In 2050, climate change will have increased the frequency of extreme high temperatures across the globe. Estimations of the impact of this on populations’ vital rates of performance—including growth, development and fitness—have already been made. But these projections forget to consider a critical factor that may change outcomes, according to David Vasseur, associate professor of ecology and evolutionary biology at Yale University and featured speaker at the Life Sciences Center on Jan. 31 as part of this term’s Cramer Seminar Series.
While past models adjusted mean temperatures to reflect the future effects of global warming, they neglected the impacts of accompanying changes in climate variability and skewness.
“Their model assumes that organisms experience a mean temperature, plus some range that represents their average seasonality or the average diurnal temperature range,” Vasseur said. “They didn’t take into account daily, weekly or seasonal extreme events that might have important consequences for populations.”
When Vasseur spoke on the same topic in Kansas City last week, his audience members mentioned the recent drastic hot-cold weather changes in their area. A difference of 60 degrees had prevailed on one day between the minimum and maximum temperatures.
“What’s going on there is variability,” said Vasseur.
Vasseur further demonstrated the prevalence of temperature variability on a graph of Hanover’s average temperatures across months over the past 50 years. Last year’s conditions were superimposed on the graph, with the gap between daily temperature maxima and minima filled in. A significant portion of the temperatures fell outside the band of average temperatures from the last five decades.
In addition, Vasseur proved that leaving out climate variability can profoundly effect mathematical calculations of the mean. According to Jensen’s inequality, the value of a non-linear function defined at the mean is different from the mean value of that function.
“If we have some function representing performance as a consequence of an environmental attribute, of course we can go in and read off what the performance is,” Vasseur said. “But if the environment is swapping between two different states, a minimum and maximum, our performance is also swapping between the values given by the function of the minimum and maximum. If we draw those across and get the average performance for that organism, we now get a different answer.”
Vasseur applied Jensen’s inequality to a model of environmental fitness versus temperature, adjusting the standard deviation of temperature to make the climate more variable and examining its impact on the average fitness of an organism.
“The effect of a warmer temperature on fitness is much more powerful than a cooler temperature—it’s not symmetric,” Vasseur said.
The skewness of temperatures, which Vasseur later found to impact predictions of population performance only minimally, was visualized on contour plots of thermal performance curves comparing relative long-term performance to mean environmental temperature. Variation in temperature degraded organisms’ performance, as they could not use extreme temperatures as well. More skewed environments bent the downwards-sloping ridge on the contour more quickly.
“What this ridge represents is the transition from when warming is good, to when it becomes bad,” Vasseur said. “So more skewed environments have a much sharper transition; the effective climate warming becomes bad much faster.”
In order to determine the actual impact of changes in temperature variance and skewness on population performance predictions instead of only changes in the mean, Vasseur set up a study using a thermal performance curve for a specific citrus pest. He measured the rate of change in performance that resulted from “infinitesimal” changes in the mean temperature, and did the same when altering variance and skewness, holding mean constant.
“We can break that apart given what we’ve learned about how the mean, variance and skewness independently affect things,” Vasseur said. “When we do that, we can attribute certain proportions of the global change in temperature distributions to either the mean or variance. We didn’t need to consider the skewness, because it doesn’t cause performance to really generate a change.”
The study found that changes in the mean temperature accounted for only 32 percent of the shift in temperature distributions. Environmental variance, in addition, was responsible 54 percent of the change.
“Variance alone explained much, much more of the response of population to climate change,” Vasseur said.
In the study, a “smoothing approach” was taken, in which a running average of the data was used to account for organisms’ ability to temporarily protect themselves from extreme events. As a result, the studies’ final findings were even more conservative than their initial estimates, Vasseur said.
“Insects might be able to avoid harsh events for half a day,” Vasseur said. “But they can’t spend three days hiding from the climate, so we took an approach where we tried to incorporate at least some of that information.”
Still, there are ways to improve the data in their study, according to Vasseur. The disparity between organisms’ body temperature and that of the surrounding air should be factored in for more accurate results, by examining influences such as heating and cooling, behavioral thermal regulation and habitat use. Vasseur is currently focusing more research on the acclimation and adaptation responses of organisms to climate changes. While some organisms, such as bacteria, can acclimate to temperature changes in a matter of hours, it may take others generations of development.
“Bacteria can acclimate in hours, probably through the up-regulation of different enzymes that will allow them to better use the temperature they are in,” Vasseur said. “The amount of time it takes, of course, depends on biology.”
Increasing temperature variability, which has been neglected so far in predictions of future global warming’s impact, affects resulting population performance far more than increasing mean temperatures. Source: http://commons.wikimedia.org/wiki/File:Sunlight_(8351294732).jpg