ArticlePDF Available

The push and pull of climate change causes heterogeneous shift in avian elevational ranges


Abstract and Figures

Projected effects of climate change on animal distributions primarily focus on consequences of temperature and largely ignore impacts of altered precipitation. While much evidence supports temperature‐driven range shifts, there is substantial heterogeneity in species' responses that remains poorly understood. We resampled breeding ranges of birds across three elevational transects in the Sierra Nevada Mountains, USA, that were extensively surveyed in the early 20th century. Presence–absence comparisons were made at 77 sites and occupancy models were used to separate significant range shifts from artifacts of false absences. Over the past century, rising temperature pushed species upslope while increased precipitation pulled them downslope, resulting in range shifts that were heterogeneous within species and among regions. While 84% of species shifted their elevational distribution, only 51% of upper or lower range boundary shifts were upslope. By comparison, 82% of range shifts were in a direction predicted by changes in either temperature or precipitation. Species were significantly more likely to shift elevational ranges than their ecological counterparts if they had small clutch sizes, defended all‐purpose territories, and were year‐round residents, results that were in opposition to a priori predictions from dispersal‐related hypotheses. Our results illustrate the complex interplay between species‐specific and region‐specific factors that structure patterns of breeding range change over long time periods. Future projections of increasing temperature and highly variable precipitation regimes create a strong potential for heterogeneous responses by species at range margins.
Twentieth century climate change and resultant expected range shifts for three resurveyed regions of the Sierra Nevada of California. (a) Elevational transects showing locations of resurvey sites superimposed on topography and Grinnell's life zones. (b) Changes in average annual temperature (red arrows) and precipitation (blue arrows) between 1900–1930 and 1980–2006 for survey sites in each region. Arrows point from the average historical climate at a site to the average modern climate at the site. (c) Differences between the elevation of each site and the nearest neighbor elevation based on 20th century changes in temperature (red arrows) and precipitation (blue arrows). Positive differences in nearest neighbor elevation (arrows above the black line) indicate that a species at a particular site would need to shift upslope to stay as close as possible to historic climatic average conditions at that site. (d) Agreement and disagreement between temperature- and precipitation-based nearest neighbor elevation change predictions for survey sites (black dots) in each region. The number of survey sites with concordant (blue and green, see legend) and discordant (red and purple, see legend) predictions are shown as numbers within each set of boxes. The number of concordant and discordant sites summed across all regions (located next to the legend) illustrates that species at 60% of sites (n = 77) experienced opposing climatic pressures from temperature and precipitation over the 20th century.
Content may be subject to copyright.
The push and pull of climate change causes
heterogeneous shifts in avian elevational ranges
*Department of Environmental Science, Policy & Management, University of California, Berkeley, CA 94720, USA, Museum of
Vertebrate Zoology, University of California, Berkeley, CA 94720, USA, Department of Integrative Biology, University of
California, Berkeley, CA 94720, USA
Projected effects of climate change on animal distributions primarily focus on consequences of temperature and lar-
gely ignore impacts of altered precipitation. While much evidence supports temperature-driven range shifts, there is
substantial heterogeneity in species’ responses that remains poorly understood. We resampled breeding ranges of
birds across three elevational transects in the Sierra Nevada Mountains, USA, that were extensively surveyed in the
early 20th century. Presenceabsence comparisons were made at 77 sites and occupancy models were used to sepa-
rate significant range shifts from artifacts of false absences. Over the past century, rising temperature pushed species
upslope while increased precipitation pulled them downslope, resulting in range shifts that were heterogeneous
within species and among regions. While 84% of species shifted their elevational distribution, only 51% of upper or
lower range boundary shifts were upslope. By comparison, 82% of range shifts were in a direction predicted by
changes in either temperature or precipitation. Species were significantly more likely to shift elevational ranges than
their ecological counterparts if they had small clutch sizes, defended all-purpose territories, and were year-round res-
idents, results that were in opposition to a priori predictions from dispersal-related hypotheses. Our results illustrate
the complex interplay between species-specific and region-specific factors that structure patterns of breeding range
change over long time periods. Future projections of increasing temperature and highly variable precipitation
regimes create a strong potential for heterogeneous responses by species at range margins.
Keywords: birds, California, climate change, elevational range shift, occupancy models, precipitation, Sierra Nevada
Received 9 February 2012 and accepted 27 June 2012
Global climate change is expected to shift the distribu-
tions of organisms, with predicted consequences of
large-scale extinctions (Thomas et al., 2004; Colwell
et al., 2008; La Sorte & Jetz, 2010; Sinervo et al., 2010)
and formation of novel assemblages (Roy et al., 1996;
Williams, 2007; Stralberg et al., 2009). With few excep-
tions (e.g., Bonebrake & Mastrandrea, 2010; Crimmins
et al., 2011), the focus has been on attribution of pole-
ward and upslope shifts of species ranges to increasing
temperature (Thomas & Lennon, 1999; Parmesan &
Yohe, 2003; Root et al., 2003; Moritz et al., 2008; Chen
et al., 2011). However, this belies a more complex real-
ity; up to 25% of the species examined worldwide have
shifted ranges equatorially or downslope, and ranges of
an additional 1030% of species have not shifted
(Parmesan & Yohe, 2003; Chen et al., 2011). A limited
understanding of the complexities underlying these
heterogenous, species-specific responses to climate
change prevents accurate predictions of response to
future climate change (Buckley et al., 2010).
Numerous hypotheses have arisen seeking to explain
variation among species’ responses to climate change.
Downslope movements could be caused by climate-
induced changes to competitive species interactions
(Hughes, 2000; Lenoir et al., 2010), land-use changes
(Archaux, 2004), changes in nontemperature environ-
mental gradients (Tingley et al., 2009; Zimmermann
et al., 2009; Crimmins et al., 2011), and stochastic fluctu-
ations in population size (Lenoir et al., 2010). Nonmov-
ement, or range stability, may result from adaptation of
local populations to new climates (Rodrı
´guez-Trelles &
´guez, 1998; Parmesan et al., 2005), an inability to
disperse (Davis et al., 1998), an insufficient amount of
climate change to push species out of their fundamental
niche (Tingley et al., 2009), or a temporal lag in move-
ment response (Svenning et al., 2008). Linking these
hypotheses are the different intrinsic ecological traits
Correspondence: Morgan W. Tingley, Woodrow Wilson School,
Robertson Hall, Princeton University, Princeton, NJ 08544, USA,
tel. + 510 590 6526, fax + 609 258 6082, e-mail: mtingley@princeton.
©2012 Blackwell Publishing Ltd 3279
Global Change Biology (2012) 18, 3279–3290, doi: 10.1111/j.1365-2486.2012.02784.x
held by diverse species assemblages. Past studies had
moderate success relating species-specific patterns of
range movements to life history and species’ traits, such
as body size, habitat requirements, and fecundity
(Perry et al., 2005; Moritz et al., 2008; Po
¨yry et al., 2009).
In most cases, however, the statistical power to deter-
mine these relationships was greatly limited (Angert
et al., 2011), and results may have been obscured by the
effects of false absences on occurrence data (Link &
Nichols, 1994; Ke
´ry, 2004; Ke
´ry et al., 2006). Given the
multitude of mechanisms and processes potentially
driving range change, temporally and spatially repli-
cated surveys across taxa are needed to test these
hypotheses (Parmesan et al., 2005).
We quantified the impacts of temperature and precip-
itation changes over the last century on ranges of breed-
ing birds along three broad elevational transects located
primarily in US National Park and US Forest Service
lands in the Sierra Nevada mountains of California
(Fig. 1a). Systematic surveys were originally done 80
100 years ago by Joseph Grinnell and colleagues.
Although annual minimum and maximum tempera-
tures have increased on average between 1 and 2 °C
throughout the Sierra Nevada over the last century
(Bonfils et al., 2008), there is substantial spatial variation
in both temperature and precipitation change (Fig. 1b).
Substantial warming occurred in the southern and cen-
tral Sierra Nevada, while the northern portion experi-
enced either marginal warming (low elevations) or local
cooling (high elevations). Precipitation generally
increased over the same time period, with the greatest
change in the north and at higher elevations (Fig. 1b).
Temperature and precipitation changes over the past
century in many parts of the Sierra Nevada yield oppos-
ing expectations as to whether species should move ups-
lope or downslope if species shift distributions to track
their climatic niche (Grinnell, 1917; Brown et al., 1996;
Tingley et al., 2009). Increased temperature should push
species upslope, but increased precipitation should pull
them downslope. This arises because precipitation gen-
erally increases with elevation in montane systems, but
temperature decreases (Fig. 1b). As a result, temporal
increases in precipitation will shift precipitation-based
climatic niches downslope, whereas warming will shift
thermal-based climatic niches upslope. Topography and
localized weather conditions can create nonlinearities in
these general patterns, leading to variable and local
effects of climate on species. To consider the potential
for temperature and precipitation to alternatively push
and pull breeding bird distributions in alternate direc-
tions, we formulated species- and limit-specific predic-
tions for how upper- and lower-elevation boundaries
should shift independently over time, as determined
separately by changes in average mean temperature and
precipitation. Expectations were derived from the
difference in elevation between each survey site and the
nearest modern-climatic neighbor within each region
(see Methods, Table S1). The majority of sites have mod-
ern-temperature nearest neighbors at higher elevations,
as expected given the average warming trend, but many
sites have modern precipitation nearest neighbors at
lower elevations (Fig. 1c). Expected range shifts from
temperature and precipitation changes over the past
century were in opposing directions at 60% of survey
sites, primarily at higher elevations (Fig. 1d).
Given the recent climatic history within the Sierra
Nevada, our goals were to: (1) test for an overall
upward shift in elevational range expected by average
warming; (2) test whether directional shifts in elevation
were better explained by site-specific temperature or
precipitation changes; and (3) determine if species’
traits can explain variation in movement responses. To
quantify range change from historical data, we conser-
vatively excluded shifts that may be due to false
absences or that represent normal, minor range fluctua-
tions. In addition, we hypothesized that dispersal- and
colonization-related traits should be positively related
to range movements (Angert et al., 2011), including
migration during the nonbreeding season, large clutch
size, large home range size, small body size, low territo-
riality, and a generalist diet.
Materials and methods
Collection and sampling of field data
Bird observations were collected as part of the Grinnell
Resurvey Project (Moritz et al., 2008; Tingley et al., 2009), a
multiyear endeavor to revisit historical vertebrate sampling
sites throughout the state of California. A total of 77 his-
torical survey sites were revisited, as well as seven addi-
tional sites that were sampled only contemporarily. Sites
were distributed across three elevational cross-sections of
the Sierra Nevada mountain range, from north to south:
Lassen, Yosemite, and Southern Sierra (Fig. 1a). All sites
contained characteristic ‘west-slope Sierran’ vegetation com-
munities (i.e., Central Valley riparian, oak woodland, Sier-
ran mixed conifer, yellow pine forest, lodgepole and true
fir forests, and alpine) and excluded Great Basin and Sono-
ran desert habitat. Elevational ranges were 802751 m for
Lassen sites, 653226 m for Yosemite sites, and 613356 m
for Southern Sierra sites. Over 87% of survey sites were
located on permanently protected lands, with 66% of sites
owned by the federal government.
Historical bird observations were conducted between 1911
and 1929 as part of regular biotic surveys by Joseph Grinnell,
Tracy Storer, and seven other researchers from the Museum of
Vertebrate Zoology (MVZ), University of California, Berkeley
(Grinnell & Storer, 1924; Grinnell et al., 1930). Survey effort
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3280 M. W. TINGLEY et al.
Southern Sierra
(b) (c)
All sites,
all transects
Resurvey site
Lower Sonoran
Upper Sonoran
18 41
Southern Sierra Yosemite Lassen
0 500 1500 2500 3500
Site Elevation (m)
NNE (m)
Site Elevation (m)
Precipitation (cm)
500 1500 2500 3500
0 500 1500 2500 3500
NNE (m)
Precipitation (cm)
0 500 1500 2500 3500
0 500 1500 2500 3500
NNE (m)
0 500 1500 2500 3500
Precipitation (cm)
Fig. 1 Twentieth century climate change and resultant expected range shifts for three resurveyed regions of the Sierra Nevada of Cali-
fornia. (a) Elevational transects showing locations of resurvey sites superimposed on topography and Grinnell’s life zones. (b) Changes
in average annual temperature (red arrows) and precipitation (blue arrows) between 19001930 and 19802006 for survey sites in each
region. Arrows point from the average historical climate at a site to the average modern climate at the site. (c) Differences between the
elevation of each site and the nearest neighbor elevation based on 20th century changes in temperature (red arrows) and precipitation
(blue arrows). Positive differences in nearest neighbor elevation (arrows above the black line) indicate that a species at a particular site
would need to shift upslope to stay as close as possible to historic climatic average conditions at that site. (d) Agreement and disagree-
ment between temperature- and precipitation-based nearest neighbor elevation change predictions for survey sites (black dots) in each
region. The number of survey sites with concordant (blue and green, see legend) and discordant (red and purple, see legend) predic-
tions are shown as numbers within each set of boxes. The number of concordant and discordant sites summed across all regions
(located next to the legend) illustrates that species at 60% of sites (n=77) experienced opposing climatic pressures from temperature
and precipitation over the 20th century.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
was focused on different geographical areas for certain years,
with primary sampling periods being: 19241928 for Lassen;
19151919 for Yosemite; and 1911 for Southern Sierra. Histori-
cal surveys were taken opportunistically using a precursor of
the line-transect method (Bibby et al., 2000). Surveys listed all
species encountered, providing reliable detection/nondetec-
tion data. All survey observations, as well as details on loca-
tion, extent, duration, and weather, were recorded in field
notebooks held at the MVZ (available online at http://bscit. A total of 266 historic sur-
veys were used as baseline data on avian occurrence, with each
of the 77 sites surveyed up to 17 times (median =3 visits).
Modern resurveys were done using point counts along a
line transect that followed, as closely as possible, the same sur-
vey route taken by historical observers. When field notes
lacked details to establish exact historic routes, routes were
placed following named geographical features in habitats
matching those described in the field notes to achieve our goal
of determining whether sites in an elevation band were occu-
pied. Whenever possible we matched habitat descriptions in
the original field notes. Sites with mixed land-use were subject
to the same class of current and historical land-use (e.g., graz-
ing or agriculture). For resurveys we used variable-distance
point counts (Ralph et al., 1995) lasting 7 min. Observation
points were separated by a minimum of 250 m and varied in
number per site depending on the extent of the historical route
(median =10 points over 2.5 km). Five trained primary
observers collected data as part of 1- or 2-person survey teams
with temporal sampling as follows: Lassen, 20062007; Yosem-
ite, 20032004; and Southern Sierra, 20082009. A total of 251
modern surveys were conducted at 84 sites, with each site sur-
veyed a maximum of 5 times (median =3). All modern occur-
rence records are archived online in the MVZ’s collections
database (, and are available
by arrangement through the authors.
Bird surveys characterized elevational ranges of species
during the breeding season. Historically, survey dates ranged
between 25 March and 2 October, with 87% conducted
between 1 May and 31 August. Modern resurveys visited sites
around the same time as historical surveys, but were concen-
trated within the breeding season (dates ranged between 3
May and 25 August). Sites were typically surveyed within one
breeding season (79% historic, 80% modern), and most were
surveyed within 1 week (historic 66%, modern 53%); the
remaining sites (2021%) were surveyed across two or more
years. Further information on the temporal structure of sur-
veys during both time periods is presented in Appendix S1.
As both migrating birds and postbreeding dispersal of juve-
niles could potentially bias inference on breeding ranges for
both time periods, observations were excluded from analysis
if either the individual detected was a juvenile or was clearly
in migration (as determined by behavior, plumage, and expert
Focal species selection
A total of 223 bird species were recorded in at least one sur-
vey, but not all species occurred across all three regions. We
created independent focal species lists for each region, includ-
ing species detected at a minimum of 10% of sites within both
eras. We additionally excluded all nonpasserines, as most
were observed sporadically, except those in five families:
Odontophoridae, Phasianidae, Columbidae, Trochilidae, and
Picidae. Our final region-specific species list tallied 78 species
for Lassen, 78 species for Yosemite, and 73 species for South-
ern Sierra. Combining the three regions resulted in 99 focal
species, of which 53 were common to all region lists.
Modeling of elevational ranges
We used a ‘single-season’ occupancy model probability frame-
work (MacKenzie et al., 2003, 2006) to simultaneously estimate
a probability of detection (p) and a probability of occupancy
(w) for each species. To explore whether occupancy changed
over time, we used an ‘unpaired-site’ model (Tingley & Beis-
singer, 2009), which tests for a temporal (‘era’) effect as a
covariate of wwithin a single-season model. To account for
both heterogeneity in detection and false absences, we tested
six parameterizations of pmodels using two different covari-
ates. The variable era allowed the probability of detection to
differ by time period (historic vs. modern), while Julian day
(jday, with linear and squared effects) was used to test if p
changed over the survey season. Other candidate variables
were considered (including habitat and intraera observer-spe-
cific effects), but preliminary analyses concluded they were
much less important in explaining heterogeneity in detectabil-
ity (Appendix S1), so were not included in candidate models
to reduce model set complexity (Burnham & Anderson, 2002).
Continuous covariates were standardized to a mean of zero
and standard deviation of one. All combinations of era, jday,
and jday
were used, in addition to a null (intercept-only)
model. The detection model employed is described as:
defining a probability of detection (p) for species i, at site j, for
survey k.
Following Moritz et al. (2008), single-season occupancy
parameterizations sought to examine how occupancy changed
over time (era), over elevation (elevation and elevation
), and, in
this case, among the three regions (defined by two dummy
variables, R1 and R2). As sites consisted of survey routes that
covered a range of elevations, we assigned a single elevation
value to each site defined by the mean elevation of point count
stations comprising each site. We tested 25 different wparam-
eterizations, which included all combinations of these covari-
ates along with all two- and three-way interaction terms and a
null (intercept-only) model (full model set listed in Table S2).
Our final model set combined all 6 pparameterizations with
all 25 wparameterizations, resulting in 150 model combina-
tions of pand wthat were run for each species. Unconditional,
model-averaged values of p
and w
were calculated using
AIC weights (w
) of each model, resulting in one composite
model for each species (Burnham & Anderson, 2002; Moritz
et al., 2008). All occupancy models were run in R version 2.13
(R Development Core Team, 2011) based on code modified
from Royle & Dorazio (2008).
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3282 M. W. TINGLEY et al.
Measurement of elevational ranges and estimation of
We used a combination of restrictions to conservatively esti-
mate significant range change at elevational limits of distribu-
tions. First, we used a P
(probability of false absence) test on
both lower- and upper-elevation range boundaries within
each region to eliminate apparent shifts that could be due to
imperfect detection of individuals at range margins. P
lates the probability that a species was present but not
detected at a set of sites at a range limit (Moritz et al., 2008;
Tingley & Beissinger, 2009), and can be expressed as:
Pfa ¼Y
where p
is the probability of detection at the ith survey of site
j, and p
jis the probability of detecting the species over nsur-
veys at site j. The probability of false absence is then calcu-
lated across msites with nondetections to estimate the chance
the species was present at all of those sites and escaped detec-
tion. Range limit shifts with a P
0.05 were statistically sig-
nificant. In addition, we considered statistically significant
results of P
tests to be ecologically meaningful if the magni-
tude of the shift was greater than 10% of the species’ historical
elevational range within a region (e.g., if a historical range
was from 500 to 1500 m, then range change at either limit
would need to be greater than 100 m) following Moritz et al.
(2008). Plots of historic and modern occupancy and the results
of P
tests for each species analyzed appear in Appendix S2.
Simulations of our ability to detect actual range shifts given
our sampling structure suggest that our methods are highly
conservative, with a Type I error rate of approximately 1%
(see Appendix S1). This conservatism is also robust to poten-
tial violations of the closure assumption within resurveys at
sites a possibility that has been suggested to bias occupancy
results (Rota et al., 2009). While our restrictive conditions for
assigning elevational range change likely underestimate the
true number of range shifts (simulations presented in Appen-
dix S1), our conservative methods are warranted given our
inferential goals as well as the difficulty of working with his-
torical data (Tingley & Beissinger, 2009).
Climatic nearest neighbor predictions
Trends in average annual climate (Fig. 1b) were ascertained
from the Parameter-elevation Regressions on Independent
Slopes Model (PRISM) (Daly et al., 2002). We used surfaces
with 1 91 km, or 30 arc-second, spatial resolution (received 26
January 2009 from C. Daly) to quantify both historical (1910
1930) and modern (19892009) average annual temperature
and precipitation at each of our survey locations. Average
annual temperature at our sites was highly correlated in both
time periods with other temperature variables (max annual
temperature: ρ=0.97; min annual temperature: ρ=0.96), and
average annual precipitation was similarly highly correlated
with other precipitation variables (precipitation of wettest
month: ρ=0.98; precipitation of driest month: ρ=0.73).
Estimation of nearest neighbor elevations followed estab-
lished methods (Ackerly et al., 2010) by measuring the Euclid-
ian distance between the historic climate (temperature or
precipitation) at a single site and the modern climate at a
regional set of sites (Fig. S1). The regional set comprised all
191 km PRISM grid cells within a geographic area defined
by a minimum convex polygon plus 20 km buffer surround-
ing all survey localities in a region. Nearest climatic neighbors
were identified by measuring the climatic distance between
modern climate cells and the historic climate at a survey site
(Fig. S1a). The climatically nearest 5% of cells were pooled
and their elevations were averaged, to account for local heter-
ogeneity in climate and elevation (Fig. S1b). Survey sites were
only compared to climates of cells within the same region.
Comparing the elevation of each site and the average elevation
of its 5% modern-climatic nearest neighbors allowed the crea-
tion of site-specific predictions of elevation change for each
survey site (Figs. 1c, S1b) and for every grid cell within
each region (Fig. 1d). Upper- and lower-limit predictions for
each species were based on the predictions for the actual sites
where a species had its historical upper and lower limit.
Mixed-model hypothesis testing
Generalized Linear Mixed Models (GLMM) were used for two
analyses of patterns of range change. First, we sought to
understand the environmental factors related to upslope vs.
downslope movement patterns. A binary response variable
was used to indicate whether a range limit had moved ups-
lope (value =1) or downslope (value =0) for species that had
significant range shifts. All range limits, regions, and species
were pooled together. Explanatory variables used included:
(1) range limit (categorical: upper or lower); (2) survey region
(categorical: Lassen, Yosemite, Southern Sierra); (3) tempera-
ture-based predictions of upslope or downslope movement
based on nearest neighbor analysis (categorical: upslope or
downslope); (4) precipitation-based predictions of movement
based on nearest neighbor analysis (categorical: upslope or
downslope); and (5) regional population trend for species in
the second half of the 20th century (continuous: percent
change in population per year, signs reversed for lower limit
data). Population trend values were derived from Sierra
Nevada-specific estimates of the North American Breeding
Bird Survey for 19662008 (Sauer et al., 2008), a continent-wide
annual survey of breeding bird populations.
Second, GLMMs were used to examine how well individual
species’ traits explained whether or not species moved,
regardless of direction. Whereas movement direction may be
related to climate change, the ability and motivation to move
may be a species-specific trait. Life history data were compiled
for all species based on accounts from The Birds of North Amer-
ica Online (Poole, 2005). We tested: (1) migratory status (three
levels: permanent resident; short-distance migrant; and long-
distance migrant); (2) mean mass of breeding adult (average
of male and female masses of California subspecies, when
available); (3) territory type (i.e., whether an all-purpose
‘Type-A’ (Nice, 1941) territory is defended); (4) mean home
range size of breeding individuals; (5) mean clutch size of
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
females; and (6) diet breadth (i.e., omnivore vs. specialized
diet). Data on all trait measures were available for 86 of the 99
focal species (Table S3). These six life history traits were
included in a GLMM analysis of the probability of an eleva-
tional range shift combined with (7) elevation zone, as deter-
mined by the historical classification of life zones in Grinnell’s
monographs (Grinnell & Storer, 1924; Grinnell et al., 1930) fol-
lowing Moritz et al. (2008). In a two-part process, each species
trait was first added as a factor on its own and then in combi-
nation with elevation zone. Subsequent models tested multi-
ple traits by adding traits individually in the order ranked by
the initial AIC scores. Traits were added in this forward step-
wise process until additional traits no longer improved AIC
scores. All GLMM models were fit using a logit link and
included species identity as a random effect. All models were
run in R (R Development Core Team, 2011) using the package
‘lme4’ (Bates & Maechler, 2011).
Significant changes in one or both range limits were
found for 84% of bird species across all regions of the
Sierra Nevada, but the direction of change was hetero-
geneous (46% and 53% of shifts were upslope for lower
and upper limits, respectively), as were responses
across the three regions (Fig. 2). Upward shifts ranged
from 161 to 1320 m for lower limits and 218 to 2503 m
for upper limits, whereas downward shifts ranged from
113 to 1557 m for lower limits and 127 to 1567 m for
upper limits. The largest range shift in any direction
was observed in the Savannah Sparrow (upper limit
shifted up 2503 m in Southern Sierra). Species that
shifted an elevational range limit upslope >1 km in any
region included American Goldfinch, Downy Wood-
pecker, Pine Siskin, Black Phoebe, Bushtit, Mourning
Dove, Purple Finch, Red-winged Blackbird, White-
breasted Nuthatch, Chipping Sparrow, Lark Sparrow,
Northern Mockingbird, Song Sparrow, and Western
Meadowlark. Species that shifted an elevational range
limit downslope >1 km in any region included Downy
Woodpecker, Black-chinned Hummingbird, Bewick’s
Wren, House Finch, American Robin, and Violet-green
Swallow. Appendix S2 shows results for all species and
The naı
¨ve expectation of ‘moving up’ was supported
for bird species in the Yosemite region, which
expanded their upper limits upslope (one-sided bino-
Species proportion
Lower limit
Species proportion
Southern Sierra
Species proportion
Upper limit
Elevation (km)
5 101520253035404550556065707580859095
Range expansion
Range contraction
Elevation (km)
5 101520253035404550556065707580859095
Elevation (km)
5 101520253035404550556065707580859095
Species ordered by elevation
Southern Sierra
Fig. 2 Upslope and downslope range shifts by birds over 80100 years. (a) Significant elevation change in species range by limit for
three regions in the Sierra Nevada. Species shifted upslope more than downslope for two region-limit comparisons (*P<0.1,
**P<0.05). (b) Graphical summary of all range shifts for each species (n=99), showing the historical range (gray bar), plus expansions
(or colonizations; green) or contractions (or extinctions; red) over the ensuing century. Species numbers (x-axis) correspond to the spe-
cies list in Table S1.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3284 M. W. TINGLEY et al.
mial test, n=35, p=0.02), and was marginally
supported for species in the Lassen region that con-
tracted their lower range limits upslope (one-sided
binomial test, n=12, p=0.07). Across all regions,
however, neither upper nor lower range limits signifi-
cantly shifted upslope more frequently than downslope
(one-sided binomial tests, upper: n=103, p=0.28;
lower: n=61, p=0.78). Of the 53 species analyzed in
all three regions, only 5 species significantly shifted in
the same direction throughout the Sierra Nevada for a
single range metric: Red-breasted Sapsucker, Fox Spar-
row, Lazuli Bunting, and Spotted Towhee shifted
upslope in all three regions while Ash-throated Fly-
catcher shifted downslope in all three regions.
The direction of shifts in elevational range limits
between 1911 and 2009 were best explained by account-
ing for movement expectations based on both tempera-
ture and precipitation (Table 1). The best-fitting GLMM
model of directional shifts included both temperature-
and precipitation-based nearest neighbor expectations.
Summing AIC weights across models (Burnham &
Anderson, 2002), the weight of evidence was nearly
twice as strong for precipitation-based expectations
(cumulative w
=0.92) as temperature-based expecta-
tions (cumulative w
=0.56). However, models with
either temperature alone (cumulative w
=0.04) or pre-
cipitation alone (cumulative w
=0.39) were not as
strongly supported as models that included both
(cumulative w
=0.52), suggesting that localized range
shifts may result from only one climatic factor, but that
both temperature and precipitation influence shifts
over a larger geographic range.
Range limit was an important factor affecting direc-
tion of change (cumulative w
=0.84), with upper limits
significantly more likely to shift upslope than lower
limits. Region received little model support (cumulative
=0.22) after controlling for site-specific climatic
trends using nearest neighbor expectations. Moreover,
range shifts seemed unaffected by late 20th century
population trends of species, which had less support
across all models than climate (cumulative w
Thus, elevational range expansions or contractions
were primarily related to climate-based expectations
and were not confounded by regional population
changes (Thomas & Lennon, 1999).
Although avian ranges did not shift upslope consis-
tently, they did shift in accord with climate-driven
responses. Eighty-two percent of all range shifts were
in a direction expected from either temperature- or pre-
cipitation-based nearest neighbors (Fig. 3). Over half of
the species that shifted were subject to opposing direc-
tional forces by temperature and precipitation. For
these species, movement in one direction potentially
represents species-specific sensitivity. Temperature-
expected shifts were predominantly upslope (two-sided
binomial test, n=91, p=0.002), whereas precipitation-
expected shifts tended to be downslope (two-sided
binomial test, n=94, p=0.079). Thus, while the
direction of climate-induced range shifts may appear
‘idiosyncratic,’ shift directions were consistent with
dual climatic niche factors that are thought to fre-
quently limit range boundaries (Grinnell, 1917; Tingley
et al., 2009; Wiens et al., 2010).
Over half of the species in each region were not
found to have significantly shifted their elevational
range, despite regional climatic expectations to do so
(Table S1). Several species traits were strong predictors
of range shifts, but not in the direction of a priori expec-
tations. Clutch size and territory type were included in
top models for both upper and lower range limit move-
ments (Table 2). Species with small clutches were more
likely to shift range than species with large clutches
(Fig. 4a), and species holding all-purpose (type A) terri-
tories had 3.3 times greater odds of shifting their upper
range limits than less territorial species (i.e., non-type
Table 1 Generalized linear mixed models examining the pat-
tern of upslope and downslope movements of 99 bird species
in three regions of the Sierra Nevada, California, over the last
Model name
limit +temp +
5106.6 223.2 0.0 0.23
limit +precip 4 107.8 223.6 0.4 0.19
limit +temp +
precip +pop. trend
6106.4 224.7 1.6 0.11
limit +precip +
pop. trend
5107.5 225.0 1.8 0.09
limit +region +
temp +precip
7105.6 225.1 2.0 0.09
limit +region +
temp +precip +
pop. trend
8105.3 226.7 3.5 0.04
precip 3 110.5 226.9 3.7 0.04
*Explanatory variables hypothesized to affect whether a spe-
cies shifted upslope or downslope included the range limit
being tested (‘limit’; i.e., upper range limit or lower range
limit), the region of the Sierra Nevada (‘region’; i.e., Lassen,
Yosemite, or Southern Sierra), the expected upslope or
downslope shift based on the temperature-based (‘temp’) or
precipitation-based (‘precip’) nearest neighbor analysis, and
species-specific population trend derived from regional sur-
vey data 19662008 (‘pop.trend’). Explanatory factors were
modeled as fixed effects, and all models included species as a
random effect.
Only models with DAIC values less than 4 (indicating
strongly supported models) are shown (see Table S4 for all
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
A) and 2 times greater odds of shifting their lower
range limits (Fig. 4b). There were also differences in the
probability of lower-limit range shifts among species
depending on their migratory status: short-distance
migrants and year-round residents were more likely to
shift lower limits than long-distance migrants (Fig. 4c).
Finally, upper limit shifts tended to occur more often
for dietary specialists, although intramodel support for
this trait was marginal (Wald test: P=0.13).
We present strong evidence for 20th century elevational
shifts in breeding distributions for birds in montane
regions of western North America. When viewed in
combination with contemporaneous studies of central
Sierra Nevadan mammals (Moritz et al., 2008), butter-
flies (Forister et al., 2010), and vegetation (Crimmins
et al., 2011), a clear pattern emerges of how recent cli-
mate change has drastically altered the elevational dis-
tributions of montane species. Our results highlight,
however, that elevational change is not unidirectional;
rather, there is substantial variation in the direction and
magnitude of elevational shifts both among species and
within species among regions. While there is a detect-
able signal of species shifting up, consistent with a cen-
tury of average warming temperatures, our results
caution that climate change impacts on species’ ranges,
including likely future shifts, are context dependent,
with species- and site-specific differences.
Meeting predictions: temperature vs. precipitation
Our results demonstrate that site-specific expectations
of the direction of elevational shift, based on both tem-
perature and precipitation changes at a site, were sub-
stantially more successful at predicting observed shifts
than the uniform hypothesis that all species should
shift upslope. Only 51% of significant range shifts
pooled across regions and species were upslope. How-
ever, 82% of significant range shifts were in accordance
with expectations from each species’ nearest climatic
neighbors based on both temperature and precipitation
changes (Fig. 3). Although the northern (Lassen) region
barely warmed on average over the last century, show-
ing localized areas of marginal warming and cooling
Percent of shifts
Yes No Ye s No Yes No
Upslope Downslope
Fig. 3 Agreement between observed elevation shifts and cli-
matic expectations of range shifts derived from modern-temper-
ature and modern-precipitation nearest neighbor elevations.
Observed upslope (shaded) and downslope (solid) shifts can be
divided into groups that were in agreement with either temper-
ature or precipitation (red bars and blue bars, respectively), or
in agreement with either climatic parameter (green bars). Data
are aggregated for upper and lower range limits and for all
three regions. Not represented are the 302 occasions where a
species did not substantially shift a range limit within a region.
Clutch size
Probability of range shift
12 5 10
Territory type
Probability of range shift
Atype Other
(b) Upper
Probability of range shift
Fig. 4 Relationships between species’ traits and the probability
of a range shift (regardless of direction). Upper and lower limits
were more likely to shift for (a) species with small clutches, and
for (b) strongly territorial (A-type) species. Lower limits were
significantly more likely to shift for (c) year-round resident spe-
cies of California, particularly short-distance migrants. Error
bars represent 95% confidence intervals around estimated
parameter means.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3286 M. W. TINGLEY et al.
(Fig. 1b), the proportion of bird species shifting there
was comparable to the other two regions that
experienced substantial warming (Fig. 2). Thus, the
northern Sierra Nevada illustrates the failure of the uni-
form expectation that warming alone will predict range
The biggest improvement to models of elevational
range shift came from incorporating precipitation
changes into directional expectations. Precipitation
change led to expectations of downward movements
that were opposed to most temperature-based expecta-
tions of upward shifts. While precipitation explained a
greater proportion of shifts, directional shifts on the
whole were best explained by both temperature and
precipitation together (Table 1, Fig. 3). Thus, the oppos-
ing push of rising temperature driving species upslope
and the pull of increased precipitation driving species
downslope aptly describes the majority of 20th century
avian elevational shifts in the Sierra Nevada mountains.
Variability in species’ responses to climate
Precipitation and temperature are likely to vary in their
magnitude of influence among species and across sites.
Generally, ranges of low-elevation species may be more
limited by biotic factors (e.g., species interactions),
whereas ranges of high-elevation species may be more
limited by abiotic factors (Brown et al., 1996). However,
using different inferential methods, Tingley et al. (2009)
found that low-elevation birds in the Sierra Nevada
were more likely to shift their occurrence in climate
space toward more favorable precipitation conditions,
whereas high-elevation species were more likely to
shift toward favorable temperature conditions. Consis-
tent with this pattern, using our geographically
expanded and more species-rich data set we found that
lower limits had directional shifts best described by
precipitation, whereas upper limits had directional
shifts best described by temperature.
Our results also highlight the importance of local cli-
matic contexts in creating variability in measured eleva-
tional shifts. Of the 53 species that we tested for range
shifts in all three regions, 11 species shifted range
boundaries in opposing directions across two regions.
This variation may be explained by species responding
to site-specific climate change in each region. For
instance, in the Southern Sierra, which is the warmest
and the driest region (Fig. 1b), precipitation explained
a greater proportion of range limit movements than
temperature (Table S1). For five of the 11 species (Fox
Sparrow, Hermit Thrush, Bewick’s Wren, Bushtit, and
House Finch), diverging directional response is explic-
itly predicted by precipitation, and it is predicted for
two additional species (Warbling Vireo and White-
breasted Nuthatch) by temperature (Table S1). Grinnell
(1917), posited that different factors limit a species’ dis-
tribution across its geographic range. Modern theory
concurs, suggesting that different biotic and abiotic fac-
tors can switch from nonlimiting to range limiting quite
rapidly, with only small changes in the balance of the
factors (Gaston, 2009). Given differing climatic regimes
in our three regions and contrasting climate histories
over the last century, our results support these theoreti-
cal expectations.
To shift or not to shift
Of critical importance to conservation are reasons why
species do not shift their ranges given climate change
(Dawson et al., 2011). In the Sierra Nevada, 10 of 53
(19%) species analyzed across all three regions did not
shift by any metric in any region. Although there are
numerous theoretical reasons why species may not shift
in response to climatic change, we found that certain
species traits were associated with range shifts (Fig. 4).
However, these relationships were generally opposed
to aprioripredictions from dispersal-related hypotheses.
We found species were more likely to shift their range
Table 2 Generalized linear mixed models testing species traits in relation to whether species shifted elevational range limits in
any direction
Range limit
Model name
Lower migratory status +territoriality +clutch size 5 100.6 213.1 0.0 0.42
migratory status +territoriality 4 101.9 213.7 0.6 0.31
migratory status 3 103.4 214.8 1.6 0.19
migratory status +elevational zone 4 103.3 216.6 3.5 0.07
Upper territoriality +clutch size +diet +elevational zone 4 114.8 241.6 0.0 0.50
territoriality +clutch Size +elevational zone 3 115.9 241.8 0.2 0.46
*Species traits were tested separately for each range limit.
Models were initially built testing species traits individually. Composite models testing multiple traits were ad hoc tested subse-
quently with only those traits that lowered AIC score. For explanations of life history covariates, see Methods.
Only models with DAIC values less than 4 (indicating strongly supported models) are shown (see Table S5 for all models).
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
if they had smaller clutches, defended all-purpose terri-
tories for feeding and breeding, and were year-round
residents of California (i.e., short-distant or nonmigra-
tory species).
An alternative hypothesis is that the likelihood of a
range shift relates to behavioral plasticity over the life
span of an individual. For instance, neotropical
migrants have shorter life spans than resident species
(Martin, 1995), and clutch size is typically inversely
related with longevity (Sæther, 1988; Martin, 1995). Cli-
mate change may impact bird ranges through indirect
pathways by changing food availability and nest suc-
cess (Sanz et al., 2003; Both et al., 2006), both of which
can impact breeding site fidelity (Greenwood &
Harvey, 1982; Hoover, 2003). Long-lived birds have
more opportunities to incorporate past knowledge of
breeding success into the selection of future nest sites.
Moreover, if phenological shifts in food availability are
a key proximate cause of differential population
declines across a range (Both et al., 2006, 2011), species
defending feeding territories might experience greater
pressure to search for and defend climatically favorable
(and thus food-resource favorable) territories than
species that travel great distances in search of food.
Our species’ trait analysis (Table 2; Fig. 4) suggests that
it is not the physical ability to disperse that prevents
bird species from shifting their elevational range.
Rather, there may be a complex interplay between the
differential effects of climate change and phenological
shifts on nest-site selection among birds of different life
history patterns.
Origins of a heterogeneous response
Our results suggest that a heterogeneous mixture of ele-
vational range shifts for birds of the Sierra Nevada over
the past century (Fig. 2) resulted from the combined
effects of: (1) temperature pushing species upslope and
precipitation pulling them downslope; (2) variation
among species in their relative sensitivity to tempera-
ture and precipitation; (3) spatial variation in recent cli-
mate change; and (4) differing propensities to shift
depending on species traits. Our results also demon-
strate that site-specific expectations of the direction of
elevational shift, based on the climatic history of a site,
will be substantially more successful at predicting
observed shifts than the naı
¨ve expectation of upward
movements, and could explain why studies with lim-
ited geographical sampling have not always shown
such trends (e.g., Archaux, 2004). Our results also high-
light the importance of accounting for precipitation in
climate-change impact studies, as has recently been
demonstrated by Bonebrake & Mastrandrea (2010) for
butterflies and Crimmins et al. (2011) for trees. Despite
examples of precipitation having direct impacts on the
population growth and survival rates of birds (Sanz,
2002; Schaub et al., 2005; Robinson et al., 2007; Seamans
& Gutie
´rrez, 2007) and other terrestrial vertebrates
(King et al., 1991; Frick et al., 2010; Warner et al., 2010),
studies of range shifts remain largely within a tempera-
ture-centric paradigm.
The failure of most empirical studies of climate-
change impacts to include precipitation and other cli-
matic dimensions highlights critical research needs in
global change biology. The future of predictive mod-
eling of climate-change impacts may lie in coupling
environmental niche-based models of species
distributions with mechanistic relationships between
environmental suitability and fitness (Kearney & Por-
ter, 2009; Brook et al., 2010). This requires a better
understanding of the direct effects of climate change
on physiology (Kearney & Porter, 2004), the indirect
effects of climate change on habitat change (Crim-
mins et al., 2011), and the interactions of precipitation
and temperature change, for example, through net
primary productivity (Tingley et al., 2009). Although
the past century has seen increased precipitation
throughout much of California and North America,
predictions for future precipitation remain highly
uncertain (IPCC, 2007). Determining how climatic
forces push and pull species in opposition or in
agreement requires a more nuanced view of climate
change impacts, and holds the key to predicting
which species are subject to increasing threats and
where species turnover will be greatest.
This research was funded by the National Science Foundation
(DEB 0640859), a George Melendez Wright Climate Change Fel-
lowship from the National Park Service to MT, the Museum of
Vertebrate Zoology and Department of Environmental Science,
Policy and Management at UC Berkeley, and the Sierra Nevada
Inventory and Monitoring program. GIS work on ArcGIS
supported by a donation by the ESRI Conservation Program.
Comments from J. Brashares, R. Bowie, two anonymous review-
ers, and members of the Grinnell Resurvey Project at UC Berke-
ley improved this manuscript. We are indebted to all those who
helped collect field data, including resurvey leaders A. J. Shultz
and T. J. Feo, and data collectors A. Greene, L. Huey, N. Najar,
P. Newsam, F. Ratcliff, K. Rowe, P. Title, D. Wait, and S. Wein-
stein, as well as those who provided access to private lands,
including R. Tollefson and A. Sheehey.
Ackerly DD, Loarie SR, Cornwell WK, Weiss SB, Hamilton H, Branciforte R, Kraft
NJB (2010) The geography of climate change: implications for conservation bioge-
ography. Diversity and Distributions,16, 476487.
Angert AL, Crozier LG, Rissler LJ, Gilman SE, Tewksbury JJ, Chunco AJ (2011) Spe-
cies’ traits weakly predict recent range shifts. Ecology Letters,14, 677689.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3288 M. W. TINGLEY et al.
Archaux F (2004) Breeding upwards when climate is becoming warmer: no bird
response in the French Alps. Ibis,146, 138144.
Bates D, Maechler M (2011) lme4: linear mixed-effects models using S4 classes. R
package version 0.999375-36. Available at:
age=lme4 (accessed 1 November 2011).
Bibby CJ, Burgess ND, Hill DA, Mustoe SH (2000) Bird Census Techniques.Academic
Press, London.
Bonebrake TC, Mastrandrea MD (2010) Tolerance adaptation and precipita-
tion changes complicate latitudinal patterns of climate change impacts. Proceed-
ings of the National Academy of Sciences of the United States of America,107,
Bonfils C, Santer BD, Pierce DW et al. (2008) Detection and attribution of temperature
changes in the mountainous Western United States. Journal of Climate,21, 6404
Both C, Bouwhuis S, Lessells CM, Visser ME (2006) Climate change and population
declines in a long-distance migratory bird. Nature,441,8183.
Both C, Van Turnhout CAM, Bijlsma RG, Siepel H, van Strien A, Foppen RPB (2011)
Avian population consequences of climate change are more severe for long-dis-
tance migrants in seasonal habitats. Proceedings of the Royal Society of London Series
B-Biological Sciences,277, 12591266.
Brook BW, Akc¸akaya HR, Keith DA, Mace GM, Pearson RG, Arau
´jo MB (2010) Inte-
grating bioclimate with population models to improve forecasts of species extinc-
tions under climate change. Biology Letters,5, 723725.
Brown JH, Stevens GC, Kaufman DM (1996) The geographic range: size, shape,
boundaries, and internal structure. Annual Review of Ecology and Systematics,
27, 597623.
Buckley LB, Urban MC, Angilletta MJ, Crozier LG, Rissler LJ, Sears MW (2010)
Can mechanism inform species’ distribution models? Ecology Letters,13, 1041
Burnham KP, Anderson DR (2002) Model Selection and Multimodel Inference. Springer-
Verlag, New York.
Chen I-C, Hill JK, Ohlemu
¨ller R, Roy DB, Thomas CD (2011) Rapid range shifts
of species associated with high levels of climate warming. Science,333, 1024
Colwell RK, Brehm G, Cardelu
´s CL, Gilman AC, Longino JT (2008) Global warming,
elevational range shifts, and lowland biotic attrition in the wet tropics. Science,322,
Crimmins SM, Dobrowski SZ, Greenberg JA, Abatzoglou JT, Mynsberge AR (2011)
Changes in climatic water balance drive downhill shifts in plant species’ optimum
elevations. Science,331, 324327.
Daly C, Gibson WP, Taylor GH, Johnson GL, Pasteris P (2002) A knowledge-based
approach to the statistical mapping of climate. Climate Research,22,99113.
Davis AJ, Jenkinson LS, Lawton JH, Shorrocks B, Wood S (1998) Making mistakes
when predicting shifts in species range in response to global warming. Nature,
391, 783786.
Dawson TP, Jackson ST, House JI, Colin PI, Mace GM (2011) Beyond predictions: bio-
diversity conservation in a changing climate. Science,332,5358.
Forister ML, McCall AC, Sanders NJ et al. (2010) Compounded effects of climate
change and habitat alteration shift patterns of butterfly diversity. Proceedings of the
National Academy of Sciences of the United States of America,107, 20882092.
Frick WF, Reynolds DS, Kunz TH (2010) Influence of climate and reproductive timing
on demography of little brown myotis Myotislucifugus.Journal of Animal Ecology,
79, 128136.
Gaston KJ (2009) Geographic range limits: achieving synthesis. Proceedings of the Royal
Society of London Series B-Biological Sciences,276, 13951406.
Greenwood PJ, Harvey PH (1982) The natal and breeding dispersal of birds. Annual
Review of Ecology and Systematics,13,121.
Grinnell J (1917) The niche-relationship of the California thrasher. Auk,34, 427433.
Grinnell J, Storer TI (1924) Animal Life in the Yosemite: An account of the mammals, birds,
reptiles, and amphibians in a cross-section of the Sierra Nevada. University of California
Press, Berkeley.
Grinnell J, Dixon J, Linsdale JM (1930) Vertebrate natural history of a section of northern
California through the Lassen Peak region. University of California Press, Berkeley.
Hoover JP (2003) Decision rules for site fidelity in a migratory bird, the Prothonotary
Warbler. Ecology,84, 416430.
Hughes L (2000) Biological consequences of global warming: is the signal already
apparent? Trends in Ecology & Evolution,15,5661.
IPCC (2007) Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II
and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. (eds Pachauri RK, Reisinger A), pp. 2675. IPCC, Geneva.
Kearney M, Porter WP (2004) Mapping the fundamental niche: physiology, climate,
and the distribution of a nocturnal lizard. Ecology,85, 31193131.
Kearney M, Porter WP (2009) Mechanistic niche modelling: combining physiological
and spatial data to predict species’ ranges. Ecology Letters,12, 334350.
´ry M (2004) Extinction rate estimates for plan populations in revisitation studies:
importance of detectability. Conservation Biology,18, 570574.
´ry M, Spillmann JH, Truong C, Holderegger R (2006) How biased are esti-
mates of extinction probability in revisitation studies? Journal of Ecology,94,
King WJ, Festa-Bianchet M, Hatfield SE (1991) Determinants of reproductive success
in female Columbian ground squirrels. Oecologia,86, 528534.
La Sorte FA, Jetz W (2010) Projected range contractions of montane biodiversity
under global warming. Proceedings of the Royal Society B-Biological Sciences,277,
Lenoir J, Gegout JC, Guisan A et al. (2010) Going against the flow: potential mecha-
nisms for unexpected downslope range shifts in a warming climate. Ecography,33,
Link WA, Nichols JD (1994) On the importance of sampling variance to investigations
of temporal variation in animal population size. Oikos,69, 539544.
MacKenzie DI, Nichols JD, Hines JE, Knutson MG, Franklin AB (2003) Estimating site
occupancy, colonization, and local extinction when a species is detected imper-
fectly. Ecology,84, 22002207.
MacKenzie DI, Nichols JD, Royle JA, Pollock KH, Bailey LL, Hines JE (2006) Occu-
pancy Estimation and Modeling, Academic Press, Burlington.
Martin TE (1995) Avian life history evolution in relation to nest sites, nest predation,
and food. Ecological Monographs,65, 101127.
Moritz C, Patton JL, Conroy CJ, Parra JL, White GC, Beissinger SR (2008) Impact of a
century of climate change on small-mammal communities in Yosemite National
Park, USA. Science,322, 261264.
Nice MM (1941) The role of territory in bird life. American Midland Naturalist,26,
Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts
across natural systems. Nature,421,3742.
Parmesan C, Gaines S, Gonzalez L, Kaufman DM, Kingsolver J, Townsend PA, Saga-
rin R (2005) Empirical perspectives on species borders: from traditional biogeogra-
phy to global change. Oikos,108,5875.
Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate change and distribution shifts
in marine fishes. Science,308, 19121915.
Poole A (ed) (2005) The Birds of North America Online. Cornell Laboratory of Ornithol-
ogy, Ithaca.
¨yry J, Luoto M, Heikkinen RK, Kuussaari M, Saarinen K (2009) Species traits
explain recent range shifts of Finnish butterflies. Global Change Biology,15,
R Development Core Team (2011) R: A Language and Environment for Statistical
Computing. Available at: (accessed 11 January 2012).
Ralph CJ, Droege S, Sauer JR (1995) Managing and monitoring birds using point
counts: standards and applications. In: Monitoring bird populations by point counts
(eds Ralph CJ, Droege S, Sauer JR), pp. 161169. General Technical Report PSW-
GTR-149, USDA Forest Service, Albany.
Robinson RA, Baillie SR, Crick WQP (2007) Weather-dependent survival: implica-
tions of climate change for passerine population processes. Ibis,149, 357364.
´guez-Trelles F, Rodrı
´guez MA (1998) Rapid micro-evolution and loss of chro-
mosomal diversity in Drosophila in response to climate warming. Evolutionary
Ecology,12, 829838.
Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Finger-
prints of global warming on wild animals and plants. Nature,421,5760.
Rota CT, Fletcher RJJ, Dorazio RM, Betts MG (2009) Occupancy estimation and the
closure assumption. Journal of Applied Ecology,46, 11731181.
Roy K, Valentine JW, Jablonski D, Kidwell SM (1996) Scales of climatic variability and
time averaging in Pleistocene biotas: implications for ecology and evolution.
Trends in Ecology & Evolution,11, 458463.
Royle JA, Dorazio RM (2008) Hierarchical Modeling and Inference in Ecology. Academic
Press, San Diego.
Sæther B-E (1988) Pattern of covariation between life-history traits of European birds.
Nature,331, 616617.
Sanz JJ (2002) Climate change and breeding parameters of great and blue tits through-
out the western Palearctic. Global Change Biology,8, 409422.
Sanz JJ, Potti J, Moreno J, Merino S, Frias O (2003) Climate change and fitness compo-
nents of a migratory bird breeding in the Mediterranean region. Global Change Biol-
ogy,9, 461472.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
Sauer JR, Hines JE, Fallon JE (2008) The North American Breeding Bird Survey, Results
and Analysis 19662007. Version 5.15.2008. USGS Patuxent Wildlife Research Cen-
ter, Laurel.
Schaub M, Kania W, Ko
¨ppen U (2005) Variation of primary production during winter
induces synchrony in survival rates in migratory white storks Ciconiaciconia.Jour-
nal of Animal Ecology,74, 656666.
Seamans ME, Gutie
´rrez RJ (2007) Sourcesof variability in spotted owl population growth
rate:testing predictions usinglong-termmark-recapturedata. Oecologia,152,5770.
Sinervo B, Me
´ndez-de-la-Cruz F, Miles DB et al. (2010) Erosion of lizard diversity by
climate change and altered thermal niches. Science,328, 894899.
Stralberg D, Jongsomjit D, Howell CA, Snyder MA, Alexander JD, Wiens JA, Root TL
(2009) Re-shuffling of species with climate disruption: a no-analog future for Cali-
fornia birds? PLoS One,4, e6825.
Svenning JC, Normand S, Skov F (2008) Postglacial dispersal limitation of widespread
forest plant species in nemoral Europe. Ecography,31, 316326.
ThomasCD, Lennon JJ (1999) Birdsextend their ranges northwards. Nature,399,213213.
Thomas CD, Cameron A, Green RE et al. (2004) Extinction risk from climate change.
Nature,427, 145148.
Tingley MW, Beissinger SR (2009) Detecting range shifts from historical species
occurrences: new perspectives on old data. Trends in Ecology & Evolution,24, 625
Tingley MW, Monahan WB, Beissinger SR, Moritz C (2009) Birds track their Grinnel-
lian niche through a century of climate change. Proceedings of the National Academy
of Sciences of the United States of America,106, 1963719643.
Warner DA, Jorgensen CF, Janzen FJ (2010) Maternal and abiotic effects on egg mor-
tality and hatchling size of turtles: temporal variation in selection over seven
years. Functional Ecology,24, 857866.
Wiens JJ, Ackerly DD, Allen AP et al. (2010) Niche conservatism as an emerg-
ing principle in ecology and conservation biology. Ecology Letters,13, 1310
Williams JW (2007) Novel climates, no-analog communities, and ecological surprises.
Frontiers in Ecology and Environment,5, 475482.
Zimmermann NE, Yoccoz NG, Edwards TCJ et al. (2009) Climatic extremes improve
predictions of spatial patterns of tree species. Proceedings of the National Academy of
Sciences of the United States of America,106, 1972319728.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
Figure S1. Schematic of how climatic nearest neighbors lead to predictions of elevational shifts.
Appendix S1. Supplementary methods.
Appendix S2. Plots showing detections and nondetections by elevation and survey period for 99 bird species in each of three Sierra
Nevadan regions.
Table S1. Observed historical range, significant change in elevational range over time, and predicted change in range based on tem-
perature and precipitation nearest neighbors.
Table S2. Descriptions and parameterizations of 25 occupancy (w) models with main, additive and multiplicative effects.
Table S3. Life history and other species characteristics used in traits analysis.
Table S4. Full results of GLMM analysis in explaining the pattern of upslope and downslope movements of 99 bird species in three
regions of the Sierra Nevada, California, over the last century.
Table S5. Full results of GLMM analyses for tests of species traits that relate to whether species shifted elevational ranges in any
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the
authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
©2012 Blackwell Publishing Ltd, Global Change Biology,18, 3279–3290
3290 M. W. TINGLEY et al.
... The standard approach to forecast range shifts is to fit an ENM in the present, transfer it to the future, and compare predictions to estimate loss and gain in the areas suitable for a species (Elith et al., 2010). Although the comparison between current and forecasted distributions can predict potential range shifts, incorporating climatic data from the last several decades into ENMs allows researchers to understand if a species may be already reconfiguring their geographic range (Iknayan and Beissinger, 2018;Rapacciuolo et al., 2014;Rowe et al., 2015;Smith et al., 2019;Tingley et al., 2012;VanDerWal et al., 2013). ...
... Despite this, the practice does not provide insights into whether species are already responding to climate change. Additionally, climate does not affect animal and plant distributions uniformly across their entire geographic ranges (Beever et al., 2016;Rowe et al., 2015;Smith et al., 2019;Tingley et al., 2012). For instance, changes in environmentally marginal areas, which usually correspond to the geographical distributional limits of the species (Pironon et al., 2017), could lead to the emergence of newly suitable areas or the disappearance of areas (that become unsuitable). ...
... As environmental niches differ between species, the varying geographical responses to climatic changes depend mainly on the loss and gain of combinations of environmentally suitable conditions particular to a species (Wiens et al., 2009). Ongoing climate change has already impacted the distributional limits of species worldwide (Huang et al., 2017;Iknayan and Beissinger, 2018;Rapacciuolo et al., 2014;Rowe et al., 2015;Tingley et al., 2012;VanDerWal et al., 2013;Zu et al., 2021), and the suitability trends detected for C. mexicanus over the last several decades suggest that this species is probably not an exception. The patterns observed along the distributional limits using a time series of model predictions indicate that the suitability for this species does not follow a general range shift pattern (i. ...
... Mixed trends have also been found on the effect of movement habits on range edge shifts. Hockey et al. (2011) and Laube et al. (2013) found that latitudinal shifters are more likely migrants and nomad birds, whilst Tingley et al. (2012) found similar results for resident birds. Lastly, Brommer (2008) found negligible impact of migratory strategy on Finnish birds shifts compared to other traits such as body size, diet composition and habitat specificity. ...
... The evidence of recent range shift as a response to climate change and other disturbances is undeniable (Chen et al., 2011;Parmesan & Yohe, 2003) and the attempts to discriminate among range expanders and contractors are numerous (e.g. Angert et al., 2011;Auer & King, 2014;Bradshaw et al., 2014;La Sorte & Thompson III, 2007;Sunday et al., 2015;Tingley et al., 2012;Yang et al., 2020). Our work differs from previous efforts to tease out the differences between these groups of species in that (1) it was set in a bounded environment, where species are physically constrained in their shifts and (2) we accounted for geographical constraints, environmental preferences and species traits (including trophic characteristics) within the same analytical framework. ...
Full-text available
Aim: Despite the strong evidence of species range shifts as a response to environmental change, attempts to identify species traits that modulate those shifts have been equivocal. We investigate the role of species traits and environmental preferences on birds' range shifts in Great Britain, an island where dispersal is limited by the English Channel and the North Sea. Location: Great Britain (England, Scotland and Wales). Taxa: Birds (Aves). Time Period: 1968-2011. Methods: Using 404,949 occurrence records from two time periods, we investigated the potential drivers of leading and rear range edge shifts of breeding birds using phy-logenetic linear mixed models. We hypothesized that shifts are influenced by species' trophic and morphological traits, dispersal abilities and environmental preferences, but also by the geographical boundaries of Great Britain. Results: Geographical boundaries-the distance from the northern or southern boundaries of Britain-accounted for most of the variability in range edge shifts. Species traits and environmental preferences emerged as relevant drivers of range shifts only for northern and Passeriform species. Northern habitat specialist, those with more predators and those sensitive to precipitation were more likely to shift their rear edge poleward. For Passeriformes, habitat generalists, species with smaller dispersal capabilities, under higher predatory pressure or associated with forest and grassland were more likely to shift their rear edge poleward. Main Conclusions: While geographical boundaries impose constraints on range shifts in British birds, the subtle effects of species traits and environmental preferences emerge as relevant predictors for Northern and passeriform species' rear edge shifts. This highlights the importance of accounting for geographical boundaries when predicting species responses to global change. Differential range shifts of species across different trophic levels could result in the reorganization of biotic interactions, with consequences for ecosystem structure and stability.
... One way species may respond to climate change is by shifting their elevational range to track the environmental conditions (e.g., temperature). Previous research has documented species shifting upslope in response to climate change and has linked these shifts to the extirpation of species occurring at the highest elevations as they run out of opportunities to shift upslope (Marris 2007;Sekercioglu et al. 2008;Tingley et al. 2012;Freeman et al. 2018). In contrast, species at the lowest elevations may gain habitat as areas at higher elevations become suitable and areas at the lowest elevations remain suitable (Freeman et al. 2018). ...
... In contrast, species at the lowest elevations may gain habitat as areas at higher elevations become suitable and areas at the lowest elevations remain suitable (Freeman et al. 2018). As such, the net habitat loss resulting from the difference between shifts at the upper extent of a given species' elevational range and that at the lower elevational range is critical to understanding species long-term conservation trajectory (Tingley et al. 2012;Mamantov et al. 2021). ...
Full-text available
Understanding how species will respond to a rapidly changing global climate is requisite to conserving biodiversity. Though habitat losses from human development and land use change remain the most critical threats to biodiversity globally, some regions, such as low-lying islands, are particularly vulnerable to the effects of climate change. Despite this vulnerability, there may be opportunities for imperiled species on islands to adapt to the effects of climate-induced sea level rise. To understand how the response to rising seas may influence the amount of future habitat, we investigated shifts in the elevational range of the endangered silver rice rat (Oryzomys palustris natator; hereafter “rice rat”), a species endemic to tidal environments of the Lower Florida Keys, USA. We quantified fine-scale habitat use using radio telemetry of collared animals, first in 2004, and again in 2021, thus spanning a 17-year period during which the local sea level rose by 0.142 m. We observed a shift in the elevational range limits of rice rats which closely mirrored the rise in sea level, and that this apparent ability to adapt to rising sea level decreased the extent of habitat loss in subsequent decades. However, over longer time scales (~ 100 yrs), the extent of habitat loss from sea level rise outpaced rice rats’ ability to adapt. As such, the conservation of biodiversity on low-lying islands hinges on the ability of the global community to decrease anthropogenic greenhouse gas emissions and mitigate the associated consequences for the global climate. Otherwise, conservation practitioners will be increasingly forced to make difficult decisions about how to conserve imperiled species on low-lying islands.
... Although altitudinal gradients are more geographically restricted, greater temperature shifts may occur across small changes in elevation compared to much larger latitudinal distances (La Sorte et al. 2014;Verheyen et al. 2019). Therefore, species may demonstrate heterogeneous, nonlinear variation in their morphology due to variable clinal differences in abiotic factors across their latitudinal and/or altitudinal ranges (Tingley et al. 2012;Pinsky et al. 2013;Verheyen et al. 2019). Changes in morphology may also be related to changes in vegetation and dietary shifts due to climate warming over time rather than a direct response to climate change (Caumul and Polly 2005;Millien et al. 2017;gigliotti et al. 2020). ...
Mammals are predicted to vary in body size following Bergmann’s rule, with individuals found at higher latitudes in colder temperatures being larger in size compared to conspecifics occurring at lower latitudes in warmer temperatures. Body size is similarly expected to vary temporally, with a decrease in size through time due to recent climate warming. While Bergmann’s rule is well-supported in mammals, there is increasing evidence of exceptions to the rule. Here, we present patterns of size variation in 17 North American mammal species using five morphological traits (condylobasal skull length, skull width, maxillary toothrow length, body weight, and head-and-body length) to determine if size varies predictably for each species in space and time. We found little support for a widespread Bergmannian pattern for these species at a broad spatial scale (across North America) and a contemporary temporal scale (the past 120 years). The effects of latitude or year on each trait were highly variable with three types of responses: an increase, a decrease, or no change in size across space or through time. Spatial size trends were detected more often than temporal size trends, as the temperature range was significantly larger in space than through time. Body weight (the most variable trait) and head-and-body length were more likely to conform to Bergmann’s rule than craniodental measurements. We did not detect any changes in size variability with latitude, and our study species either increased or decreased in size variability over time. Our findings demonstrate that size variation in mammals is highly context-dependent. As such, caution is needed when using rules of body size variation to predict the future response of species to climate warning while valid in theory, it is likely too simplistic of an approach.
... Identifying factors that influence geographic range limits of animal species is a longstanding goal in ecology and evolutionary biology. Elevational shifts in species' ranges are of special interest in the context of global climate change (Moritz et al. 2008;La Sorte and Jetz 2010;Rowe et al. 2010;Tingley et al. 2012;McCain et al. 2021;Storz and Scott 2023), and the conservation implications of such shifts underscore the need for accurate information about contemporary range limits. However, in the most mountainous regions of the planet -where potential elevational range limits are the highest -the upper limits of species' ranges are often poorly demarcated due to a paucity of survey data. ...
Full-text available
In the world's highest mountain ranges, uncertainty about the upper elevational range limits of alpine animals represents a critical knowledge gap regarding the environmental limits of life and presents a problem for detecting range shifts in response to climate change. Here we report results of mountaineering mammal surveys in the Central Andes, which led to the discovery of multiple species of mice living at extreme elevations that far surpass previously assumed range limits for mammals. We live-trapped small mammals from ecologically diverse sites spanning >6700 m of vertical relief, from the desert coast of northern Chile to the summits of the highest volcanoes in the Andes. We used molecular sequence data and whole-genome sequence data to confirm the identities of species that represent new elevational records and to test hypotheses regarding species limits. These discoveries contribute to a new appreciation of the environmental limits of vertebrate life.
Correlative species distribution models are widely used to quantify past shifts in ranges or communities, and to predict future outcomes under ongoing global change. Practitioners confront a wide range of potentially plausible models for ecological dynamics, but most specific applications only consider a narrow set. Here, we clarify that certain model structures can embed restrictive assumptions about key sources of forecast uncertainty into an analysis. To evaluate forecast uncertainties and our ability to explain community change, we fit and compared 39 candidate multi‐ or joint species occupancy models to avian incidence data collected at 320 sites across California during the early 20th century and resurveyed a century later. We found massive (>20,000 LOOIC) differences in within‐time information criterion across models. Poorer fitting models omitting multivariate random effects predicted less variation in species richness changes and smaller contemporary communities, with considerable variation in predicted spatial patterns in richness changes across models. The top models suggested avian environmental associations changed across time, contemporary avian occupancy was influenced by previous site‐specific occupancy states, and that both latent site variables and species associations with these variables also varied over time. Collectively, our results recapitulate that simplified model assumptions not only impact predictive fit but may mask important sources of forecast uncertainty and mischaracterize the current state of system understanding when seeking to describe or project community responses to global change. We recommend that researchers seeking to make long‐term forecasts prioritize characterizing forecast uncertainty over seeking to present a single best guess. To do so reliably, we urge practitioners to employ models capable of characterizing the key sources of forecast uncertainty, where predictors, parameters and random effects may vary over time or further interact with previous occurrence states.
Full-text available
The Karner blue butterfly ( Lycaeides melissa samuelis , or Kbb), a federally endangered species under the U.S. Endangered Species Act in decline due to habitat loss, can be further threatened by climate change. Evaluating how climate shapes the population trend of the Kbb can help in the development of adaptive management plans. Current demographic models for the Kbb incorporate in either a density-dependent or density-independent manner. We instead created mixed density-dependent and -independent (hereafter “endo-exogenous”) models for Kbbs based on long-term count data of five isolated populations in the upper Midwest, United States during two flight periods (May to June and July to August) to understand how the growth rates were related to previous population densities and abiotic environmental conditions, including various macro- and micro-climatic variables. Our endo-exogenous extinction risk models showed that both density-dependent and -independent components were vital drivers of the historical population trends. However, climate change impacts were not always detrimental to Kbbs. Despite the decrease of population growth rate with higher overwinter temperatures and spring precipitations in the first generation, the growth rate increased with higher summer temperatures and precipitations in the second generation. We concluded that finer spatiotemporally scaled models could be more rewarding in guiding the decision-making process of Kbb restoration under climate change.
A key question in biology concerns the extent to which distributional range limits of species are determined by intrinsic limits of physiological tolerance. Here, we use common‐garden data for wild rodents to assess whether species with higher elevational range limits typically have higher thermogenic capacities in comparison to closely related lowland species. Among South American leaf‐eared mice (genus Phyllotis ), mean thermogenic performance is higher in species with higher elevational range limits, but there is little among‐species variation in the magnitude of plasticity in this trait. In the North American rodent genus Peromyscus , highland deer mice ( Peromyscus maniculatus ) have greater thermogenic maximal oxygen uptake () than lowland white‐footed mice ( Peromyscus leucopus ) at a level of hypoxia that matches the upper elevational range limit of the former species. In highland deer mice, the enhanced thermogenic in hypoxia is attributable to a combination of evolved and plastic changes in physiological pathways that govern the transport and utilization of O 2 and metabolic substrates. Experiments with Peromyscus mice also demonstrate that exposure to hypoxia during different stages of development elicits plastic changes in cardiorespiratory traits that improve thermogenic via distinct physiological mechanisms. Evolved differences in thermogenic capacity provide clues about why some species are able to persist in higher‐elevation habitats that lie slightly beyond the tolerable limits of other species. Such differences in environmental tolerance also suggest why some species might be more vulnerable to climate change than others. image
High mountain habitats are globally important for biodiversity. At least 12% of birds worldwide breed at or above the treeline, many of which are endemic species or species of conservation concern. However, due to the challenges of studying mountain birds in difficult-to-access habitats, little is known about their status and trends. This book provides the first global review of the ecology, evolution, life history and conservation of high mountain birds, including comprehensive coverage of their key habitats across global mountain regions, assessments of diversity patterns along elevation gradients, and adaptations for life in the alpine zone. The main threats to mountain bird populations are also identified, including climate change, human land use and recreational activities. Written for ecologists and naturalists, this book identifies key knowledge gaps and clearly establishes the research priorities needed to increase our understanding of the ecology of mountain birds and to aid in their conservation.
Full-text available
A guide to data collection, modeling and inference strategies for biological survey data using Bayesian and classical statistical methods. This book describes a general and flexible framework for modeling and inference in ecological systems based on hierarchical models, with a strict focus on the use of probability models and parametric inference. Hierarchical models represent a paradigm shift in the application of statistics to ecological inference problems because they combine explicit models of ecological system structure or dynamics with models of how ecological systems are observed. The principles of hierarchical modeling are developed and applied to problems in population, metapopulation, community, and metacommunity systems. The book provides the first synthetic treatment of many recent methodological advances in ecological modeling and unifies disparate methods and procedures. The authors apply principles of hierarchical modeling to ecological problems, including * occurrence or occupancy models for estimating species distribution * abundance models based on many sampling protocols, including distance sampling * capture-recapture models with individual effects * spatial capture-recapture models based on camera trapping and related methods * population and metapopulation dynamic models * models of biodiversity, community structure and dynamics * Wide variety of examples involving many taxa (birds, amphibians, mammals, insects, plants) * Development of classical, likelihood-based procedures for inference, as well as Bayesian methods of analysis * Detailed explanations describing the implementation of hierarchical models using freely available software such as R and WinBUGS * Computing support in technical appendices in an online companion web site.
Full-text available
Species distribution models (SDMs) use spatial environmental data to make inferences on speciesÕ range limits and habitat suitability. Conceptually, these models aim to determine and map components of a speciesÕ ecological niche through space and time, and they have become important tools in pure and applied ecology and evolutionary biology. Most approaches are correlative in that they statistically link spatial data to species distribution records. An alternative strategy is to explicitly incorporate the mechanistic links between the functional traits of organisms and their environments into SDMs. Here, we review how the principles of biophysical ecology can be used to link spatial data to the physiological responses and constraints of organisms. This provides a mechanistic view of the fundamental niche which can then be mapped to the landscape to infer range constraints. We show how physiologically based SDMs can be developed for different organisms in different environmental contexts. Mechanistic SDMs have different strengths and weaknesses to correlative approaches, and there are many exciting and unexplored prospects for integrating the two approaches. As physiological knowledge becomes better integrated into SDMs, we will make more robust predictions of range shifts in novel or non-equilibrium contexts such as invasions, translocations, climate change and evolutionary shifts.
This chapter gives results from some illustrative exploration of the performance of information-theoretic criteria for model selection and methods to quantify precision when there is model selection uncertainty. The methods given in Chapter 4 are illustrated and additional insights are provided based on simulation and real data. Section 5.2 utilizes a chain binomial survival model for some Monte Carlo evaluation of unconditional sampling variance estimation, confidence intervals, and model averaging. For this simulation the generating process is known and can be of relatively high dimension. The generating model and the models used for data analysis in this chain binomial simulation are easy to understand and have no nuisance parameters. We give some comparisons of AIC versus BIC selection and use achieved confidence interval coverage as an integrating metric to judge the success of various approaches to inference.