SPATIAL ECOLOGY OF THE CASCADES FROG: IDENTIFYING DISPERSAL,
MIGRATION, AND RESOURCE USES AT MULTIPLE SPATIAL SCALES.
By
Justin Matthew Garwood
A Thesis
Presented to
The Faculty of Humboldt State University
In Partial Fulfillment
Of the Requirements for the Degree
Master of Science
In Natural Resources: Wildlife
December 2009
SPATIAL ECOLOGY OF THE CASCADES FROG: IDENTIFYING DISPERSAL,
MIGRATION, AND RESOURCE USES AT MULTIPLE SPATIAL SCALES.
by
Justin M. Garwood
Approved by the Master's Thesis Committee:
______Dr. Hartwell H. Welsh Jr., Major Professor Date
______Dr. Luke T. George, Committee Member Date
______Dr. Matthew D. Johnson, Committee Member Date
______Dr. Terry D. Roelofs, Committee Member Date
______Dr. Gary L. Hendrickson, Graduate Coordinator Date
______Dean for Research and Graduate Studies Date
ABSTRACT
Spatial Ecology of the Cascades Frog: Identifying Dispersal, Migration, and Resource Uses at Multiple Spatial Scales.
Justin M. Garwood
This study combined both intensive and long-term sampling to provide
information on Cascades frog, Rana cascadae, landscape use at multiple spatial scales in the Trinity Alps Wilderness of Northern California. Mark-recapture and radio telemetry
were used from 2003 to 2008 to determine key ecological components of R. cascadae life
history, resource use, spatial patterns, migrations and dispersal for all post-metamorphic
age groups. Rana cascadae often used different aquatic resources for breeding, summer
foraging and overwintering. Resources were commonly found to be spatially or
temporally separated and frogs were observed to move seasonally among them. Lentic
habitats fed by groundwater springs were used extensively for both breeding and overwintering. Seasonal migration events were common among isolated habitats. Single sites were not likely to contain self-sustaining sub-populations, but contribute to a matrix of required seasonal resources across a patchy landscape. The majority of dispersal events occurred between patches located inside a single basin, however, one percent of frogs dispersed among four of six neighboring basins. Consequently the conservation of
R. cascadae populations requires management of key resources, processes, and an integration of several spatial scales that reflect the whole range of life history attributes.
iii
ACKNOWLEDGEMENTS
This study could not have been possible without the vision, dedication and support of my advisor, Dr Hartwell H. Welsh, Jr. I am grateful to have studied under such an influential figure on the ecology of herpetofauna in the Pacific Northwest. I thank you
Hartwell for your open door and always engaging conversations. My committee: Dr Luke
George, Dr Matthew Johnson, and Dr Terry Roelofs provided support and expert knowledge in the design, analysis, and completion of this thesis. I thank Becky Howard for her help with designing and managing this studies data base. Jim Hotchkiss and Eric
Haney provided essential GPS equipment and thorough training for accurate mapping and spatial data collection. Betsy Bolster deserves special appreciation for seeking matching funds over the course this study which allowed me to play in the Klamath
Wilderness longer and answer more complex ecological questions.
Words fail to describe my appreciation for my colleagues and friends who were by my side and supported me in the field throughout this study: Clara Wheeler, Ryan
Bourque, and Monty Larson. These three joined along to capture herpetofauna with pure enthusiasm, climbing that brutal trail countless times, wading through ice covered lakes, surveying streams in the black of night, and snowshoeing across steep ridges. I thank you all for stepping off the trail with me, deep in the Klamath Mountains. We wore the soles off our boots in the pursuit of discovering new ecological phenomena, and have many more adventures to come. I also offer special thanks to those individuals who volunteered their time on field adventures: Terra Fuller, Nate Nieto, James Bettaso, Laura
iv
Burkholder, Don Ashton, Rebecca Studebaker, Jim Hotchkiss, Mel Dean, Joshua Dorris,
William Wheeler, Abby Wheeler, Kara Cox, Jacob Ehlerding, Molly Alles, Jon Stead,
Karen Pope, Cheryl Bondi, Betsy Bolster, Cara McGary, Erin Hannelly, Brian Jennings,
Cain Adams, Reuben Koontz, Michael Van Hattem, Colin Anderson, Marisa Parish, and
Nick Van Vleet.
I have been humbled by the intellectual support from many of my friends and colleagues who kept some of my wild ideas between the lines and helped generate
countless others: Jamie Bettaso, Clara Wheeler, Don Ashton, Ryan Bourque, Garth
Hodgeson, Karen Pope, Monty Larson, Michael Van Hattem, and Seth Ricker.
I thank my family for supporting me in my endeavors throughout my life. I
especially thank my parents, Paul and Lorna Garwood, for providing a warm shower, a
cold beer, and good company for the whole crew after many long trips in the Trinities. I
owe special recognition and praise to Rebecca Studebaker for her unwavering
compassion and support throughout the writing of this thesis. I thank you Rebecca for
your help in the field, analysis discussion, thesis review, and keeping me positive
throughout.
Last, I would like to thank the organizations that funded this research: the
California Department of Fish and Game, the National Fish and Wildlife Foundation, the
US forest Service, PSW, Redwood Sciences Laboratory, the World Conservation Union,
Amphibian Specialist Group, and the USGS Amphibian Research and Monitoring
Initiative.
v
TABLE OF CONTENTS
Page
ABSTRACT ...... iii
ACKNOWLEDGEMENTS ...... iv
TABLE OF CONTENTS ...... vi
LIST OF TABLES ...... viii
LIST OF FIGURES ...... ix
LIST OF FIGURES ...... ix
LIST OF APPENDICIES ...... xiii
INTRODUCTION ...... 1
MATERIALS AND METHODS ...... 7
Study Area Description ...... 7
Surveys and Marking ...... 12
Radio Telemetry ...... 15
Resource Data Collection ...... 16
Data Analysis ...... 18
Seasonal Use of Space and Resources ...... 21
Movements of Immature Frogs ...... 25
Breeding Dispersal ...... 25
Seasonal Migrations ...... 26
Interbasin Dispersal ...... 28
vi
RESULTS ...... 29
Surveys and Captures ...... 29
Summer Use of Space ...... 29
Seasonal Space Use by Adults ...... 31
Winter Space Use by Juveniles and Adults ...... 35
Use of Aquatic Resources ...... 35
Seasonal Resource Use ...... 38
Movements of Immature Frogs ...... 44
Dispersal ...... 45
Adult Seasonal Migrations ...... 48
Movement Rates and Routes ...... 64
Interbasin Dispersal ...... 66
DISCUSSION ...... 70
Spatial Dynamics ...... 70
Resource Use ...... 72
Movements by Immature Frogs ...... 76
Adult Seasonal Migrations ...... 77
Dispersal ...... 78
Interbasin Dispersal ...... 80
References ...... 85
Appendices ...... 94
vii
LIST OF TABLES
Table Page
1. Summary of spatial, temporal and demographic information used for analyses of life history attributes of Rana cascadae in the Southeast Trinity Alps Wilderness, California, 2003 to 2008. See text for variable and analysis descriptions...... 19
2. Volume of intersection index for utilization distributions of separate age groups of Rana cascadae during the summer season from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Young of previous year individuals were captured and marked from 2003 to 2005 only...... 30
3. Volume of intersection index between seasonal population utilization distributions of adult Rana cascadae from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. In 2006 the winter season was omitted due to a small (N = 37) sample size...... 33
4. Volume of intersection index between seasonal population utilization distributions of juvenile and adult Rana cascadae from 2003 to 2005 in Echo Lake Basin, Trinity Alps Wilderness, California. In 2006 the winter season was omitted due to a small (N = 6) juvenile sample size...... 36
5. Summary of the total surface area of water by patch and season for Echo Lake Basin, Trinity Alps Wilderness, ...... 39
6. Mean percentage of annual summer Rana cascadae captures in Echo Lake Basin, Trinity Alps Wilderness, California, 2003 to 2006. Data were summarized by sampling zone hydrologic categories identified in Figure 3. Means are % (± 1 Standard Error)...... 41
7. Dispersal probabilities for adult reproducing Rana cascadae with: A) first time breeding adults with juvenile capture histories, and B) experienced breeding adults between two or more breeding seasons. Dispersal probabilities are summarized for the seven patch complexes identified in Figure 3 within Echo Lake Basin, Trinity Alps Wilderness, California...... 46
8. Interbasin dispersal propensity by Rana cascadae among six basins in the southeast Trinity Alps Wilderness, California, from 2003 to 2008...... 67
viii
LIST OF FIGURES
Figure Page
1. Adult female Rana cascadae (Length: 70mm) photographed on 16 August 2007 at Green Springs Meadows, Echo Lake Basin, Trinity Alps Wilderness, California. This frog was a resident at Lower Van Matre Meadows as a juvenile from 2004 to 2006 and then dispersed across the basin 1.17 km (240 m elevation gain) as an adult to breed at Snowmelt Pond in 2007...... 3
2. Study region located in the southeast Trinity Alps Wilderness, Klamath Mountains, California. Black dotted lines represent catchment boundaries for major drainage basins. Basins and aquatic resources included in the study are labeled...... 8
3. Distribution of surveyed aquatic habitats within Echo Lake Basin, Trinity Alps Wilderness, California. Seven patch complexes are outlined with dotted lines. Three large patch complexes (grouped isolated patches that were located in close proximity) include: VMM = Van Matre meadow complex, DPC = Deep Creek meadow complex, CAM = Cascade meadow complex. Within patch complexes, fifteen isolated patches and 73 sampling zones are represented by a 4 letter code. The first three letters represent the patch ID; the last letter represents the sampling zone ID within a specific patch. Stream, pond, and lake sampling zones are separated by black hatch lines...... 11
4. Fixed kernel utilization distributions for different age groups of the Rana cascadae population during summer from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Young of previous year: density function of 416summer Young of previous year captures, Juvenile: density function of 900 juvenile summer locations, Adult: density function of 793 adult summer capture locations. The height of each density function depicts the relative probability of an individual occurring at each location within the study area by age group. Colored halo lines represent the boundary of the area that contains 95% of the volume of a probability density distribution. Bottom panel shows base map of aquatic habitats within Echo Lake Basin in black...... 32
ix
5. Fixed kernel utilization distributions for adult Rana cascadae population during the spring (breeding), summer and winter from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Breeding: density function of 195 R. cascadae egg masses, Summer: density function of 793 adult summer locations, Overwintering: density function of 297 adult capture locations just prior to the onset of winter. The height of each density function depicts the relative probability of an individual occurring at each location within the study area by season. Colored halo lines represent the boundary of the area that contains 95% of the volume of a probability density distribution. Bottom panel shows base map of aquatic habitats within Echo Lake Basin in black...... 34
6. Mean percentage of annual egg production and captures of Rana cascadae (separated by sex and age group) found in each aquatic resource category summarized across all seasons within Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2006. The “Large Deep Pond” category represents Snowmelt and Eden ponds. Use proportions are based on 275 egg masses and 4,838 frog captures. The young of previous year age group is represented from 2003 to 2005, although present, none were captured in subsequent years...... 37
7. Map of Echo Lake Basin, Trinity Alps Wilderness, California, displaying seasonal hydrological properties of 73 sampling zones used by Rana cascadae from 2003 to 2008. Cross hatched zones and dotted stream segments dried completely by late summer so they were unavailable for winter use. Areas within green colored zones and solid line streams were used for both summer and overwintering habitats. Areas within solid blue zones were used year round. Yellow diamonds indicate where groundwater springs emerge. Orange circles represent traditional breeding areas used by R. cascadae...... 40
8. Map depicting dispersal directions of 84 first time breeding Rana cascadae between eight isolated patch complexes from 2003 to 2008 in the southeast Trinity Alps Wilderness, California. Arrows at the end of lines indicate dispersal direction. Dotted lines surrounding two patch complexes delineate the border of the two patch complexes that produced dispersing individuals: Van Matre Meadow complex and Cascade Meadow complex. Asterisk symbols indicate R. cascadae breeding locations used consistently throughout the study area...... 47
9. Cumulative dispersal distance curves for all first time breeding female (N = 27) and male (N = 57) Rana cascadae that were originally marked as juveniles and subsequently captured as adults at active breeding sites in Echo Lake Basin, Trinity Alps Wilderness, California...... 49
x
10. Maps of Echo Lake Basin, Trinity Alps Wilderness, California, showing capture points (open circles) and connecting lines of 279 individual adult male (A) and 143 individual adult female (B) Rana cascadae captured two or more times from 2003 to 2006. Orange circles represent breeding sites. An individual frog can be ...... 51
11. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between overwintering and breeding seasons from 2003 to 2007 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years...... 52
12. Examples of Long-term mark-recapture histories of seasonal migrations for seven adult female Rana cascadae in Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2008. All subsequent capture locations of individuals at specific habitat patches were within 25 m of original capture locations. Asterisks represent breeding locations used all years...... 53
13. Examples of Long-term mark-recapture histories of seasonal migrations for eight adult male Rana cascadae in Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2008. All subsequent capture locations of individuals at specific habitat patches were within 25 m of original capture locations. Asterisks represent breeding locations used all years...... 55
14. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between breeding and summer seasons from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years...... 56
15. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results. 57
16. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summers of 2003 and 2004 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results...... 58
xi
17. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results. 59
18. Examples of migration routes and timing between seasonal resources for three telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results. 60
19. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between summer and winter seasons from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years...... 62
20. Connecting lines of nineteen individual Rana cascadae that dispersed over ridges separating four neighboring watersheds in the southeast Trinity Alps Wilderness, California. Numbers associated with white lines represent the number of individuals that moved in a particular pattern; arrows indicate movement direction. The overall air distance linking habitat patches across the landscape through dispersal was 5.2 km, indicating high potential gene flow across this landscape. Top inset depicts a cross section of the lowest elevations along ridges at passes (yellow arrows) that dispersing frogs likely used as corridors. Two basins where interbasin movements were not observed are not indicated in figure...... 68
21. Examples of two population structure models for amphibians in mountainous regions. (A) ‘Valley-Mountain’ model for Columbia spotted frogs (Rana luteiventris) adapted from Funk et al. (2005b) with three gene flow scenarios: (i) Low-elevation populations with high gene flow, (ii) high-elevation populations with little to no gene flow over ridges, (iii) restricted gene flow between low and high- elevation populations. (B) Proposed ‘Mountain-Island’ model for Cascades frogs (R. cascadae) in the Klamath mountains, California. Two gene flow scenarios: (i) moderate gene flow across ridges between small high-elevation populations, (ii) little to no gene flow between populations separated by major drainages dipping below 1200 meters elevation...... 84
xii
LIST OF APPENDICIES
Appendix Page
A. Empirical model displaying timing and duration of specific life history attributes for all age groups of Rana cascadae from 2003 to 2008 in Echo Lake Basin, Trinity Alps Wilderness, California. Some life history attributes overlap due to geographic variation in snowmelt and or site availability/ elevations within the basin. Timing and duration of each life history attribute included data from 2006 (extremely wet year with 251% snowpack), and data from 2007 (dry year with 38% snowpack) so extreme annual variability is represented. Vertical arrows in first year stages represent major development transitions...... 94
B. Summary table of attributes, movement, and fate of 51 Rana cascadae monitored by radio telemetry for years 2003 and 2004 in Echo Lake Basin, Trinity Alps Wilderness, California...... 95
C. Water temperature profiles of two shallow spring fed ponds in Echo Lake Basin, Trinity Alps Wilderness, California, during the winter of 2007/ 2008 used annually from 2003 to 2008 by Rana cascadae for overwintering. Grey line represents the water temperature profile of a data recorder 45 cm below the pond surface suspended under 15 cm of loose silt. Black line represents the temperature profile of a data recorder in a second pond located 93 cm below the surface in open water 10 cm above the bottom. Temperatures were recorded every two hours continuously from 07 October 2007 to 04 July 2008...... 97
xiii
INTRODUCTION
Declines in amphibian populations have been noted worldwide (Stuart et al. 2004)
and the ranid frog complex of western North America are among the most severely
impacted of all ranid species (Lannoo 2005). In light of investigations documenting declines of the Cascades frog, Rana cascadae, in California (Fellers and Drost 1993,
Welsh et al. 2006, Fellers et al. 2008) renewed attention has been placed on the conservation of this species. Recent studies have identified several threats including: habitat loss and alteration (Fellers et al. 2008), pesticides (Davidson et al. 2002), exotic predators (Welsh et al. 2006, Pope et al. 2008, Pope 2008) and disease (Garcia et al.
2006). These threats are also identified among the leading causes of amphibian declines worldwide (Collins and Storfer 2003). In response, R. cascadae has been listed as a
California species of special concern since 1994 (Jennings and Hayes 1994) and has been listed as “sensitive” by the USDA Forest Service since 1998 (United States Department of Agriculture Forest Service 1998). R. cascadae has also been listed as “near threatened” on the global International Union for Conservation of Nature Red List since 2004
(Hammerson and Pearl 2004).
Rana cascadae distributions are largely associated with montane and sub-alpine
landscapes (Pearl and Adams 2005). Their range extends from the northern Sierra
Nevada Mountains, north throughout the Cascades Ranges of California, Oregon and
Washington. Two isolated populations occur west of the Sierra Nevada/Cascade range in
the Olympic Peninsula in Washington and the Klamath mountains in Northern California.
1 2 Research on the genetic structure of R. cascadae (excluding the Klamath mountains) found California populations are distinctly different from those in Oregon and
Washington. The two populations have likely been separated since the glaciation during the Pleistocene epoch (Monsen and Blouin 2003). Known extant populations of R. cascadae in California appear to be restricted to elevations above 1220 m (Welsh et al.
2006, Fellers et al. 2008, J. Garwood pers. obs.) creating a highly fragmented “island” distribution.
My study took place in the Klamath range where the majority of the frog’s are found within 315,400 hectares of protected lands, which include the Trinity Alps,
Russian, and Marble Mountain Wilderness areas (Welsh et al. 2006). Disjunct Klamath populations also occur in the Mount Eddy region and both slopes of the Shasta-Trinity
Divide. In the Shasta/Lassen region (Southern Cascades range), small remnant populations of R. cascadae are presently extremely isolated from one another with many historic sites now absent of frogs (Fellers et al. 2008).
Rana cascadae was first considered a unique species among the western ranid frog complex by Slater (1939). This species can be identified from other western ranids by well-defined, inky black dorsal spots and a yellow upper jaw stripe that extends almost to the shoulder (Olson 2005, Figure 1). Rana cascadae is considered to be medium sized among western ranid frogs. In this study, I found females up to 81 mm snout-urostyle length and weighing as much as 56 g when gravid. Males grew up to 67 mm SUL and weighed as much as 28 g.
3
Figure 1. Adult female Rana cascadae (Length: 70mm) photographed on 16 August 2007 at Green Springs Meadows, Echo Lake Basin, Trinity Alps Wilderness, California. This frog was a resident at Lower Van Matre Meadows as a juvenile from 2004 to 2006 and then dispersed across the basin 1.17 km (240 m elevation gain) as an adult to breed at Snowmelt Pond in 2007.
4 Across their range, R. cascadae reproduce once annually, immediately after surface waters begin to thaw (Sype 1975, Briggs 1976, Nussbaum et al. 1983). Egg masses containing 300-500 ova are typically oviposited on shallow benches or alcoves of lentic waterbodies (Briggs 1976, Nussbaum et al. 1983). Rana cascadae have a single year larval development period, with tadpoles metamorphosing into frogs by late summer
(Sype 1975, Nussbaum et al. 1983, Briggs 1987), (Appendix A).
All lifestages of R. cascadae are highly aquatic, requiring year round water to survive. During winter, frogs are inactive for up to seven months under ice and deep snowpack. Little is known about overwintering habitat requirements beyond observations of individuals overwintering in muddy substrates associated with a spring fed pond in
Oregon (Briggs 1987). Rana cascadae have used a wide variety of habitats beyond those used for breeding. For example, frogs have been observed in all types of aquatic environments including ponds, meadows, deep lakes and creeks by Bury and Major
(1997) in Washington, Brown (1997) in Oregon and Welsh et al. (2006) in California.
These observations demonstrate R. cascadae uses a variety of often patchily distributed habitats, and suggests that individuals may disperse and migrate seasonally for specific life history attributes such as breeding, summer foraging and overwintering.
Understanding the spatial-temporal dimension of animal populations, and the interplay between environmental heterogeneity and individual movement, has long been considered an extremely important aspect of a species spatial dynamics (Turchin 1998).
Movements provide key information regarding how animals use the environment, including migration patterns, dispersal, and resource use (Sinsch 1990, Semlitsch 2008).
5 Determining the scale at which populations function has become central to managing amphibian species (Trenham et al. 2001, Petranka et al. 2004, Smith and Green 2005,
Petranka and Holbrook 2006). However, dispersal and migration behaviors can operate at different spatial and temporal scales, so it is essential to understand each component separately. For example, Semelitsch (2008) found dispersal and migration terms have been used interchangeably to identify amphibian movement patterns in the past, which has lead to confusion in describing specific population dynamics and managing species.
Since R. cascadae lives in patchily distributed aquatic environments, comprehensive information on spatial ecology is vital to understanding their life history.
Studies on other pond breeding montane ranid frogs existing in patchily distributed habitats indicate species life histories can be complex, with populations having individuals migrating seasonally between separate resources (Pope et al. 2000, Pope and
Matthews 2001, Pilliod et al. 2002). Dunning et al. (1992) identified this life history strategy as Landscape Complementation, where two or more non-substitutable resources are separated, thus requiring animals to move among separate patches to survive.
Dispersal may also be operating at different spatial-temporal scales than migration, so dispersal’s role in population dynamics may be underestimated for many amphibians as a result of study areas being too small (Smith and Green 2005).
Currently, the only empirical information on R. cascadae movement is from a study on amphibian recolonization in the destruction zone at the base of Mount St.
Helens volcano after the 1980 eruption. Two marked R. cascadae moved 0.75 and 1.2 kilometers between waterbodies (Crisafulli et al. 2005). Due to this limited information,
6 dispersal capabilities are also not well understood. Crisafulli et al. (2005) found R. cascadae took up to seven years to colonize new habitats 3.7 km from source populations that survived the 1980 Mount St. Helens eruption. A microsatellite genetics study on R. cascadae in Oregon and Washington also found gene flow sharply dropped at distances greater than 10 km suggesting strong genetic isolation among populations (Monsen and
Blouin 2004).
The goals of this study were to provide detailed information on vital life-history components of R. cascadae at the scale of an entire catchment and its adjacent watersheds. Specific objectives were to, 1) identify spatial-temporal trends in demographic groups, 2) determine natal dispersal rates and its inverse natal site fidelity,
3) determine seasonal migration rates and long term adult migratory patterns, 4) identify seasonal resource use, and 5) identify minimum dispersal rates of individuals dispersing to adjacent basins. The results from this study will provide comprehensive information on migration, dispersal, and associated seasonal resource use for this species. They will also provide needed life history information in California where populations of R. cascadae are currently in decline and (or) are extremely isolated from one another.
MATERIALS AND METHODS
Study Area Description
My study was conducted with a large-scale component including a group of six
sub-alpine basins. In addition, I included a small-scale component nested in a single
basin, Echo Lake Basin. The six neighboring basins are separated by steep rocky
ridgelines, in the southeastern portion of the Trinity Alps Wilderness, Klamath
Mountains, California (40˚55’14” N, 122˚53’33” W, Figure 2). Elevations of sampling
sites throughout the extensive study area ranged from 1946 to 2310 m. The geology of
the study area is dominated by the serpentine rich Trinity ultramafic pluton which is thought to have formed in the late Jurassic period (Lipman 1964). Other rocks present in lesser amounts include various small granite intrusions, gabbro and schists. The majority of the topography consists of steep elevational gradients on ridge slopes composed of bare rock outcrops and expansive talus fields. The topography of the region has been shaped by periodic glaciation, most recently from two separate events within the
Pleistocene epoch (Sharp 1960). As a result, superficial glacial deposits and moraines are scattered throughout the area, adding to the complex topography. This rugged terrain is typical throughout the sub-alpine portion of the Klamath Mountains.
The climate of the Klamath Mountains is Mediterranean, characterized by wet, cool winters and dry, warm summers. The local expression of this climate regime is remarkably variable due to a strong west to east moisture and temperature gradient
7 8
Figure 2. Study region located in the southeast Trinity Alps Wilderness, Klamath Mountains, California. Black dotted lines represent catchment boundaries for major drainage basins. Basins and aquatic resources included in the study are labeled.
9 caused by proximity to the Pacific Ocean. Steep elevational gradients further influence temperature and spatial pattern of precipitation, through orographic effects (Skinner et al.
2006). The dominant source of precipitation falls as snow from November through May, with the overall May 1 average equaling 210 cm from 1946 to 2007 (California
Department of Water Resources 2007). From 2003 to 2008, regional snow pack was
170%, 125%, 120%, 251%, 38%, and 110%, respectively, of the May 1 average
(California Department of Water Resources 2009). Precipitation from June through
October usually falls as rain from localized thunderstorms. During this six year study, total rainfall from June through October (the active period for R. cascadae) ranged from
0.4 to 5.5 cm (mean: 2.6 cm) annually. Air temperatures during this period ranged from -
11.1 to 31.6º C (mean: 12.7º C) at the nearest weather station (Red Rock Mountain, ~10 km north, Elevation: 2042 m). In contrast, air temperatures when frogs were overwintering, November to May, were much cooler and ranged from -19.4 to 24.4º C
(mean: 0.9º C).
The flora of the Trinity Alps has been described in detail by Ferlatte (1974).
Floristic zones well represented in this region include open sub-alpine forest, montane chaparral, and sub-alpine meadows which are composed of serpentine tolerant plant species. Thin patches of western white pine (Pinus monticola), foxtail pine (P. balfouriana) and Jeffrey pine (P. jeffreyi) dominate the sparse forest canopy, which contains a total of six conifer species. Huckleberry oak (Quercus vaccinifolia) and angelica (Angelica arguta) dominate dry exposed slopes of montane chaparral. Meadow patches contain a variety flowering plants, but are dominated by sedges (Cyperaceae) and
10 grasses (Poaceae), along with corn lily (Veratrum californicum), white-flowered schoenolirion (Schoenolirion album), California pitcher plant (Darlingtonia californica), meadow goldenrod (Solidago canadensis elongata), and mountain spiraea (Spiraea densiflora).
The aquatic resources in this region are highly diverse in both morphology and their respective annual hydrologic cycles. Surface water is derived from three main sources: snowpack, groundwater, and rainfall; all providing spatial and temporal variation in water available to sustain aquatic features. As snow melts in spring, meltwater fills lakes, ponds and streams to maximum capacity. During this time, most aquatic features are connected through a complex network of temporary and permanent streams. As summer progresses, groundwater becomes the primary water source, with many snowmelt derived features completely drying, restricting the connections between aquatic resources. Although some large ponds and a few lakes are present in the study region, aquatic features are dominated by sub-alpine wet meadow complexes (Figure 2). These meadows collectively contain thousands of small ponds and extensive stream networks along with occasional Darlingtonia fens.
Echo Lake Basin, site of the intensive study, is a medium sized glacial cirque basin (342 ha) encompassed by steep jagged peaks reaching elevations up to 2497 m.
These peaks are among some of the highest in the Klamath Province. The basin is west facing and drains into Stuart’s Fork, a major tributary of the upper Trinity River (Figure
2). A diverse assemblage of aquatic features create a matrix of patchy “island” aquatic habitats accounting for 3.5% of the basin area (Figure 3). These features are separated by
11
Figure 3. Distribution of surveyed aquatic habitats within Echo Lake Basin, Trinity Alps Wilderness, California. Seven patch complexes are outlined with dotted lines. Three large patch complexes (grouped isolated patches that were located in close proximity) include: VMM = Van Matre meadow complex, DPC = Deep Creek meadow complex, CAM = Cascade meadow complex. Within patch complexes, fifteen isolated patches and 73 sampling zones are represented by a 4 letter code. The first three letters represent the patch ID; the last letter represents the sampling zone ID within a specific patch. Stream, pond, and lake sampling zones are separated by black hatch lines.
12 inhospitable dry rocky slopes, steep talus fields and post-glacial moraine piles. During each spring 3.5% of the basin surface area contains an estimated 3.12 hectares of surface water including: 63 groundwater springs, one lake, 588 ponds, and 13.9 km of streams.
Most of the aquatic features are within wet meadow complexes, with the exception of three isolated ponds and one lake (Figure 3). Aquatic features occur at 1960 to 2279 m in elevation and are located on slopes ranging from 0-48% (mean: 15.5%).
Surveys and Marking
The majority of survey effort for R. cascadae focused on the small-scale study in
Echo Lake Basin. Here, systematic visual encounter surveys for R. cascadae were
conducted approximately every two weeks throughout the annual active period (May to
October) from 2003 to 2006. All habitat patches in Echo Lake Basin were surveyed
resulting in a complete census of all defined aquatic habitats (Figure 3). In 2007 and
2008, complete censuses were restricted to the breeding season in the spring, with
additional limited survey effort occurring during each summer and fall.
Habitat patches in five basins proximal to Echo Lake Basin were surveyed on a
limited basis over the course of this study for the large-scale study component. These
areas include Red Mountain Meadows, Siligo Meadow complex, Billy-Be-Damned Lake,
Deer Lake, and an unnamed meadow system and adjacent pond, hereafter referred to as
Atlantis (Figure 2). Of these proximal sites, Red Mountain Meadows received the most
survey effort with surveys conducted all six years. Each of the other sites were surveyed
13 in four years (Siligo Meadow complex, Deer Lake), or three years (Billy-Be-Dammed
Lake and Atlantis).
Visual encounter surveys were conducted during calm and warm periods when conditions are most effective for detecting diurnal herpetofauna in temperate regions (Crump and Scott 1994, Thoms et al. 1997). Most surveys were conducted between 1000 and 1900 h, when sites were exposed to direct sunlight. Adverse weather conditions were avoided, especially when wind and rain limited visibility, or when air temperatures were below 12º C. Visual encounter surveys consisted of one to two observers walking the perimeter of all lentic habitats and the banks of all streams (Crump and Scott 1994, Thoms et al. 1997). Lentic sites, with surface areas greater than 0.2 hectares were surveyed by two observers due to the complex irregular shorelines and adjacent talus fields. One surveyor walked the shallow littoral zone, while a second searched on land within two meters of the shoreline. Streams greater than 0.5 m wide were also searched with two observers, with one surveyor walking each bank. Flooded portions of meadows were searched using a zigzag pattern (Thoms et al. 1997) after all confined lentic and stream habitats were searched.
Juvenile and adult frogs were captured. However, young-of-the-year frogs present each fall were not captured to reduce possible handling and marking stress on metamorphosing individuals going into their first winter. All R. cascadae were captured by hand or dip net. When captured, individual frogs were placed in numbered pint size plastic freezer bags filled with water. Capture locations were marked with a numbered pin flag and individuals were returned to their capture location after processing. All R.
14 cascadae locations were mapped using a Trimble Geoexplorer III global positioning system, (Sunnyvale, California).
To assess habitats associated with reproduction, egg mass surveys (Crouch and
Paton 2000) were conducted during each spring throughout Echo Lake Basin. Surveys occurred weekly throughout each breeding season to determine the total basin-level reproductive output. On cloudy days, polarized glasses were utilized to avoid surface glare. Individual egg masses were labeled with numbered pin flags to avoid double counting on subsequent visits.
All captured frogs were weighed to the nearest gram with a 60 g Pesola spring scale (accuracy ± 3%). Snout-urostyle lengths (SUL) were measured to the nearest mm using metric dial calipers (accuracy ± 1 mm). Measurements of SUL were standardized by “maxing out” SUL lengths of individual frogs. This was achieved by placing frogs on a flat surface and applying gentle lateral and vertical pressure on the animal when measuring. This method of measurement minimized SUL length variation by not allowing frogs to contract during measurement.
Immature R. cascadae (less than 40 mm SUL) were marked from 2003 through
2005 using a biocompatible Visual Implant Elastomer (Northwest Marine Technologies,
Shaw Island, Washington, USA). Visual Implant Elastomer has no reported negative effects on survival, growth and behavior of other amphibians, and has recently been preferred over toe clipping (Nauwelaerts et al. 2000, Bailey 2004, Wahbe et al. 2004). All new captures from this group were cohort marked with a color and digit code to identify individuals to a specific patch. Cohort marks for each patch were changed annually so
15 frogs could be identified to the patch and year marked. Visual Implant Elastomer was discontinued after 2005 because a sufficient sample size had been obtained for intended analyses. All captured R. cascadae greater or equal to 40 mm SUL were individually marked with Passive Integrated Transponder (PIT) tags (TX1400L, Biomark Inc., Boise,
Idaho, USA) throughout the study period. PIT tags have recently become a method of choice for marking medium to large size anurans (Ferner 2007). PIT tags were inserted into a 2 mm wide, V-shaped incision through the dermis using stainless steel dissecting scissors sanitized in 90% ETOH prior to each marking (Pope and Matthews 2001). The incision was made on the dorsal surface, anterior and lateral to the sacral hump. PIT tags were inserted under the skin at the incision point and then pushed posterior to the sacral hump.
Radio Telemetry
Radio telemetry was used to study movement rates, migration routes, and
microhabitat use of adult R. cascadae. Radio transmitters (BD-2 Holohil Systems, Ltd.,
Carp, Ontario, Canada) were attached to 30 female and 21 male R. cascadae from 20
June through 3 October in 2003, and 18 June through 10 October in 2004. In 2003,
transmitters were fitted to individuals using a polyester ribbon waist belt designed by
Pilliod et al. (2002). Some frogs developed dermal abrasions from this attachment
method (see Appendix A), so in 2004 transmitters were fitted to frogs by using glass seed
beads threaded with an elastic string waist belt (Muths 2003). To control frogs while
fitting transmitter belts, a holding device designed by Bourque (2007) was utilized.
16 Transmitters weighed on average 4.3 ± 0.15 g of the body weight. This is well under 10% of the body weight rule recommended by Richards et al. (1994).
Frogs were tracked every two to three days using a TR-4 receiver fitted with an
“H” style antenna (Telonics, Inc. Mesa, Arizona). Frogs were captured once weekly to perform visual health inspections for potential skin abrasions caused by transmitter attachment. When belt abrasions were identified, fine belt adjustments were made or transmitters were removed from individuals altogether to avoid further injury. Individual frogs were tracked for varying lengths of time, but overall, tracking sessions occurred from June through October, encompassing the entire annual activity period for frogs in the study area. Frog locations and habitat information were recorded using a GPS.
Position averaging was employed with 100 points taken at each location. All locations were post-processed using the nearest available base station to increase accuracy to 1-3 meters using Trimble Pathfinder Office 2.8 (Sunnyvale, California). Frogs were often sedentary and found within one meter of previous locations. In these instances only one set of coordinates was used to eliminate error generated from multiple GPS coordinates of the same location.
Resource Data Collection
Spatial attributes were mapped using a GPS, including water body shoreline perimeters, wet meadow perimeters, stream networks, spring source locations and R. cascadae breeding sites (Figure 3). Defined stream channels were mapped by walking a single path along the length of each channel. Wet meadow perimeters were defined by
17 placing pin flags at the edge of the wetted area and then walking the perimeter using a
GPS. All permanent streams and large ponds within these meadow perimeters were also mapped. Spatial data was differentially corrected using a nearby base station and further ground-truthed with high resolution satellite imagery and site visits to maximize mapping accuracy.
All lentic and most lotic features (~83%) in Echo Lake Basin were characterized and measured to establish site-specific seasonal hydroperiod dynamics. A total of 3.6 km of streams that were mainly intermittent and isolated steep gully features were not measured (~17%). Measurements were collected three times (spring, summer and fall) within each patch and summarized for 73 pre-established sampling zones within the 15 patches as indicated in Figure 3. These sampling zones were used to define local seasonal hydrological conditions that occur within patches. In meadow patches, sampling zones were identified as approximately 50 m long belts of varying widths. Stream and pond or lake patch sampling zones were based on 50 m transects along the stream or shoreline
(Figure 3). All aquatic features within these sampling zones were measured. Pond patches were characterized as spring-fed or runoff ponds based on their respective water sources.
Spring-fed ponds were directly connected to, or formed by, a nearby spring source, whereas runoff ponds were depressions filled by spring snowmelt meltwater and summer rainfall.
Lentic feature (e.g. spring pond, runoff pond or lake) measurements included surface area and maximum depth in each zone for each visit. Surface area was determined by multiplying the average of three width measurements by the length. Lotic
18 feature (e.g., springs and streams) measurements included surface area and maximum depth in each zone for each visit. Pond features were categorized into two groups based on size and depth. Large deep ponds were defined by being greater than or equal to 1.5 m deep and having a surface area greater than or equal to 200 m2. Waterbodies smaller and
shallower were defined as small shallow ponds. Sampling zone inventories were
collected seasonally. The spring season inventory was collected during peak snowmelt
runoff to represent the maximum catchment capacity of each patch. The summer
inventory was collected mid-season to represent average summer conditions. The winter
inventory occurred late in the season to approximate conditions as frogs became inactive
at overwintering locations. Measurements represent conditions for moderate to wet water
years as sampling zone hydrological data were collected in 2003 when snowpack reached
170% of normal. The Humboldt State University Institutional Animal Care and Use
Committee (IACUC) approved the methods employed for this project (permit ID:
02/03.W.106-A).
Data Analysis
All spatial variables were calculated in a geographic information system (ArcGIS
9.3, Environmental Systems Research Institute Inc. 2008) using spatial analyst and
Hawth’s Tools, an ArcMap kernel and spatial analysis extension (Beyer 2004). Data
analysis categories and associated parameters are organized in Table 1 to accompany
descriptions in the following sections. I used S-PLUS v. 6.0 (S-Plus 2001) and NCSS v.
2004 (Hintze 2001) to perform statistical operations. Error calculations reported with
Table 1. Summary of spatial, temporal and demographic information used for analyses of life history attributes of Rana cascadae in the Southeast Trinity Alps Wilderness, California, 2003 to 2008. See text for variable and analysis descriptions.
Analysis Measurement Unit Scale of Analysis R. cascadae Gender Temporal Scale1 Timeframe age group Population Space Use X, Y Coordinates Echo Lake Basin YOPY2 ─ S/ Annual 2003-2005 Juvenile ─ S, W/ Annual 2003-2005 Adult ─ B, S, W/ Annual 2003-2006 Aquatic Resource Inventory Aquatic feature Patch and zone NA NA B, S, W 2003 (Echo Lake Basin)
Aquatic Resource Use Aquatic feature Echo Lake Basin Eggs/ Larvae ─ Annual 2003-2006 (categorical) YOPY2 ─ Annual 2003-2005 Juvenile ─ Annual 2003-2006 Adult ♀, ♂ Annual 2003-2006
Seasonal Resource Use Zone Echo Lake Basin Eggs/ Larvae ─ B/ Annual 2003-2006 YOPY2 ─ S, W/ Annual 2003-2005 Juvenile ─ S, W/ Annual 2003-2006 Adult ♀, ♂ B, S, W/ Annual 2003-2006
Movements of Immature Frogs Distance (m) Echo Lake Basin YOPY2 ─ S, W 2003-2005 Juvenile ─ Annual 2003-2006
Breeding Dispersal (FTB) Patch complex/ Echo Lake Basin and Juvenile, adult ♀, ♂ B 2003-2007 (EB) distance (m) Red Mountain Meadows Adult ♀, ♂ B 2003-2007
Seasonal Migrations Patch/ Echo Lake Basin Juvenile ─ W/ Annual 2003-2005 distance (m) Adult ♀, ♂ B, S, W/ Annual 2003-2006
Annual Migration Patterns Patch Echo Lake Basin Adult ♀, ♂ Study/ Annual 2003-2008
Movement Rates and Routes Telemetry/ Echo Lake Basin Adult ♀, ♂ S, W 2003-2004 distance (m) Interbasin Dispersal Distance (m) Multiple basins/ regional YOPY2-adult ♀, ♂ Study 2003-2008 1B = Breeding, S = Summer foraging, W = Overwintering, Annual = Data summarized annually, Study = Spans entire study duration. 2Young of previous year. 19 20 mean values represent ± 1 standard error. All statistical tests were considered significant at α less than 0.05.
Rana cascadae captures were placed into one of four age groups, and juveniles were sexed when possible so age and gender could be addressed in the analysis (Table 1).
Age categories included eggs or larvae, young of previous year, juveniles and adults. I defined “young of previous year” as frogs that have survived their first winter and were living in their second year of life. Since young R. cascadae grow relatively fast (Briggs
1978), size cutoffs separating young of previous year from older juveniles were generated using length frequency histograms of captured individuals for each season separately.
Like many other anurans, R. cascadae exhibit sexual size dimorphism as adults (Monnet and Cherry 2002) with females attaining larger sizes than males. Although male R. cascadae developed nuptial pads at 45mm SUL in this study area, I augmented this secondary sexual characteristic by also using reproductive behavior to determine age group. During each breeding season (2003-2008), only active adult frogs exhibiting courtship behavior were observed congregated around breeding sites. Based on minimum sizes of these animals, I determined the smallest size of adult frogs to be 50 mm SUL for males, and 58 mm for females. Furthermore, all measured amplexing individuals (n =
66), as well as gravid fall females, exceeded the established minimum size cutoffs.
The spatial scale of this study varied based on specific questions (Table 1). One could overestimate dispersal by including local movements among patches in close proximity. To determine between patch breeding dispersal, independent patches were defined as: 1) separated by distances greater than 100 meters, and 2) collectively
21 supporting all lifestages of R. cascadae. These criteria resulted in the dependant patches being grouped into three larger but separate complexes (Figure 3). For example, Upper
Van Matre Meadows is an isolated patch but is in close proximity (i.e. < 100 meters) to others so it was grouped into the Van Matre meadow complex. As a result, 15 identified patches were reduced to seven for the dispersal analysis (Figure 3). Dispersal was considered a hierarchal process, with dispersal within Echo Lake Basin (Figure 3) referred to as intrabasin dispersal and dispersal between basins (Figure 2) referred to as interbasin dispersal.
In contrast to dispersal, the dependent patch scale was used to quantify seasonal migrations in order to address seasonal movements among non-breeding resources.
Seasonal migration was defined as any movement greater than 100 m that occurred between patches between two adjacent seasons. Population space use and distribution was analyzed at the whole-basin scale in Echo Lake Basin only to reflect the area assumed to support all major population-level functions of a viable population.
Seasonal aquatic resource use data were collected at an intermediate “zone scale”
(Figure 3) within Echo Lake Basin, so seasonal changes in resource availability could be detected at a local, within-patch scale. Aquatic habitat use was defined as the habitat type where each animal was captured (e.g. lake, spring pond, stream).
Seasonal Use of Space and Resources
To assess seasonal and age specific differences in the use of space by R. cascadae in the Echo Lake Basin, utilization distributions were calculated using bivariate fixed-
22 kernel estimators (Worton 1989). The utilization distribution describes the relative probability of resource use, at a location relative to all others in the space of inference, within a specific timeframe (Van Winkle 1975). The utilization distribution is typically used for home range analysis of individual animals (Kernohan et al. 2001). The use of utilization distribution was adapted to describe and compare seasonal distributions of R. cascadae within Echo Lake Basin. Specifically, utilization distributions were calculated for seasonal distribution data separately for young of previous year, juvenile, and adult R. cascadae annually from 2003 to 2006 (Table 1). utilization distributions could not be determined for immature lifestages (young of previous year and juveniles) during the breeding season because many non-breeding habitats were not available to survey due to deep snowpack. By in large, immature lifestages were inactive during the breeding season, even at breeding sites with active adults present. Winter season utilization distributions for young of previous year were also not determined because yearly sample sizes were insufficient (Seaman et al. 1999).
Any calculation of space use from multiple observations can suffer from autocorrelation. That is, an animal’s location at time t is dependent on t – 1 (Swihart and
Slade 1985). To minimalize autocorrelation, only one capture per individual, per season, per year was used when calculating each utilization distribution. Also, during the breeding season, adult female R. cascadae are elusive resulting in very few captures. Egg mass counts and locations were used during the breeding season to represent the relative adult population distribution during this period. When more than one capture of an individual occurred in the summer, only the capture closest to the middle of the season
23 was used to reflect summer locations. If an individual was captured more than once during the fall, only the capture closest to winter was included.
The smoothing parameter or bandwidth, is considered the most important component effecting the outcome of any kernel analysis (Silverman 1986, Seaman et al.
1999, Gitzen et al. 2006). Kernel estimators using large bandwidth values can overly smooth a utilization distribution estimate where small bandwidth values can produce negatively biased utilization distribution estimates (Gitzen et al. 2006). The R. cascadae population I studied exists in a matrix of isolated patches (Figure 3) separated by large areas of inhospitable terrain. Analysis trials using traditional approaches to estimating ideal bandwidths, e. g. least-squared cross-validation, vastly overestimated population space use into areas outside of patches. To account for the landscape physiognomy, a global fixed kernel estimate for all R. cascadae captures was generated by providing bandwidth values until the resulting utilization distribution best fit the overall available habitat. The resulting bandwidth (= 50) was then used to build kernel generated utilization distributions for all demographic groups by season and year.
The volume of intersection index statistic, developed by Seidel (1992), was used to compare the degree of overlap between population-level seasonal utilization distributions of young of previous year, juvenile, and adult R. cascadae in Echo Lake
Basin. The volume of intersection index calculates the volumetric overlap of two utilization distributions by
∞ Volume of intersection , , , ∞
24 where ƒUD1 is the utilization distribution of population component 1 and ƒUD2 is the
utilization distribution of population component 2. The integral in the equation calculates
the volume of the intersection between the two three dimensional planes (Seidel 1992,
Kernohan et al. 2001). The volume of intersection statistic is bounded between 0 and 1,
with a value of 0 equaling no space use overlap of two utilization distributions and 1,
complete space use sharing of two utilization distributions (Millspaugh et al. 2000). The
resulting volume of intersection score can be interpreted as the percentage of three
dimensional overlap occurring between two utilization distributions.
Due to the diversity of size and type of aquatic resources, direct comparisons
based on quantity of hydrological features could not be explored. For example, Echo
Lake and some permanent ponds contained disproportionably large surface areas of water
relative to meadows with small ponds. The total surface area of water was summarized
for each of the 73 sampling zones within the 15 patches in Echo Lake Basin (Figure 3) to
determine how much water was available for breeding, summer and winter seasons. The
proportion of water available for the summer and winter seasons was summarized relative
to the maximum in the spring. Maximum depths in each sampling zone were determined
for the winter season and represent the deepest water feature available for overwintering.
Each sampling zone was categorized as providing breeding, summer foraging, or
overwintering resources. These categories were based on seasonal hydrology and the
presence of breeding (i.e. egg masses, tadpoles) or frogs during a given season. Sampling
zones that were dry in a particular season were not considered available for use at that
time. Many sampling zones were used for multiple seasons, so rankings were based on
25 the total use of each zone over all seasons. Resource use information was collected for each egg mass and frog captured and recorded as found in a lake, large deep pond (< 1m), small shallow pond or stream. These data were summarized from 2003 to 2006 due to uneven survey effort for non-breeding seasons after these years.
Movements of Immature Frogs
Movement and fidelity of young of previous year and juveniles were determined
separately to examine possible differences between these demographic groups. To assess
the average distance young of previous year and juveniles were found from breeding
sites, I measured the distance of each capture from the nearest breeding site for each age
group from 2003 to 2005. Movement frequency and distances between patches were
determined separately for each immature age group because of differences in marking
method. The young of previous year group received only a site and year “batch” mark
using Visual Implant Elastomer, so a limited number of individuals could be determined
to move among patches. For example, the number of young of previous year that dispersed to a new patch could only be determined as the maximum number of captured individuals carrying a unique mark on a given survey for the whole year. Movements of juvenile frogs were limited to dispersal analysis.
Breeding Dispersal
For this study, dispersal was defined as any permanent movement away from a
patch containing at least one breeding site. Breeding dispersal rates were calculated
separately for first time breeders and for experienced breeders (Trenham et al. 2001,
26 Gamble et al. 2007). Since an individual cannot be defined as a successful disperser until it survives and attempts to breed at a non-natal site (Gamble et al. 2007), strict limits were imposed to the data used for this analysis. To estimate successful dispersal probabilities for first time breeders, two capture history criteria were required, 1) adult individuals needed to be captured at a breeding site during a breeding season, and 2) these adults needed to original site/ cohort marks as juveniles to identify their patch of origin. Successful dispersal was calculated for each patch by determining the percentage of adult frogs that attempted to breed in patches different from their patch of origin. For example, Deep Creek Meadow complex (Figure 3) had 50 unique juveniles that were later captured as breeding adults. Thirty of these individuals were captured elsewhere, so the probability of successful dispersal was estimated as 30/(20 + 30), or 60%. Dispersal rates for experienced breeders were determined similar to first time breeders, using adults captured at active breeding sites two or more years after marking to determine the probability a frog used two or more patches to breed.
Seasonal Migrations
Seasonal migrations were defined as an individual movement between separate
patches separated by distances greater than or equal to 100 m that occurred between two
adjoining seasons. For example, an adult migration to a breeding site was determined by
calculating the distance between a frogs initial location at a overwintering site in 2005,
and its breeding site location in 2006. Summer and fall migrations were determined based
27 on capturing an individual between breeding and summer, or between summer and fall seasons, within the same year.
In order to address seasonal migrations (between spring [breeding], summer and fall), cutoff dates for R. cascadae captures were established. Due to wide geographic variation in R. cascadae reproduction timing in Echo Lake Basin (up to 6 weeks), dates between breeding and summer seasons were determined independently for each patch.
Rana cascadae are explosive breeders (< 10 days) at individual patches (Sype 1975), so the beginning of summer at each patch was defined as beginning one week after the last egg mass was laid annually. Captures at individual patches prior to this time were considered within the breeding season. The start of the fall/winter season was determined to begin on September 15th annually. By this date, most seasonal summer habitats were
dry, daily air temperatures decreased relative to summer, and most frogs were staging at
overwintering locations.
The leptokurtotic distribution of the movement distance data warranted the use of
non-parametric tests (Zar 1999). Kruskal-Wallis analysis of variance (ANOVA) on ranks
test was used to test if mean migration distances by adult frogs differed by season
(breeding, summer and winter). Differences between groups were determined using the
Bonferroni corrected Kruskal-Wallis Z test. Mann-Whitney U tests were used to
determine if mean movement by season differed between male and female frogs.
Captures after the breeding season in 2007 were excluded from migration, dispersal, and
site fidelity analysis because not all patches were surveyed with equal effort during this
field season.
28 Movement rates of adult R. cascadae were best described by comparing daily locations of telemetered individuals. Since most radio tracking intervals were once daily, calculated migration rates are likely an underestimate of the frog’s true travel but reflect daily activity. Average daily movement rates were compared between seasons (spring, summer and fall) and sex using analysis of variance (ANOVA). Post hoc comparisons of groups were performed using the Tukey Kramer multiple comparisons (Zar 1999).
Interbasin Dispersal
Minimum observed dispersal rates for interbasin dispersing frogs was determined
for each basin. This was determined by dividing the total number of individual frogs that
dispersed to new basins by the total number of uniquely marked frogs from the basin of
origin. Dispersal distance was calculated as the minimum straight line distance a frog
could travel between the last capture in the basin where it originated and its initial capture in the destination basin. If movement paths bisected absolute barriers, such as vertical rock cliffs or summits, paths were adjusted to the nearest low point on the ridge. Since this mountainous landscape is characterized with steep elevation relief, movement paths were corrected using a 10 meter resolution digital elevation model in a geographic information system (ArcGIS 9.3, Environmental Systems Research Institute Inc. 2008) to approximate dispersal on the land surface.
RESULTS
Surveys and Captures
A total of 610 systematic visual encounter surveys over 50 census periods were conducted in Echo Lake Basin and Red Mountain Meadows from 2003 to 2007. These surveys encompassed the entire annual active period of R. cascadae (May through
October) with limited effort in 2007. Additionally, 17 opportunistic surveys were conducted at a subset of sites in 2008. Overall, the study resulted in 7,348 frog captures over the six year period. Of these captures, a total of 1,980 individuals were identified and marked: 1,188 with PIT tags, and 1,075 with Visual Implant Elastomer. Three hundred and sixty seven individuals tagged with Visual Implant Elastomer received PIT tags after attaining larger sizes. Additionally, 353 egg masses were counted over the six years at 40 lentic waterbodies.
Summer Use of Space
Post-metamorphic age groups of R. cascadae were most widely distributed during the summer months in Echo Lake Basin. Population level 95% isopleths of the fixed kernel analysis indicated that young of previous year had the most confined summer distribution of the three age groups, averaging only 44 and 53 percent of the estimated patch area used by juveniles and adults, respectfully (Table 2). Juveniles and adults had similar areas of estimated space use during the summer utilizing an average of 20.61 and
19.06 hectares squared, respectfully (Table 2).
29 30
Table 2. Volume of intersection index for utilization distributions of separate age groups of Rana cascadae during the summer season from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Young of previous year individuals were captured and marked from 2003 to 2005 only.
Summer utilization distribution 95% 95% VOI2 comparisons Isopleth1 Isopleth1 Summer (YOPY3 with Juveniles) YOPY3 Juvenile 2003 12.60 23.43 0.46 2004 9.73 26.03 0.40 2005 8.58 20.31 0.30
Average 10.30 23.26 0.39
Summer (YOPY3 with Adults) YOPY3 Adult 2003 12.60 15.86 0.26 2004 9.73 21.13 0.36 2005 8.58 21.51 0.35
Average 10.30 19.5 0.32
Summer (Juveniles with Adults) Juvenile Adult 2003 23.43 15.86 0.47 2004 26.03 21.13 0.59 2005 20.31 21.51 0.57 2006 12.68 17.73 0.44 Average 20.61 19.06 0.52
1 95 % volume contour (squared hectares). This value represents the boundary of the area that contains 95% of the points used to generate the population scale fixed kernel estimate. 2Volume of intersection index ranges between 0: two seasonal population utilization distributions with no overlap in area used, and 1: two identical seasonal population utilization distributions with complete overlap in area used. 3Young of previous year.
31 Overall, the combined population utilization distributions indicate a substantial difference in the intensity of space use among young of previous year, juvenile, and adult age groups during summer (Figure 4). Small volume of intersection indices between summer utilization distributions illustrate this point, with the largest annual summer volume of intersection comparison between two utilization distributions not exceeding
0.59 (mean = 0.42 ± 0.03)(Table 2). The greatest overlap occurred between juvenile and adults, with volume of intersection varying annually from 0.44 to 0.59, while the smallest summer utilization distribution overlap occurred between young of previous year and adults with volume of intersection varying annually from 0.26 and 0.36. The volume of intersection space use overlap between young of previous year and juvenile utilization distributions were also small, varying annually from 0.30 to 0.46 (Table 2).
Seasonal Space Use by Adults
Adult R. cascadae showed substantial differences with respect to space used
among the breeding, summer, and winter seasons. Population-level 95% isopleths of
fixed kernels indicate that adults were narrowly distributed during the breeding season,
equaling only 28 (summer) and 47 (winter) percent of the patch area used during the other seasons (Table 3). This pattern is evident in the visual depiction of the seasonal utilization distributions in Figure 5, with the widest distribution occurring during the summer, followed by the winter and breeding seasons.
The volume of intersection indices for adult R. cascadae indicate large shifts in space use occurred seasonally (Figure 5). Annual volume of intersection indices
Adult
Juvenile
Young of previous year
Base Map N
100 meters
Figure 4. Fixed kernel utilization distributions for different age groups of the Rana cascadae population during summer from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Young of previous year: density function of 416summer Young of previous year captures, Juvenile: density function of 900 juvenile summer locations, Adult: density function of 793 adult summer capture locations. The height of each density function depicts the relative probability of an individual occurring at each location within the study area by age group. Colored halo lines represent the boundary of the area that contains 95% of the volume of a probability density distribution. Bottom panel shows base map of aquatic habitats within Echo Lake Basin in black.
32 33 Table 3. Volume of intersection index between seasonal population utilization distributions of adult Rana cascadae from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. In 2006 the winter season was omitted due to a small (N = 37) sample size.
Adult seasonal utilization distribution 95% 95% VOI2 comparisons Isopleth1 Isopleth1 Adult Adult Breeding with Summer (adults) breeding summer 2003 6.20 15.86 0.32 2004 4.96 21.13 0.37 2005 5.33 21.51 0.37 2006 4.91 17.73 0.39
Average: 5.35 19.06 0.36 Adult Adult Breeding with Winter (adults) breeding winter 2003 6.20 8.43 0.30 2004 4.96 14.50 0.41 2005 5.33 12.36 0.39
Average: 5.50 11.76 0.37 Adult Adult Summer with Winter (adults) summer winter 2003 15.86 8.43 0.63 2004 21.13 14.50 0.66 2005 21.51 12.36 0.57
Average: 19.50 11.76 0.62
1 95 % volume contour (squared hectares). This value represents the boundary of the area that contains 95% of the points used to generate the population scale fixed kernel estimate. 2Volume of intersection index ranges between 0: two seasonal population utilization distributions with no overlap in area used, and 1: two identical seasonal population utilization distributions with complete overlap in area used.
Overwintering
Summer Foraging
Breeding
N
Base Map
100 meters
Figure 5. Fixed kernel utilization distributions for adult Rana cascadae population during the spring (breeding), summer and winter from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Breeding: density function of 195 R. cascadae egg masses, Summer: density function of 793 adult summer locations, Overwintering: density function of 297 adult capture locations just prior to the onset of winter. The height of each density function depicts the relative probability of an individual occurring at each location within the study area by season. Colored halo lines represent the boundary of the area that contains 95% of the volume of a probability density distribution. Bottom panel shows base map of aquatic habitats within Echo Lake Basin in black.
34 35 comparing breeding and summer, and breeding and winter utilization distributions were low with values not exceeding 0.41 (Table 3). The highest volume of intersections on average occurred between summer and winter utilization distributions, with volume overlap values ranging from 0.57 to 0.66 annually (Table 3).
Winter Space Use by Juveniles and Adults
The juvenile population was narrowly distributed during the winter. Mean
population level 95% isopleths of fixed kernels equaled only 47 percent of the patch area
used by juveniles during summer (Table 4). However, the mean winter juvenile isopleth
area equaled 92% of the mean winter area used by adults (Table 4) indicating a general
population-level contraction in space use during the winter. Large shifts in space use by
juveniles between summer and winter were apparent with annual volume of intersection
values ranging from 0.39 to 0.56 (Table 4). Differences in annual space use were
apparent between juveniles and adults during winter, with volume of intersection values
for these age groups ranging from 0.39 to 0.58.
Use of Aquatic Resources
Overall, aquatic resources used by R. cascadae in Echo Lake Basin varied widely
by both age and sex (Figure 6). The majority of egg masses produced annually in the
basin were deposited in small shallow ponds (55-78%) followed by large deep ponds (17-
33%). In contrast, very few egg masses were deposited annually in Echo Lake (4-12%)
and none were observed in streams. The majority (58-78%) of annual young of previous
year captures were found in small shallow ponds suggesting strong site fidelity to natal
36 Table 4. Volume of intersection index between seasonal population utilization distributions of juvenile and adult Rana cascadae from 2003 to 2005 in Echo Lake Basin, Trinity Alps Wilderness, California. In 2006 the winter season was omitted due to a small (N = 6) juvenile sample size.
95% 95% Utilization distribution comparison VOI2 Isopleth1 Isopleth1 Summer Winter Summer with Winter (juveniles) juvenile juvenile 2003 23.43 8.95 0.39 2004 26.03 14.86 0.56 2005 20.31 8.53 0.48
Average: 23.26 10.78 0.48 Winter Winter Winter (adults with juveniles) adult juvenile 2003 8.43 8.95 0.58 2004 14.50 14.86 0.50 2005 12.36 8.53 0.39
Average: 11.76 10.78 0.49
1 95 % volume contour (squared hectares). This value represents the boundary of the area that contains 95% of the points used to generate the population scale fixed kernel estimate. 2 Volume of intersection index ranges between 0: two seasonal population utilization distributions with no overlap in area used, and 1: two identical seasonal population utilization distributions with complete overlap in area used.
37
Figure 6. Mean percentage of annual egg production and captures of Rana cascadae (separated by sex and age group) found in each aquatic resource category summarized across all seasons within Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2006. The “Large Deep Pond” category represents Snowmelt and Eden ponds. Use proportions are based on 275 egg masses and 4,838 frog captures. The young of previous year age group is represented from 2003 to 2005, although present, none were captured in subsequent years.
38 ponds during the first year after metamorphosis. Notwithstanding, some movement away from natal sites was observed with 20-33% of young of previous year captured along streams. Many juveniles (40-44%) were also found at small shallow ponds, though more were found in streams (43-52%) than any other habitat.
Adult frogs had the highest diversity of aquatic habitat use of all age classes
(Figure 6). In contrast to immature frogs, adults were found regularly at Echo Lake and in large deep ponds as well as small shallow ponds and streams. Among adults, females used streams twice as much as males (Figure 6). Adult females were captured at stream habitats on average 40% (23-57%) of the time compared to 21% (13-30%) for males.
Based on annual captures, adult male R. cascadae were found in lentic habitats more often (70-87%) than any other post-metamorphic group.
Seasonal Resource Use
The availability of aquatic resources in Echo Lake Basin changed considerably
throughout the active period of R. cascadae with many features drying completely each
year (Table 5). By winter, only 56% of spring-season water surface area remained
(Figure 7). Moreover, removing Echo Lake, the largest waterbody in the basin, only 23%
of the water surface area remained by winter. Overall, 56% of the 73 sampling zones
were classified as perennial by containing aquatic features all year (Table 6, Figure 7).
The remaining sampling zones (44%) were ephemeral, with all features drying
completely by fall. Out of 73 zones, 23 (32%) were used for breeding, 73 (100%) were
used during summer and 41 (56%) were used during winter (Figure 7). Many sampling
Table 5. Summary of the total surface area of water by patch and season for Echo Lake Basin, Trinity Alps Wilderness, California from June to October of 2003. Individual patches are identified in Figure 3.
Season: Breeding Summer Winter Maximum % of % of Water Water Max Winter Water Maximum Maximum Surface Surface Feature Surface 2 2 Surface 2 2 Surface 4 2 1 Area (m ) 3 Area (m ) 3 Depth (m) Patch Id. Patch Type Area (m ) Area Area Black Bear Meadows Meadow 487 221 45 149 31 0.25 Cascade Meadows Meadow 822 397 48 204 25 0.88 Corn Lily Meadows Meadow 430 157 37 0 0 ─ Deep Creek Meadows Meadow 3553 2240 63 1145 32 1.24 East Van Matre Meadows Meadow 1391 446 32 361 26 0.62 Green Springs Meadow 794 568 72 178 22 0.36 Lower Van Matre Meadows Meadow 646 265 41 256 40 0.62 Mossy Springs Meadow 433 271 63 89 21 1.15 Middle Van Matre Meadows Meadow 1294 333 26 191 15 0.5 Upper Van Matre Meadows Meadow 581 381 65 168 29 0.34 Penthouse Ponds Meadow 388 217 56 186 48 0.7 Echo Lake Lake 11,698 11,527 99 11,465 98 5.1 Eden Pond Pond 278 215 77 140 50 0.8 Snowmelt Pond Pond 2483 489 20 0 0 ─ Van Matre Creek Stream 477 255 53 201 42 0.31 Mean: 1717 1199 53 982 32 1.0 Total Surface Area: 25,755 17,982 70 14,735 57 Total Surface Area without Echo Lake: 14,057 6455 46 3270 23 1Total surface area of water features by patch during the breeding season when water-bodies have maximum catchments and streams are flowing at bankfull capacity. 2Total surface area of water features by patch during the mid-summer and late fall seasons. 3Percent of the spring maximum surface area of water features by patch in the mid-summer and late fall seasons. 4Maximum depth of the deepest water feature by patch during the late fall season.
39 40
Figure 7. Map of Echo Lake Basin, Trinity Alps Wilderness, California, displaying seasonal hydrological properties of 73 sampling zones used by Rana cascadae from 2003 to 2008. Cross hatched zones and dotted stream segments dried completely by late summer so they were unavailable for winter use. Areas within green colored zones and solid line streams were used for both summer and overwintering habitats. Areas within solid blue zones were used year round. Yellow diamonds indicate where groundwater springs emerge. Orange circles represent traditional breeding areas used by R. cascadae.
41 Table 6. Mean percentage of annual summer Rana cascadae captures in Echo Lake Basin, Trinity Alps Wilderness, California, 2003 to 2006. Data were summarized by sampling zone hydrologic categories identified in Figure 3. Means are % (± 1 Standard Error).
Zone Classifications Perennial (N = 41 zones) Ephemeral (N = 32 zones) Breeding Summer Breeding Summer Summer and and Only Age Group/ Sex Winter1 Winter Summer (n = 19 zones) (n = 22 zones) (n = 4 zones) (n = 28 zones) YOPY2 Mean annual 59.2 (8.2) 25.7 (9.1) 5.11 (2.9) 9.9 (3.5) Range 49.8-75.6 8.5-39.2 0.6-10.5 5.4-16.7 Juvenile Mean annual 24.3 (1.4) 45.9 (3.3) 3.9 (0.8) 25.9 (2.2) Range 20.8-27.0 38.7-53.1 2.9-6.3 21.3-30.7 Adult Male Mean annual 35.2 (1.0) 41.0 (1.5) 15.3 (2.3) 8.5 (2.2) Range 33.0-37.4 38.2-44.2 9.7-20.0 4.9-14.4 Adult Female Mean annual 38.8 (6.1) 45.2 (10.1) 5.8 (1.6) 10.2 (3.0) Range 30.3-57.0 30.3-52.1 4.5-8.2 4.1-18.2
1Includes all Echo Lake zones and captures. 2Young of previous year.
42 zones dried completely by fall and overwintering locations were sometimes isolated features separated by areas with xeric conditions within patches that had contained aquatic features earlier in the year (Figure 7).
From 2003 to 2008, a total of 40 individual lentic sites were used for reproduction by R. cascadae throughout Echo Lake Basin, including 29 spring-fed ponds, 10 runoff ponds, and one lake. These breeding sites were located in 12 distinct locations throughout the basin, each containing from one to nine individual breeding features (Figure 7). Many breeding ponds in meadows were found in close proximity, with some being hydrologically connected during spring. Larger waterbodies, including Echo Lake,
Snowmelt Pond and Eden Pond, were relatively isolated from other breeding sites (Figure
7). The minimum distance separating breeding areas ranged from 114 to 528 m (mean =
245 m).
Rana cascadae were most widely distributed during summer, using all types available aquatic resources in Echo Lake Basin. In addition to perennial resources, some frogs utilized sites during the summer that completely dried by fall, requiring frogs to move to and away from these environments seasonally. Captures, by age group and sex, at both perennial and ephemeral resources are summarized in Table 6. Annual summer captures of young of previous year frogs were strongly associated with habitat zones containing perennial features (up to 85% of total annual captures). This was especially evident for habitat zones with breeding sites, which accounted for 59% of annual young of previous year captures. In contrast, only 24% of juvenile frogs were captured in perennial breeding zones during this period. Juvenile frogs were mostly found in non-
43 breeding perennial zones (46%) and ephemeral zones that dried (26%) during this period.
Summer captures of both adult male and female frogs were similar, with most captured in perennial habitat zones (76% and 84% of the total summer captures, respectively).
However, the majority of adult male and female captures occurred at non-breeding perennial zones, indicating directed movement away from breeding sites.
From mid-September through October, as temperatures cooled and ephemeral summer use areas contracted, most R. cascadae were found concentrated in areas with adequate water as they staged for overwintering (Figure 7). At some areas, this was as little as a single isolated spring pond that measured less than one square meter of surface area that was greater than 50 m from other waterbodies. Other overwintering areas contained multiple sites in close proximity, or were hydrologically connected. Based on late fall surveys (2003-2006), 69% of all frogs were observed using lentic and 31% using lotic environments (N = 548). The highest percentage of captures were found in spring ponds (38% ± 8.4) followed by Echo Lake (26% ± 9.0) and runoff ponds (5% ± 3.2). In lotic environments, overall captures were slightly higher in surface streams (16% ± 4.4) than spring fed streams (15% ± 2.7).
Besides Echo Lake, the deepest available overwintering habitat in the basin was
1.24 m or less. Only three of the 13 patches containing overwintering habitats had maximum depths exceeding one meter (Table 5). All stream habitats had maximum depths less than one meter. With the exception of Echo Lake and Eden Pond, most overwintering areas were directly connected to spring sources. Additionally, most breeding areas were directly connected to perennial springs (Figure 7) and many frogs
44 also used these areas for overwintering. For example, at Deep Creek Meadows 34 adult frogs were captured during late fall from 2003 to 2006 using spring-fed ponds. In subsequent breeding seasons, from 2004 to 2007, 18 (50%) of these animals were recaptured while engaging in breeding courtship in these ponds. At another patch (Mossy
Springs), only two small spring ponds (< 1 meter surface area) remained by fall and the patch became hydrologically isolated by ~ 50 m. One of these spring-fed ponds was searched on 8 October 2006 and two inactive juvenile frogs were found 0.5 and 0.6 m down a 15 cm diameter hole. This spring-fed pond was surrounded by a thick moss mat and the two frogs were found partially burrowed into these mosses.
Movements of Immature Frogs
The young of previous year age group of R. cascadae had distinct spatial patterns
that differed the most from all other age groups. Most of the 1095 young of previous year
captures were in close proximity to breeding sites. For example, 69% of captures were
within 25 m and 94% within 100 m of breeding sites. The furthest capture a young of
previous year frog was found away from a breeding site was 312 m. Movement of young
of previous year frogs between patches was rare with nine out of 544 (1.7%) marked
individuals moving distances greater than 100 m. These nine individuals moved a
minimum of 130 to 439 m (mean: 238 ± 38 m) between habitat patches.
In contrast to young of previous year, juvenile R. cascadae captures were
distributed farther away from breeding sites. Only 31% of juveniles were captured within
25 m of a breeding site and 78% within 100 m. Of the 540 juvenile frogs captured more
45 than once between the years of 2003 and 2006, 55% moved more than 100 m from their original locations. The mean net distance for juvenile frogs that moved more than 100 m was 338 ± 12.5 m (n = 296). Of the 121 juvenile frogs that had three years between locations, 70% had net movement distances greater than 100 m. The mean net movement for this group was 447 ± 29.2 m. Fifty-three juvenile frogs moved more than 500 m, and seven moved distances more than 1000 m. The maximum observed distance moved by a juvenile was 1886 m for females and 1534 m for males respectively.
Each fall, juvenile frogs were generally inactive compared to summer resulting in few late season captures. Thus, migration rates could only be assessed for 90 individuals.
Migrations of juveniles between separate summer and overwintering patches (> 100 meters apart) occurred, on average, only 17% of the time from 2003 to 2005. Migration distances averaged 245 m, Range: 134-359 m. Ten out of the fifteen observed juvenile migrations were from summer habitats that had dried completely by fall requiring frogs to move to permanent aquatic overwintering locations.
Dispersal
Dispersal rates of R. cascadae among the seven patch complexes in Echo
Lake Basin that survived to reproduction were exceptionally high. A total of 84 out of
164 (51%) adult frogs captured for the first time at active breeding sites originated from different patches as juveniles (Table 7, Figure 8). This indicates true dispersal (Semlitsch
2008). Dispersal rates varied by patch, with rates varying from 50 to 71% in the three patches having greater than 100 marked individuals. The two patches (Penthouse Ponds
Table 7. Dispersal probabilities for adult reproducing Rana cascadae with: A) first time breeding adults with juvenile capture histories, and B) experienced breeding adults between two or more breeding seasons. Dispersal probabilities are summarized for the seven patch complexes identified in Figure 3 within Echo Lake Basin, Trinity Alps Wilderness, California.
First-time Breeders Experienced Breeders Patch Total Returning Dispersing Dispersal Total Returning Dispersing Dispersal Total Complex juveniles breeders breeders probability number of breeders breeders probability number of marked (2004- (2004- (%)3 emigrating (2003- (2003- (%)3 emigrating (2003- 2008)1 2008)2 breeders4 2008)5 2008)6 breeders4 2006) Deep Creek Meadow complex 311 20 30 60 11 26 2 7.1 3 Van Matre Meadow complex 407 35 36 50.7 7 16 3 15.8 0 Eden Pond 33 0 0 0 2 8 0 0 0 Penthouse Ponds 75 4 0 0 2 10 0 0 0 Cascade Meadow complex 122 4 10 71.4 15 9 4 30.8 0 Echo Lake 81 17 7 29.2 20 30 1 3.2 4 Snowmelt Pond 44 2 1 33.3 23 50 1 2 4 Total: 1073 80 84 51.2 75 149 11 6.9 11
1 Number of individuals that returned to their natal patch complex to breed as adults. 2 Number of individuals that dispersed from their natal patch complex to breed elsewhere. 3 Dispersal probability: calculated as the number of dispersing breeders divided by the sum of known alive breeding adults from a particular patch complex. 4 Number of breeders a specific patch complex received from other patch complexes. 5 Number of individuals that returned to the same patch complex to breed between two or more years. 6 Number of individuals that dispersed from a patch complex used for breeding and breed elsewhere in subsequent years.
46 47
Figure 8. Map depicting dispersal directions of 84 first time breeding Rana cascadae between eight isolated patch complexes from 2003 to 2008 in the southeast Trinity Alps Wilderness, California. Arrows at the end of lines indicate dispersal direction. Dotted lines surrounding two patch complexes delineate the border of the two patch complexes that produced dispersing individuals: Van Matre Meadow complex and Cascade Meadow complex. Asterisk symbols indicate R. cascadae breeding locations used consistently throughout the study area.
48 and Eden Pond) that produced no observed first time breeding dispersers were those that contained the most isolated breeding habitats in the basin (Figure 3). Notably, four male first time breeders (5%) dispersed over a ridge to Red Mountain Meadows located in a proximal basin south of Echo Lake Basin (see Interbasin Dispersal section).
Females dispersed more often on average (66%) than males (46%), and moved significantly farther (Mann-Whitney U test: Z = 1.97, N = 84, P = 0.02). Mean dispersal distances were 669 ± 55 m for females and 546 ± 40 m for males respectively (Figure 9).
The maximum recorded dispersal distance for first time breeders was 1170 m for females and 1294 m for males. Of 84 dispersing frogs, a significant proportion (85%) relocated to patches that were higher in elevation than those they originated from (proportion test: Z =
-6.22, N = 84, P > 0.001) (Figure 8).
In contrast to first time breeders, experienced breeders had high annual fidelity to breeding sites. Of individuals captured in two or more breeding seasons only 11 (5 females and 6 males) out of 149 (7%) dispersed to new breeding sites among different years (Table 7). Dispersal probabilities for experienced breeders varied by patch and was low (< 8%) for patches with sample sizes exceeding 16 individuals. Notably, seven frogs
(six male and one female) were captured at active breeding sites over six years of the study and all remained faithful to the same breeding locations from 2003 through 2008.
Adult Seasonal Migrations
Based on recaptures of marked adult Rana cascadae in Echo Lake Basin, seasonal
migrations between patches occurred often, with migration patterns to and from breeding
49
100
Female Male
80
60
40
Dispersers of Percentage 20
0 0 200 400 600 800 1000 1200 1400
Dispersal Distance (m)
Figure 9. Cumulative dispersal distance curves for all first time breeding female (N = 27) and male (N = 57) Rana cascadae that were originally marked as juveniles and subsequently captured as adults at active breeding sites in Echo Lake Basin, Trinity Alps Wilderness, California.
50 areas being the most frequent for both males and females (Figure 10). A total of 320 unique adult R. cascadae had 733 instances where seasonal migration among breeding, summer and overwintering sites could be assessed. Overall, 43% of these cases resulted in frogs migrating between different patches between seasons. Of the frogs that migrated, travel distances to breeding sites were significantly greater (mean: 319 ± 19 m) than those to overwintering areas (mean: 231 ± 17 m) (Kruskal-Wallis one way ANOVA on ranks test: H = 19.8, P < 0.001). In contrast, migration distances traveled to summer use areas
(mean: 251 ± 11 m) were not significantly different from those to breeding or overwintering areas (Kruskal-Wallis Z-test, Bonferroni correction: Z = 2.39).
During the breeding season, adult R. cascadae captures were exclusively around the 12 lentic breeding areas in Echo Lake Basin. Many of these individuals wintered within, or in close proximity (< 100 m), to these breeding sites (n = 76). However, almost half (47% of females and 45% of males) completed spring migrations each year directly after spring emergence between separate overwintering and breeding areas (Figure 11).
Of the frogs that migrated, females moved significantly farther than males to reach breeding sites (Mann-Whitney U test: Z = 2.84, P = 0.004). Females moved a maximum of 1001 m (mean: 491 ± 79 m), while males moved up to 844 m (mean: 266 ± 17 m).
The longest round-trip migration observed for an adult female between breeding and overwintering areas was 1980 m (Figure 12, frog 3). This female wintered in a stream 239 m elevation below her breeding site (Snowmelt Pond) on the opposite side of the basin for two consecutive years. This frog was then captured three years later at
Snowmelt Pond indicating, at minimum, a third breeding migration and breeding site
51
Figure 10. Maps of Echo Lake Basin, Trinity Alps Wilderness, California, showing capture points (open circles) and connecting lines of 279 individual adult male (A) and 143 individual adult female (B) Rana cascadae captured two or more times from 2003 to 2006. Orange circles represent breeding sites. An individual frog can be represented by more than one line.
52
55 55
50 50 45 Female 45 Male 40 N = 32 40 N = 108 35 35
30 30
25 25
20 20
15 15
10 10 Percentage of Frogs of Percentage 5 5
0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
Migration distance from overwintering to breeding habitats (m)
Figure 11. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between overwintering and breeding seasons from 2003 to 2007 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years.
53
Figure 12. Examples of Long-term mark-recapture histories of seasonal migrations for seven adult female Rana cascadae in Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2008. All subsequent capture locations of individuals at specific habitat patches were within 25 m of original capture locations. Asterisks represent breeding locations used all years.
54 fidelity spanning 6 years. Remarkably, all other 11 established breeding areas in the basin were closer to this frog’s overwintering site than where she chose to breed. The farthest breeding migration completed by a male between breeding and overwintering sites was
1607 m (Figure 13, frog 7). This male was captured at an active breeding site (Deep
Creek Meadows) in spring of 2003. Five days later this male was recaptured at Echo
Lake, a distance of 886 m and 240 m elevation above his original capture point. This rapid movement was completed when most of the basin was covered in deep snow.
Many adult R. cascadae completed migrations (> 100m) directly after breeding to other areas for summer foraging. Only 40% of females (n = 32) and 50% of males (n =
127) were captured within 100 m of their spring breeding sites during summer. Of the frogs that migrated, there was no significant difference between adult male and female migration distances (Mann-Whitney U test: Z = 0.48, P = 0.63). Females moved up to
989 m (mean: 279 ± 24 m), while males moved up to 591 m (mean: 241 ± 8 m) to reach summer use areas (Figure 14). During, and directly after, the breeding season, much of
Echo Lake Basin remained saturated from receding snowpack and soils remained damp with most streams still flowing. Based on these conditions, it appeared frogs were not restricted in moving to other patches separated by inhospitable terrain absent of aquatic features during this time. For example, six radio tagged individuals (Figure15, Figure 16,
Figure 17 [frog 13], and Figure 18 [frogs 8 and 9]) demonstrated rapid migrations from breeding sites directly after breeding subsided.
By fall, the majority of adult frogs (73% of females [n = 53] and 68% of males [n
= 120]) were captured within 100m of their summer locations. This suggests that many
55
Figure 13. Examples of Long-term mark-recapture histories of seasonal migrations for eight adult male Rana cascadae in Echo Lake Basin, Trinity Alps Wilderness, California, from 2003 to 2008. All subsequent capture locations of individuals at specific habitat patches were within 25 m of original capture locations. Asterisks represent breeding locations used all years.
56
55 55
50 50 45 Female 45 Male 40 N = 79 40 N = 255 35 35
30 30
25 25
20 20
15 15
10 10 Percentage of Frogs of Percentage 5 5
0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
Migration distance from breeding to summer foraging habitats (m)
Figure 14. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between breeding and summer seasons from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years.
57
Figure 15. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results.
58
Figure 16. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summers of 2003 and 2004 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results.
59
Figure 17. Examples of migration routes and timing between seasonal resources for two telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results.
60
Figure 18. Examples of migration routes and timing between seasonal resources for three telemetered Rana cascadae during the summer of 2003 in Echo Lake Basin, Trinity Alps Wilderness, California. Inset shows the proximity of migration examples in the basin. See Appendix B for more information on individual frog movement results.
61 summer use areas away from breeding sites were suitable and used for overwintering.
However, many areas dried up as the fall approached, causing frogs to move to areas with sufficient water for overwintering. Of the frogs that migrated from summer to winter areas (> 100 m), there was no significant difference in distance moved between males and females (Mann-Whitney U test: Z = 0.93, P = 0.35). Migrating females moved up to
1158m (mean: 317 ± 67 m) while migrating males moved up to 427m (mean: 200 ± 10 m) to reach overwintering areas (Figure 19).
While many adult R. cascadae exhibited strong migration patterns among seasonal resources throughout Echo Lake Basin, many of these individuals repeated the same or similar migration patterns between patches for multiple years. Figures 12 and 13 show many examples of these patterns by both adult male and female frogs. The strongest example of annual seasonal migratory patterns was between Echo Lake and
Snowmelt Pond. Snowmelt Pond dried completely by late summer each year forcing frogs to find permanent water-bodies for overwintering. Thus, Snowmelt Pond might reflect the highest percentage of annual adult migrations of all patches because it was consistently the most productive breeding site in the basin. Echo Lake had the strongest influence on Snowmelt Pond with 75% of all adult Snowmelt Pond captures coming from
Echo Lake. The remaining 25% came from patches further away. During early spring, many adult frogs migrated overland from Echo Lake to Snowmelt Pond prior to breeding
(Figures 17 and 18). Most frogs returned to Echo Lake directly after the breeding season for summer foraging and overwintering. Ninety two out of 165 (56%) individual adults captured at Echo Lake completed at least one round-trip breeding migration between
62
55 55
50 50 45 Female 45 Male 40 N = 73 40 N = 176 35 35
30 30
25 25
20 20
15 15
10 10 Percentage of Frogs of Percentage 5 5
0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
Migration distance from summer foraging to overwintering habitats (m)
Figure 19. Frequency of migration distances traveled by adult female (left column) and adult male (right column) Rana cascadae between summer and winter seasons from 2003 to 2006 in Echo Lake Basin, Trinity Alps Wilderness, California. Each migration represents a movement that occurred between the two successive seasons. Density line represents approximate smoothed frequency distribution. Individual frogs may have been counted more than once between different years.
63 these patches from 2003 to 2008. Of the frogs that migrated between these two areas, at least 51% completed annual migrations twice, 26% three times, 14% four times. At least three (3.3%) completed the annual migration at least five times (Figure 13, frog 4).
Lastly, at least four adult migrating frogs completed this migration in 2003 and again in
2008 spanning six breeding seasons.
Although most experienced breeders exhibited consistent breeding migration patterns between specific patches (Table 7), variation in annual migration strategies for frogs with long capture histories occasionally occurred. Based on capture histories of 34 gravid adult females in amplexus, one was found in amplexus at Snowmelt Pond in 2003 and then again in amplexus at Echo Lake in 2004. Five male frogs were also observed actively engaged in breeding activities at Echo Lake during an entire year while they had completed breeding migrations to Snowmelt Pond in prior and subsequent years.
Alternatively, another adult male frog (Figure 13, frog 5) exhibited a strict migration pattern between Deep Creek Meadows and Lower Van Matre Meadows for three straight years. This male then dispersed to Cascade Meadows for the following breeding season, located 570 m away from his previous breeding site. Finally, a few frogs showed a nomadic distribution, with no apparent migration pattern among years. For example, an adult female (Figure 12, frog 6) moved among five different patches over three years, all of which contained active breeding sites. A male frog (Figure 13, frog 3) also displayed nomadic tendencies, visiting three different patches with active breeding sites over four years. These irregular movements are evidence that some adult frogs choose to breed at different sites in different years.
64 Movement Rates and Routes
Most daily movements were short and localized around proximate aquatic features (Figures 15, 16, 17, 18). Daily movement rates of adult male (n = 35) and female frogs (n = 51) did not differ significantly ANOVA (F = 0.31, P = 0.74). However, movement rates of the sexes combined differed by season (F = 4.07, P = 0.02), with daily rates significantly higher in spring (mean: 9.9 ± 1.3 m) than in fall (mean; 4 ± 1.8 m).
Summer movement rates of adults (mean: 6.9 ± 1.3 m) were not significantly different from spring or fall (Tukey Kramer multiple comparisons test).
Most radio tracked individuals completed at least one long-distance migration between two patches (Figures 15, 16, 17, 18). Movement rates were calculated for eight radio tagged frogs which exhibited 10 seasonal migrations from 114 to 334 m among patches. Most migrations were rapid (3.9 to 15.3 m/hour within 24 hours). The fastest overall movement rate observed was 30.3 m/ hour within a 1.75 hour period. However, two frogs (Figures 17 and 18, frogs 8 and 13) took up to five days to move overland between patches during spring. Two of 10 migrations were observed between 20:00 and
10:00, suggesting these frogs conducted their migrations at night. Furthermore, both of these migrations occurred overland during August, showing R. cascadae will make rapid movements over land during dry periods. The fastest long-range movement rate (> 500 m) detected was an adult male that traveled from Deep Creek Meadows to Echo Lake, a minimum distance of 839 m, with a 238 m elevation gain, in less than 4.6 days (Figure
13, frog 7). This dispersal event was determined from a mark-recapture basin census, so the actual movement rate was likely underestimated.
65 Examples of movement routes used by radio tagged R. cascadae are displayed in
Figures 15, 16, 17, and 18). Overall, frogs appeared to choose the shortest and most direct routes while moving within and between patches in Echo Lake Basin. Throughout the R. cascadae active period, many patches were hydrologically connected through the basins expansive stream network. Many radio tagged frogs used these streams as corridors between patches (e.g., Figure 13, frog 6). However, overland movements of radioed frogs were also common, especially when sites were completely isolated or as stream corridors dried.
Overland routes sometimes occurred over difficult terrain with steep inclines.
Two radio tagged frogs were found navigating through steep, to near vertical, terrestrial inclines during an August rain in 2003. One individual (Figure 15, frog 6), climbed 20 m up a stream bank (70º slope) composed of bare soil and embedded boulders. This frog remained near the top of this bank within a rodent burrow for three days prior to returning to the stream directly below. Another individual (Figure 18, frog 9), navigated 220 m overland which included a descent down a steep (> 70º slope) couloir in a rocky outcrop south of Echo Lake. This frog was found on 2 August perched midway up a near vertical cliff section approximately four meters high, and subsequently navigated down the following day.
Overland movements usually shortened the distance and elevation gradients traveled between patches considerably. For example, the stream distance between Echo
Lake and Snowmelt Pond is 760 m, whereas the overland distance is ~250 m.
Furthermore, if individuals were to use the stream route to move between these sites it
66 would require a steep decent of 112 m followed by a steep 125 m climb. In the early spring, the outlets of both sites, as well as the rocky knoll separating them, are completely covered in snow up to 3.5 m deep. Through telemetry, it was determined that R. cascadae migrated prior to breeding between these patches in deep snowpack along a narrow snowmelt crack between a rock outcrop and snowfield (< 1.5 m wide) (Figure 18, frog 8).
The extent to which adult male and female frogs migrated overland between Echo Lake and Snowmelt Pond is evident in Figure 10 where at least 180 unique round-trip migrations between these sites occurred from 2003 to 2008.
Interbasin Dispersal
Recaptures of marked R. cascadae in the six basins showed that dispersal
occurred between basins. Nineteen (1%) of the 1955 marked R. cascadae from 2003 to
2007 dispersed over three steep headwall ridges to colonize new habitats in four out of
six neighboring basins from 2003 to 2008 (Table 8). Travel routes likely followed low
points in saddles associated with mountain passes and avoided jagged ridge sections and
peaks (Figure 20). Average incline for headwall slopes below the three ridges equaled
33° (range: 29° - 62°). The minimum distances traveled by interbasin dispersing frogs averaged 1177 m (range: 736 m - 1886 m). The maximum vertical gain by a frog that left lower Siligo Basin to relocate to Echo Lake was 308 m, the second highest patch in Echo
Lake Basin. Land features separating basins included steep and rocky talus terrain
lacking permanent exposed aquatic features. At least 16 individuals moved distances
greater than 500 m over land (Figure 20). Of the four frogs captured before and after
67 Table 8. Interbasin dispersal propensity by Rana cascadae among six basins in the southeast Trinity Alps Wilderness, California, from 2003 to 2008.
Basin of origin Number Total Dispersing Dispersal Individuals of marked individuals1 Probability2 Received3 surveys (2003-2007) (%) Deer Creek 22 18 1 5.6 0 Siligo Meadows 16 334 8 2.4 4 Echo Lake 56 1528 10 0.7 9 Little Deep Creek 5 59 0 0.0 0 Billy-Be-Dammed 4 4 0 0.0 0 Stony Creek 44 12 0 0.0 6 Total: 147 1955 19 1.0 19
1 Number of marked individuals that dispersed to different basins from their origin. 2Dispersing individuals divided by the sum of uniquely marked individuals from the basin of origin. This value represents the minimum observed propensity for Rana cascadae to colonize new basins. 3Number of marked individuals a basin received from another basin through dispersal.
68
2300 Deer Creek Pass Little Stonewall Pass Stonewall Pass (m)
2150 Deer Creek Upper Siligo Echo Lake Stony
Elevation Basin Basin Basin Creek Basin 2000 5.2 Km 1
1 Echo 5 2 Lake 1 2 3 1 1 1 1
1* Deer Lake
R. Cascadae Breeding Site Luella Lake Basin Ridge Boundary
Figure 20. Connecting lines of nineteen individual Rana cascadae that dispersed over ridges separating four neighboring watersheds in the southeast Trinity Alps Wilderness, California. Numbers associated with white lines represent the number of individuals that moved in a particular pattern; arrows indicate movement direction. The overall air distance linking habitat patches across the landscape through dispersal was 5.2 km, indicating high potential gene flow across this landscape. Top inset depicts a cross section of the lowest elevations along ridges at passes (yellow arrows) that dispersing frogs likely used as corridors. Two basins where interbasin movements were not observed are not indicated in figure.
*frog did not disperse over a pass or between two basins but between two isolated sites in a single basin; this movement adds to the overall landscape connectivity via R. cascadae movements.
69 dispersing in the same summer, movements occurred between 03 July and 14 September with the shortest travel interval observed within 19 days (03 July and 22 July).
Eight individuals dispersed as juveniles prior to their first breeding season, while three male frogs dispersed as adults. The remaining eight interbasin dispersing frogs grew from juveniles to adults between captures, so age at dispersal could not be determined.
Sex ratios of interbasin dispersing frogs (n = 19) were equal (9 females, 9 males [one undetermined]) suggesting unbiased interbasin dispersal by gender. Final destinations for all 19 dispersing frogs were within 50 m of consistently utilized R. cascadae breeding sites (Figure 20). Additionally, six male frogs (32%) that dispersed to new basins were captured at breeding sites as adults while engaging in breeding activity. Of the eight frogs captured over multiple years after dispersing, all remained in the same patch up to four years later. Although it could not be determined if all dispersals resulted in successful reproduction, potential gene flow from these animals through interbasin connectivity spanned an air distance of 5.2 km (Figure 20).
DISCUSSION
The complex life cycles (Wilbur 1980) and dynamic movement ecologies
(Semlitsch 2008) of many amphibians are challenging to understand. However, dissecting
the nature of these processes is vital to understanding the appropriate scales at which they
operate if we are to effectively conserve these organisms (Morales and Ellner 2002,
Petranka et al. 2004). My study is the first on the ecology of R. cascadae that included
the entire array of annual activity patterns of a population over multiple years across an
entire catchment and its adjacent watersheds. Based on their annual life cycle (Appendix
A) in an environment with extreme annual environmental fluctuations, R. cascadae have a diverse range of strategies at different temporal and spatial scales to address life history requirements.
Spatial Dynamics
The spatial arrangement of resources across a landscape can have profound
effects on species distributions (Dunning et al. 1992, Rickets 2001). The volume of
intersection analysis in Echo Lake Basin revealed relatively low overlap in space use
among both seasonal (mean volume of intersection: 0.36 - 0.62) and age-specific (mean
volume of intersection: 0.32 - 0.52) populations of R. cascadae. In addition, comparisons
of age class and seasonal isopleths showed population range sizes varied greatly between
seasons and age groups. The overall low overlap and varying space use among population
sub-groups demonstrates that the spatio-temporal distribution of this population is
70 71 complex and changes with demographic status and resource needs among seasonally available resources.
During summer, when all age groups could be compared, juvenile and adult R. cascadae populations had similar range sizes, but only showed moderate utilization distribution space use overlap (mean volume of intersection: 0.52). During summer overall population range sizes were largest for both adults and juveniles, indicating these groups were widespread throughout the basin during the time when aquatic resources were most abundant. In contrast, during summer young of previous year used approximately half of the total area used by juveniles or adults with captures strongly associated with breeding areas. The lowest observed summer utilization distribution overlap occurred between young of previous year and adult frogs (mean volume of intersection: 0.32). Juveniles shared slightly more space with young of previous year than adults.
Unlike many other ranid frogs, R. cascadae metamorphose just prior to the onset of winter, restricting growth and the timing to move to other resources. After surviving their first winter, young of previous year may have limited abilities to move because of their small size, and metabolic costs associated with water balance (Sinsch 1990). This may explain why young of previous year largely remained close to breeding areas during summer. During the breeding and winter seasons, adults used approximately one quarter and one half of the space used during the summer, respectively. High adult densities during the breeding season can be attributed to specific courtship behaviors. However, moderate densities at overwintering sites may be because temporary hydrological features
72 disappear by the fall. The widespread distribution of juveniles and adults during summer could be the result of expanded seasonal resource availability and secondarily may be influenced by predator avoidance. The main predators of R. cascadae in the region are two species of garter snake (Thamnophis sirtalis, T. atratus) (Garwood and Welsh 2005,
Pope et al. 2008). Breeding sites consistently had the highest number of snake observations relative to other areas in the basin (Garwood and Welsh 2007).
Changes in space use by adults and juveniles between summer and winter seasons were similar. There was a general range contraction to approximately half of the area used during the summer by both as they congregated near overwintering habitats. The overlap in winter space use was moderate for adults and juveniles (mean volume of intersection: 0.49), contrary to what was expected due to the limited available winter resources. Adults, especially males, tended to overwinter near breeding sites. Because juveniles are not reproductive, they may have been less influenced to use these sites for overwintering and select from a wider variety of locations.
Resource Use
Rana cascadae appear to be well adapted to a wide variety of habitats and were
found using all available aquatic resources in Echo Lake Basin. However, these resources
were used in different proportions by different age groups and genders. For example,
both juvenile and adult female R. cascadae used streams nearly twice as often as young
of previous year and adult males. Egg masses and young of previous year were found
predominantly in small shallow ponds. Young of previous year were consistently found
73 associated with or near lentic breeding areas, apparently having high site fidelity to their natal sites during the second year. Unlike other western ranid frogs living in lower elevations, R. cascadae metamorphose just prior to winter, restricting their ability to disperse before winter. I did not observe metamorphosed individuals moving between sites in the fall, however, some breeding sites dried prior to the winter suggesting that some individuals may have moved to perennial overwintering sites, or perished.
Adult R. cascadae were most narrowly distributed during breeding seasons in
Echo Lake Basin, with aggregations of adult male R. cascadae calling at 12 separate breeding complexes annually. Adult female frogs remained cryptic during this time, however their egg masses were readily detected at breeding sites. Breeding complexes were relatively isolated from one another, indicating that each breeding complex supported its own sub-population. R. cascadae have been identified as explosive breeders, with the breeding season occurring within three to fourteen days at specific locations in Oregon (Sype 1975, Briggs 1987). The breeding season lasted no more than seven days at individual breeding sites for the duration of this study. Based on their explosive reproductive behavior, most adult R. cascadae only remained at local breeding sites for a few days before relocating to summer foraging areas once breeding subsided.
Echo Lake Basin contained 588 lentic waterbodies, of which 40 were used for reproduction. Small shallow spring-fed ponds were found to be important habitats for reproduction. These accounted for 72% of breeding sites. During early spring, these spring-fed ponds were commonly available before those with surface water only, with adjacent snow and ice melting earlier than at other aquatic features. Spring-fed ponds
74 were also more likely to contain water late in the season relative to surface water ponds and thereby may have the effect of increasing survival to metamorphosis. However, not all frogs used spring-fed ponds, and 20 % of the egg masses produced from 2003 to 2008 were placed in Snowmelt Pond, a large temporary pond that dried every year causing extensive pre-metamorphosis mortalities during each fall.
During summer, R. cascadae took advantage of all available aquatic resources in
Echo Lake Basin. Most notably, juvenile and adult age groups were found using aquatic resources in ephemeral zones often during summer. These areas may provide different or more food sources associated with the dryer upland vegetation. For example, grass hoppers, a common prey item of R. cascadae (Larson, M. D. 2009, personal communication) appeared to have higher densities in areas containing less standing water. Specific information on the diet of R. cascadae, and the availability of food resources, may help explain which summer aquatic resources are the most productive for foraging.
Based on late fall and early spring captures of individual R. cascadae, a variety of perennial aquatic resources are used for overwintering in the Echo Lake Basin. During fall of 2003, patch water surface area was reduced to 32% of average spring values. This resulted in a large reduction in aquatic resources available for overwintering. Frogs were generally found most commonly in lentic habitats during fall. However, almost one third of all captures were found using streams. These results differ from studies on other high- elevation ranid frogs, including the Sierra yellow-legged frog (R. sierra) and the
Columbia spotted frog (R. Luteiventris). These ranid frogs were found using lentic
75 habitats exclusively for overwintering, particularly in lakes exceeding 2 meters in maximum depth (Pope and Matthews 1999, Pilliod et al. 2002). However, Bradford
(1983) found R. sierra frogs and tadpoles survived overwintering in a stream section that was 1.8 meters deep. In my study, Echo Lake was the only patch that exceeded two meters in depth and produced more R. cascadae captures (26%) during the late fall than any other patch. Notably, ten out of the 12 remaining overwintering patches contained features less than 1 meter deep, suggesting specific hibernacula may not necessarily be dependant upon water depth.
Both lotic and lentic aquatic habitats, derived from groundwater springs produced
53 % of all fall frog captures, indicating these springs likely have an important influence on suitability of overwintering habitat. A study in Oregon excavated inactive overwintering R. cascadae from muddy substrates in a spring-fed section of a pond
(Briggs 1987). This suggests springs were important features in this area. I often captured inactive R. cascadae in shallow loose muddy substrates of spring ponds just prior to winter. Burrowing into mud may insulate frogs from colder ambient temperatures and ice. For example, water temperature data loggers buried in two spring ponds during an entire winter indicate physical conditions remained well above freezing and relatively stable (Appendix C). However, average water temperatures were highest and most stable
(mean temperature: 5.19 °C ± 0.003) at a data logger located 45 cm below a pond surface and under 15 cm of loose mud substrate compared with another logger located in a pond twice as deep in open water, 10 cm above the mud substrate (mean temperature: 4.83 °C
± 0.01). Lamoureux and Madison (1999) found green frogs (Rana clamitans)
76 predominantly used unfrozen groundwater springs and streams less than 1 meter deep for overwintering habitats, and they suggested dissolved oxygen levels were more stable than in waters lacking flowing water. Oxygen depletion during the winter in iced over waterbodies of the Sierra Nevada mountains has been known to kill R. muscosa
(Bradford 1983) in waters determined to not be influenced by springs.
Movements by Immature Frogs
Movements were distinctly different among young of previous year and juvenile
R. cascadae indicating that different activity patterns exist among immature life stages.
For example, young of previous year rarely moved between patches and were found most often in or near breeding sites that may have been where they originated. Juveniles were widely distributed throughout Echo Lake Basin and often moved between different habitats. The two groups of immature R. cascadae also used habitats in much different proportions. This may help explain their differences in movement. In general, juvenile R. cascadae had low site fidelity, that decreased over time. This low site fidelity was also evident in the analysis of breeding success in these juvenile frogs as they matured. I could find no evidence of a strong seasonal movement pattern for juveniles as was observed with adults. It appears that juveniles have a nomadic movement pattern, as a result of
their lack of requiring breeding habitats. Juvenile amphibians typically disperse
(unidirectional movement) and round trip migrations are largely considered to be an adult
behavior (Semiltsch 2008). Notwithstanding, I found that some juvenile R. cascadae
appeared to express a seasonal migration which was likely a response to summer water
77 drying up. Thus, juveniles were required to move to perennial sites where they tended to overwinter. Overall, the limited migratory strategy of juvenile R. cascadae in Echo Lake
Basin appeared to be driven by hydroperiod.
Adult Seasonal Migrations
This study demonstrated individual adult R. cascadae will often conduct seasonal
migrations between separate breeding, summer foraging, and overwintering resources.
These results are consistent with other studies on high-elevation ranid frogs (Pope and
Matthews 2001, Pilliod et al. 2002) that revealed similar seasonal migratory patterns.
These migration patterns appear to highlight two ecological processes described as landscape complementation and landscape supplementation (Dunning et al. 1992).
Landscape complementation occurs where separate, non-substitutable resources are spatially and or temporally isolated, thus requiring some animals to migrate seasonally.
Landscape supplementation occurs when animals move among patches with substitutable resources. That is, each patch offers all resources to survive, yet animals still move between patches. I suggest both ecological processes are operating with R. cascadae, based on locations of seasonal resources (i.e. breeding, summer foraging, overwintering), and the dynamic changing hydrological conditions in this study region. Some sites provided year round habitats, and many frogs did not migrate. However, on average 43% of adult frogs conducted seasonal migrations between ephemeral and perennial resources.
Migration to supplemental and complementary resources may be a behavioral response to
78 ecological factors potentially affecting survival such as, resource quality, decreased conspecific competition, predation, or decreased exposure to diseases.
The strongest migration patterns were those to and away from breeding sites demonstrating the significant influence breeding locations have on the seasonal distributions of R. cascadae. It is not clear why adult females migrated farther and more often than adult males between overwintering locations and breeding sites. This pattern was also observed by Pilliod et al. (2002) on a movement study on Columbia spotted frogs. The authors indicated females moved to areas with higher quality food resources and hypothesized females could move further than males because of their larger body size. Similarly, my study found females to be much larger in size than males and thus may require specific foraging areas with abundant food. However, a detailed diet study is needed to test this hypothesis. It is unlikely male R. cascadae are less able than females to move between patches in Echo Lake Basin. I observed many extensive movements in my study. Male R. cascadae may more often choose to remain closer to breeding sites in order to maximize the probability of finding a mate during the short breeding season
(Wells 1977).
Dispersal
A few empirical studies exist that quantify age-specific dispersal rates for aquatic
breeding amphibian populations living among multiple isolated patches; these studies
also had robust sample sizes of marked individuals (> 5000 marked individuals). Breden
(1987) found 27% of first time breeding Fowler’s toads (Anaxyrus fowleri) dispersed
79 away from natal ponds. Berven and Grudzien (1990) found 18 % of first time breeding wood frogs (Rana sylvatica) dispersed to ponds other than their origin whereas experienced breeders had perfect site fidelity. Gill (1978) found experienced breeding
Red-spotted newts (Notophthalmus viridescens) also showed perfect site fidelity. A study by Gamble et al. (2007) found 9% of first time breeding marbled salamanders
(Ambystoma opacum) dispersed to new ponds followed by only 3.2% for experienced breeders.
I found similar dispersal trends for R. cascadae, with first time breeders dispersing far more often (51%) than experienced breeders (7%). However, R. cascadae had exceptionally higher average First time breeder dispersal rates than reported in other species. Funk et al. (2005a) used multistrata models to estimate dispersal rates of marked
Columbia spotted frogs among separate patches in two basins. They found juvenile dispersal varied from one to 62% whereas annual adult dispersal probabilities approximated zero. Although the study did not determine successful dispersal through confirmed reproduction, it demonstrates dispersal rates can be high in other ranid frogs.
The high breeding dispersal rate of first time breeders in Echo Lake Basin demonstrates that landscapes containing patchy resources can lack strong demographic independence. However, this may depend on the spatial structure of the patches (Dunning et al. 1992, DeWoody et al. 2005). For example, most patches in Echo Lake Basin were connected through dispersal, yet two small and isolated patches (Penthouse Ponds and
Eden Pond) produced no apparent dispersers over the length of the study. Both patch size and isolation are important indicators to overall landscape connectivity in terrestrial
80 ecosystems (Moilanen and Nieminen 2002). This may explain the observed lack of connectivity of these patches relative to other patches in the basin.
Interbasin Dispersal
The effect a landscape has on movement capabilities of amphibians is complex
and not well understood. The term “long-distance dispersal” identifies dispersal by
organisms beyond a specific threshold that has ecological and evolutionary consequences which differ from those for dispersal below this threshold (Nathan et al. 2003). Crossing
or not crossing a boundary between suitable habitats is the first behavioral component leading to dispersal (Baguette and Van Dyck 2007). Permeability of the landscape between suitable habitats can be dependent on physical and biological constraints for a particular species (Mazerolle and Desrochers 2005). This can lead an overall resistance on dispersal behavior by the landscape (Forman 1995). In my study, within-basin dispersal estimates averaged 51%, whereas between-basin observed dispersal was far less
(1%). Thus, two dispersal patterns exist for R. cascadae. One could view these patterns as a product of scale, with ridgelines acting as dispersal boundaries for most individuals.
Those individuals that dispersed between local patches within Echo Lake Basin exhibited a local population dispersal pattern, whereas interbasin dispersers are more likely part of a metapopulation (Smith and Green 2005). Recently, landscape-scale patterns of dispersal and gene flow have been closely linked to movement behavior in other aquatic breeding amphibians (Funk et al. 2005a, Lowe et al. 2008) indicating dispersal information is valuable for predicting patterns of gene flow. However, a robust investigation into
81 regional gene flow of R. cascadae at fine scales would be required to indicate if interbasin movements occur between separate populations (Funk et al. 2005b, Zamudio and Weiczorek 2007).
Studies documenting movements of aquatic breeding amphibians over complex landscapes are rare. Twitty (1966) studied homing abilities of Red-bellied newts (Taricha rivularis). He found 81% of relocated individuals returned to their stream of origin after being displaced between drainages over 3 km away. These drainages were divided by a steep mountain ridge exceeding 300 m elevation. Individual newts were often intercepted in route at ridge top pitfall traps. The lack of empirical studies documenting ridge dispersal of aquatic breeding amphibians may be an artifact of study design. Smith and
Green (2005) reviewed amphibian movements and found most studies covered areas too small to reveal unbiased population-level movement patterns. For example, many temperate amphibian mark-recapture studies are confined to single basins based on assumed barriers to dispersal (e.g. mountain ridges) and may underestimate the dispersal potential of an organism.
Semlitch (2008) argued that many previous studies and review papers on amphibian movements have failed to adequately differentiate between dispersal and migration processes, thus making results difficult to discern. Since dispersal is largely considered a juvenile life history behavior (Lemckert 2004, Semlitch 2008), long-range movements could be missed if juvenile age groups are not studied. For example, popular anuran marking techniques, such as radio transmitters or PIT tags, require a minimum size at marking, thus eliminating juveniles from many studies (Richards et al. 1994,
82 Ferner 2007). Additionally, marking the juvenile component in a population can be costly and time intensive if factors such as high species fecundity, low capture probability, or large population sizes exist.
To my knowledge, my study provides the first empirical evidence of successful dispersal by an aquatic breeding frog over steep mountain ridges. Ridge movements by R. cascadae demonstrate the potential for population connectivity between multiple basins can be high, even when the landscape is inhospitable by lacking aquatic features and vegetative cover over long distances. By surveying six basins, I found that dispersal among basins was not a rare event, with 19 individuals (1%) moving between four basins within five years. These results suggest that gene flow among adjacent basins is sufficient to maintain genetic diversity between sub-sets of a larger metapopulation (Mills and
Allendorf 1996) by having greater than one individual per generation dispersing over mountain passes. Based on the minimal survey effort in proximal basins surrounding
Echo Lake Basin, this appears to be a minimum estimate for potential gene flow. For example, the only proximal site surveyed regularly was Red Mountain Meadows where six out of 15 individual frogs were originally captured in Echo Lake Basin.
Recent molecular studies on amphibians have tested the effect ridges have on genetic variation. Specifically, gene flow was assessed between regional populations occurring at both low and high elevations (Funk et al. 2005b, Giordano et al. 2007). Gene flow was found to be lowest among high elevation populations in both studies, suggesting dispersal across mountain ridges was limited to absent. Funk et al. (2005a) found juvenile
Columbia spotted frogs had many juveniles dispersing from low to high elevation
83 populations. Funk et al. (2005b) developed a conceptual model to explain differences in gene flow based on altitudinal and topographic features (Figure 20A). This “mountain- valley” model was also supported by Giordano et al. (2007) who found the same landscape genetic patterns with Long-toed salamanders (Ambystoma macrodactylum).
Current California populations of R. cascadae are restricted to elevations exceeding 1200 meters (Welsh et al. 2006, Fellers et al. 2008) causing the species to have a highly fragmented “sky island” distribution. For example, populations in the Lassen region have undergone a major range contraction. Only two out of the six remnant populations are separated by less than 12 km (Fellers et al. 2008). Consequently, the likelihood of connectivity via dispersal among most remaining populations in the region is extremely low. In addition, the Klamath mountains contain several low-elevation river canyons that may limit gene flow. I propose an alternative gene flow model for R. cascadae (Figure 20B). This hypothesized “mountain-island” model suggests that gene flow occurs often among high-elevation populations that are not separated by deep low- elevation canyons, below the species lower elevational limits. This hypothesis is also supported by the findings of Monsen and Blouin (2004) who found a sharp drop in gene flow among R. cascadae populations in Oregon and Washington separated by distances greater than 10 km. However, future studies using molecular techniques could assess the degree to which low-elevation canyons influence gene flow.
84
Figure 21. Examples of two population structure models for amphibians in mountainous regions. (A) ‘Valley-Mountain’ model for Columbia spotted frogs (Rana luteiventris) adapted from Funk et al. (2005b) with three gene flow scenarios: (i) Low-elevation populations with high gene flow, (ii) high-elevation populations with little to no gene flow over ridges, (iii) restricted gene flow between low and high-elevation populations. (B) Proposed ‘Mountain-Island’ model for Cascades frogs (R. cascadae) in the Klamath mountains, California. Two gene flow scenarios: (i) moderate gene flow across ridges between small high-elevation populations, (ii) little to no gene flow between populations separated by major drainages dipping below 1200 meters elevation.
REFERENCES
Baguette, M. and H. Van Dyck. 2007. Landscape connectivity and animal behavior: functional grain as a key determinant for dispersal. Landscape Ecology 22: 1117- 1129.
Bailey, L. L. 2004. Evaluating elastomer marking and photo identification methods for terrestrial salamanders: marking effects and observer bias. Herpetological Review 35: 38-41.
Berven, K. E. and T. A. Grudzien. 1990. Dispersal in the wood frog (Rana sylvatica): implications for genetic population structure. Evolution 44: 2047-2056.
Beyer, H. L. 2004. Hawth's analysis tools for ArcGIS. Available at: http://www.spatialecology.com/htools. (Contacted June 2009)
Bourque, R.M. 2007. An alternative method for restraining frogs. Herpetological Review 38: 48-49.
Bradford, D. F. 1983. Winterkill, oxygen relations, and energy metabolism of a submerged dormant amphibian, Rana muscosa. Ecology 64: 1171-1183.
Breden, F. 1987. The effect of post-metamorphic dispersal on the population genetic structure of Fowler's toad, Bufo woodhousei fowleri. Copeia 1987: 386-395.
Briggs, J. L. 1976. Breeding biology of the Cascades frog, Rana cascadae. Herpetological Review 12: 75.
Briggs, J. L. 1978. An asymptotic growth model allowing seasonal variation in growth rates, with application to a population of the Cascades frog Rana cascadae (Amphibia, Anura, Ranidae). Journal of Herpetology 12: 559-564.
Briggs, J. L. 1987. Breeding biology of the Cascades frog, Rana cascadae, with comparisons to R. aurora and R. pretiosa. Copeia 1987: 241-245.
Brown, C. 1997. Habitat structure and occupancy patterns of the montane frog, Rana cascadae, in the Cascade Range, Oregon, at multiple scales: Implications for population dynamics in patchy landscapes. Masters Thesis. Department of Forest Science, Oregon State University, Corvallis, Oregon.
85 86 Bury, R. B., and D. J. Major. 1997. Integrated sampling for amphibian communities in montane habitats. In: Olson D. H., W. P. Leonard, and R. B. Bury (editors). Sampling amphibians in lentic habitats. p 75-82. Olympia (WA): Northwest Fauna 4.
California Department of Water Resources. 2007. Historical course data. California Resources Agency, Division of Flood Management, California Cooperative Snow Surveys. Available at: http://cdec.water.ca.gov/snow/. 1416 9th Street, Sacramento, California, 95814. Contacted on May 10, 2008.
California Department of Water Resources. 2009. California Data Exchange Center, Snow course measurements. Available at http://cdec.water.ca.gov/cgi- progs/rpts1/COURSES). 1416 9th Street, Sacramento, California, 95814. Contacted on June 1, 2009.
Collins, J. P. and A. Storfer. 2003. Global amphibian declines: Sorting the hypotheses. Diversity and Distributions 9: 89-98.
Crisafulli, C. M., L. S. Trippe, C. P. Hawkins, and J. A. MacMahon. 2005. Amphibian responses to the 1980 eruption of Mount St. Helens. Pages 193-198 in V. H. Dale, F. J. Swanson, and C. M. Crisafulli, editors. Ecological responses to the 1980 eruption of Mount St. Helens. Springer-Verlag, New York, New York.
Crouch, W. B. and W. C. Paton. 2000. Using egg-mass counts to monitor wood frog populations. Wildlife Society Bulletin 28: 895-901.
Crump, M. L. and N. J. J. Scott. 1994. Visual encounter surveys. Pages 84-91 in W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L.-A. C. Hayek, and M. S. Foster, editors. Measuring and monitoring biological diversity: Standard methods for amphibians. Smithsonian Institution Press, Washington D.C.
Davidson, C., B. H. Shaffer, and M. R. Jennings. 2002. Spatial tests of the pesticide drift, habitat destruction, UV-B, and climate-change hypotheses for California amphibian declines. Conservation Biology 16: 1588-1601.
DeWoody, Y. D., Z. Feng, and R. K. Swihart. 2005. Merging spatial ant temporal structure within a metapopulation model. The American Naturalist 166: 42-55.
Dunning, J.B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in complex landscapes. Oikos, 65: 169-175.
Environmental Systems Research Institute Inc. 2008. ArcGIS v. 9.3. Redlands, California.
87 Fellers, G. M. and C. A. Drost. 1993. Disappearance of the Cascades frog, Rana cascadae at the southern end of its range, California, USA. Biological Conservation 65: 177-181.
Fellers G.M., K.L. Pope, J.E. Stead, M.S. Koo, and H.H. Welsh, Jr. 2008. Turning population trend monitoring into active conservation: can we save the Cascades frog (Rana cascadae) in the Lassen region of California? Herpetological Conservation and Biology 3: 28-39.
Ferlatte, W. J. 1974. A flora of the Trinity Alps of northern California. University of California Press, Berkeley and Los Angeles, California.
Ferner, J. W. 2007. A review of marking and individual recognition techniques for amphibians and reptiles. Herpetological Circular No. 35, Society for the Study of Amphibians and Reptiles, Salt Lake City, Utah.
Forman, R. T. T. 1995. Some general principles of landscape and regional ecology. Landscape Ecology 10: 133-142.
Funk, W. C., M. S. Blouin, P. S. Corn, B. A. Maxell, D. S. Pilliod, S. Amish, and F. W. Allendorf. 2005b. Population structure of Columbia spotted frogs (Rana luteiventris) is strongly affected by the landscape. Molecular Ecology 14: 483- 496.
Funk, W. C., A. E. Green, P. S. Corn, and F. W. Allendorf. 2005a. High dispersal in a frog species suggests that it is vulnerable to habitat fragmentation. Biology Letters 1: 1-4.
Gamble, L. R., K. McGarigal, and B. W. Compton. 2007. Fidelity and dispersal in the pond-breeding amphibian, Ambystoma opacum: implications for spatio-temporal population dynamics and conservation. Biological Conservation 139: 247-257.
Garcia T. S., Romansic, J.M., Blaustein A. R. 2006. Survival of three species of anuran metamorphs exposed to UV-B radiation and the pathogenic fungus Batrachochytrium dendrobatidis. Diseases of Aquatic Organisms 72: 163-169.
Garwood, J. M. and H. H. Welsh Jr. 2005. Natural History Notes: Rana cascadae (Cascades frog). Predation. Herpetological Review 36: 165
Garwood, J. M. and H. H. Welsh. 2007. Ecology of the Cascades frog (Rana cascadae) and interactions with garter snakes and nonnative trout in the Trinity Alps Wilderness, California. Final report for the California Department of Fish and Game and the National Fish and Wildlife Foundation, Arcata, California. 87 pp.
88 Gill, D. E. 1978. The metapopulation ecology of the Red-spotted newt, Notophthalmus viridescens (Rafinesque). Ecological Monographs 48: 145-166.
Giordano, A. R., B. J. Ridenhour, and A. Storfer. 2007. The influence of altitude and topography on genetic structure in the long-toad salamander (Ambystoma macrodactylum). Molecular Ecology 16: 1625-1637.
Gitzen, R. A., J. J. Millspaugh, B. J. Kernohan. 2006. Bandwidth selection for fixed- kernel analysis of animal utilization distributions. Journal of Wildlife Management 70: 1334-1344.
Hammerson, G. and C. Pearl. 2004. Rana cascadae. IUCN Red List of Threatened Species. Available at www.iucnredlist.org. Rue Mauverney 28, 1196, Gland, Switzerland. Contacted on 22 August 2009.
Hintze, J. 2004. NCSS and PASS. Number Cruncher Statistical Systems. Kaysville, Utah.
Jennings, M. R. and M. P. Hayes. 1994. Amphibian and reptile species of special concern in California. California Department of Fish and Game, Sacramento, California.
Kernohan, B. J., R. A. Gitzen, and J. J. Millspaugh. 2001. Analysis of animal space use and movements. Pages 125-166 in J. J. Millspaugh and J. M. Marzluff, editors. Radio tracking and animal populations. Academic Press Inc., San Diego, California.
Lamoureux, V. S. and D. M. Madison. 1999. Overwintering habitats of radio-implanted green frogs, Rana clamitans. Journal of Herpetology 33: 430-435.
Lannoo, M. 2005. Amphibian declines: The conservation status of United States species. University of California Press, Berkeley, California.
Larson, M. D. 2009. Personal communication. United States Department of Agriculture, Forest Service, Pacific Southwest Research Station, Redwood Science Laboratory, 1700 Bayview, Arcata, CA 95521.
Lemckert, F. L. 2004. Variations in anuran movements and habitat use: implications for conservation. Applied Herpetology 1:165-181.
Lipman, P. W. 1964. Structure and origin of an ultramafic pluton in the Klamath Mountains, California. American Journal of Science 262: 199-222.
89 Lowe, W. H., M. A. McPeek, G. E. Likens, and B. J. Cosentino. 2008. Linking movement behavior to dispersal and divergence in plethodontid salamanders. Molecular Ecology 17: 4459-4469.
Mazerolle, M. J. and A. Desrochers. 2005. Landscape resistance to frog movements. Canadian Journal of Zoology 83: 455-464.
Mills, L. S. and F. W. Allendorf. 1996. The One-migrant-per-generation rule in conservation and management. Conservation Biology 10: 1509-1518.
Millspaugh, J. J., G. C. Brundige, R. A. Gitzen, and K. J. Raedeke. 2000. Elk and hunter space use sharing in South Dakota. Journal of Wildlife Management 64: 994- 1003.
Moilanen, A. and M. Nieminen. 2002. Simple connectivity measures in spatial ecology. Ecology 83: 1131-1145.
Monnet, J. M. and M. I. Cherry. 2002. Sexual size dimorphism in anurans. Proceedings of the Royal Society of London, Series B: 269: 2301-2307.
Monsen, K. J. and M. J. Blouin. 2003. Genetic structure in a montane ranid frog: restricted gene flow and nuclear-mitochondrial discordance. Molecular Ecology: 3275-3286.
Monsen, K. J. and M. J. Blouin. 2004. Extreme isolation by distance in a montane frog Rana cascadae. Conservation Genetics 5: 827-835.
Morales, J. M. and S. P. Ellner. 2002. Scaling up movements in heterogeneous landscapes: the importance of behavior. Ecology 83: 2240-2247.
Muths, E. 2003. A radio transmitter belt for small ranid frogs. Herpetological Review 34:345-348.
Nathan, R., G. Perry, J. T. Cronin, A. E. Strand, and M. L. Cain. 2003. Methods for estimating long-distance dispersal. Oikos 103: 261-273.
Nauwelaerts, S., J. Coeck, and P. Aerts. 2000. Visible implant elastomers as a method for marking adult anurans. Herpetological Review 31: 154-155.
Nussbaum, R. A., E. D. Brodie Jr., and R. M. Storm. 1983. Amphibians and reptiles of the Pacific Northwest. University of Idaho Press, Moscow, Idaho.
90 Olson, D. H. 2005. Cascades frog: Rana cascadae (Slater, 1939). Pages 214-217 in L. C. Jones, W. P. Leonard, and D. H. Olson, editors. Amphibians of the Pacific Northwest. p. 214-217. Seattle Audubon Society, Seattle, Washington.
Pearl, C. A. and M. J. Adams. 2005. The Cascades frog (Rana cascadae). Pages 538-540 in M. J. Lannoo, editor. Status and conservation of U.S. amphibians. University of California Press, Berkeley, California.
Petranka, J. W. and C. T. Holbrook. 2006. Wetland restoration for amphibians: should local sites be designed to support metapopulations or patchy populations? Restoration Ecology 14: 404-411.
Petranka, J. W., C. K. Smith, and A. F. Scott. 2004. Identifying the minimal demographic unit for monitoring pond-breeding amphibians. Ecological Applications 14: 1065- 1078.
Pilliod, D. S., C. R. Peterson, and P. I. Ritson. 2002. Seasonal migration of Columbia spotted frogs (Rana luteiventris) among complementary resources in a high mountain basin. Canadian Journal of Zoology 80: 1849-1862.
Pope, K. L. 2008. Assessing changes in amphibian population dynamics following experimental manipulations of introduced fish. Conservation Biology 22: 1572- 1581.
Pope, K. L., J. M. Garwood, H. H. Welsh Jr., and S. P. Lawler. 2008. Introduced trout indirectly impact native amphibians via facilitation of a shared native predator. Biological Conservation 141: 1321-1331.
Pope, K. L. and K. M. Matthews. 2001. Movement ecology and seasonal distribution of mountain yellow-legged frogs, Rana muscosa, in a high-elevation Sierra Nevada basin. Copeia 2002: 787-793.
Pope, S. E., L. Fahrig, and H. G. Merriam. 2000. Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81: 2498-2508.
Richards, S. J., U. Sinsch, and R. Alford. 1994. Radio tracking. Pages 155-158 in W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Heyek, and M. S. Foster, editors. Measuring and monitoring biological diversity: Standard methods for amphibians. Smithsonian Institution Press, Washington, DC.
Rickets, T. H. 2001. The matrix matters: effective isolation in fragmented landscapes. The American Naturalist 158: 87-99.
91 Seaman, D. E., J. J Millspaugh, B. J. Kernohan, G. C. Brundige, K. J. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimates. Journal of Wildlife Management 63: 739-747.
Seidel, K. D. 1992. Statistical properties and applications of a new measure of joint space use for wildlife. Masters Thesis, University of Washington, Seattle Washington.
Semlitsch, R. D. 2008. Differentiating migration and dispersal processes for pond- breeding amphibians. Journal of Wildlife Management 72: 260-267.
Sharp, R. P. 1960. Pleistocene glaciation in the Trinity Alps of northern California. American Journal of Science 258: 305-340.
Silverman, B. W. 1986. Density estimation for statistics and data analysis. Chapman and Hall, London, United Kingdom.
Sinsch, U. 1990. Migration and orientation in anuran populations. Ethology, Ecology, and Evolution 2: 65-79.
Skinner, C. N., A. H. Taylor, and J. K. Agee. 2006. Klamath Mountains bioregion. Pages 170-194 in: Fire in California’s ecosystems. N. G. Sugihara, J. W. van Wagtendonk, J. Fites-Kaufman, K. E. Shaffer, A. E. Thode, editors. University of California Press, Berkeley, California.
Slater, J. R. 1939. Description and life-history of a new Rana from Washington. Herpetologica 1: 145-149.
Smith, A. M. and D. M. Green. 2005. Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28: 110-128.
S-Plus. 1999. User’s manual, version 2000. Data Analysis Products Division, Math Soft, Seattle Washington.
Stuart, S. N., J. S. Chanson, A. C. Neil, B. E. Young, A. S. L. Rodrigues, D. L. Fischman, and R. W. Waller. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783-1786.
Swihart, R. K. and N. A. Slade. 1985. Testing for independence of observations in animal movements. Ecology 66: 1176-1184.
92 Sype, W. E. 1975. Breeding habits, embryonic thermal requirements and embryonic and larval development of the Cascade frog, Rana cascadae (Slater 1939). Ph. D. Dissertation. Department of Zoology, Oregon State University, Corvallis, Oregon.
Thoms C., C. C. Corkran, and D. H. Olson. 1997. Basic amphibian survey for inventory and monitoring in lentic habitats. Pages 35-46 in: D. H. Olson, W. P. Leonard, R. B. Bury, editors. Sampling amphibians in lentic habitats. Society of Northwestern Vertebrate Biology, Olympia, Washington.
Trenham, P. C., W. D. Koenig, and H. B. Shaffer. 2001. Spatially autocorrelated demography and interpond dispersal in the salamander Ambystoma californiense. Ecology 82: 3519-3530.
Turchin, P. 1998. Quantitative analysis of movement: measuring and modeling population redistribution in animals and plants. Sinauer Associates, Sunderland, Massachusetts.
Twitty, V. C. 1966. Of scientists and salamanders. W. H. Freeman and Company, San Francisco, California.
United States Department of Agriculture, Forest Service. 1998. Regional Forester Sensitive Species List, Region 5. United States Department of Agriculture, Forest Service unpublished report. Vallejo, California.
Van Winkle, W. 1975. Comparison of several probabilistic home range models. Journal of Wildlife Management 39: 118-123.
Wahbe, T. R., F. L. Bunnell, and R. B. Bury. 2004. Terrestrial movements of juvenile and adult tailed frogs in relation to timber harvest in coastal British Columbia. Canadian Journal of Forestry Research 34: 2455-2466.
Wells, K. D. 1977. The social behaviour of anuran amphibians. Animal Behaviour 25: 666-693.
Welsh, H. H. Jr., K. L. Pope, and D. Boiano. 2006. Sub-alpine amphibian distributions related to species palatability to non-native salmonids in the Klamath Mountains of northern California. Diversity and Distributions 12: 298-309.
Wilbur, H. M. 1980. Complex life cycles. Annual Review of Ecological Systematics 11: 67-93.
Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home- range studies. Ecology 70: 164-168.
93 Zamudio, K. R. and A. M. Wieczorek. 2007. Fine-scale spatial genetic structure and dispersal among spotted salamander (Ambystoma maculatum) breeding populations. Molecular Ecology 16: 257-274.
Zar, J. H. 1999. Biostatistical analysis, Fourth edition. Prentice Hall, Inc., Upper Saddle River, New Jersey.
APPENDICES
Appendix A. Empirical model displaying timing and duration of specific life history attributes for all age groups of Rana cascadae from 2003 to 2008 in Echo Lake Basin, Trinity Alps Wilderness, California. Some life history attributes overlap due to geographic variation in snowmelt and or site availability/ elevations within the basin. Timing and duration of each life history attribute included data from 2006 (extremely wet year with 251% snowpack), and data from 2007 (dry year with 38% snowpack) so extreme annual variability is represented. Vertical arrows in first year stages represent major development transitions.
First Year Stages Eggs Incubation
Larvae Nutrition
Metamorphosis Nutrition
Over wintering
Juveniles Nutrition
Over wintering Over wintering
Adults Extreme Extreme Dry Year Breeding Wet Year Low ELEV. High ELEV. Nutrition
Over wintering Over wintering
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Month
94
Appendix B. Summary table of attributes, movement, and fate of 51 Rana cascadae monitored by radio telemetry for years 2003 and 2004 in Echo Lake Basin, Trinity Alps Wilderness, California.
Radio Tracking Maximum Weight Weight Duration Dist Sum Displacement ID. Sex SUL (g) % (g) Dates Monitored (Days) Fixes (m)a (m)b Fate 1 F 56.2 16.8 5 13 Aug-1 Sep 03 19 12 210 141 Radio battery expired, could not locate 2 F 73.7 42.4 2.4 17 Aug-3 Oct 03 47 13 252 179 Radio battery expired, could not locate 3 M 57 19.7 5.2 6 Jul-1 Sep 03 57 33 518 306 Suspect Thamnophis spp. predation 4 M 57.7 18.7 4.5 21 Jun-22 Sep 03 93 45 263 117 Radio battery expired, could not locate 5 M 59.3 21.5 4.7 24 Jun-22 Aug 03 59 35 1171 260 Released, abrasion on stomach 6 F 65.3 32.5 3.1 21 Jun-12 Aug 03 52 25 747 592 Released, abrasion on hip and stomach 7 F 53.5 15.9 5.3 20 Jun-12 Aug 03 53 31 204 70 Released, abrasion on urostyle and stomach 8 M 57.6 19.5 4.4 24 Jun-5 Aug 03 42 30 829 408 Released, abrasion on stomach 9 M 56.5 21.7 3.9 20 Jun-12 Sep 03 84 37 1342 255 Radio battery expired, could not locate 10 F 61.2 24.5 4.2 23 Aug-15 Sep 03 23 9 409 337 Predation: Thamnophis sirtalis-Length: 70cm 11 F 60 27.7 3.7 23 Jun-4 Aug 03 42 26 348 261 Released, abrasion on dorso-lateral folds 12 M 57.7 22.5 4.5 21 Jun-23 Jun 03 2 3 57 49 Predation: Thamnophis atratus- Length: 85cm 13 M 58.8 20.9 4.1 20 Jun-24 Aug 03 65 40 1132 373 Radio battery expired, could not locate 14 F 68.5 28.8 3.5 18 Jul-15 Sep 03 59 32 839 424 Released 15 F 70.9 30.7 3.3 18 Jul-5 Aug 03 18 14 316 88 Predation: Thamnophis sirtalis- Length: 72cm 16 M 58.2 17.9 5.7 13 Aug-2 Oct 03 50 20 247 80 Released 17 F 59.7 22.3 4.6 24 Jun-17 Jul 03 23 13 221 161 Released, abrasion on dorso-lateral folds 18 M 60 17 6 13 Aug-2 Oct 03 50 21 368 57 Released, abrasion on dorso-lateral folds 19 F 64.1 28.6 3.6 8 Jul-4 Aug 03 27 16 263 66 Released, abrasion on side and stomach 20 F 63 27.5 3.7 5 Jul-17 Jul 03 12 7 185 102 Radio battery expired, could not locate 21 M 58.9 20.9 3.9 18 Jun-3 Jul 04 15 8 19 11 Shed radio 22 M 60.1 21.3 3.8 19 Jun-29 Jun 04 10 7 347 253 Released 23 F 63.6 23 4.6 19 Jun-20 Jun 04 2 2 218 218 Shed radio 24 F 75.5 39.1 2.7 20 Jun-26 Jun 04 6 3 3 3 Shed radio 25 F 71.3 35.4 3 20-Jun-04 1 1 ─ ─ Shed radio 26 F 73.2 31 3.4 20 Jun-26 Jun 04 6 3 11 10 Shed radio 27 M 58.9 18.5 5.7 21 Jun-13 Jul 04 22 10 141 120 Radio battery expired, could not locate 28 M 63.9 21.9 4.8 21 Jun-2 Jul 04 11 6 73 36 Shed radio (continued) 95
Appendix B. Summary table of attributes, movement, and fate of 51 Rana cascadae monitored by radio telemetry for years 2003 and 2004 in Echo Lake Basin, Trinity Alps Wilderness, California (continued).
Radio Tracking Maximum Weight Weight Duration Dist Sum Displacement ID. Sex SUL (g) % (g) Dates Monitored (Days) Fixes (m)a (m)b Fate 29 F 74.7 36.6 2.9 21 Jun-3 Oct 04 104 38 416 271 Released 30 M 59.8 20.4 5.1 21 Jun-10 Oct 04 111 42 357 67 Shed radio, found later and tracked again. 31 M 60 20.1 5.2 21 Jun-3 Aug 04 43 18 91 34 Released, abrasion on right side 32 F 62.7 20.9 5 26 Jun-28 Jul 04 32 16 217 162 Predation: Thamnophis sirtalis 33 F 77.9 35.5 3 27 Jun-10 Oct 04 105 43 830 191 Released 34 F 67 22.5 4.7 27 Jun-9 Oct 04 104 44 408 112 Released 33 F 77.9 35.5 3 27 Jun-10 Oct 04 105 43 830 191 Released 34 F 67 22.5 4.7 27 Jun-9 Oct 04 104 44 408 112 Released 35 F 66.9 27.9 3.6 29 Jun-10 Oct 04 103 39 259 59 Released 36 M 55.8 13.8 5.8 2 Jul-8 Oct 04 98 37 1009 326 Released 37 F 63 20.7 5.1 10 Jul-17 Jul 04 7 6 49 26 Predation: Thamnophis atratus 38 F 67.8 17.9 5.9 12 Jul-8 Aug 04 27 16 301 77 Predation: Thamnophis sirtalis 39 M 56.2 17 6.2 12 Jul-17 Jul 04 5 5 18 10 Shed radio 40 F 62.7 19 5.5 13 Jul-27 Aug 04 45 22 138 63 Predation: Thamnophis sirtalis- Length: 73cm 41 F 56.8 16.3 4.7 23 Jul-10 Oct 04 79 31 211 100 Released 42 M 61.1 19.8 5.3 23 Jul-10 Oct 04 79 30 272 136 Released 43 F 70 25.9 4.1 23 Jul-8 Aug 04 16 8 15 4 Mortality of unknown cause, desiccated 44 F 76.5 39.8 2.7 26 Jul-3 Oct 04 69 27 264 132 Shed radio 45 M 59.3 18.5 5.4 28 Jul-3 Aug 04 7 2 10 10 Radio battery expired, could not locate 46 F 67.4 24.4 4.1 6 Aug-4 Sep 04 30 2 14 14 Released, abrasion on dorso-lateral folds 47 M 58.8 16.6 6.3 18 Aug-4 Sep 04 17 6 15 5 Released 48 F 73.6 31.5 3.2 20 Aug-8 Oct 04 49 15 50 14 Released 49 F 73.8 37.9 2.6 4 Sep-10 Oct 04 36 13 140 71 Released 50 F 63.9 22.5 3.6 9 Sep-10 Oct 04 31 10 174 112 Released 51 M 56.6 17 6.2 19-Jun-04 1 1 ─ ─ Shed radio aCumulative distance traveled by an individual over its entire tracking period. bDistance separating the furthest two telemetry locations for each individual representing the maximum displacement.
96
Appendix C. Water temperature profiles of two shallow spring fed ponds in Echo Lake Basin, Trinity Alps Wilderness, California, during the winter of 2007/ 2008 used annually from 2003 to 2008 by Rana cascadae for overwintering. Grey line represents the water temperature profile of a data recorder 45 cm below the pond surface suspended under 15 cm of loose silt. Black line represents the temperature profile of a data recorder in a second pond located 93 cm below the surface in open water 10 cm above the bottom. Temperatures were recorded every two hours continuously from 07 October 2007 to 04 July 2008.
6.5
6.0
5.5
5.0
4.5
Water Temperature (C) 4.0
3.5 01-Oct-07 01-Dec-07 01-Feb-08 01-Apr-08 01-Jun-08 01-Aug-08 Date
97