ArticlePDF Available

Spatial variation in demographic processes and the potential role of hybridization for the future

Authors:

Abstract and Figures

Context Spatial variation in life history traits plays a crucial role in shaping the current and future dynamics of populations, particularly in systems where expanding hybrid zones could further shape population structure. The demographic responses of local populations to fine-scale habitat heterogeneity have consequences for species at a broader scale and demographic responses often vary across spatial scales. Objectives We evaluated spatial variation in population size and demographic traits (e.g., survival, individual growth, movement, and reproduction) of a montane endemic species of lungless terrestrial salamander across elevation and stream distance gradients representing broad and fine spatial scales, respectively. Methods Using 4 years of mark-recapture and count data from the Plethodon shermani × P. teyahalee hybrid system, whereby phenotypic hybrids occur at mid-elevations between low and high elevation congeners, we modeled demographic rates across environmental gradients and spatial scales using a combination of tools including individual growth models, and a spatially explicit Cormack-Jolly Seber model and Integrated Population Model. Results We found that high elevation animals grow faster and move more, especially far from streams, likely as a result of local microclimate conditions. Survival was highest but recruitment rates were lowest at low elevations and significantly declined with distance to stream. We also found that phenotypic hybrids at low elevations had higher survival probabilities. Conclusions Our study reveals nuanced spatial variation in demographic rates that differ in magnitude depending on the scale at which they are assessed. Our results also suggest that animals exhibit demographic compensation across abiotic gradients, underscoring the need for future conservation and management efforts to implement spatially explicit and dynamic strategies to match the demographic variation exhibited by populations across space.
Content may be subject to copyright.
Vol.: (0123456789)
1 3
Landsc Ecol
https://doi.org/10.1007/s10980-022-01503-y
RESEARCH ARTICLE
Spatial variation indemographic processes
andthepotential role ofhybridization forthefuture
MeaghanR.Gade · QingZhao·
WilliamE.Peterman
Received: 19 August 2021 / Accepted: 28 July 2022
© The Author(s), under exclusive licence to Springer Nature B.V. 2022
a montane endemic species of lungless terrestrial
salamander across elevation and stream distance
gradients representing broad and fine spatial scales,
respectively.
Methods Using 4years of mark-recapture and count
data from the Plethodon shermani × P. teyahalee
hybrid system, whereby phenotypic hybrids occur at
mid-elevations between low and high elevation con-
geners, we modeled demographic rates across envi-
ronmental gradients and spatial scales using a com-
bination of tools including individual growth models,
and a spatially explicit Cormack-Jolly Seber model
and Integrated Population Model.
Results We found that high elevation animals grow
faster and move more, especially far from streams,
likely as a result of local microclimate conditions.
Survival was highest but recruitment rates were
lowest at low elevations and significantly declined
with distance to stream. We also found that pheno-
typic hybrids at low elevations had higher survival
probabilities.
Conclusions Our study reveals nuanced spatial vari-
ation in demographic rates that differ in magnitude
depending on the scale at which they are assessed.
Our results also suggest that animals exhibit demo-
graphic compensation across abiotic gradients, under-
scoring the need for future conservation and man-
agement efforts to implement spatially explicit and
dynamic strategies to match the demographic varia-
tion exhibited by populations across space.
Abstract
Context Spatial variation in life history traits plays
a crucial role in shaping the current and future
dynamics of populations, particularly in systems
where expanding hybrid zones could further shape
population structure. The demographic responses of
local populations to fine-scale habitat heterogene-
ity have consequences for species at a broader scale
and demographic responses often vary across spatial
scales.
Objectives We evaluated spatial variation in popu-
lation size and demographic traits (e.g., survival,
individual growth, movement, and reproduction) of
Supplementary Information The online version
contains supplementary material available at https:// doi.
org/ 10. 1007/ s10980- 022- 01503-y.
M.R.Gade(*)· W.E.Peterman
School ofEnvironment andNatural Resources, The Ohio
State University, Columbus, OH43201, USA
e-mail: meaghan.gade@yale.edu
Present Address:
M.R.Gade
Department ofEcology andEvolutionary Biology, Yale
University, NewHaven, CT, USA
Q.Zhao
School ofNatural Resources, University ofMissouri,
Columbia, MO, USA
Q.Zhao
Bird Conservancy oftheRockies, FortCollins, CO, USA
Landsc Ecol
1 3
Vol:. (1234567890)
Keywords Climate change· Demography· Hybrid·
Life history· Plethodon· Spatially explicit integrated
population model· True survival
Introduction
Landscapes are inherently heterogeneous in both
biotic and abiotic features across spatial scales and
life history traits are expected to covary in response to
such heterogeneity (Wiens etal. 1993). Microhabitats
are a strong driver of life history variation because
environmental conditions experienced by individu-
als at a fine scale can be very different from condi-
tions measured at a broader scale. For example, the
breeding success of red-billed choughs (Pyrrhocorax
pyrrhocorax) differs between individual nest sites but
not across the larger landscape suggesting local-scale
variation in habitat likely drives this demographic
parameter (Reid et al. 2006; Sullivan and Vierling
2012). The role of microhabitats is particularly sali-
ent for small-bodied organisms such as amphibians or
small mammals, which interact with the landscape at
a fine-scale due to their restricted dispersal distance
and microhabitat requirements. As such, small-bod-
ied organisms often experience a microhabitat buff-
ered from broader landscape conditions (Riddell etal.
2021). Understanding fine-scale spatial heterogeneity
of demographic rates provides a clearer understand-
ing of species dynamics across a broader landscape
undoubtedly facilitating more informed species moni-
toring and management efforts; yet such information
can be difficult to ascertain.
Many significant challenges for measuring fine-
scale demographic rates exist including time, labor,
and cost intensive sampling, thereby limiting our
inference on species dynamics. Additionally, tradi-
tional analyses such as Cormack–Jolly–Seber (CJS)
models which statistically model demographic rates
from capture-mark-recapture (CMR) studies can-
not distinguish between emigration and mortality,
thereby severely biasing apparent survival estimates
(Marshall etal. 2004; Schaub and Royle 2014). The
development of spatially explicit CJS (sCJS) mod-
els ameliorate this bias by incorporating individual
spatial capture histories which provide estimates
of true, instead of the traditional apparent survival
(Schaub and Royle 2014). However, with this statisti-
cal advancement comes the challenging requirement
of large amounts of CMR data to obtain reliable
estimates. Integrated population models (IPM) have
been increasingly applied in ecological research as an
elegant solution to intensive, long-term data collec-
tion as they can estimate latent parameters that have
not been measured directly (Schaub and Abadi 2011).
IPMs also leverage numerous data sources including
CMR, repeated counts, reproduction, among others,
to account for multiple sources of uncertainty to esti-
mate demographic parameters more accurately and
precisely (Schaub and Abadi 2011). Moreover, spa-
tially explicit IPMs (SEIPM) allow for heterogeneity
in reproduction and survival processes based on the
spatial variation in the landscape, density depend-
ent processes, and dispersal between habitat patches
(Zhao 2020). SEIPMs are more flexible in data inputs
and can accommodate large amounts of the more
logistically simple count data with smaller amounts
of the more intensive CMR data. The advent of
advanced and integrated models provides opportunity
to more robustly assess fine-scale demographic rates
which will be particularly beneficial with the looming
threats of climate change.
The effects of climate change are spatially heter-
ogenous and hypothesized to promote life history
variation, changes, or tradeoffs (Urban etal. 2016).
To better understand climate change effects, we can
evaluate life history variation across elevational gra-
dients which offer a unique “space-for-time” substi-
tution because of the steep abiotic gradients across
small geographic distances (Pickett 1989; Blois etal.
2013). Life history tradeoffs do exist across eleva-
tion; for example, the cooler temperatures and greater
seasonality at high elevations often favor fewer but
higher condition offspring (Bears et al. 2009; Cam-
field etal. 2010). In addition to life history tradeoffs
across elevations that may be exacerbated or altered
by climate change, species are also tracking their
climate niches by moving up in elevations (Hick-
ling etal. 2006). In some species, a consequence of
upward range shifts has resulted in modified (e.g.,
Taylor etal. 2014) and/or novel (e.g., Garroway etal.
2010) hybrid zones between high and low elevation
congeners. Hybridization can provide a unique solu-
tion to climate change by providing genetic varia-
tion that may facilitate local adaptation within a sin-
gle to few generations (Chunco 2014). In fact, there
are reports of hybrid taxa outperforming parental
species, especially in extreme, marginal, or novel
Landsc Ecol
1 3
Vol.: (0123456789)
environments (Taylor etal. 2015). For organisms that
lack evolutionary capacity to keep pace with climate
change due to long generation times, low gene flow,
or narrow physiological tolerances (Ficke etal. 2007),
hybridization could be an important mechanism to
allow population persistence in the future (Chunco
2014).
Hybridization has been documented across many
taxa and is expected to increase in occurrence with
climate change (Chunco 2014). Amphibians, particu-
larly lungless salamanders in the family Plethodonti-
dae, are particularly susceptible to climate change
effects in part because of their reliance on cool and
wet conditions to facilitate gas exchange across the
skin (Feder and Londos 1984). In the Southern Appa-
lachian Mountains, where the highest abundance and
species richness of terrestrial Plethodon spp. exists,
there is a well-known hybrid zone between the low
elevation P. teyahalee and the high elevation replace-
ment species, P. shermani, in the Nantahala Moun-
tains of North Carolina (Hairston et al. 1992; Walls
2009). Hairston et al. (1992) showed that the phe-
notypic hybrid zone between these two species has
moved upward, with the proportion of individuals
with traits of P. teyahalee, increasing with elevation
over a 16-year period. Walls (2009) attributed this
upward shift to increasing temperatures across eleva-
tions. To understand the role hybridization may play
in the survival of species requires a detailed knowl-
edge of demographic vital rates of parental species
and their hybrids. Yet for many Plethodon salaman-
ders, we know virtually nothing about their fine-scale
demographic vital rates, especially as they relate to
spatial habitat variation (but see Caruso and Rissler
2018).
In this study, we use three complementary statisti-
cal models to understand fine–scale spatial patterns in
population size and demographic rates (e.g., individ-
ual growth, survival, dispersal, reproduction) of the
red-legged salamander P. shermani and their hybrid
populations. The Southern Appalachian Mountains
are characterized by stark abiotic gradients that
shape the distribution and abundance of terrestrial P.
shermani across scales. Gade and Peterman (2019)
found that at low elevations, where the regional cli-
mate is hot and dry, salamanders are distributed and
in the highest abundance near stream sides. Stream
sides offer a microhabitat that is cooler and moister
than the surrounding landscapes. Conversely, at high
elevations, salamanders are distributed more uni-
formly across the landscape because the regional cli-
mate at high elevations is more broadly cool and wet
(Gade and Peterman 2019). The dynamics of broad-
scale elevation and local-scale stream distance gra-
dients interact to drive the distribution of these sala-
manders, yet we do not understand how population
dynamics vary across these gradients. We hypoth-
esized that high elevation animals have higher indi-
vidual growth, survival, and reproduction due to the
cooler and more moist conditions in comparison to
low elevation. We also hypothesized that low-eleva-
tion salamanders have higher demographic vital rates
in areas near streams compared with areas far from
streams due to the buffered microclimate conditions
provided by streams. Finally, we expect phenotypic
hybrid animals will have different vital rate from
‘pure’ P. shermani. Using 4years of CMR data and
3years of count data, we attempt to provide an under-
standing of both fine-scale and landscape-scale spatial
dynamics of P. shermani population demographics.
Methods
Capture-Mark-Recapture surveys
We initiated a long-term spatial capture-mark-recap-
ture (sCMR) study in 2017 to assess salamander
demographics and life history across elevation and
moisture gradients. Twelve plots, 10m×10m in size,
gridded off in 1-m2 sections, were spatially arranged
to capture an elevation and moisture gradient across
Wayah Mountain, Macon County, North Carolina
(35.158, − 83.574). Specifically, two replicate sets
of three plots that were close (< 5 m), medium dis-
tance (100–150m), and far (> 190m) from a stream
were situated at low (900 m) and high elevations
(~ 1500 m), respectively, yielding a total of 12 plots
(Fig.1). Stream distances were selected to capture the
greatest heterogeneity in the landscape with respect to
stream-mediated moisture. To minimize extraneous
variation, forest stand, slope, aspect, and stream order
were standardized to the best extent possible among
plots (Supporting Information Table1).
We conducted nocturnal area-constrained surveys
of each plot whereby at least two observers exhaus-
tively searched and hand captured all species of
surface-active salamanders, recording their specific
Landsc Ecol
1 3
Vol:. (1234567890)
capture location to within 1 m2. While we captured
and marked all species encountered, the present study
focuses exclusively on P. shermani and P. sherm-
ani × P. teyahalee hybrids. We conducted at least
three (maximum = five) surveys each year from 2017
to 2020 during salamander active season, from May
to August. Surveys were conducted across haphazard
days to capture all possible weather conditions, and
at least 7days elapsed between each survey occasion.
Captured salamanders were placed in a sealable bag
with moist leaf litter and transported to Highlands
Biological Station, approximately 60 km away. At
each plot during each survey, we measured environ-
mental covariates including surface soil temperature,
soil temperature at 10cm below ground, air tempera-
ture, and relative humidity using an infrared ther-
mometer (Raytek MT4), soil temperature probe, and
Kestrel 5200.
584
1650
Elevation (m)
Fig. 1 Locations of the salamander count plots (red circles) and mark-recapture plots (black triangles) on Wayah Mountain in the
Nantahala National Forest in western North Carolina
Table 1 Parameter estimates for the growth model of P.
shermani
Parenthetical high and low refer to elevation-specific intercept
estimates. L is the asymptotic size, K is growth rate. CRI rep-
resents credible intervals and f indicates percent of the esti-
mates that lies to one side of zero. Stream distance, hybrid
score, and precipitation are continuous covariates of K with
elevation-specific estimates
Parameter Mean 2.5% CRI 97.5% CRI f
L (High) 66.01 64.38 67.85 1.00
L (Low) 73.51 71.67 75.50 1.00
K (High) 0.56 0.48 0.65 1.00
K (Low) 0.46 0.39 0.54 1.00
Stream (High) −0.03 −0.07 0.00 0.99
Stream (Low) 0.02 −0.01 0.05 0.89
Hybrid (High) 0.00 −0.04 0.36 0.54
Hybrid (Low) 0.02 −0.01 0.04 0.84
Precipitation (High) 0.21 0.10 0.32 1.00
Precipitation (Low) 0.06 −0.02 0.15 0.93
Landsc Ecol
1 3
Vol.: (0123456789)
Salamanders were housed at Highlands Biologi-
cal Station in a 10°C environmental chamber to limit
their metabolism (Connette 2014) until processing
occurred within 24 h of each survey. We anesthe-
tized each salamander in an Orajel solution (1g/1L)
and uniquely marked them with a visual implant
elastomer (Northwest Marine Technologies, LLC).
Visual implant elastomers are commonly used to tag
amphibians and have minimal effects on survival and
are more permanent than other marking techniques
(e.g. toe clipping; Bailey 2004; Oropeza–Sánchez
et al. 2020). We also recorded morphometric data
including snout-vent-length (SVL), sex, and a pho-
tograph of the dorsum of each individual using an
iPhone 6S + . Sex was determined based on the pres-
ence of a mental gland and or visual testes for males
(Gillette and Peterson 2001). If neither were apparent,
we made expert determination based on snout shape
with males having more square shape when compared
to females (Dawley 1992). We then returned salaman-
ders to their capture location (within 1m) within 48h
of initial capture.
Count surveys
We established 87-paired (total plots = 174) count
plots, each 3 × 3 m in size situated across elevation
(700–1550m) and stream-distance (0–250m) gradi-
ents(Fig.1). Each plot was visually surveyed for sur-
face active salamanders between 2130 and 0230 EST.
We determined the age class (juvenile or adult) of
each counted individual by expert visual estimation
of size based on snout-vent-length (SVL), whereby
adults were > 45mm and juveniles otherwise, which
could provide information for reproduction (i.e., age-
ratio). Plots were surveyed four times in 2017 and
once in both 2018 and 2019. More specific details of
the count data can be found in Gade and Peterman
(2019).
Quantifying hybrid phenotype
Our study region was within a known hybrid zone
between P. teyahalee and P. shermani, (Hairston etal.
1992; Highton and Peabody 2000; Walls 2009). Our
low elevation sites at 900 m overlapped the hybrid
zone but were not low enough to capture ‘pure
parental P. teyahalee whereas the high elevation sites
captured only ‘pure’ parental P. shermani. Plethodon
teyahalee are distinctly characterized by white spots
across their body and P. shermani are characterized
by red coloration on the legs. Thus, we quantified
hybrid phenotype from animals photographed in the
mark-recapture plots only by estimating the percent
of each leg covered with red coloration and count-
ing the number of white dots on 4 body regions: (1)
head, (2) anterior dorsum, (3) posterior dorsum, and
(4) tail (Supporting Information Fig.1A). We loaded
the percent red and spotting scores into a Principal
Components Analysis to estimate a composite meas-
ure of hybrid phenotype. Dimension 1 accounted for
49.2% of the variation, and we subsequently used
the Dimension 1 hybrid scores for each individual in
the individual growth and sCJS modeling discussed
below. Hybrid scores ranged from −4.28 to 2.20 with
negative values indicating more hybrid characteristics
and positive values indicating more ‘pure’ P. sherm-
ani characteristics (Supporting Information Fig.1B).
While phenotypic hybrids were only observed at low
elevations, there was variation in the hybrid score
across elevation due to the variation in leg coloration.
Importantly, our scoring method accounts only for
the observed phenotypic traits of hybrids, neglecting
genetic variability or gene combinations, and should
thus be considered a metric of coarse phenotypic
hybridization. Our phenotypic hybrid scoring proce-
dure is similar to those used by Hairston etal. (1992),
Walls (2009), and a long-term study at Coweeta
Hydrological Laboratory (J. Maerz, personal commu-
nication) which use a discrete scoring system for two
traits, white spots and red leg coloration, each ranging
from 0 to 3 (pure P. shermani scores 0 for white spots
and 3 for red legs). Our scoring procedure differs
by creating a continuous, multivariate score ranging
from more phenotypically hybrid (i.e., more P. teya-
halee characteristics) to ‘pure’ P. shermani.
Modeling approaches
Estimating individual growth athigh andlow
elevations
We estimated the effect of elevation, stream dis-
tance, precipitation, and hybrid score on individual
growth of captured P. shermani during the active sea-
son using Fabens formulation of the von Bertalanffy
growth model (Fabens 1965). We used the aver-
age SVL of an individual if it was captured multiple
Landsc Ecol
1 3
Vol:. (1234567890)
times over a single year and we only included ani-
mals recaptured between years. Since the sCMR plots
were arranged at high and low elevation, we included
elevation as a binary variable and stream distance
as a continuous variable. We calculated the cumula-
tive precipitation for the active season 01 May to 31
August for each survey year from the North Carolina
Climate Retrieval and Observation Network of the
Southeast Database (https:// clima te. ncsu. edu/ cronos).
The growth function was defined for individual i at
time t as:
where asymptotic size, L, was estimated as a function
of categorical elevation (ELE; low, high) as:
SVL0i,t is size at first capture and follows a Uni-
form distribution with a minimum of 10 and maxi-
mum of 80, L0 was estimated from a Normally dis-
tributed prior with a mean of 60 and precision of 0.01
while Li was estimated from a Normally distributed
prior with a mean of 0 and precision of 0.01. K rep-
resents the individual growth rate, and I is the annual
scaling interval between captures. We considered K
as a function of categorical elevation (high/low), and
continuous distance to stream (STR), precipitation
(PREC), and hybrid score (HYB) such that:
Given the locations of our mark-recapture plots at
high and low elevations, this model formulation was
necessary to allow estimates of stream distance at
the low and high elevation sites separately. We have
continuous variation in both stream distance and
hybrid scores, so wanted to understand the roles of
these factors at the high and low elevation plots. We
used vague normal priors for all growth rate covari-
ates with a mean of 0 and precision of 0.01. We fit the
model in JAGS (Plummer 2003) using jagsUI (Kell-
ner 2017). Following a burn in of 1000 and adapta-
tion phase of 5000, 5 MCMC chains were run for
12,000 iterations, thinned at a rate of 5 for a total pos-
terior sample of 18,220.
(1)
SVL
0i,t=SVL0i,t1+
{
LSVL0i,t1×
[
1exp
(
Ki,t×
I
365 )]}
(2)
L=L0+Li×ELEi
(3)
K
i,t=𝛽
[K]
0
+𝛽
[K]
1
×ELEi+𝛽
[K]
2
×STRi+𝛽
[K]
3
×ELEi×STRi+𝛽
[K]
4
×PRECt+𝛽
[K]
5
×HYBs+𝛽
[K]
6
×ELEi×HYB
s
We then used the growth rates to estimate time
to maturity for salamanders at both high and low
elevations, with the knowledge that hybrid animals
only exist at low elevation. At high elevations, we
estimated time to maturity based on the size of the
smallest gravid female P. shermani as estimated
by Connette (2014), 49.4mm, because we did not
capture any gravid females during our surveys. For
low elevations, we estimated the size of maturity
for P. shermani hybrids to be 57mm, based on the
nearest reported estimate of size at maturity of a
closely related species to P. teyahalee, P. glutino-
sis, in Giles County, VA (Highton 1962). We then
averaged the mature sizes of P. glutinosis and P.
shermani (53mm) because hybrid individuals often
exhibit traits intermediate of their parent species
(Seehausen 2004). We used the average hatchling
size at high elevation (19.15 mm) and low eleva-
tion (22.09mm) as the starting size to generate two
growth curves (Connette 2014).
Estimating survival andmovement probability athigh
andlow elevations
We used a spatial Cormack-Jolly-Seber (sCJS)
model (Schaub and Royle 2014) to examine the spa-
tial variation in survival and movement probability.
A sCJS model differs from traditional CJS models
by incorporating locations of each individual rela-
tive to the study area; in our study, spatial loca-
tions refer to the 1-m2 grid cell each salamander
was captured. The model assumes that death, birth,
immigration, and emigration could occur between,
but not within years and equal recapture probabil-
ity across all individuals and years. The true latent
status of an individual i at time t,
zi,t
(1 for alive and
0 for dead), was modeled with a Bernoulli distribu-
tion given the status in the previous time step and
survival probability such that:
Landsc Ecol
1 3
Vol.: (0123456789)
in which survival probability (
𝜙i,t
) was modeled as
a function of categorical elevation (high/low), con-
tinuous stream distance, and individual hybrid scores
such that:
Survival estimates were scaled to annual survival
by including a term indicating the length of time
between capture periods.
The change in location (G) for individual i between
time t (Gi,t) and time t + 1 (Gi,t+1) is modeled using a
random walk (Turchin and Thoeny 1993) such that:
in which the variance
𝜒2
i
is a function of covariates
such that:
Thus, the metric of dispersal is a measure of activ-
ity center variance between survey periods (Munoz
etal. 2016; Schaub and Royle 2014).
The observation process of the model which is
conditional on survival and presence in the study area
is indicated by a step function of whether individual
i is Space-for-Time Substitution as an Alternative to
Long-Term Studies inside (ri,t = 1) or outside (ri,t = 0)
of the study area at time t. The observation process is
then expressed as:
where pi,t is the recapture probability of individual i
at time t. We modeled recapture probability as a func-
tion of air temperature (ATEMP), relative humidity
(HUMID), and surface soil temperature (STEMP)
such that:
(4)
zi,t|
z
i,t1
Bernoulli
(
z
i,t1
×𝜙
i,t1)
(5)
logit
(𝜙i,t)=𝛽
[𝜙]
0+𝛽
[𝜙]
1×ELEi+𝛽
[𝜙]
2×STRi+𝛽
[𝜙]
3×ELEi×STRi
+
𝛽[𝜙]
4
×HYB
i
+𝛽[𝜙]
5
×ELE
i
×HYB
i
(6)
Normal
G
,𝜒
(7)
log
(𝜒2
i
)=𝛽
[
𝜒
]
0
+𝛽
[
𝜒
]
1
×ELE
i
+𝛽
[
𝜒
]
2
×STR
i
+𝛽
[
𝜒
]
3
×ELE
i
×STR
i
+𝛽
[
𝜒
]
4
×HYB
i
+𝛽
[
𝜒
]
5
×ELE
i
×HYB
s
(8)
y
i,t|
z
i,t
,r
i,t
Bernoulli
(
z
i,t
×r
i,t
×p
i,t)
(9)
logit(
p
i,t)
=𝛽
[
p
]
0
+𝛽
[
p
]
1
×ATEMP
t
+𝛽
[
p
]
2
×HUMID
t
+𝛽
[
p
]
3
×STEMP
t
We used vague normal priors for all covariates with a
mean of 0 and precision of 0.01. We fit the model in JAGS
(Plummer 2003) using jagsUI (Kellner 2017). Following
a burn in of 75,000 and adaptation phase of 60,000, 10
MCMC chains were run for 165,000 iterations, thinned at
a rate of 5 for a total posterior sample of 305,000.
Estimating demographic rates acrossacontinuous
landscape
We used a spatially explicit integrated population
model (SEIPM) to estimate initial population size,
spatiotemporally varying demography (survival, per
capita reproduction rate) in relation to habitat and
density covariates, as well as movement (emigration
and immigration) (Zhao 2020). This model jointly
analyzes count and CMR data by combining a spa-
tially explicit dynamic N-mixture model (Zhao etal.
2017) with a multistate capture-recapture model, in
which the counts of adults inform population size, the
counts of juveniles inform reproduction, and CMR
data inform survival and movement. The particular
benefit of the SEIPM in this study is that it allowed
us to leverage multiple data sets that spanned differ-
ent spatial, temporal, and information coverage for
more robust demographic estimation. Unlike the sCJS
model, SEIPM allowed for estimations along con-
tinuous stream and elevation gradients because of the
spatial distribution of the count plots.
For the purposes of the SEIPM, we split the land-
scape into 100 × 100m grids and combined the counts
of all plots within a grid to reduce the total number
of plots from 174 to 50 and reduce overall variation.
The model with reduced number of plots resulted in
improved mixing of MCMC chains and more efficient
sampling of posterior distributions, in comparison to
Landsc Ecol
1 3
Vol:. (1234567890)
the same model with the original number of plots. Six
of the 12 capture-recapture plots overlapped spatial
locations with count plots (Fig.1). We did not include
hybrid phenotype score in the SEIPM because we did
not photograph salamanders from the count plots.
The SEIPM is a hierarchical Bayesian model con-
taining a process model which describes how local
population sizes vary spatially and temporally and an
observation model which describes how count and
CMR data are obtained. The process model assumes
that the local population size in the first year at habi-
tat patch s, denoted Ns,1, follows a Poisson distribution
such that
Ns,1
Poisson
(
𝜆
[0]
s)
, in which
𝜆[0]
s
is a func-
tion of continuous elevation and stream distance such
that:
which the residuals
𝜀[0]
s
follow a Normal distribu-
tion with mean 0 and standard deviation
𝜎[0]
. For
subsequent years, the model assumes that varia-
tion in local population sizes are a consequence of
demographic processes including survival, repro-
duction, immigration, and emigration, and thus have
Ns
,
t=Rs
,
t+Ss
,
tEs
,
t+Is
,
t
, where Rs,t is the number
of individuals that are reproduced in habitat patch s
and year t, Ss,t is the number of individuals in habi-
tat patch s that survived from year t-1 to t, Es,t is the
number of individuals that emigrated from habitat
patch s in year t, and Is,t is the number of individu-
als that immigrated into habitat patch s in year t. We
further assumed that
Rs,t
follows a Poisson distribu-
tion such that
Rs
,
tPoisson(Ns
,
t
1
×𝛾s
,
t
1
)
, in which
the per capita reproduction rate
𝛾s,t1
(i.e., age-ratio)
is a function of local population size, elevation and
stream distance such that:
which the residuals
𝜀[R]
s,t
follows a Normal distribution
with mean 0 and standard deviation
𝜎[R]
.
We assumed that
Ss,t
follows a Binomial distribution
such that
Ss,tBinomial(Ns,t1,
𝜔
s,t1)
, in which the
(10)
log(
𝜆[0]
s)
=𝛽
[0]
0
+𝛽
[0]
1
×ELE
s
+𝛽
[0]
2
×STR
s
+𝛽
[0]
3
×ELE
s
×STR
s
+𝜀[0
]
s
(11)
log(
𝛾
s,t)
=𝛽
[R]
0
+𝛽
[R]
1
×
(
N
s,t
𝜆[0]
s)
𝜆[0]
s
+𝛽
[R]
2
×ELE
s
+𝛽
[R]
3
×STR
s
+𝛽
[R]
4
×ELE
s
×STR
s
+𝜀
[R]
s,t
#l
(
𝛾
s,t)
survival probability
𝜔s,t1
is a function of local popula-
tion size, elevation and stream distance:
which the residuals
𝜀[S]
s
,
t
follows a Normal distribution
with mean 0 and standard deviation
𝜎[S]
.
We further assumed that
Es,t
followed a Binomial
distribution such that:
in which
𝜅
is the probability of emigration given that
an individual survived. Immigrant individuals are
assumed to be emigrants from other patches, and the
number of immigrants was calculated as
where
Mj
,
s
,
t
is the number of individuals that
moved from grid j to grid s at year t and fol-
lows a Multinomial distribution such that
Mj,s,tMultinomial(Ej,t,wj,s)
. The movement rate
between grid cells j and s is a function of the distance
between the centroids of these two grids (
dj,s
):
where θ is the decay parameter and the error term
𝜀[M]
j,s
follows a Normal distribution with mean 0 and stand-
ard deviation
𝜎[M]
.
The observation model assumes that the age-struc-
tured counts are generated for adults and juveniles sepa-
rately without individual identification. Counts for adults
at site s and year t during survey k informs local popula-
tion size
Ns,t
such that
Y[a]
s,t,k
Binomial(N
s
,
t
,p
[obs]
t)
, and
counts for juveniles informs reproduction
Rs,t
such that
Y[j]
s,t,k
Binomial(R
s
,
t
,p
[obs]
t)
, in which the detection
(12)
log
(𝜔s,t)=𝛽
[S]
0+𝛽
[S]
1×(Ns,t𝜆
[0]
s)𝜆
[0]
s+𝛽
[S]
2×ELE
s
+𝛽[S]
3
×STR
s
+𝛽[S]
4
×ELE
s
×STR
s
+𝜀[S]
s,t
(13)
Es,t
Binomial
(
S
s,t
,𝜅
)
(14)
I
s,t=
n
j=1
Mj,s,t
(15)
w
j,s= exp
(
𝜃×dj,s+𝜀[M]
j,s
)
n
s=1
exp
(
𝜃×dj,s+𝜀[M]
j,s
)
Landsc Ecol
1 3
Vol.: (0123456789)
probability
p[obs]
t
is a random effect such that
logit(
p[obs]
t
)
follows a Normal distribution with mean
𝜇[obs]
and standard deviation
𝜎[obs]
. We included an offset
term (search area/100 m2) in both the adult and juvenile
observation sub-model to account for variable number of
count plots in each 100-m2 grid. The capture-recapture
process also assumes a capture probability p
[cap]
t
which
again is a random effect such that
logit(
p[cap]
t
)
follows a
Normal distribution with mean
𝜇[cap]
and standard devia-
tion
𝜎[cap]
. The likelihood of the individual encounter his-
tory is based on
𝜔s,t
,
𝜅
,
wj,s
, and
p[cap]
t
.
We used vague normal priors for all covariates with
a mean of 0 and precision of 0.01. We fit the model in
JAGS (Plummer 2003) using jagsUI (Kellner 2017). Fol-
lowing a burn in of 2500 and adaptation phase of 5000,
15 MCMC chains were run for 35,000 iterations, thinned
at a rate of 15 for a total posterior sample of 30,495.
Model fit for each of the above models was evaluated
based on the clear and consistent mixing of MCMC trace
plots and all parameters having a Gelman-Ruben (Rhat)
statistics below 1.05. For all estimates and covariates, we
considered support for the direction of the effect to be
meaningful if > 85% of the posterior density was to one
side of zero. Fully annotated code for each of the above-
described models can be found at https:// github. com/
meagh anreg ina/ Pleth odon- IPM- SCJS.
Results
Across 15 surveys over 4 years, we captured 1,712
individual P. shermani, recapturing 447 animals at
least once (range: 1–6; 26.1% recapture rate). The aver-
age SVL of all P. shermani captured was 53.66 mm
(range: 20.83–77.70 mm). The average SVL for adult
female P. shermani across all sites was 59.31 mm
(range: 37.15–77.70) and adult males was 57.98 (range:
40.24–74.45). Out of 753 animals captured and photo-
graphed across the 6 low elevation sites, 730 (97.50%)
exhibited some level of phenotypic hybrid traits (i.e.,
white spots and reduced red leg coloration). No high
elevation animals exhibited phenotypic hybrid traits
including white spots; however, the amount of red col-
oration on the legs of high elevation animals varied.
Estimating Individual Growth at High and Low
Elevations — Salamanders from high elevations had
higher growth rates (K) (mean: 0.56, 95% CRI: 0.48,
0.65) than those from low elevation (mean: 0.46,
95% CRI: 0.39, 0.54) (Table1). There was a negative
effect of distance from stream on growth at high ele-
vations (mean: −0.03, 95% CRI: −0.07, −0.01) but
the effect of stream distance was positive at low ele-
vation (mean = 0.02, 95% CRI: − 0.01, 0.06). There
was also a positive effect of precipitation at both
high (mean = 0.21, 95% CRI: 0.10, 0.32) and low
elevations (mean = 0.06 95% CRI: − 0.02, 0.15) on
growth. At low elevations, animals with more hybrid
characteristics had reduced growth (mean = − 0.02,
95% CRI: −0.05, 0.02), Salamanders at high eleva-
tion reached a lower asymptotic size (66.01 mm,
95% CRI: 64.28, 67.85) than those at low elevations
(73.52 mm, 95% CRI: 71.65, 75.50) (Table 1). The
mean estimated age of maturity at high elevation is
1.81years (95% CRI: 1.77, 1.84) and 2.02years (95%
CRI: 1.98, 2.12) at low elevations (Fig.2).
Estimating Survival and Movement Probability at
High and Low Elevations—Across our survey period,
the average annual survival of salamanders was 0.595
(95% CRI: 0.51, 0.69) at high elevation and 0.87 (95%
CRI: 0.74, 1.00) at low elevations. At high elevation,
there was no effect of stream distance on survival
(mean = 0.06, 95% CRI: −0.50, 0.52) and stream dis-
tance had a negative but marginal effect on survival
at low elevations (mean = − 1.05, 95% CRI: -9.77,
3.15). At low elevations, animals with more hybrid
characteristics had increased survival probability
(effect size mean = −3.02, 95% CRI: −7.54, −0.65)
(Fig.3A) and decreased movement probability (effect
size mean: 1.22, 95% CRI: 0.80, 1.64) (Fig.3C). At
high elevations, ‘pure’ P. shermani had increased
survival probability (mean = 1.16, 95% CRI: −0.06,
2.61) (Fig.3B) and decreased movement probability
(mean: −0.64, 95% CRI: −0.87, −0.29) (Fig.3D)
especially relative to the hybrids at low elevation.
Salamanders from high elevation had a mean move-
ment of 1.82m (95% CRI: 0.06, 6.50) while those at
low elevation had a mean movement of 3.32m (95%
CRI: 0.11, 12.00) (Fig.4A). There was greater vari-
ance in movement at low elevations (10.34; 95% CRI:
6.97, 14.83) than high elevations (3.11, 95% CRI:
2.58, 3.79) (Fig.4B). The distance from stream had
a marginally negative effect on movement at high ele-
vation (−0.08, 95% CRI: −0.23, 0.06) but a substan-
tial positive effect at low elevation (0.32, 95% CRI:
-0.04, 0.68) (Fig.4A). The average individual recap-
ture probability across all sites was 0.16 and recapture
Landsc Ecol
1 3
Vol:. (1234567890)
was positively influenced by relative humidity and
surface soil temperature and negatively influenced by
air temperature (Supporting Information Fig.2)
Estimating Demographic Rates Across a Continu-
ous Landscape—Our models show that initial popu-
lation size increased with elevation (mean = 0.53,
95% CRI: 0.31, 0.75) and decreased with the dis-
tance from stream (mean = −0.12, 95% CRI: −0.36,
0.13) without an interaction between elevation and
stream distance (mean = 0.02, 95% CRI: − 0.25,
0.28). The average survival across all sites was 0.43
(95% CRI: 0.32, 0.51) and was not density depend-
ent (mean = 0.18, 95% CRI: − 1.88, 1.90). There
was a significant negative interaction between eleva-
tion and stream distance on survival, meaning that
survival was highest at low elevation near streams
and declined with the distance from stream at low
elevations (Fig.5). Per capita reproduction averaged
0.25 (95% CRI: 0.15, 0.42) across the landscape
and was negatively influenced by population density
(mean = −0.97, 95% CRI: −1.79, −0.21). There was
also a significant positive interaction between eleva-
tion and stream distance on per capita reproduction
rate whereby reproduction increases with elevation
and at low elevation, reproduction decreases with
stream distance (Fig. 6). Emigration probabilities
were low (mean = 0.01, 95% CRI: 0.00, 0.07), indi-
cating a low probability that an individual will move
from their 100-m grid cell; however, the low value of
the decay parameter in dispersal (mean = 0.01, 95%
CRI: 0.00, 0.01) indicates that individuals moved a
long distance if they dispersed. The average detection
probability for animals across all count plots was 0.27
and average recapture probability across all sMRC
sites was 0.34.
Discussion
Estimating demographic variation in populations
across space provides insight into how landscape het-
erogeneity shapes ecological processes (Gurevitch
et al. 2016). Using elevation and stream distance to
capture both broad and fine scale temperature and
moisture gradients, we reveal nuanced spatial patterns
in population demography. Our results show that
demographic rates vary across elevation broadly as
well as across fine-scale stream distance gradients
especially at low elevation where regional climate
is more spatially heterogeneous. We estimated that
movement and survival are highest at low elevations
and at low elevations, both individual growth and
movement increase with stream distance. High eleva-
tion salamanders grow faster and reach respective
size at maturity approximately two and half months
earlier than low elevations animals, but low elevation
animals reach a larger sizeoverall. The larger size at
low elevation is likely attributed to the larger hatch-
ling sizes relative to P. shermani and higher growth
rates of phenotypic hybrid animals found at our low
Fig. 2 Growth projection
for high elevation (dotted
line) and low elevation
(solid line) Plethodon
shermani based on theindi-
vidual growth model. The
horizontal lines represent
minimum size (in SVL) at
maturity for high elevation
‘pure’ P. shermani (blue)
and low elevation pheno-
typic hybrid P. shermani
(red)
Low Elevation
High Elevation
Landsc Ecol
1 3
Vol.: (0123456789)
elevation sites. Hairston (1983) similarly noted that
P. teyahalee reach a larger maximum size than P. jor-
dani (a closely related species to P. shermani).
Interestingly, our estimates for age at maturity
across elevations were lower than previous estimates
for other Plethodon (Hairston 1983, Howard and
Maerz 2022), and somewhat overlap estimates from
Connette (2014), who measured P. shermani from
the same elevation and estimated time to maturity for
males at 1.69 (95% CRI 1.59–1.79) and females at
2.53 (CRI: 2.38–2.68) years. It is increasingly appar-
ent that time to maturity in Plethodon is spatially and
temporally variable and closely tied to environmental
conditions, namely precipitation (Connette et al.
2015; Caruso & Rissler 2018; Howard & Maerz
2022). However, large adult P. shermani are dispro-
portionately active under drier conditions which may
introduce sampling bias in growth and maturity esti-
mates (Connette etal., 2015). Generally, there is very
limited data on vital rates within or across Plethodon
spp., and much of the available data is likely biased
from time, space, and methodological constraints
(Howard and Maerz 2022). Future work should
focus on long-term, spatially heterogeneous vital rate
estimation.
Fig. 3 The effect of hybrid
score on survival prob-
ability at A low elevations
and B high elevations and
movement probability at
C low elevations and D
high elevations from 5000
random draws of the pos-
terior of the sCJS model.
Negative hybrid scores
indicate more characteris-
tics of P. teyahalee and thus
represent more phenotypic
hybridization
Landsc Ecol
1 3
Vol:. (1234567890)
Our results also suggest that precipitation has a
much larger effect on individual growth at high ele-
vations relative to low elevations (Table 1). Larger
bodied salamanders, such as the phenotypic hybrids
at low elevations, have higher resistance to water
loss (Riddell & Sears 2015) which may result in less
sensitivity to precipitation (Table1). However, given
our model estimates and previous work (Caruso
etal. 2020; Caruso and Rissler, 2018; Connette etal.
2015), it appears that precipitation influences individ-
ual growth rates more strongly than spatial variation
in either moisture or temperature at any scale. This
could have significant demographic consequences
especially considering the projected increasein vari-
ation of precipitation in the Southern Appalachi-
ans by the end of the century (Kunkel et al. 2020).
Demographic consequences could include increased
variation in annual growth rates leading to changes
in time to maturity, reduced lifetime fecundity, and
changes in population size structures in the future.
Continued monitoring of these populations is neces-
sary to determine the demographic consequences of
climate-change induced precipitation changes.
Terrestrial plethodontid salamanders are known
to have small home ranges and limited dispersal
and movement, often moving less than 3m in their
lifetime (Cabe et al. 2007; Caruso & Rissler 2018;
Muñoz et al. 2016). We found low movement prob-
ability across the landscape as a whole, but P. sherm-
ani from high elevations moved less and with less
variation than low elevation hybrid animals. Move-
ment rates are likely tied to temperature. Low eleva-
tions are warmer than high elevations, especially
far from streams, where we estimated the highest
Parameter Estimate Parameter Estimate
Mean Distance (Low)
Mean Distance (High)
Stream Distance (Low)
Stream Distance (High)
Variance in
Movement (Low)
Variance in
Movement (High)
AB
Fig. 4 Posterior mean densities for the A mean movement dis-
tance and stream distance covariates of P. shermani and B the
variance in movement estimated in the sCJS. The parenthetical
High and Low refer to the two elevation intercepts estimated.
The shaded region indicates 50% of the posterior
Landsc Ecol
1 3
Vol.: (0123456789)
movement probability (Fig.4A) (Gade & Peterman
2019). Temperature appears to increase surface activ-
ity probability in Plethodon (Gade etal. 2020), other
salamander families (Johnson etal. 2010), and other
ectotherms like lizards and fish (Xiang et al. 1996,
Petty et al. 2012) potentially because of increased
metabolic demands, which undoubtably climate
change will exacerbate. Increased surface activity to
supplement metabolic demand may come at a cost to
reproductive effort, a pattern we observed in animals
at low elevations (Fig. 6). Movement may also be
tied to the quality or quantity of resources available
(poorer quality prey at low elevations) or population
density; decreased population density at low eleva-
tions may lead to intraspecific competitive release and
increased movement probability (Gade & Peterman
2019).
Our complementary sCJS and SEIPM models
both estimate that annual P. shermani survival prob-
ability decreases at higher elevations. Caruso and
Rissler (2018) described similar elevational trends
in survival in a closely related species, P. montanus,
and showed survival was positively associated with
temperature. Temperature is often linked with higher
energy assimilation and survival rates, but only to an
optimal temperature, which when surpassed, rates
decline rapidly (Caruso & Rissler 2018; Clay & Gif-
ford 2018). In our system, we see further evidence
for this quadratic-like relationship whereby survival
probability dropped to 0.19 in animals found at low
elevations and far from streams (Fig.5), where condi-
tions are the hottest and driest on the landscape (Gade
& Peterman 2019), creating high vapor pressure defi-
cits that exacerbate water loss and provide unsuit-
able conditions for terrestrial salamanders (Peterman
& Gade 2017; Riddell etal. 2017). The hot and dry
conditions far from streams provide a combination of
abiotic stressors that salamanders may not be able to
survive. Additionally, although there appears to be a
strong positive effect of stream distance on survival at
high elevations (Fig.5), this trend is due to increased
uncertainty in model estimates (Supplemental Fig.4)
as these regions were not well sampled due to logisti-
cal constraints (Gade & Peterman 2019).
We found a positive relationship between per
capita reproduction and elevation (Fig. 6) which
may be a contributing factor to the higher abundance
of P. shermani at high elevations observed by Gade
and Peterman (2019) as well as other related Pleth-
odon spp (Caruso et al., 2020). Higher reproductive
rates at high elevations may serve as a compensatory
mechanism for the lower estimated survival prob-
ability at high elevations (Muths etal. 2011; Villellas
etal. 2015; Buckley etal. 2021). Negative covariance
between vital rates can occur to maintain demo-
graphic performance over environmental gradients
800
1000
1200
1400
Elevation (m)
Stream Distance (m)
0
50
100
150
0
0.2
0.4
0.6
0.8
1
Survival
Fig. 5 Interaction between elevation and stream distance on
the survival probability of P. shermani estimated from the
SEIPM. Survival probability is highest at low elevation and
decreases with stream distance at lower elevations
800
1000
1200
1400
Elevation (m)
Stream Distance (
m)
0
50
100
150
0
1
2
3
4
5
Reproduction Rate
Fig. 6 Interaction between elevation and stream distance on
per capita reproduction rate of P. shermani estimated from the
SEIPM. Per capita reproduction increases with elevation and at
low elevations, decreases with stream distance
Landsc Ecol
1 3
Vol:. (1234567890)
and contribute to local adaptation to environmen-
tal conditions (Angilletta et al. 2003). Interestingly,
studies from other taxa find per capita reproduction
to be lower at high elevations. However, lungless
salamanders are evolutionarily tied and rely on the
cooler and wetter conditions that high elevations of
the Southern Appalachians provide, likely promot-
ing adaptation forhigher reproductive rates at higher
elevations. We observed other vital rate tradeoffs that
may contribute to energy  compensation and local
adaptation across temperature and moisture gradients.
For example, lungless salamanders must constantly
regulate and invest energy into water loss resistance,
which is achieved through capillary bed regeneration
or regression (Riddell et al. 2019). With low eleva-
tion animals moving more and experiencing higher
desiccation probability, it is possible that a significant
amount of energy is diverted from reproduction and
invested into water loss resistance for survival pur-
poses. As such, animals at the cooler and wetter high
elevations can invest less energy to water loss regula-
tion, allowing more energy investment into reproduc-
tion. While these tradeoffs are likely driven and main-
tained by microclimatic trends, there does appear to
be a climatic breaking point where conditions surpass
any ability of salamanders to maintain vital rates, as
we saw in areas at low elevations far from streams
(Figs. 5, 6). The complex spatial patterns of vital
demographic rates provide valuable insight into com-
pensations between life-history and energy allocation
across a landscape with multiple abiotic gradients.
Our study focuses on the spatial variation in demo-
graphic rates across abiotic gradients that represent
the range of environmental conditions experienced
by individuals. While this spatial approach offers
an understanding into the role of abiotic gradients
on demography, it does neglect the temporal aspect
of demographic rates. Due to the somewhat limited
timescale of our study (4years) especially relative to
the lifespan of Plethodon (~ 10 years; Staub, 2016),
we do not capture enough temporal variation to make
the mostrobust estimates. Importantly, demographic
rates are influenced at different time scales. For exam-
ple, survival is often affected by short term expo-
sure to extreme environmental conditions whereas
fecundity and reproduction tend to be an integrated
response to longer-term environmental stochastic-
ity (Levins 1968; Gilchrist 1995; Buckley et al.
2021). Our present study substitutes space-for-time to
understand how elevation and stream distance gradi-
ents influence population vital rates at fine and broad
spatial scales. As such, we are limited in our inference
of how stochasticity in weather events (i.e., droughts,
heat waves) affect vital rates in the long term. Contin-
ued monitoring of these populations will be necessary
to disentangle temporal effects on demography.
Hybridization may allow species to survive in
rapidly changing environments by promoting pheno-
typic and genotypic novelty more quickly than typi-
cal evolutionary mechanisms (Arnold 1997; Riese-
berg et al. 2003). Hybridization with populations
preadapted to emerging environments offers a unique
avenue for species to survive climate change (Chunco
2014). In our study, phenotypic hybrids at low eleva-
tions had a higher survival probability than “pure” P.
shermani (Fig. 3A) at the same elevation. Warmer
temperatures predicted with climate change may
benefit hybrid P. shermani because P. teyahalee are
preadapted to occupy warmer and drier microhabitats
(Hairston etal. 1992), and larger bodied salamanders
have higher resistance to water loss. Additionally, P.
teyahalee tend to occupy warmer but moist micro-
habitats that minimize water loss (Farallo etal. 2020),
which may potentially balance the costs of higher
temperatures. Importantly, however, survival esti-
mates decline with stream distance at low elevation,
where temperatures are high and moisture is low, a
microhabitat that appears unsuitable even for hybrid
individuals (Fig.5). Thus, hybridization may only be
beneficial if moist microhabitats are maintained in
the future. Despite our results suggesting phenotypic
hybrids move less than more phenotypically pure ani-
mals at low elevations (Fig.3) broadly, salamanders
at low elevation move more than those at high eleva-
tion which is likely tied to the higher temperatures at
low elevation. Similarly, Walls (2009) attributed the
upward expansion of the hybrid zone to tempera-
ture, providing evidence that over time with climate
change the hybrid zone will likely continue to expand
upwards. Contemporary hybridization may facilitate
plethodontid diversification through increased spe-
ciation and decreased extinction rates (Patton et al.
2020) and may be a creative strategy for rapid adapta-
tion to novel stressors expected with climate change.
It will be critical to continue monitoring hybrid popu-
lations to understand the role that hybridization may
play in salamander population dynamics and species
persistence.
Landsc Ecol
1 3
Vol.: (0123456789)
Our study demonstrates the importance of evaluat-
ing how scales interact to drive species demographic
rates. Robustly sampled fine-scale estimates provide
valuable information but may not necessarily repre-
sent landscape-scale patterns. Estimation of landscape-
scale demographic rates has previously been limited
by time, costs, and logistics. SEIPM’s provide a rigor-
ous and coherent framework for maximally leveraging
disparate data sources and a promising tool for future
spatial demographic studies. Ultimately, our results
underscore the need for future studies to evaluate vital
rates and life history across relevant environmental gra-
dients (Howard and Maerz 2022). Finally, our study
shows that as climate change progresses, salamanders
will likely implement life history tradeoffs to moderate
their responses. However, tradeoffs and demographic
compensation have limits. Once temperature and mois-
ture thresholds are surpassed, salamanders will no
longer be able to persist and therefore, conservation
efforts should focus on maintaining both cool and moist
microhabitats.
Acknowledgements We thank the many people who contrib-
uted to field collection including Philip Gould, Katie Greene,
Addison Hoven, Andrew Wilk, Nicole Cooke, Winter Gary,
Kate Donlon, Amelia Gray, and Kasey Foley. We also thank
members of the Peterman Lab group, Suzanne Gray, Lau-
ren Pintor, Chris Tonra, and three anonymous reviewers for
their helpful comments on earlier version on the manuscript.
We also thank the staff at Highlands Biological Station for their
logistical support.
Author contributions MRG and WEP: conceived the ideas
and designed the methodology. MRG: collected the data. MRG
and QZ: analyzed the data and MRG lead the writing of the
manuscript. All authors contributed equally to drafts of the
manuscript and have given final approval for publication.
Funding This work was supported by The Herpetologists
League E.E. Williams grant and multiple Grant-in-aids from
Highlands Biological Station.
Data availability All data is available on GitHub link:
https:// github. com/ meagh anreg ina/ Pleth odon- IPM- SCJS .
Code availability Code is available on GitHub: https://
github. com/ meagh anreg ina/ Pleth odon- IPM- SCJS
Declarations
Conflicts of interest The authors have no conflicts of interest
to declare.
Ethical approval This research was conducted following The
Ohio State University IACUC protocol #2016A00000026, with
permission from the US Forest Service Permit number NAN
45716.
References
Angilletta MJ, Wilson RS, Navas CA, James RS (2003) Trade-
offs and the evolution of thermal reaction norms. Trends
Ecol Evol 18(5):234–240
Arnold M (1997) Natural Hybridization and evolution. Oxford
University Press, Oxford
Badyaev AV, Ghalambor CK (2001) Evolution of life histories
along elevational gradients: trade-off between parental
care and fecundity. Ecology 82:2948–2960
Bailey LL (2004) Evaluating elastomer marking and photo
identification methods for terrestrial salamanders: mark-
ing effects and observer bias. Herpetol Rev. 35:38–41
Balasubramaniam P, Rotenberry JT (2016) Elevation and lati-
tude interact to drive life-history variation in precocial
birds: a comparative analysis using galliformes. J Anim
Ecol 85:1528–1539
Barton NH (2013) Does hybridization influence speciation? J
Evol Biol 26:267–269
Bears H, Martin K, White GC (2009) Breeding in high-eleva-
tion habitat results in shift to slower life-history strategy
within a single species. J Anim Ecol 78:365–375
Blois JL, Williams JW, Fitzpatrick MC etal (2013) Space can
substitute for time in predicting climate-change effects on
biodiversity. Proc Natl Acad Sci U S A 110:9374–9379.
Buckley LB, Schoville SD, Williams CM (2021) Shifts in the
relative fitness contributions of fecundity and survival in
variable and changing environments. J Exp Biol. https://
doi. org/ 10. 1242/ jeb. 228031
Cabe PR, Page RB, Hanlon TJ etal (2007) Fine-scale popu-
lation differentiation and gene flow in a terrestrial sala-
mander (Plethodon cinereus) living in continuous habitat.
Heredity (edinb) 98:53–60
Camfield AF, Pearson SF, Martin K (2010) Life history varia-
tion between high and low elevation subspecies of horned
larks Eremophila spp. J Avian Biol 41:273–281
Caruso NM, Rissler LJ (2018) Demographic consequences of
climate variation along an elevational gradient for a mon-
tane terrestrial salamander. Popul Ecol. https:// doi. org/ 10.
1002/ 1438- 390X. 1005
Caruso NM, Staudhammer CL, Rissler LJ (2020) A demo-
graphic approach to understanding the effects of climate
on population growth. Oecologia 193:889–901
Chown SL, Hodgins KA, Griffin PC et al (2015) Biologi-
cal invasions, climate change and genomics. Evol Appl
8:23–46
Chunco AJ (2014) Hybridization in a warmer world. Ecol Evol
4:2019–2031
Chunco AJ, Jobe T, Pfennig KS (2012) Why do species co-
occur? a test of alternative hypotheses describing abiotic
differences in sympatry versus allopatry using spadefoot
toads. PLoS ONE. https:// doi. org/ 10. 1371/ journ al. pone.
00327 48
Landsc Ecol
1 3
Vol:. (1234567890)
Clay TA, Gifford ME (2018) Energy assimilation in a polymor-
phic salamander, Plethodon angusticlavius. Source J Her-
petol 52:269–272
Connette GM (2014) Individual, Population and Landscape-
scale effecs of timber harvest on the red-legged salaman-
der (Plethodon shermani). Dissertation 1–126
Connette GM, Crawford JA, Peterman WE (2015) Climate
change and shrinking salamanders: alternative mecha-
nisms for changes in plethodontid salamander body size.
Glob Chang Biol 21:2834–2843
Dawley EM (1992) Sexual dimorphism in a chemosensory
system: the role of the vomeronasal organ in salamander
reproductive behavior. Copeia 1992:113
Fabens AJ (1965) Properties and fitting of the Von Berta-
lanffy growth curve. Growth 29:265–289
Farallo VR, Muñoz MM, Uyeda JC, Miles DB (2020) Scal-
ing between macro- to microscale climatic data reveals
strong phylogenetic inertia in niche evolution in pletho-
dontid salamanders. Evolution 74:979–991
Feder ME, Londos PL (1984) Hydric Constraints upon
Foraging in a Terrestrial Salamander, Desmognathus
ochrophaeus. Oecologia 64:413–418
Ficke AD, Myrick CA, Hansen LJ (2007) Potential impacts
of global climate change on freshwater fisheries. Rev
Fish Biol Fish 17:581–613
Gade MR, Connette GM, Crawford JA, Hocking DJ, Maerz
JC, Milanovich JR, Peterman WE (2020) Predicted
alteration of surface activity as a consequence of climate
chage. Ecology 101(11):e03154
Gade MR, Peterman WE (2019) Multiple environmental gra-
dients influence the distribution and abundance of a key
forest-health indicator species in the Southern Appala-
chian Mountains. Landsc Ecol, USA
Garroway CJ, Bowman J, Cascaden TJ etal (2010) Climate
change induced hybridization in flying squirrels. Glob
Chang Biol 16:113–121
Gilchrist GW (1995) Specialists and generalists in changing
environments. I. Fitness landscapes of thermal sensitiv-
ity. Am Nat 146:252–270
Gillette JR, Peterson MG (2001) The benefits of transpar-
ency: Candling as a simple method for determining sex
in red-backed salamanders (Plethodon cinereus). Herpe-
tol Rev 32:233–235
Gurevitch J, Fox GA, Fowler NL, Graham CH (2016) Land-
scape demography: population change and its drivers
across spatial scales. Q Rev Biol 91:459–485
Hairston NG (1983) Growth, survival and reproduction of
plethodon jordani : trade-offs between selective pres-
sures. Copeia 4:1024–1035
Hairston NG, Wiley RH, Smith CK, Kneidel KA (1992) The
dynamics of two hybrid zones in Appalachian salaman-
ders of the genus Plethodon. Evolution 46:930–938
Hickling R, Roy DB, Hill JK etal (2006) The distributions of
a wide range of taxonomic groups are expanding pole-
wards. Glob Chang Biol 12:450–455
Highton R (1962) Geographic variation in the life history of
the slimy salamander. Copeia. https:// doi. org/ 10. 2307/
14411 85
Highton R, Peabody RB (2000) Geographic Protein Variation
and Speciation in Salamanders of the Plethodon jordani
and Plethodon glutinosus Complexes in the Southern
Appalachian Mountains with the Description of Four
New Species. In: Bruce RC, Jaeger RG, Houck LD (eds)
The Biology of Plethodontid Salamanders. Springer,
US, pp 31–93
Howard JS, Maerz JC (2022) Review and synthesis of esti-
mated vital rates for terrestrial salamanders in the family
Plethodontidae. Ichthyl Herpetol 109(4):929–939
Johnson JR, Johnson BB, Shaffer HB (2010) Genotype and
temperature affect locomotor performance in a tiger sal-
amander hybrid swarm. Funct Ecol 24:1073–1080
Kellner K (2017) jagsUI: a wrapper around ‘rjags’ to stream-
line ‘JAGS.’ R Package Version 1(4):9
Kozak KH, Wiens JJ (2006) Does niche conservatism pro-
mote speciation? a case study in North American sala-
manders. Evolution 60:2604
Kunkel KE, Easterling DR, Ballinger A, etal (2020) North
Carolina Climate Science Report. North Carolina Insti-
tute for Climate Studies, 233 pp.
Levins R (1968) Evolution in Changing Environments.
Princeton Universtiy Press, Princeton
Marshall MR, Diefenbach DR, Wood LA, Cooper RJ
(2004) Annual survival estimation of migratory song-
birds confounded by incomplete breeding site-fidelity:
study designs that may help. Anim Biodivers Conserv
27:59–72
Muñoz DJ, Hesed KM, Grant EHC, Miller DAW (2016) Evalu-
ating within-population variability in behavior and fitness
for the climate adaptive potential of a dispersal-limited
species. Plethodon Cinereus Ecol Evol. https:// doi. org/ 10.
1002/ ece3. 2573
Muths E, Scherer RD, Pilliod DS (2011) Compensatory effects
of recruitment and survival when amphibian populations
are perturbed by disease. J Appl Ecol 48:873–879
Oropeza-Sánchez MT, Sandoval-Comte A, García-Bañuelos P
etal (2020) Use of visible implant elastomer and its effect
on the survival of an endangered minute salamander.
Anim Biodivers Conserv 43:187–190
Patton AH, Margres MJ, Epstein B et al (2020) Hybridizing
salamanders experience accelerated diversification. Sci
Rep 10:1–12
Peterman WE, Gade MR (2017) The importance of assessing
parameter sensitivity when using biophysical models: a
case study using plethodontid salamanders. Popul Ecol.
https:// doi. org/ 10. 1007/ s10144- 017- 0591-4
Pickett S (1989) Space-for-time substitution as an alternative to
long-term studies. In: Likens GE (Eds) Long-Term Stud-
ies in Ecology: Approaches and Alternatives. Springer.
New York. pp 110–135
Plummer M (2003) JAGS: A program for analysis of Bayesian
graphical models using Gibbs sampling. In: Proceedings
of the 3rd International Workshop on Distributed Statisti-
cal Computing (DSC2003). pp 20–22
Reid JM, Bignal EM, Bignal S etal (2006) Spatial variation in
demography and population growth rate: the importance
of natal location. J Anim Ecol 75:1201–1211
Riddell EA, Apanovitch EK, Odom JP, Sears MW (2017) Phys-
ical calculations of resistance to water loss improve pre-
dictions of species range models. Ecol Monogr 87:21–33
Riddell EA, Iknayan KJ, Hargrove L etal (2021) Exposure to
climate change drives stability or collapse of desert mam-
mal and bird communities. Science 371:633–636
Landsc Ecol
1 3
Vol.: (0123456789)
Riddell EA, Roback E, Wells C etal (2019) Thermal cues drive
plasticity of desiccation resistance in montane salaman-
ders with implications for climate change. Nat Commun.
https:// doi. org/ 10. 1017/ cbo97 81316 105542. 007
Riddell EA, Sears MW (2015) Geographic variation of resist-
ance to water loss within two species of lungless salaman-
ders : implications for activity. Ecosphere 6:1–16
Rieseberg LH, Raymond O, Rosenthal DM etal (2003) Major
ecological transitions in wild sunflowers facilitated by
hybridization. Science 301:1211–1216
Schaub M, Abadi F (2011) Integrated population models: A
novel analysis framework for deeper insights into popula-
tion dynamics. J Ornithol 152:S227–S237
Schaub M, Royle JA (2014) Estimating true instead of apparent
survival using spatial Cormack-Jolly-Seber models. Meth-
ods Ecol Evol 5:1316–1326
Seehausen O (2004) Hybridization and adaptive radiation.
Trends Ecol Evol 19:198–207
Staub NL (2016) The age of plethodontid salamanders: a short
review on longevity. Copeia 104:118–123
Sullivan SMP, Vierling KT (2012) Exploring the influences of
multiscale environmental factors on the American dipper
Cinclus mexicanus. Ecography (cop) 35:624–636
Taylor SA, Larson EL, Harrison RG (2015) Hybrid zones: win-
dows on climate change. Trends Ecol Evol 30:398–406
Taylor SA, White TA, Hochachka WM et al (2014) Climate-
mediated movement of an avian hybrid zone. Curr Biol
24:671–676
Todd Petty J, Hansbarger JL, Huntsman BM, Mazik PM (2012)
Brook trout movement in response to temperature, flow,
and thermal refugia within a complex Appalachian river-
scape. Trans Am Fish Soc 141:1060–1073
Turchin P, Thoeny WT (1993) Quantifying dispersal of south-
ern pine beetles with mark-recapture experimentsand a
diffusion model. Ecol Appl 3:187-198
Urban MC, Bocedi G, Hendry AP etal (2016) Improving the
forecast for biodiversity under climate change. Science
353:8466
Villellas J, Doak DF, García MB, Morris WF (2015) Demo-
graphic compensation among populations: What is it,
how does it arise and what are its implications? Ecol Lett
18:1139–1152
Walls SC (2009) The role of climate in the dynamics of a
hybrid zone in Appalachian salamanders. Glob Chang
Biol 15:1903–1910
Wiens JA, Stenseth NC, Van HB, Ims RA (1993) Ecological
mechanisms and landscape ecology. Oikos 66:369
Willis BL, Van Oppen MJH, Miller DJ etal (2006) The role
of hybridization in the evolution of reef corals. Annu Rev
Ecol Evol Syst 37:489–517
Xiang J, Weiguo D, Pingyue S (1996) Body temperature, ther-
mal tolerance and influence of temperature on sprint speed
and food assimilation in adult grass lizards, Takydromus
septentrionalis. J Therm Biol 21:155–161
Zhao Q (2020) On the sampling design of spatially explicit
integrated population models. Methods Ecol Evol. https://
doi. org/ 10. 1111/ 2041- 210X. 13457
Zhao Q, Royle JA, Boomer GS (2017) Spatially explicit
dynamic N-mixture models. Popul Ecol 59:293–300
Publisher’s Note Springer Nature remains neutral with regard
to jurisdictional claims in published maps and institutional
affiliations.
Springer Nature or its licensor holds exclusive rights to this
article under a publishing agreement with the author(s) or other
rightsholder(s); author self-archiving of the accepted manuscript
version of this article is solely governed by the terms of such
publishing agreement and applicable law.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Population models depend on reliable estimates of vital rates, yet for many taxa, such estimates and how they vary in response to spatial or temporal environmental gradients are lacking. The goal of this review was to determine whether existing estimates of vital rates for temperate, direct-developing plethodontid salamanders (subfamily Plethodonti-nae) could be used to reasonably project values for populations or species where such estimates are lacking, or whether current estimates are biased in a manner that limits their utility. We synthesized current knowledge of stage-specific survival rates, age-and size-at-maturity at first clutch, and clutch frequency. We tested for expected correlations among published vital rates (e.g., age at maturity and survival) and between vital rates and factors such as body size or latitude. We used matrix projection models to judge whether published estimates were reasonably possible for stable salamander populations. The largest number of published vital rates were for clutch size, clutch frequency (proportion of females with clutches), size at maturity or first clutch, and age at maturity or first clutch, though the latter vital rate is primarily inferred from size distribution and growth rate data. Among these vital rates, we found expected correlations with body size and latitude suggesting these rates were reasonable and somewhat predictable among species or populations. In contrast, there were few estimates of egg hatch rate or juvenile or adult survival. Hatch and survival rate estimates were widely variable; estimates seldom included measures of uncertainty, but when uncertainty measurements were included, they were generally high. Based on projection models, few survival estimates were likely unbiased or realistic for stable populations given other salamander vital rates. Additionally, few studies quantified how vital rates vary with spatial or temporal environmental gradients. We outline the key knowledge gaps that limit basic demographic modeling of these remarkably common, influential, and otherwise well-studied salamanders, and make recommendations for future research efforts.
Article
Full-text available
Microhabitat matters Understanding how our warming climate affects vulnerable species is of paramount importance. However, predicting responses is complicated because species are complex and may adapt or respond in distinct ways. Riddell et al. compared a century-old dataset on species richness in the Mojave against modern surveys to measure climate-related changes in bird and small mammal communities. They found little change in mammal richness or occupancy but large declines across birds. They attribute these differences to differences in microclimate opportunities: Specifically, mammals can mitigate temperature impacts through burrowing, whereas birds are generally more exposed. Science , this issue p. 633
Article
Full-text available
Amphibian life history traits are affected by temperature and precipitation. Yet, connecting these relationships to population growth, especially for multiple populations within a species, is lacking and precludes our understanding of amphibian population dynamics and distributions. Therefore, we constructed integral projection models for five populations along an elevational gradient to determine how climate and season affect population growth of a terrestrial salamander Plethodon montanus and the importance of demographic vital rates to population growth under varying climate scenarios. We found that population growth was typically higher at the highest elevation compared to the lower elevations, whereas varying inactive season conditions, represented by the late fall, winter and early spring, produced a greater variation in population growth than varying active season conditions (late spring, summer, and early fall). Furthermore, survival and growth were consistently more important, as measured by elasticity, compared to fecundity, and large females had the greatest elasticity compared to all other body sizes. Our results suggest that changing inactive season conditions, especially those that would affect the survival of large individuals, may have the greatest impact on population growth. We recommend future experimental studies focus on the inactive season to better elucidate the mechanisms by which these conditions can affect survival.
Article
Full-text available
It is important to understand metapopulation dynamics and underlying demographic processes in heterogeneous landscapes. Traditionally demographic parameters are estimated using capture‐recapture data that can be difficult to collect. Spatially explicit dynamic N‐mixture models allow inference for demographic parameters, including dispersal, using count data of unmarked animals, but these models have only been shown effective under constant demographic parameters and dispersal between adjacent local populations. In this study I aimed to compensate the weakness of spatially explicit dynamic N‐mixtures and multistate capture‐recapture models by jointly analyzing count and capture‐recapture data. This spatially explicit integrated population model allows for spatiotemporal variation of demographic parameters in relation to environmental and density covariates and dispersal between any local populations. I conducted simulations to evaluate this model (1) for species with distinct life histories under different detection and capture probabilities, (2) when spatial sampling intensity varied, (3) when the length of survey period varied, (4) when the robust sampling design was adopted or not, and (5) when auxiliary information is partially available. I also provided an empirical example of Gadwall (Mareca strepera) metapopulation dynamics in North American prairies. The results showed that the model provided unbiased parameter estimates under a variety of ecological and sampling conditions, even when the spatial sampling intensity of capture‐recapture survey was low (20% of the patches) with the complement of count data (≥ 60% of the patches). Also, the model only required a relatively short survey period (6~8 years) to provide unbiased inferences. The robust sampling design was not necessary for the model to provide unbiased inferences when spatial counts were intense, but became critical when spatial counts were sparse. Parameter estimates remain unbiased when auxiliary information is partially available. The model showed that Gadwall had low emigration probability (11.8%) but could disperse more than 200 km. Based on the results, I provide recommendations about the tradeoff between spatial sampling intensity, length of survey period, and the use of the robust sampling design when applying this model in real world studies. The model could have wide applications in the interface of metapopulation ecology and landscape ecology.
Article
Full-text available
Utilización de implantes visibles de elastómero y sus efectos en la supervivencia de una salamandra enana en peligro de extinción El estudio poblacional de especies amenazadas requiere técnicas de marcaje que no afecten a la supervivencia de los individuos. En este estudio, evaluamos la efectividad de los implantes visibles de elastómero (VIE, en sus siglas en inglés) en la identificación y supervivencia de individuos de la salamandra Parvimolge townsendi. Comparamos tres grupos de salamandras sometidos a diferentes tratamientos: Intervenido, Intervención Simulada y Testigo. No se observaron diferencias de mortalidad entre los grupos (con dos, ninguno y un individuo, respectivamente), pero se observó la migración del implante en cuatro de los 10 individuos intervenidos. Aunque los VIE no tienen un efecto significativo en la supervivencia, debería considerarse la migración de los implantes antes de emplearlos en estudios poblacionales.
Article
Full-text available
Whether hybridization generates or erodes species diversity has long been debated, but to date most studies have been conducted at small taxonomic scales. Salamanders (order Caudata) represent a taxonomic order in which hybridization plays a prevalent ecological and evolutionary role. We employed a recently developed model of trait-dependent diversification to test the hypothesis that hybridization impacts the diversification dynamics of species that are currently hybridizing. We find strong evidence supporting this hypothesis, showing that hybridizing salamander lineages have significantly greater net-diversification rates than non-hybridizing lineages. This pattern is driven by concurrently increased speciation rates and decreased extinction rates in hybridizing lineages. Our results support the hypothesis that hybridization can act as a generative force in macroevolutionary diversification.
Article
Organisms respond to shifts in climate means and variability via distinct mechanisms. Accounting for these differential responses and appropriately aggregating them is central to understanding and predicting responses to climate variability and change. Separately considering fitness components can clarify organismal responses: fecundity is primarily an integrated, additive response to chronic environmental conditions over time via mechanisms such as energy use and acquisition, whereas survival can be strongly influenced by short-term, extreme environmental conditions. In many systems, the relative importance of fecundity and survival constraints changes systematically along climate gradients, with fecundity constraints dominating at high latitudes or altitudes (i.e. leading range edges as climate warms), and survival constraints dominating at trailing range edges. Incorporating these systematic differences in models may improve predictions of responses to recent climate change over models that assume similar processes along environmental gradients. We explore how detecting and predicting shifts in fitness constraints can improve our ability to forecast responses to climate gradients and change.
Article
Wildlife are faced with numerous threats to survival, none more pressing than that of climate change. Understanding how species will respond to changing climate behaviorally, physiologically, and demographically is a cornerstone of many contemporary ecological studies, especially for organisms, such as amphibians, whose persistence is closely tied to abiotic conditions. Activity is a useful parameter for understanding the effects of climate change because activity is directly linked to fitness as it dictates foraging times, energy budgets, and mating opportunities. However, activity can be challenging to directly measure, especially for secretive organisms like plethodontid salamanders which only become surface active when conditions are cool and moist due to their anatomical and physiological restrictions. We estimated abiotic predictors of surface activity for the seven species of the Plethodon jordani complex. Five independent data sets collected from 2004‐2017 were used to determine the parameters driving salamander surface activity in the present‐day, which were then used to predict potential activity changes over the next 80 years. Average active season temperature and vapor pressure deficit were the strongest predictors of salamander surface activity and, without physiological or behavioral modifications, salamanders were predicted to exhibit a higher probability of surface activity during peak active season under future climate conditions. Temperatures during the active season likely do not exceed salamander thermal maxima to cause activity suppression and, until physiological limits are reached, future conditions may continue to increase activity. Our model is the first comprehensive field‐based study to assess current and future surface activity probability. Our study provides insights into how a key behavior driving fitness may be affected by climate change.