Something is always lost in translation from the lab to the real world—a fact that’s captured in the concept of “ecological validity,” the extent to which the conditions in an experiment match those you’d encounter in the wild. One way around this problem is to skip the lab entirely and search for patterns in naturally occurring data. Cyclists, who gather extensive data about every pedal stroke with their power meters, make particularly good subjects for this kind of analysis, as illustrated by a new study in the International Journal of Sports Physiology and Performance.
A team of Spanish researchers coordinated by David Barranco-Gil of the Universidad Europea de Madrid pooled eight years of data from 74 world-class cyclists (48 men and 26 women) and asked a simple question: how does air temperature affect performance? The answers offer some useful insights about the differences between cycling and running, and about the differences (or lack thereof) between how men and women respond to heat and cold.
To assess performance, the researchers first divided the power data into temperature ranges in 5-degree-Celsius increments from below 5 degrees Celsius (41 degrees Fahrenheit) to above 35 degrees Celsius (95 degrees Fahrenheit), using the built-in temperature sensors on the riders’ bike computers. In each temperature range, they then searched for the highest average power produced over four different durations: 5 seconds, 30 seconds, 5 minutes, and 20 minutes. This is the approach the same researchers used to evaluate the effects of altitude on performance in another recent study, and it has the advantage of including race data from big events like the Tour de France, where the cyclists are presumably digging as deep as they can.
Here’s an example of the data for the five-minute efforts, with the men’s data shown by black circles and women by white circles:
As you’d expect, the power outputs drop a bit when the temperature gets too high. There are a bunch of different physiological reasons for this, including changes in heart and nervous system function as you get hotter and possibly dehydrated.
Performance also drops when it gets too cold. This too makes intuitive sense, and you can come up with plausible physiological reasons like changes in muscle properties as temperatures drop. Surprisingly, though, previous laboratory studies have produced mixed results about whether cold temperatures hurt endurance performance. In practice, it’s not clear whether the drop-off in the new study is due to some fundamental physiological limitation, or simply reflects the fact that it’s really unpleasant to try to cycle up a mountain when the temperatures are hovering around freezing. Either way, the pros ride slower when it’s cold.
When you aggregate the results, you find that performance is maximized over a broad range between about 50 and 77 degrees Fahrenheit. (The upper threshold of 77 degrees isn’t as precise as it sounds: 77 degrees Fahrenheit is 25 degrees Celsius, one of the 5-degree-Celsius increments.) Compare that to a study I wrote about last fall that aggregated a century’s worth of race results and meteorological data to find the optimal temperature for running. The conclusion in that case was a sweet spot of 50 to 63.5 degrees Fahrenheit.
Why would cyclists be able to handle higher temperatures better than runners? Probably because cyclists riding outdoors generate their own cooling wind—a factor that was notably overlooked in many lab-based studies of cycling and heat tolerance.
Overall, there was no statistically significant difference between male and female cyclists. When each data set was analyzed separately, however, there were differences. The optimal temperature range for men only was 50 to 86 degrees Fahrenheit, while for women only it was 41 to 77 degrees Fahrenheit. This might indicate that women were better able to tolerate cool temperatures without slowing down, while men handled hot temperatures better. Again, you can come up with plausible physiological explanations for why this should be true, related to body surface area, size differences, the insulative properties of body fat, sweat gland distribution, and so on. But to be honest, you could argue it both ways. For now, the data is just a hint that there might be something interesting for future study, not proof one way or the other.
For the record, if you’re planning a ride in hot or cold conditions and want to estimate the handicap, here’s how much power output decreased from peak levels:
- Below 41 degrees Fahrenheit: 9 to 17 percent (men); 10 to 18 percent (women)
- 41 to 50 degrees Fahrenheit: 2 to 7 percent (men); no significant difference (women)
- 77 to 86 degrees Fahrenheit: no significant difference (men); 3 to 6 percent (women)
- 86 to 95 degrees Fahrenheit: 3 to 5 percent (men); 5 to 7 percent (women)
- Above 95 degrees Fahrenheit: 9 to 13 percent (men); 14 to 16 percent (women)
The ranges (e.g. 9 to 17 percent) indicate the power drop-offs in the four different durations analyzed, between 5 seconds and 20 minutes. Overall, the patterns are very similar for the different durations, though it looks to me as though the drop-off in sprinting is more pronounced in cold conditions.
A final point to note: the analysis here considers only air temperature. In reality, the effects of weather on endurance performance also include humidity, solar radiation, and wind speed. That’s probably one of the reasons the sweet spot is so wide: cool days feel OK if it’s calm and sunny, warm days aren’t so bad if it’s dry and overcast. If you held all those other factors constant, I suspect there would be a narrower range of air temperatures that maximize performance. Still, my overall takeaway from these results is that it has to be pretty darn hot or cold before I can use the temperature as an excuse for a poor performance.
Hat tip to Chris Yates for additional research. For more Sweat Science, join me on Twitter and Facebook, sign up for the email newsletter, and check out my book Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance.