Dillon Michael E, Woods H Arthur, Wang George, Fey Samuel B, Vasseur David A, Telemeco Rory S, Marshall Katie, Pincebourde Sylvain
*Department of Zoology and Physiology and Program in Ecology, University of Wyoming, Laramie, WY 82071, USA
Division of Biological Sciences, University of Montana, Missoula, MT 59812, USA.
Integr Comp Biol. 2016 Jul;56(1):14-30. doi: 10.1093/icb/icw024. Epub 2016 Jun 1.
Over the last few decades, biologists have made substantial progress in understanding relationships between changing climates and organism performance. Much of this work has focused on temperature because it is the best kept of climatic records, in many locations it is predicted to keep rising into the future, and it has profound effects on the physiology, performance, and ecology of organisms, especially ectothermic organisms which make up the vast majority of life on Earth. Nevertheless, much of the existing literature on temperature-organism interactions relies on mean temperatures. In reality, most organisms do not directly experience mean temperatures; rather, they experience variation in temperature over many time scales, from seconds to years. We propose to shift the focus more directly on patterns of temperature variation, rather than on means per se, and present a framework both for analyzing temporal patterns of temperature variation and for incorporating those patterns into predictions about organismal biology. In particular, we advocate using the Fourier transform to decompose temperature time series into their component sinusoids, thus allowing transformations between the time and frequency domains. This approach provides (1) standardized ways of visualizing the contributions that different frequencies make to total temporal variation; (2) the ability to assess how patterns of temperature variation have changed over the past half century and may change into the future; and (3) clear approaches to manipulating temporal time series to ask "what if" questions about the potential effects of future climates. We first summarize global patterns of change in temperature variation over the past 40 years; we find meaningful changes in variation at the half day to yearly times scales. We then demonstrate the utility of the Fourier framework by exploring how power added to different frequencies alters the overall incidence of long-term waves of high and low temperatures, and find that power added to the lowest frequencies greatly increases the probability of long-term heat and cold waves. Finally, we review what is known about the time scales over which organismal thermal performance curves change in response to variation in the thermal environment. We conclude that integrating information characterizing both the frequency spectra of temperature time series and the time scales of resulting physiological change offers a powerful new avenue for relating climate, and climate change, to the future performance of ectothermic organisms.
在过去几十年里，生物学家在理解气候变化与生物体表现之间的关系方面取得了重大进展。这项工作大部分都集中在温度上，因为它是保存最完好的气候记录要素，在许多地方预计未来还会持续上升，而且它对生物体的生理、表现和生态有着深远影响，尤其是构成地球上绝大多数生命的变温生物。然而，现有的关于温度与生物体相互作用的许多文献都依赖于平均温度。实际上，大多数生物体并不会直接经历平均温度；相反，它们会在从秒到年的许多时间尺度上经历温度变化。我们建议将重点更直接地转向温度变化模式，而不是温度均值本身，并提出一个框架，用于分析温度变化的时间模式，并将这些模式纳入对生物体生物学的预测中。特别是，我们主张使用傅里叶变换将温度时间序列分解为其组成的正弦波，从而实现时域和频域之间的转换。这种方法提供了：（1）标准化的方式来可视化不同频率对总时间变化的贡献；（2）评估温度变化模式在过去半个世纪中如何变化以及未来可能如何变化的能力；（3）明确的方法来操纵时间序列，以提出关于未来气候潜在影响的“如果……会怎样”的问题。我们首先总结过去40年全球温度变化的模式；我们发现在半日到年的时间尺度上变化有显著意义。然后，我们通过探索添加到不同频率的功率如何改变高低温长期波动的总体发生率，来证明傅里叶框架的实用性，并发现添加到最低频率的功率大大增加了长期热浪和寒潮的概率。最后，我们回顾了关于生物体热性能曲线响应热环境变化的时间尺度的已知情况。我们得出结论，整合表征温度时间序列频谱和由此产生的生理变化时间尺度的信息，为将气候及气候变化与变温生物的未来表现联系起来提供了一条强大的新途径。
