Integrative and Comparative Biology
Integrative and Comparative Biology, volume 56, number 1, pp. 110–119
doi:10.1093/icb/icw005
Society for Integrative and Comparative Biology
SYMPOSIUM
Beyond the Mean: Biological Impacts of Cryptic Temperature
Change
Kimberly S. Sheldon1,* and Michael E. Dillon*
*Department of Zoology and Physiology and Program in Ecology, University of Wyoming, Laramie, WY 82071, USA
From the symposium ‘‘Beyond the Mean: Biological Impacts of Changing Patterns of Temperature Variation’’ presented
at the annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2016 at Portland, Oregon.
1
E-mail: [email protected]
Synopsis Studies have typically used shifts in mean temperatures to make predictions about the biotic impacts of climate
change. Though shifts in mean temperatures correlate with changes in phenology and distributions, other hidden, or
cryptic, changes in temperature, such as temperature variation and extreme temperatures, could pose greater risks to
species and ecological communities. Yet, these cryptic temperature changes have received relatively little attention because
mean temperatures are readily available and the organism-appropriate temperature response is often elusive. An alternative to using mean temperatures is to view organisms as physiological filters of hourly temperature data. We explored
three classes of physiological filters: (1) nonlinear thermal responses using performance curves of insect fitness, (2)
cumulative thermal effects using degree-day models for corn emergence, and (3) threshold temperature effects using
critical thermal maxima and minima for diverse ectotherms. For all three physiological filters, we determined the change
in biological impacts of hourly temperature data from a standard reference period (1961–90) to a current period
(2005–10). We then examined how well mean temperature changes during the same time period predicted the biotic
impacts we determined from hourly temperature data. In all cases, mean temperature alone provided poor predictions of
the impacts of climate change. These results suggest that incorporating high frequency temperature data can provide
better predictions for how species will respond to temperature change.
Introduction
Global temperatures have increased by 0.7ºC in the
last century and are projected to increase by 2–6º C
by 2100 (IPCC 2014). These recent increases in mean
temperatures have had measurable impacts on populations, species, and communities (Parmesan and
Yohe 2003; Sinervo et al. 2010; Chen et al. 2011;
Tewksbury et al. 2011; Goodman et al. 2012;
Gibson-Reinemer et al. 2015) leading to dire predictions for the biological impacts of projected increases
in mean temperatures (Deutsch et al. 2008; IPCC
2014; Urban 2015). However, mean temperatures
have not increased everywhere—some locations
have experienced little change or decreases in mean
temperatures over time—and some organisms have
shown little to no response to changing mean temperatures (Parmesan and Yohe 2003; Thackeray et al.
2010; Chen et al. 2011). In either case, however,
mean temperatures are unlikely to capture the full
story. Even in locations where mean temperatures
have not changed over time, other aspects of the
temperature signal that are biologically relevant,
such as seasonal variation, repeated exposures, or
extreme temperatures (Marchand et al. 2005;
Sheldon and Tewksbury 2014; Marshall et al. 2015;
Marshall and Sinclair 2015; Sheldon et al. 2015) have
been altered due to climate change (Meehl and
Tebaldi 2004; Rahmstorf and Coumou 2011; Wang
and Dillon 2014; Dillon et al. this issue). These other
temperature changes are cryptic, or hidden, and,
until recently (Clusella-Trullas et al. 2011;
Paaijmans et al. 2013; Vasseur et al. 2014), have
been largely ignored by biologists studying the impacts of climate change (Supplementary Fig. S1). Our
analysis of the top ecology journals (following methods in HilleRisLambers et al. 2013) reveals that in the
Advanced Access publication April 13, 2016
ß The Author 2016. Published by Oxford University Press on behalf of the Society for Integrative and Comparative Biology. All rights reserved.
For permissions please email: [email protected].
Impacts of cryptic temperature change
most recent 12-year period, the number of publications assessing biological impacts of temperature variation increased eight-fold, but there were still twice
as many publications focused on mean temperatures
(Supplementary Fig. S1). This difference in the use of
mean versus variation in temperature in part reflects
the lack of high-resolution temporal data available
from climatologists, who usually provide only aggregated versions of the raw data they collect over time
for use by other researcher (e.g., WorldClim,
Climatic Research Unit) (Hijmans et al. 2005).
Are changes in mean temperatures over time good
enough for predicting impacts of climate change? In
some cases, biological impacts appear to be tightly
correlated with changes in mean temperatures (Dunn
and Winkler 1999; Walther et al. 2002; Parmesan and
Yohe 2003; Gordo and Sanz 2005; Chen et al. 2011)
such that incorporating temperature variation would
add little predictive power. In most cases, however,
we don’t know if mean temperatures are good
enough because we have not yet explored how predictions are altered when we incorporate highresolution data. The small but growing body of
literature beginning to ask this question (e.g., Wu
et al. 2014) suggests that mean temperatures are
not sufficient for accurately predicting impacts of
climate change because incorporating temperature
variation often better explains (Paaijmans et al.
2010) or more strongly predicts (Niehaus et al.
2012; Paaijmans et al. 2013; Vasseur et al. 2014)
changes in the biological trait of interest. This
emerging viewpoint likely reflects the nonlinear relationships between temperature and many biological
traits (Bozinovic et al. 2011; Dell et al. 2011;
Amarasekare and Savage 2012). Thermal performance curves describing these relationships allow
one to transform temperatures into performance
traits, providing a powerful approach for building
mechanistic models of climate change impacts
(Kearney and Porter 2009; Kingsolver et al. 2011;
Dangles et al. 2013) and revealing the importance
of temperature variation (Estay et al. 2011; Niehaus
et al. 2012; Lawson et al. 2015).
We hypothesized that aggregate metrics of temperature may fail to detect or may underestimate biological impacts of changing climates because the
underlying temperature variation matters (Fig. 1).
To test this hypothesis, we first identified three
types of organismal responses to temperature that
could show different responses to changes in temperature variation: (1) fitness curves, (2) thermal tolerance thresholds, and (3) degree-day accumulation.
111
Fitness curves, like other thermal performance
curves (TPCs), are nonlinear and generally built
from measurements at a series of constant temperatures (Huey and Stevenson 1979; Scheiner 2002).
TPCs allow us to estimate the effects of temperature
variation under the assumptions that (1) temperatures experienced over a short time period will lead
to the same performance outcome as performance
measured after chronic exposure to the same temperatures, and (2) that thermal history does not influence the performance curve (Niehaus et al. 2012;
Kingsolver et al. 2015). Though these assumptions
are likely false (Estay et al. 2014), we still largely
lack both empirical data testing them (but see
Niehaus et al. 2012 and references therein) and reasonable approaches to discarding them (Kingsolver
and Woods 2016; see contributions in this issue).
With these shortcomings in mind, this approach is
nevertheless useful for uncovering the biological importance of changes in temperature variation (Fig. 1;
see e.g., Estay et al. 2011; Vasseur et al. 2014; Lawson
et al. 2015), which will require both an understanding of the patterns of thermal variation in the environment and the mechanisms by which animals cope
with such variation (Bozinovic et al. 2011).
Thermal thresholds are ubiquitous in organismal
physiology—many higher level physiological processes show step changes at particular temperatures
rather than continuously varying with temperature
(Heinrich 1974; Hazell and Bale 2011; Macmillan
and Sinclair 2011). Critical thermal limits are one
such example of thermal thresholds for which we
have excellent data for diverse organisms across latitude (Sunday et al. 2014). Very small changes in
temperature variation could push organisms from
just within to just beyond thermal limits, with potentially huge biological consequences (Fig. 1; Huey
et al. 2009; see Williams et al., this issue).
Finally, many phenological responses, such as
spring emergence from overwintering and leaf flushing, depend on thermal history. Specifically, these
critical life history events are tightly correlated with
the accumulation of degree-days above a sensitivity
threshold (Taylor 1981; Logan and Powell 2001;
Schwartz 2013). Changes in temperature variation,
particularly at key times of the year, could alter the
timing of degree-day accumulation, shifting phenological events independent of the effects of mean temperatures (Fig. 1; Crozier 2006; Cleland et al. 2007).
Using empirical descriptions of these thermal responses, we transformed global temperature data at
different temporal resolutions (hourly, daily, and
112
K. S. Sheldon and M. E. Dillon
Fig. 1. Temperature variation can affect nonlinear performance metrics, threshold responses, and responses to thermal history. A single
year of hourly temperature data from Portland, OR, USA is plotted on the bottom left (8396 temperature measurements, gray), with
mean monthly temperatures (12 measurements) and mean annual temperature overlaid (black line). Note that mean monthly temperatures fail to capture many temperature excursions beyond the upper and lower thermal limits (CTmax and CTmin, respectively).
Temperatures can be filtered through a nonlinear performance curve (bottom right) to estimate integrated performance over time (top
right), yielding different integrated performance distributions for the different sampling frequencies. The top left figure shows accumulated growing degree-days over the course of the year, estimated based on hourly and mean monthly temperatures. At the beginning
of the year, monthly temperatures underestimate growing degree-days, but by the middle of the year, monthly degree-day estimates
occur 27 days earlier than hourly estimates.
monthly) to outcomes for organisms in terms of fitness, crossing of critical thermal thresholds, and accumulation of degree-days. By comparing the effects
of different temporal resolutions of temperature data
on these outcomes, we show that incorporating temperature variation can strongly alter assessments of
climate change impacts across these three broad categories of thermal responses.
Climate data
From a total of 26,639 ‘‘isd-lite’’ weather stations
(Lott et al. 2001), we followed Dillon et al. (2010)
and kept only those weather stations (2457 total)
that sampled at least every 6 h on average (with no
seasonal gaps) throughout a standard reference
period (1961–90) and a more current 5-year period
(2005–10) with complete and quality-checked data
(see Supplementary Fig S2). Our goal was to understand whether incorporating temperature variation
alters estimates of climate change impacts.
Therefore, for all analyses, we estimated anomalies
of biological traits for the current period (2005–10)
and the standard reference period (1961–90) for each
individual weather station. We then averaged these
anomalies within 5 5º latitude by longitude grid
113
Impacts of cryptic temperature change
cells (n ¼ 613) to account for uneven spatial distribution of weather stations (Supplementary Fig. S2).
Averaging all stations within grid cells reduces spatial
resolution, but is critical for subsequent analyses
(Dillon et al. 2010). In addition, our goal was to
show how temporal, not spatial variation in temperature, affects predictions of the impacts of climate
change (for analysis of spatial variation in temperature, see Pincebourde and Suppo, this issue). Thus,
this approach allowed us to explore questions related
to temporal variation in temperature while still accounting for uneven distribution of weather stations.
Nonlinear thermal responses to
temperature variation
Using supporting data from Deutsch et al. (2008), we
modeled the temperature dependence of population
growth rate, r, for 38 insect species from different
latitudes based on three parameters: critical thermal
maximum (CTmax) and minimum (CTmin) and optimal temperature (Topt) (Supplementary Fig. S3).
We modified the curves to allow relative population
growth rates to drop below zero at high temperatures
following Kingsolver et al. (2013). We determined
the best-fit relationship between each parameter
and latitude (Supplementary Fig. S4), which allowed
us to generate a theoretical insect fitness curve for
every weather station based on the station’s location.
Using a similar approach to other researchers
(Deutsch et al. 2008; Vasseur et al. 2014), we then
estimated the change in fitness caused by changes in
temperature at each weather station by subtracting
fitness integrated over the standard reference period
from fitness integrated over the current period. We
estimated these changes in fitness based both on
mean monthly and on hourly temperature data for
each individual station over the entire year for tropical stations (latitudes 523.58) and over the warmest
6 months of the year for temperate stations (latitude
423.58) given that most species at higher latitudes
are not active year-round (Ragland and Kingsolver
2008; Sheldon and Tewksbury 2014). We then examined the average change in r for all experimental
weather stations in a 5 5º latitude by longitude
grid of the globe (n ¼ 613).
Fitness increased by as much as 10% and decreased
by as much as 42% from the standard reference
period to the current period when estimated based
on mean monthly temperatures (Fig. 2a). When we
used hourly temperature data, fitness changes were
less pronounced, ranging from 22% decreases to
11% increases (note the muted colors of Fig. 2b compared to Fig. 2a). The differences in fitness estimates
between the two temporal resolutions (i.e., mean
monthly minus hourly estimates for each station)
were striking. Relative to hourly temperature data,
mean monthly temperature data over- and underestimated fitness impacts by as much as 11% and 22%,
respectively (Fig. 2c, d). Thus, aggregate data (mean
monthly temperatures) gave quite different estimates
of fitness impacts of climate change from those based
on hourly temperature data. We are not suggesting
that fitness values actually change on an hourly basis,
but simply that ignoring temperature variation when
filtering through nonlinear performance curves
can profoundly alter integrated estimates of
performance.
Threshold responses to temperature variation
Some biological responses may be most sensitive to
extreme temperatures. Climate metrics that aggregate
data into mean monthly temperatures may miss
these extreme events and subsequent biological consequences. We used critical thermal limits (CTmax
and CTmin) to examine how aggregation of temperature data affects our ability to detect when temperatures exceed species’ threshold limits.
We used supporting data from Sunday et al. 2014
in which critical thermal maxima (CTmax) and
minima (CTmin) were compiled for a variety of ectothermic taxa (n ¼ 197). We used model fits for
CTmax and CTmin as a function of latitude as listed in
Sunday et al. (2014; Supplementary Fig. S5) to assign
critical thermal limits to each weather station based
on its latitude. We ignored the cold limit factor for
CTmin because we assumed that species are in diapause or hibernating when temperatures are below
CTmin.
For each weather station, we calculated monthly
mean maxima and minima for the entire standard
reference period and current period. We then calculated the change from the standard reference period
to the current period in the proportion of months
where the monthly mean maxima and minima temperatures were above or below CTmax or CTmin, respectively (see Appendix 1 in Supplementary
Materials for details). For comparison, we then did
the same calculation using daily temperature data to
once again calculate the change in the proportion of
months where temperatures were above or below
CTmax or CTmin, respectively. We then averaged
these changes in threshold temperature crossings
for all weather stations within a 5 5º latitude by
longitude grid across the world (n ¼ 613).
Based on mean monthly maximum temperatures,
the vast majority of grid cells showed no change
114
K. S. Sheldon and M. E. Dillon
Fig. 2. Change in fitness from the standard reference period (1961–90) to the current period (2005–2010). Results show change in
fitness using mean monthly temperatures (A) and hourly temperatures (B). Squares are average change in fitness from all weather
stations in a 5 58 latitude by longitude grid (n ¼ 613). Colors show where fitness has increased (red), decreased (blue), or remained
the same (white). The difference in mean monthly and hourly temperature estimates (C and D) indicate where monthly mean
temperatures overestimated (red), underestimated (blue), or showed no difference (white) in change in fitness relative to hourly
temperature data.
from the standard reference period to the current
period in the proportion of months exceeding
CTmax (Fig. 3a). Of 613 gridded locations, two locations had 15% increases in the proportion of months
exceeding CTmax, seven locations showed slight decreases, but the majority (95%) showed no change
(Fig. 3a). In contrast, for estimates based on daily
maximum temperatures, most locations (62%)
showed a change in the proportion of months exceeding CTmax, with increases of up to 20% and decreases of as much as 7% (Fig. 3b).
Based on mean monthly minimum temperatures,
43% of locations showed a change in the proportion
of months dropping below CTmin with increases and
decreases of as much as 5% and 15%, respectively
(Fig. 4a; Note that the decrease of 15% is not shown
in Fig. 5 because mean temperatures at the site
shifted by 78C and the point is, thus, an outlier).
The greatest changes happened in north temperate
regions (Fig. 4a). For estimates based on daily minimum temperatures, all three regions (78% of all
locations) showed changes in the proportion of
months below CTmin, with increases of up to 18%
and decreases of as much as 27% (Fig. 4b).
Cumulative thermal responses to
temperature variation
Some biological responses, including many phenological events, depend on an accumulation of temperatures (Taylor 1981; Logan and Powell 2001;
Schwartz 2013). To examine how biological responses that depend on thermal history may be differentially altered by changes in temperature
variation, we asked whether mean temperatures predict growing degree-days (GDDs) of plant emergence, a critical life history event. GDDs are a
measure of heat accumulation and can be used to
predict when plants will reach growing milestones
(e.g., emergence, maturity) during the year
(McMaster and Wilhelm 1997). We calculated
GDDs by summing temperatures over time between
lower and upper threshold temperatures, Tbase, and
TUT, respectively. Because corn (Zea mays L.) is
grown worldwide (Paliwal 2000), good data exist
Impacts of cryptic temperature change
on the relationship between degree-days and corn
emergence, and impacts of climate change on crop
production is of global importance (Sakschewski et
al. 2014), we used corn to test whether changes in
mean temperatures predict climate-driven shifts in
corn emergence date. We set Tbase and TUT to
108C and 308C, respectively, which approximate development requirements for corn and other major
food crops (e.g., wheat) (Paliwal 2000).
We used hourly temperature data to calculate the
average calendar date during the standard reference
period and the current period when corn reached the
GDD requirements for emergence (100 GDDs). For
temperate regions, we determined the calendar date
it took to reach emergence starting with 1 January
in the Northern Hemisphere and 1 July in the
Southern Hemisphere. For tropical regions, we took
the average of the days it took to accumulate GDDs
for emergence starting on both 1 January and 1 July.
We averaged the change in the number of days it
took to reach GDD requirements from the standard
reference period to the current period for all experimental weather stations in a 5 58 latitude by
longitude grid of the globe (n ¼ 613).
Corn emerged earlier with increased mean temperature in north temperate regions ( ¼ 1.32, p ¼
0.01). Some north-temperate sites, however, had a
delayed time to emergence even when mean temperatures increased by as much as 28C (Fig. 5). In tropical and south-temperate areas, corn did not emerge
earlier even when mean temperatures increased substantially (p ¼ 0.16 and p ¼ 0.18, respectively).
Discussion
Changes in mean monthly temperatures are overwhelmingly used to make predictions about the biotic impacts of climate change. Mean temperatures do not,
however, capture all aspects of temperature that are
biologically relevant (e.g., Paaijmans et al. 2013;
Camacho et al. 2015; Marshall and Sinclair 2015).
Based on our results, predictions of the impacts of
climate change differed when we incorporated temperature variation. Specifically, mean monthly temperatures changed our estimates of fitness and
proportion of time spent beyond critical thermal
limits relative to hourly and daily temperatures, respectively. In addition, mean monthly temperatures
showed poor predictive power for estimating changes
in phenology (i.e., corn emergence date). The difference in predictions of the impacts of climate change is
due to the cryptic changes that are not captured by
mean temperatures coupled with the nature of physiological filters. The nonlinear, threshold, and
115
Fig. 3. Change in proportion of months spent beyond critical
thermal maxima as a function of shifts in mean temperatures
from 1961 to 2010. Each point represents the change in the
proportions of months with at least one temperature above
critical thermal maximum when we used mean monthly (A) or
daily (B) maximum temperature data from all weather stations in
a 5 58 latitude by longitude grid (n ¼ 613). Points indicate
north temperate (black), tropical (gray), and south temperate
(white) sites.
cumulative relationships between temperature and organismal responses mean that temperature variation
matters and that ignoring it changes predictions of
biotic impacts (Niehaus et al. 2012; Vasseur et al.
2014).
When we estimated change in fitness and time
spent above or below critical thermal limits, we
saw very different results based on mean monthly
temperature compared to results based on higher
frequency temperature data (Figs. 2–4). In general,
predictions based on mean monthly temperatures
tended to show much stronger fitness impacts,
both positive and negative, relative to hourly temperatures (Fig. 2). Mean temperature defines where
on the fitness curve a species resides and, due to the
nonlinear nature of the curve, this can have large
implications for how variation around the mean affects fitness (Vasseur et al. 2014; Lawson et al. 2015).
A species in a cooler environment will have smaller
changes in fitness associated with variation around
the mean than a species living in a warmer environment (Deutsch et al. 2008; Dillon et al. 2010). Thus,
mean temperatures have fitness implications, but
knowledge of variation around the mean is also
116
K. S. Sheldon and M. E. Dillon
Fig. 5. Change in corn emergence date from the standard reference period (1961–90) to the current period (2005–10) as a
function of shifts in mean temperatures from 1961–2010. Each
point represents the average change in date of corn emergence
from all weather stations in a 5 58 latitude by longitude grid
(n ¼ 613). Points indicate north temperate (black), tropical
(gray), and south temperate (white) sites.
Fig. 4. Change in proportion of months spent below critical
thermal minima as a function of shifts in mean temperatures from
1961 to 2010. Each point represents the change in the proportions of months with at least one temperature below critical
thermal maximum when we used mean monthly (A) or daily (B)
minimum temperature data from all weather stations in a 5 58
latitude by longitude grid (n ¼ 613). Points indicate north temperate (black), tropical (gray), and south temperate (white) sites.
key for predicting climate-driven changes in fitness
(Niehaus et al. 2012; Vasseur et al. 2014)
In contrast to fitness, predictions for the time
spent above or below threshold temperatures,
which approximates the impact of extreme events,
were dampened when we used mean monthly
temperatures compared to daily temperature data
(Figs. 3 and 4). Thus, predictions based on aggregated maximum and minimum temperature data
could result in a largely different perception of population- and species-level health and vulnerability
when compared to approaches that incorporate temperature variation.
Many organismal responses to temperature are
based on thermal history (Taylor 1981; Logan and
Powell 2001; Schwartz 2013) and often show correlations with shifts in mean temperatures (Walther et
al. 2002; Parmesan and Yohe 2003). The timing of
these phenological events is critical in many mutualistic interactions, and accurate predictions are essential for anticipating change in species’ interactions
associated with climate change (Cleland et al. 2007;
Araujo and Luoto, 2007). Based on our estimates of
phenological response to hourly temperatures, mean
monthly temperatures explained little of the variation we saw in date of phenological events (Fig. 5).
This suggests that despite observed correlations between mean temperatures and phenology, mean temperatures will not provide an accurate picture of the
timing of phenological events in response to climate
change (Visser et al. 2006; Yang and Rudolf 2010).
In this study, we focused on incorporating temperature variation as one way to improve predictions
of the biotic impacts of climate change. However, we
recognize that in some instances, studies are constrained by available historical data and, thus, mean
temperatures are the only available data with which
to make predictions (e.g., Youngstadt et al. 2014). In
addition, the impact of temperature variation will
vary by the organism of interest such that smaller
ectotherms may be more affected by hourly shifts
in temperature compared to larger ectotherms with
greater thermal inertia (Bell 1980; Woods et al. 2015)
or endotherms, which have fundamentally different
physiology compared to ectotherms (Huey et al.
2012). However, temperature variation should be incorporated whenever possible and appropriate. In
addition, several other factors could foster more
accurate predictions depending on the species of
interest, including greater spatial resolution of temperature data for small species that use microclimates
(Faye et al. 2014; Woods et al. 2015) and knowledge
of population-level variation in species’ traits
(Higgins et al. 2014; Buckley et al. 2015). Weather
station data, such as those used in this study, provide
air temperatures that do not capture the temperatures organisms experience or the spatial variation
in those temperatures (Potter et al. 2013; Faye et
Impacts of cryptic temperature change
al. 2014; Sears and Angilletta 2015). Microclimate
data (Bonebrake et al. 2014) and estimates of operative temperatures (Bakken 1992; Camacho et al.
2015) would provide greater spatial resolution of
the temperatures species experience in their environment and their body temperatures in those environments (Zamora-Camacho et al. 2015). This
information will be critical for understanding and
predicting biological impacts of changing temperatures (Woods et al. 2015; Pincebourde et al., this
issue) []. Finally, traits that may be affected by temperature are not uniform across species’ ranges
(Higgins et al. 2014). Incorporating intraspecific variation among populations of key traits will facilitate
better assessments of climate change impacts
(Valladares et al. 2014).
The impact of temperature change on species and
ecological communities will depend on shifts in
mean temperatures and on changes in temperature
variation around those means (Paaijmans et al. 2010,
2013; Vasseur 2014). Continuing to collect and use
temperature data of high temporal and spatial resolution and expanding our knowledge of organismal
thermal physiology, particularly population-level variation, will allow us to appropriately filter those temperatures, thereby producing more realistic
predictions of the biotic impacts of climate change.
Supplementary data
Supplementary data available at ICB online.
Acknowledgments
We thank Art Woods, Emile Faye, and an anonymous reviewer for helpful comments and Joel
Kingsolver and Ray Huey for fruitful discussions.
Funding
K.S.S. was funded by National Science Foundation
(Postdoctoral Research Fellowship 1306883). M.E.D.
was funded by NSF (IOS 1457659). The symposium
was funded by NSF (IOS 1545787), the Company of
Biologists, and The Society for Integrative and
Comparative Biology.
References
Amarasekare P, Savage V. 2012. A framework for elucidating
the temperature dependence of fitness. Am Nat 179:178–91.
Araujo MB, Luoto M. 2007. The importance of biotic interactions for modeling species distributions under climate
change. Glob Ecol Biogeogr 16:743–53.
Bakken GS. 1992. Measurement and application of operative
and standard operative temperature in ecology. Am Zool
32:194–216.
117
Bell CJ. 1980. The scaling of the thermal inertia of lizards.
J Exp Biol 86:79–85.
Bonebrake TC, Deutsch CA. 2012. Climate heterogeneity
modulates impact of warming on tropical insects. Ecology
93:449–55.
Bozinovic F, Bastı́as DA, Boher F, Clavijo-Baquet S, Estay SA,
Angilletta MJ. Jr 2011. The mean and variance of environmental temperature interact to determine physiological tolerance and fitness. Physiol Biochem Zool 84:543–52.
Buckley LB, Ehrenberger JC, Angilletta MJ. 2015.
Thermoregulatory behaviour limits local adaptation of
thermal niches and confers sensitivity to climate change.
Funct Ecol 29:1038–47.
Camacho A, Rodrigues MT, Navas C. 2015. Extreme
operative temperatures are better descriptors of the thermal
environment than mean temperatures. J Therm Biol
49:106–11.
Chen IC, Hill JK, Ohlemuller R, Roy DB, Thomas CD. 2011.
Rapid range shifts of species associated with high levels of
climate warming. Science 333:1024–26.
Cleland EE, Chuine I, Menzel A, Mooney HA, Schwartz MD.
2007. Shifting plant phenology in response to global
change. Trends Ecol Evol 22:357–65.
Clusella-Trullas S, Blackburn TM, Chown SL. 2011. Climatic
predictors of temperature performance curve parameters in
ecotherms imply complex responses to climate change. Am
Nat 177:738–51.
Crozier L, Dwyer G, Fagan AEWF, DeAngelis EDL. 2006.
Combining population-dynamic and ecophysiological
models to predict climate-induced insect range shifts. Am
Nat 167:853–66.
Dangles O, Herrera M, Mazoyer C, Silvain JF. 2013.
Temperature-dependent shift in herbivore performance
and interactions drive non-linear changes in crop damages.
Glob Change Biol 19:1056–63.
Dell AI, Pawar S, Savage VM. 2011. Systematic variation in
the temperature dependence of physiological and ecological
traits. Proc Natl Acad Sci USA 108:10591–96.
Deutsch C, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor
C, Haak DC, Martin PR. 2008. Impacts of climate warming
on terrestrial ectotherms across latitude. Proc Natl Acad Sci
USA 105:6668–72.
Dillon ME, Wang G, Huey RB. 2010. Global metabolic impacts of recent climate warming. Nature 467:704–06.
Dillon ME, Woods HA, Wang G, Fey SB, Vasseur DA,
Telemeco RS, Marshall KE, Pincebourde S. 2016. Life in
the frequency domain: the biological impacts of changes
in climate variability at multiple time scales. Int Comp Biol.
Dunn PO, Winkler DW. 1999. Climate change has affected
the breeding date of Tree Swallows throughout North
America. Proc R Soc B Biol Sci 266:2487–90.
Estay SA, Clavijo-Baquet S, Lima M, Bozinovic F. 2011.
Beyond average: an experimental test of temperature variability on the population dynamics of Tribolium confusum.
Pop Ecol 53:53–58.
Estay SA, Lima M, Bozinovic F. 2014. The role of temperature
variability on insect performance and population dynamics
in a warming world. Oikos 123:131–40.
Faye E, Herrera M, Bellomo L, Silvain J-F, Dangles O. 2014.
Strong discrepancies between local temperature mapping
118
and interpolated climatic grids in tropical mountainous agricultural landscapes. PLoS ONE 9(8):e105541.
Gibson-Reinemer DK, Sheldon KS, Rahel FJ. 2015. Climate
change creates rapid species turnover in montane communities. Ecol Evol 5:2340–47.
Goodman RE, Lebuhn G, Seavy NE, Gardali T, Bluso-Demers
JD. 2012. Avian body size changes and climate change:
warming or increasing variability? Glob Change Biol
18:63–73.
Gordo O, Sanz JJ. 2005. Phenology and climate change: a
longterm study in a Mediterranean locality. Oecologia
146:484–95.
Hazell SP, Bale JS. 2011. Low temperature thresholds: are chill
coma and CTmin synonymous? J Insect Physiol
57:1085–89.
Heinrich B. 1974. Thermoregulation in endothermic insects.
Science 185:747–56.
Higgins JK, MacLean HJ, Buckley LB, Kingsolver JG. 2014.
Geographic differences and microevolutionary changes in
thermal sensitivity of butterfly larvae in response to climate.
Funct Ecol 28:982–89.
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005.
Very high resolution interpolated climate surfaces for global
land areas. Int J Clim 25:1965–78.
HilleRisLambers J, Ford K, Haak DC, Horwith M, Ettinger
AK, Miner BE, Rogers HS, Sheldon KS, Water S, Yang S,
et al. 2013. Accidental experiments: ecological and evolutionary insights and opportunities derived from anthropogenic change. Oikos 122:1649–61.
Huey RB, Deutsch CA, Tewksbury JJ, Vitt LJ, Hertz PE, Pérez
HJA, Garland T. 2009. Why tropical forest lizards are vulnerable to climate warming. Proc R Soc B Biol Sci
276:1939–48.
Huey RB, Kearney MR, Krockenberger A, Holtum JAM, Jess
M, Williams SE. 2012. Predicting organismal vulnerability
to climate warming: roles of behaviour, physiology and
adaptation. Philos Trans R Soc Lond B Biol Sci
367:1665–79.
Huey RB, Stevenson RD. 1979. Integrating thermal physiology
and ecology of ectotherms: a discussion of approaches. Am
Zool 19:357–66.
IPCC 2014. Climate change 2014: impacts, adaptation, and
vulnerability. Cambridge: Cambridge University Press.
Jiguet F, Julliard R, Thomas CD, Dehorter O, Newson SE,
Couvet D. 2006. Thermal range predicts bird population
resilience to extreme high temperatures. Ecol Lett
9:1321–30.
Kearney M, Porter W. 2009. Mechanistic niche modelling:
combining physiological and spatial data to predict species’
ranges. Ecol Lett 12:334–350.
Kingsolver JG, Diamond SE, Buckley LB. 2013. Heat stress
and the fitness consequences of climate change for terrestrial ectotherms. Funct Ecol 27:1415–23.
Kingsolver JG, Higgins JK, Augustine KE. 2015. Fluctuating
temperatures and ectotherm growth: distinguishing
non-linear and time-dependent effects. J Exp Biol
218:2218–2225.
Kingsolver JG, Woods HA. 2016. Beyond thermal performance curves: modeling time-dependent effects of thermal
stress on ectotherm growth rates. Am Nat 187:283–94.
K. S. Sheldon and M. E. Dillon
Kingsolver JG, Woods HA, Buckley LB, Potter KA, MacLean
HJ, Higgins JK. 2011. Complex life cycles and the responses
of insects to climate change. Integr Comp Biol 51:662–65.
Lawson CR, Vindenes Y, Bailey L, van de Pol M. 2015.
Environmental variation and population responses to
global change. Ecol Lett 18:724–36.
Logan JA, Powell JA. 2001. Ghost forests, global warming, and
the mountain pine beetle (Coleoptera: Scolytidae). Am
Entomol 47:160–173.
Lott N, Baldwin R, Jones P. The FCC integrated surface
hourly database, a new resource of Global Climate Data.
(http://www1.ncdc.noaa.gov/pub/data/techrpts/tr200101/
tr2001-01.pdf (National Climatic Data Center, 2001).
Macmillan HA, Sinclair BJ. 2011. Mechanisms underlying
insect chill-coma. J Insect Physiol 57:12–20.
Marchand FL, Mertens S, Kockelbergh F, Beyens L, Nijs I.
2005. Performance of High Arctic tundra plants improved
during but deteriorated after exposure to a simulated extreme temperature event. Glob Change Biol 11:2078–2089.
Marshall DJ, Rezende EL, Baharuddin N, Choi F, Helmuth B.
2015. Thermal tolerance and climate warming sensitivity in
tropical snails. Ecol Evol 5:5905–19.
Marshall KE, Sinclair BJ. 2015. The relative importance of
number, duration and intensity of cold stress events in
determining survival and energetics of an overwintering
insect. Funct Ecol 29:357–66.
Meehl GA, Tebaldi C. 2004. More intense, more frequent, and
longer lasting heat waves in the 21st century. Science
305:994–97.
Niehaus AC, Angilletta, MJ Jr, Sears MW, Franklin CE,
Wilson RS. 2012. Predicting the physiological performance
of ectotherms in fluctuating thermal environments. J Exp
Biol 215:694–701.
Parmesan C, Yohe G. 2003. A globally coherent fingerprint of
climate change impacts across natural systems. Nature
421:37–42.
Paaijmans KP, Blanford S, Bell AS, Blanford JI, Read AF,
Thomas MB. 2010. Influence of climate on malaria transmission depends on daily temperature variation. Proc Natl
Acad Sci USA 107:15135–39.
Paaijmans KP, Heinig RL, Seliga RA, Blanford JI, Blanford S,
Murdock CC, Thomas MB. 2013. Temperature variation
makes ectotherms more sensitive to climate change. Glob
Change Biol 19:2373–80.
Paliwal RL. 2000. Tropical maize: improvement and production. Rome: Food and Agricultural Organization of the
United Nations.
Pincebourde S, Murdock CC, Vickers M, Sears MW. 2016.
Microclimate variations across fine spatial scales can shape
the response of organisms to global change. Int Comp Biol
Pincebourde S, Suppo C. 2016. The vulnerability of tropical
ectotherms to warming is modulated by the microclimatic
heterogeneity. Int Comp Biol
Potter KA, Woods HA, Pincebourde S. 2013. Microclimatic
challenges in global change biology. Glob Chang Biol
19:2932–39.
Ragland GJ, Kingsolver JG. 2008. Evolution of thermotolerance in seasonal environments: the effects of annual temperature variation and life-history timing in Wyeomyia
smithii. Evolution 62:1345–57.
Impacts of cryptic temperature change
Rahmstorf S, Coumou D. 2011. Increase of extreme events in
a warming world. Proc Natl Acad Sci USA 108:17905–09.
Sakschewski B, von Bloh W, Huber V, Müller C, Bondeau A.
2014. Feeding 10 billion people under climate change: how
large is the production gap of current agricultural systems?
Ecol Model 288:103–11.
Scheiner SM. 2002. Selection experiments and the study of
phenotypic plasticity. J Evol Biol 15:889–98.
Schwartz MD, editor. 2013. Phenology: an integrative environmental science. Dordrecht: Springer Netherlands.
Sears MW, Angilletta MJ. 2015. Costs and benefits of thermoregulation revisited: both the heterogeneity and spatial
structure of temperature drive energetic costs. Am Nat
185:E94–102.
Sheldon KS, Leaché AD, Cruz FB. 2015. The influence of
seasonality in temperature on elevational range size across
latitude: a test using Liolaemus lizards. Glob Ecol Biogeogr
24:632–41.
Sheldon KS, Tewksbury JJ. 2014. The impact of seasonality in
temperature on thermal tolerance and elevational range size
of tropical and temperate beetles. Ecology 95:2134–43.
Sinervo B, Méndez-de-la-Cruz F, Miles DB, Heulin B,
Bastiaans E, Villagrán-Santa Cruz M, Lara-Resendiz R,
Martı́nez-Méndez N, Calderón-Espinosa ML, Meza-Lázaro
RN et al. 2010. Erosion of lizard diversity by climate change
and altered thermal niches. Science 328:894–99.
Sunday JM, Bates AE, Kearney MR, Colwell RK, Dulvy NK,
Longino JT, Huey RB. 2014. Thermal-safety margins and
the necessity of thermoregulatory behavior across latitude
and elevation. Proc Natl Acad Sci USA 111:5610–15.
Taylor F. 1981. Ecology and evolution of physiological time in
insects. Am Nat 117:1–23.
Tewksbury JJ, Sheldon KS, Ettinger AK. 2011. Moving farther
and faster in a warming world. Nat Clim Change 1:396–97.
Thackeray SJ, Sparks TH, Frederiksen M, Burthe S, Bacon PJ,
Bell JR, Botham MS, Brereton TM, Bright PW, Carvalho L,
et al. 2010. Trophic level asynchrony in rates of phenological change for marine, freshwater and terrestrial environments. Glob Change Biol 16:3304–13.
Urban MC. 2015. Accelerating extinction risk from climate
change. Science 348:571–73.
Valladares F, Matesanz S, Guilhaumon F, Araújo MB,
Balaguer L, Benito-Garzón M, Cornwell W, Gianoli E,
119
Kleunen M, Naya DE, et al. 2014. The effects of phenotypic plasticity and local adaptation on forecasts of species range shifts under climate change. Ecol Lett 17:1351–
64.
Vasseur DA, DeLong JP, Gilbert B, Harley CDG, McCann KS,
Savage VM, Tunney TD, O’Connor MI. 2014. Increased
temperature variation poses a greater risk to species than
climate warming. Proc R Soc B Biol Sci 281:20132612.
Visser ME, Holleman LJM, Gienapp P. 2006 Shifts in caterpillar biomass phenology due to climate change and its
impact on the breeding biology of an insectivorous bird.
Oecologia 147:164–72.
Walther GR, Post E, Convey P, Menzel A, Parmesan C,
Beebee TJC, Fromentin JM, Hoegh-Guldberg O, Bairlein
F. 2002. Ecological responses to recent climate change.
Nature 416:389–395.
Wang G, Dillon ME. 2014. Recent geographic convergence in
diurnal and annual temperature cycling flattens global thermal profiles. Nat Clim Chang 4:988–92.
Williams CM, Buckley LB, Sheldon KS, Vickers M, Pörtner HO, Dowd WW, Gunderson AR, Marshall KE, Stillman JH
2016. Biological impacts of thermal extremes: mechanisms
and costs of functional responses matter. Int Comp Biol.
Woods HA, Dillon ME, Pincebourde S. 2015. The roles of
microclimatic diversity and of behavior in mediating the
responses of ectotherms to climate change. J Therm Biol
54:86–97.
Wu TH, Shiao SF, Okuyama T. 2014. Development of insects
under fluctuating temperature: a review and case study. J
Appl Entomol 139:592–99.
Yang LH, Rudolf VHW. 2010. Phenology, ontogeny and the
effects of climate change on the timing of species interactions. Ecol Lett 13:1–10.
Youngsteadt E, Dale AG, Terando AJ, Dunn RR, Frank SD.
2014. Do cities simulate climate change? A comparison of
herbivore response to urban and global warming. Glob
Change Biol. doi: 10.1111/gcb.12692.
Zamora-Camacho FJ, Rubiño-Hispán MV, Reguera S,
Moreno-Rueda G. 2015. Thermal dependence of sprint
performance in the lizard Psammodromus algirus along a
2200-meter elevational gradient: cold-habitat lizards do
not perform better at low temperatures. J Therm Biol
52:90–96.