Taking the temperature of climate change

Antonella Zanobetti

June 23, 2015 — Antonella Zanobetti, principal research scientist in the Department of Environmental Health, discusses a new study that found that people appear to adapt over time as temperatures creep higher, but also may face increased mortality risk from extreme temperature swings—and their level of risk may depend on where they live.

What did you discover that hadn’t been known before?

It’s long been known that there is an effect of both heat and cold on mortality and that heat-related mortality is lower in warmer U.S. cities, and cold-related mortality is lower in colder cities. But, until now, no one had looked at how temperature changes over time were affecting health across the nation.

We looked at the relationship between mortality and temperature by region, from the 1960s until now, to see how that relationship is changing. We looked at six different U.S. regions, grouping cities into those regions according to their seasonal temperatures and humidity. We found that, as average summer temperatures increased, the effect of very warm temperatures on mortality decreased. But we found the opposite effect with very cold temperatures—as average winter temperature increased, so did mortality. We also found that temperature-related mortality varied by region. For example, along the Pacific coast, the climate is milder than in other parts of the country, so when there’s an extreme temperature event there, people appear to be more susceptible.

We read a lot in the news about people dying during heat waves. But if people are adapting to warmer summer temperatures, as your paper suggests, does that mean we don’t need to worry so much about times of extreme heat?

No! It’s true that there’s been a protective effect from warming temperatures—because people are acclimating over time to the rise in temperatures, and air conditioning use has grown over time. Our results still show an increased mortality due to both heat and to cold temperature; even if these are modest changes their impact is not trivial. Moreover, there is a different effect by region, so adaptation is different depending on where people live. But looking forward 50 years, the biggest problem might be temperature variability—swings in temperature—which seems to have a stronger effect on mortality than average temperature, and which we would expect to increase with climate change.

How might current risk models change based on your findings?

Current risk assessments for climate change assume that the relationship between temperature and daily deaths in cities will not change, but our estimates show they already have, and our calculations will allow risk assessors to extrapolate how much they will change further in the future. We’ve shown that these changes will vary across the U.S. and that, in general, there will be more cold-related deaths and fewer heat-related deaths in the future than current models predict.

Karen Feldscher