-
extracted text (extracttext:extracted_text)
-
Projecting climatic suitability for a high-elevation amphibian
in the Cascades range: a case study of the
Cascade frog (Rana cascadae)
by
Dylan Kubina
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2017
©2017 by Dylan Kubina. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Dylan Kubina
has been approved for
The Evergreen State College
by
________________________
John Withey, Ph. D.
Member of the Faculty
________________________
Date
iv
Abstract
Projecting climatic suitability for a high-elevation amphibian
in the Cascades range: a case study of the
Cascade frog (Rana cascadae)
Dylan Kubina
Amphibians in the Cascades Region are declining. In montane wetlands of the Cascades
Range, several species of amphibians have been decreasing in number since the late
1800s primarily due to the stocking of Eastern brook trout (Salvelinus fontinalis) that
took place in originally fishless lakes. The various trout species introduced to these lakes
often prey on the different life stages of the mountain-dwelling amphibians. The present
study focuses on the threat of anthropogenic climate change for the Cascade frog (Rana
cascadae). Cascade frogs occur in high mountain lakes from northern California through
Oregon, and northeast Washington. To examine the effects of climate change, I took
1,600 records from a variety of databases to represent known occurrences of Cascade
frogs. I used bioclimatic features from Worldclim.org as inputs into the species
distribution modeling (SDM) software MaxEnt to project the current probability of
occurrence for Cascade frogs. In addition, I used four Representative Concentration
Pathways (RCPs) from Worldclim.org to project the suitability of where Cascade frogs
may occur. MaxEnt predicts high probabilities of occurrence of Cascade frogs in Mount
Rainier National Park, North Cascades National Park, Olympic National Park, and as far
south as Lassen Volcanic National Park and the Mount Shasta region. MaxEnt showed
that future projections of Cascade frogs dwindling across their Cascades Range
distribution, with increases in suitability moving toward northern California in RCPs 4.5
and 6.0. The most dramatic decreases in Cascade frog suitability was in RCP 8.5; all the
Cascade frogs’ hot spots showed decreases in suitability, with new suitability appearing
in northeast Washington.
v
Table of Contents
List of Figures ....................................................................................................... vi
List of Tables ....................................................................................................... vii
Acknowledgements ................................................................................................1
Chapter 1: Introduction and Literature Review.................................................2
Introduction ............................................................................................................2
Abiotic and Biotic Factors in High Mountain Wetlands that Affect Amphibians of
the Western United States .....................................................................................7
Introduction ..................................................................................................7
Trout Introduction History ...........................................................................8
Species of Importance ................................................................................11
The Cascade Frog Biology.........................................................................12
Factors Relating to Decline of Cascade Frogs ...........................................12
Abiotic Factors Affecting Amphibians ......................................................15
Biotic Factors Affecting Amphibians ........................................................17
Climate Change Factors Affecting Amphibians ........................................20
Amphibian Studies in the Western US ......................................................22
Habitat Segregation ...............................................................................22
Paleolimnological Observations ............................................................23
Removal Restoration Techniques ...........................................................24
Trout Removal Experiment .....................................................................26
Amphibian and Trout Abundance Findings ............................................29
Low Elevation/ Artificial Ponds ..............................................................29
Conclusion ...................................................................................................30
vi
Gaps in the Research ....................................................................................31
Amphibian Survey in an undocumented Montane Wetland ........................31
Chapter 2: Geographic patterns of a high-elevation amphibian in the Cascades
range: A case study of the Cascade frog (Rana cascadae) ................................34
Abstract ..................................................................................................................34
Introduction ............................................................................................................36
Materials and Methods ...........................................................................................39
Geographic limits of distribution .................................................................39
Elevational limits of distribution .................................................................41
Habitat Types of the Cascade Frog ..............................................................41
Preliminary Research: The Enchantments Basin ........................................42
Data Collection ............................................................................................43
Data Analysis ...............................................................................................44
Chapter 3. Results ................................................................................................47
MaxEnt Current Suitability Projection ..................................................................47
Model Performance ................................................................................................47
Jackknife Analysis .................................................................................................48
MaxEnt 2070 Projections ..........................................................................................48
Chapter 4. Discussion and Conclusion ...............................................................57
Reasons for Decline ...............................................................................................57
MaxEnt Explanations .............................................................................................58
Potential Changes to the Analysis..........................................................................60
Future Ecological Implications for Wetland Biota ................................................60
Literature Cited ...................................................................................................63
Appendix ...............................................................................................................73
v
List of Figures
Figure 1. General Model of Dissimilarity Time Series .........................................................25
Figure 2. Range of Cascade Frogs ........................................................................................41
Figure 3. Cascade Frog Spatial Distribution amongst the Cascades Region .........................49
Figure 4. Current Suitability ..................................................................................................50
Figure 5. Omission and Predicted Area for 0 ........................................................................51
Figure 6. Sensitivity vs. Specificity .......................................................................................51
Figure 7. Jackknife Analysis .................................................................................................52
Figure 8. Climate Projection Scenario 2.6 for 2070 ..............................................................53
Figure 9. Climate Projection Scenario 4.5 for 2070 ..............................................................54
Figure 10. Climate Projection Scenario 6.0 for 2070 ............................................................55
Figure 11. Climate Projection Scenario 8.5 for 2070 ............................................................56
vi
List of Tables
Table 1. Trout Occupying the Montane Wetlands of Western North America ....11
Table 2. Amphibians Occupying the Montane Wetlands of Western North America .....11
Table 3. Bioclimatic Factors .................................................................................45
Table 4. ICPP Global Mean Representative Concentration Pathway Temperature
Projections..............................................................................................................46
vii
Acknowledgements
This study took over a year and a half and would not have been possible if it weren’t for
the support of my mentors, teachers, and peers. I’d like to give a huge thanks to my thesis reader
John Withey for the constant support throughout the last year. John read sections of my thesis
and helped me to learn GIS as well as understand concepts having to do with Species Modeling
Distribution and its associated software, MaxEnt. I would also like to thank Erin Martin for
getting me starting with crafting my prospectus in the earlier stages, David Geerharts for getting
me on the right foot with the use of ArcGIS and ArcMap, Ryan Hobbs and Desiree Anderson for
the constant support and aid through ArcGIS. Thanks to Monika Lapinsky, Frank Stevick, and
Jenny Welch for always peer-reviewing my sections. Lastly, thank you to Eric Jonsson, Alex
Case-Cohen, Stephen Leitheiser, and Erik Johnson for accompanying me on my trips to the
Enchantments. Every ounce of help and inspiration I got for my thesis would not have been
possible without these great people.
1
Chapter 1: Introduction and Literature Review
Introduction
Amphibian declines are a global phenomenon. Amphibians are more threatened
than taxa such as birds or mammals (Stuart et al. 2004) (Table 2). Montane wetlands
(also referred to as high mountain lakes, or mountain wetlands) are particularly affected
as climate change is reducing habitat for amphibians in these regions (Lee et al., 2014).
Warming climate has led to habitat drying and reduced hydroperiods, which is believed
to have promoted the decline of many amphibian species (Carey & Alexander, 2003;
Ryan et al., 2014, Lee et al., 2014).
In the Cascade Mountain Range of Western North America, climate change coincident with the introduction of novel predatory fishes in the 1800s, namely several
species of trout, has also led to a decrease in amphibians (Bahls, 1992; Drake & Naiman,
2000; Knapp, Corn, and Schindler, 2001; Pilliod & Peterson, 2001; Pister, 200; Ryan et
al., 2014). Precipitation, snowpack, and evaporation are the most important factors
reflecting the warming climate in the montane wetlands. Once fishless lakes are now
facing many disturbances, however this study will focus specifically on the Cascade frog
(Rana cascadae) which occupies the Cascades Range of the western U.S. The present
study aims to examine both abiotic and biotic factors in relation to how climate change
may be influencing the habitats of the decreasing Cascade frog.
Montane wetlands are among the most sensitive ecosystems to climate change
(Lee et al., 2014). They include lakes found in high elevations with surrounding subalpine to alpine regions. The regions offer habitat to a large range of species and are part
of crucial hydrologic and geochemical processes (Lee et al., 2014). Humans use the land
2
more frequently in lowland wetlands, which causes those areas to sustain more impacts.
Relatively untouched by human development, montane wetlands areas are more affected
by climate change impacts. Montane wetland areas have a diverse vegetation and fauna,
which make them unique and unblemished in relation to many other parts of the world
(Ryan et al., 2014; Lee et al., 2014). Unfortunately the introduction of trout to has
affected many of these ecosystems. State fisheries managers estimated 95% of roughly
16,000 large mountain lakes in the western U.S. were naturally fishless, but now almost
all the lakes contain fish (Bahls, 1992; Drake and Naiman, 2000, Knapp et al., 2001).
Now, 60% of smaller ponds and lakes contain fish (Ryan et al., 2014). Around 60% of
the total number of lakes and 95% of deeper lakes (>3 m) and larger (>2 ha) lakes contain
trout. There is little research on base stocking programs (Bahls, 1992).
Climate change typically affects the mountainous ecosystem most severely
because of the unique abiotic factors such as high elevation and permanent snow.
Wetlands themselves are important globally because they are critical for both natural
communities and human habitat. Since wetlands are so vulnerable to the impacts of
climate change, focused research is necessary to address why species such as the Cascade
frog are in decline (Lee et al., 2014).
The Cascade frog is an endemic flagship species of the Cascades, a volcanic
mountain range of the Western US. Cascade frogs also have an isolated population
system in the Olympic Mountains. Looking at frog populations in montane wetlands of
the Cascades Range over the last 100 years has indicated significant decline in most
mountain-dwelling frogs since the late 1800s (Bahls, 1991; Knapp et al., 2001; Pilliod
and Peterson, 2001; Pister, 2001; Ryan et al., 2014). Although the Cascade frog is not
3
federally listed, climate change coupled with other abiotic and biotic influences could
lead to serious future declines for the Cascade frog. The Cascade frog is listed as near
threatened (NT) by the IUCN (IUCN, 2004; NatureServe, 2017).
Abiotic features are crucial when discussing an amphibian’s population status in
high mountain lakes. Climate change influences abiotic features such as precipitation
and temperature most directly, leading to earlier snowmelt at more rapid rates. Increased
warming in all seasons leads to earlier snowmelt, which in turn, leads to loss of mountain
snowpack and increased evapotranspiration and increased soil-moisture in late summer
(Lee et al., 2014). Duration and severity of cold weather of both the surroundings and
water temperature are important abiotic factors affecting Cascades high mountain lakes.
Lastly, hydrological (i.e. groundwater recharge, surface water storage, and filtration) and
geochemical cycles (nutrient cycling, carbon sequestration, sediment and contaminant
transport) exist that greatly affect populations of the Cascade frog. Montane wetland
ecosystems tend to be understudied, so looking in greater detail can help provide links to
biotic features and climate change (Ryan et al., 2014; Lee et al., 2014).
Biotic Factors
Biotic factors affect Cascade frogs through processes such as nutrient cycling and
food webs (Ryan et al., 2014). Climate change is proposed to influence biotic factors as
well such as warmer temperatures year-round leading to earlier and faster wetland
drawdown rates. This causes water surface levels to go down and therefore take away
habitat from species in addition to amphibians (Ryan et al., 2014; Lee et al., 2014). Biotic
factors that heavily influence Cascade frogs include vegetation, the amphibian chytrid
4
fungus, and introduced trout (Bahls, 1992; Knapp et al., 2001; Pister, 2001; Ouellet et al.,
2005; De León, Vredenburg, and Piovia-Scott, 2017). Vegetation provides habitat and a
food source for Cascade frogs (Ryan et al., 2014). At the lower elevations of the Cascade
frog range, well-developed coniferous forest dominates the habitat (US EPA, 2017).
Cascade frogs in Crater Lake tend to reside along the shoreline areas of ponds, tarns and
lakes, where they may conceal themselves under woody debris surrounded by moss and
other wetland plants (Farner and Kezer, 1953).
Prominent among biotic factors thought to affect Cascade frogs is the fungal
disease, chytridiomycosis, that prominently affect amphibians globally (Ouellet et al.,
2005; De León, Vredenburg, and Piovia-Scott, 2017). This fungus has profound effects on
amphibian populations globally as it is thought responsible for the extirpation of up to
200 species of amphibians. The chytrid fungus (Batrachochytrium dendrobatidis or Bd)
that causes chytridiomycosis has dramatically affected amphibians in Central America,
but it may be the biggest threat to montane populations of amphibians globally. Studies
reveal that montane-dwelling amphibians like the Cascade frog such as the two species of
Mountain yellow-legged frogs (Rana muscosa and Rana sierra) are species prominently
affected by the amphibian chytrid fungus (Ouellet et al. 2005). In northern California,
chytrid fungus has been speculated to be the cause of large declines in distribution of the
Cascade frog. Laboratory experiments show that post-metamorphic Cascade frogs are
susceptible to Bd (Garcia et al., 2006; Piovia-Scott et al., 2015). The most dramatic
declines of Cascade frogs in northern California are thought to be because of Bd (Fellers
et. al., 2008; Pope et al., 2014; Piovia-Scott et al., 2015). Although Bd is widespread
throughout the range (Adams et al.. 2010; Piovia-Scott et al., 2011), declines have not
5
been observed everywhere (Pear et al., 2009) suggesting that there is variation in the
effect of Bd on populations of this host amphibian (Piovia-Scott et al., 2015).
Aquatic taxa are also important biotic elements of the Cascade frog landscape.
Introduced trout represent an especially high risk to frogs due to predator abilities. Trout
presence trout has led to trophic cascades as well as drops in occupancy levels (Bahls,
1992; Knapp et al., 2001; Pister, 2001). Trophic cascades occur due to consequences of
novel predator introduction. In this case, the trout eat the Cascade frog’s eggs, thus
reducing the number of Cascade frogs. The balance of the food web changes due to
larger, native predators that would have typically fed on Cascade frog life stages having
less to feed from, so they also experience declines in population. Strong feedbacks take
place between fish, herbivores, and algae (Drake & Naiman, 2000). These feedbacks are
related to the addition of top predators that alter lake food chains through trophic
cascades. Perturbations also take place in food-chain structure and primary production
(Carpenter and Kitchell, 1993; Drake & Naiman, 2000).
Climate Change Influences
The Cascades are seeing climate change from human influences. Anthropogenic
climate change appears to be eliminating crucial habitat for amphibians that occupy these
mountain lakes. In fact, the changing climate is thought to place amphibians; in this case,
the Cascade frog, at risk of being extirpated (Ryan et al., 2014). It has been predicted
that by the 2080s, (Lee et al. 2015) climate change will cause a reduction in wetland
habitat availability for many species, namely amphibians (Lee et al., 2014). Climate
change effects can be seen impacting amphibian populations already in many sections of
6
the western U.S. (Carey and Alexander, 2003; Ryan et al., 2014; Lee et al., 2014). In the
Cascades Region, many amphibians could be losing habitat crucial for breeding and
overwintering (Ryan et al., 2014).
The first section of this thesis will investigate existing peer-reviewed literature to
ascertain what information exists for amphibians in the western US. After I describe
abiotic and biotic factors in some detail, I will discuss my modeling results and explore
projections of how areas of climatic suitability may be changing for Cascade frogs. I will
discuss reasons for decline, future suggestions, and future implications for the Cascade
frog and mountain-dwelling amphibians in general.
Literature Review: Abiotic and Biotic Factors in High Mountain
Wetlands that Affect Amphibians of the Western United States
Introduction
Amphibians in the high-elevation lakes of the Western United States (US) are
dwindling now more than ever due to climate change and the introduction of trout during
the 1800s. Climate change affects abiotic and biotic factors (i.e. vegetation), potentially
leaving amphibians with reduced suitable habitat. Evidence increasingly suggests that
climate change leads to a decline in amphibians, however, more research is needed to
demonstrate that some aspect of climate change is the causal factor. Some data have
revealed indirect effects of climate change, for example, on the earlier initiation of
breeding activities (Carey & Alexander, 2003). Since amphibian population decline
reflecting climate change has received relatively little attention since the turn of this
century, this thesis focuses on examining how different climate change scenarios may
7
influence Cascade frog distribution, and consider that in context of trout introduction and
other abiotic factors (Carey & Alexander, 2003; Linder et al., 2003).
Understanding historical accounts from previously published works is crucial to observe
why certain events are taking place. This literature review covers a history of the trout
introduction of the 1800s, which have been introduced to many montane lakes in the
western United States. After I discuss this historical background, I will discuss important
factors likely to affect Cascade frogs, including the impending changes in the climate,
and other abiotic as well as biotic factors. I will use past research to describe six different
types of montane wetland amphibian investigations: 1) experiments based on habitat
segregation, 2) paleolimnological observations, 3) removal restoration techniques, 4)
trout removal experiments, and 5) amphibian and trout abundance findings, and 6) low
elevation/artificial ponds
Trout Introduction History
A variety of fish-stocking clubs brought fish into the montane wetlands
throughout the 1800s and 1900s. The US Forest Service, which was established in 1905,
stocked trout for recreational purposes. While they stocked trout for sport, other local
fish and hunting clubs introduced trout illegally for resource purposes for years and
continue to do so to this day (Bahls, 1992; Pister, 2001; Ryan et al., 2014). Drake and
Naiman (2000) discovered from personal observation that fish in the Mount Rainier
National Park (MORA) had been stocked illegally.
The elevated lakes within the mountains of the western U.S. are landscapes dating
from the end of the Pleistocene epoch. Recession of glaciers, which took place over
8
10,000 years ago during this epoch, created many lakes impassible to entry by fishes
because of barriers. The fishless lake’s biota was fragile due to set up of nutrient cycling
occurring in the lakes. Lakes that are fishless contain important habitat, refuge, and
routes of recolonization for plants, invertebrates, and vertebrates (Drake & Naiman,
2000). Other semi-aquatic species, that could negotiate those barriers, such as
amphibians, prospered here thousands of years ago in lake environments devoid of fish
(Knapp et al. 2001).
European settlers facilitated the introduction of trout into the lakes for sport
fishing (Bahls, 1992; Drake & Naiman, 2000; Knapp et al., 2001; Pilliod & Peterson,
2001 and Pister, 2001). Due to technological advances, the U.S. Forest Service could
introduce trout at more rapid rates (Bahls, 1992). In the 1940s, aircraft became a common
method of transporting the trout into the lakes. A variety of trout species have been
stocked in western lakes over the years (Table 1). In 1964, stocking came to a halt in
Wilderness lakes that had been stocked prior to their Wilderness designation due to
federal guidelines (Bahls, 1992). Fish stocking ceased in 1963 in MORA because the
Leopold Report (Leopold et al., 1963), which tried to urge against the trout introductions
in naturally fishless lakes solely for the sake of recreation (Drake & Naiman, 2000).
Leopold was more concerned for the conservation of the parks rather than the prospect of
sports and resource management gain. Again, in the 1970s and 1980s, fish stocking was
stopped in most national parks in the Western U.S. (Knapp et al., 2001). However, the
U.S. Forest Service continued stocking fish in other unprotected areas in the montane
West (Landres et al., 2001; Knapp et al., 2001). Roughly 5% of about 16,000 mountain
lakes in the Western US were estimated to contain fish, but now almost all the lakes in
9
these montane areas contain some species of trout (Bahls, 1992; Drake & Naiman, 2000).
Historical introductions occurred in a near vacuum of scientific knowledge on the
ecological impacts of trout in montane still-water systems. Native Americans may have
had knowledge about the biota found in these high mountain regions, but this information
was not transmitted to European settlers, who planted trout for the sake of sport and
resources. When trout additions first started, European settlers were interested in using
the land available to them to their benefit (Pister, 2001). Amphibians might have fared
better had the trout species died off over time. Instead, entire trophic systems underwent
changes. Fish continued to survive after stocking ceased (Knapp et al. 2001). The
management of trout continues to be a controversial issue because many people desire the
ability to fish recreationally in the high mountain lakes, while others realize how harmful
these non-native predators are to the ecosystem and how amphibian occupancy levels are
depressed because of fish-altered trophic cascades (Bahls, 1992; Knapp et al. Schindler,
2001; Pister, 2001).
10
Table 1. Trout Occupying the Montane Wetlands of Western North America
Standard English Name
(aka Common Name)
Use in High Lakes
Stocking
Scientific Name
Cutthroat Trout
Oncorhynchus clarki
Infrequent
Rainbow Trout
Oncorhynchus mykiss
Widespread & Frequent
Kokanee Salmon
Oncorhynchua nerka
Selected Larger Lakes
Eastern Brook Trout
Salvelinus confluentus
Widespread & Frequent
Table 2. Amphibians Occupying the Montane Wetlands of Western North America
Standard English Name (aka
Common Name)
Scientific Name
Distribution
Long-toed Salamander
Ambystoma macrodactylum
CA, ID
Northwestern Salamander
Ambystoma gracile
BC, CA, OR, WA
Western Toad
Bufo boreas
AK, BC, CA, CO, ID,
MN, NV, OR, UT, WA,
WY
Yosemite Toad
Bufo canorus
CA
Southern Treefrog
Pseudacris hypochondriaca
BJ, CA
Northwestern Treefrog
Pseudacris pacifica
BC, OR, WA
Pacific Treefrog
Pseudacris regilla
CA, OR
Cascade Frog
Rana cascadae
CA, OR, WA
Columbia Spotted Frog
Rana luteiventris
AK, BC, CA, ID, MN,
NV, OR, UT, WA, WY
Southern Yellow-legged Frog
Rana muscosa
CA
Sierran Yellow-legged Frog
Rana sierrae
CA, NV
Rough-skinned Newt
Taricha granulosa
AK, BC, CA, OR, WA
11
Province or State Abbreviations are: Alaska (AK), British Colombia (BC), Baja
California (BJ), California (CA), Colorado (CO), Idaho (ID), Montana (MN), Nevada
(NV), Oregon (OR), Utah (UT), Washington (WA), and Wyoming (WY).
The Cascade Frog Biology
The Cascade frog is a high mountain frog, which live for around 5 years and can
survive under 30 feet of snow. Mottled yellow, tan, brown-olive skin with black spots on
the yellow areas allows the amphibian to stay relatively hidden in the temporary pools
where it hides (Farner & Kezer, 1953; Nussbaum et al., 1983; NatureServe, 2017). The
frog breeds in spring-summer, from March to August, directly after ice and snow melt
(Stebbins, 1985; NatureServe, 2017). Most eggs are laid out over a few days in a pond
and each female lays around 300-500 eggs, most often in aggregations. Tadpoles
metamorphose into frogs about 2-3 months after the female laid their eggs. It is
speculated that frogs breed after their third hibernation (Nussbaum et al., 1983;
NatureServe, 2017). Overwintering takes place in late fall. Cascade frogs are found
congregated in spring-fed ponds and perennial streams (Garwood, 2009; Pope et al.,
2014). The frogs were found overwintering in deep, loose silt at the bottom of a pond
(Briggs, 1987; Pope et al., 2014).
Factors Relating to Decline of Cascade Frogs
Factors influencing declines in amphibians globally are diverse (Stuart et al.
2004). Factors thought to influence the decline of Cascade frogs represent a subset, most
prominently introduced trout, solar UVB radiation, fungal pathogens, and loss of meadow
habitat from fire suppression (Hayes and Jennings 1986, Fellers and Drost 1993,
Blaustein et al. 1994, Fite et al. 1998, Adams et al. 2001). Declines in Lassen Volcanic
National Park, which are particularly severe, most likely reflect a combination of local
12
factors, including (1) presence of non-native predatory fishes that have restricted
available habitat and limited dispersal of frogs, (2) gradual loss of open meadows and
associated aquatic habitats, and (3) loss of breeding habitat due to drought (Fellers and
Drost 1993); and 4) based on more recent data, the amphibian chytrid fungus (PioviaScott et al. 2015). Climate change is an abiotic shift resulting from changes in biotic
influences from humans. Climate change influences both abiotic and biotic factors,
which in turn affects amphibian habitat. Moreover, this region in Northern California
may be more strongly influenced by climate change than areas further North in its
geographic range (i.e., Oregon and especially northern Washington). Cascade frogs have
declined greatly near Lassen Volcanic National Park at the southern end of the range in
northern California (Fellers and Drost, 1993). Jennings and Hayes (1994) and Fellers and
Drost (1993) estimated that this species is extirpated from about 99 percent of its
southernmost population clusters (Mt. Lassen and surroundings) and 50 percent of the
total historical distribution in California. Cascade frogs in their northern range of Oregon
and Washington appear to face fewer/less severe threats posed by climate change and
habitat loss. Declines in Oregon were cited by Nussbaum et al. (1983) and Blaustein and
Wake (1990), but other data found in the literature do not suggest low occupancy rates
(Brown 1997, unpublished data cited by Pearl and Adams 2005). In Oregon, it has been
estimated that a 22% decline in Cascade frog populations has occurred over the past
hundred years (Fite et al. 1998). Cascade frog declines may be related to the sensitivity
of their eggs to increased levels of ultraviolet-B radiation resulting from ozone depletion
(Blaustein et al. 1994). However, this hypothesis is controversial because the abiotic
spectral characteristics of natural waters likely shields eggs from most detrimental
13
physiological effects of UVB (Palen et al. 2002).
Water quality, physical characteristics and nutrient concentrations are among the
most important factors scientists have identified that influence amphibian habitats (Ryan
et al., 2014). These abiotic factors as also interact with biotic factors such as vegetation
and fauna. Abiotic features that appear to be important include precipitation and
temperature, whereas important biotic features of influence include vegetation that the
Cascade frog thrives off and fungal diseases. After all, the Cascade frog hides under
vegetation and needs it in the lakes to support their nutrient cycling. The frog lives in
sphagnum bogs and lay their eggs in submerged vegetation (Briggs, 1987). They tend to
eat algae, detritus, and plant tissue from the vegetation (NatureServe, 2017). Fungal
diseases, which directly affect amphibian health, have spread globally (Piovia-Scott,
2015). Changes due to climate change and introduction of fish are influencing changes in
the ecosystem that do not just impact the Cascade frog, but other species as well.
Various political and conservation entities have given a status to Cascade frogs
reflecting the level of threats and the overall risk to the species. For example, both the
States of California and Oregon list it as a Species of Special Concern, a non-legal
designation indicating that listing of the species may be justified at some point. The
species has no such designation in the State of Washington. Further, the US Fish and
Wildlife Service consider the Cascade frog a Species of Concern, a designation that
carries a meaning like the state-level Species of Special Concern. Based on the IUCN
Red List (2004), the Overall Threat Impact to the Cascade frog has a near threatened
(NT) ranking. Collectively, all designations for the Cascade frog indicate a level of
concern that, though not severe, bears watching.
14
Abiotic Factors Affecting Amphibians
People tend to overlook montane wetland ecosystem’s abiotic factors in literature
regarding amphibians (Ryan et al., 2014; Lee et al., 2014). For example, these factors
can provide powerful insight on habitat changes and loss of biodiversity (Ryan et al.,
2014). Temperature and precipitation are two of the most important abiotic features that
are expected to change the most because of climate change. High elevation coupled with
warming temperatures are factors that scientists and ecologists take into consideration
when performing montane research. This pattern can lead to substantial changes in
amphibian habitat due to the rapid and earlier recession of snow (Carey and Alexander,
2003; Ryan et al., 2014; and Lee et al., 2014). Climate change also influences water
storage, nutrient cycling and carbon sequestration. These processes are not unique to the
high mountain wetlands, but they do negatively affect these sensitive, yet unique
landscapes. Water storage experiences changes related to earlier snowmelt. Snowpack
has declined by greater than 50 percent in some regions over the last half century. An
altered hydrology system means there will be shifts in nutrient transfer, which could
throw off the entire ecosystem. Also, soil inundation can take place, resulting in different
rates of carbon sequestration and release (Ryan et al., 2014; Lee, 2014). Ryan et al.
(2014) used object-based remote sensing is used to categorize wetlands and to classify
the distribution of wetlands within different regions. This method was seen in Ryan et al.
(2014) study with the use of Variable Infiltration Capacity (VIC), which is a coarse-scale
model projection of change in how likely it is for drying to occur in intermediate
hydroperiods wetlands in Washington state (Hamlet et al., 2005; Ryan et al. and 2014;
Lee et al., 2014). There are different types of ponds within the montane wetlands,
including: 1) hydrologic intermediate pools that are places that hold water for most years,
15
but may sometimes dry up, 2) perennial ponds, which are usually inundated by water, but
with fluctuations in volume, which allows for a broader range of species, and 3)
permanent ponds, which always have water in them and provide habitat for the broadest
number of species (Ryan et al., 2014; Lee et al., 2014). These ponds all contain varying
species of amphibians. Perennial ponds are more biodiverse and contain permanent
obligate or facultative aquatic amphibian adult life stages. The Cascade frog can be
observed in either intermediate or perennial ponds. Intermediate ponds are also at a high
risk to climate change because they are vulnerable to drying up and experience a
hydroperiods anywhere from around 2-5 years. The shorter the hydroperiods, the more at
risk the pond.
The physical characteristics that make up most high mountain lakes have changed
significantly over the years. Drake and Naiman (2000) quantified a lake’s environmental
conditions using dissimilarity in diatom composition and relative abundance.
Dissimilarities in diatom assemblages can help elucidate past biological details of lakes
or ponds, and trajectories of change. Dissimilarity analysis can provide an index of
disturbance and identify baseline conditions (Drake & Naiman, 2000). Dissimilarity
analysis of diatoms is based on findings from sediment samples. Sediments revealed that
for any particular lake, organic matter was consistent within each core, but an increase in
organic matter was evident in recent years. This could reflect in lake productivity or
organic terrestrial input. Lastly, invertebrate’s assemblages in lakes varied, but variation
did not appear to reflect fish introduction.
Observing the connections between abiotic and climate change factors is
fundamental to understanding the Cascade frog’s response to climate change in montane
16
wetlands because the two are inextricably intertwined. Elevation, temperature, physical
characteristics of the lakes, water quality, water storage and carbon sequestration were all
factors researchers considered when analyzing why amphibian populations globally and
regionally are declining. Emerging diseases, UV radiation and synergistic interactions
between these factors and others have been shown to contribute to amphibian declines
(Vredenburg, 2014). Gradients in predation are also associated with the aforementioned
hydrologic characteristics and codetermine the distribution of species in varying habitat
types (Welborn et al., 1996; Snodgrass et al., 2000).
Biotic Factors Affecting Amphibians
Biotic factors can be as important as abiotic factors when assessing amphibian
occupancy in the montane wetlands. Many types of biotic factors that can influence the
population status of amphibians; this literature review focuses on vegetative features,
animal species that are part of the ecosystem, interactions with predators (including
trout), and fungal disturbances. Vegetation, aquatic and terrestrial species are the main
categories of biotic factors that influence montane amphibians.
Cascade frogs occur in wet mountain meadows, mossy bogs, and in patchy
coniferous forests. They tend to lay their eggs in shallow water or among submerged
vegetation (Briggs, 1987). The frog adults eat mostly invertebrates, algae, detritus, plant
tissue and small organisms. Egg masses can be found at the surface of shallow water
where emergent vegetation is present. They have also been found in free-floating in
lakes (Garwood et al., 2009; Pope et al., 2014)
17
Cascade frogs share the ponds with many other animal species. The predation of
fish can exclude many native species from viable habitat. Cascade frogs often encounter
other amphibians such as Long-toed salamanders, Western toads, Northwestern
salamander, mosquito larvae, caddis flies and mayfly larvae. The effects of climate
change and trout introduction may also lead to loss of biodiversity and loss of
environment. Habitat diversity is crucial when maintaining high levels of biodiversity.
Climate change can limit the levels of biodiversity in montane wetlands, so montane
wetlands are at risk now more than ever to losing vertebrate and invertebrate species
(Chesson, 2000; Whittaker, 2001; Ryan et al., 2014)
Different species interact with each other in diverse ways, with predation being a
leading cause to changes in the amount of habitat biodiversity (Knapp et al., 2001).
Gradients in predation determine the patterns of amphibian life cycles (Welborn et al.,
1996; Snodgrass et al., 2000, Ryan et al., 2014). Ryan et al. (2014) explains how the
presence of large salamanders and some dragonfly larvae, both predators, can limit larvae
production of smaller amphibians. This will result in less breeding and rearing in the
habitats that the predators occupy (Hoffman et al., 2003; Ryan et al., 2014).
The trout introduction of the 1800s led to altered trophic patterns, manifest as
trophic cascades, leading to changes in primary productivity, food chain structure, and
processes, such as nutrient cycling (Knapp et al., 2001; Kats and Ferrer, 2003, Ryan et
al., 2014). Trophic cascades arise when a novel predator is introduced, in this case the
trout, and that disproportionally consume the early life stages of the amphibians, thus
reduced the likelihood the enough individuals will survive to adulthood to reproduce.
Such a shift reduces the resource base for larger, natural, predators of the amphibians,
18
which may decline as well if they are unable to switch to alternate prey or emigrate.
However, alteration of trophic pathways is not the only consequence of trout
introduction: interactions between fishes, herbivores and algae, which are strong, can
also be altered (Drake & Naiman, 2000). The predation on the Cascade frog by trout
results in changes in the entire biological structure of montane wetlands. Intense
predation by introduced trout has led to alterations in the food web and in nutrient
dynamics (Ryan et al., 2014). Pond-breeding amphibians as well as invertebrates have
adapted to fluctuating environments. Over time, amphibians have enhanced their natural
resilience. Amphibians have large clutch sizes buffer populations per location. This
means they persist through poor years and rebound as soon as conditions improve (Ryan
et al., 2014).
Emerging fungal diseases represent a relatively novel factor influencing
amphibian populations. For example, the amphibian chytrid fungus (Batrachochytrium
dendrobatidis), the causal agent of the fungal disease, chytridiomycosis, has devastated
amphibians in the Central American tropics, including several montane forms (Berger et
al., 1998). This fungus may be the biggest threat to the world’s montane amphibians. In
California, the two species of Mountain Yellow-legged frog appear to be naturally more
susceptible to the chytrid fungus than many other frogs (Ouellet et al., 2005). Temperate
zones, such as the Pacific Northwest (PNW), must also deal with chytridiomycosis. The
Mountain Yellow-legged frogs in California are two of several hundred amphibians
globally that have been affected by this disease. The fungus attacks keratinized areas of a
frog’s body. The fungus does not severely affect tadpoles because they are only heavily
keratinized on their jaw sheaths and tooth rows. In contrast, post-metamophic frogs have
19
keratin-rich skin and suffer worse infections (De León, Vredenburg, and Piovia-Scott,
2017).
Varied biodiversity, trophic processes, amphibian life cycle patterns and fungi
make up biotic features experienced in the high mountain lakes. These influence abiotic
factors and vice versa. Changes in temporal pulses of peak water may strongly influence
several biotic features. This affects local pond metabolism and primary productivity, the
structure of plant communities and the patterns of wildlife connectivity (Mitsch &
Gosselink, 2007; Lee et al., 2014).
Climate Change Factors Affecting Amphibians
Climate change may be negatively impacting montane wetlands (Ryan et al.,
2014; Lee et al., 2014). To understand these impacts, the historical backdrop to montane
wetland formation must be understood. Montane wetlands reflect repeated glaciation
during the Pleistocene era (defined as the time that began 1.8 million years ago and lasted
until about 11,700 years ago). When the glaciers receded and the most recent ice age
occurred, most lakes and ponds we see today were formed. These lakes, mostly devoid
of fish (Knapp et al., 2001), provided habitat for both aquatic and terrestrial species (Lee
et al., 2014).
Climate change may be affecting the way the different amphibians breed and
survive. Snow, a vital part of the hydrologic cycle in the western US, appears to be one of
the pivotal abiotic factors being affected by climate change. In montane wetlands,
amphibians may be negatively impacted by the reduced snowpack and more rapid
snowmelt earlier in the year. Anthropogenic climate change has resulted in warming air
20
temperatures and decreased precipitation as snow (more as rain), which melts available
snowpack more rapidly, and increases plant evapotranspiration and soil moisture stress in
the later summer months (IPCC, 2007; Ryan et al.,2014; Lee et al., 2014). High
mountain lakes may be drying up earlier or, in some cases, permanently due to this
warming weather pattern. Water will be at lower levels in the wetlands, which can
increase amphibian mortality from drought. Earlier seasonal rates of wetland drawdown
and increased drought frequency in the summer months especially have been shown to
place amphibians at risk (Carey & Alexander, 2003; Ryan et al., 2014; Lee et al., 2014).
The undesirable effects of climate change are anticipated to be felt strongly across
the Western U.S., including the PNW, with higher temperatures and much more severe
summer droughts (IPCC, 2007; Ryan et al., 2014; Lee et al., 2014). Until recently, most
studies conducted in the Western U.S. have tended to focus on the effects of trout
introduction that began in the 1800s, however, few studies have focused on climate
change as the main factor in amphibian declines. Carey & Alexander (2003) focused on
the rapid warming that took place during the latter half of the 20th century. Temperatures
now rise at about 0.5°C with an increase in the severity of weather patterns across the
world (Easterling et al., 2000; Carey & Alexander, 2003). This is a fact that must not be
taken lightly, especially for the sake of declining amphibians.
There are many factors that can lead to climate change in the montane wetlands,
pollution being a main factor. Greenhouse Gases (GHG) are gas that trap heat in the
atmosphere: carbon dioxide, methane, nitrous oxide, and fluorinated gases such as
hydrofluorocarbons. Changes in GHG emissions and aerosols can lead to alterations in
the energy balance of the climate system, therefore leading to potential climate change
21
(ICPP, 2007; US EPA, 2015). Besides GHG emissions, motorization of transport has
created congestion and pollution in large cities, which leads to climate change (IPCC,
2007). Ryan et al. (2014) illustrate the significance of using simulations to predict
hydrologic changes, which can be used to demonstrate interactions with pollution.
Increased levels of air pollution and pesticides were also highlighted in Vredenburg’s
(2004) discussion about the endangered mountain yellow-legged frogs. Wetland
dynamics and hydrologic features can be forecasted over broad geographic regions using
predictive modeling programs, such as object-based remote sensing. The predictions
have emphasized not only the impact that trout introduction has had on wetland
ecosystems, but the clear negative impacts felt in relation to climate change (Halabisky et
al., 2011; Ryan et al., 2014). Coupling these climate-predictive models to amphibian
occupancy patterns, it become possible to identify vulnerable areas and meta-population
or landscape-scale resilience of wetlands and amphibians.
Amphibian Studies in Montane Wetlands Performed in the Western U.S.
Habitat Segregation
Hoffman et al. (2003) performed intensive research on both the Long-toed
salamander and the Northwestern salamander within MORA addressing what influenced
their distributions, and was meaningful in demonstrating now two species can partition
their habitat. Physical characteristics, water quality and nutrient concentrations of the 27
sites and 22 ponds and lakes were measured the Northwestern salamander tended to
occupy larger, deeper lakes at somewhat lower elevations that contained coarse woody
debris, whereas the Long-toed salamander tended to be found in smaller, shallower water
22
lakes and ponds at higher elevations in areas with substantial emergent and aquatic
vegetation habitat in the surrounding area. This habitat segregation reflects the life
history and interspecific interactions of each species. The most valuable take-away from
this habitat segregation-based study, is that larvae from montane populations develop
much more slowly than populations from lower elevations. The effects of climate change
might cause the populations to develop at even slower rates in higher elevations. It does
not help that higher elevations are more vulnerable to the changes of earlier snowmelt
and overall warmer temperatures. The amphibians dwelling in MORA lakes and ponds
may not have the time to adapt to the rapid changes that are taking place in higher
elevations, thus leading to declines in distribution (Hoffman et al., 2003).
Paleolimnological Observations
Paleolimnological studies analyze indicators in lake sediments, such as diatoms or
invertebrates to draw conclusions about, or make comparisons to, a lake’s current state.
Drake and Naiman (2000) indicate that fish introductions in MORA have led to many
detrimental effects revolving around amphibians and climate change. These
paleolimnological studies were conducted 20-30 years after fish removal had taken place.
Dissimilarity analyses of diatoms were used to compare the current physical conditions of
the lakes to how they looked before the trout were removed. Eight lakes were examined;
diatoms were used as the indicator because they are very sensitive and can reflect both
damage and recovery in lakes, as scientific experiments and analyses have demonstrated
(Drake & Naiman, 2000). The article stressed just how bad sport fish are to lakes that
previously never had any fish in them. The dissimilarities showed that the introduction
of trout led to many state changes in lakes. There were dramatic changes in diatom
23
communities, but the researchers also observed that there were several sustained state
changes which took place hundreds of years before European settlement. Although these
lake systems have a great deal of capacity to adapt to changes in the climate, the
resiliency is still limited. At some point in time the system loses the ability to absorb
change and will shift into another condition. Ecological conditions in stocked lakes may
have passed the threshold of change, which exceeds the bounds of resiliency (Drake &
Naiman, 2000).
Removal Restoration Techniques
Drake and Naiman (2000) also conducted a removal experiment to see if lakes
could return to predisturbance conditions. Previous studies lacked sediment records. In
1996 and 1997, sediments cores were collected in the deepest sections of the eight lakes.
Half the lakes had previously introduced fish; the other half had never had fish in them.
The researchers chose to observe seven lakes; three lakes occupied by trout, two that
were previously stocked, but that now do not have any trout in them due to removal and
two that were never stocked. An additional ‘restored’ lake that was not one of the
original four lakes was added in two years after the start of the study; this lake once had
fishes, but then they were removed. Sediment cores provided a 480-year-old record from
the eight lakes based on the 210-lead isotope (210Pb). A general model of dissimilarity
(SDC), which was calculated by comparing each assemblage in the 480-year old core
samples to lake-specific baseline values (Drake & Naiman, 2000).
24
Figure 1. General Model of Dissimilarity (SCD) Time Series. The model is
general and not based on real data. SDC analysis is typically based on
comparing proportions of species in the different strata. As time
increases, assemblages become more dissimilar. When an SCD
comparing two different assemlages surpassed 0.15, the two assemblages
are considered different (Drake and Naiman, 2000).
With proper baseline conditions, meaning diatom floras were relatively stable
between 315 and 90 years before being present in all lakes, then the lake is at a critical
value of 0.15 generated in the general model of dissimiliarty (SCD) time series. As time
increases, assembledges become more dissimilar. When two assemblages exceed this
critical value of 0.15, the two assemblages are considered different. Several lakes in
MORA are undergoing fish removal programs. Since 1973, fishes have either
disappeared naturally or have been removed by angling (Drake & Naiman, 2000). The
two lakes that were known to once contain fishes and now that currently have do not
contain any fishes (Eunice and Owyhigh) contained diatoms, which did not return to, nor
show a trend toward pre-disturbance conditions. The explanation for why these lakes are
25
not showing a trend toward making their way back to pre-disturbance conditions is
because recovery may take longer than 20-30 years since fish were removed from the
lakes. Also, ecological conditions in stocked lakes might have experienced a threshold
such that the bounds of resiliency (i.e., returning to pre-disturbance conditions) might
have been exceeded, this means the lakes will not return spontaneously. Lastly, other
disturbances such as lakeshore vegetation may have confounded the fish effect and
affected diatoms as well (Drake & Naiman, 2000).
The Drake & Naiman (2000) experiment suffered from high variabilities in
density and composition of preserved invertebrates in the MORA lakes, so this variable
has the potential to mask identifying changes due to fish introduction. This experiment
focused on general dissimilarities and compositions of diatoms of the lakes to determine
variation as well as the effects of fish introduction with relation the removal restoration
projects, but the experiment did not address amphibians. Perhaps extending the study to
include both abiotic characteristics as well as biotic characteristics of this ecosystem
could aid in explaining why amphibians are declining.
Trout Removal Experiment
A study performed by Vredenburg (2004) focused on removal of previously
stocked trout lakes in the High Sierras of California. Although fish have been
disappearing or have been removed through angling since 1973 in MORA, no researchers
have performed a trout removal experiment of this magnitude (Drake & Naiman, 2000;
Vredenburg, 2004). Vredenburg (2004) was aware of other potential factors, such as
diseases; UV radiation and climate change and increased levels of air pollution and
26
pesticide use; and that these could have synergistic effects on frogs in montane wetlands.
But Vredenburg (2004) also realized that many more studies have been done on these
factors without the consideration of the predation effects of introduced trout. Studies that
took place before Vredenburg’s trout removal experiment considered mostly the role
local factors play. The Mountain yellow-legged frogs of the High Sierras had been
declining dramatically over the last couple of decades (Vredenburg, 2004). Addressing
the effects of introduced trout in the High Sierras was important because the High Sierras
were a focal area during the trout introduction interval beginning in the late 1800s.
The experiment was lengthy (8 years), which reflects the amount of time needed
to see the response in a before-and-after-impact design. Additionally, it reflects the large
number of water bodies utilized (n = 81). The study site was the Sixty Lakes Basin in
Kings Canyon National Park, California, which was chosen largely because Mountain
yellow-legged frogs were still relatively abundant in several fishless lakes in this basin.
Mountain yellow frogs were visual encounter surveyed, and gill nets were used to
determine whether fish were present in each specific lake that was chosen. To obtain a
baseline, the average number of frogs and tadpoles per 10 m of shoreline in lakes with
and without introduced trout were first compared (Vredenburg, 2004). Next, a trout
removal experiment was conducted in a serial fashion in a series of lakes to see if trout
limit both the size and distribution of the mountain yellow-legged frog. Interestingly,
Mountain yellow-legged frogs and tadpoles began to increase in every lake in which fish
removal had occurred; reference (or control lakes) that either already had no fish or from
which fish were not removed showed essentially no change over time. Vredenburg
(2004) concluded that with trout removal, frog populations could recover if source
27
populations existed nearby from which colonization could occur. This study was
particularly telling because a study of this type has only been performed once previously,
but not to this magnitude (Knapp et al., 2001; Vredenburg, 2004).
In a study conducted in North Cascades National Park (NOCA), Tyler and
colleagues (1998a) focused on the Long-toed salamander and the physical properties of
45 studied lakes. The researchers sampled seventeen lakes two or more times in a year, 8
lakes once a year over 2 or more years, and 20 lakes once. A snorkeling method was
used to survey for salamanders, which was beneficial because Long-toed salamander
sometimes like to hide out under or in the substrate. Kjeldahl-N (TKN) concentration (a
method for quantitative determination of organic nitrogen) was an important factor
because in lakes without fish, crustacean zooplankton (a focal food of Long-toed
salamanders) was positively related to TKN concentration. The more zooplankton found
in the lakes (fish-containing or fishless), the greater the number of long-toed salamanders.
When TKN was low(<0.045mg/L), no significant difference existed in the abundance of
larval Long-toed salamanders in lakes with and without fish, but when TKN was higher
(>0.045mg/L), larvae densities of this salamander were higher in the lakes where trout
were not reproducing (Tyler et al., 1998a). The important part here is that one must not
only look at the abundance levels to get the entire picture of what is going on here, but
the natural abiotic and biotic factors and processes as well (Tyler et al., 1998a; Hayes &
Jennings, 1986). This NOCA study only focused on one species, where other species
could be important to fully understand the observed patterns. TKN was the only nutrient
measured, but other nutrients might have refined understanding the pattern. It might
make for a more thorough experiment if Tyler et al. (1998a) observed factors other
28
nutrients besides TKN, or perhaps factors related to climate change taking place in
NOCA.
Amphibian and Trout Abundance Findings
Though it does not address Cascade frogs, the following research is important
because it focuses on the trout introduction in relation to amphibians in high lakes outside
the Cascade Mountains. Although several amphibian trout introduction relationship
studies have been done on small numbers of bodies of water, Pilliod & Peterson (2001)
conducted a study that focused on an entire mountain range in the Salmon River
Mountains of Idaho. Columbia spotted frogs and Long-toed salamanders were the
studied species. The Salmon River Mountains of Idaho, which are part of the Rocky
Mountains, have a climate with long winters (up to 8 months) with heavy snowfall
(Pilliod & Peterson, 2001). Understanding the climate and elevational footprint of an
area is essential to understanding its seasonal pattern of suitability for amphibians.
Surveys were done for both amphibians and fishes at 101 sites. Three different types of
trout (O. clarki, O. mykiss, and O. m. aguabonita) were recorded at 43 lakes; 42.6% of all
sites surveyed. Surveyed showed that fish-occupied sites had fewer amphibians. Lack of
pre-stocking data provides uncertainty as to whether low numbers already existed in the
lakes that had been stocked with fish later. This study suggests that amphibians in fishoccupied lakes will continue to decline in number (Pilliod & Peterson, 2000).
Low Elevation/Artificial Ponds
Another study examined Long-toed salamanders and Northwestern salamander
relative abundances (Tyler et al., 1998b), both species prevalent in montane wetlands.
29
Salamanders were observed in a controlled lab setting using artificial ponds with and
without fish (Tyler et al., 1998b). Long-toed salamander larvae were collected from low
elevation ponds in Benton County, Oregon in 1996. Rainbow trout (O. mykiss) obtained
from Oregon State University were used in the experimental (i.e., fish) treatments and
control ponds were chosen randomly. The researchers measured 20 larvae snout-vent
lengths (SVL) and selected physical characteristics, such as water temperature and
substrate abundance, which were obtained daily. Both salamander species fared better in
the ponds without trout than in the ponds, which contained trout. These conclusions are
based on larval growths of both Long-toed and Northwestern salamanders. Both species
showed longer SVLs in control ponds than in treatment ponds after the experiments.
Although this experiment facilitated measurement that was easy to control and was not
done in a natural amphibian habitat setting, the study did lead to findings that show
reduced foraging opportunities associated with refuge use. This refuge use could lead to
avoided encounters with trout or competition between larvae and trout for a limited food
resource. This then could result in limited growth of larval Long-toed and Northwestern
salamanders. The actual mechanism responsible for reduction in growth of these larval
salamanders in the presence of trout was unable to be determined in this experiment
(Tyler et al., 1998b).
Conclusion
Gaps in the Research
30
Only the sparsest of scientific information exists on amphibian populations before
the introduction of trout in the 1800s. This condition requires that we take indirect
approaches, such as the work of Drake and Naiman (2000), to infer the condition of
fishless lakes and match that with similar condition in current amphibian-occupied
fishless lakes.
Tyler and colleagues (1998b) indicated that many studies on eastern
Ambystomatid larvae exist. Most East Coast findings regarding declining amphibians
have been like West Coast findings (Petranka, 1983; Semlitsch 1988). What could this
mean for different species of amphibians? Focusing on salamanders may help inform
what is happening on different species of amphibians in major amphibian taxa (i.e.
particular frog or salamander families). Considerable information exists regarding trout
of other fish introductions in relation to amphibians, but few of those data have integrated
the effects of climate change. However, many aspects of climate change are uncertain.
However, we do know that many amphibians, because of their water-dependent life
histories, are likely to be vulnerable. We do know that air temperatures are rising and are
expected to continue to rise. We anticipate that rising air temperatures will accelerate
snow melt and increase rates of evaporation, which may accelerate the seasonal shrinkage
or drying of aquatic habitats. How fast changes in hydrology will affect montane
wetlands amphibian populations needs greater region-specific focus.
Amphibian Survey in an Undocumented Montane Wetland
My research included a preliminary amphibian survey I was a part of in the
summer of 2016. The survey took place in the Enchantments Basin of the Wenatchee
31
Okanogan/Wenatchee National Forest. This forest deals with disturbances such as forest
fires and landslide. No amphibian surveys had been previously done in the region.
While there, I teamed up with the US Forest Services, specifically Shannon Claeson of
the Wenatchee Station. On our first trip was from June 20-24; we observed Columbia
spotted frogs tadpoles, juveniles, and adults, and Long-toed salamander larvae. I went
back twice after that, once from July 11-15, and again in from August 15-19 and
observed the same species; larvae, juvenile, and adult Columbia Spotted frogs, and larvae
of Long-toed salamanders. The larvae and tadpoles could be found hiding under rocks on
the edges of the ponds or sometimes swimming near the center of the ponds. The
juvenile adult frogs were often seen on the edges near rocks or hiding under vegetation.
The amphibians typically hid under the rocks near the shores of the lakes and ponds.
Because the core Enchantments Basin sits near the anticipated elevation limit of
amphibian abundance, it seemed crucial to explore the area because of lack of
documentation of frogs found prior in the area. The amphibians were spotted in the
Snow Lakes at about 1524 m (5000 ft) elevation. There is permanent snow only in the
highest areas (i.e. 2133 m (7000) ft elevation) in the Upper Enchantments Core Lakes
(there were roughly 15 core lakes with many various unnamed ponds and tarns). This
area provided an opportunity to assess how abiotic factors versus climate change factors,
such as earlier and more intense snow melt runoff, may affect amphibian abundance. If
no amphibians are present in areas from which the snow disappears late, it could have
more to do with this region not having sufficient seasonal breadth for amphibians to
survive in. Without previous knowledge of amphibians in this area, interpreting my
findings is difficult. There is not any information which exists regarding whether
32
amphibians were abundant in the Enchantments or any of the Cascade montane wetlands
for that matter. With climate change advancing rapidly, understanding how these nearpristine landscapes are changing is crucial.
33
Chapter 2: Article Manuscript
Abstract
Geographic patterns of a high-elevation amphibian
in the Cascades range a case study of the
Cascade frog (Rana cascadae)
Dylan Kubina
The present study focuses on the threat of anthropogenic climate change in the montane
wetlands and modeled its potential effect on Cascade frog (Rana cascadae) distribution.
I first gathered all available verifiable records (~1,600) to estimate the current suitability
of Cascade frogs. I then ran a predictive Species Distribution Modeling (SDM) software,
MaxEnt to estimate Cascade frog suitability under different climate change scenarios. I
obtained 19 bioclimatic features for modeling from Worldclim.org, and used these inputs
to estimate Cascade frogs’ suitability under four future climate scenarios (Year averages
for 2050 and 2070). MaxEnt output revealed that habitat highly suitable for Cascade frog
was more frequent in Mount Rainier National Park, Olympic National Park, and as far
south as Lassen Volcanic National Park. Overall, my training and testing datasets
appeared similar and both datasets were significantly different from random data based
on the Specificity versus Sensitivity output, indicating that a non-trivial model. Using
Representative Concentration Pathway (RCP) climate change scenarios, MaxEnt showed
that future projections reveal Cascade frogs dwindling across their geographic range, with
overall decreases in suitability in Mount Rainier and the Olympics. The most dramatic
overall decrease in suitable Cascade frog habitat was in RCP 8.0 (2070). RCP 2.6 was the
most similar scenario to the current suitability, while RCP 4.5 and 6.0 displayed growing
34
suitability in northern California with RCP 8.0 showing the sparsest suitability of
Cascade frogs overall.
35
Introduction
Montane wetlands are keystone ecosystems in the Cascades Mountains. Over the
last 150 years, two significant changes have affected these montane wetlands and their
occupant amphibians. The first was when fish were introduced to previously fishless
lakes, which began in the 1800s. The second is anthropogenic climate change, which
though not recognized until recently, promises to dwarf all previous disturbances to
montane wetlands (Carey & Alexander, 2003, Ryan et al., 2014, Lee et al., 2014).
Because these two factors (climate change and introduced fish) interact, amphibians in
montane wetlands are at particular risk. This is unsettling because these wetlands also
provide important ecological services for both the assemblage of species living in them as
well as human society (Ryan et al., 2014; Lee et al., 2014). Importantly, these montane
wetlands provide food to aquatic and terrestrial species, and influence local hydrologic
processes (i.e. groundwater recharge, surface water storage, and filtration), and
geochemical cycles (Lee et al., 2014). They also tend to be understudied despite their
potentially high sensitivity to climate change. There are few data available for montane
wetlands relative to other ecosystem types. Montane wetlands may be experiencing
declines in amphibian populations because of anthropogenic climate change and the
introduction of trout in the 1800s. This thesis focuses on climate change factors affecting
montane amphibians because amphibians in montane wetlands may be particularly
sensitive to climate change, and research in montane wetlands that explores habitat
changes as a function of climate change is sparse (Burkett and Kusler, 2000, Ryan et al.,
2014, Lee et al., 2014).
36
Temperature and precipitation are the two most important abiotic climatic factors
that affect amphibians. Abiotic factors are understudied in montane landscapes (Ryan et
al., 2014, Lee et al., 2014). High elevations may experience disproportionately warmer
air temperatures than lower elevation locations. Hence, water storage, nutrient cycling
and carbon sequestration processes, may be differentially affected in montane landscapes
(Carey & Alexander, 2003, Ryan et al., 2014, Lee et al., 2014). If so, montane
amphibians may be losing their habitat disproportionately faster than elsewhere.
Climate change influences changes in biotic factors, which in turn affect Cascade
frogs. Vegetation which makes up the frog’s habitat and food sources, animal species
that meet the Cascade frog, and fungi leading to diseases leading to the extirpation of
over 200 amphibian species are the largest factors, which influence Cascade frogs. Trout
did not exist in most high mountain lakes of the Cascades before the 1800s (Bahls, 1992).
Trout, a potential predator of amphibians, now occupy almost all the lakes found in
Cascade frog geographic range (Bahls, 1992). Fungi, specifically the amphibian chytrid
fungi that causes chytridiomycosis, not only affects species such as the High Sierras
Mountain yellow-legged frog as well as the Cascade frog in the PNW (DeLeón,
Vredenburg, and Piovia-Scott, 2017) may be the biggest threat to the world’s entire
montane amphibian population (Ouellet et al., 2005).
Climate change influences both abiotic and biotic factors in the montane
wetlands. Mountain lakes are vulnerable to rapid environmental changes stemming from
climate change due to their high elevations which lead to permanent snow, which is
melting earlier on in the year Ryan et al., 2014). Snow is an increasingly important part
of the hydrologic cycle in the Western U.S. Montane wetlands are sensitive to hydrologic
37
drivers that determine the rate and balance of water inflow and seasonal fluctuations in
pond levels (Ryan et al., 2014; Lee et al., 2014). In vulnerable montane wetlands,
amphibians are likely to be negatively impacted by the more rapid snowmelt resulting in
a reduced later season snowpack because of warming temperatures (IPCC, 2007, Ryan et
al., 2014, Lee et al., 2014). Increased warming results from human disturbances that add
greenhouse gases (GHGs), leading to greater warming and shifts from less precipitation
as snow and more precipitation as rain, which both lead reduced snow packs, increased
evapotranspiration and soil moisture stress in the later summer months (IPCC, 2007,
Ryan et al., 2014, Lee et al., 2014). Areas in the montane wetlands will be drier due to
warmer weather and wetland water levels will be lower, so more amphibians may beat
the risk of successful metamorphosing (Ryan et al., 2014, Lee et al., 2014). Earlier rates
of wetland drawdown and increased frequency of summer drought is already occurring,
both of which appear to be detrimental to amphibians. These negative symptoms of
climate change are expected to worsen in the Western U.S., that is reflecting even higher
summer temperatures and much more severe droughts (IPCC, 2007, Ryan et al., 2014,
Lee et al., 2014).
As a PNW montane endemic, the Cascade frog represents the ideal species for
observing the declining patterns among amphibian populations in the Cascades Region.
This species lives in sphagnum bogs and lay their eggs in submerged vegetation (Briggs,
1987). They tend to eat algae, detritus, and plant tissue from the vegetation
(NatureServe, 2017). The frog was once more abundant, but due to the fish stocking of
the late 1800s (Bahls, 1992) and potential abiotic and biotic features as well as
anthropogenic climate change, the species has suffered declines in habitat distribution
38
(Knapp et al., 2001, Ryan et al., 2014). The Cascade frog was historically especially
abundant in Lassen Volcanic National Park, California, which is located at the Southern
end of its range. Fellers & Drost (1993) conducted research in Lassen Volcanic National
Park to find very few frogs when compared to the much larger number of frogs found in
the past. Additional data demonstrates that a similar pattern is taking place in breeding
grounds in various montane wetlands in the western U.S. (Ryan et al., 2014, Lee et al.,
2014).
Extensive research has been done on the trout stocking in high mountain lakes of
the western U.S., however only recently have a few articles focused on anthropogenic
climate change as a real threat to the amphibians of the Cascades (Cary & Alexander,
2003; Ryan et al., 2014; Lee et al., 2014). If populations continue to decline, the Cascade
frog could be put on the endangered species list. The goal of this research is to illustrate
that climate change may influence habitat suitability in the Cascades montane landscapes,
which may decrease Cascade frog occupancy. First, I will describe my methods used to
obtain Cascade frog data. Next, I will show where Cascade frogs occurred
geographically using MaxEnt species distribution modeling. Then I will also use MaxEnt
to make predictions about Cascade frog suitability under different climate change
scenarios. I will close by providing suggestions of how to refine the model, how to test
predictions, and what actions should be taken in the future.
Materials and Methods
Geographic limits of distribution:
39
The Cascade frog range begins in northern California in Lassen Peak National
park and runs through the Cascades of Oregon, north into Washington (Figure 2). A
population also exists in the Olympic National Park of Washington (which is not a part of
the Cascade Range). Populations in the Olympic Mountains of Washington and the
Trinity Alps, Mount Shashta, and Mount Lassen are disjunct from the primary
distribution along the main Cascades Range (Pearl and Adams, 2005; NatureServe,
2017). Lassen Volcanic National Park contains smaller numbers of Cascade frogs with
warmer year-round temperatures than Oregon and Washington.
40
Figure 2. Range of Cascade Frogs (IUCN, 2004)
Elevational limits of distribution:
The Cascade frog lives at elevations of around 1524 m (5000 ft). Historical
locations at low elevations in Washington suggest that the Cascade frog may have once
been more broadly distributed (Leonard et al., 1993). Historical elevation range extended
from around 400 to 2,500 meters (1312 ft- 8202 ft) (NatureServe, 2017).
Habitat Types of the Cascade Frog
41
The Cascade frog lives in sphagnum bogs and lay their eggs in submerged
vegetation (Briggs, 1987). The frogs do not live out their days where they were bred, but
rather near edges of ponds and streams. Overwintering takes place in late fall. Cascade
frogs are found congregated in spring-fed ponds and perennial streams (Garwood, 2009;
Pope et al., 2014). The frogs were found overwintering in deep, loose silt at the bottom
of a pond (Briggs, 1987; Pope et al., 2014).
Preliminary Research: The Enchantments Basin
I surveyed the Enchantments Lakes Basin in the summer of 2016 to allow me to
get a sense of montane wetland habitats. The survey, conducted in conjunction with the
U.S. Forest Services and WDFW, is of special interest as the Enchantments sit on an
anticipated elevation boundary of amphibian occupancy. This means that if amphibians
exist here they are strongly influenced seasonal by abiotic factors. I traveled to the
Enchantments Basin on three separate occasions from the dates June 20-24, July 11-15,
and August 15-19, 2016. No previous amphibian surveys had been conducted in the
Enchantments prior to this study. To reach the high lakes, a trek to around 1,524 m
(5,000 ft) elevation was necessary to see where the first high-elevation amphibians
occurred. Surveys were conducted upwards on to around 2,133 m (7,000 ft), where
permanent snow and near-freezing water was apparent. Our surveys in the Enchantments
documented amphibians using standard visual encounter techniques along shorelines and
streams using a dip net. I measured amphibians trapped in nets and released them back
into the habitat. I documented the species by keying them out per features such as snout
to vent length and physical features including spot size, color, and eye placement.
Tadpoles were measured as well. I recorded notes on dominant vegetation type, animals
42
present, water quality based on substrate, temperature of water, and the size of lakes. I
used visual encounters to observe vegetation animal species, water quality and lakes over
a meter. I used a digital thermometer to measure water temperature in degrees Celsius.
Although there were not any amphibians onwards of 7,000 ft in the Upper Enchantments
lakes, there were amphibians in the 5,000 ft zone, which contained Upper and Lower
Snow Lakes. There was also an intermediate pond between Lower and Upper Snow
Lake, which I called Little Upper Snow Lake. The amphibians I did find, all in the Snow
Lakes complex, around 5000 ft elevation, were Columbia spotted frogs, Pacific tree
frogs, and Long-toed salamander larvae.
Data Collection
I assembled Cascade frog data from several sources. These included the Vertnet
database, and databases obtained from the Mount Rainier National Park headquarters, the
Olympic National Park headquarters, and USGS. Vertnet, a National Science
Foundation-funded vertebrate database, had the most extensive database, but it was
entirely based on verifiable museum specimens. Remaining databases included a mix of
survey data (the majority) and specimen-verifiable information1. Most location records
in the databases were georeferenced, but points lacking georeferenced data or for which
the geo-referencing data were too vague were not used in the analysis dataset. Over
5,000 georeferenced records existed from all four databases combined, but this included
duplicates. I removed duplicates for the analysis dataset. I kept latitude, longitude pairs
and dates when the specimens were collected or documented for the GIS portion of the
results2. I converted all georeferenced locations in UTM zone 10N to WGS84 latitude,
longitude coordinates. Once converted, the latitude, longitude pairs depicted a spatial
43
distribution of where the frogs occurred (Figure 2). I elected to keep the points that were
missing dates if they were georeferenced on my maps, I left the Central Valley of
California on my maps to show lands outside of the Cascade frog range.
There were some points that I was skeptical of due to the location on the map and
other parameters suggesting they were possible misidentified frogs and their locations
were in areas unlikely to be occupied by Cascade frogs, so I deleted these points. Details
of these points are provided in Appendix 2.
Data Analysis
I used the screened georeferenced data points from the databases as inputs for
MaxEnt use. I used parallel georeferenced 19 bioclimatic features from Worldclim.org
(an ESRI supported website) as inputs as well. Worldclim.org derives variables from
monthly temperature and rainfall values to generate more biologically meaningful
factors. Bioclimatic variable across georeferenced locations represent interpolations of
observed data over the period 1960-1990. Worldclim.org is a climate data site the
Environmental Systems Research Institute (ESRI) uses, which one uses for ecological
modeling and GIS. ESRI develops ArcGIS software, which applies the science of using
maps to reveal deeper insight into data. Worldclim.org uses a set of global climate layers
(gridded climate data) with a spatial resolution of about one km2. The data are 30 arcseconds resolution downloaded in 30 × 30 degree footprints. I chose to include the rasters
for the 19 of the bioclimatic factors from Worldclim.org (Table 3).
44
Table 3. Bioclimatic Factors from (Worldclim.org, 2017)
Code
Bioclimatic Feature
BIO1
Annual Mean Temperature
BIO2
Mean Diurnal Range (Mean of monthly (max temp - min temp))
BIO3
Isothermality (BIO2/BIO7) (* 100)
BIO4
Temperature Seasonality (standard deviation *100)
BIO5
Max Temperature of Warmest Month
BIO6
Min Temperature of Coldest Month
BIO7
Temperature Annual Range (BIO5-BIO6)
BIO8
Mean Temperature of Wettest Quarter
BIO9
Mean Temperature of Driest Quarter
BIO10
Mean Temperature of Warmest Quarter
BIO11
Mean Temperature of Coldest Quarter
BIO12
Annual Precipitation
BIO13
Precipitation of Wettest Month
BIO14
Precipitation of Driest Month
BIO15
Precipitation Seasonality (Coefficient of Variation)
BIO16
Precipitation of Wettest Quarter
BIO17
Precipitation of Driest Quarter
BIO18
Precipitation of Warmest Quarter
BIO19
Precipitation of Coldest Quarter
I used Bioclimatic factors for environmental inputs in MaxEnt analysis. The
future data were drawn from IPCC5 climate projections from global climate models
45
(GCMs) for four representative concentration pathways (RCPs) or scenarios (Table 4).
These are the most recent GCM climate projections that researchers used in the Fifth
Assessment IPCC report (IPCC, 2014). I used Worldclim.org 1.4 downscaled and
calibrated (also known as bias corrected) using its software as my baseline ‘current’
climate. I used CCSM4 GCM, also known as CC for short for the year 2070. These files
are in 30 seconds’ resolution. The time periods for 2070 represent, respectively, averages
over 2061-2080.
Table 4. AR5 Global Warming Increase (°C) Projections for 2081-2100
Scenario
Mean and likely Range
RCP 2.6
1.0 (0.4 to 1.6)
RCP 4.5
1.4 (0.9 to 2.0)
RCP 6.0
1.3 (0.8 to 1.8)
RCP 8.5
2.0 (1.4 to 2.6)
Across all RCPs for 2081-2100, global mean temperature is projected to rise by
0.3 to 4.8°C (ICPP, 2015).
46
Chapter 3. Results
Current Spatial Distribution
The current distribution of the Cascade frog is based on the collective
georeferenced points from the database (Figure 3). Current spatial distribution shows
Cascade frogs spanning the Cascades Range and the Olympics. The more dots that are
clustered together, the more known occurrences are present in that region. There are high
occurrences of Cascade frogs in the Olympics, NOCA, MORA, the Cascades that run
through Oregon, and Lassen Peak.
MaxEnt Current Suitability Projection
The current suitability of Cascade frogs is shown with a scale bar, which explains
areas of green representing little to no Cascade frog occurrences and areas of red
representing areas of high suitability of Cascade frogs.
Model Performance
Omission and Predicted Area for 0 and Sensitivity vs. Specificity were generated
with the given parameters: random seed, random test percentage of 25%, one replicate,
and subsample replicated run type. These graphs explain the current suitability
projection. In both graphs, area under the curve’s (AUCs) for the Training and Test data
appear to be correct because 1) One they are very close to each other, meaning the test
data agree that the training data produced a good model; and 2) both the training and
testing data sets are substantially different from the random data set, meaning the model
carries significant weight (meaning) (Figures 5 and 6). The area under the curve in this
case is closer to 1, which indicates a better model performance and predicted probability
of occurrence using sample size (Young, Carter, and Evangelista, 2011).
47
Jackknife Analysis
Jackknife analysis compares the training data in isolation to the training gain with
all the variables included. Jackknife is represented in terms of regularized training gain
for 0 (Figure 7). The jackknife analysis shows that precipitation for the driest quarter
(BIO17), precipitation for the warmest quarter (BIO18), mean temperature of the driest
quarter (BIO9), and mean temperature of the warmest quarter (BIO10) had the most
influence on the projections for current suitability (Figure 4). The Jackknife shows the
training gain of each variable had the model been run in isolation.
MaxEnt 2070 Projections
The future scenarios, which consider RCP 2.6, 4.5, 6.0, and 8.5 for the year 2070
show where suitable areas where Cascade frogs might exist based off bioclimatic features
as inputs. Projections of Cascade frogs’ suitability for Scenario 2.6 shows high
suitability in the Olympics and MORA. There are slight decreases in suitability when
compared to Current Suitability (Figure 4) in these hot spots (Figure 8). Scenario 4.5
areas of high suitability in the Olympics, MORA, Oregon, and new high suitability in
northeast Washington and Lassen Peak (Figure 9). When compared to Figure 4.5, there
is new suitability in northeast Washington and northern California with decreases in the
Olympics and MORA. Scenario 6.0 shows areas of high suitability in the Olympics,
MORA, and Lassen Peak (Figure 10). When compared to Figure 4, there is a decreased
suitability in the Olympics and MORA. The high suitability in northern California is a
newly suitable area. Lastly, the Scenario 8.5 depicts low suitability in MORA, the
Olympics, Oregon, and California, with high suitability in northeast Washington (Figure
11).
48
Olympics
NOCA
MORA
Lassen Peak
Figure 3. Cascade Frog Occurrences in the Cascades Region
49
Figure 4. Current Suitability for Cascade Frogs
50
Figure 5. Omission and Predicted Area Omission and Predicted Area. for 0 for a random
seed, random test percentage of 25%, one replicate, and subsample replicated run type.
The output format is logistic with a 10% threshold percentage; this is for the Current
Suitability Projection.
51
Figure 6. Sensitivity vs. Specificity. Training and testing datasets for a random seed,
random test percentage of 25%, one replicate, and subsample replicated run type. The
output format is logistic. The training data has an Area under the curve (AUC) of 0.936,
an AUC of 0.925 for test data and an AUC of 0.5 for random prediction.
Figure 7. Jackknife of regularized training gain. The Jackknife shows the training gain of
each variable and the influence said variable had when MaxEnt was run in isolation, and
compares the training data ran in isolation to the training gain including all the variables.
This is for Current Suitability.
52
Figure 8. Scenario 2.6 for Year 2070
53
Figure 9. Scenario 4.5 for Year 2070
54
Figure 10. Scenario 6.0 for Year 2070
55
Figure 11. Scenario 8.5 for Year 2070
56
Chapter 4. Discussion and Conclusion
Specific changes in temperature and rainfall values from the associated RCP
climate change scenarios led to overall decreases in suitability as well as increases in
newly suitable areas of Cascade frogs. These monthly temperatures and rainfall factors
were projected to change based on future climate change patterns. These changes
influence an apparent decrease in amphibian populations from recent literature. My
findings support these results by demonstrating most of the future raster outputs result in
dwindling suitability for the Cascade frog. In each scenario, precipitation was reduced in
the spring and summer months and overall temperatures were warmer overall (Table 4).
The bioclimatic variables: BIO9 (Mean temperature of driest quarter), BIO10 (Mean
temperature of warmest quarter), BIO12 (Annual precipitation), BIO17 (Precipitation of
driest quarter) and BIO18 (Precipitation of warmest quarter) (Table 3) are the bioclimatic
factors that had the most influence (Figure 7) on the current suitability projection (Figure
4). Without data on population trends from past years, I cannot conclude that Cascade
frogs are declining. The purpose of this discussion is to fill some of the gaps from the
literature review with my findings. I will do this by going more in depth with how my
findings complement the literature. I then describe what my MaxEnt findings mean.
Lastly, I will provide suggestions about what I might model differently, what additions I
would make to the study, and lastly suggestions on what actions should take place.
Reasons for Decline
The Cascade frog’s Red List Category and Criteria is listed as Near Threatened
(IUCN, 2004), meaning there should be cause for concern that the frog may end up on the
57
Endangered Species List. Declining patterns of Cascade frogs are seen in northern
California from RCP 2.6 and 8.5 (Figures 8 and 11). The MaxEnt generated maps,
except for RCP 4.5 and 6.0 (Figures 9 and 10) support the past literature by recognizing
that there was once more suitable habitat in these regions compared to the Current
Suitability (Figure 4).
Ryan and colleagues (2014) interpreted fish introduction and climate change to
be the two largest factors resulting in amphibian decline in montane environments. Fish
introductions which begin in the late 1800s, was not addressed in my study, but was
certainly recognized as a factor as to why amphibians in the western U.S. are dwindling.
My study was similar Ryan’s in assessing future climate scenarios. I used MaxEnt while
Ryan et al. study utilized Variable Infiltration Capacity (VIC) as a form of wetland
projection modeling software. Ryan and colleagues (2014) also demonstrated how the
current climate change trajectory is already affecting summer water availability, resulting
in summers being warmer and having less precipitation.
The spring months when Cascade frogs breed is crucial because with the warmer
weather in winter, ice melts sooner, which triggers amphibians, like the Cascade frogs, to
breed. The warmer spring months and overall less amount of precipitation projected
from the bioclimatic features and climate change models suggests a pattern that Cascade
frogs overwintering and breeding patterns are being changed. This will in turn affect
their life cycles. Different areas in the Cascades include montane wetlands, and they are
sensitive to snowpack volume, snow runoff, direct precipitation and evapotranspiration.
MaxEnt Explanation
58
Projections of Cascade frog locations were determined from the environmental
inputs, (i.e. bioclimatic features). I collected data from four different databases and had
to convert different projections, which could result in inadvertent errors. Comparing the
current spatial suitability (Figure 4) to Climate Change Scenario 2.6 for 2070 (Figure 8)
shows that Cascade frogs will result in lower suitability in both MORA and the
Olympics. RCP 2.6 differed slightly from the Current Suitability projection, but was still
overall similar. Scenario 4.5 shows lower suitability in MORA and the Olympics, with
new higher suitability in northeast Washington, Oregon, and northern California. (Figure
9). Scenario 6.0 showed decreases in suitability in MORA and the Olympics, but a newly
suitable area became evident slightly in northeast Washington and in northern California
as well (Figure 10). Lastly, Scenario 8.5 shows the overall greatest decrease in suitability
throughout the Cascade frogs’ Range in the Cascades. There is also new suitability in
northeast Washington (Figure 11). This may be because the global temperature is
expected to rise the most for this scenario (Table 4). The area of higher bioclimatic
suitability by 2070 in northern California under RCP 4.5 (Figure 9) is characterized by
BIO14 (more precipitation in the driest month) and BIO18 (more precipitation in the
warmest quarter). It would make sense for northern California and northeastern
Washington to show higher suitability for RCP 4.5, 6.0, and 8.5 because California and
northeastern Washington have very hot and dry climates in the summer. More
precipitation in the summer (when it is the warmest and driest) may cause for this spikes
in suitability for the Cascade frog. It would be beneficial to implement a habitat mask
into MaxEnt to take into account where Cascade frogs actually exist based on habitat
features. Cascade frogs will most likely not have higher suitability in northeast
59
Washington and northern California in the future because these places will have droughts
and experience the impacts of climate change much more than more temperate regions
such as mid-western/western Washington and Oregon.
Potential Changes to the Analysis
In the future, I would expand my study to include waterbody data with and
without fish descriptors (Ryan et al. 2014). This may allow for an independent
opportunity to evaluate the interaction of fish presence and climate change may affect
Cascade frog occupancy in wetlands. I would also like to use object-based remote
sensing; VIC simulations of historical wetland dynamics may enable identifying
trajectories of change for different type of wetlands that can be integrated into the
modeling. For future study, I would like to incorporate other environmental features into
MaxEnt, such as distance from the water, snow cover, elevation, and slope are all
potentially important factors when considering the Cascade frog’s life history. I would
also like to run MaxEnt multiple times for different parameters and different percent
thresholds. This would change the Jackknife and Sensitivity analyses and assist me in
identifying the best models. A habitat mask would also be an important addition to this
MaxEnt analysis in order to project where Cascade frogs might occur in the future.
Future Ecological Implications for Wetland Biota
The Cascade frog, once abundant in the Southern Cascade Range and Klamath
Mountains of California, has become extremely rare in the Southern Cascades. However,
up to very recently, the species is still abundant in the Klamath Mountains (Pope et al.,
60
2014). Many potential reasons exist for the decline of Cascade frogs, I explored climate
change using MaxEnt to predict how Cascade frog habitat might change in the future.
The future projections, especially the 2070 projections, do not look good for the species
because highly suitable habitat will essentially vanish from the northern California where
it is currently abundant (specifically in the Klamath region of Oregon and Washington).
Climate change factors, such as decreased precipitation in summer months, earlier
breeding times due to earlier onset of snow melting and warmer temperatures may be
causing the frogs to decline as well as lose their habitat. Cascade frogs are adapted to
living at elevation in snowy regions. My trip to the Enchantments Basins allowed me to
take a closer look at the habitat where Cascade frogs may live. During my three trips
(once in June, July and August) allowed me to see just how fast the snow melted;
whether those habitats are ultimately colonized by Cascade or Columbia spotted frogs as
seasonal snow melts earlier remains to be seen. The research done over the last 100 years
shows that the indicator amphibian species are declining in number. Actions should be
taken to ensure more amphibians appear on the endangered species list. The focus of my
thesis was to examine current Cascade frog suitability and to project where their habitat
might in the future year 2050. Time still exists to protect this species as well as other
amphibians living in high mountain lakes. We also cannot ignore that the climate is
changing rapidly—there are earlier and faster rates of wetland drawdown, reduction in
water availability and an increased frequency of complete drying (Ryan et al., 2014).
Hope may still exist for other high mountain-dwelling amphibians if the proper steps are
taken for conservation. I concluded that Cascade frogs probably do not exist in the Lower
Lakes complex of the Enchantments Basin, although this is type of montane wetland
61
habitat where they can be found. Targeted fish removal programs and amphibian
monitoring programs should be implemented. Targeted fish removal programs may aid
in revitalizing amphibian populations, while amphibian monitoring programs can better
keep tabs on amphibian populations. This is not just about one frog, rather an entire
ecosystem and all the species living there.
62
Literature Cited
Adams, M., Schindler, D., and Bury, R. (2001). Association of amphibians with
attenuation of ultraviolet-B radiation in montane ponds. Oecologia, 128(4); 519525.
Adams, M. J., Chelgren, N. D., Reinitz D., Cole, R. A., Rachowicz, L. J., and Galvan, S.
(2010). Using occupancy models to understand the distribution of an amphibian
pathogen Batrachochytrium dendrobatidis. Ecological Applications 20:289–302.
Bahls, P. (1992). The status of fish populations and management of high mountain lakes
in the western United States. Northwest Science, 66(3):183–193.
Berger, L., Speare, R., Daszak, P., Green, D. E., Cunningham, A. A., Goggin, C. L.,
Slocombe, R., Ragan, M. A., Hyati, A. D., McDonald, K. R., Hines, H. B., Lips,
K. R., Marantelli, G., and Parkes, H. (1998). Chytridiomycosis causes amphibian
mortality associated with population declines in the rain forests of Australia and
Central America. Proceedings of the National Academy of Sciences of the United
States of America, 95(15), 9031-9036. doi:10.1073/pnas.95.15.903
Blaustein, A. R., and Wake, D. B. (1990). Declining amphibian populations: a global
phenomenon?. Trends in Ecology & Evolution 5(7):203–204.
Blaustein, A. R., Hoffman, P. D., Hokit, D. G., Kiesecker, J. M., Walls, S. C. and Hays,
J.B. (1994). UV repair and resistance to solar UV-B in amphibian eggs: a link to
population declines. Proceedings of the National Academy of Sciences
91(5):1791–1795.
Briggs, J.L. (1987). Breeding biology of the Cascades frog, with comparisons to R.
aurora and R. pretiosa. Copeia. 1987: 241–245.
63
Burkett, V., and Kusler, J. (2000). Climate change: potential impacts and interactions in
wetlands of the United States. Journal of the American Water Resources
Association 36(2):313–20.
Carey, C., and Alexander, M. A. (2003). Climate change and amphibian declines: Is there
a link? Diversity and Distributions 9(2):111–21.
Carpenter, S. R., and Kitchell, J. F. (1993). The trophic cascade in lakes. Cambridge
University Press, New York.
Chesson P. 2000. General theory of competitive coexistence in spatially-varying
environments. Theoretical Population Biology 58: 211–37.
De León, M. E., Vredenburg, V. T., and Piovia-Scott, J. (2017). Recent emergence of a
chytrid fungal pathogen in California Cascades Frogs (Rana cascadae). Ecohealth
14(1):155–161. https://doi.org/10.1007/s10393-016-1201-1
Drake, D. C., and Naiman, R. J. (2000). An evaluation of restoration efforts in fishless
lakes stocked with exotic trout. Conservation Biology 14(6):1807–1820.
Easterling, D. R., Karl, T. R., Gallo, K. P., Robinson, D. A., Trenberth, K. E., and Dai, A.
(2000). Observed climate variability and change of relevance to the biosphere.
Journal of Geophysical Research: Atmospheres, 105(D15), 20101–20114.
https://doi.org/10.1029/2000JD900166
Fellers, G. M., Pope, K. L., Stead, J. E., Koo, M. S., and Welsh, H. H. (2008). Turning
population trend monitoring into active conservation: can we save the Cascades
frog (Rana cascadae) in the Lassen region of California? Herpetological
Conservation Biology 3:28–39.
Garwood, J.M. (2009). Spatial ecology of the Cascades frog: identifying dispersal,
64
migration, and resource uses at multiple spatial scales. Arcata, CA: Humboldt
State University. 97 p. M.S. thesis
Elith, J., and Leathwick, J. (2009). Species Distribution Models: Ecological explanation
and prediction across space and time. Annual Review of Ecology, Evolution, and
Systematics (40):677–697.
https://doi.org/10.1146/annurev.ecolsys.110308.120159
Elith, J., Phillips, S. J., Hastie, T., Dudik, M., Chee, Y. E., and Yates, C. J. (2011). A
statistical explanation of MaxEnt for ecologists. Diversity and Distributions
17(1):43–57.
Esri: GIS Mapping Software, Spatial Data Analytics & Location Platform. (n.d.).
Retrieved May 26, 2017, from http://www.esri.com/
Farner, D. S., and Kezer, J. (1953). Notes on the amphibians and reptiles of Crater Lake
National Park. The American Midland Naturalist 50(2):448–462.
https://doi.org/10.2307/2422104
Fellers, G. M., and Drost, C. A. (1993). Disappearance of the Cascades frog Rana
cascadae at the southern end of its range, California, USA. Biological
Conservation 65(2):177–181. https://doi.org/10.1016/0006-3207(93)90447-9
Fellers, G. M., Pope, K. L., Stead, J. E., Koo, M. S., and Welsh, H. H. (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.
Fite, K. V., Blaustein, A., Bengston L., and Hewitt, H. E. 1998. Evidence of retinal light
damage in Rana cascadae: A declining amphibian species. Copeia 1998(4):906–
65
914.
Garcia, T.S., Romansic, J. M., and 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.
Google Maps. 2017. Retrieved April 11, 2017, from https://www.google.com/maps
Hamlet, A. F., Mote, P. W., Clark, M. P., and Lettenmaier, D. P. (2007). Twentiethcentury trends in runoff, evapotranspiration, and soil moisture in the western
United States. Journal of Climate 20(8):1468–1486.
Halabisky, M., Moskal, L. M., Gillespie, A., and Hannam, M. (2016). Reconstructing
semi-arid wetland surface water dynamics through spectral mixture analysis of a
time series of Landsat satellite images (1984–2011). Remote Sensing of
Environment 177:171–183. https://doi.org/10.1016/j.rse.2016.02.040
Hayes, M.P., and Jennings, M.R. (1986). Decline of ranid frog species in western North
America: are bullfrogs (Rana catesbeiana) responsible? Journal of Herpetology
20(4):490–509.
Hijmans, R.J., S.E. Cameron, J.L. Parra, P.G. Jones and A. Jarvis, 2005. Very high
resolution interpolated climate surfaces for global land areas. International Journal
of Climatology 25(15):1965–1978
Hoffman, R. L., Larson, G. L., and Brokes, B. J. (2003). Habitat segregation of
Ambystoma gracile and Ambystoma macrodactylum in mountain ponds and lakes,
Mount Rainier National Park, Washington, USA. Journal of Herpetology
37(1):24–34.
IPCC (Intergovernmental Panel on Climate Change). (2007). Climate change 2007:
66
impacts, adaptation and vulnerability. Contribution of Working Group II to the
Fourth Assessment IPCC Report of the Intergovernmental Panel on Climate
Change. Geneva, Switzerland.
IPCC (Intergovernmental Panel on Climate Change). (2015). Climate Change 2014:
Mitigation of Climate Change: Working Group III Contribution to the IPCC Fifth
Assessment Report. Cambridge: Cambridge University Press.
doi:10.1017/CBO9781107415416
IUCN, Conservation International, and NatureServe. (2004). Global Amphibian
Assessment. IUCN, Conservation International, and NatureServe, Washington DC
and Arlington, Virginia, USA.
IUCN (International Union for Conservation of Nature), Conservation International &
NatureServe. (2004). Rana cascadae. The IUCN Red List of Threatened Species.
Version 2016-3
Kats, L. B. and Ferrer, R. P. (2003). Alien predators and amphibian declines: review of
two decades of science and the transition to conservation. Diversity and
Distributions 9(2):99–110.
Knapp, R. A., Corn, P. S., and Schindler, D. E. (2001). The introduction of nonnative fish
into Wilderness Lakes: Good intentions, conflicting mandates, and unintended
consequences. Ecosystems 4(4):275–278.
Landres, P., Meyer, S., and Matthews, S. (2001). The Wilderness Act and fish stocking:
An overview of legislation, judicial interpretation, and agency implementation.
Ecosystems 4(4):287–295.
Lee, S.-Y., Ryan, M. E., Hamlet, A. F., Palen, W. J., Lawler, J. J., and Halabisky, M.
67
(2014). Projecting the hydrologic impacts of climate change on montane
wetlands. PLoS ONE 10(9):e.0136385. https://doi.org/10:
10.1371/journal.pone.0136385
Leopold, A. S., Cain, S. A., Cottam, C. M., Gabrielson, I. N., and Kimball, T. L. (1963).
Wildlife management in the national parks. Transactions of the North American
Wildlife Conference (28):28–45.
Mitsch, W.J., and Gosselink, J.G. (2015). Wetlands. Fifth edition. John Wiley & Sons,
Inc, New York: New York, USA. 747 pp.
NatureServe. 2017. (online at: http://www.natureserve.org/)
NPS (U.S. National Park Service). Homepage. (Online at:
https://www.nps.gov/index.htm, accessed 19 May 2017)
Nussbaum, R. A., Brodie, E. D., & Storm, R. M. (1983). Amphibians and reptiles of the
pacific northwest. Moscow, Idaho: University Press of Idaho.
Ouellet, M., Mikaelian, I., Pauli, B., Rodrigue, J., and Green, D. (2005). Historical
evidence of widespread chytrid infection in North American amphibian
populations. Conservation Biology 19(5):1431–1440.
Palen, W. J., Schindler, D. E., Adams, M. J., Pearl, C. A., Bury, R. B., and Diamond, S.
A. (2002). Optical characteristics of natural waters protect amphibians from UVB. B in the US Pacific Northwest. Ecology 83(11):2951–2957.
Pearl, C. A., and Adams, M. J. (2005). Rana cascadae Slater 1939: Cascade frog, Pp.
538-540. In; Lannoo, M (editors), Amphibian Declines: the Conservation Status
of United States’ Species. University of California Press, Berkeley, California,
USA.
68
Petranka, J. W. (1983). Fish predation: A factor affecting the spatial distribution of a
stream-breeding salamander. Copeia 1983(3):624–628.
Pilliod, D. S., and Peterson, C. R. (2001). Local and landscape effects of introduced trout
on amphibians in historically fishless watersheds. Ecosystems 4(4):322–333.
Pister, E. P. (2001). Wilderness fish stocking: History and perspective. Ecosystems
4(4):279–286.
Piovia-Scott J, Pope KL, Lawler SP, Cole EM, Foley JE. (2011). Factors related to the
distribution and prevalence of the fungal pathogen Batrachochytrium
dendrobatidis in Rana cascadae and other amphibians in the Klamath Mountains.
Biological Conservation 144:2913–2921.
Piovia-Scott, J., Pope, K., Worth S.J., S., Rosenblum, E. B., Poorten, T., Refsnider, J.,
Rollins-Smith, L. A., Reinert, L. K., Wells, H. L., Rejmanek, D., and Lawler, S.
(2015). Correlates of virulence in a frog-killing fungal pathogen: evidence from a
California amphibian decline. ISME Journal: Multidisciplinary Journal of
Microbial Ecology 9(7):1570–1578. doi:10.1038/ismej.2014.241
Pope, K. L., Brown, C., Hayes, M., Green, G., and Macfarlane, D. (2014). Cascades Frog
Conservation Assessment. Pacific Southwest Research Station, United States
Forest Service: Albany, CA, USA.
Pope, K. L., and Long, J. W. (2014). Chapter 6.4—Lakes: Recent Research and
Restoration Strategies, Pp. 373–390. In: Science Synthesis to Support
Socioecological Resilience in the Sierra Nevada and Southern Cascade Range. US
Department of Agriculture, Forest Service, Pacific Southwest Research Station,
General Technical Report GTR-247.
69
Ryan, M. E., Palen, M.J., Adams, M. J., and Rochefort, R. M. (2014). Amphibians in the
climate vice: loss and restoration of resilience of montane wetland ecosystems in
the western US. Frontiers in Ecology and the Environment 12(4):232–240.
Semlitsch, R. D. (1988). Allopatric distribution of two salamanders: effects of fish
predation and competitive interactions. Copeia 1988(2):290–298. doi:
10.2307/1445868
Snodgrass, J. W., Komoroski, M. J., Bryan, A. L., Jr., and Burger, J. (2000).
Relationships among isolated wetland size, hydroperiod, and amphibian species
richness: implications for wetland regulation. Conservation Biology 14(2):414–
419.
Stebbins, R. C., National Audubon Society, & National Wildlife Federation. (1985). A
field guide to western reptiles and amphibians: Field marks of all species in
western north America, including Baja California. Boston: Houghton Mifflin.
Stebbins, R. C. (2003). A field guide to western reptiles and amphibians (3rd ed.).
Boston: Houghton Mifflin.
Stuart, S. N., Chanson, J. S., Cox, N. A., Young, B. E., Rodrigues, A. S., Fischman, D.
L., and Waller, R. W. (2004). Status and trends of amphibian declines and
extinctions worldwide. Science 306(5702):1783–1786.
Tyler, T., Liss, W. J., Ganio, L. M., Larson, G. L., Hoffman, R., Deimling, E., &
Lomnicky, G. (1998a). Interaction between introduced trout and larval
salamanders (Ambystoma macrodactylum) in high-elevation lakes. Conservation
Biology 12(1):94–105.
Tyler, T. J., Liss, W. J., Hoffman, R. L., & Ganio, L. M. (1998b). Experimental analysis
70
of trout effects on survival, growth, and habitat use of two species of
ambystomatid salamanders. Journal of Herpetology 32(3):345–349.
https://doi.org/10.2307/1565448
US EPA (U.S. Environmental Protection Agency). Ecoregions of Western Washington
and Oregon. (Online at:
ftp://newftp.epa.gov/EPADataCommons/ORD/Ecoregions/reg10/ORWAFront90.
pdf, accessed 19 May 2017)
US EPA, (2015, December 23). Overview of Greenhouse Gases [Overviews and
Factsheets]. Retrieved June 7, 2017, from
https://www.epa.gov/ghgemissions/overview-greenhouse-gases
Vredenburg, V. T., and Wake, D. B. (2004). Reversing introduced species effects:
Experimental removal of introduced fish leads to rapid recovery of a declining
frog. Proceedings of the National Academy of Sciences 101(20):7646–7650.
Welborn, G.A., Skelly, D.K., and Werner, E.E. (1996). Mechanisms creating community
structure across a freshwater habitat gradient. Annual Review of Ecology and
Systematics, (27):337–363.
Whittaker RJ, Willis KJ, and Field R. 2001. Scale and species richness: towards a
general, hierarchical theory of species diversity. Journal of Biogeography 28:
453–70.
WorldClim - Global Climate Data | Free climate data for ecological modeling and GIS.
(n.d.). Retrieved May 26, 2017, from http://worldclim.org/
Young, N., Carter, L., and Evangelista, P. (2011). A MaxEnt Model v3.3.3e Tutorial
71
(ArcGIS v10). Natural Resource Ecology Laboratory, Fort Collins, Colorado,
USA.
72
Appendix
1. Databases included information from a group of different herpetologists and researchers
showing the locations where Cascade frogs were documented, the date collected,
institution, collection ID, institution code, preservation status, the age/class, sex,
catalogue number, recorder name, count, occurrence marks, latitude and longitude World
Geodetic System 1984 (WGS84), elevation, the museum or institute which identified the
specimen, the date collected, and the nomenclatural code. The locations varied
substantially and included regions from all over the Cascades and the Olympics. The
Cascade frog data from the MORA headquarters, the Olympic National Park
headquarters, and United States Geological Survey (USGS) contained uniform factors
such as record ID, source, source contact, program, survey type, the date collected, the
park and its associated code, the site name, elevation, life stage, count, Universal
Transverse Mercator (UTM) zone 10N location data and notes regarding township,
section, and range.
2. These points included (48.71866, -122.352255), (47.197465, -122.534167), (45.407626, 122.570384), (45.467161, -122.790306), (42.760341, -123.701202), (42.02212, 123.459195). The first point was in Bellingham, WA, right off interstate five. With
further speculation, this point had an elevation of roughly 2500 feet. This appears to be
relatively low elevation for Cascade frogs. W.C. Brown identified the next frog(s)
documented in Pierce County, off Chambers Creek at an elevation of roughly 500 feet.
The low elevation and surrounding neighborhood- from the Google maps timeline photo
73
from 1990 displays that this is not ideal habitat for Cascade frogs (GoogleMaps, 2017).
The next point lies in Clackamas County, Oregon with an elevation of 170 feet and
appears to be right in the middle of the street. The following point was in Beaverton,
Tualatin Hills Park and Recreation District, Fanno Creek Park, Fanno Creek Pond,
Oregon. The next frogs existed in Douglas County, Oregon in or near the Kelsay River
with an elevation of 4146 feet. The point was identified in Josephine County, Bolan
Lake, Oregon at an elevation of 5450 feet. I removed this point because it occurred
outside of the Cascades, where Cascade frogs are normally found. The final count and
sample size for the data set was roughly 1,700 points.
74