ArticlePDF Available

Climate-associated decline of body condition in a fossorial salamander


Temperate ectotherms have responded to recent environmental change, likely due to the direct and indirect effects of temperature on key life cycle events. Yet, a sub- stantial number of ectotherms are fossorial, spending the vast majority of their lives in subterranean microhabitats that are assumed to be buffered against environmen- tal change. Here, we examine whether seasonal climatic conditions influence body condition (a measure of general health and vigor), reproductive output, and breed- ing phenology in a northern population of fossorial salamander (Spotted Salamander, Ambystoma maculatum). We found that breeding body condition declined over a 12- year monitoring period (2008–2019) with warmer summer and autumn temperatures at least partly responsible for the observed decline in body condition. Our findings are consistent with the hypothesis that elevated metabolism drives the negative associa- tion between temperature and condition. Population-level reproduction, assessed via egg mass counts, showed high interannual variation and was weakly influenced by autumn temperatures. Salamander breeding phenology was strongly correlated with lake ice melt but showed no long-term temporal trend (1986–2019). Climatic warm- ing in the region, which has been and is forecasted to be strongest in the summer and autumn, is predicted to lead to a 5%–27% decline in salamander body condition under realistic near-future climate scenarios. Although the subterranean environment offers a thermal buffer, the observed decline in condition and relatively strong ef- fect of summer temperature on body condition suggest that fossorial salamanders are sensitive to the effects of a warming climate. Given the diversity of fossorial taxa, heightened attention to the vulnerability of subterranean microhabitat refugia and their inhabitants is warranted amid global climatic change.
Letter to the Editor
1703 Can carbon storage in West Antarctic fjords have an impact on climate change, following glacier
William R. Hunter
e7 Response to concerns regarding the role of Antarctic blue carbon habitats in a global context
Nadescha Zwerschke, David K. A. Barnes and Chester J. Sands
1705 Socio-ecological drivers of multiple zoonotic hazards in highly urbanized cities
Matthew A. Combs, Pallavi A. Kache, Meredith C. VanAcker, Nichar Gregory, Laura D. Plimpton,
Danielle M. Tufts, Maria P. Fernandez and Maria A. Diuk-Wasser
Primary Research Articles
1725 Climate-associated decline of body condition in a fossorial salamander
Patrick D. Moldowan, Glenn J. Tattersall and Njal Rollinson
1740 Rapid evolution generates synergism between multiple stressors: Linking theory and an
evolution experiment
James A. Orr, Pepijn Luijckx, Jean-François Arnoldi, Andrew L. Jackson and Jeremy J. Piggott
1753 Projecting coral responses to intensifying marine heatwaves under ocean acidification
Shannon G. Klein, Nathan R. Geraldi, Andrea Anton, Sebastian Schmidt-Roach, Maren Ziegler,
Maha J. Cziesielski, Cecilia Martin, Nils Rädecker, Thomas L. Frölicher, Peter J. Mumby,
John M. Pandolfi, David J. Suggett, Christian R. Voolstra, Manuel Aranda and Carlos. M. Duarte
1766 Putting the Pacific marine heatwave into perspective: The response of larval fish off southern
California to unprecedented warming in 2014–2016 relative to the previous 65 years
Andrew R. Thompson, Noah J. Ben-Aderet, Noelle M. Bowlin, Dovi Kacev, Rasmus Swalethorp and
William Watson
1786 Seasonal to decadal spatiotemporal variations of the global ocean carbon sink
Min Zhang, Yangyan Cheng, Ying Bao, Chang Zhao, Gang Wang, Yuanling Zhang, Zhenya Song,
Zhaohua Wu and Fangli Qiao
1798 Decadal trend of plankton community change and habitat shoaling in the Arctic gateway
recorded by planktonic foraminifera
Mattia Greco, Kirstin Werner, Katarzyna Zamelczyk, Tine L. Rasmussen and Michal Kucera
contents continued on inside back cover
Global Change Biology is indexed by ISI Alerting Services, Current Contents®
Agriculture Biology
& Environmental Sciences, Science Citation Index®, Research Alert®, CABS, CABSonline,
EMBASE, Elsevier Biobase, Elsevier GeoAbstracts and Scopus.
Information on this journal can be accessed at
This journal is available online at Wiley Online Library.
Visit to search the articles
and register for table of contents e-mail alerts.
MARCH 2022
ISSN 1354
VOL UM E 28 | N UM BE R 5 | M AR CH 2 02 2
Previous Total Pages: 316 Spine Width: 9.5 mm Paper type: UPM Gloss 70 Current Total Pages: 294 Spine Width: 14.9 mm Paper type: UPM Gloss 70 gsm
Slimming salamanders in a warming climateSlimming salamanders in a warming climate
California drought and ecosystem dynamicsCalifornia drought and ecosystem dynamics
Climate change and boreal forests site indexClimate change and boreal forests site index
Ecological drivers of urban zoonotic hazardsEcological drivers of urban zoonotic hazards
GCB_v28_i5_Cover.indd 1GCB_v28_i5_Cover.indd 1 1/28/2022 7:48:26 AM1/28/2022 7:48:26 AM
Glob Change Biol. 2022;28:1725–1739.
1725© 2021 John Wiley & Sons Ltd.
Global climate change is unequivocal, and rates of observed and
predicted temperature increases are greater at higher latitudes
(Intergovernmental Panel on Climate Change, 2014). High- and
mid- latitude regions have experienced rapid environmental change
over the past half- century, including increases in mean air and sur-
face temperature, extreme high temperature events, decreases in
the number of frost- free days, and decreases in ice cover (Bar tolai
et al., 2015; Cohen et al., 2014). The short post- glacial history of
the Northern hemisphere mid- latitudes, combined with regional
geology, climate, and varied habitat types, makes this region the
Received: 21 March 2021 
Accepted: 27 April 2021
DOI: 10.1111/gcb.15766
Climate- associated decline of body condition in a fossorial
Patrick D. Moldowan1,2,3 | Glenn J. Tat t ers a ll4| Njal Rollinson1,2
1Depar tment of Ecology a nd Evolutionary
Biolog y, University of Toronto, Toronto,
ON, Canada
2School of the Environment , Univer sity of
Toronto, Toronto, ON, Canad a
3Algonquin Wildlife Research Station,
Whitney, ON, Canada
4Depar tment of B iological Scie nces, Brock
University, St. Catharines, ON , Canada
Patrick D. Moldowa n, Depa rtme nt
of Ecology and Evolut ionar y Biology,
University of Toronto, Toronto, ON M5S
3B2, Canada.
Email: pat rick.
Funding information
This research was financially supported
by Ontar io Parks; Town of Huntsville
Environmental Research Bursa ry,
graduate student research award from the
Centre fo r Globa l Change Science at t he
University of Toronto, Beatrice and Ar thur
Minden G raduate Research Fellowship at
the Unive rsity of Toronto School of the
Environment, an d a CGS- D award from
the Natural Scie nces and Enginee ring
Research Council of Canada (NSERC) to
PDM; NSERC Discover y Grant (NSERC
RGPIN- 05814) to GJT; Connaught Early
Career Award and N SERC Discovery
Grant (RG PIN- 2016- 06469) to NR; and
the Research Excursion Program f rom the
Faculty of Arts and Science, Universit y of
Tor on to.
Temperate ectotherms have responded to recent environmental change, likely due
to the direct and indirect effects of temperature on key life cycle events. Yet, a sub-
stantial number of ectotherms are fossorial, spending the vast majority of their lives
in subterranean microhabitats that are assumed to be buffered against environmen-
tal change. Here, we examine whether seasonal climatic conditions influence body
condition (a measure of general health and vigor), reproductive output, and breed-
ing phenology in a northern population of fossorial salamander (Spotted Salamander,
Ambystoma maculatum). We found that breeding body condition declined over a 12-
year monitoring period (2008– 2019) with warmer summer and autumn temperatures
at least partly responsible for the observed decline in body condition. Our findings are
consistent with the hypothesis that elevated metabolism drives the negative associa-
tion between temperature and condition. Population- level reproduction, assessed via
egg mass counts, showed high interannual variation and was weakly influenced by
autumn temperatures. Salamander breeding phenology was strongly correlated with
lake ice melt but showed no long- term temporal trend (1986– 2019). Climatic warm-
ing in the region, which has been and is forecasted to be strongest in the summer
and autumn, is predicted to lead to a 5%– 27% decline in salamander body condition
under realistic near- future climate scenarios. Although the subterranean environment
offers a thermal buffer, the obser ved decline in condition and relatively strong ef-
fect of summer temperature on body condition suggest that fossorial salamanders
are sensitive to the effects of a warming climate. Given the diversity of fossorial taxa,
heightened attention to the vulnerability of subterranean microhabitat refugia and
their inhabitants is warranted amid global climatic change.
Ambystoma maculatum, body condition, climate change, ecophysiology, metabolic rate,
microhabitat, phenolog y, reproduction
northern range edge of many cold- adapted ectotherms (Lesbarrères
et al., 2014; Seburn & Bishop, 2007; Sunday et al., 2011). Some of
these temperate, cold- adapted species may respond to climate
change with latitudinal and altitudinal range shifts, and indeed,
such shifts are increasingly documented among terrestrial species
(Chen et al., 2011; Hughes, 2000; Walther et al., 2002). However,
not all taxonomic groups are highly mobile and capable of rapid or
long- distance dispersal in response to unfavorable environment al
conditions. For instance, the slow dispersal of many ectothermic
vertebrates, including some amphibians, relative to the accelerating
rate of climate change imposes hard limitations on geographic range
shifts as a response to climate change (Blaustein et al., 2010).
Amphibians— especially high latitude, mid latitude, and northern
range edge populations— are expected to respond strongly to climate
change due to their specific reproduc tive requirements, complex
life histories, and often short breeding period (Daszak et al., 2005;
Fitzpatrick et al., 2020). Climate- induced shifts in reproductive tim-
ing and biology are very well reported for anurans (Beebee, 1995a,
1995b; Gibbs & Breisch, 2001; Klaus & Lougheed, 2013; McCarty,
2001; Sheridan et al., 2018; Terhivuo, 1988; Todd et al., 2011;
Tryjanowski et al., 2006; Walpole et al., 2012). Of broader concern,
however, is that climatic change is negatively impacting amphibian
populations by reducing net resource acquisition rates of individu-
als and/or inducing physiological (temperature and/or water) stress,
leading to lower body condition and curtailed reproductive efforts.
Decreases in anuran body size, body condition, and fecundity, for
example, have been attributed to increases in temperatures or proxy
measures thereof (Benard, 2015; Lannoo & Stiles, 2017; Reading,
2007; Sheridan et al., 2018). Amphibian responses to a changing en-
vironment are complex and have been shown to vary in a number of
ways. For example, warmer winters were correlated with an increase
in body size in two species of European ranid frogs (Tryjanowski
et al., 2006). Annual drought severity, including the interplay of tem-
perature and precipitation, was correlated with decline in Crawfish
Frog (Lithobates areolatus) body condition and fecundity (Lannoo
& Stiles, 2017). Regional changes in frost- free days and precipita-
tion across Nor th America explain sex- specific responses in body
size, including body size increases and decreases, in the Wood Frog
(Lithobates sylvaticus) over the past century (Sheridan et al., 2018).
Climatic warming and drying have been implicated in reducing body
size of plethodontid salamanders in eastern Nor th America (Caruso
et al., 2014; but see Connette et al., 2015). In sum, there is strong
evidence that the breeding biology and life cycles of amphibians,
among other taxa, are responding to rapid climate change.
The cryptic and relatively fossorial nature of many amphibians
has generally impeded research into relationships between climate
change and biology (but see Beebee, 1995a, 1995b; Kirk et al., 2019;
Kusano & Inoue, 2008; Todd et al., 2011). Yet, fossorial lifest yles are
common in amphibians: an estimated 11% of known amphibian spe-
cies (723 spp.) associate with underground environments, including
7% of frogs (397 spp. of Anura), 23% of salamanders (145 spp. of
Caudata), and 97% of caecilians (181 spp. of Gymnophiona; Oliveira
et al., 2017). Although non- fossorial amphibians are likely to bear
the brunt of rapid warming, fossorial species have been largely over-
looked in climate change studies, perhaps owing to the untested
assumption that they are safeguarded by living in a thermally buf f-
ered underground environment. Subterranean- dwelling Ambystoma
salamanders, for instance, demonstrate strong dependence on
temperature and precipitation to signal aboveground migration and
breeding, but bouts of surface activity are shor t (Blanchard, 1930;
Douglas, 1979; Sexton et al., 1990). Thus, subterranean microhab-
itat use may mediate exposure to thermal and moisture extremes,
thereby allowing individuals to persist within their physiological
tolerances in otherwise inhospitable ambient environments (Moore
et al., 2018; Schef fers et al., 2013; Schef fers, Edwards, et al., 2014).
Fossorial lifestyles in amphibians are, in fact, broadly associated with
temperate regions characterized by climatic variabilit y, high tem-
perature variation, and/or low precipitation (Oliveira & Scheffers,
2019), suggesting that below- ground microhabitat use mediates cli-
matic extremes, allowing persistence in challenging above- ground
If fossorial salamanders demonstrate sensitivity to climatic
change, then there could be broad repercussions, as salamanders
contribute important ecological roles, including trophic structuring,
leaf litter and soil nutrient cycling, and carbon sequestration (Best
& Welsh, 2014; Davic & Welsh, 2004). The goal of this study is to
explore the thermal sensitivity of body condition (i.e., general health,
vigor) and key lifecycle events of a fossorial amphibian, leveraging
a long- term dataset on the Spotted Salamander (Ambystoma macu-
latum; Shaw, 1802). Since 1986, Spotted Salamander reproductive
biology and breeding phenology has been studied non- continuously
in their northern distribution in Algonquin Provincial Park, Ontario,
Canada. Concurrent with global and regional climate change,
changes in seasonal temperature and precipitation have occurred in
Algonquin Provincial Park over the past half- centur y (Lemieux et al.,
2007; Ridgway et al., 2018; Waite & Strickland, 2006). Our first
objective was to estimate the relationship between annual climate
(temperature, precipitation, and ice conditions) and body condition,
population- level reproductive output, and breeding phenology for
the Spotted Salamander. Our second objective was to predict trends
in salamander breeding body condition under realistic near future
(2050– 2100) climate scenarios. Our third objective was to esti-
mate longitudinal trends in the reproductive biology of the Spotted
2.1  |  Study species and area
The Spotted Salamander (A. maculatum) is a medium- bodied mem-
ber of the mole salamander group and is widespread in the forests of
eastern North America. The species is a not true burrower (Semlitsch,
1983), instead relying on the burrows created by small mammals (e.g.,
Blanaria, Peromyscus, Tamia s ; Faccio, 2003; Kleeberger & Werner,
1983; Madison, 1997; Montieth & Paton, 2006; Windmiller, 1996).
Brief periods of aboveground activity, limited to late winter/early
spring breeding and autumn movement, are strongly cued by weather,
namely rainfall and spring thaw at northern latitudes (Baldauf, 1952;
Madison, 1997; Sexton et al., 1990). Spotted Salamanders are not
freeze- tolerant and overwinter below the frost line (Madison, 1997).
Breeding occurs in shallow water and females can lay multiple egg
masses during a single spring reproductive period (Petranka, 2010).
Long- term observations suggest that females in our study population
do not breed annually and/or are highly transient at the breeding site
(P.D. Moldowan, unpublished data).
Bat Lake in western Algonquin Provincial Park is a naturally
acidic (pH 4.6) and fishless kettle bog with permanent water (3.4 ha
open- water surface area; mean depth: 4.5 m; maximum depth:
8.3 m; Hoeniger, 1986) in the low latitude boreal wetland region
(Zoltai et al., 1988) of Ontario, Canada. The region is characterized
by a relatively high elevation (300– 600 m above sea level) that is
wetter and cooler than the surrounding landscape (150 m). The ac-
tive season for Spotted Salamanders in the region, defined as the
non- freezing period during which salamanders are post- breeding
and presumed feeding, spans approximately ordinal day 140– 288
(May 20 October 15). This northernly location within the Spotted
Salamander geographic range, local elevation, and prevailing climatic
means that the Bat Lake salamander population effectively lives
near a climatic range edge.
2.2  |  Data collection
The breeding biology of Spotted Salamander was studied non-
continuously from 1986 to 2019 at Bat Lake. Prior to 2008, breed-
ing phenology data (first egg- lay date) were sporadically collected
by various observers (1986, R .G. Tozer and D. Strickland in Oldham
& Weller, 1989; 1992– 1993, D.C. Cunnington, unpublished; 2004–
2007 author G.J. Tattersall). Between 2008 and 2019, the capture–
mark– recapture study and reproductive monitoring were conducted
annually. Each year (N = 12; 2008– 2019) we conducted salamander
trapping using non- baited aquatic funnel traps and egg mass counts,
commencing at lake ice- of f and continuing daily throughout the
duration of the 3- to 4- week spring breeding period. Each year we
distributed traps equidistantly (spaced 8– 10 m apart) around the pe-
rimeter of the lake at fixed locations. Captured salamanders were
measured for snout– vent length (SVL) and body mass. Female repro-
ductive status (gravid or spent) was assessed by visually inspecting
the venter for abdominal distension and visible ova. Individuals were
marked with a single toe clip, digitally photographed, cataloged in
a photo identification database, and recaptures were identified by
their unique spot pattern using Interactive Individual Identification
System Classic software (v. 4.0.2., den Hartog & Reijns, 2020; van
Tienhoven et al., 2007). Trapping ended following three consecutive
days of no salamander captures, coinciding with the mass emigration
of adults out of the lake at the end of the breeding season, or until
water temperature at submerged trap depth reached 10°C (as to not
risk severe hypoxia in submerged salamanders).
We conducted whole lake egg mass counts, a measure of
population- level reproductive output, throughout the annual sam-
pling period by paddling the vegetated perimeter of the lake and
counting egg masses with the aid of polarized sunglasses. Egg mass
counts ceased when egg mass abundance plateaued and/or sal-
amander captures declined sharply in traps, signifying the end of
2.3  |  Hypothesis testing and statistics
We were interested in whether climatic conditions prior to breeding
influence female and male breeding body condition, population- level
reproductive output (egg mass abundance), and spring breeding phe-
nology. We therefore identified plausible environmental predictor
variables related to the breeding biology and life history of temper-
ate amphibians (Table S1), including environmental conditions from the
preceding summer, autumn, winter, and the contemporary spring pe-
riod. These predictors included mean temperature (°C) and total pre-
cipitation (mm) of the summer (July and August), autumn (September
and October), pre- winter (November), and spring periods (April and/
or May), as well as the number of days from lake surface freeze to
thaw (ice- on duration, a proxy for the length of winter and hence sala-
mander dormancy), and the ordinal date of majority lake ice- off (i.e.,
first breeding opportunity for pond- breeding salamanders; Table S1).
Climate data were retrieved from the Algonquin Provincial Park East
Gate climate station (45°32′N, 78°16′W; Environment and Climate
Change Canada: www.clima te.weath located approximately
30 km east of Bat Lake. Annual ice- off dates were known for Bat Lake,
although annual ice- on dates were not available. We used ice- on date
and ice- on duration from Lake of Two Rivers (2007 2019), a long- term
monitoring site within two kilometers of Bat Lake (R.G. Tozer, unpub-
lished data), as a proxy for Bat Lake. Annual ice- off dates for Bat Lake
and Lake of Two Rivers for the period of data overlap (2008– 2019)
were highly correlated (Pearson correlation: r = 0.93, p < 0.0001).
Statistical analyses were completed in R statistical software
(version 3.6.3, R Development Core Team, 2020). We used an
information- theoretic approach (Burnham & Anderson, 2002;
Burnham et al., 2011; Nakagawa & Schielzeth., 2013) and developed
an ecologically relevant set of candidate models for each response
variable, with each unique candidate set addressing one of the three
major response variables in our study (body condition, reproductive
output, and breeding phenology); below, we broadly describe these
hypotheses for each variable (Tables S2– S4). All models in each can-
didate set were fit with maximum likelihood then ranked according
to the second- order Akaike's information criterion (AICc; Burnham
& Anderson, 2002). The best- ranked models in each set (
Burnham & Anderson, 2002) were carried forward for model aver-
aging (using subset or “natural averaging,” as opposed to full- model
averaging; Burnham & Anderson, 2002; Symonds & Moussalli, 2011)
using the R package MuMIn (version 1.43.17; Bartoń, 2020).
We tested all response variables for temporal (i.e., year over
year) autocorrelation because salamanders may skip years and
accumulate energ y for reproduc tion over multiple seasons (e.g.,
Bulahova & Berman, 2017; Peacock & Nussbaum, 1973; Yartsev
& Kuranova, 2015). We examined the predictor variables of each
model for multicollinearity by calculating variance inflation factors
(VIF) using the R package car (version 3.0.10, Fox & Weisberg, 2019).
All predictors were retained in the models unless stated otherwise.
2.4  |  Body condition
Salamander body condition at the time of breeding (late April through
mid- May) reflects their state following overwintering, which at the
latitude and elevation of our study site means that salamanders have
presumably not fed for the preceding 5– 6 months, from the onset
of freezing temperatures (November) of the preceding year. Female
and male breeding body condition for the period of 2008– 2019
was calculated as the scaled mass index (SMI; Peig & Green, 2009).
The SMI standardizes the mass of an individual to the mean body
size of all individuals in the population (female and male datasets
were treated independently) while accounting for the SVL– mass al-
lometric relationship (Peig & Green, 2009). The SMI accurately re-
flects the energy stores (fat) of adult newt s and other amphibians
(MacCracken & Stebbings, 2012) and output values are interpretable
in the same unit of measure as input dat a (Peig & Green, 2009). To
calculate SMI, we plotted ln- transformed mass and ln- transformed
SVL, then fitted a line of best fit using standardized major axis re-
gression (Peig & Green, 2009). The slope of this best fit line was
carried forward as the scaling exponent for the SMI calculation. We
used mean SVL (L0, as per Peig & Green, 2009) for females and males
in calculating SMI for each sex.
We reasoned that the summer and/or autumn temperature in the
preceding year may affect fossorial salamander biology by altering
ground temperatures, and thus af fecting metabolic rate and/or the
extent of thermal stress of salamanders while underground. In the
absence of a compensator y response in resource acquisition by sal-
amanders, this would result in a negative relationship between tem-
perature and body condition, and no expected relationship between
precipitation and body condition. We further reasoned that summer
and/or autumn temperature from the preceding year may also affect
forest productivity, thereby altering food availability when salaman-
ders forage aboveground. If productivity and resource acquisition
outweighed a metabolic or stress response, then this would manifest
as a positive relationship between body condition and temperature;
furthermore, under the productivity hypothesis, we also expect a
positive relationship between body condition and precipitation.
Next, we reasoned that the pre- winter environment (November
temperature and precipitation from the preceding year) would be
related to annual variation in the length of the foraging period. We
predicted that November temperature of the previous year would
be positively related to body condition, and that November precip-
itation would be negatively related to condition, as precipitation in
November is overwhelmingly snow (climate data did not differen-
tiate between snow and rain). We predicted that ice- on duration
(i.e., winter duration) would be negatively related to body condition,
reasoning that longer winter periods represent a longer non- feeding
dormancy period and possibly a shorter pre- November foraging pe-
riod. We reasoned that warm temperature and high rainfall during
late winter/early spring of the contemporaneous breeding year
would result in earlier thaw, an earlier onset of salamander breeding,
and shor tened overwinter fasting period. Therefore, we predicted a
positive relationship between body condition and contemporaneous
April temperature and precipitation.
Finally, we included ordinal date (i.e., day of year) of capture as
a covariate in all models, as salamanders do not feed during breed-
ing and are expending energy continuously. We expected that body
condition would decline with ordinal capture date. Reproductive
status was included as a grouping variable (yes, gravid; no, spent)
in all models of female body condition, as some females were cap-
tured before laying and others post- laying. No interactions among
variables were included in any model.
Ultimately, these hypotheses and their combinations resulted in
a set of 17 ecologically plausible models to explain each female and
male body condition from 2009 to 2019 (Tables S1 and S2; body
condition data from 2008 were retained for time series analyses but
excluded from linear mixed- effect models because relevant precip-
itation data were not available); female and male data were fit sep-
arately. Linear mixed- effects models were fit using the R Package
lme4 (version 1.1.23; Bates et al., 2015). Both body condition model
sets included individual ID, unique day (unique date of capture; a
unique day assignment across all years in the dataset that addresses
error unique to a sampling day), and year as random effects to ac-
count for the non- independence of these factors on measures of
condition (Table S2). Data standardizations were used to improve
the biological interpretability of regression coefficients and relative
importance of predictors: The response variable (SMI) was scaled
(centered then divided by the standard deviation), and all predictor
(fixed effect) variables were centered prior to modeling (Schielzeth,
2.5  |  Body condition: Long- term temporal and
environmental relationships
We investigated whether a temporal change in body condition oc-
curred over the study period (2008– 2019). We used a separate
model for each sex with random effects of year, individual ID, unique
day, and fixed effects of ordinal date, year, and for females, repro-
ductive status; below we refer to this model as the “Base Model”
for each sex, as the model contains the fixed effect of year but no
environmental covariates (Table 1). We inferred a temporal change
in body condition if confidence intervals on the fixed ef fect of year
(βYear ) did not overlap zero.
Next, we explored which environmental covariates, if any, were
responsible for driving temporal changes in body condition. Drawing
from our model selection procedure with the body condition model
set, we sequentially substituted each environmental covariate
identified as important (i.e., where model- averaged confidence in-
tervals did not overlap zero; Figure 1) into the base model, separately
for each sex ( Table 1). For each substitution, we qualitatively exam-
ined how the parameter estimate for the fixed effect of year (βYea r)
changed relative to that of the base model. Large changes in βYear r el-
ative to the base model suggested that an environmental covariate
played a role in driving changes in body condition over time. Because
we were interested only in parameter estimates, all models were fit
with restricted maximum likelihood.
2.6  |  Body condition: Forecasting body condition
under climate change
Our analyses suggested that increased temperature, particularly
summer temperature, was associated with a temporal decline in
salamander body condition, and that this association may be damp-
ened by increases in summer precipitation (see Section 3; Table 1).
Drawing from this result, we used model- averaged parameter esti-
mates to forecast male body condition under three realistic near-
future (2050– 2100) climate scenarios for Algonquin Provincial Park
(Lemieux et al., 2007; Ridgway et al., 2018): up to 24% increase in
mean summer precipitation and up to 3°C increase in each mean
summer temperature, mean autumn temperature, and concurrent
mean summer and autumn temperature. For simplicity, only male
condition was used for forecasting as females had a much lower
sample size coupled with multiple reproductive states.
2.7  |  Egg mass counts
We assumed that the maximum observed egg mass count at the
peak of breeding (2009– 2019) was a suitable estimate of population-
level reproductive output, as egg mass count should be related to
both the number of breeding females and the number of clutches
produced per female. Population- level reproductive output may
be linked to climatic conditions that favor resource acquisition and
body condition, and therefore, the same hypotheses that describe
variation in body condition apply to egg mass counts. We included
one additional fixed effect in our model set, ice- off date, reasoning
that the timing of thaw may be related to how many salamanders ar-
rive to breed (Table S1). Statistic al modeling of egg mass counts was
relatively simple, given that egg mass count was taken as one value
(maximum) per year. Thus, we developed a set of 18 candidate mod-
els (Table S3). We compared models using AICc, as above. We also
used linear regression to test whether maximum egg mass counts
changed during 2008– 2019.
2.8  |  Spring breeding phenology
Our field observations suggested that annual variation in spring
weather patterns is associated with variation in salamander breeding
date. Median breeding date, retrospectively calculated as the ordinal
date at which half of the maximum egg masses were deposited at Bat
Lake, was used as the response variable in the candidate model set.
TAB LE 1 Description of models and results exploring the sensitivity of the parameter βYea r, the slope of Spot ted Salamander ( Ambystoma
maculatum) body condition over year, to environmental covariates. The base model for each sex does not feature an environmental
covariate, whereas subsequent environmental models consider a single environmental covariate; after the addition of the covariate, the
qualitative change in the magnitude and direction of βYea r relative to the base model is described. In all models, random effects are individual
ID, unique day, and year (defined Table S1)
Model description Fixed effects Period βYea r ± SE
% change in βYe ar
relative to base model
Base model F0Reprod status, ordinal date, year 2008– 19 −0.114 ± 0.0470
Env model F1Reprod st atus, ordinal date, year, summer temp 2008– 19 −0.0625 ± 0.0445 −45%
Env model F2Reprod st atus, ordinal date, year, autumn temp 2008– 19 −0.102 ± 0.0552 −11%
Env model F3Reprod st atus, ordinal date, year, ice in
2008– 19 −0.117 ± 0.0519 +3 %
Env model F4Reprod status, ordinal date, year, summer
2009– 19 −0.135 ± 0.0456a+30% a
Base model M0Ordinal date, year 2008– 19 −0.146 ± 0.0482
Env model M1Ordinal date, year, summer temp 2008– 19 −0.106 ± 0.0509 −27 %
Env model M2Ordinal date, year, autumn temp 2008– 19 −0.128 ± 0.0556 −12%
Env model M3Ordinal date, year, ice in dur ation 2008– 19 −0.141 ± 0.0530 −3%
Env model M4Ordinal date, year, april temp 2008– 19 −0.125 ± 0.0654 −14%
Env model M5Ordinal date, year, summer precip 2009– 19 −0.186 ± 0.0473a+22% a
aPrecipitation data were not available in 2007 for pairing with 2008 salamander body condition data. Therefore, Env model F4 and Env model M4
are based on the interval 20 09– 2019. The percent change in βYe ar ± SE estimates is also based on the interval 2009– 2019, where the base model
βYear ± SE estimate for females is −0.103 ± 0.0560 and for males is −0.152 ± 0.0594.
Median breeding date was selected because it is more robust to out-
liers compared to start date (Carter et al., 2018) and captures the pe-
riod of environmental conditions responsible for the onset of major
breeding activit y. We reasoned that mean temperature and summed
precipitation in March, April, and May, as well as Ice- off Date, could
affect breeding phenology (2008– 2019). We expected tempera-
tures to be negatively associated with breeding date, as high tem-
perature accelerates snow and ice melt. We expected ice- off date
to be positively correlated with breeding date, as breeding cannot
begin until ice is off the breeding pond. Lastly, we expected precipi-
tation in March and April (usually comprising snow) to be positively
associated with breeding date, and precipitation in May (cool rain) to
stimulate activit y and lead to earlier breeding date. Ultimately, these
hypotheses and their combinations resulted in a set of six candidate
models (Table S4).
Linear regression was used to test for temporal trends in repro-
ductive phenology using two dataset s: annual first egg- lay dates
(non- continuous study, 1986– 2019, N = 19 years) and annual first
egg- lay, median egg- lay, and peak egg- lay dates (focal study period,
2008– 2019, N = 12 years). Finally, because ice- off date was a very
strong predictor of salamander breeding (see Section 3), we examined
trends in ice- off phenology at Bat Lake (20 08– 2019, N = 12 years)
and nearby Lake of Two Rivers (1973– 2019, N = 44 years available;
R.G. Tozer, unpublished data).
2.9  |  Salamander metabolic rate estimation
During the period of August 2017 to September 2020, dataloggers
(HOBO TidbiT® v2) recorded temperature every 2 h at three soil
depths (shallow/surface, 0 m; mid- depth, 0.5 m; deep, 1.0 m) in the
forest surrounding Bat Lake. We used soil thermal profiles from Bat
Lake and the Spotted Salamander temperature- metabolic rate re-
lationship repor ted by Whitford and Hutchison (1967) to estimate
metabolic rate for Bat L ake Spotted Salamanders. See the Supporting
Information file for full metabolic rate estimation methods.
Over the period of 2008– 2019, we recorded 1004 capture events
of 956 individual female and 3247 capture events of 2176 individual
male Spotted Salamanders. Male- biased capture rates likely resulted
from sampling methodolog y (aquatic funnel traps disproportionately
FIGURE 1 Parameter estimates from the best- ranked model set (
≥ 0.95) for (a) female* (black) and male (gray) Spotted Salamander
body condition (scaled mass index, SMI; Peig & Green, 2009); (b) egg mass abundance (population- level reproductive output). Adjusted
standard error (Burnham & Anderson, 2002: section 4.3.3; Bartoń, 2020: 33, 42) was used to construct 95% confidence inter vals of
the parameter estimates. In (a) temperature estimates were multiplied by 0.1 before plotting. Also, April precipitation and mean April
temperature were present in the best- ranked models of the male dataset, but absent from the corresponding female models (Table S5).
*Gravid parameter estimate (0.49 ± 0.11, 95% CI) not plotted among female body condition parameter estimates because of figure scaling
[Colour figure can be viewed at wileyonlinelibrar]
intercept mate searching males), differential breeding turn- out of
the sexes, and/or a male sex bias in the population. Body condition
sample sizes are therefore much larger for males than for females.
3.1  |  Body condition
The best supported model predicting female body condition in-
cluded summer temperature, summer precipitation, ice- on (winter)
duration, and reproductive status (Model FBC13: w = 0.60; Table S5).
Although there was model selection uncertainty, Model FBC13 was
still five to eight times more likely to be the best model compared
to the next- best models in the candidate set (Table S5). As with
females, male body condition was best predicted by a model that
included summer temperature, summer precipitation, and ice- on du-
ration (MBC13: w = 0.39, Table S5). A model that included summer
temperature, autumn temperature, and November temperature was
also competitive in explaining male body condition (MBC8: w = 0.19).
Among model averaged parameter estimates, summer tempera-
ture, autumn temperature, ice- on duration, ordinal date, and repro-
ductive status (for females) were important predictors of female and
male body condition (Figure 1a). Higher mean summer and autumn
temperatures, a longer winter duration, and later spring capture date
negatively affected both female and male body condition (Figure 1a).
We found no significant temporal (i.e., year over year) autocorrela-
tion in mean female body condition when tested with all three data-
sets (gravid, spent, and pooled datasets; Figure S1a– c), but mean
male body condition demonstrated marginally significant temporal
autocorrelation at a 1- year time lag (Figure S1d).
3.2  |  Body condition: Long- term temporal and
environmental relationships
We found that female body condition declined during 2008– 2019
(βYear ± SE = −0.114 ± 0.0470; Figure 2). Substituting summer tem-
perature into the base model resulted in a 45% reduction in the mag-
nitude of the parameter estimate for year (βYea r = −0.0625 ± 0.0445),
whereas little change in βYear was observed when substituting au-
tumn temperature, or ice- in duration (Table 1). Using data from
2009 to 2019 (as no 2007 precipitation data were available to pair
with 2008 condition data), body condition declined across time for
females, albeit marginally (βYear = −0.104 ± 0.0560) and substitut-
ing summer precipitation into the base model slightly increased the
magnitude of βYear ( βYear = −0.135 ± 0.0456, a 30% increase; Table 1).
Results were broadly similar for males. Body condition declined
during 2008– 2019 in males (βYe ar = −0.146 ± 0.0482; Figure 2), and
substitution of summer temperature into the base model resulted in
a 27% reduction in the magnitude of βYe ar (βYea r = −0.106 ± 0.0509).
We observed modest changes in βYear when substituting autumn
temperature, ice- in duration, and April temperature into the base
model ( Table 1). Using data from 2009 to 2019 (again, no 2007 pre-
cipitation data were available to pair with 2008 condition data), body
condition declined across time for males (βYear = −0.152 ± 0.0594)
and substituting summer precipitation increased the magnitude of
βYear in the model (βYear = −0.186 ± 0.0473, a 22% increase; Table 1).
In sum, holding summer temperature constant tended to dampen
the temporal decline in condition in females and males, suggesting
summer temperature played a role in driving the temporal decline in
condition; holding summer precipitation constant tended to exacer-
bate the decline in condition in males and females, suggesting that
annual variation in summer precipitation moderates the temporal
decline in condition.
3.3  |  Body condition: Forecasting body condition
under climate change
Assuming a linear relationship between temperature and body
condition loss (and no compensator y response by salamanders),
warming of Mean Summer Temperature by 1– C is predicted to
reduce male breeding body condition by 5%– 14% (−0.25 to −0.75
SD of body mass; Figure 3a) for an average breeding condition male
± SD = 9.7 ± 1.8 g, all breeding males 2008– 2019). Warming of
mean autumn temperature by 1– 3°C is predicted to cause a body
condition decline of 5%– 9% for an average condition male (−0.25 to
−0.50 SD of body mass; Figure 3b). Simultaneous increases in both
mean summer and autumn temperature of 1– 3°C are expected to be
additive and expedite body condition loss, with declines of 10%– 27%
FIGURE 2 Body condition, measured as scaled mass index (SMI,
Peig & Green, 2009), for breeding adult female (black) and male
(gray) Spotted Salamanders (Ambystoma maculatum), Bat Lake,
Algonquin Provincial Park, Canada (2008– 2019). Time series body
condition data for females and males with best fit lines and 95%
confidence intervals from base models F0 and M0, respectively
(Table 1) [Colour figure can be viewed at]
(−0.50 to −1.50 SD of body mass; Figure 3c). Realistic concurrent in-
creases in precipitation are predicted to offer very little to no buffer-
ing of temperature- driven body condition decline (Figure 3).
3.4  |  Reproductive output
Maximum eg g mass abundance (2008– 2019) showed substan-
tial interannual variation across breeding seasons, varying by
approximately 400% (Figure 4). A model containing autumn tem-
perature best explained annual egg mass abundance; however, over-
all support and fit was weak for this model (Model RO5: w = 0.38,
= 0.25) and there was considerable overall model selection un-
certainty (Table S6). Model- averaged parameter estimates suggested
autumn temperature was positively associated with maximum egg
mass abundance, although uncertainty was high (Figure 1b; also see
Figure S2). We detected no temporal autocorrelation in annual egg
mass abundance (Figure S1e).
FIGURE 3 Forecast of male Spotted Salamander (Ambystoma maculatum) body condition under realistic climate scenarios (2050– 2100)
for western Algonquin Provincial Park: (a) increasing summer precipitation (up to 36 mm, or 24% increase in mean summer precipitation) and
increasing mean summer temperatures (increase up to 3°C); (b) increasing summer precipitation and increasing mean autumn temperatures;
and (c) increasing summer precipitation and increasing mean summer and autumn temperatures (synchronous increases up to 3°C in both
seasons). Using 2008– 2019 climate data as a baseline, the SD in mean annual summer and autumn temperature is 1.1°C and the SD in
mean summer precipitation is 50 mm. Therefore, realistic near- future warming of 2– C would be approximately ≥2 SD above current
seasonal temperatures, whereas realistic near- future summer precipitation increases (20%– 24%, equivalent to 30– 36 mm) are <1 SD of
current values. Near- future climate scenarios are based on Lemieux et al. (20 07) and Ridgway et al. (2018) [Colour figure can be viewed at]
FIGURE 4 Time series of Spotted Salamander (Ambystoma maculatum) egg mass abundance at Bat Lake, Algonquin Provincial Park
(2008– 2019). (a) Daily egg mass counts starting at first egg lay date and extending until maximum (peak) values. During 20 08– 2019, the
earliest breeding star t date was ordinal date 98 (April 7, 2010), latest breeding start date was day 129 (May 9, 2019), and the mean first
breeding date was day 116 (April 26). Egg mass count values that extended post- peak were excluded to improve visual clarity of the plot.
(b) Maximum egg mass count per annum [Colour figure can be viewed at]
(a) (b)
3.5  |  Spring breeding phenology
During 2008– 2019, the mean first breeding date of Spotted
Salamanders at Bat Lake was ordinal date 116 (April 26). Across this
period, first breeding date varied by 31 days (earliest breeding date:
day 98, April 7, 2010; latest: day 129, May 9, 2018). Salamander first
breeding date (
= 0.29, p = 0.04), median breeding date (
= 0.39,
p = 0.02), and peak breeding date (
= 0.51, p = 0.006) became sig-
nificantly later during 2008– 2019 (Figure 5a); however, the longer
time series (1986– 2019) showed no trend in first breeding date
= 0, p = 0.54; Figure 5b).
Model selection showed strong suppor t for Model BP6 in which
ice- off date predicted median breeding date (w = 0.93,
= 0.67;
Table S6). Model averaging was not conducted because Model BP6
was strongly supported. High levels of collinearity (VIF values up
to 6.2) were identified among predictors in models of the breeding
phenology model set (Table S4). The variable(s) with high VIF val-
ues (VIF > 3.0) in a model were removed and model selection was
rerun. The overall outcome of model selection was unaffected, as
strong suppor t remained for Model BP6 as the best- fitting model in
the model set.
In subsequent analysis of ice- off dates, we detec ted a marginally
significant delay in Ice- of f date for Bat Lake (
= 0.27, p = 0.047) at
a rate of 1.3 days per year over the 2008– 2019 monitoring period
(β ± SE = 1.3 ± 0.58 SE). There was no change in ice- off date for the
nearby long- term reference site, Lake of Two Rivers, during 1972
2019 (
= 0, p = 0.90, β ± SE = 0.012 ± 0.099).
3.6  |  Salamander metabolic rate estimation
Ground surface temperature around Bat Lake ranged mostly between
20 and 25°C, occasionally reaching 30°C , in the months of July and
August (Figure 6a). Temperatures were 10– 15°C at depths of 0.5–
1.0 m during summer (Figure 6a). Using soil temperature recorded at
multiple depths and the Spotted Salamander temperature– metabolic
rate relationship reported by Whitford and Hutchison (1967), sala-
manders at shallow depths (14.8°C surface mean temperature) have
an estimated metabolic rate 1.3 times that of individuals at mid-
depths (10.8°C mean temperature at 0.5 m depth) and 1.6 times
that of individuals deep underground (9.1°C mean temperature at
1.0 m depth; Figure 6c) during the active period (May 20– October
15). Metabolic rate estimation highlighted the thermal sensitivity of
these fossorial salamanders. For instance, given the Q10 value re-
ported for the temperature range of 10– 15°C (Q10 = 4.37; Whitford
& Hutchison, 1967), a change of 1, 2, or 3°C results in a 16%, 34%, or
56% change in salamander metabolic rate, respectively.
In this study, we explored the association between annual climatic
variation, body condition, population- level reproductive output, and
breeding phenology in a fossorial salamander over a 12- year period.
First, we found that interannual variation in climate is associated
with body condition and breeding phenology of salamanders, but
climatic variation was not strongly associated with our measure of
reproductive output. Second, body condition of females and males
declined during the study period, and a slight increase in summer
temperature during the study period is likely par tly responsible for
the decline. Third, associations between temperature and body con-
dition were predicted to be non- trivial under realistic near- future
climate change, with up to a 27% decline in body condition for a 3°C
seasonal temperature increase and little evidence that precipitation
increases will moderate these temperature effects.
Global amphibian diversity is highly structured across vertical
strata (fossorial, terrestrial, arboreal), with a large propor tion of
temperate and arid species occupying subterranean environment s
(Oliveira et al., 2017; Oliveira & Schef fers, 2019). Microhabitat re-
fugia in combination with behavioral thermoregulation is expected
FIGURE 5 Breeding phenolog y of the Spotted Salamander (Ambystoma maculatum) at Bat Lake, Algonquin Provincial Park. (a) First egg-
laying date (
= 0.29, p = 0.04), median egg- laying date (
= 0.39, p = 0.02), and peak egg- laying date (
= 0.51, p = 0.006) have become
significantly later during the focal study period (2008– 2019). (b) Extended dataset of first breeding date (1986– 2019) shows no temporal
trend (best fit line;
= 0, p = 0.54). Ice- of f dates at Bat Lake (2008– 2019) and nearby reference site, Lake of Two Rivers (1986– 2019 shown)
are highly correlated (r = 0.93, p < 0.0001) during years of data overlap (2008– 2019). Note close synchrony (typically ≤24 h) between Bat
Lake ice- off and first salamander egg- laying
(a) (b)
to be crucial for individual survival as well as population and spe-
cies persistence under climate change (Fitzpatrick et al., 2020; Huey
& Buckley, 2018; Lara- Reséndiz et al., 2021; Moore et al., 2018;
Scheffers et al., 2013; Scheffers, Edwards, et al., 2014). Amphibians
are susceptible to direct threat s (thermal and hydric stress), indirect
threats (habitat change, food availability, disease risk), and interac-
tions thereof associated with climate change (Blaustein et al., 2010;
Lertzman- Lepofsky et al., 2020; Reading, 2007; Rohr & Palmer,
2013), but fossorial species may be buffered from strong climate
effects, at least in the short term. This is important because sala-
manders, and amphibians more broadly, contribute to numerous
ecological functions (Best & Welsh, 2014; Davic & Welsh, 2004),
particularly in the forest s of eastern Nor th America. Never theless,
given the global distribution, diversity, and ecological roles of fosso-
rial amphibians, heightened attention to the vulnerability of micro-
habitat refugia is warranted amid global climatic change.
Our study shows that fossorial salamanders are sensitive to tem-
perature increases, with temperature from the preceding summer
and autumn negatively affecting breeding body condition. The di-
rection of these relationships is consistent with the interpretation
that reduced body condition may be driven by elevated metabo-
lism and/or thermal stress in the summer. Direct monitoring in the
forest understory around Bat Lake demonstrated that ground sur-
face temperature ranged consistently between 20 and 25°C in July
and August (Figure 6a). Notably, surface temperatures sometimes
reached 30°C, which would greatly elevate metabolism (Whitford
& Hutchinson, 1963, 1967) and likely induce stress, as these tem-
peratures approach the thermal critical maximum of the Spotted
Salamander (Gatz, 1971, 1973; Hutchinson, 1961; Pough & Wilson,
1970). Spotted Salamanders often spend the summer and autumn
periods at shallow depth (2.5– 3 cm below the surface of the leaf
litter; Faccio, 2003) in horizontal burrows and runways of small
mammals, transitioning to use of deeper vertical burrows for over-
wintering (Madison, 1997; Montieth & Paton, 2006; salamanders
at depths of 30– 122 cm are known during summer, Gordon, 1968;
Kleeberger & Werner, 1983). Although temperatures are buffered
in the subterranean environment, thermal profiles at depths of 0.5–
1.0 m at Bat Lake indicated that temperatures still reached 10– 15°C
during summer (Figure 6a). The temperature sensitivity of metabolic
expenditure of Spotted Salamanders at 10– 15°C is high (Q10 = 4.37
at 10– 15°C, based on metabolic rate measured across a temperature
range of 5– 30°C; Whitford & Hutchison, 1967), which could explain
the association between reduced body condition and high active
seasonal temperatures. Using soil temperature data, we estimated
that salamanders at warmer shallow depths have a metabolic rate
1.3– 1.6 times that of individuals at cooler deeper depths (0.5– 1.0 m
depth; Figure 6c). Furthermore, temperature increases of 1– C are
estimated to result in sizable increases in salamander metabolic rate
(16%– 56% change in met abolic rate), suggesting substantive ener-
getic consequences within these commonly encountered spring and
summer temperatures. Spotted Salamander microhabitat selection
during the active season, including use of areas with deeper leaf
litter, high plant and natural object cover, and high small mammal
burrow densities, is consistent with preferences for cool damp con-
ditions that would reduce metabolic rate (Faccio, 2003; Montieth &
Paton, 2006). Soil temperature profiles (Figure 6a) and metabolic rate
FIGURE 6 Thermal environment and estimated metabolic rate
of the Spotted Salamander (Ambystoma maculatum). (a) Averaged
thermal profile at shallow (surface, 0 m), mid- depth (0.5 m), and deep
(1.0 m) soil depths from t wo temperature monitoring s tations at Bat
Lake, Algonquin Provincial Park. (b) Estimated Spotted Salamander
metabolic rate based on soil thermal profiles. Estimates based on
the empirical temperature– metabolic rate relationship reported
by Whitford and Hutchison (1967; see Supporting Information). (c)
Mean estimated metabolic rate of Spotted Salamanders at varying
soil depth across the active and inactive season. For our study
population, the active season spanned from ordinal day 140– 288
(May 20– October 15) and was defined by the non- freezing period
during which salamanders were post- breeding and presumed
feeding (In (a) note that that the underground thermal gradient
flips direction in mid- May and mid- October). The numerical values
presented above each bar are mean soil temperatures at the
respective depth and activit y period. See Supporting Information
for additional information about metabolic rate estimation [Colour
figure can be viewed at]
2018 2019 2020
Shallo w
2018 2019 2020
Metabolic rate ( l O
Active Inactive
Shallow MidDeep ShallowMid Deep
Metabolic rate ( l O
estimates (Figure 6b) suggest that remaining at mid- depths during
the inactive (winter) period offers slight advantage for minimizing
metabolism while also avoiding freezing temperatures (Figure 6b,c).
During the 12- year study period, the breeding body condition of
spent females and males significantly declined (Figure 2). During this
period, our study area experienced a slight increase in mean summer
temperature ( β ± SE = 0.14 ± 0.081°C, p = 0.11, n = 12 years; Table
S8). In our models, holding summer temperature constant dampened
the temporal decline in body condition in both sexes (Table 1) and it
therefore seems likely that increases in summer temperature are at
least partly responsible for the temporal decline in body condition.
Longer term temperature trends for Algonquin Park show significant
temperature increases (1975– 2006, Waite & Strickland, 2006; 1915–
2016, Favot et al., 2019). The parameter estimates for temperature–
condition relationships in our models (Figure 1a), and projec ted
temperature increases for Algonquin Park over the next half- century
(Lemieux et al., 2007; Ridgway et al., 2018), suggest that this trend
in declining body condition is expected to continue (Figure 3). We
did not model female body condition under near- future climate sce-
narios given their multiple reproductive states and a smaller sample
size. However, given the synchronous decline of body condition in
both sexes over the past 12 years (Figure 2) and similar seasonal
temperature sensitivities of body condition in both sexes (Figure 1a),
it is reasonable to expect that female body condition will continue
to decline in a warming climate, as forecasted for males (Figure 3). If
female body condition declines continue, individual- and population-
level consequences may begin to manifest, such as reduced egg size,
clutch size, reproductive frequency, and survival.
The negative relationship between body condition and ice- on
(winter) duration can be reasonably attributed to extended fasting in
years of late ice- off. Similarly, later spring capture dates resulted in
reduced body condition of females and males, a result corroborated
in similar salamander field studies (Homan et al., 2018; Strickland
et al., 2015), and likely related to their non- feeding status and en-
ergy expenditure throughout winter (Figure 6b) and during early
spring migration and breeding. Although we did not forecast how
changes in ice- on duration may affect body condition (largely due to
the lack of any projection available), it seems likely that warmer fall
temperatures will be associated with decreases in winter duration in
future climates, and hence relatively warm temperatures in the over-
wintering environment (Figure 6a), which may partly offset the en-
ergetic benefit of a shortened winter duration. Furthermore, based
on parameter estimates (Figure 1a), a change in ice- on duration of
approximately 20 days is associated with the same change in body
condition as a 1°C change in summer temperature, suggesting that
temperature effects during the active season would overwhelm any
benefit associated with shor ter winter (fasting) duration.
Our forecast of salamander body condition assumed that sala-
manders do not demonstrate a compensator y response to warmer
temperatures, although behavioral thermoregulation in response to
elevated temperatures is a reasonable expectation (Huey & Buckley,
2018; Kearney et al., 2009). Spot ted Salamanders often reside in
shallow horizontal burrows during the active season (Faccio, 2003;
Madison, 1997) but could become more reliant on vertical small
mammal burrows that offer cooler refuge. Although a shift from
horizontal to vertical habitat use during the active season may re-
lieve salamanders from high temperatures and elevated metabolism,
sheltering deeper underground may fur ther complicate energy ac-
quisition and/or retention by stymieing near- surface foraging oppor-
tunities (Gordon, 1968).
Mean autumn temperature was positively correlated with egg
mass abundance (Figure 1; Figure S2). A positive association be-
tween autumn temperature and egg mass count s suggests energetic
gain at the individual level during autumn, but this is contradictory to
our analyses incorporating individual- level information, which sug-
gest that autumn temperature was negatively correlated with female
body condition. We suspect, for several reasons, that the positive as-
sociation between autumn temperature and egg mass counts is spu-
rious. First, the frequency of female reproduction in our population
is likely not annual (as with Homan et al., 2018; Husting, 1965) and
this may confound interpretation of temperature and population-
level reproductive output because individuals are amassing energy
over multiple years. Second, philopatry at the breeding site may be
low, but interpretation of egg mass abundance implies that Bat Lake
is the only breeding site used by salamanders within this population.
Although pond- breeding amphibians are reputed for their philopatry
to breeding sites (e.g., Gamble et al., 2007; Gill, 1978; Semlitsch,
2008; Vasconcelos & Calhoun, 2004), it is also evident that breeding
individuals use alternative sites (e.g., colonization of created ponds;
Denoël et al., 2017; Patrick et al., 2008; Petranka et al., 2003). Also,
temperature may influence reproduction at a finer scale, such as egg
size and clutch size (e.g., Fraser, 1980) or clutch partitioning (female
Spotted Salamanders are reported to lay two to four egg masses per
reproductive bout; Petranka, 2010), rather than at population- level
through breeding turnout. In sum, it seems unlikely that autumn
temperature is positively associated with egg mass counts, and that
more detailed, individual- level data on reproduction may be needed
to estimate associations between climate and reproductive output .
We did not find evidence to suggest that salamander breed-
ing phenology was advancing concurrent with climatic warming.
Although Algonquin Park and surrounding regions have experi-
enced a warming trend of approximately 1°C in the past half- centur y
(Lemieux et al., 2007; Ridgway et al., 2018), much of the warming
has been concentrated in the summer and autumn (Favot et al.,
2019; Waite & Strickland, 2006). Given that salamander breeding
phenology is closely associated with spring conditions, primarily
ice- off, and that there has been little to no significant advancement
of spring thaw conditions in Algonquin Park, salamander breeding
phenology remains unchanged over the past several decades in the
higher elevation uplands of western Algonquin Park. Over our pe-
riod of continuous monitoring (2008– 2019), Spotted Salamanders
demonstrated phenological delay in reproductive timing. Ariet ta
et al. (2020) also reported delayed breeding (2000– 2020) in another
early- spring breeding amphibian, the Wood Frog, in the northeast-
ern United States. These authors cite seasonally heterogeneous
warming temperatures, later snow accumulation, and longer spring
snow and ice persistence for delayed breeding, analogous to that
observed in Algonquin Park (Favot et al. 2019). Signals of delayed
breeding in shor ter term datasets of amphibians from northeast-
ern North America (Arietta et al., 2020; this study) are in contrast
to longer datasets, which tend to show mixed responses of am-
phibian breeding phenology to climate change: no change (North
America: Blaustein et al., 2001; Green, 2017; Kirk et al., 2019; Klaus
& Lougheed, 2013; Todd et al., 2011) or advancement (eastern north
America: Gibbs & Breisch, 2001; Todd et al., 2011; Walpole et al.,
2012; Klaus & Lougheed, 2013; northern Europe: Terhivuo, 1988;
Beebee, 1995a, 1995b; Tryjanowski et al., 2006; Japan: Kusano &
Inoue, 2008). Amphibians clearly demonstrate a range of species, re-
gional, and temporal phenological responses to climate change (e.g.,
Muths et al., 2017; Okamiya et al., 2021).
Fossorial species are difficult to study and are underrepre-
sented in research at the interface of organismal biology and climate
change. Our findings show that fossorial species can be susceptible
to rapid warming. Subterranean microhabitats, such as burrows, are
buffered from surface conditions, but these environments are not
completely insulated from aboveground climate. Large numbers of
species live in subterranean microhabitats, especially where already
extreme ambient environmental conditions prevail. Many fossorial
species are ecosystem engineers, creating thermal refugia on the
landscape upon which whole ecological communities are depen-
dent (Doody et al., 2021; Pike & Mitchell, 2013), emphasizing the
importance of protecting burrowing biota. Landscape- scale changes
in species distributions, such as latitudinal and altitudinal shifts,
in response to climate change are increasingly recognized. In con-
trast, comparatively little is known about thermal habitat structure
and species responses to environmental change on the microhab-
itat scale (Scheffers et al., 2013; Scheffers, Edwards, et al., 2014;
Scheffers, Evans, et al., 2014), particularly for subterranean envi-
ronments (Lara- Reséndiz et al., 2021; Moore et al., 2018). Fossorial
ectotherms and amphibians in particular may be at elevated risk as
their thermal microhabitat warms. Our study draws attention to the
vulnerability of subterranean environments and their inhabitants
amid global climatic change.
PDM sincerely thanks DL LeGros for a warm welcome into field bi-
ology and a dozen years of salamander sleuthing; LA Rye and JD
Litzgus for encouragement and suppor t that made several years of
this project possible. We thank RG Tozer for collecting and provid-
ing access to long- term ice data and JA Leivesley for assistance with
troubleshooting R code. We greatly appreciate the help of numerous
research students (and other volunteers) for field data collection, in-
cluding: DL LeGros, SP Boyle, O Butty, JWD Connoy, D Crawford,
EA Francis, G H- Y Gao, N Hrynko, JA Leivesley, DI Mullin, S Paiva, D
Ravenhearst, C Rouleau, M Terebiznik, H Vleck, L Warma, with spe-
cial thanks to SJ Kell and T Wynia for many years of data collection
and frozen fingers. We appreciate the help of R Eckenswiller, TM
Winegard, and the Algonquin Wildlife Research Station for accom-
modation and field support. We thank P Gelok, J Hoare, A Lake, B
Steinberg, L Trute, Algonquin Provincial Park, and Ontario Parks for
permits, funding, and logistical support of this research.
Patrick D. Moldowan, Glenn J. Tattersall, and Njal Rollinson conceived
of the study; Patrick D. Moldowan and Glenn J. Tattersall conducted
field data collection; Patrick D. Moldowan and Glenn J. Tattersall
curated the dat aset; Patrick D. Moldowan, Glenn J. Tattersall, and
Njal Rollinson analyzed the data; Patrick D. Moldowan, Glenn J.
Tattersall, and Njal Rollinson funded the study; Patrick D. Moldowan
and Njal Rollinson wrote the manuscript; Patrick D. Moldowan,
Glenn J. Tattersall, and Njal Rollinson edited and approved the final
version of the manuscript.
The data that suppor t the findings of this study will be openly
available in the Zenodo research data repository at http://doi.
org/10.5281/zenodo.4993600 (reference number 4993600) follow-
ing publication.
Patrick D. Moldowan 0-0003-2852-794X
Arietta, A. Z. A. , Freide nbu rg, L. K., Urban, M. C., Rodrig ues , S. B.,
Rubenstein, A., & Skelly, D. K. (2020). Phenological delay de-
spite warming in wood frog Rana sylvatica reproduct ive timing:
A 20- year study. Ecography, 42, 1 11 . htt ps://doi. org/10 .1111/
Baldauf, R. J. (1952). Climatic f actor s influencing the breeding migration
of the Spotted Salamander, Ambystoma maculatum. Copeia, 1952 ,
178– 181.
Bartolai, A . M., He, L., Hurst, A. E., Mortsch, L., Paehlke, R., & Scavia, D.
(2015). Climate change as a driver of change in the Great Lakes St.
Lawrence River Basin. Jour nal of Great Lakes Research, 41– 1, 45– 58 .
Bartoń, K. (2020). MuMIn: Multi- model inference. R Package Version,
1(43), 17. https://CRAN.R- proje ge=MuMIn
Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear
mixed- effects models using lme 4. Journal of Statistical Sof tware, 67,
1– 48. JSS.V067.I01
Beebee, T. J. C. (1995a). Amphibian breeding and climate. Nature, 374,
219– 220.
Beebee, T. J. C. (1995b). Ever- earlier migr ations by alpine newts (Tri tur us
alpestris) living wild in Britain. The British Herpetological Society
Bulletin, 51, 5– 6.
Benard, M. F. (2015). Warmer winters reduce frog fecundity and shift
breeding phenology, which consequently alters lar val development
and metamorphic timing. Global Change Biology, 21, 1058– 1065.
Best, M. L., & Welsh Jr., H. H. (2014). The trophic role of a forest sala-
mander: Impacts on invertebrates, leaf litter retention, and the
humification process. Ecosphere, 5, 1– 19. h ttps://doi.or g/10.18 90/
ES13- 00302.1
Blanchard, F. N. (1930). The stimulus to the breeding migration of the
Spotted Salamander, Ambys toma maculatum (Shaw). The Am erican
Naturalist, 64, 154– 167. 06
Blaustein, A. R., Belden, L. K., Olson, D. H., Green, D. M., Root,
T. L., & Kieseeker, J. M. (20 01). Amphibian breeding and
climate change. Conservation Biology, 15, 1804– 1809. https://doi.
org/10.1046/j.1523- 1739.20 01.0 03 07.x
Blaustein, A . R., Walls, S. C., Bancroft, B. A ., L awler, J. J., Searle, C . L., &
Gervasi, S . S. (2010). Direc t and indirect effects of climate change
on amphibian populations. Diversity, 2, 281– 313. https://doi.
Bulahova, N. A., & Berman, D. I. (2017). Reproductive cycle of females
and reproduction of the Siberian Salamander (Salamandrella keyser-
lingii, Caudata, Hynobiidae). Biology Bulletin, 44, 688– 699. https:// 35901 7070068
Burnham, K. P., & Anderson, D. R. (2002). Model selection and multi-
model inference: A practical information- theoretic approach (2nd ed.).
Burnham, K. P., Anderson, D. R., & Huy vaer t, K. P. (2011). AIC model
selection and multimodel inference in behavior al ecolog y:
Some background, observations, and comparisons. Behavioral
Ecology and Sociobiology, 65, 23– 35.
5- 010- 1084- z
Carter, S. K ., Saenz, D., & Rudolf, V. H. W. (2018). Shifts in phenological
distributions reshape interaction potential in natural communities.
Ecology Letters, 21, 1143– 1151.
Caruso, N. M., Sears, M. W., Adams, D. C., & Lips, K. R. (2014). Widespread
rapid reductions in body size of adult salamanders in response to
climate change. Global Change Biology, 20, 17511759. https://doi.
Chen, I.- C., Hill, J. K., Ohlemüller, R., Roy, D. B., & Thomas, C. D. (2011).
Rapid r ange shifts of spe cies associated with high levels of climate
warming. Science, 333, 1024– 1026. http s:// /10.1126/scien
Cohen, J., Screen, J. A., Furtado, J. C., Barlow, M., Whittleston, D.,
Coumou, D., Francis, J., Dethloff, K ., Entekhabi, D., Overland,
J., & Jones, J. (2014). Recent Arctic amplification and ex treme
mid- latitude weather. Nature Geoscience, 7, 627 637. https://doi.
Connet te, G. M., Crawford, J. A., & Peterman, W. E. (2015). Climate
change and shrinking salamanders: Alternative mechanisms for
changes in plethodontid salamander body size. Global Change
Biology, 21, 2834– 2843.
Daszak, P., Scot t, D. E., Kilpatrick, A. M., Faggioni, C ., Gibbons, J. W., &
Porter, D. (2005). Amphibian population declines at Savannah River
site are linked to climate, not chytridiomycosis. Ecolog y, 86, 3232–
3237. 0598
Davic, R. D., & Welsh Jr., H. H. (2004). On the ecological roles of salaman-
ders. Annual Review of Ecology, Evolution, and Systematics, 35, 405–
435. ev.ecols ys.35.112202.130116
den Har tog, J. E., & Reijns, R. A . (2020). I3S: Interactive individual identifi-
cation sys tem.
Denoël, M., Dalleur, S., Langrand, E., Besnard, A ., & Cayuela, H. (2017).
Dispersal and alternative breeding site fidelit y strategies in an
amphibian. Ecography, 41 , 1543– 1555.
Doody, J. S., Soennichsen, K. F., James, H., McHenry, C., & Clulow, S.
(2021). Ecosystem engineering by deep- nesting monitor lizards.
Ecology, 102 (4).
Douglas , M. E. (1979). Migration and sexual selection in Ambystoma jef-
fersonianum. Canadian Journal of Zoolog y, 57, 2303– 2310. https:// 299
Faccio, S. D. (2003). Postbreeding emigration and habitat use by Jefferson
and Spotted Salamanders in Vermont. Journal of Herpetolog y, 37,
479– 489. 02A
Favot, E. J., Rühland, K. M., DeSellas, A. M., Ingram, R., Paterson, A. M.,
& Smol, J. P. (2019). Climate variabilit y promotes unprecedented
cyanobacterial blooms in a remote, oligotrophic Ontario lakes:
Evidence from paleolimnolog y. Journal of Paleolimnolog y, 62, 3152. 3- 019- 00074 - 4
Fitzpatrick, M. J., Por ter, W. P., Pauli, J. N., Kearney, M. R., Notaro, M., &
Zuckerberg, B. (2020). Future winters present a complex energetic
landscape of decreased costs and reduced risk for a tree- tolerant
amphibian the Wood Frog (Lithobates sylvaticus). Global Change
Biology, 26, 6350– 6362.
Fox, J., & Weisberg, S. (2019). An R companion to applied regression (3rd
ed.). Sage Publications.
Fraser, D. F. (1980). On the environmental control of oocy te maturation
in a plethodontid salamander. Oecologia, 46, 302– 307. ht tps://doi.
org/10.1007/BF003 46256
Gamble, L. R., McGarigal, K., & Compton, B. W. (2007). Fidelity and
dispersal in the pond- breeding amphibian, Ambystoma opacum:
Implications for patio- temporal population dynamics and con-
servation. Biological Conservation, 139, 247– 257. ht tps://doi.
Gatz, A . J. (1971). Critical thermal maxima of Ambystoma maculatum
(Shaw) and Ambystoma jeffersonianum (Green) in relation to time of
breeding. Herpetologica, 27(2), 157– 200.
Gatz, A. J. (1973). Intraspecific variations in critical thermal maxima of
Ambystoma maculatum. Herpetologica, 29, 264– 268.
Gibbs, J. P., & Breisch, A. R. (2001). Climate warming and calling phenology
of frogs near Ithaca, New York, 1900– 1999. Conservation Biology,
15, 1175– 1178. 173 9.20 01 .01 50 0
41175. x
Gill, D. E. (1978). The metapopulation ecolog y of the Red- spotted Newt,
Notophthalmus viridescens (Rafinesque). Ecological Monographs, 48,
145– 166. ht tps://
Gordon, R. E. (1968). Terrestrial ac tivit y of the Spotted Salamander,
Ambystoma maculatum. Copeia, 196 8, 879– 880. https://doi.
Green, D. M. (2017). Amphibian breeding phenology trends under cli-
mate change: Predicting the past to forecast the future. Global
Change Biology, 23, 646– 656. /10.1111/gcb.13390
Hoeniger, J. F. M . (1986). Decomposition studies in t wo central
Ontario lakes having surf icial pHs of 4.6 and 6.6. Applied and
Environmental Microbiology, 52, 489– 497.
aem.52.3.489- 497.1986
Homan, R. N., Holger son, M. A., & Biga, L. M. (2018). Long- term demo-
graphic study of a Spotted Salamander (Ambystoma maculatum)
population in central Ohio. Herpetologica, 74, 109– 116. htt ps://doi.
org/10.1655/Herpe tolog ica- D- 17- 00067.1
Huey, R . B., & Buckley, L. B. (2018). Biological buffers and the impacts
of climate change. Integrative Zoology, 13, 349– 354. https://doi.
org/10.1111/1749- 4877.12321
Hughes, L. (2000). Biological consequences of global warming: Is the
signal already apparent? Trends in Ecolog y & Evolution, 15, 56– 61. - 5347(99)01764 - 4
Husting, E. L. (1965). Survival and breeding struc ture in a population
of Ambystoma maculatum. Copeia, 196 5, 352– 362. https://doi.
Hutchinson, V. H. (1961). Critical thermal maxima in salamanders.
Physiological Zoology, 34, 92– 125.
Intergovernmental Panel on Climate Change (IP CC). (2014). Climate
change 2014: Synthesis report. Contribution of working groups I, II and
III to the fifth assessment report of the Intergovernmental Panel on
Climate Change [Core Writing Team, R. K. Pachauri, & L . A . Meyer
(Eds.)]. IPCC .
Kearney, M., Shine, R., & Porter, W. P. (2009). The potential for behavioral
thermoregulation to buf fer “cold- blooded” animals against climate
warming. Proceedings of the National Acade my of Sciences of the
United States of America, 106 , 3835– 3840.
pnas.08089 13106
Kirk, M. A., Galatowitsch, M. L., & Wissinger, S. A . (2019). Seasonal dif-
ferences in climate change explain a lack of multi- decadal shif t in
population characteris tics of a pond breeding salamander. PLoS
One, 14, e0222097. al.pone.0222097
Klaus, S. P., & Lougheed, S. C . (2013). Changes in breeding phenology of
eastern Ontario frogs over four decades. Ecolog y and Evol ution, 3,
835– 8 45. /10.1002/ece3.501
Kleeberger, S. R., & Werner, J. K. (1983). Post- breeding migration
and summer movement of A mbystoma maculatum. Journal of
Herpetology, 17, 176– 177.
Kusano, T., & Inoue, M . (2008). Long- term trends toward earlier breed-
ing of Japanese amphibians. Journal of Herpetology, 42, 608– 614. 002R1.1
Lannoo, M. J., & Stiles, R. M. (2017). Effec ts of shor t- term climate
variation on a long- lived frog. Copeia, 105, 726– 733. https://doi.
org/10.1643/CH- 16- 449
Lara- Reséndiz, R. A ., Galina- Tessaro, P., Sinervo, B., Miles, D. B., Valdez-
Villavicencio, J. H., Valle- Jiménez, F. I., & Méndez- de La Cruz, F. R.
(2021). How will climate change impact fossorial lizard species? Two
examples in the Baja California Peninsula. Journal of Thermal Biology,
95, 102811. bio.2020.102811
Lemieux, C. J., Scott, D. J., Gray, P. A., & Davis, R. G. (2007). Climate
change and Ontario's provincial parks. Towards an adaptation strat-
egy. Ontario Ministry of Natural Resources, Applied Research and
Development Branch. Climate change research repor t CCRR- 06.
Lertzman- Lepofsk y, G. F., Kissel, A. M., Sinervo, B., & Palen, W. J. (2020).
Water loss and temperature interact to compound vulnerability to
climate change. Global Change Biology, 26 , 4868– 4879. https://doi.
Lesbarrères, D., Ashpole, S. L ., Bishop, C. A ., Blouin- Demers, G., Brooks,
R. J., Echaubard, P., Govindarajulu, P., Green, D. M., Hecnar,
S. J., Her man, T., Houlahan, J., Litzgus, J. D., Mazerolle, M. J.,
Paszkowski, C. A ., Rutherford, P., Schock, D. M., Storey, K. B., &
Lougheed, S. C . (2014). Conservation of herpetofauna in nor thern
landscapes: Threats and challenges from a Canadian perspective.
Biological Conservation, 170 , 48– 55. ht tps://
MacCracken, J. G ., & Stebbings, J. L. (2012). Test of a body condition
index with amphibians. Journal of Herpetolog y, 46, 346– 350. https:// 292
Madison , D. M. (1997). The emigration of radio- implanted spotted sal-
amanders, Ambystoma maculatum . Journal of Herpetology, 31,
542 551.
McCar ty, J. P. (2001). Ecological consequences of recent climate change.
Conservation Biology, 15, 320– 331.
j.1523- 1739.20 01 .0150 0 2320.x
Montieth, K. E., & Paton, P. W. C. (20 06). Emigration behavior of Spotted
Salamanders on golf courses in southern Rhode Island. Journal of
Herpetology, 40, 195– 205. /10.1670/130- 04A.1
Moore, D., Stow, A., & Kearney, M. R. (2018). Under the weather?
The direct effects of climate warming on a threatened des-
ert lizard are mediated by their activity phase and burrow
system. Journal of Animal Ecology, 87, 660– 671. https ://doi.
org/10.1111/1365- 2656.12812
Muths, E., Chambert, T., Schmidt, B. R., Miller, D. A . W., Hossack, B. R .,
Joly, P., Grolet, O., Green, D. M., Pilliod, D. S., Cheylan, M., Fisher,
R. N., McCaffery, R. M., Adams, M. J., Palen, W. J., Arntzen, J. W.,
Garwood, J., Fellers, G., Thirion, J.- M., Besnard, A., & Campbell
Grant, E. H. (2017). Heterogenous responses of temperate- zone
amphibian populations to climate change complicates conserva-
tion planning. Scientific Reports, 7, 17102.
s4159 8- 017- 17105 - 7
Nakagawa, S., & Schielzeth, H. (2013). A general and simple method
for obtaining R2 from generalized linear mixed- effects mod-
els. Methods in Ecology and Evolution, 4, 133– 142. ht tps://doi.
org/10.1111/j.2041- 210x.2012.0 0261.x
Okamiya, H., Hayase, N., & Kusano, T. (2021). Increasing body size and
fecundity in a salamander over four decades, possibly due to global
warming. Biological Journal of the Linnean Society, 132, 634– 642. nnean/ blaa201
Oldham, M. J., & Weller, W. F. (1989). Ontario herpetofaunal summar y
1986, technical supplement. Ontario Field Herpetologists.
Oliveira, B. F., São- Pedro, V. A., Santos- Barrera, G., Penone, C., & Costa,
G. C. (2017). AmphiBIO, a global database for amphibian ecolog-
ical traits. Scientific Data, 4, 170123.
Oliveira, B. F., & Scheffers, B. R. (2019). Vertical stratific ation influences
global patterns of biodiversity. Ecography, 42, 249– 258. https://doi.
Patrick , D. A., Calhoun, A. J. K ., & Hunter Jr., M. L . (2008). The impor-
tance of understanding spatial population structure when evaluat-
ing the effects of silviculture on Spot ted Salamanders ( Ambystoma
maculatum). Biological Conservation, 141, 807– 814. htt ps://doi.
Peacock, R. L., & Nussbaum, R. A. (1973). Reproductive biology and
population stric ture of the western red- backed salamander,
Plethodon vehiculum. Journal of Herpetology, 7, 215– 224. https://doi.
Peig, J., & Green, A . J. (2009). New per spectives for estimating
body condition from mass/length data: The scales mass index
as an alternative method. Oikos, 118 , 1883– 1891. https://doi.
org/10.1111/j.1600- 0706 .20 09.17643.x
Petranka, J. W. (2010). Salamanders of the United States and Canada.
Smithsonian Books.
Petranka, J. W., Kennedy, C. A., & Murray, S. S. (2003). Response of
amphibians to restor ation of a southern Appalachian wetland: A
long- term analysis of community dynamic s. Wetland s, 23, 1030–
1042. 5212(20 03) 023[1030:ROATR
Pike, D. A., & Mitchell, J. C. (2013). Burrow- dwelling ecosystem
engin ee rs provide ther ma l refugia thr oughout the lan dsca pe .
Animal Conservation, 16 , 694– 703.
Pough, F. H., & Wilson, R. E. (1970). Natural daily temperature stress, de-
hydration, and acclimation in juvenile Ambystoma maculatum (Shaw)
(Amphibia: Caudat a). Physiological Zoology, 43, 194– 205. https:// ool.43.3.30155529
R Core Development Team. (2020). R: A language and environment for sta-
tistical computing. R Foundation for Statistical Computing. http://
www.R- proje
Reading, C. J. (20 07). Linking global warming to amphibian declines
through its effects on female body condition and sur vivor-
ship. Oecologica, 151, 125– 131. 0044
2- 006- 0558- 1
Ridgway, M., Smith , D., & Middel, T. (2018). Climate warming projec tions
for Algonquin Provincial Park. Ontario Ministry of Natural Resources
and Forestry, Science and Research Branch. Science and research
information report IR- 14.
Rohr, J. R., & Palmer, B. D. (2013). Climate change, multiple stressors,
and the decline of ectotherms . Conservation Biology, 27, 741– 751.
Scheffers, B. R., Brunner, R. M., Ramirez, S. D., Shoo, L. P., Diesmos,
A., & Williams, S. E. (2013). Thermal buffering of microhabitats
is a critical factor mediating warming vulnerability of frogs in the
Philippine biodiversity hotspot . Biotropica, 45, 628– 635. https://
Scheffers, B. R., Edwards, D. P., Diesmos, A., Williams, S. E., & Evans,
T. A . (2014). Microhabitats reduce animal's exposure to cli-
mate extremes. Global Change Biology, 20, 495– 503. https://doi.
Scheffers, B. R., Evans, T. A., Williams, S. E., & Edwards, D. P. (2014).
Microhabitats in the tropics buffer temperature in a globally coher-
ent manner. Biology Letters, 10, 20140819.
Schielzeth, H. (2010). Simple means to improve the interpretability of re-
gression coefficients. Methods in Ecolog y and Evolution, 1, 103– 113. 210X.2010.00 012. x
Seburn, C. N. L ., & Bishop, C. A. (2007). Ecology, conservation, and sta-
tus of reptiles in Canada. Blue Jay, 66(4).
bluej ay5610
Semlit sch, R. D. (1983). Burrowing ability and behavior of salamanders
of the genus Ambystoma. Canadian Journal of Zoology, 61, 616– 620. 082
Semlit sch, R. D. (2008). Differentiating migration and dispersal processes
for pond- breeding amphibians. The Jour nal of W ildlife Management,
72, 260– 267. 082
Sexton , O. J., Phillips, C., & Bramble, J. E. (1990). The effects of tempera-
ture and precipitation on the breeding migration of the Spotted
Salamander ( Ambystoma maculatum). Copeia, 19 90, 781– 787. 46443
Shaw, G . (1802). General zoology or systematic natural histor y. Volume III,
Part 1. Amp hibia. Thomas Davison.
Sheridan, J. A., Caruso, N. M., Apodaca, J. J., & Rissler, L. J. (2018). Shifts
in frog size and phenology: Testing prediction of climate change on
a widespread anuran using data from prior to rapid climate warm-
ing. Ecology an d Evolution, 8, 1316– 1327.
Strickland, J. C., Bahram, C. H., Harden, L. A., Pittman, S. E., Kern, M. M.,
& Dorcas, M. E. (2015). Life- history costs of reproductive behaviors
in a wetland- breeding amphibian. Journal of Freshwater Ecology, 30,
435– 444. 060.2014.982725
Sunday, J. M., Bates, A. E., & Dulvey, N. K. (2011). Global analysis of
thermal tolerance and latitude in ectotherms. Proceedings of the
Royal Society B: Biological Sciences, 278, 1823– 1830. https://doi.
Symonds, M. R. E., & Moussalli, A. (2011). A brief guide to model se-
lection, multimodel inference and model averaging in behavioural
ecology using Akaike's information criterion. Behavioral Ecology
and Socio biolog y, 65, 13– 21. ht tps://
5- 010- 1037- 6
Terhivuo, J. (1988). Phenology of spawning for the Common Frog (Rana
temporaria L.) in Finland from 1846 to 1986. Annales Zoologici
Fennnici, 25, 165– 175.
Todd, B. D., Scott, D. E., Pechmann, J. H. K., & Gibbons, J. W. (2011).
Climate change correlates with rapid delays and advancements
in reproductive timing in an amphibian communit y. Pro ceedings of
the Royal Society B: Biological Sciences, 278, 2191– 2197. ht tps://doi.
Tryjanowski, P., Sparks, T., Rybacki, M., & Berger, L. (2006). Is body
size of the water frog Rana esculenta complex responding to cli-
mate change? Naturwissenschaften, 93, 110– 113. https://doi.
org/10.1007/s0011 4- 006- 0085- 2
van Tienhoven, A. M., den Hartog, J. E., Reijns, R. A., & Peddemors,
V. M. (2007). A computer- aided program for pattern- matching
of natural marks on the spotted raggedtooth shark Carcharias
taurus. Journal of Applied Ecology, 44, 273– 280. https://doi.
org/10.1111/j.1365- 26 64. 200 6.01273.x
Vasconcelos, D., & Calhoun, A. J. K . (2004). Movement pat terns of
adult and juvenile Rana sylvatica (LeConte) and Ambystoma macu-
latum (Shaw) in three restored seasonal pools in Maine. Journal of
Herpetology, 38, 551– 561. 03A
Waite, T. A., & Strickland, D. (2006). Climate change and the demo-
graphic demise of a hoarding bird living on the edge. Proceedings of
the Royal Societ y B: Biological Sciences, 273, 2809– 2813. https://doi.
Walpole, A . A., Bowman, J., Tozer, D. C., & Badzinski, D. S. (2012).
Community- level re sponse to climate change: Shifts in anuran call-
ing phenology. Herpetological Conser vation and Biology, 7, 249– 257.
Walther, G.- R., Post, E., Convey, P., Menzel, A., Parmesan, C ., Beebee, T.
J. C., Fromentin, J.- M., Hoegh- Guldberg, O., & Bairlein, F. (2002).
Ecological responses to recent climate change. Nature, 416 , 389–
395. 389a
Whitford, W. G ., & Hutchison, V. H. (1963). Cutaneous and pulmonary gas
exchange in the Spot ted Salamander, Ambystoma maculatum. The
Biologic al Bulletin, 124, 344– 354.
Whitford, W. G., & Hutchison, V. H. (1967). Body size and metabolic
rate in salamanders. Physiological Zoology, 40, 127133. https://doi.
org/10.1086/physz ool.40.2.30152447
Windmiller, B. S. (1996). The pond , the forest, and the city: Spotted
Salamander ecolog y and conser vation in a human- dominated land-
scape. PhD dissertation, Tufts University.
Yartsev, V. V., & Kuranova, V. N. (2015). Seasonal dynamics of male and fe-
male reproductive s ystems in the Siberian Salamander, Salamandrella
keyserlingii (Caudata, Hynobiidae). A sian Herpetological Research, 6,
169– 183. /10.16373/ j.cnki.ahr.140033
Zoltai, S. C., Taylor, S. J., Jeglum, J. K., Mills, G. F., & Johnson, J. D.
(1988). Wetlands of boreal Canada in Wetlands of Canada, by
National Wetlands Working Group, Canada Committee on Ecological
Classification (Chapter 4, pp. 97– 154). Polyscience Publications Inc.
Additional supporting information may be found online in the
Suppor ting Information section.
How to cite this article: Moldowan, P. D., Tattersall, G. J., &
Rollinson, N. (2022). Climate- associated decline of body
condition in a fossorial salamander. Global Change Biology, 28,
1725– 173 9.
Climate change has already had wide‐ranging effects on populations, including shifts in species’ ranges, phenology, and body size. While some common patterns have emerged, the direction and magnitude of responses vary extensively among populations as well as across life stages within populations. Understanding consequences of climate change and predicting future responses at the population level require experimental tests of how warmer temperatures affect life history traits, including growth rate, development time, and reproductive output. Here, we tested how experimental warming affected life history from larval development and survival to adult reproductive maturity and investment in mole salamanders, Ambystoma talpoideum. We found that a small temperature increase (~1°C) experienced during larval development had complex consequences: density‐dependent effects on growth and body mass, density‐independent effects on fat storage, and no effects on survival and reproductive investment. While warming reduced growth rates, size at maturity, and fat storage, salamanders in both warmed and control conditions had similar survival and reproductive investment in their first year. However, costs of smaller body size and lower fat reserves may limit overwintering survival and/or future reproduction. Our study highlights differential effects of warming across life history traits and multifaceted population responses to climate change. This work motivates future studies to examine variation in response to climate change across life stages and life history traits.
Full-text available
As the current biodiversity crisis approaches levels comparable to the rates of the five historical mass extinctions, increasing attention has focused on how to stop or slow species loss and preserve ecosystem function. The impact of the loss of an individual species on communities and ecosystems is heterogeneous, however. Removing some species has negligible effects while the removal of others can be catastrophic. Metaphorically, the scenario can be likened to Jenga, a popular block‐balancing game in which players build a tower of wooden pieces, analogous to a dynamic ecosystem (de Ruiter et al. 2005).
Full-text available
Across all taxa, amphibians exhibit some of the strongest phenological shifts in response to climate change. As climates warm, amphibians and other animals are expected to breed earlier in response to temperature cues. However, if species use fixed cues such as daylight, their breeding timing might remain fixed, potentially creating disconnects between their life history and environmental conditions. Wood frogs Rana sylvatica are a cold‐adapted species that reproduce in early spring, immediately after breeding ponds are free of ice. We used long‐term surveys of wood frog oviposition timing in 64 breeding ponds over 20 yr to show that, despite experiencing a warming of 0.29°C per decade in annual temperature, wood frog breeding phenology has shifted later by 2.8 d since 2000 (1.4 d per decade; 4.8 d per °C). This counterintuitive pattern is likely the result of changes in the timing of snowpack accumulation and melting. Finally, we used relationships between climate and oviposition between 2000 and 2018 to hindcast oviposition dates from climate records to model longer‐term trends since 1980. Our study indicates that species can respond to fine‐grained seasonal climate heterogeneity within years that is not apparent or counterintuitive when related to annual trends across years.
Full-text available
There is considerable variation among studies that evaluate how amphibian populations respond to global climate change. We used 23 years of annual survey data to test whether changes in climate have caused predictable shifts in the phenology and population characteristics of adult spotted salamanders (Ambystoma maculatum) during spring breeding migrations. Although we observed year-to-year correlation between seasonal climate variables and salamander population characteristics, there have not been long-term, directional shifts in phenological or population characteristics. Warm winters consistently resulted in early migration dates, but across the 23-year study, there was no overall shift towards warmer winters and thus no advanced migration timing. Warm summers and low variability in summer temperatures were correlated with large salamander body sizes, yet an overall shift towards increasing body sizes was not observed despite rising summer temperatures during the study. This was likely due to the absence of long-term changes of within-year variation in summer temperatures, which was a stronger determinant of body size than summer temperature alone. Climate-induced shifts in population characteristics were thus not observed for this species as long-term changes in important seasonal climate variables were not observed during the 23-years of the study. Different amphibian populations will likely be more resilient to climate change impacts than others, and the probability of amphibians exhibiting long-term population changes will depend on how seasonal climate change interacts with a species' life history, phenology, and geographic location. Linking a wide range of seasonal climatic conditions to species or population characteristics should thus improve our ability for explaining idiosyncratic responses of species to climate change.
Full-text available
Dickson Lake, located in Algonquin Provincial Park, Ontario, is a remote, oligotrophic lake, where cyanobacterial blooms of the genus Dolichospermum (Ralfs ex Bornet & Flahault) P. Wacklin, L. Hoffmann and J. Komárek were reported for the first time in the fall of 2014, and subsequently in the late spring of 2015. To investigate the potential environmental triggers of these bloom events, we assessed long-term trends in water quality using a multi-proxy paleolimnological approach, examining sedimentary diatoms, chironomids, cladocerans, spectrally inferred chlorophyll a, and cyanobacterial akinetes preserved in a 210Pb-dated sediment core. Assemblage changes were modest in all biological proxies. A subtle increase in the abundance of warm-water chironomid taxa (Topt > 15 °C) commences in the year ~ 2000, with further increases in the most recent years of the sediment record (~ 2013–2015). End-of-summer volume-weighted hypolimnetic oxygen concentrations (CI-VWHO), inferred from chironomid remains, reveal a decline in oxygen concentrations over the last two decades coincident with the highest levels of sedimentary chlorophyll a and cyanobacterial akinetes in the sediment record. These paleolimnological findings corroborate observed reports of the onset of cyanobacterial blooms in Dickson Lake in late 2014 and are consistent with increasingly favourable bloom-forming conditions over the past few decades that are related to warmer air temperatures, sharp declines in wind speed, and a lengthening of the ice-free season by 2 weeks since 1975. It is plausible that late ice-out and a quick onset to stratification in 2014 may have resulted in incomplete spring mixing, early onset of hypolimnetic anoxia, and increased internal nutrient loading, that, occurring during a period when climate conditions were particularly ideal for cyanobacterial proliferation, may have fueled the unprecedented algal blooms in this remote lake. Collectively, the factors causing algal blooms in remote lakes such as Dickson Lake are not yet fully understood, and it is worrisome that with continued warming the triggering conditions may become a more common feature of Algonquin Park and other minimally impacted Boreal Shield lakes in the coming years.
Full-text available
Climate change has changed the phenologies of species worldwide, but it remains unclear how these phenological changes will affect species interactions and the structure of natural communities. Using a novel approach to analyse long‐term data of 66 amphibian species pairs across eight communities, we demonstrate that phenological shifts can significantly alter the interaction potential of coexisting competitors. Importantly, these changes in interaction potential were mediated by non‐uniform, species‐specific shifts in entire phenological distributions and consequently could not be captured by metrics traditionally used to quantify phenological shifts. Ultimately, these non‐uniform shifts in phenological distributions increased the interaction potential for 25% of species pairs (and did not reduce interaction potential for any species pair), altering temporal community structure and potentially increasing interspecific competition. These results demonstrate the potential of phenological shifts to reshape temporal structure of natural communities, emphasising the importance of considering entire phenological distributions of natural populations.
Recent climate change has been shown to affect phenotypic traits, such as body size and fecundity, in some animals. It is important to assess the response of a species to climate change for predicting a population’s future. We compared historic and contemporary body size and clutch size measurements in the lentic breeding salamander, Hynobius tokyoensis, collected from a wide range of latitudes in its geographical range and concluded that the species has gone through significant increases in body size and clutch size over the last four decades. Although a decrease in body size due to climate change is well documented for other species, reports of an increase in body size are rare. In addition, we found that increases in temperature and precipitation were constant regardless of latitude, but that the ratios of increase in body size and clutch size were greater in high-latitude populations. Our results suggest that, even within a species, the magnitude of the response to climate change depends on the geography of the population.
Global climate change and the associated erosion of habitat suitability are pervasive threats to biodiversity. It is critical to identify specific stressors to assess a species vulnerability to extinction, especially in species with distinctive natural histories. Here, we present a combination of field, laboratory, and modeling approaches to evaluate the potential consequences of climate change on two endemic, fossorial lizards species (Anniella geronimensis and Bipes biporus) from Baja California, Mexico. We also include soil type in our models to refine the suitable areas using our mechanistic models. Results suggest that both species are at high risk of extinction by global climate change based on the thermal habitat suitability. The forecast for species persistence is most grave under the RCP8.5 scenario. On the one hand, suitable habitat for A. geronimensis diminishes at its southern distribution, but potential suitable expands towards the north. On the other hand, the suitable habitat for B. biporus will contract significantly with a concomitant reduction in its potential distribution. Because both species have low mobility and are restricted to low elevation, the potential for elevational and latitudinal dispersal to mitigate extinction risk along the Baja California Peninsula is unlikely. In addition each species has specialized thermal requirements (i.e., stenothermic) and soil type preferences to which they are adapted. Our ecophysiological models in combination with the type of soil are fundamental in developing conservation strategies.
Winter climate warming is rapidly leading to changes in snow depth and soil temperatures across mid‐ and high‐latitude ecosystems, with important implications for survival and distribution of species that overwinter beneath the snow. Amphibians are a particularly vulnerable group to winter climate change because of the tight coupling between their body temperature and metabolic rate. Here, we used a mechanistic microclimate model coupled to an animal biophysics model to predict the spatially‐explicit effects of future climate change on the wintering energetics of a freeze‐tolerant amphibian, the Wood Frog (Lithobates sylvaticus), across its distributional range in the eastern United States. Our below‐the‐snow microclimate simulations were driven by dynamically downscaled climate projections from a regional climate model coupled to a one‐dimensional model of the Laurentian Great Lakes. We found that warming soil temperatures and decreasing winter length has opposing effects on Wood Frog winter energy requirements, leading to geographically heterogeneous implications for Wood Frogs. While energy expenditures and peak body ice content were predicted to decline in Wood Frogs across most of our study region, we identified an area of heightened energetic risk in the northwestern part of the Great Lakes region where energy requirements were predicted to increase. Because Wood Frogs rely on body stores acquired in fall to fuel winter survival and spring breeding, increased winter energy requirements have the potential to impact local survival and reproduction. Given the geographically variable and intertwined drivers of future under‐snow conditions (e.g. declining snow depths, rising air temperatures, shortening winters), spatially‐explicit assessments of species energetics and risk will be important to understanding the vulnerability of subnivium‐adapted species.
Ectotherm thermal physiology is frequently used to predict species responses to changing climates, but for amphibians, water loss may be of equal or greater importance. Using physical models, we estimated the frequency of exceeding the thermal optimum (T opt) or critical evaporative water loss (EWLcrit) limits, with and without shade‐ or water‐seeking behaviours. Under current climatic conditions (2002–2012), we predict that harmful thermal (>T opt) and hydric (>EWLcrit) conditions limit the activity of amphibians during ~70% of snow‐free days in sunny habitats. By the 2080s, we estimate that sunny and dry habitats will exceed one or both of these physiological limits during 95% of snow‐free days. Counterintuitively, we find that while wet environments eliminate the risk of critical EWL, they do not reduce the risk of exceeding T opt (+2% higher). Similarly, while shaded dry environments lower the risk of exceeding T opt, critical EWL limits are still exceeded during 63% of snow‐free days. Thus, no single environment that we evaluated can simultaneously reduce both physiological risks. When we forecast both temperature and EWL into the 2080s, both physiological thresholds are exceeded in all habitats during 48% of snow‐free days, suggesting that there may be limited opportunity for behaviour to ameliorate climate change. We conclude that temperature and water loss act synergistically, compounding the ecophysiological risk posed by climate change, as the combined effects are more severe than those predicted individually. Our results suggest that predictions of physiological risk posed by climate change that do not account for water loss in amphibians may be severely underestimated and that there may be limited scope for facultative behaviours to mediate rapidly changing environments.