Climate Change and Health: Extreme Heat Methods

Map

The map shows the average number of extreme summer heat days a year, from 2007 to 2016, rounded to the nearest whole integer. The categories mapped are: more than 14 extreme heat days per year (red); greater than 9 to 14 extreme heat days per year (orange); 9 or fewer extreme heat days per year (yellow); and insufficient data to perform the calculation (gray).

“Extreme heat days” are defined as those days from June 1 to August 31 in the years 2007 to 2016 on which the maximum temperature exceeded the 90th-percentile value. We used the June to August daily maximum temperatures from the 1961 to 1990 reference period for the same monitoring station to calculate the 90th percentile. The 90th percentile value is among the more common ways to define extreme heat and was likewise used in NRDC’s 2011 heat-mapping analysis.¹

There are a total of 92 days from June 1 to August 31, so the expected number of days that would exceed the 90th percentile value is 9.2 days. (Or, rounding to the nearest whole number, 9 days.) Therefore, for the 2007 to 2016 study period, locations with more than 9 days per summer, on average, with temperatures above that location’s 90th-percentile reference value were considered to have “more than expected” numbers of days of extreme heat. The highest category shown on the map (more than 14 days) gives a sense of those locations that now experience more than two weeks’ worth of extreme heat days per year.

Note that “extreme heat” is defined by local temperatures at each site; the map does not compare temperatures in one part of the country to those in another region.

Data sources

Historical temperature data for all cooperative weather stations for all years were downloaded from the Global Historical Climatology Network (GHCN), formerly the National Climatic Data Center.² Geographic detail for stations was also downloaded from GHCN, which defines cooperative stations as “U.S. stations operated by local observers which generally report max/min temperatures and precipitation. National Weather Service (NWS) data are also included in this dataset. The data receive extensive automated + manual quality control.”

Calculations

For each station, we calculated the 90th percentile for the maximum daily temperature during the reference period of June, July, and August of 1961 to 1990. We then computed the total number of days on which the maximum daily summer temperature exceeded the 90th-percentile reference for that station for each year in the most recent decade, from 2007 to 2016. County-level averages of the number of extreme heat days were then calculated by averaging station-level data from stations within the county and from each of the 10 years.

Stations were excluded from the analysis if: (a) they had data for less than 75 percent of possible reference days available; (b) one entire year from the 2007 to 2016 study period was missing; or (c) they had data for less than 75 percent of possible study period days available.

Population

Estimates of population living in counties with more than nine extreme heat days per year (i.e.,"more than expected") were taken from the 2015 U.S. Census.³

Health data

Heat-related health data largely are from the Centers for Disease Control and Prevention’s (CDC) National Environmental Public Health Tracking Network,⁴ which tallies deaths with heat specifically listed as a direct or indirect cause, including deaths from heatstroke.⁵ This method significantly underestimates the true number of heat-related deaths, however. For example, a deadly heart attack during a heat wave may not be counted as a “heat-related death,” even if exposure to extreme heat triggered the heart attack.⁶ In addition, since the CDC does not report data in years when the total number of heat-related deaths is determined to be too low, it may be underestimating the health impact.

Citations

1. World Meteorological Organization/World Health Organization, Heatwaves and Health: Guidance on Warning-System Development (WMO-No. 1142), http://www.who.int/globalchange/publications/WMO_WHO_Heat_Health_Guidance_2015.pdf?ua=1 (accessed August 22, 2017).
2. NOAA, Global Historical Climatology Network (GHCN), https://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets/global-historical-climatology-network-ghcn (accessed June 15, 2017).
3. U.S. Census Bureau, “2011–2015 American Community Survey 5-Year Estimates,” 2015, https://factfinder.census.gov (accessed June 5, 2017).
4. Centers for Disease Control and Prevention, National Environmental Public Health Tracking Network, 2017, https://ephtracking.cdc.gov/showHome.action (accessed June 15, 2017).
5. Centers for Disease Control and Prevention, National Environmental Public Health Tracking Network, “Indicators and Data,” 2017, https://ephtracking.cdc.gov/showIndicatorPages (accessed June 15, 2017).
6. Centers for Disease Control and Prevention, “Picture of America Report: Heat-Related Illness,” 2017, https://www.cdc.gov/pictureofamerica/pdfs/Picture_of_America_Heat-Related_Illness.pdf (accessed August 15, 2017).

Credits

Project teams: NRDC Science Center, Communications Department, Climate & Clean Air program

Researchers and writers (alphabetical order): Tom Barkley, Juanita Constible, Kim Knowlton, Will Shafer

Reviewers (alphabetical order): Sean Alcorn, Hunter Jones, Linda Rudolph, Sayantan Sarkar, Christina Swanson

Editor: Sarah Engler

Data analysis and map production (alphabetical order): Alex Hart, Kate McKenney, ZevRoss Spatial Analysis