How much does heat slow a marathon? What the race-result studies
Who this applies to: finishers of big-city road marathons (Boston, New York, Twin Cities, Grandma's, Richmond, Hartford, Vancouver in one study; Paris, London, Berlin, Boston, Chicago, New York in the other). The studies are of race results and weather, not lab trials; they describe groups of runners, not what will happen to one person.
What we found
- Ely and colleagues (Medicine & Science in Sports & Exercise, 2007) compared marathon results with the wet-bulb globe temperature (WBGT) on race day, across seven marathons and up to 36 years of each race.
- The top three men were slower than the course record by 1.7% when WBGT was 5.1 to 10 degrees C, 2.5% at 10.1 to 15 degrees C, 3.3% at 15.1 to 20 degrees C and 4.5% at 20.1 to 25 degrees C (means).
- The top women followed a similar trend, 3.2%, 3.2%, 3.8% and 5.4% slower than the course record over the same four bands, but the differences among the bands were not statistically significant for women.
- The 25th, 50th, 100th and 300th place finishers slowed more than the fastest runners as WBGT increased: slower runners pay a larger heat penalty.
- The same group's review (Sports Medicine, 2007) summarises 140 race-years of data: marathon times slow progressively once weather warms above 5 to 10 degrees C WBGT, men and women are affected similarly, and slower runners suffer a greater penalty than elite runners.
- El Helou and colleagues (PLoS One, 2012) analysed 1,791,972 finishers' performances in six major marathons from 2001 to 2010. Air temperature had the largest influence of all weather factors; the optimal temperature differed by performance level, and above it running speed fell and withdrawal rates rose. Humidity, dew point, pressure and most pollutants had no significant effect once temperature was counted (ozone did, perhaps through temperature).
- A related study (Ely et al., 2007, MSSE) found that the fastest marathons were run at about 12 degrees C air temperature, and cloud cover or low sun was not associated with fast times.
Worked example (our arithmetic, using the elite figures as a floor)
The gap between the coolest and the warmest band for the top men is 4.5% minus 1.7%, which is 2.8 percentage points.
- For a runner whose cool-day time is 3 hours (180 minutes), 2.8% is about 5.0 minutes.
- For a 4-hour runner (240 minutes), 2.8% is about 6.7 minutes.
- For a 5-hour runner (300 minutes), 2.8% is about 8.4 minutes.
Because the studies found slower finishers lose more than the elite, these minutes are a lower bound for recreational runners, not a prediction.
What this means in practice
- Check the forecast WBGT or temperature, not just the air temperature, the week before a race; adjust goal pace for warm bands rather than chasing a cool-day plan.
- The evidence is observational (race results against weather), so it shows the average slowing in a crowd, not an individual's limit. Heat illness is a separate safety question: follow race medical advice.
Sources
- Ely MR, Cheuvront SN, Roberts WO, Montain SJ. Impact of weather on marathon-running performance. Med Sci Sports Exerc 2007. https://doi.org/10.1249/mss.0b013e31802d3aba
- Ely MR, Cheuvront SN, Montain SJ. Marathon performance in thermally stressing conditions. Sports Med 2007. https://doi.org/10.2165/00007256-200737040-00012
- El Helou N et al. Impact of environmental parameters on marathon running performance. PLoS One 2012. https://pmc.ncbi.nlm.nih.gov/articles/PMC3359364/
- Ely MR et al. Neither cloud cover nor low solar loads are associated with fast marathon performance. Med Sci Sports Exerc 2007. https://doi.org/10.1249/mss.0b013e318149f2c3
What we are unsure of: the WBGT bands are race-day conditions at the start or averaged; individual acclimatisation, pacing and hydration were not measured.