Introduction to the Symposium: Beyond the Mean: Biological Impacts of Changing Patterns of Temperature Variation

“Variation is the hard reality, not a set of imperfect measures for a central tendency. Means and medians are the abstractions.”1 –Stephen Jay Gould Beetles aside, biologists have an inordinate fondness for means. Means and other measures of central tendency help us to describe complex phenomena in simple ways, and they facilitate quantitative comparisons of traits among treatment groups, populations, and species. For organisms, however, means are abstractions; they confront only variation. In the context of organism–environment interactions, means are used to characterize not only the biological traits of interest but also the environments in which those traits are expressed. But variation in traits among individuals (e.g., Careau et al. 2014 ) and in environments across space ( Sears et al. 2011 ; Woods et al. 2015 ) and time ( Niehaus et al. 2012 ; Wang and Dillon 2014 ) can profoundly alter the outcomes of organism–environment interactions (e.g., Svancara et al. 2002 ; Williams 2008 ; Bonebrake and Deutsch 2011 ; Montalto et al. 2014 ; Boersma et al. 2016 ). Measuring and predicting the biological impacts of climate change has become one of the pressing ecological challenges of our time, and the field has been dominated by central tendencies. Mean values of behavioral, physiological, and ecological traits paired with averages of temperature, precipitation, snowpack, etc. provide inferences about past and predictions about future climate impacts ( Walther et al. 2002 ; Field et al. 2014 ). Increasingly, however, ecologists are recognizing that variation, in both biology and climate, is as important, if not more, in determining the effects of climate change ( Paaijmans et al. 2013 ; Vasseur et al. 2014 ; Lawson et al. 2015 ; Chan et al. 2016 ). In this symposium and associated journal issue, we ask a particular question: What are the biological impacts of changing patterns of temperature variation? We focused on temperature because it is perhaps the most measured environmental variable with the best-known effects on organisms ( Kingsolver 2009 ). In addition to traditional papers from symposium speakers, the issue includes three synthetic papers that address the biological effects of: (1) spatial variation in temperature (Pincebourde et al. 2016, this issue), (2) temporal variation in temperature (Dillon et al. 2016, this issue), and (3) rare but extreme events (Williams et al. 2016a, this issue). These synthetic contributions were written by working groups of symposium speakers and attendees. The symposium focus on variation complements the 2014 HETEROCLIM conference that took place in Loches, France, as well as several recent Society for Integrative and Comparative Biology symposia examining other aspects of organismal responses to climate change: the 2011 symposium titled “A synthetic approach to the response of organisms to climate change: The role of thermal adaptation” ( Sears and Angilletta 2011 ), and the 2013 symposium titled “Physiological responses to simultaneous shifts in multiple environmental stressors: relevance in a changing world ( Todgham and Stillman 2013 ). Several contributions in this issue (Pincebourde et al. 2016, this issue; Pincebourde and Suppo 2016, this issue) highlight the paucity of studies that explicitly measure spatial variation in temperature, particularly at scales relevant to organisms (cf Potter et al. 2013 ). The “grain” of spatial variation with respect to the size of the organism ( Levins 1968 ) will likely determine whether that variation is a resource to be exploited or a threat to be managed. From a meta-analysis, Pincebourde et al. (2016, this issue) found similar levels of spatial heterogeneity in temperature across scales from single leaves to landscapes. Their random walk model suggests that, up to a point, high spatial heterogeneity helps organisms thermoregulate, highlighting the need to incorporate spatial variation into climate models and predictions ( Sears et al. 2011 ; Woods et al. 2015 ). Detailed measurements at multiple spatial scales suggest that microhabitats may allow small tropical organisms to escape the heat ( Tewksbury et al. 2008 ): even during the hottest parts of the day, more than 70% of potential micro-habitats had surface temperatures that would permit activity by small ectotherms (Pincebourde and Suppo 2016, this issue). At any location on a landscape, temperatures also vary in time from seconds to decades to geologic eras (Dillon et al. 2016, this issue; Poertner et al. 2016, this issue). Changes in temporal variation in temperature associated with climate change ( Wang and Dillon 2014 ; Dillon et al. 2016, this issue) can drastically alter our estimates of how much climate will affect organisms (Sheldon and Dillon 2016, this issue). Our ability to estimate past and to predict future effects will be greatly enhanced by new approaches to measuring and manipulating variation at different timescales, as illustrated by the frequency domain approach outlined by Dillon et al. (2016, this issue). Over longer timescales, changes in the magnitudes of climate oscillations may alter community composition, as more variable climates should select for broader thermal tolerances (eurythermy; Poertner et al. 2016, this issue). Collectively, our synthetic symposium efforts have revealed that the relative importance of variation over different timescales will depend on: (1) organism size ( Stevenson 1985 ; Woods et al. 2015 ) and life history ( Levins 1968 ; Gilchrist 1995 ; Chan et al. 2016 ), and (2) the timescales over which thermal performance curves shift (Williams et al. 2016a, this issue; Dillon et al. 2016, this issue). Even when they are rare, extreme temperatures may disproportionately affect the physiology, ecology, and evolution of organisms and populations (Buckley and Huey 2016, this issue; Dowd et al. 2015 ). Changes in temperature variation at different timescales alter the probability that an organism will experience an extreme and perhaps lethal event (Dillon et al. 2016, this issue). Even when not lethal, extremes can have profound consequences; exposure to sublethal extremes induces a cascade of physiological responses that can have carryover effects leading to increased tolerance of or susceptibility to subsequent exposures to extremes (Williams et al. 2016a, this issue). Williams et al (2016b, this issue) pose the novel hypothesis that, over longer timescales, metabolic compensation (a macrophysiological “rule”; Gaston et al. 2009 ) may be a direct response to adaptation to extremes rather than a compensatory response to maintain growth rates. And both lethal (Buckley and Huey 2016, this issue) and sublethal (Williams et al. 2016a, this issue) extremes may shape the evolution of thermal performance curves much more than do temperatures around the mean (that are most often experienced by most organisms most of the time). Clearly, we need to study extremes explicitly using both theory and experiments. Altogether, the papers in this issue call for us to meld climate science and biology in ways that move us beyond means. Doing so will require novel approaches to methods, experiments, and theory. We need new, more standardized methods for collecting meteorological and microclimatic data to characterize spatial and temporal variation of temperature and other abiotic factors at scales more relevant to organisms. We need to develop new theory for translating abiotic conditions into organismal and population performance, and we need more and better experiments elucidating the effects of both fluctuating and extreme temperatures ( Colinet et al. 2015 ) on the plasticity and evolution of thermal performance curves ( Kingsolver and Woods 2016 ). Understanding the forces that shape performance curves over multiple timescales will be key to making better predictions about how climate change will affect organisms. We thank Lynn Martin for helpful comments on the manuscript. The symposium was made possible by generous support from the National Science Foundation (IOS-1545787), the Society for Integrative and Comparative Biology, and The Company of Biologists. Michael E. Dillon was supported by NSF IOS-1457659, and H. Arthur Woods was supported by NSF IOS-1341485.