tadpoles find open water in order to consume bacteria, algae and detritus (McAllister &
Leonard, 1997). After 13 to 16 weeks, the OSF tadpoles metamorphose into juvenile
frogs (McAllister & Leonard, 1997). In some OSF sites, juvenile frogs move into ponds
along with the adult frogs for the summer months (Hallock, 2013); however, they have
also been known to use shallower habitats (M. Hayes, personal communication).
OVERWINTERING SEASON
Overwintering for this species is considered the time between October and
February of each year, based on studies conducted on overwintering characteristics at
Conboy Lake National Wildlife Refuge (Hayes et al., 2001) and Trout Lake (Hallock &
6
Pearson, 2001) in Eastern Washington; and Dempsey Creek in western Washington
(Risenhoover et al., 2001).
In a study of 11 adult female radio-telemetered OSFs at Dempsey Creek in
lowland western Washington, 5 adults remained relatively active, while 6 others were
generally sedentary, with a total average movement of 6.7 m per day (Risenhoover et al.,
2001). Ninety-five percent of the habitat utilized by OSFs in this study was palustrine
wetland, with varying levels of vegetation and cover Table 1.2 (Risenhoover et al., 2001).
For 90% of the observations made in this study, ice was not encountered (Risenhoover et
al., 2001).
Table 1.2. Frequency of OSF habitat locations during overwintering in lowland western
Washington
Table adapted from Risenhoover et al. (2001)
In comparison, a study of overwintering habitat was conducted at Conboy NWR
(Hayes et al., 2001), which is a higher elevation (550-561 m [1,804-1,840 ft.]) OSF site
(Hallock, 2013). In this study, 10 individual male and female adult OSFs were pit tagged,
ice was prevalent, and significantly more movement was observed before ice occurred
7
(Hayes et al., 2001). Furthermore, frogs were found to utilize several vegetation types
including floating vegetation, upland vegetation, and OSFs were observed in open water
or ice (Hayes et al., 2001).
Based on the results of Risenhoover et al. (2001), in lowland western Washington,
it appears that temperature may play an important role in the movement of OSF across
the landscape, although these results were not significant. Hayes et al. (2001) did find a
significant difference in movement where frogs tended to move more during pre-ice
conditions than during icy conditions. These results warrant further investigation into the
movement and habitat utilization during the overwintering period. Furthermore, both of
these studies had small sample sizes (10-11 individuals) and may not adequately
represent the overwintering habitat utilization of the OSF.
NON-BREEDING SEASON & MANAGEMENT
Multiple studies, some using radio telemetry, have described the home ranges and
habitat utilization of various OSF adult populations during the breeding and
overwintering seasons. They also demonstrated what habitat regions individual
populations are utilizing at metamorphosed and juvenile life stages. However, adult nonbreeding habitat requirements remain widely unknown, as much of the research into
Oregon spotted frog habitat utilization has focused on breeding. This emphasis on
breeding habitat has also been the primary focus of management objectives for the
species in western Washington. However, breeding only occurs during the late
winter/early spring months (Watson et al., 2003), which leaves the rest of the OSF annual
cycle less understood.
8
In an attempt to understand home range of the OSF throughout the year, Watson,
et al. (2003) attached radio telemetry devices to individual frogs at the previously
discussed Dempsey creek, a single site in lowland western Washington. Watson et al.
(2003) tracked a total of 60 adult OSF at varying times intervals between 1997-1999.
However, only 18 of these individual frogs were tracked during the non-breeding, active
dry season. This work revealed that OSFs move to small, deep remnant pools during the
dry season in June-August (Watson et al., 2003). During the dry season, these wetlands
often decrease in area or dry out due to sun exposure and reduced precipitation, and
remnant pools are typically some of the few remaining aquatic habitats. These results
indicate that the home range of this species drastically decreases (2 to 4 times) during the
dry summer months (Watson et al., 2003; Hallock, 2013).
This study was valuable in indicating locations where OSF reside during the nonbreeding season. However, these results may not adequately represent the non-breeding
habitat utilization of the OSF across its geographic range, as it was conducted at a single
study site. Furthermore, given the highly aquatic nature of the OSF, it is possible that
these remnant pools may be a limiting factor as related to habitat requirements for this
species, a possibility that warrants further investigation. Research on OSF adult nonbreeding, active season habitat is lacking for multiple sites and populations, and warrants
further investigation. This thesis study will fill some of these gaps by characterizing the
habitat used by adults during the non-breeding summer season at a different site in
western Washington.
9
PRIMARY HABITAT REQUIREMENTS & MANAGEMENT
Remnant OSF populations in Washington typically require palustrine wetlands,
which are connected to stream networks (Hallock, 2013). “The perennial creeks and
associated network of intermittent tributaries provide aquatic connectivity between
breeding sites, active season habitat and overwintering habitat” (Hallock, 2013). These
systems also provide a constant flow of oxygenated water. This may be especially
important for the species during hot summer months where water tends to become
stagnant (Hallock, 2013). The wetlands where OSFs live include a mix of aquatic bed,
emergent, scrub-shrub, and forested areas (Hallock, 2013). The dynamic habitat
requirements of the OSF, and the heavy alteration to the majority of sites make
management of this species and their habitat difficult. “Proper management of the
remaining isolated frog populations requires site-specific knowledge of vegetation
characteristics, home range, and seasonal changes in hydrology that may affect
movements” (Watson et al., 2003). In order to successfully manage this species, it is vital
that its habitat-range requirements are understood, and that management is adapted
accordingly.
10
HABITAT THREATS & MANAGEMENT
A number of factors contribute to the decline in OSF populations, many of which
are caused by habitat alterations from anthropogenic effects. Isolation, low effective
population sizes, exotic flora and fauna, altered landscape and hydrology, and changes in
water chemistry are just a few contributors (Watson et al., 2003; Hallock, 2013). The
primary concern, which will likely produce the best results for this species’ population, is
in preserving and restoring quality habitat. Since resilient habitats and ecosystems tend to
have resilient species populations, it is important to focus on preserving and restoring
quality habitat for the benefit of all species in OSF associated wetlands.
Reed Canary Grass has greatly altered the majority of primary habitat for this
species, and has become a primary focus in breeding habitat restoration for the OSF in
Washington State (Kapust et al., 2012). Introductions of invasive American Bullfrogs in
OSF habitat have altered the predatory dynamics concerning the OSF, and have
contributed to their decline (Pearl et al., 2004). Finally, contaminant levels and low water
conductivity in their habitat may have contributed to high mortality in both embryonic
and adult life phases (Marco et al., 1999).
While these issues have major management implications, restoration of OSF
habitat must first be guided by well-researched documentation of what habitats OSFs are
using. Furthermore, before altering their habitat for restoration, each population’s genetic
health must be assessed.
11
STUDY RATIONALE
Research on OSF genetics has provided information about genetic variation across
their range (Blouin et al., 2010), setting the stage for further analyses. That research was
range-wide, but OSF management is often focused on small isolated populations.
Therefore, conservation requires site-specific understanding of OSF habitat, population
genetics, and local landscape processes. Understanding habitat utilization patterns and
population structure is critical to developing appropriate strategies to manage these small
populations.
The feasibility with which OSF egg masses can be detected has led to a focus on
breeding surveys to gain general knowledge about OSF population sizes, and track
trends. As a consequence, in Washington State, the nature of breeding habitat is
reasonably well understood. However, the habitats where OSFs breed are largely
ephemeral, as in lowland western Washington, they breed at high water during late
winter. When the water recedes from breeding sites, adult OSFs are thought to move to
different habitats for the non-breeding season. Furthermore, since adult OSFs are often
cryptic (Hallock, 2013), the nature of non-breeding habitat and their linkage to breeding
sites is largely unknown for most sites in western Washington.
This study examines the non-breeding habitat utilization of OSF adults during the
2014 active season (summer to early fall), by using genetics to spatially link these adults
to eggs laid earlier in 2014 at specific oviposition sites (breeding habitat) at West Rocky
Prairie (WRP). WRP is a known, Washington Department of Fish and Wildlife (WDFW)
managed, OSF site in the upper Black River drainage in western Washington. Genetic
12
data was used to infer parentage and characterize this population. Both the genetic and
habitat components of this study will help guide management activities at WRP and other
sites with similar habitat characteristics. In particular, information obtained from this
study characterizes OSF non-breeding active-season habitat in a manner useful to habitat
managers responsible for ensuring survival of this population in the long-term. If nonbreeding active season habitat somehow limits the OSF population size at WRP, this
research can inform the appropriate direction for management of this population, and
potentially other populations. Furthermore, linking seasonal habitats utilized via
parentage analysis can identify the habitat footprint of the biologically effective, rather
than the total OSF population at the site. This study includes two key components,
breeding and non-breeding habitat utilization across temporal and spatial scales, and
genetic linkage between those habitats.
13
CHAPTER 2
HABITAT UTILIZATION
METHODS
FIELD METHODS
WEST ROCKY PRAIRIE STUDY AREA
West Rocky Prairie (WRP) is located in Thurston County in lowland western
Washington (Figure 2.1.1), and includes a wetland complex with two areas of focus,
hereafter referred to as East Side and West Side marshes (Tyson & Hayes, 2014). WRP, a
verified OSF site since 1999, has been under state ownership (WDFW) since 2006, and is
one of approximately 50 locations where the OSF resides across its geographic range
(Tyson & Hayes, 2014). The site has been the focus of numerous studies for the OSF,
including a series of controlled studies on the response of OSF oviposition to mowing of
the invasive Reed Canary Grass (Kapust et al., 2012; Tyson & Hayes, 2014).
14
Figure 2.1.1. Map of Washington, and the location of West Rocky Prairie. Headwaters
include Allen Creek (West Side) and Beaver Creek (East Side), which are connected to
the Black River. Red boxes indicate the geographic locations of the West Area and East
Area where surveys in this study were conducted. Figure adapted from Tyson & Hayes
(2014).
West Rocky Prairie Wildlife Area encompasses multiple habitat types including
prairie, forested area, and wetlands. Figure 2.1.2 includes the area of WRP ownership by
WDFW, and the locations of known OSF oviposition sites (Tyson & Hayes, 2014). The
West and East side marshes include study sites for Reed Canary Grass management,
while the Central West and Central East sites are not currently a part of those ongoing
studies. These sites are surveyed annually for OSF egg masses.
15
Figure 2.1.2. Aerial photograph of the WRP land ownership. Orange boxes indicate
locations of detected OSF egg masses in recent years. Figure adapted from Tyson &
Hayes (2014)
At the west side of the West Rocky Prairie site, there are two small ponds in
relatively close proximity to this population’s oviposition habitat. Based on preliminary
observations, adults appear to use one of these ponds (South Pond) (M. Hayes,
unpublished data), and hence, may utilize the second pond (Tilley Pond). Figure 2.1.3 is a
detailed map of the West Side Survey Area.
16
Figure 2.1.3. Aerial photograph of the West Side Survey Area and the locations of the
South Pond where OSF adults have been previously observed, and the location of Tilley
Pond where adult OSFs may be present.
*World Imagery Base Map by ESRI 2015. Map developed by Chelsea Waddell, 2015.
This study included multiple components, which began with WDFW directed
annual egg-mass surveys during the 2014 breeding season (February to March). During
these egg-mass surveys, in which I was a major participant, a total of 218 eggs were
collected from 109 egg clusters throughout the WRP site for a separate study of gene
flow. Availability of this genetic data was instrumental to my work, and the methods used
17
in this study are described below. In order to determine adult post-breeding (JulySeptember 2014) habitat use, myself, with the assistance of volunteers and WDFW
employees, surveyed both ponds, the intervening marsh, and the surrounding marsh
footprint, to capture adults. I used these surveys to gather both habitat and genetic data
for adults. With the help of the Molecular Genetics Laboratory (MGL) at WDFW, I
subsequently performed a genetic parentage analysis to determine the locations and
habitat preferences of the parents of offspring sampled during breeding season (See
Genetics Chapter).
BREEDING HABITAT SURVEYS
At the West Rocky Prairie site, OSFs began laying their eggs in the West Side
marsh on March 1st, 2014 (Tyson & Hayes, 2014). Surveys for egg masses are typically
done using Visual Encounter Surveys (VES), where surveyors walk parallel to each
other, 1-2 meters apart, scanning in front and to the left and right of the observer. Once an
egg mass is detected, it is marked with a pin flag. Egg mass fidelity, number of clusters
and masses, and habitat measurements are taken, and entered into a PDA (Personal
Digital Assistant) Excel spreadsheet (Tyson & Hayes, 2014). Measurements include GPS
points in Decimal Degrees, number of egg masses at the location, air and water
temperature in C°, water depth in cm, developmental stage (Gosner, 1960), and percent
mortality (Tyson & Hayes, 2014). The sites were typically revisited, depending on the
accessibility of the site, until the OSF tadpoles hatched.
18
Eggs were collected for genetic analysis from egg mass clusters located in six
representative locations. The West Side was partitioned into 3 locations: West North,
West Central, and West South. The other 3 were the East Side, Central West, and Central
East areas. See Figure 2.1.2 for site locations. Two eggs were collected randomly from
randomly selected egg masses at each site for genetic analysis, and stored at room
temperature in cryogenic tubes filled with DNA-grade ethanol (M. Hayes, unpublished
data).
NON-BREEDING HABITAT SURVEYS
Field surveys were always conducted by two or more people, which included
myself with the help of volunteers and/or WDFW employees. Field surveys for adult
OSFs during the non-breeding season were conducted in two seperate sessions. The first
session began on July 22nd, 2014 and continued through August 10th, 2014. The second
session began on September 5th, 2014 and continued until September 19th, 2014. Three
standard survey methods were used to conduct this research. Each of the three methods
described here were conducted throughout the field survey component of the study.
A study-specific PDA (Personal Digital Assistant) with an excel spreadsheet was
used for all data collection throughout the study. Equipment for these survey methods
included a Garmin GPS, digital thermometer, large dip nets for each surveyor, chestwaders for each surveyor, a waterproof digital camera, an iPhone 4S, and equipment
vests. Additional equipment was required for occasions where adults were captured and
included a bendable ruler, a digital scale, Ziploc bags, Sharpie markers, ethanol resistant
pens, and sterile Epicentre® Catch-All™ Buccal Swabs.
19
Visual Encounter Surveys
Visual encounter surveys (VES) were conducted by slowly walking the entire
marsh and pond areas (Figure 2.1.4), or by using an inflatable fishing tube where areas
were too deep to walk. These VES involve one or more individuals walking parallel to
each other, 1-2 meters apart, scanning the area to the left, right and in front of them for
post-metamorphic OSF. When an individual was seen, a dip net or hand capture was used
to capture the animal, depending on the configuration of the habitat and accessibility to
the animal (Figure 2.1.4).
Figure 2.1.4. Visual encounter surveys were conducted to capture OSF adults. On the left
is an image of a walking visual encounter survey in a narrow channel. The middle image
shows the capture of a frog using the dip-netting technique. The image on the right is an
adult male OSF captured using the aforementioned dip net technique. Photo Credits:
Sierra Blakeley & Cameron Smith.
Surveys conducted in deep water channels and deep ponds included visual
encounter surveys using an inflatable fishing tube (Figure 2.1.5).
20
Figure 2.1.5. Floating Visual Encounter Surveys were conducted in areas where water
was too deep to survey with Walking Visual Encounter Surveys. Photo Credit: Julie A.
Tyson.
The floating method was used to survey both the centers and edges of each pond, as
the ponds and some channels were too deep to survey by walking, even with chest
waders. Similar to the walking VES surveys, when adult OSFs were observed, they were
captured by hand or with a dip net depending on animal positioning and nearby
vegetation structure. Captured individuals were processed (see Animal Processing
section) in situ.
All adult OSFs observed were recorded and an attempt was made to capture all
observed adults. Captured OSF adults were processed according to the approach
described below (see Animal Processing section). Once each animal was processed,
individuals were immediately released behind the individual who captured it when
21
surveys were conducted in channels and the marsh. When animals were captured in the
ponds, they were placed in inflated zip lock bags with water and kept in the shade. This
helped decrease the likelihood of recapturing the same individual during the same survey
session. These individuals were then released back into the ponds and marsh unharmed
when the capture session was complete.
Additionally, habitat measurements (see Habitat Characteristic Measurements
below) were taken from random locations across the surveyed landscape (see Habitat
Results Figures 2.2.2. & 2.2.3) where adults were not detected. This information was
collected to determine OSF habitat preference during the non-breeding season in 2014.
Capture of adult OSFs was also done using float-enhanced minnow traps. The
traps were set up with short (2-3 meter) aquatic drift fences to increase the probability of
captures (Olson et al., 1997) (Figure 2.1.6).
Figure 2.1.6. Floating minnow traps. On the left is an image of multiple minnow traps
placed in sections of Tilley Pond where they were connected with drift nets. On the right
is an image of float enhanced minnow traps connected to a drift net in the East Channel
of WRP, and an individual reaching into the trap through a zipper opening to retrieve the
species within it. Photo credit: Sierra Blakely.
22
Traps were left open overnight and were checked within 12 hours (on the
following day). Minnow traps were also left open while surveyors were present and
conducting walking and floating surveys. Minnow traps were set up at intervals greater
than 10 meters apart across the different aquatic habitats (marsh, ponds) at West Rocky
Prairie. Trapped animals were then processed (see Animal Processing) in situ, and
released 2-3 meters away from the trap to minimize the likelihood of recapturing the
same individual during the same survey session. Minnow traps were discontinued during
the study, as they did not capture more than 2 adults over the entire study, and the
monitoring effort they required was substantial. Although they were discontinued, they
were particularly useful for finding other species that are present in the wetland. These
species included the Olympic Mudminnow (Novumbra hubbsi), Three-Spined
Stickleback (Gasterosteus aculeatus), and larval stage northwestern salamander
(Ambystoma gracile), which is shown in a minnow trap in Figure 2.1.7. Other species
encountered across the wetland include the Common Garter Snake (Thamnophis sirtalis),
Northern red-legged frog (Rana aurora), and leech (Hirudinea). Juvenile OSF,
metamorphosed juvenile OSF, and OSF tadpoles were also observed. Appendix A
contains the common and scientific names of observed flora and fauna at WRP during
OSF adult non-breeding habitat surveys.
23
Figure 2.1.7. Incidental Species in minnow traps: Larval Northwestern Salamander.
Photo Credit: Sierra Blakeley.
Animal Processing
Animals were processed according to standard WDFW protocols approved for
handling amphibians in the field (Beaupre et al., 2004). Once OSF adults were captured
using the aforementioned survey methods, they were processed using the following
template. For display mapping purposes, a GPS coordinate was taken for the location of
each animal captured in Decimal Degrees to the 6th decimal place. Air and water
temperature, and weather condition (e.g., mist, rain, sun, cloud cover) were also
immediately recorded at the location of observed or captured OSF adults. Temperature is
an important covariate for understanding habitat utilization and also influences the
likelihood of detection (M. Hayes, Personal Communication). Additionally, the general
location of the individual was noted. These categorical notes included whether the animal
was on land, in the water, or on the bank, whether they were in full sun or shade, and if
they were on or near vegetation. For each captured adult, I also measured snout-vent
24
length (SVL) in millimeters where the animal was gently pressed on the ruler with the
snout at 0mm, and measured to the vent (or tail). Animals with SVL greater than 50mm
(M. Hayes, personal communication) were included in the study, as lengths >45mm
indicate that males have reached sexual maturity (Hallock, 2013 via C. Pearl, personal
communication) (Figure 2.1.8).
Figure 2.1.8. Snout-to-vent measurements: conducted for each animal and was measured
in millimeters. Photo Credit: Chelsea Waddell.
Additionally, shank length (knee to heel) was measured in millimeters (Figure
2.1.9), and mass was measured for each animal in grams (Figure 2.1.10) to determine
their body condition (Yahnke et al., 2013).
25
Figure 2.1.9. Shank measurement: taken in millimeters (mm). Photo Credit: Chelsea
Waddell
Figure 2.1.10. Mass was measured in grams for each captured animal. Photo Credit:
Chelsea Waddell
26
In addition to the measurements described above, I determined each captured
animal’s gender (Figure 2.1.11). For the Oregon spotted frog, gender is typically
identified by looking at the presence or absence of nuptial pads, which are only present
on males (Hallock, 2013). Males use these nuptial pads to latch on to females (ampelxus)
during oviposition.
Figure 2.1.11. Male & Female Identification. Male (left) has a nuptial pad in the location
of the thumb, and female (right) does not have the nuptial pad present. Photo Credit:
Chelsea Waddell.
Photographs of the dorsal pattern for each captured adult were taken and used in a
study-specific photo-book, in the form of a power point, based on the protocol
established by WDFW. Adult OSF dorsal patterns remain similar across years and serve
as a useful tool for rapid identification of individual OSF adults (M. Hayes, unpublished
data). See Figure 2.1.12 for an example of dorsal pattern identification of an individual
captured on July 14, 2014 and August 10, 2014. These photographs can be used to
identify whether animals have been captured during previous years or previous survey
sessions, and will be integrated into a master photo-book, which has already been
27
established for this site. The study-specific Power-Point photo-book was particularly
useful for identifying individuals who had already been sampled for genetics in this
study. The photo-book was updated prior to each field survey session.
Figure 2.1.12. Adult Oregon spotted frog dorsal patterns is a useful tool for rapid
identification of individuals. The photograph (left) is a male captured on July 14, 2014;
the photograph (right) is the same male recaptured on August 10, 2014. Photo Credit:
Chelsea Waddell.
For each captured animal, all of the above measurements were taken before
determining whether the animal had previously been captured during the study. The
Power Point photo-book consisted of each individual captured at their first location of
capture, their SVL, shank length, mass, gender, general location of capture, unique
identifier code, and date of initial capture and sampling. See Figure 2.1.13 for an example
of the photo-book.
28
Figure 2.1.13. For each sampled individual, a power point slide was developed. This
included (from left to right) their Snout Vent Length (SV), Shank length, Mass, general
location of capture, unique identifier code, gender, date of initial capture and sampling,
and image of the animal. Photo Credit: Chelsea Waddell.
The captured animal would be compared, based on dorsal pattern, to the images
within the photo-book, and would either be deemed a recapture (an animal that had
previously been captured) or a new individual. If the animal were deemed a recaptured
individual, it would not be sampled for buccal cells. However, if the individual were
deemed a unique new individual, it would then have its mouth swabbed for buccal cells.
Field Sampling For Genetic Material
All captured and processed OSF adults, as described above, were also sampled for
genetics by conducting the buccal sampling technique outlined in Pidancier et al. (2003),
29
Poschadel & Moller (2004), and Gallardo et al. (2012). Buccal sampling is a less invasive
alternative to the more commonly used toe-clipping method for collecting tissue samples
from amphibians. Tissue sampling included swabbing each captured unique individual’s
mouth with duplicate buccal swabs (Epicentre® Catch-All™ Sample Collection Swab).
Mouth swabs were done only when the animal was deemed a unique individual based on
In-Field dorsal pattern recognition, as describe above. Swabs were dried immediately,
and stored at -20°C. See Appendix B for full protocol for buccal swabbing using the
Epicentre© Catch-All™ Sample collection Swabs.
Habitat Characteristic Measurements
Multiple covariates were measured for locations where adults were captured and for
locations where adults were not detected. In order to determine what habitat adults prefer
based on their detectability during the survey, vegetation type, temperature, cloud cover,
OSF juvenile abundance, and incidental amphibian and fish species presence were
collected for all areas surveyed to determine which areas the adults preferred over others.
Survey Area by Region & Organisms Observed
During field surveys, areas were partitioned by geographic location. The locations
included: Small (South) Pond, East Channel, Clearing North of East Channel, North
Channel, West Channel, Area West of West Channel, Tilley Pond, Pond Adjacent to
Tilley Road, and the East Side Survey Area. In order to quantify the number of times
each area was visited, the master Excel spreadsheet was filtered based on these locations,
and the sum of the number of days each site was visited was quantified. Additionally, the
types of other species observed in these areas were listed.
30
ANALYSIS OF EFFORT METHODS
TOTAL EFFORT
Calculation of total effort during the detection surveys was based on start and end
times from each survey, and summary statistics were calculated using Microsoft Excel.
Specifically, the average number of hours for each survey was calculated by dividing the
sum of hours for each survey by the number of survey days. The sum of the product of
the number of hours a single survey took, and the number of surveyors present on that
day, was used to calculate the Total Survey Hours. Effort per Animal was calculated by
dividing Total Survey Hours by the number of adult Oregon spotted frogs captured,
which included new captures and recaptures. This Effort per Animal calculation is based
on the effort of each surveyor, and not based on the number of frogs captured within each
survey.
CATCH PER UNIT TIME
Additionally, a calculation of catch per unit time was calculated by partitioning all
of the survey observations into two-hour time intervals. For each time interval, the
percentage of times an animal was captured during that time interval was calculated
based on the total number of observations that occurred at that time interval. The total
number of days animal observations made for each time interval was then summed.
31
SPATIAL ANALYSIS METHODS
All habitat distribution maps of captured and sampled adults were done using
ArcGIS 10.2 (ESRI, 2015). The same World Imagery base map (ESRI, 2015) was used
for every map developed in this study. All attributes were compiled from collected field
data. Methods used for making these maps include point selection, and polygon
formation based on GPS points collected in the field. The maps included in the Habitat
Results section show the entire surveyed area, locations where adults were captured and
sampled, locations where egg masses were observed and sampled, and vegetation types
throughout the surveyed area. All points were based on GPS data collected with a Garmin
handheld GPS unit; the average error for each point was 11.86 feet, based on the total
number of points taken with the error noted (n=184).
SURVEY AREA & VEGETATION TYPE
Based on survey points, polygons were generated using the Create Features tool in
ArcMAP 10.2. Polygons of the survey area were generated based first on viewing the
ESRI World Imagery base map (ESRI, 2015), and then based on survey points gathered
during field collection. Survey Area polygons were then broken up into 5 land cover
types based on the dominant species observed; open water, reed canary grass, sedge,
scrub/shrub/willow, and other (See Appendix A for scientific names). Using the
Snapping Tool, each adjacent polygon was snapped to the one next to it to ensure the area
of each land cover type equaled the sum of the total survey area. For each land cover
type, multiple polygons were made adjacent to each other, depending on the orientation
of other nearby polygons. To remove excess lines between polygons, the Dissolve tool in
32
ArcMAP 10.2 was used. To calculate the ‘Other’ land cover type, the polygons generated
for open water, reed canary grass, sedge, and scrub/shrub/willow were merged, using the
Merge tool in ArcMAP 10.2. The Erase tool was then used to erase the merged area from
the total survey area; this generated the polygon for “other” in ArcMAP 10.2.
The areas of each land cover type, and the total survey area, were then calculated
by using the Editing Tool and the Calculate Geometry tool in the attributes table of
ArcMAP 10.2 (ESRI, 2015). Area was calculated in both acres (ac) and square meters
(m2). The difference between the survey area and the sum of all land-cover-type polygons
was then calculated to demonstrate the accuracy of the snapping tool and polygon
formulation.
ADULT LOCATIONS
With the total number of adults observed or captured during the study, the Select by
Location tool in ArcMAP 10.2 (ESRI, 2015) was used to determine which land cover
type the animals were observed in. These were then separated into captured and
observed, and a map was generated to convey the locations of each observed or captured
adult, and the land-cover types within the survey area.
33
HABITAT UTILZATION
RESULTS
EGG-MASS SURVEYS
Egg-mass surveys were conducted between February and March 2014, and eggs
(n=218) for genetic analysis were collected from 109 egg masses at six locations. Figure
2.2.1 shows the locations where egg masses were observed. A total of 336 egg masses
were observed among these sites, with the highest amount observed (n total =288) at the
west side survey area (west north, west central, west south). For the east side survey area
(n total =15), central east (n total =24), central west (n total =9).
34
Figure 2.2.1. WDFW employees and volunteers at West Rocky Prairie collected egg
masses between February and March 2014: egg masses were coded based on the six
locations demarcated by orange boxes.
*World Imagery Base Map by ESRI 2015; West Rocky Prairie WDFW Management Resource Boundary; coordinates
collected by WDFW employees and volunteers, 2014. Map developed by Chelsea Waddell (2015).
NON-BREEDING HABITAT SURVEY AREA
A total of 29.1 acres were surveyed during the adult survey component of the
study. Surveyed areas included the west and east sides of West Rocky prairie, and a small
pond off Tilley road (Figure 2.2.3), which is southwest of Tilley pond (Figure 2.1.3 in
Methods). These surveyed areas represent land types that are inundated with water during
the wet season, in addition to areas of exploration where adult OSF surveys had not been
done in the past. Figure 2.2.2 is a map of the overall survey area and the West Rocky
35
prairie state ownership polygon, while Figure 2.2.3 is zoomed in to the west and east side
survey areas, respectively. Areas to the northwestern side of the west side survey area
include exploratory survey expeditions for open water sources, where the hope was to
find adult Oregon spotted frogs.
Figure 2.2.2. Map representative of the entire survey area within the Washington state
owned West Rocky Prairie. Surveys were partitioned into two general areas: West Side
and East Side, with particular attention to the West Side area.
*World Imagery Base map by ESRI 2015; West Rocky Prairie WDFW Management Resource Boundary; coordinates
collected by Chelsea Waddell, volunteers, and WDFW employees, 2014. Map developed by Chelsea Waddell (2015).
36
Figure 2.2.3. Map representative of the West Side Survey Area & Pond South of Tilley
Pond (right) and the East Side Survey Area (left) area within the Washington state owned
West Rocky Prairie. Surveys were partitioned into two general areas: West Side and East
Side, with particular attention to the West Side area.
World Imagery Base Map by ESRI 2015; coordinates collected by Chelsea Waddell, volunteers, and WDFW
employees, 2014. Map made by Chelsea Waddell (2015).
EFFORT
Adult capture surveys were conducted between July 22nd, 2014 and September
19th, 2014, with a total of 28 survey days. The surveys were partitioned into two sessions.
The first session was between July 22nd and August 10th, which included 17 survey days.
The second session was between September 4th, 2014 and September 19th, 2014, which
included 11 survey days. Between two and four surveyors were present each survey day.
A total of 13 surveyors participated in the adult capture component of the study, two of
them were WDFW employees, and others included volunteers and myself. The average
length of each survey was 5.76 ± 2.08 hours, which were typically performed during the
mornings (Table 2.2.1). The total survey hours, calculated based on total person hours of
effort, was 395.3 hours. Additionally, 194 new and recaptured adults were captured
during the adult survey component of the study, each of which took an average of 2.04
37
hours to capture based on the total person hours. However, an average of 7.07 hours of
effort was required to capture new adults, or those that had not yet been captured during
the study. See Table 2.2.1 below for summary table of the outlined results.
Table 2.2.1. Summary of effort per adult captured
Surveyors
(n)
Survey
Days
(n)
Average
Survey
Hours/
Day
Standard
Dev.
13
28
5.76
±2.08
Surveyor/Day
(n)
Total
Survey
Hours
Adults
Captured
(n)
Person
Hours/
adult
captured
Person
Hours/
new
adult
captured
2 to 4
395.3
194
2.04
7.07
Additionally, observations per unit time were calculated in order to determine in
which time interval the most captures occurred (Table 2.2.2.). Most observations
occurred between 10:00 and 11:59 hours. This time period also had the most captures of
adult Oregon spotted frogs. However, the highest percentages of adults observed per total
observations were between 16:00 and 19:59 hours, indicating that these times may be
best for capturing adult Oregon spotted frogs. Although, this result was likely confounded
by the lower number of days (n=6) surveys were occurring during those time periods
when compared to time intervals between 8:00 and 15:59. Additionally, those later times
also had a lower number of survey locations where observations occurred.
Table 2.2.2. Summary of effort per adult partitioned by two-hour time intervals
Total Observations (n)
68
185
120
53
16
3
Days (n)
17
24
21
19
5
1
Animal Observations (n)
32
94
36
15
14
3
% Animal Observations
47.1%
50.8%
30.0%
28.3%
87.5%
100.0%
38
ADULT LOCATIONS BY LAND COVER TYPE
No adult Oregon spotted frogs were observed or captured in the East Side survey
area or the pond next to Tilley Road. Adults were either observed or captured in the East
Channel, the clearing North of the East Channel, the West Channel, the North Channel,
Tilley Pond, and in the small pond in the West Side Survey Area (Figure 2.2.5).
Although, the majority of adults captured were present in the small South Pond, which is
10×6 meters (Tyson & Hayes, 2014), and varied in depth over the course of the study
(Figure 2.2.4). Of the 194 captured adults in this study, 161 or 82.99% of them were
captured in the small pond.
Figure 2.2.4. Small Pond in the West Side Survey Area of West Rocky Prairie; Image
taken on July 22nd, 2014.
39
The second most abundant location where adult Oregon spotted frogs were
captured was the East Channel. A total of 25 adults (12.89%) were captured there, and
many of those individuals were located in the part of the East Channel directly adjacent
to the small pond, as can be seen in Figure 2.2.5. The other locations where adults were
captured include the West Channel, where 5 adults (2.58%) were captured, the pond
North of the East Channel, where 2 adults (1.03%) were captured, and the North Channel,
where 1 adult (0.52%) was captured.
Figure 2.2.5 shows the geographic locations of the 194 captured adults, and the
locations of observed, but missed adults. Adults were only observed in open-water landcover types (Table 2.2.3), and the point within the small pond represents 161 captured
adults and observations of missed adult Oregon spotted frogs.
40
Figure 2.2.5. Locations of observed and captured adults, and the land cover types they
were observed in.
*World Imagery Base Map by ESRI 2015; coordinates collected by Chelsea Waddell, volunteers, and WDFW
employees, 2014. Map developed by Chelsea Waddell (2015).
41
LAND COVER
Land cover type was partitioned into five distinct categories of dominant
vegetation or land cover: scrub/shrub/willow, sedge, reed canary grass, open water, and
other (See Appendix A for scientific names). The other land cover type generally
represents an area that did not have survey points associated with it, or the land cover
type was not distinguishable based on the 2015 ESRI World Imagery base map used for
developing the land cover type polygons. The land cover types are depicted below
(Figures 2.2.6, 2.2.14, & 2.2.16) in the form of three maps, each of which describes one
of the three survey areas assessed in this study; they include the West Side Survey Area,
East Side Survey Area, and the Pond/Clearing near Tilley Road.
ANIMAL PRESENCE
The area of each land cover type was calculated for all areas surveyed for this
study. The total survey area was 29.1 acres based on the survey-area polygon, which
included the West Side Survey Area, the East Side Survey Area, and the Pond adjacent to
Tilley Road (Table 2.2.3). Each land-cover-type polygon was snapped to other adjacent
land-cover types and the survey-area polygon, and encompassed a total of 28.77 acres, or
98.9% of the survey area. This result indicates that the snapping method used for
describing the land cover types throughout the entire survey area was successful.
All 194-captured adult Oregon spotted frogs were observed in the open water land
cover classification (Table 2.2.3).
42
Table 2.2.3. Area of land cover type and the number of captured adults within each.
Scrub/Shrub/Willow
0
11.0
Reed Canary Grass
0
7.8
Sedge
0
2.8
Open Water
194
5.6
Other
0
1.5
Total
194
29.1
SURVEY AREAS
While field surveys were being conducted, survey areas were partitioned into
regions based on geographical features. These regions, and the number of times they
were visited, are outlined in Table 2.2.4. The types of species detected within each area,
and whether minnow traps were present, are also described in Table 2.2.4. Regions with
higher abundance of adult OSF were visited more frequently than those where adult OSF
were infrequently, or not detected, as the study objective was to determine where the
parents of egg masses resided during the non-breeding season.
43
Table 2.2.4. Survey Regions: Number of surveys, use of minnow traps, and diversity of
species present within each region.
Area West of West
Channel
4
No
Juvenile OSF, Metamorphosed OSF
Pond South of
Tilley Pond
2
No
Northern Red-legged Frog
East Channel
15
Yes
Adult OSF, Juvenile OSF, Metamorphosed OSF,
OSF tadpole, Northern Red-legged Frog, Olympic
Mudminnow, Three-Spined Stickleback, Common
Garter Snake, Leech, Northwestern Salamander
East Side
2
No
Juvenile OSF, Northern Red-legged Frog
North Channel
6
No
Adult OSF, Juvenile OSF, Metamorphosed OSF
Clearing North of
North Channel
7
Yes
Adult OSF, Juvenile OSF, Metamorphosed OSF,
OSF tadpole, Olympic Mudminnow, Three-Spined
Stickleback, Northwestern Salamander
Small Pond
13
Yes
Adult OSF, Juvenile OSF, Olympic Mudminnow,
Common Garter Snake, Northwestern Salamander
Tilley Pond
8
Yes
OSF Adult (observed), Juvenile OSF, OSF Tadpole,
Olympic Mudminnow, Three-Spined Stickleback,
Northwestern Salamander, Common Garter Snake
West Channel
10
No
Adult OSF, Juvenile OSF, Metamorphosed OSF,
Olympic Mudminnow, Northern Red-legged Frog
Area Between West
Channel and East
Channel
>2
No
No animals observed
See Appendix A for scientific names.
44
WEST SIDE SURVEY AREA
Figure 2.2.6 is a map of the land-cover types found throughout the West Side
survey area. Each area was surveyed a minimum of 3 times on the West side, with the
most frequent surveys occurring at the small pond. Since these surveys were conducted
during the dry season in western Washington, many of the areas that are inundated within
the wetland during oviposition time did not have water present. However, areas in the
northwestern side of the West Side survey area, which is dominated by
scrub/shrub/willow, was frequently inundated with water. The clearings west of the
diagonal portion of the West Channel were areas of particular exploration for adult
Oregon spotted frogs, and had not been surveyed for adults prior to this study. These
areas were chosen through examination of aerial photographs that indicated the presence
of water, which was confirmed by our surveys. Observationally, the amount of water
within the open water land cover type decreased substantially between the survey session
conducted from July to August, and the session conducted from August to September,
2014.
45
Figure 2.2.6. West Side Survey Area: Land Cover Type.
*World Imagery Base map by ESRI 2015; coordinates collected by Chelsea Waddell, volunteers, and WDFW
employees, 2014. Map developed by Chelsea Waddell (2015).
46
Small Pond
The Small Pond was the second most surveyed area (n=13), as this area had the
highest abundance of detected adult OSF. Many species were present in this area, as
demonstrated in Table 2.2.4. Additionally, minnow traps were set up in the small pond,
but they yielded very few adult OSF. The small pond was surveyed predominantly by
walking VES, but Floating VES was also used. Observationally, water levels decreased
in the small pond between the time of the first survey session (July to August) and the
second survey session (August to September) (Figure 2.2.7).
Figure 2.2.7. Images of the water level change in the Small Pond. The photograph on the
left was taken on Aug. 10th, 2014. The photograph on the right was taken on Sept. 19th,
2014. Photo Credit: Chelsea Waddell & Cameron Smith.
East Channel
Portions of the East channel were surveyed most (n=15). This is, in part, due to
the fact that the East Channel area directly adjacent to the Small Pond was surveyed at
the same time as the small pond. Additionally, minnow traps were set out in this area.
However, they yielded very few adult OSF and a high abundance of other species, as well
as juvenile, metamorphosed, and tadpole OSF. All species were also detected within the
47
East Channel, including the Common Garter Snake (Thamnophis sirtalis). According to
M. Hayes (personal communication), Common Garter Snakes prey upon young OSF.
This area typically had water present (Figure 2.2.8), and was traversed using the walking
Visual Encounter Surveys (VES). The northern extent was narrow with heavy vegetation,
typically had water present, and was generally more difficult to traverse (Figure 2.2.8).
During the second survey session (August to September 2014), water had disappeared
from large portions of the East Channel.
Figure 2.2.8. East Channel: Image on the left is a portion of the southern extent of the
East Channel taken on July 25th, 2015 (Photo Credit: Sierra Blakeley). Image in the
center is a portion of the northern extent of the East Channel taken on Aug. 5th, 2014.
Image on the right is the East Channel on September 9th, 2014. Photo Credit: Chelsea
Waddell & Cameron Smith.
Area Between East and West Channel
The area between the East and West Channel was surveyed two times (Table
2.2.4). Although, in order to get to and from the West Channel and the Area West of the
West Channel, the area North of the small pond between the East and West Channels had
to be crossed. For this reason, many more visits to this area occurred than were actually
included in a formal survey. No aquatic animals were detected in this region, although
some shallow water on the North end was present during the first survey session.
Walking VES was used to survey this area.
48
West Channel
The West Channel was surveyed on 10 occasions, and a diversity of species was
encountered. The channel typically had water present, and at times the water was too
deep to survey using chest waders. Therefore, both Walking VES and Floating VES were
used there. Figure 2.2.9 below shows 3 images of the West Channel: One at a southern
location past a beaver dam where adults were captured, another at the central area where
adults were also captured, and a third at the northern extension where an adult was
captured. Water levels also decreased in the West Channel, but this was most noticeable
in the northern extent of the channel.
Figure 2.2.9. West Channel. Top left is and image taken on Sep. 12th, 2014 of the West
Channel at a southern location past a beaver dam where adults were observed. Top right
is the West Channel at a central location taken on Aug. 8th, 2014. Bottom left image,
taken on Aug. 7th, 2014, is of the northern extension of the West Channel. Bottom right
image, taken on Sep. 11th, 2014 is of the northern extension of the West Channel. Photo
Credit: Chelsea Waddell & Cameron Smith.
49
Clearing North of North Channel
A clearing North of the North Channel was visited 7 times during the surveys
(Table 2.2.4), and numerous species were detected there including adult OSF. This area
typically had water present, and connected to the southern and northern extent of the East
Channel (Figure 2.2.10). Minnow traps were set out at this location and Walking VES
was used for surveys. This area also had an observable decrease in water levels during the
second survey session (August to September).
Figure 2.2.10. Clearing North of North Channel: The photograph on the left is of the
main section of the clearing and was taken on Aug. 7th, 2014. The photograph in the
center is of the channel leading to where the corner of the East Channel and North
Channel meet it. The image on the right was taken on Sep. 5th, 2014. Photo Credit:
Chelsea Waddell & Cameron Smith.
North Channel
The North Channel, which connects the East and West Channels, was surveyed 6
times and a single adult OSF was captured there (Table 2.2.4). This area typically had
water present, but no minnow traps were used. This area was, at times, difficult to
traverse because of its depth, and the presence of a large, impassable willow in the center.
However, Walking VES surveys were used to survey this area. Figure 2.2.11 is two
images of the North Channel taken from its connection point with the West Channel,
50
showing what the channel looked like during the first session (July to August) and the
second session (August to September).
Figure 2.2.11. North Channel. Image on the left was taken on Aug. 7th, 2014. Image on
the right was taken on Sep. 11th, 2014. Photo Credit: Chelsea Waddell & Cameron Smith.
Area West of West Channel
The area west of the West Channel encompasses a large area dominated by dense,
tall vegetation with occasional clearing and channels. This area was surveyed 4 times
throughout the survey, and only juvenile and metamorphosed OSFs were detected in the
clearings and channels (Table 2.2.4). These areas typically had water present with the
heavy scrub/shrub/willow (discussed above) growing out of the water. Figure 2.2.12
shows what the dense vegetation and clearings looked like at the time of the surveys. The
clearings during the first survey session had water in them. However, during the second
survey session, the water was no longer present. These areas were typically very difficult
to traverse, which may have influenced detectability of OSF there.
51
Figure 2.2.12. Area West of West Channel: Top right and left, taken on Aug. 1st, 2014
are of the scrub/shrub/willow dominant vegetation in this area. The image on the bottom
left is of one of the clearings on Aug. 7th, 2014. The image on the bottom right is of one
of the clearings on Sep. 11th, 2014. Photo Credit: Chelsea Waddell & Cameron Smith.
Tilley Pond
A primary focus of this project was to determine whether adult Oregon spotted
frogs resided in Tilley Pond and it’s connecting channels (Figure 2.2.13). However, after
an exhaustive effort (8 surveys), only one adult Oregon spotted frog was observed. An
effort was made to capture this adult, but failed due the amount of vegetation present.
Many species were detected in Tilley Pond (Table 2.2.4). Walking VES was used
to survey the edges and connected channels of this pond, but Floating VES was used to
survey the center of the pond, as it was too deep to traverse in chest waders.
52
Observationally, water levels decreased in Tilley pond during the second survey session
(August to September, 2014), especially around the edges of the pond.
Figure 2.2.13. Tilley Pond on July 25th, 2014. Photo Credit: Sierra Blakeley
POND SOUTH OF TILLEY POND
(aka. Pond Adjacent to Tilley Road; Beaver Creek Pond)
The pond next to Tilley Road is directly south of Tilley pond and was approached
as an exploratory measure to see if adult Oregon spotted frogs were present. The
surveyed area was heavily dominated with deep open water (Figure 2.2.14), with
channels leading to more open water on the east end of the pond area. The pond is
connected to Beaver Creek, which runs east to west in the area directly under Tilley
Road, in lowland western Washington.
53
Figure 2.2.14. Pond South of Tilley Pond and adjacent to Tilley Road: Land Cover
Type. Tilley road, which is represented in the map West of the pond.
*World Imagery Base map by ESRI 2015; coordinates collected by Chelsea Waddell, volunteers, and WDFW
employees, 2014. Map developed by Chelsea Waddell (2015)
54
The area (Figure 2.2.15) was surveyed two times (Table 2.2.4). This pond had
water present during both surveys. Northern Red-legged frogs were observed in the pond,
but no adult or juvenile OSF were observed or captured there. Floating VES were used to
survey the area. No minnow traps were used in this area.
Figure 2.2.15. Pond South of Tilley Pond. Image taken on Sep. 12th, 2014. Photo Credit:
Cameron Smith.
EAST SIDE SURVEY AREA
The East Side Survey Area (Figure 2.2.16) was visited twice (Table 2.2.4) during
the adult survey time, as this area was particularly difficult and dangerous to traverse.
The second survey was the most extensive survey of this area. The area included very
deep mud, at times deeper than chest height. Land-cover type was assessed based on the
methods described previously, and areas of open water often included large flowering lily
pads, swarming with yellow jackets. Areas indicated as other were designated when
survey points were not taken in those areas, and the vegetation was not distinguishable
based on the ESRI World Imagery base map used (ESRI, 2015).
55
Figure 2.2.16. East Side Survey Area: Land Cover Type.
*World Imagery Base Map by ESRI 2015; coordinates collected by Chelsea Waddell, volunteers, and WDFW
employees, 2014. Map developed by Chelsea Waddell (2015).
To survey the East side survey area (Figure 2.2.17), walking and floating VES
were used. Northern red-legged frogs were the most abundant species observed, and
juvenile OSFs were observed.
56
Figure 2.2.17. East Side Survey Area. All three images were taken on Sep. 10th, 2014.
Photo Credit: Chelsea Waddell.
DORSAL PATTERN RECOGNITION RESULTS
During field surveys, dorsal patterns were used to distinguish individuals who had
been previously sampled, and those that had not yet been captured and sampled. During
the first survey session, a PDA (Personal Digital Assistant) was used to view the pictures,
however it is likely that the low resolution of the screen caused misidentification of
individuals that had been previously sampled; thus causing repeat sampling of
individuals. During the second survey session, images of previously captured individuals
were viewed on an iPhone 4S, which had a substantially higher resolution and,
observationally, made rapid identification simpler. Using this method in the field, 81
adult OSF were sampled, and deemed unique new individuals.
When adults were compared in the office on higher resolution screens in
September 2014, the identification of repeat individuals was noticeably faster, and 58 of
the 81 sampled adults were deemed unique, new individuals. Finally, when compared to
the genetic results, 56 individuals were deemed unique individuals (See Chapter 3
Results for details on the results of this comparison).
57
CHAPTER 3
GENETICS
INTRODUCTION
OVERVIEW OF POPULATION GENETICS & AMPHIBIAN DECLINE
Population genetics is a growing and complex field in which genetic analysis
informs researchers about the status of species populations. Areas of considerable
concern in population genetics include gene flow, inbreeding depression, heterozygosity,
allelic richness, and effective population size. These analyses can be performed at
multiple spatial and temporal scales, and with varying objectives. Commonly, species’
populations tend to decrease first at their extended range, causing them to decline inward
to the center of their range (McKenzie et al., 2005). This pattern is evident in the Oregon
spotted frog, as its current range is now much smaller than its historical one. OSF
populations have declined in recent decades, and occupy only 10-30% of their original
range (Blouin et al., 2010). The species historically persisted in southern British
Columbia, western Washington, western Oregon, and northern California; it is now
believed to be extinct in California, and parts of western Oregon (Blouin et al., 2010).
58
AMPHIBIAN POPULATION DECLINE
Amphibians are considered the most imperiled of the vertebrates, with 41% of
them threatened with extinction (Monastersky, 2014). Many issues contribute to the
decline of amphibians worldwide, but they can generally be broken down into two classes
of factors, deterministic and stochastic. These factors affect amphibian population health,
and can act additively or synergistically (Storfer et al., 2009). Class 1 factors, or
deterministic factors, include habitat alteration and the introduction of invasive and nonnative species. These factors can cause declines in food availability, and invasive species
may prey on native amphibians or compete for resources (Storfer et al., 2009). Class 2
factors, or stochastic factors, include global climate change, infectious disease, and
environmental contaminants (Storfer et al., 2009).
As discussed in Chapter 1,
deterministic and stochastic factors affecting OSF include the invasion of Reed Canary
Grass, invasion of bullfrogs, and susceptibility to contaminants. Given the tremendous
diversity of issues amphibian populations are facing, it is critical to understand how their
populations are functioning in order to best manage them. Population genetics serves an
integral role in endangered species management and has been increasingly used for many
declining species, including the Oregon spotted frog (Blouin et al., 2010).
Habitat alteration and fragmentation can have major effects on amphibian
populations. Landscape genetics has been used to address issues of gene flow (exchange
of genes) among populations. Habitat loss and fragmentation can restrict the dispersal,
and exchange of genes (i.e., gene flow), which is important for maintaining genetic
diversity among populations (Storfer et al., 2009). Restrictions to gene flow can cause
populations to have high susceptibility to inbreeding, which can cause further
59
demographic problems commonly associated with small population sizes (Storfer et al.,
2009); this is also the case with the decline of the Oregon spotted frog (Blouin et al.,
2010). Furthermore, when populations become small and isolated, they can lack genetic
variability and thus, are less able to adapt to future environmental changes such as
climate change and increased fragmentation (Storfer et al., 2009). With increasing habitat
alteration and fragmentation, maintaining connectivity between habitats will enable gene
flow and, ideally, functional populations. However, research needs to be conducted to
determine how these populations function on large and small scales. The integration of
population genetics with wildlife management undoubtedly increases the means with
which we can manage threatened and endangered species.
The primary emphasis of OSF management has been to mitigate loss of habitat
and fragmentation. These efforts have been supported with genetic information, which
inform managers about their population structure, and how populations across their range
differ genetically. However, more needs to be learned from looking at the small, isolated
or distinct populations.
OREGON SPOTTED FROG GENETICS: CURRENT KNOWLEDGE
Genetic research on OSF has focused on understanding the divergence between
populations and the population structure of the OSF across its range. This section directly
addresses the current knowledge about genotypic variation across the OSFs geographic
range. General knowledge of this information is critical to effectively managing their
populations, and for interpreting the methods and results of this study.
60
Blouin et al. (2010) compared the genetic variability and health of the Oregon
spotted frog to a related ranid species called the Cascade frog (Rana cascadae). Both
species share much of the same general geographic range, but Cascade frogs are more
abundant (Blouin et al., 2010) and generally occur at higher elevations. Based on 23
sampled populations, three major hierarchical groups, or clades, of Oregon spotted frogs
exist across their range; a northern clade, a central cascades clade, and a southern
Klamath basin clade (Blouin et al., 2010). Figure 3.1.1 is a map developed by Blouin et
al. (2010) showing the OSF sample sites where genetic material was collected from OSF;
there are circles around two of the three clades on the southern extent of the map.
Locations north of the central cascades clade are considered part of the northern clade
(Blouin et al., 2010).
Figure 3.1.1. Sampled Range: Map of OSF sampling locations representing their
geographic range, and clades. Figure adapted from Blouin et al. (2010).
61
Furthermore, hierarchical substructure was found within these three groups. Four
subgroups exist within the northern clade, but weaker subgroup structure is evident in the
central cascade and southern Klamath clades (Blouin et al., 2010). A Principle
Coordinate analysis shows the genetic distances based on the allelic frequencies of all
685 individuals sampled in these three major groups and their substructures (Figure
3.1.2).
Figure 3.1.2. Principle Coordinates Analysis showing the Oregon spotted frog’s three
genetically distinct clades; genetic distances are based on the divergence of allelic
frequency. Figure adapted from Blouin et al. (2010).
62
Given the amount of substructure in the northern clade, special attention should
be paid to these populations (Blouin et al., 2010). The level of genetic distance between
these groups indicates low connectivity and minimal gene flow between them. Gene flow
is very small beyond 10km, and the distances between populations are typically larger
than this (Blouin et al., 2010). It is therefore important to maintain healthy populations
within the six subgroups (4 Northern, 1 Central, 1 South) because genetic rescue from
nearby populations is not likely to occur. According to Blouin et al. (2010), these six
subgroups should be considered Ecologically Significant Units (ESU), and therefore
should be considered distinctly different for purposes of conservation (Blouin et al.,
2010). The population of interest in my study is located within the Chehalis clade. While
understanding this species divergence across its range provides critical information about
the range-wide population structure, it does not paint the entire picture of individual
population structures and functions.
Blouin et al. (2010) tested for deviations from the Hardy-Weinberg and genotypic
equilibrium, which is the assumption that there is a constant level of genetic variation
from generation to generation, for each of the populations (n=23). They determined that
all populations tested showed little genotypic disequilibrium between loci (locations of a
gene) and were in Hardy-Weinberg equilibrium (Blouin et al., 2010). They also tested for
allelic richness (number of alleles which are alternative forms of a gene, AR) and
heterozygosity (H e ). They found a mean H e =0.31, suggesting that 31% of loci
characterized were heterozygous instead of homozygous, and an AR 15 =2.46, indicating
that in a population size of 15 individuals, the mean number of alleles per loci was 2.46
63
(Blouin et al., 2010). These results indicate lower allelic richness and heterozygosity in
OSF than found in the comparison species, Cascade frogs (Blouin et al., 2010).
Effective population size (effectively breeding adults, N e ) is a foundational
principle used by conservation genetics. Based on a single season analysis, the OSF is
thought to have especially small effective (0.1-0.4) population sizes when compared to
their census population (Total population, N) (Phillipsen et al., 2009). Fluctuation in
population size and variance in family size are thought to impact effective population
size, especially in pond breeding amphibians like the OSF (Phillipsen et al., 2009). The
influence of habitat factors, especially in breeding habitat, is likely the cause of OSF
boom and bust population changes from year to year (Phillipsen et al., 2009).
Additionally, their family sizes may vary greatly, since females only lay 1 egg mass per
year, and males only fertilize one egg mass per year (Phillipsen et al., 2009). These egg
masses are often susceptible to freeze, desiccation, and disease based on the conditions
that year. However, single population studies are not fully indicative of general effective
population sizes across the OSF’s range, and should be expanded to areas across their
range. In part, this project adds to our current understanding of effective population sizes
(N e ) across OSF’s range. Furthermore, the addition of parentage analyses, looking at the
linkage between parents and offspring within a single population, can be extremely useful
for understanding family relationships between individual OSFs in a small population.
Knowledge of population structure across the OSF geographic range has
established precedence for further studies. This is especially the case for small,
potentially isolated populations, as these populations are the primary focus of
management.
64
APPLICATIONS OF POPULATION GENETICS
As species continue to decline worldwide, we must integrate knowledge from
multiple conservation fields to most adaptively conserve and manage biodiversity.
Population genetics is a burgeoning field which has substantially advanced our
knowledge of how declining and healthy populations function. Its applications are vast in
the conservation community, and it is especially applicable to conserving declining
amphibian populations such as the Oregon spotted frog. Furthermore, there are sub-fields
within population genetics and genomics, such as parentage analyses, which can help
answer specific hypotheses.
PARENTAGE ANALYSIS OVERVIEW
Parentage analyses have been widely used by ecologists in diverse fields to obtain
knowledge about wildlife population structures and the behavior of these populations.
Parentage analyses use genotypes of individuals to assign paternity and maternity
(Frankham et al., 2003). It has become a prevalent practice in the field of molecular
ecology and has advanced quickly (Jones & Ardren, 2003). “Patterns of parentage play a
central role in the study of diverse ecological and evolutionary topics, such as sexual
selection, patterns of dispersal and recruitment, estimation of quantitative genetic
parameters, and conservation biology” (Jones et al., 2010). Additionally, parentage
information helps managers understand the impacts of inbreeding, determine the effective
population size, and verify pedigree so that the species can be managed based on their
genetics (Frankham et al., 2003).
65
In the 1980s, DNA fingerprinting advanced the field, and parentage analyses were
often conducted to determine the behavioral ecology of bird populations (Jones &
Ardren, 2003). When microsatellites were discovered (See Methods: Microsatellites),
they quickly became the chromosomal section of choice in parentage analyses. Until
recently, parentage analyses were predominantly conducted with avian and fish
populations, this was primarily due to a lack of identified microsatellites for other, less
commonly studied species (Jones & Ardren, 2003). Currently, the field is still growing to
incorporate more areas of the genome, and more advanced computational analyses.
Parentage analyses utilize DNA obtained from the focal organisms and, ideally,
DNA from both parents and offspring should be obtained. “The basis of paternity comes
down to the fact that in the absence of mutation, a child receives one allele matching each
parent at every genetic locus examined” (Butler, 2005). Microsatellites are commonly
used for parentage analyses because they follow the rules of Mendelian segregation,
where a child receives one allele from each parent (Jones et al., 2010). However, it is
possible for both parents to share alleles. For this, there are various statistical approaches
to determining parents of specific offspring (Butler, 2005). The methods used for
conducting these types of analyses are discussed in the Methods section of this chapter.
Methods begin with obtaining the tissue samples from the organism, followed by
laboratory methods for extracting (See Methods DNA Extraction Section) the DNA from
those samples, and amplifying them via PCR (See Methods Polymerase Chain Reaction
Section). Finally, the parentage-analysis methods are discussed, and the methods for
determining the population’s N e , and allelic diversity.
66
GENETICS
METHODS
Field, laboratory, and analytical methods were used for the genetic components of
this project. Genetic samples from adults and egg masses were collected in the field (as
described in Habitat Chapter Methods). Once the samples were collected in the field and
frozen at -20°C, a series of laboratory methods were performed.
Staff in the Molecular Genetics Laboratory (MGL) at Washington Department of
Fish and Wildlife (WDFW) completed the genetic analysis of two individual eggs from
109 egg clusters. Laboratory analysis for all adults was done in the same laboratory as the
eggs. I performed the bench work for the adults with the assistance of Cherril Bowman, a
senior research technician in the MGL at WDFW.
A description of the methods used to collect and store buccal swabs, followed by
the methods commonly used for population genetics studies is below. I then discuss the
methods used to conduct the laboratory component of the analysis, and finally I describe
the analyses used to assess parentage with the program CERVUS and population
structure with the program COLONY.
FIELD METHODS
Buccal swabs (Epicentre©) were used to collect samples from each adult Oregon
spotted frog by swabbing the inside and back of the mouth. Mouth swabs were performed
in duplicate for each animal in order to ensure high yields of DNA in the extraction
process. These swabs were dried immediately in the field, and stored at -20°C according
67
to the label. See Appendix B for field method protocol. A total of 162 samples were
collected from adult Oregon spotted frogs during this study (n=81 x duplicate samples),
and a total of 218 offspring (n=109 egg masses x 2 offspring).
SOURCES OF DNA
Obtaining DNA samples from organisms is a key component of conducting an
analysis such as the one described here. Beyond collecting the samples, there is a lot of
basic research that goes into determining which genes will give enough power to assess
population structure or conduct parentage analyses, for example, and finally developing
primers to isolate them. There are currently three ways of targeting regions or types of
DNA that are commonly used in genetic analysis for population studies: microsatellites,
mitochondrial DNA, and Single Nucleotide Polymorphisms (SNP). These common
regions are typically chosen based on the objective of the study, and are used to represent
genetic differentiation between individuals within a population. These approaches
represent genetic variability differently, as they can represent chromosomal DNA
(microsatellites, SNP), mitochondrial DNA, or areas across the genome (SNP). For this
study, microsatellites were used because the primer sets to isolate them have already been
established for the Oregon spotted frog. A discussion of mitochondrial DNA and Single
Nucleotide Polymorphisms can be found in Appendix C. Additionally, microsatellites are
commonly used for parentage analyses (Phillipsen et al., 2009; Blouin et al., 2010), and
there were 12 markers (loci) available through the MGL. These characteristics
sufficiently give this study a high degree of power.
68
MICROSATELLITES (MICROSATELLITE LOCI)
Microsatellite loci consist of tandem repeats of sequences which are between 1
and 6 nucleotides long and repeat between 5 and 100 times (Allendorf et al., 2013; Jehle
& Arntzen, 2002). Microsatellites are present in every eukaryotic genome and typically
occur in large numbers (Jehle & Arntzen, 2002). In population genetics, genome
mapping, and parentage analyses, microsatellites are the most commonly used DNA
markers (Allendorf et al., 2013). Microsatellites are very common among similar species
and therefore primers can frequently be used more universally than other loci types
(Allendorf et al., 2013). Furthermore, microsatellites tend to have high mutation rates due
to slippage during DNA replication, and show high levels of genetic diversity, even in
small populations (Allendorf et al., 2013; Frankham et al., 2003). ). Microsatellite loci are
commonly used in amphibian population genetics and have been used to understand the
genetic variability of Oregon spotted frog populations (Blouin et al., 2010; Phillipsen et
al., 2009). They are also used for looking at genetic diversity of other threatened species
(Frankham et al., 2003).
69
LABORATORY METHODS: MICROSATTELITES
DNA EXTRACTION: SOLID PHASE
Once samples were collected from adult OSFs and stored, the DNA needed to be
extracted because samples contained substances other than DNA (Butler, 2005). There
are multiple methods for extracting DNA from samples, but Solid Phase extraction using
silica bead columns, enables high-throughput DNA extractions, and is widely available
for purchase from Qiagen (Butler, 2005). First, tissue was removed from the swabs, and
lysed using a proprietary lysate solution (Qiagen).
For this study, Qiagen DNA extraction kits (DNeasy Blood and Tissue Kits) were
used for extraction from mouth swabs and eggs. Extracted DNA was then stored in 96
well plates at 4°C in the short term, and -80°C for the long term (Butler, 2005). DNA was
extracted independently from all of the mouth swabs collected in the field to minimize
cross contamination between duplicate samples.
POLYMERASE CHAIN REACTION (PCR)
After the DNA was extracted from samples, PCR amplifications were run with 12
previously developed markers (loci), some of which were used by Blouin et al. (2010).
These fluorescently labeled primers were used to isolate and amplify the microsatellites.
“PCR is an enzymatic process in which a specific region of DNA is replicated over and
over again to yield many copies of a particular sequence” or region (Butler, 2005).
Primers were annealed to the 3’ and 5’ ends of each DNA strand, and billions of copies of
the region of interest were produced.
70
Depending on the type of primer used, the annealing temperature differed (Table
3.2.1). For this reason, the primer sets were run in Multiplex, or combinations of different
primers in the PCR reactions based on the annealing temperature requirements of the
primer sets. Multiplexing streamlined this process by allowing me to use fewer resources
and PCR blocks while working in the laboratory (Table 3.2.1). Total volume for each
PCR reaction was 10μL, with the following final concentrations: 1μL template genomic
DNA, 1.5mM MgCl 2 , 20μM dNTPs, 1X Promega PCR buffer, 50μM Promega Gotaq®
(taq polymerase), and diH2O (Deionized). Some of the PCR reactions followed a “touchdown” protocol, while others were amplified using a GO reaction (Table 3.2.1). Touchdown PCR began with an initial two minute denature at 94°C; then 3 cycles of 94°C for
30 seconds, followed by 30 seconds of annealing temperatures (Table 3.2.1), then 72°C
for 1 minute; this process was then repeated 36 times; finally, the reaction was held at
72°C for 10 minutes, and held at 10°C in the PCR block until they were stored at 4°C for
preparation for the 3730. GO PCR began with an initial two minute denature at 94°C;
then 39 cycles of 94°C for 30 seconds, then varying annealing temperatures (Table 3.2.1)
for 30 seconds, then 72°C for 1 minute; finally, the reaction was held at 72°C for 10
minutes, and held at 10°C in the PCR block until they were stored at 4°C in preparation
for the 3730.
71
Table 3.2.1. Table represents: Column 1 – either the primer sets were run in multiplex or
on their own, based on annealing temperature (Column 6). Column 2 signifies the marker
type used, Column 3 is the fluorescent dye tag color (red, green, blue, or yellow), Column
4 is the primer number. Column 5 represents the type of PCR reaction performed.
Multiplex/
Single
Marker
Dye
Label
Primer
Number
Reaction
Annealing
Temperature °C
RP26
VIC
2748, 2749
TouchDown
SFC120
NED
2760, 2761
58°- 55°
SFC134
VIC
2762, 2763
RP415
NED
2756, 2757
TouchDown
57°- 50°
RP17
PET
2742, 2743
RP15
6FAM
2740, 2741
RP461
VIC
2758, 2759
TouchDown
50°- 45°
RP22
NED
2744, 2745
S1
RP193
6FAM
2752, 2753
Basic GO
50°
S2
RP23
VIC
2746, 2747
TouchDown
50°- 45°
S3
RP385
6FAM
2754, 2755
Basic GO
50°
S4
RP3
VIC
2750, 2751
Basic GO
50°
M1
M2
M3
3730 GENETIC ANALYZER
The 3730, like many other genetic analyzers, uses fluorescent dye tags, which flag
the primers, or other regions of the amplified DNA (Butler, 2005). PCR products were
loaded with primers in sets (or multiplexes) in 96 well plates (Butler, 2005). The 3730
Analyzer at the MGL uses a series of capillaries to detect the fluorescence and size of the
DNA fragments (C. Bowman, personal communication, 2014). The output file has
72
colored (blue, green, red, yellow) label peaks (DNA size and quantity) for each present
allele for each animal sample (Butler, 2005). The actual size of these DNA fragments,
and DNA locus genotypes (alleles), must then be calculated using an algorithm called the
Local Southern Method in a genotyping software program called GeneMapper (Butler,
2005). The samples in this study were run on ABI 3730 DNA Analyzer either in
multiplexes or as single reaction sets. Alleles were sized based on number of base pairs
and calculated using the local southern method based on the GS500LIZ_3730 internal
lane size standard.
LOCAL SOUTHERN METHOD & GENEMAPPER
With microsatellites, the number of tandem repeats that occur at each locus
indicates the allele (Butler, 2005). These peaks are then compared to an allelic sizing
ladder, which includes known sizes for each allele (Butler, 2005). The loci, which look
like colored peaks in the file output, are sized using an “internal sizing standard”, called
the Local Southern Method (Butler, 2005). In this study, the internal sizing standard was
the GS500LIZ_3730. For each allele, this method calculates the size of two peaks on both
sides of the unknown peak being measured (Butler, 2005). The product was an allele
genotype, which is the size/number of tandem repeats for an allele (Butler, 2005). For
this study, this process was done in a program called GeneMapper, a commonly used
program for scoring microsatellites (Butler, 2005).
Occasionally, scoring errors occur with this and other programs, typically from
background fluorescence from other microsatellites run in the same matrix/multiplex
(Butler, 2005). For this reason, I first allowed GeneMapper to score the alleles based on
73
its own programming, then systematically went through the generated scores to see if I
agreed with them. Finally, I checked all scores with an experienced technician, Cherril
Bowman.
The markers used for this study (n=12) and allele lengths were previously
established and have been used in previous analyses conducted with the MGL. These
allele lengths were in the form of bins, or shaded areas on the screen at a certain value.
For example, the allele type 208 for the marker RP18 represents 208 base pairs, this allele
was identified prior to my study, and it was therefore considered a bin. Each scored allele
was put into these previously identified bins (alleles); no new alleles were discovered in
the West Rocky Prairie adult population. Finally, these scored alleles were put together in
a series of numerical values for each allele in a spreadsheet, giving me an output of the
sample (or animal) genotype, based on the microsatellites used (Butler, 2005).
Within the same marker, 2 alleles would be scored. If there was a single peak at a
number of base pairs, the individual was scored as homozygous. This means that both
parents, according to Mendelian genetics, contributed the same allele. For example, for
the loci marker RP15, a homozygous individual would have a score of 200, 200.
However, if there were two peaks for an individual, that individual was deemed
heterozygous. For the loci marker RP15, a heterozygous individual would have a score of
200, 208, which means that each of the individual’s parents contributed different alleles.
One element of scoring error may be the presence of null alleles, which are the absence
of one microsatellite from one parent due to mutations at the primer’s annealing site
(Chapuis & Estoup, 2007). This can cause a misrepresentative higher ratio of
homozygous alleles (Chapuis & Estoup, 2007), and is represented in the allelic richness
74
file outputs in parentage programs (see Parentage Below). These scores (number of base
pairs) were then exported as a spreadsheet, which were then used to conduct multiple
genetic analyses, including parentage assignments. Spreadsheets for both offspring and
adult genotypes were developed, and used for parentage analyses. These spreadsheets
included the sampled individuals, each of the 12 markers, and the allele scores.
MATCHING DUPLICATE SAMPLES
In order to determine whether the same adults were sampled more than once, an
MS Excel plug-in, Microsatellite-Toolkit, which looks at repeat genotypes, was used. I
used this method to determine the accuracy of the dorsal pattern recognition method used
in the field component of the study. The output matches samples based on the scored
alleles of all tested samples. Between two samples, it gives a score, which is the percent
of alleles that match each other within the samples, the number of alleles that were
compared, and the number of alleles that match.
PARENTAGE ANALYSIS: METHODS USED
Based on the literature, natural history, collection methods, and expert advice by
Kenneth Warheit Ph.D and Maureen Small Ph.D, a number of methods were used to
determine parental assignments for this project. As described below, there are five
different types of methods to choose from for this type of analysis (Jones et al., 2010),
and each researcher uses varying methods, depending on the objective of their study. The
types of approaches used for this project include categorical allocation using a likelihood
approach with the program CERVUS 3.0.7; Parental reconstruction, a commonly used
75
method for amphibians, was performed using the program FRANz, although the results
of this analysis are not reported (Riester et al., 2009); and finally, the program COLONY
(Jones & Wang, 2009) was used to determine parentage, the relationships between
individual adults, and the effective population size. This section begins with a discussion
of the two overarching methods used by researchers for parentage analyses. It then details
the specific methods and programs used to run the parentage analysis in this study, and
the methods and program used to determine the effective population and adult sibling
relationships. See Appendix D for a discussion of the other types of methods and
programs that can be used to meet different objectives for parentage analyses. Appendix
D outlines the methods I also used for determining parentage using the programs FRANz
and COLONY, as this information is only briefly discussed here, and not reported.
The two overarching approaches to conducting a parentage analysis are exclusion
and likelihood. Parentage analysis by exclusion is preferable and relatively simple to
understand, but is typically difficult to achieve. Therefore, multiple approaches to
parentage analysis by likelihood have been designed for use when exclusion is not
achievable.
EXCLUSION
Exclusion is considered the simplest technique in parentage analysis, and “is
based on Mendelian rules of inheritance”, where each parent contributes a unique allele
to the offspring (Jones & Ardren, 2003). The genotypes of a candidate parent are
compared with that of the focal offspring (Jones et al., 2010). “Any candidate parent who
fails to share at least one allele with the offspring at any locus is eliminated from
76
consideration” (Jones et al., 2010). While perfect exclusion is the ideal, it is often
difficult to achieve because parents may share, and therefore contribute the same alleles
to their offspring. There are also multiple weaknesses to strict exclusion approaches, as
genotyping errors, null alleles, and mutations often produce false exclusions (Jones &
Ardren, 2003). These errors also become more common as datasets become larger
because of scoring errors (discussed in GeneMapper section). When assigning parents to
offspring, exclusion of individuals is attempted first, to exclude as many parents as
possible. For the remaining non-excluded parents, parentage assignment/allocation is
performed using statistical likelihood methods.
LIKELIHOOD
Likelihood methods either “assign progeny to nonexcluded parents based on
likelihood scores derived from their genotypes” (Jones & Ardren, 2003), or they use a
posterior probability to assign progeny (Jones et al., 2010). To assign parents to offspring
based on likelihood, a likelihood-ratio is calculated for each adult using a frequentist
statistical approach based on hypothesis testing. A likelihood ratio is calculated as the
likelihood of paternity or maternity of a sampled adult/parent compared to the likelihood
of paternity or maternity of an arbitrary adult/parent (Marshall et al., 1998). Alternatively,
posterior probability uses a Bayesian approach, where parents are allocated with a
probability using a known maternal or paternal genotype. Currently, there are five
methods used for parental allocation, all of which have likelihood and posterior statistical
approaches. Categorical Allocation and Sibship (Sibling Relationship) Reconstruction
were specifically used for this project; other methods are discussed in Appendix D.
77
CATEGORICAL ALLOCATION & CERVUS 3.0.7 ANALYSIS
For parentage analysis, categorical allocation is the most commonly used and is a
“method to choose the single most likely parent from a group of nonexcluded putative
parents” (Jones et al., 2010). This is the primary method used for assigning parents to
offspring in this project. Similar to exclusion, this method also requires a set of two
candidate parents and a single offspring, and it serves as an excellent alternative for cases
where perfect exclusion cannot be achieved (Jones et al., 2010). Both posterior
probability and likelihood approaches can be used for categorical allocation (Jones et al.,
2010); a likelihood approach was used for this study. Both approaches adhere to
Mendelian transition probabilities, which is “the probability of the offspring’s genotype
given the genotypes of the mother and father” (Jones et al., 2010; Marshall et al., 1998).
With categorical allocation, the entire offspring is assigned to a parent within the sample
that has the highest likelihood or posterior probability of being the actual parent (Jones et
al., 2010). Categorical allocation can be applied when a parent is already known for an
offspring, or when no parents are known, as is the case with this study; it can also be used
to assign one parent or parent pairs (Jones et al., 2010). This method can deal with
scoring errors (See Laboratory Methods in this chapter) and mutations, and can calculate
confidence in parentage assignment (Jones et al., 2010).
A program called CERVUS 3.0.7 is widely used to perform this type of analysis,
and was used to determine parentage for this project. CERVUS 3.0.7 requires a number
of assumptions, which were met by this study; it analyzes markers, such as
microsatellites and SNPs (Described in Appendix C), assumes that the organism is
diploid (two alleles per locus), and assumes that markers are inherited independently
(Kalinowski et al., 2007). CERVUS 3.0.7 performs parentage analyses in three steps,
78
beginning with an Allele Frequency Analysis, followed by a Simulation of Parentage
Analysis, and finally the Parentage Analysis.
Allele Frequency Analysis
Allele Frequency Analysis is required for parentage testing that uses likelihood
(Kalinowski et al., 2007). A file with all genotypes (scored spreadsheet) from all sampled
members of the population, including offspring and parent genotypes, were imported into
CERVUS 3.0.7. From this genotype file, the frequency (how many and how often) of
each allele present in the population is calculated for each locus tested (Kalinowski et al.,
2007). It also calculates a suite of summary statistics for Hardy Weinberg Equilibrium,
and chi-squared statistics for observed and expected heterozygosity (Kalinowski et al.,
2007). This allele frequency file helps the program and researcher determine the
suitability of the loci used for the continued analysis.
Simulation of Parentage Analysis
Simulation of Parentage Analysis is “used to calculate critical values of likelihood
ratios (test statistic), so that when parentage analysis is carried out using real data, the
confidence of parentage assignments can be determined” (Kalinowski et al., 2007). The
allele frequencies file is used to run the simulation, along with the number of likely
parents in the population and the percentage of the population that was sampled. For the
population at WRP, 336 egg masses were laid in 2014; therefore, based on a 1:1 female
to male sex ratio (Phillipsen et al., 2009), about 336 males and 336 females are expected
to be in the population. The number of simulations run for this analysis was 100,000, as
79
suggested by the program to maximize power. Additionally, given that there are about
336 males and 336 females, the number of possible offspring genotypes is approximately
112,896, making 100,000 simulations an appropriate number.
Parentage Analysis
Parentage analysis uses both the simulated file and the allele frequency file to
assign offspring to parents (Kalinowski et al., 2007). For each offspring sampled, the
most likely parent was assigned based on the pre-determined level of confidence from the
simulation or the offspring is left unassigned (Kalinowski et al., 2007). The parentage
analysis that was run for this study included the parents gender based on field
observations. The minimum number of loci typed was 6, meaning that if a candidate
parent or offspring had less than 6 loci scored, they were not included in the analysis. The
confidence results were partitioned into three different categories: strict assignment (95%
confidence), relaxed assignment (80% confidence), and unassigned, which was
partitioned into two sub-categories a) the most likely candidate that was not assigned
parentage and b) unassigned (blank). In the analysis for this study, only the most likely
candidate parent was assigned to each offspring.
Parentage-analysis output files included the identification of each offspring and its
most likely mother, father, and family set. They also included a non-exclusion
probability, which is the probability that the individual is included. The number of loci
that were typed for each assigned parent and offspring, the number of loci that were
compared between the offspring and parent, and the number of loci that do not match
between the offspring and the parent were also in the output files. They also gave a pair
80
(single parent and offspring) and family (both parents and offspring) an LOD score,
which is the log of the overall likelihood score. A positive LOD score means that the
assigned parent is more likely to be the actual parent than not the actual parent (>50%
likelihood) (Kalinowski et al., 2007). An LOD score of zero means that the assigned
parent is equally likely to be the parent as it is not to be the parent (50% likelihood)
(Kalinowski et al., 2007). A negative LOD score means that the assigned parent is not
likely to be the actual parent (<50% likelihood) (Kalinowski et al., 2007). This LOD
score is directly related to a Delta score, which is given for each pair and family
(Kalinowski et al., 2007). The Delta is the difference in the LOD scores between the most
likely parent and the second most likely parent. Finally, a measure of confidence is in the
output file, where * is 95% confident assignment, + is 80% confident assignment, - is the
most likely parent that was not assigned parentage, and blank is unassigned (Kalinowski
et al., 2007).
Assignments were determined by a number of criteria that first relied on the
assignment of the same parent to both offspring within an egg mass, based on the
assumption that a single female lays an egg mass, and a single male fertilizes it
(Phillipsen et al., 2009). For example, for offspring A and B from egg mass “1”, if
different mothers were assigned to A and B, then the offspring were not included in the
final results. If offspring A and B from egg mass “1” were assigned to the same mother,
the confidence was then assessed. Parent:offspring pairs with confidence as the most
likely unassigned parent, 80% confidence, and 95% confidence were included; those left
unassigned (blank) were not included for further analysis. LOD scores were then looked
at, and if both offspring were assigned to the same parent with positive LOD scores they
81
were included. Negative LOD scores were infrequently included, unless it was the most
likely candidate and the other offspring was assigned with confidence (>80%) to the
same most likely candidate parent. Delta scores were then assessed. If Delta scores were
the same or higher than the LOD score for that parent:offspring pair, the pair was
included. If Delta was substantially lower than the LOD score, zero, or negative, the
parent:offspring pair was not included. If the same parent was assigned to multiple egg
masses, the parent pair was not included unless one of the assignments had >80%
confidence. These methods allowed me to determine the parent:offspring assignments
using categorical allocation and the program CERVUS.
FRANz ANALYSIS & PARENTAL RECONSTRUCTION
Besides using the program CERVUS, the program FRANz was also used to
initially assess parentage for this study. However, FRANz uses a parental reconstruction
likelihood approach, which requires >10 offspring from a single egg mass for adequate
reconstruction. For this reason, the parentage assignments from FRANz are not reported,
although a detailed description of the methodological approach FRANz uses in parental
reconstruction is outlined in the Appendix D.
SIBLING RELATIONSHIP (SIBSHIP) RECONSTRUCTION & COLONY ANALYSIS
This approach is meant for studies where parents are not available for parental
assignment, or it can be used to look at the relationships between parents. For the
purposes of this study, it was used to assess the relationships (relatedness) between
parents, and the effective population size (N e ) at WRP. This program can also be used for
parentage analysis. It uses siblings, both full and half-siblings, to reconstruct the parental
82
genotypes, much like parental reconstruction except that the offspring are not assigned
parents when they are not available (Jones et al., 2010). The siblings are assigned to
different classes of relationships, typically full and half sibling relationships (Jones et al.,
2010). Once the groups are identified, Sibship’s can be used to reconstruct parental
genotypes, and be used for parental analysis (Jones et al., 2010). Both likelihood, and
Bayesian posterior probabilities can be used for Sibship reconstructions (Jones et al.,
2010).
COLONY was primarily used to identify the sibling relationships between the
adults sampled by constructing all of the potential genotypes of sibling parents within the
population and identifying those sampled parents that were siblings. These siblings were
then compared to the parentage assignments based on CERVUS to identify whether any
sibling parents were assigned to both eggs from an egg mass; if siblings were assigned to
the same egg mass, they were not included in the parentage results. Also, these sibling
relationships demonstrate the amount of relatedness within the population. Additionally,
COLONY generated an output showing the N e based on all sampled individuals. In order
to assess the relationships between candidate parents, COLONY also ran a parentage
analysis, which is not reported here, but is discussed in Appendix D.
83
SPATIAL ANALYSIS
In ArcMap 10.2 (ESRI, 2015), parents were selected from the total number of
captured individuals, and a parent’s feature-class was made which included all of the
times each of the animals were captured. Then, for each parent, a new feature-class was
made, which included all of their capture locations. Offspring within the parentage
analysis were designated by geographical location (ex. East Side), not a specific
latitude/longitude point for each individual offspring, as this information was not
available. For this reason, the egg mass within each geographical region that had the most
eggs within a cluster was chosen to represent the geographic area, as it can be assumed
that it was the most likely to have the most eggs taken from it. For each geographical
area, the number of individual egg masses present, the number of clusters present at each
location, and the number of egg masses sampled per location was calculated. In ArcMap
10.2, the distance between parent points and offspring points (in meters) was generated
using the Point Distance Analysis tool. This method was chosen because of the variability
of the landscape and lack of knowledge about possible modes of travel between the
points. The animals could have traveled down the channels or traveled across the wetland
by any number of pathways.
84
GENETICS
RESULTS
SAMPLED ADULTS & OFFSPRING: A COMPARISON OF DORSAL PATTERN &
MICROSATELLITE TOOL KIT
Of 81 duplicate adult samples collected in the field, all were genotyped in the
laboratory to confirm whether repeat sampling of the same individual occurred and to
confirm the efficacy of dorsal pattern recognition methods. The final number of adult
individuals (n=56) was determined based on the spot pattern recognition discussed in
Chapter 2, and the microsatellite plug-in for Excel, which was discussed in the Methods
section of this chapter. In order to prepare the genetic data for an accurate parentage
analysis, duplicate parents (n=25) were removed from the parent genotype files. In order
to do this, I compared the “match individuals” based on genetics to the “matched
individuals” based on the high-resolution comparison of the dorsal pattern images. The
results of both comparisons are explained further in Table 3.3.1 below.
Three animal sets showed discrepancies between spot pattern recognition, and
genetic analysis. Based on spot pattern recognition on September 23, 2014, adults 0003,
0022, 0033, and 0048 were matched. However, the genetic analysis did not match adult
0022 to the other three animals because marker Rp15 was likely mis-scored during the
scoring process using GeneMapper. Animal number 0022 was deemed homozygous at
Rp15, while the others were deemed heterozygous for Rp15. Based on the combination
of spot pattern recognition and the likelihood that Rp15 was mis-scored, animal number
85
0022 was removed from the parentage analysis, and considered a duplicate individual.
The representative sample for this group of samples from a single adult is 0003.
Animal codes 0007, 0028, and 0043 were matched based on high-resolution spot
pattern recognition on September 23, 2014. However, due to the lack of the Rp15 and
Rp23 markers in 0043’s genotype, the microsatellite plug-in matching software did not
recognize it as a match. Animal number 0043 was removed from the parentage analysis,
as it is the same animal as the others based on the scores matching for all of the
remaining 10 markers, and spot pattern recognition. The representative sample for this
group of samples from a single adult is 0007.
The genetic analysis matched 3 sampled individuals, 0027, 0034, and 0041.
However, 0027 was initially marked in the field as a female, and 0034 and 0041 were
marked as male. Their spot patterns are nearly identical, as identified on September 23,
2014. Therefore, based on the genetics, and similarity of spot pattern, it is likely that this
animal was misidentified in the field as a female, and is actually the same individual as
0034 and 0041. The representative sample for this group of samples from a single adult is
0034. The results of the microsatellite match program, which matches individuals based
on genotype, are shown below in Table 3.3.1.
A total of 56 adult individuals and 218 offspring were genotyped at 12
polymorphic microsatellite loci for parentage analysis in this study. Of the attempted 218
genotyped offspring, 216 were successfully genotyped.
86
Table 3.3.1. A total of 81 samples were collected in the field, representing 56 unique
individual adults. Based on dorsal pattern recognition and microsatellite Excel plug-in,
this table shows the individuals that were sampled more than 1 time. This table does not
represent all 56 individuals used in the parentage analysis.
Percent Alleles
Typed/Matched
No.
Alleles
No.
Alleles
Matched
100%
24
24
97.90%
48
47
0054
100%
24
24
0005
0032
100%
24
24
0007
0028
0043
100%
32
32
0010
0036
0038
100%
36
36
0013
0051
100%
24
24
0014
0021
100%
36
36
0015
0068
100%
24
24
0019
0030
100%
24
24
0023
0035
100%
36
36
0025
0052
100%
24
24
0027
0034
0041
100%
36
36
0029
0042
0058
100%
36
36
0060
0070
100%
24
24
0074
0078
100%
24
24
0016
0026
100%
22
22
Adult
Sample
Match 1
0001
0055
0003
0033
0004
Match 2
0048
0040
0039
Match 3
0022
N total = 56; Bold Type Font: individual was identified as a repeat based on dorsal-pattern
recognition. See Sample Size results section for details. Data Generated by CERVUS
3.0.7 (Kalinowski et al., 2007), and adapted to include dorsal pattern matches.
87
EFFECTIVE POPULATION SIZE (N e ) ANALYSIS
The program COLONY (Jones & Wang, 2009), described in the Methods section
of this chapter, was used to assess the number of individuals in the effective population
based on the sampled population at WRP. COLONY calculated the N e in three different
ways. Based on our knowledge that OSF breed non-randomly, meaning that they select
their mate, and generally only mate with a single individual in a given breeding year
(Phillipsen et al., 2009), the N e calculation by COLONY based on pair-likelihood score
method, which assumes non-random mating was deemed most appropriate (Wang, 2009).
N e =25 (Table 3.3.2) is consistent with what Phillipsen et al. (2009) found, which was an
N e =36.7 (95% C.I. 19-71.9) based on calculating N e from N b . N b is the effective
population of a single breeding year based on Sibship reconstruction from egg masses
taken in a single breeding season (Phillipsen et al., 2009).
Table 3.3.2. Effective population size of sampled population at WRP: Pair likelihood
score method based on non-random mating.
Alpha:
0.02
Ne:
25
CI95 (Lower):
15
CI95 (Upper):
43
N e (Effective Population Size); CI95 (95% Confidence Interval). Table is based on
algorithms outlined in Wang (2009), and adapted from the generated N e output by
COLONY.
88
ALLELIC RICHNESS ASSESSMENT
The total number of individuals, which includes both adults and offspring,
genotyped in this study is 273. For each of the 12 loci that were used for assessing
genotypes of the WRP sampled population, the average number of alleles represented at
each locus tested in the population is A mean =3.833. The mean expected heterozygosity
(H e ), which is the average of all expected heterozygosity calculated across all loci tested
(Kalinowski et al., 2007), is H e =0.5174. See Table 3.3.3.
Table 3.3.3. Number of individuals, loci tested, average number of alleles, and mean
expected H e
Number of Individuals:
273
Number of Loci:
12
Mean Number of Alleles per Locus:
3.833
Mean Expected H e :
0.5174
H e : Heterozygosity; Table adapted from the output generated by CERVUS 3.0.7
(Kalinowski et al., 2007)
Allelic Richness (A mean ) was measured to determine the variation of alleles
present within the population at each locus observed. This information reflects whether
the representative sample of the population, at specific loci, is at Hardy Weinberg
Equilibrium. Based on chi-square goodness-of-fit analysis with Bonferroni correction,
using observed versus expected heterozygosity for each locus, two loci are out of Hardy
Weinberg Equilibrium (p>0.05 is in HW Equilibrium, p<0.05 is out of HW Equilibrium).
This is likely due to the presence of null alleles (F null) at RP15 and RP461, which are
generally caused by mutations at the annealing site of a primer, and can result in a higher
89
degree of homozygosity than actually occurs within the population (Chapuis & Estoup,
2007). Certain chi-square values were calculated with Yates Correction for continuity,
while others were not. Table 3.3.3 below, shows the number of alleles represented within
the sampled population at each locus, and the number of individuals typed for each locus.
Additionally, Table 3.3.4 demonstrates the PIC (Polymorphic Information Content),
which is the usefulness of each marker for linkage analysis (Elston, 2005) based on this
sampled population.
90
Table 3.3.4. Allelic Richness: Summary table of CERVUS Allelic Richness Output
PIC
Chi
Squared
p
value
HW
Significance
(Bonferroni
Correction)
Null
Allele
(F)
0.499
0.374
148.7596
<0.001
***
0.2734
0.101
0.096
0.091
ND
ND
ND
0.0166
271
0.738
0.801
0.773
10.9599
0.0896
NS
0.0416
6
271
0.631
0.637
0.572
1.1804
0.7577
NS
0.0046
RP23
3
270
0.593
0.602
0.533
3.4311
0.3298
NS
0.0051
RP26
5
271
0.627
0.642
0.568
3.4136
0.3322
NS
0.0112
RP3
4
271
0.435
0.434
0.401
0.1642
0.6854
NS
0.0065
RP385
4
272
0.68
0.722
0.667
1.9667
0.5793
NS
0.0283
RP415
4
269
0.632
0.601
0.529
1.5446
0.672
NS
-0.025
RP461
2
252
0.21
0.321
0.269
28.4627
<0.001
***
0.2075
SFC120
6
272
0.349
0.364
0.341
0.2233
0.6365
NS
0.0188
SFC134
2
269
0.52
0.488
0.369
0.9874
0.3204
NS
-0.033
Locus
Num.
of
Alleles
Individuals
Typed
H e Obs
H e Exp
RP15
2
268
0.873
RP17
2
268
RP193
6
RP22
Bold Type Font – Indicates the use of Yates Correction for Continuity in calculating ChiSquared values; *** Loci is not in Hardy Weinberg Equilibrium; ND: Not Done (No
result to show); NS: Not Significant (Locus is in Hardy Weinberg Equilibrium); H e Obs:
Heterozygosity Observed; H e Exp: Heterozygosity Expected. Data Generated by
CERVUS 3.0.7 (Kalinowski et al., 2007).
91
ADULT RELATEDNESS
Based on the results of the Sibship analysis conducted in COLONY (Jones &
Wang, 2009), it appears that there is a high degree of relatedness between the individuals
in the adult sampled population. Table 3.3.5 shows the number of individuals within the
sampled population (n=56) that have no sampled siblings, one sampled sibling, and two
sampled siblings. These results indicate that 55.4% of the total sampled adult population
has at least one of their full siblings within the sampled population. This level of
relatedness can cause conflict in parental assignment, as siblings will typically have very
similar genotypes, and if the actual parents are not sampled, there is a high likelihood that
the parent’s sibling (aunt/uncle) may be assigned to their niece or nephew (Wang, 2012).
This information was taken into account for the parental assignments discussed in the
Parentage Assignment Section in the methods, and in the results below.
Table 3.3.5. Sibling Relationships: Number of sampled adult individuals with siblings.
Number of Siblings
Number of Individuals
0
25
1
22
2
9
Data compiled from COLONY output (Jones & Wang, 2009)
92
PARENTAGE ASSIGNMENT RESULTS
Assignment of parents to offspring followed the methods outlined in the Methods
section of this chapter with the program CERVUS 3.0.7 (Kalinowski et al., 2007). Based
on log likelihood scores, delta scores, and confidence outputs from CERVUS 3.0.7
(Kalinowski et al., 2007) and the parameters outlined in the parentage Methods section,
parental assignments were determined. A total of 12 individual parents were assigned
offspring pairs in this study. However, no triplicate sets were assigned (mother, father,
offspring). Table 3.3.6. shows the 12 assignments, the number of loci types for both
offspring and the assigned parent, the number of loci that did not match in the
assignment, the LOD and delta scores for each assignment, and the confidence based on
parameters outlined by CERVUS 3.0.7 in (Kalinowski et al., 2007). If an assigned parent
was matched to both offspring of another egg mass, then the number of times it occurred,
the highest LOD and delta scores, and the number of loci mismatched are shown. Finally,
the number of single offspring to which each parent was assigned is shown. Due to the
high degree of relatability within this population, the best available methods were used to
make these parentage assignments. Thus, 10 of the 12 assignments are based on the most
likely, unassigned parent, with 2 of the 12 having an 80% confidence as the most likely
assigned parent.
93
Table 3.3.6. Parentage Assignments based on log likelihood approach from CERVUS 3.0.7.
Candidate Mothers
Offspring
ID
Loci
typed
FQ0010a
12
FQ0010b
12
FR0010a
FR0010b
12
12
FS0002a
12
FS0002b
12
FU0010a
FU0010b
12
12
FU0024a
12
FU0024b
12
FV0007a
12
FV0007b
12
Candidate
mother
ID
0066
0046
0017
0075
0006
0008
Loci
typed
Pair loci
compared
Pair loci
mismatching
Pair
LOD
score
12
12
0
3.13
3.13
-
12
12
0
3.81
3.81
-
11
11
11
11
0
0
6.78
4.4
6.78
2.81
+
-
12
12
0
1.08
1.08
-
12
12
0
2.38
1.36
-
12
12
12
12
0
0
2.27
2.97
2.27
2.97
-
11
11
0
5.01
3.89
-
11
11
0
0.97
0.97
-
12
12
0
4.92
1.38
-
12
12
0
2.37
0.93
-
Pair
Delta
Pair
confidence
# of
times
matched
to other
offspring
pair(s)
0
3
0
0
1
1
Highest
Pair
LOD
Delta
Loci
Mismatched
NA
NA
NA
2.71
1.61
2.4
1.61
0
NA
NA
NA
NA
NA
NA
-1.78
0
1
1.77
1.77
1.65
1.38
-2.31
0
1
# of
times
matched
to single
offspring
10
13
5
8
12
10
Table Continued on Next Page
94
Candidate Fathers
Pair
Delta
Pair
confidence
# of
times
matched
to other
offspring
pair(s)
1
Offspring
ID
Loci
typed
Candidate
father ID
Loci
typed
Pair loci
compared
Pair loci
mismatching
Pair
LOD
score
FU0007a
FU0007b
12
12
0062
12
12
12
12
0
0
5.87
4.66
5.87
0.9
+
-
FV0010a
12
12
12
0
4.78
4.78
-
FV0010b
12
12
12
0
3.78
1.52
-
FV0016a
12
FV0016b
12
FV0027a
FV0027b
12
11
FU0025a
12
FU0025b
12
FQ0006a
9
FQ0006b
12
0023
0005
0016
0004
0061
12
12
0
3.3
3.3
-
12
12
0
3.31
3.31
-
12
12
12
11
0
0
6.28
4.01
2.95
0.08
-
12
12
0
2.42
0.13
-
12
12
0
2.33
0.12
-
12
9
0
2.8
0.68
-
12
12
0
3.02
0.68
-
0
0
0
0
0
Highest
Pair
LOD
Delta
0.02
-2.89
0.02
0
1
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Loci
Mismatched
# of
times
matched
to single
offspring
7
5
3
6
8
2
Fx indicates the number and general location of the egg masses the eggs were taken from, a and b are the offspring; Candidate parents
are assigned to offspring Fx a and b; (-) Indicates that the candidate parent is the most likely parent; (+) indicates that the candidate
parent is the most likely parent with 80% confidence. Data generated by CERVUS 3.0.7. (Kalinowski et al., 2007).
95
SPATIAL ANALYSIS OF PARENT AND OFFSPRING LOCATIONS
As described in the Methods section of this chapter, the distance between each
parent and their offspring was calculated and mapped using ArcGIS 10.2 (ESRI, 2015).
Nine of the 12 parents were captured in the South Pond (henceforth, Small Pond). Of
those nine individuals, four (0062, 0004, 0075, 0006) laid their eggs in the West Central
location, which is 101.91 meters away from the pond (Table 3.3.7, Figure 3.3.1). Another
four (0023, 0005, 0016, 0008) laid their eggs in the West North location, which is 157.58
meters away from the South Pond (Table 3.3.7, Figure 3.3.1). The final animal from the
South Pond (0061) fertilized an egg mass in the East side area, which is 2,147.83 meters
away (2.15 km). Animal number 0017 was captured in 3 different locations, all within
very close proximity to each other (2 in the ECH next to the pond, and one time in the
PND). The distance for this animal was calculated for all 3 locations, and then an average
distance was taken from the 3. This animal laid eggs in the Central West location, which
is an average of 404.94 meters away. Animal number 0046 was captured in the northern
extent of the West Channel, and laid her eggs in the Central East location, 1079.07 meters
(1.08 km) away. Finally, animal number 0066 was captured in a portion of the East
Channel, close to its junction with the North Channel. The animal’s eggs were laid in the
East side oviposition area, 2230.15 meters away (2.23 km).
96
Table 3.3.7. Straight-Line Distance between Parents and Offspring
Parent
Mother/
Father
Offspring
Location
Total Egg
Clusters
at site
Masses
at site
Eggs
sampled
Parent
Location
Distance
(m)
0062
Father
West Central
171
5
56
South Pond
101.91
0023
Father
West North
112
2
56
South Pond
157.58
0005
Father
West North
112
2
56
South Pond
157.58
0016
Father
West North
112
2
56
South Pond
157.58
0004
Father
West Central
171
5
56
South Pond
101.91
0061
Father
East
15
3
30
South Pond
2147.83
0066
Mother
East
15
3
30
East
Channel
2230.15
48
North
section of
West
Channel
1079.07
404.94*
0046
Mother
Central East
24
2
0017
Mother
Central West
9
5
18
South
Pond/East
Channel
0075
Mother
West Central
171
5
56
South Pond
101.91
0006
Mother
West Central
171
5
56
South Pond
101.91
0008
Mother
West North
112
2
56
South Pond
157.58
No Parent
Assigned
NA
West South
5
2
10
NA
NA
Parent: The assigned parent & gender. Offspring location: Assigned offspring location.
Total Egg masses at site (n): egg masses present at each oviposition site. Egg sampled
(n): The number of eggs sampled at the site where the offspring came from. Distance (m):
The straight-line distance the parent traveled from the assigned oviposition site to nonbreeding summer habitat. *Indicates that the animal was captured in multiple locations.
97
Figure 3.3.1. Locations of OSF parents (n=12), and their assigned egg mass locations (n=12). Right: West Rocky Prairie showing
parents that traveled >400 meters to oviposit (n=4); From the South Pond, (n=1) parent went to the East Side, (n=1) parent went to
Central East, and (n=1) parent went to Central West; From the East Channel, (n=1) parent went to the East Side. Left: Area in Orange
box Zoomed in (n=8); From the South Pond, (n=4) parents went to the West Central oviposition site, and (n=4) parents went to the
West North oviposition Site. *World Imagery Base Map by ESRI 2015; coordinates collected by WDFW (egg masses) & Chelsea Waddell, volunteers, and WDFW
employees (adults), 2014. Map developed by Chelsea Waddell (2015)
98
CHAPTER 4
DISCUSSION
PARENT OFFSPRING LOCATIONS
The locations of parents in relation to their offspring was measured and reported
in the Genetics Chapter Results section. The distance parents (n=12) traveled between
oviposition and active summer habitat were striking for a few of the animals, with the
highest straight-line travel distance being 2.23 km. It is likely, however, that adult OSF in
this study traveled farther than estimated using the straight-line distance, as they typically
travel by water. However, Watson et al. (2003) showed that adult OSF in lowland
western Washington typically travel several hundred meters to recolonize an area during
the post breeding season, based on radio telemetry. Whereas McAllister & Walker (2003)
showed that three adults, one male and two females, traveled an estimated 2.36 km by
water between seasons at Dempsey Creek in lowland western Washington. These
consistent findings suggest that adult OSF can travel substantial distances between late
winter breeding and the summer non-breeding season. Additionally, this travel distance is
likely closely linked to habitat availability, given the high abundance of individuals
within the South Pond, and the fact that one parent traveled 2.15 km and resided in the
South Pond. These consistent findings also suggest that the parentage assignment method
coupled with the spatial analysis used in this study gave a robust and detailed view of
OSF behavior as well as population structure. The straight-line distance measurement of
distance traveled in this study is a conservative estimate of the distance these animals are
99
traveling throughout the landscape. Therefore, future spatial analysis of this data should
investigate potential pathways these parents are traveling by water, as they may be
traveling farther distances than currently realized.
Watson et al. (2003) showed that OSF movement slows during the dry summer
months (Watson et al., 2003). Capture/recapture data from this study can be used to
determine the active summer season movement range of each captured adult at West
Rocky Prairie at a future time. It is likely that the adults captured during the summer
season actually moved to those dry season habitat locations during the wet spring (post
breeding) season, based on Chelgren et al. (2008).
Interestingly, in past studies of the populations at WRP, egg masses found at the
East Side, West Side, and Central areas have been considered unique populations, and
estimates of the population size at the WRP site have been established based on the
number of egg clusters found at the East and West Sides (M. Hayes, personal
communication). Parents of offspring found across all three general locations are utilizing
habitat within the West Side survey area, so these populations should not be considered
separate. Also, the calculation of egg masses for WRP should begin to include the Central
locations. If this were done for the 2014 year, the number of offspring would be 336, and
not 303. Therefore, the estimated breeding population size is actually 672, and not 606, as
was estimated prior to this study. This indicates that the census population is likely much
higher than may have originally been estimated. Most importantly, the population at
WRP should be considered much larger than originally estimated by the West and East
Side Oviposition areas (Tyson & Hayes, 2014), where the Central Oviposition area
should also be included. The calculated effective population (N e ) represents the entire
100
population at WRP, including all unique adults and egg masses from the West, East, and
Central Oviposition areas.
POPULATION STRUCTURE
EFFECTIVE POPULATION, ALLELIC RICHNESS, & RELATABILITY
The result that 55.4% of the sampled adult population had at least one full sibling
within the sampled population indicates that some level of inbreeding may be occurring
within the West Rocky Prairie population. This is further supported by the small effective
population size (N e =25), which is consistent with, but smaller than, the effective
population size (N e =36.7) of adults at the Sunriver, Oregon population studied by
Phillipsen et al. (2009). However, the N e in these two studies is calculated differently,
which may explain this discrepancy. The Sun River population, studied by Phillipsen et
al. (2009), also had an estimated breeding population of 90 adults at the time of the study.
While small N e is typical of pond breeding amphibians (Phillipsen et al., 2009) and ranid
frogs (Hoffman et al., 2005), based on the results of the parentage analysis, the breeding
population at WRP is projected to be around 672 individuals, based on the 1:1 male to
female sex ratio used by Phillipsen et al. (2009). While these populations are in different
locations, the Sun River population has an estimated breeding population that is 13.4% of
the size of the estimated breeding population at WRP, with a comparable N e . For the
seven microsatellite markers used in Phillipsen et al. (2009), no more than 3 alleles were
present at any locus, and the authors indicated that there was potentially a high degree of
relatability in their studied population. This is consistent with our findings where we
found that the WRP population had low allelic richness (mean AR=3.833) and high levels
101
of relatability. This may indicate that genetic drift is occurring within the WRP
population. Genetic drift occurs when infrequently present alleles are lost from a
population (Kliman et al., 2008). Blouin et al. (2010) found a consistent level of mean
AR=2.46 across the OSF’s range. Given the consistency of findings between these
multiple studies, OSF populations tend to have low levels of allelic richness (AR) across
their range, although this may be based on how robust the markers are for the OSF.
Blouin et al. (2010) cautioned that isolated populations of Rana pretiosa may experience
levels of inbreeding depression, and genetic drift, and that management should focus on
mitigating these affects. Therefore, the risks of inbreeding depression and genetic drift
within this WRP population may need to be further investigated in future analyses of this
study’s results and others results.
HABITAT AVAILABILITY & POPULATION SIZE
The effective population size observed in this analysis may indicate a recent rapid
increase (bottleneck) of OSF at WRP due to increased habitat availability from
management practices (Maureen Small, personal communication). Since the summer of
2000, sporadic mowing treatments of invasive reed canary grass (See Introduction) have
been done at the WRP site to increase breeding habitat for the species. In 2000, Kapust et
al. (2012), found 107 egg masses between the West Side survey area, and the central
oviposition areas. Based on the same 1:1 male to female ratio, an estimated 214 breeding
adults were within the WRP site, although this did not include the East Side oviposition
area. Based on the observations of Tyson & Hayes (2014), and the findings from this
study, a total of 672 individuals are part of the WRP site, with 15 egg masses on the East
102
Side. The population has therefore increased more than three times since 2000, which
may be linked to the mowing treatments at WRP (Kapust et al., 2012).
The findings from our research bring to light some important issues about habitat
availability and increasing population size. There was likely a recent bottlenecking
because of a small starting population, which then likely rapidly increased since breeding
habitat availability was increased. However, the amount of non-breeding habitat appears
to be limited, according to the results of the parentage analysis component of this study.
Individuals appear to be breeding in potentially far away locations and are primarily
located within a small pond (South Pond). Therefore, if we wish to continue this
increased population trajectory, we may need to consider increasing non-breeding habitat
availability for these small populations in the form of small ponds similar to the size and
hydrology of the South Pond.
Increases in OSF populations, such as the one at WRP, are beneficial for the
recovery of this species, but the low level of allelic richness and relatedness in this
population and other OSF populations may be cause for concern. Genetic rescue may be
necessary for these rapidly increasing populations, as suggested by Blouin et al. (2010).
However, the use of captive rearing should be implemented carefully, as the addition of
locally poorly adapted individuals may actually harm established populations. This
perspective is reinforced by statements made by Blouin et al. (2010).
103
GENETIC MARKERS
Twelve microsatellite loci were used to assess relatedness, N e , and parentage in
this study. While this number of markers is typical for studies assessing the OSF (Blouin
et al., 2010; Phillipsen et al., 2009), the addition of new markers for the analysis of OSF
genetics would be beneficial. The addition of SNP (Single Nucleotide Polymorphisms),
and more microsatellite markers will likely increase the resolution and robustness of
genetic analyses for this species. The addition of more robust markers would have likely
increased our ability to define parents with higher confidence, as the level of relatedness
within the population, and the low levels of allelic diversity at each marker, did not fully
represent the distinction between candidate parents. Furthermore, the addition of new
markers will aid the future management of this federally listed species. Additionally, the
low level of diversity within the markers used, and the high level of relatedness within
the population, may have caused assigned parents to be the sampled sibling or first cousin
of the actual, un-sampled parent. However, the methods used to distinguish parents were
done with a high level of stringency. Most importantly, even if the selected candidate
parent is the sibling of the actual un-sampled parent, this still indicates that WRP inhabits
a single OSF population, and that the members of the population are moving across the
landscape to find non-breeding habitat. Investigating the site fidelity of individual
breeding frogs across breeding years, and their subsequent offspring, would further
determine if there is consistency in site selection within families to support this
conclusion. Currently, OSF adults tend to utilize similar sites across years (Tyson &
Hayes, 2014), but a more focused analysis of individual’s site fidelity may provide new
information about breeding habitat selection for this species.
104
LAND COVER CHARACTERISTICS
Based on this study, adult OSF at West Rocky Prairie preferred open water habitat
over vegetated habitat, as 100% of the frogs captured in this study were found in open
water. This is consistent with their fully aquatic nature (Green et al., 1997). However,
Watson et al. (2003) found that adult OSF were found under hardhack (Spiraea
douglasii), which was designated in this study under scrub/shrub/willow. Watson et al.
(2003) used radio telemetry to identify the locations of adult OSF, whereas this study
depended on VES (Visual Encounter Survey), which likely decreased our ability to detect
adult OSF under heavy vegetation.
Between the first survey session (July to August) and the second survey session
(August to September), there was a noticeable decrease in water availability across the
landscape (see Habitat Survey Chapter Results). This is likely due to the nature of the dry
season in lowland western Washington, and consistent with the findings of Watson et al.
(2003). This decrease in water availability may be a factor in non-breeding habitat
limitation for adult OSF. For example, animal number 017, a parent, was captured twice
in the East Channel adjacent to the South Pond during the first survey session. During the
second survey session, when that portion of the East Channel no longer had water, animal
017 was captured in the South Pond. This suggests that animals are moving to the most
available water source, and, as demonstrated by the high number of individuals found in
the small pond (South Pond), areas with year-round water sources are limited throughout
the year. Further investigation of the movement across the WRP landscape of adults in
this study may bring to light important information about the movement of adults during
the non-breeding season.
105
An initial hypothesis of this project was that adult OSF may be using Tilley Pond
during the summer season. However, only one adult OSF was observed there. This
suggests that either adult OSF were actually present in the location and not detected, or
that they are not actively utilizing that pond during the non-breeding season. Tilley pond
was quite deep, and needed to be surveyed using Floating VES. This may indicate that
adult OSF prefer more shallow water during the non-breeding summer season, as is
present in the channels and South Pond. Watson et al. (2003) found that tracked OSF
preferred remnant pools (23.6 ± 1.0 cm) in the non-breeding summer season (JuneAugust). Although, the small South Pond, with a high concentration of individuals, was
deeper (chest height) than what Watson et al. (2003) found. Photo documentation of the
South Pond water level meter was collected for this study, and an average water depth
will be computed in further analysis associated with this study.
LOCATIONS OF ADULTS
A majority (83%) of adults captured during this study were present in a small
pond, called South Pond. This pond, as discussed previously, is 10m × 6 m, and was
surveyed 13 times during the study. This predominance of adult OSF in the small pond
could indicate that there is limited suitable habitat for adult OSF at West Rocky Prairie.
This sentiment is further reinforced by the results of the spatial analysis of parents and
offspring.
Interestingly, all six of the male parents resided in the South Pond, whereas two
female parents were more scattered throughout the landscape, and one moved from the
East Channel to the South Pond. It is possible that the high number of adults within the
106
South Pond may limit resources within it, and females seeking food resources for yolk
production during summer months (M. Hayes, unpublished data) may be moving to less
desirable habitats due to overcrowding. Future investigations with the data collected
during this study can be used to determine if there are differences in habitat preferences
between adult males and females at WRP by looking at the abundance and distribution of
males and females across the landscape, and in the small pond.
HABITAT VARIABLES & THEIR EFFECTS
ON OREGON SPOTTED FROG CAPTURE
Frequently, the capture of animals was difficult depending on the orientation of
the animal, and the tools being used. Walking VES and floating VES were used to survey
for adult OSF, and either hand captures or dip nets were used to capture them. Depending
on the orientation of a prospective capture, a decision was made to use either the dip net
or hand capture. If the animal was in the middle of a channel, or otherwise in open water,
the dip net was used. Whereas when an animal was present on or near the edge of the
water on a bank, hand capture was the primary method. Minnow traps yielded very few
adult OSF, as discussed in the Habitat Survey Chapter Methods and Results section.
Watson et al. (2003) found that 88% of the adult OSF they monitored were found at the
water’s surface, indicating that some individuals may reside underwater. Given that some
adults may not have been present at the surface, new methods should be introduced in
future studies in order to account for those undetected animals. New methods could
include the use of scuba gear, or bottom trawling. Voris & Glodek (1980) used bottom
trawling to determine the habitat of File Snakes (A. granulatus) in salt-water habitats.
107
Although, these methods would likely cause a high level of disturbance to OSF and other
species present within the wetland, and consideration of these factors should be taken into
account before they are used. Information obtained by these methods could determine the
habitat utilization of adult OSF under water and the number of individuals present in
those locations. Additionally, radio telemetry would be a useful addition to a study such
as this to determine where OSF are going during the summer season, and to track parents
across the landscape for longer periods of time. Radio telemetry is a commonly used
method for OSF, and would also be useful for tracking the spatial and temporal
movement patterns of parents across the landscape.
OREGON SPOTTED FROG BEHAVIOR & DETECTABILITY
While conducting field surveys for the capture of adult Oregon spotted frogs,
many were seen basking in the sun, floating on the surface, or on the edge of a bank.
TEMPERATURE
Animals were primarily captured during the time interval between 10:00 and
12:00, which likely correlates with air and water temperature during that time of day.
Based on observation, air and water temperature played an important role in the
detectability of adult OSF, and should be investigated further through future analysis of
the data collected during this study.
DIET
It has long been thought that adult OSFs do not eat small fishes (M. Hayes,
personal communication). A single female at the South Pond regurgitated an Olympic
108
Mudminnow while having her mouth swabbed for buccal samples. This observation
brings to light new details about the diets of adult OSFs at West Rocky Prairie.
PRESENCE OF NORTHERN RED-LEGGED FROGS (Rana aurora)
Multiple adult and juvenile Northern red-legged frogs (Rana aurora) were
detected in the East Side Survey Area, and adult Northern red-legged frogs were detected
in the pond next to Tilley Road (Beaver Creek Pond). Additionally, a single adult
Northern red-legged Frog was captured in the East Channel, and juveniles were detected
in the West Channel. However, no red-legged frogs were detected in South Pond, Tilley
Pond, the North Channel, or the clearing North of the North Channel. The areas with the
most red-legged frogs detected were the small pond next to Tilley Road, and the East
Side Survey Area. It is possible that the presence of red-legged frogs may decrease the
presence of adult OSF and may account for the high level of adult OSF in areas where
red-legged frogs are not present, or in low abundance. This may also be due to
differences in non-breeding habitat preference between Northern red-legged frogs and
OSF. These phenomena should be investigated further, and may aid management in
future endeavors for finding adult OSF during the non-breeding season.
JUVENILE OREGON SPOTTED FROGS
Juvenile OSF were typically found in open water habitats and were present in all
aquatic survey areas except for the pond adjacent to Tilley Road (Beaver Creek Pond), as
described in the Habitat Survey Results Chapter. This may indicate that juvenile OSFs
have a more generalized habitat preference compared to adult OSF. Additionally, the
109
presence of juvenile OSF and lack of detection of adult OSF in the East Side Survey Area
and Tilley Pond may indicate that these are habitat types that are not preferred by adult
OSF. Future analysis of the data collected in this study may indicate that the distribution
of juveniles and adults within the WRP site differ. These observations may warrant
further research into the habitat preference differences between adult and juvenile OSF,
as they may need to be managed differently.
DORSAL PATTERN RECOGNITION METHOD
Dorsal pattern recognition is a method that has been used for OSF and Cascade
frogs by the Habitat Program at WDFW for 12 years (M. Hayes, unpublished data).
Dorsal pattern recognition was used during field surveys, for in-office animal comparison
analysis, and compared to the results of matched individuals based on genetics. This
method is certainly useful, as once OSF reach adult stages, their spot patterns remain
consistent across years (M. Hayes, personal communication). However, this method was
very time consuming in the field, as each captured animal had to be compared to all
images of previously sampled adults. The use of high-resolution devices in the field will
likely increase the number of correctly identified individuals and should be used for
future studies. If this method is coupled with genetic sampling in future studies, pictures
and swabs could be taken for all captured individuals in the field, then marked as
recapture or unique individual by analysis of the pictures in the office, rather than in the
field. This will likely decrease the in-field processing time, and increase the amount of
time devoted to capturing adult OSF by all surveyors. While the buccal swab method is
110
considered less invasive than the more commonly used toe-clipping method, it may cause
stress to the animal, and should be used for this listed species with care and caution.
MANAGEMENT IMPLICATIONS & FUTURE RESEARCH
The high abundance of adult OSF in the small South Pond, and the distances
parents are traveling through the landscape to reside there during the summer months,
suggest that non-breeding summer habitat may be limiting for adult OSF at WRP. Also,
given the noticeable presence of Northern red-legged frogs in the East Side survey area,
and the low number of OSF egg masses detected there (n=15), the parents of these egg
masses are likely traveling to the West Side to find more suitable habitat in the summer.
The addition of a duplicate small pond, similar to the size and hydrology of the South
Pond could be warranted in order to increase habitat availability for all adults within the
WRP population. The most logical location, based on the information collected in this
study, would be to build another small pond near the Central East Oviposition location.
This would provide refuge to the parents of East Side egg masses, and provide more nonbreeding habitat for parents of egg masses laid in the Central East and Central West
Oviposition location.
Future studies of OSF adult habitat utilization at WRP should investigate the
southern extent of the West Channel. Two adults were found in the West Channel near
the Central West oviposition habitat location, but they were not assigned to any offspring.
Given the variety of distances adult OSFs travel between breeding and non-breeding
seasons, there may be adults in the areas of the West Channel that were not surveyed
during this study.
111
This spatial analysis indicates a conservative estimate of the distance parent’s
travel between the breeding and non-breeding seasons at WRP, and future analysis of this
data will include estimates of their potential travel paths across the landscape. For future
studies, the addition of radio tracking devices coupled with parentage analysis could
warrant important information about the true paths parents choose to take across the
landscape in a single year.
This study provides an important snapshot of the population and spatial
distribution of the OSF population at WRP. However, N e estimates from single season
data do not take into account fluctuations in populations over time (Schmeller & Merila,
2007). Additionally, ranid frogs tend to experience boom and bust population changes
over time (Phillipsen et al., 2009). Future studies of N e in the WRP and other populations
should include multiple years of genetic information to gain a robust understanding of the
changes in effective size these populations experience across years. Also, the low allelic
richness and relatability seen in this study and others (Phillipsen et al., 2009; Blouin et
al., 2010) suggests that more robust markers are needed for this species. Genetic analysis
of threatened and endangered species can only be as good as the resources the researchers
have available. Future efforts for the development of more robust genetic markers should
be considered by any management efforts aiming to use molecular techniques in their
research. These efforts could consist of the identification and development of
polymorphic markers such as SNPs or microsatellites.
Finally, future investigation into the metabolic cost of traveling long distances for
the OSF should be conducted to determine and quantify energy expenditure, and to
112
ascertain the cost and benefit of traveling. Also, future research should address their
dietary preferences to see whether food availability is driving movement.
SPECIFIC APPLICATIONS TO THE OREGON SPOTTED FROG RANGE-WIDE
Buccal swabs, and the corresponding methods used to obtain these samples, are a
minimally invasive and effective method for obtaining tissue samples from this federally
listed species. Of the 81 duplicate samples obtained from adults in this study, DNA was
successfully extracted from all of them. This high success rate suggests that buccal swabs
are an excellent alternative to the invasive toe-clipping method, which has traditionally
been used for this species (M. Hayes, personal communication).
Dorsal pattern recognition with high-resolution devices can be a useful method
for determining whether individual OSFs are recaptures or new captures, and could
replace other, more invasive methods for determining recaptured individuals.
The use of parentage analysis coupled with spatial analysis can be useful for
tracking where individuals have moved throughout the landscape, and temporally.
Additionally, genotyping can help researchers understand the effective population (N e ) at
a site and the relationships between individuals in the sampled population. Management
of OSF is often focused on relatively isolated and site-specific populations. Genetic
information, such as what was obtained in this study, can be invaluable to management of
those small, relatively isolated populations. Information about the movement of parents
and adults can help management allocate effort toward managing the adult demographic
within these small populations.
113
The successful breeding of larvae and juvenile OSF can only be fully successful if
these individuals, as adults, have available habitat. Incorporating parent/offspring
analysis for managed small populations can glean information about the actual size, and
spatial distribution of these populations. For example, based on the results of this study,
and M. Hayes (personal communication), the East, West, and Central oviposition sites are
less likely to be viewed as distinct populations.
The emphasis on oviposition and tracking of egg mass locations does not
adequately describe this species, and specifically does not address where adult OSF are
going and what habitat they are utilizing. Monitoring oviposition sites only describes a
small, yet significant, component of the OSF lifecycle and habitat. In order to manage
this species most adequately, knowledge of where these adults are going is essential.
Additionally, since some adults at WRP are traveling >1 km to find summer habitat, this
may be an important factor in the decline of this species. Therefore, the incorporation of
summer surveys of adult non-breeding habitat within all known OSF sites, and the
collection of buccal swabs from those adults and egg masses during oviposition, is an
essential addition to inform the management of this species.
In order to best manage, and recover the Oregon spotted frog, it is essential that
we understand, monitor, and manage all habitats and aspects of its life cycle.
114
LITERATURE CITED
Allendorf, F. W., Luikart, G., & Aitken, S. N. (2013). Conservation and the Genetics of
Populations (Second Edition). Wiley-Blackwell.
Beaupre, S. J., Jacobson, E. R., Lillywhite, H. B., & Zamudio, K. (2004). Guidelines for
use of live amphibians and reptiles in field and laboratory research. (Second
Edition). Revised by the Herpetological Animal Care and Use Committee
(HACC) of the Animal Society of Ichthyologists and Herpetologists.
Blouin, M. S., Phillipsen, I. C., & Monsen, K. J. (2010). Population structure and
conservation genetics of the Oregon spotted frog, Rana pretiosa. Conservation
Genetics, 11(6), 2179–2194.
Butler, J. M. (2005). Forensic DNA Typing: Biology, Technology, and Genetics of STR
Markers (Second Edition). Burlington, MA: Elsevier.
Chapuis, M.P., & Estoup, A. (2006). Microsatellite Null Alleles and Estimation of
Population Differentiation. Molecular Biology and Evolution, 24(3), 621–631.
Chelgren, N. D., Pearl, C. A., Adams, M. J., & Bowerman, J. (2008). Demography and
movement in a relocated population of Oregon spotted frogs (Rana pretiosa):
Influence of season and gender. U.S. Department of the Interior; U.S. Geological
Survey, 2008(4).
Cushman, K. A., & Pearl, C. A. (2007). A conservation assessment for the Oregon
Spotted Frog (Rana pretiosa). USDA Forest Service and USDI Bureau of Land
Management, Oregon.
Department of the Interior: US Fish and Wildlife Service. August 29th, 2014. Federal
Register: 50 CFR Part 17, Endangered and Threatened Wildlife and Plants;
Threatened Status for Oregon spotted frog; Final Rule. 79:168.
Elston, R. C. (2005). Polymorphism Information Content. Encyclopedia of Biostatistics.
6.
Frankham, R., Ballou, J. D., & Briscoe, D. A. (2003). Introduction to conservation
genetics. Cambridge University Press.
Gallardo, C. E., Correa, C., Morales, P., Sáez, P. A., Pastenes, L., & Méndez, M. A.
(2012). Validation of a cheap and simple nondestructive method for obtaining
AFLPs and DNA sequences (mitochondrial and nuclear) in amphibians.
Molecular Ecology Resources, 12(6), 1090–1096.
Green, D. M., Kaiser, H., Sharbel, T. F., & McAllister, K. R. (1997). Cryptic species of
spotted frogs, Rana pretiosa complex, in western North America. Copeia, 1, 1–8.
115
Gosner, K. (1960). A simplified table for staging anuran embryos with notes on
identification. Herpetologica, 16(3), 183-190.
Hallock, L.A., Pearson, S. (2001). Telemetry Study of Fall and Winter Oregon spotted
frog (Rana pretiosa) Movement and Habitat Use at Trout Lake, Klickitat County,
Washington. Report to Washington State Dept. of Transportation and WA DNR
Natural Areas Program.
Hallock, L.A., (2013). DRAFT State of Washington Oregon Spotted Frog Recovery Plan.
Department of Fish and Wildlife, Olympia. WA. 1-106.
Haycock, R.D. (2000). COSEWIC assessment and status report on the Oregon spotted
frog (Rana pretiosa) in Canada. Committee on the Status of Endangered Wildlife
in Canada, Ontario, Ottawa, Canada.
Hayes, M.P. (1997). Status of the Oregon spotted frog (Rana pretiosa) in the Deschutes
Basin and selected other systems in Oregon and northeastern California with a
range wide synopsis of the species’ status. Final report prepared for The Nature
Conservancy under contract to US Fish and Wildlife Service, Portland, Oregon.
Hayes, M. P., Engler, J. D., Van Leuven, S., Friesz, D. C., Quinn, T., & Pierce, D. J.
(2001). Overwintering of the Oregon Spotted Frog (Rana pretiosa) at Conboy
Lake National Wildlife Refuge, Klickitat County, Washington, 2000-2001.
Interim Final Report to Washington Department of Transportation by Science
Team, Habitat and Wildlife Management Programs, Washington Department of
Fish and Wildlife, Olympia, Washington.
Hoffman, E. A., Schueler, F. W., & Blouin, M. S. (2004). Effective population sizes and
temporal stability of genetic structure in Rana pipiens, the northern leopard frog.
Evolution, 58(11), 2536–2545.
Jehle, R., & Arntzen, J. W. (2002). Review: microsatellite markers in amphibian
conservation genetics. Herpetological Journal, 12, 1–9.
Jones, A. G., & Ardren, W. R. (2003). Methods of parentage analysis in natural
populations. Molecular Ecology, 12(10), 2511–2523.
Jones, A. G., Small, C. M., Paczolt, K. A., & Ratterman, N. L. (2010). A practical guide
to methods of parentage analysis: TECHNICAL REVIEW. Molecular Ecology
Resources, 10(1), 6–30.
Jones, O. & Wang, J. (2009). COLONY: a program for parentage and Sibship inference
from multilocus genotype data. Molecular Ecology Resources, 10, 551-555.
Kalinowski, ST, Taper, ML & Marshall, TC. (2007). Revising how the computer
program CERVUS accommodates genotyping error increases success in paternity
assignment. Molecular Ecology, 16, 1099-1106.
116
Kapust, H. Q., McAllister, K. R., Hayes, M. P., & others. (2012). Oregon spotted frog
(Rana pretiosa) response to enhancement of oviposition habitat degraded by
invasive Reed Canary Grass (Phalaris arundinacea). Herpetological
Conservation and Biology, 7(3), 358–366.
Kliman, R., Sheehy, B. & Schultz, J. (2008). Genetic Drift and Effective Population
Size. Nature Education, 1(3):3.
Licht, L.E. (1971). Breeding Habits and Embryonic Thermal Requirements of the Frogs,
Rana Aurora and Rana Pretiosa, in the Pacific Northwest. Ecology, 52(1), 116124.
Licht, L. E. (1974). Survival of embryos, tadpoles, and adults of the frogs Rana aurora
aurora and Rana pretiosa pretiosa sympatric in southwestern British Columbia.
Canadian Journal of Zoology, 52(5), 613–627.
Marco, A., Quilchano, C., & Blaustein, A. R. (1999). Sensitivity to nitrate and nitrite in
pond-breeding amphibians from the Pacific Northwest, USA. Environmental
Toxicology and Chemistry, 18(12), 2836–2839.
Marshall, T. C., Slate, J., Kruuk, L. E. B., & Pemberton, J. M. (1998). Statistical
confidence for likelihood-based paternity inference in natural populations.
Molecular Ecology, 7(5), 639–655.
McAllister, K.R., Leonard, W.P. (1997). Washington State Status Report for the Oregon
Spotted Frog. Washington Department of Fish and Wildlife: Wildlife Management
Program, 1-47.
McAllister, K., Walker, M. (2003). An inventory of Oregon spotted frogs (Rana pretiosa)
in the upper Black River drainage, Thurston County, Washington. Unpublished
Report; Wildlife Program: Washington Department of Fish and Wildlife.
McKenzie, D.L., Nichols, J.D., Royle, J.A., Pollock, K.H., Bailey, L.L., Hines, J.E.
(2005). Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of
Species Occurrence. Elsevier: Academic Press.
Monastersky, R. (2014). Life – A Status Report. Nature, 516, 159-161.
Olson, D.H., Leonard, W.P., Bury, R.B. (1997). Sampling Amphibians in Lenthic
Habitats: Methods and Approaches for the Pacific Northwest. Society for Northwest
Vertebrate Biology. Olympia, WA.
Pearl, C.A., Adams, M.J., Bury, R.B., McCreary, B. (2004). Asymmetrical Effects of
Introduced Bullfrogs (Rana catesbeiana) on Native Ranid Frogs in Oregon.
Copeia, (1), 11-20.
117
Pearl, C. A., Adams, M. J., & Leuthold, N. (2009). Breeding habitat and local population
size of the Oregon spotted frog (Rana pretiosa) in Oregon, USA. Northwestern
Naturalist, 90(2), 136–147.
Phillipsen, I. C., Bowerman, J., & Blouin, M. (2010). Effective number of breeding
adults in Oregon spotted frogs (Rana pretiosa): genetic estimates at two life
stages. Conservation Genetics, 11(3), 737–745.
Pidancier, N., Miquel, C., Miaud, C. (2003). Buccal Swabs as a non-destructive tissue
sampling method for DNA analysis in amphibians. Herpetological Journal, 13,
175-178.
Poschadel, J.R., Moller, D. (2004). A versatile field method for sampling on small
reptiles and amphibians, applied to pond turtles, newts, frogs, and toads.
Conservation Genetics, 5, 865-867.
Qiagen. (2015). DNeasy Blood and Tissue Kits.
https://www.qiagen.com/us/products/catalog/sample-technologies/dna-sampletechnologies/genomic-dna/dneasy-blood-and-tissue-kit#productdetails
Riester, M., Stadler, P. F., & Klemm, K. (2009). FRANz: reconstruction of wild
multi-generation pedigrees. Bioinformatics, 25(16), 2134–2139.
Risenhoover, K.L., McBride, T.C., McAllister, K., Golliet, M. (2001). Overwintering
behavior of the Oregon spotted frog (Rana pretiosa) along Dempsey Creek,
Thurston County, Washington. Final Report to Washington Department of
Transportation.
Schmeller, D. S., & Merilä, J. (2007). Demographic and Genetic Estimates of Effective
Population and Breeding Size in the Amphibian Rana temporaria. Conservation
Biology, 21(1), 142–151.
Storfer, A., Eastman, J. M., & Spear, S. F. (2009). Modern molecular methods for
amphibian conservation. BioScience, 59(7), 559–571.
Tyson, J. A., Hayes, M.P. (2014). Biennial Report: 2013-2014 Oregon Spotted Frog
(Rana pretiosa) Egg Mass Surveys at West Rocky Prairie Wildlife Area and A
Review of Historical Egg Mass Surveys at This Site. Washington Department of
Fish and Wildlife.
Voris, H.K., Glodek, G.S. (1980). Habitat, Diet, and Reproduction of the File Snake,
Acrochordus granulatus, in the Straits of Malacca. Journal of Herpetology, 14(1),
108-111.
Wang, J. (2009). A new method for estimating effective population sizes from a single
sample of multilocus genotypes. Molecular Ecology, 18(10), 2148–2164.
118
Wang, J. (2012). Computationally efficient Sibship and parentage assignment from
multilocus marker data. Genetics, 191, 183-194.
Watson, J. W., McAllister, K. R., Pierce, D. J. (2003). Home Ranges, Movements, and
Habitat Selection of Journal of Herpetology, 37(2), 292–300.
Yahnke, A. E., Grue, C. E., Hayes, M. P., & Troiano, A. T. (2013). Effects of the
herbicide imazapyr on juvenile Oregon spotted frogs. Environmental Toxicology
and Chemistry, 32(1), 228–235.
Personal Communication
Bowman, Cherril. 2014-2015. Personal Communication. Technician: Molecular Genetics
Laboratory, Washington Department of Fish & Wildlife. Olympia, WA.
Hayes, Marc P. 2014-2015. Personal Communication. Senior Research Scientist, Aquatic
Research Section, Habitat Program, Washington Department of Fish & Wildlife.
Olympia, WA.
Hayes, Marc P. Unpublished Data.
Small, Maureen. 2015. Personal Communication. Fisheries Biologist/Geneticist,
Molecular Genetics Laboratory, Washington Department of Fish & Wildlife.
Olympia, WA.
Warheit, Kenneth. 2014-2015. Personal Communication. Director, Molecular Genetics
Laboratory, Washington Department of Fish & Wildlife. Olympia, WA.
Production of Maps
All maps created by Chelsea D. Waddell in this document were created using ArcGIS®
software, ArcMAP™ 10.2 with World Imagery Base Maps by Esri®.
119
APPENDICES
APPENDIX A: Standard English and Scientific Names for flora and fauna found in the
WRP Survey Area.
Fauna
Common Name
Oregon spotted frog
Olympic Mudminnow
Three-Spined Stickleback
Northwestern Salamander
Common Garter Snake
Red-Legged Frog
Leech
Scientific Name
Rana pretiosa
Novumbra hubbsi
Gasterosteus aculeatus
Ambystoma gracile
Thamnophis sirtalis
Rana aurora
Hirudinea spp.
Flora
Common Name
Sedge
Reed Canary Grass
Willow
Hard Hack
Scientific Name
Carex spp.
Phalaris arundinacea
Salix spp.
Spiraea douglasii
Grouped Classification
Sedge
Reed Canary Grass
Scrub/Shrub/Willow
Scrub/Shrub/Willow
120
APPENDIX B: Field Survey Protocol for Buccal Swab Sampling. Protocol developed by
Chelsea D. Waddell (2014) for the purposes of this study.
121
APPENDIX C: A discussion about other DNA types (mitochondrial DNA and Single
Nucleotide Polymorphisms) that can be used to assess genetic diversity in amphibian
populations.
Mitochondrial DNA (mtDNA)
Mitochondrial DNA (or mtDNA) barcoding is an even more standardized
approach than microsatellites for determining genetic differentiation between individuals
(Storfer et al., 2009). This type of DNA barcoding uses universal primers to generate
short DNA sequences, which vary among species. These sequences can be applied to
large taxonomic groups (Storfer et al., 2009). However, mtDNA is only inherited
maternally, and therefore does not give adequate representation of genetic diversity
(Storfer et al., 2009). Since male chromosomal influence is not represented in
mitochondrial DNA, this method is more appropriate for rapidly identifying specimens as
a particular species and should not be used to establish genetic diversity (Storfer et al.,
2009), or parentage. Another added challenge to using mtDNA is that these fragments
typically represent one loci; to best represent genetic diversity within an individual and
between members of the same species, multiple loci should be used (Storfer et al., 2009).
The use of multiple loci is important because genetic variability within a species cannot
only be represented by one loci, and not all members of a population across their range
will have all the same loci (Storfer et al., 2009). Therefore, the law of large numbers is
warranted, and multiple loci should be used to adequately represent a population’s
genetic variability. Many researchers still use mtDNA to look at amphibian population
genetics, but this is typically supplemental to microsatellites, as was done by Blouin et al.
(2010). Researchers can also use Single Nucleotide Polymorphisms (SNPs) to understand
genetic variability.
122
Single Nucleotide Polymorphisms (SNPs)
Single Nucleotide Polymorphisms (SNPs) are single base-pair (point mutation)
variations at particular sites along an organism’s genome (Storfer et al., 2009). “SNPs are
bi-parentally inherited and expressed, commonly have two alleles, can occur in protein
coding and non-coding regions of the genome” and are easily sequenced using PCR
methods (Storfer et al., 2009). This method is more powerful than microsatellites and is
excellent for assessing genetic distinctiveness among populations (Storfer et al., 2009).
Typically, SNP data is obtained from Next Generation Sequencing methods because of
its high yield of sequencing outputs compared to traditional sequencing methods. Next
Generation Sequencing is a relatively new method, which has been gaining popularity in
the past few years. However, the process can be very expensive, is not yet widely
available (Storfer et al., 2009), and SNPs for the Oregon spotted frog have not yet fully
been developed (K. Warheit, personal communication). Given that Next Generation
sequencing technology is not yet widely available, SNPs are less widely used (Storfer et
al., 2009), these two methods were not an option for this study.
123
APPENDIX D: This appendix outlines the types of parentage analyses that can be
performed that were not done for this study. Additionally, it details other methods that
were used to inform the results of the parentage analysis, but were not reported in the
text.
UNREPORTED METHODS USED
FRANz & Parental Reconstruction
The parental reconstruction approach “uses the genotypes of offspring in full- or
half-sib (sibling) families to reconstruct parental genotypes” (Jones et al., 2010). In the
case of amphibians, who typically lay their eggs in masses, each sibling within the mass
shares one parent, the mother. The unknown parental genotype can be estimated by
subtracting the known parent’s alleles, thus reconstructing the shared parent’s genotype
(Jones et al., 2010). When the parental reconstructions are developed, they can be
compared to the actual parents to determine which individuals have the highest likelihood
of being a parent (Jones et al., 2010). There are multiple statistical approaches to this
technique, which include complex algorithms, likelihood, and Bayesian posterior
probabilities (Jones et al., 2010). Disadvantages to this approach include the need for
hyper variable (highly polymorphic) loci, and 8-10 full and half-sibs to confidently
reconstruct parents (Jones et al., 2010).
For parental reconstruction analysis in this project, the program FRANz was used
(Riester et al., 2009). FRANz develops likelihood scores for each parent assignment,
much like CERVUS 3.0.7. However, unlike CERVUS 3.0.7, it takes into account the
sibling relationships between the offspring and assigns parents based on posterior
probabilities. FRANz develops a simulation of parent:offspring and 50,000 iterations
were run in the simulation for this study. It also generates a file of allele frequencies
within the sampled population, which includes observed and expected heterozygosity.
124
The analysis method I used to assess parentage with the FRANz output is very
similar to the method I used from CERVUS, where both offspring from a single egg mass
had to be assigned to the same parent based on posterior probability. If a parent was
assigned to both offspring from multiple egg masses, they were discarded from the
assignment. These assignments were done using a pivot table in Microsoft Excel, where
the parent with the highest likelihood’s posterior probability was displayed and compared
to the offspring.
Sibship (Sibling Relationship) Reconstruction & COLONY
Sibship reconstruction is an excellent method for reconstructing parental
genotypes, as it does not have the same numerical restrictions (8-10 offspring) as parental
reconstruction.
The program COLONY (Jones & Wang, 2009) was used in this project to
determine parentage assignments in the form of posterior probabilities. It was run two
times for parentage, in a short run and a medium run, the resulting assignments were
compared to each other and a final assessment of the assigned parents was determined.
To assess parentage using the COLONY outputs, I first compared the results of the short
run and medium run to see if they were consistent. If both offspring from a single egg
mass had to be assigned to the same parent based on the posterior probabilities in both the
short and medium run, they were included. If a parent was assigned to both offspring
from multiple egg masses, they were discarded from the assignment.
125
OTHER PARENTAGE ASSIGNMENT METHODS:
Fractional Allocation
Another parentage assignment method not used for this study is fractional
allocation, which is very similar to categorical allocation except that instead of assigning
the entire offspring to the most likely parent, it “assigns a given offspring partially to
each nonexcluded candidate parent” (Jones et al., 2010). Each offspring and parent is
assigned a percent likelihood of being matched. Fractional allocation assigns these
unexcluded parents based on the same likelihood or posterior probability calculations as
in categorical allocation (Jones et al., 2010). While there are some statistical advantages
to using this method, it is rarely used in empirical studies because it does not fully assign
parents to offspring (Jones et al., 2010). While categorical allocation will generate full
assignment of offspring to the sampled parents, fractional allocation can be extremely
useful.
Full Probability Parentage Analysis
Another method for parentage assignments is full probability parentage analysis,
but it was not used in this project. Full probability parentage analysis estimates
“population level variables of interest at the same time as the patterns of parentage” using
a single modeling approach in a Bayesian framework (Jones et al., 2010). It allows for the
incorporation of habitat and spatial information, such as physical barriers, which may be
important to the relationships of the population of interest (Jones et al., 2010). In
categorical and fractional allocation, uncertainty in parentage assignments is not
incorporated when the researcher is looking at other variables of interest (i.e. Habitat
126
restrictions), causing an inflated level of confidence (Jones et al., 2010). In full
probability analyses, the uncertainty in parentage assignments is included in “the
estimation of the population-level variables of interest”, which allows for increased
confidence (Jones et al., 2010). This incorporation of uncertainty helps researchers
answer questions about population level processes. Furthermore, in categorical and
fractional allocation, all parents in the analysis are considered equally likely to be true
parents of an offspring (Jones et al., 2010). With full probability, each parent is not
considered equally likely because it takes “into account relevant ecological information,
such as territoriality, spatial location, breeding status, etc.” (Jones et al., 2010). A major
disadvantage to this approach is that not all of this ecological and mating behavior
information can be known (Jones et al., 2010). This can result in an inaccurate model and
confidence in parentage assignment could be weakened. Full probability models should
be run in conjunction with tests for the actual probabilities and likelihoods for parentage
to be sure that the model is accurately assigning parents (Jones et al., 2010).
127