How can researchers measure heat for health research?
Drawing on Data Considerations for Estimating Ambient Heat Exposure for Environmental Epidemiological Studies (opens in new tab), Gause et al. (2026), this page serves as a guide for selecting data and metrics to study heat in the context of health research.
Explore key approaches and considerations for heat exposure research, and use the interactive tool below to receive
tailored to your individual research question and health study design.
Air temperature is the most common way researchers measure heat because it is simple, widely available, and easy to interpret.
However, other metrics exist to capture how hot it feels or how the body responds under different atmospheric conditions. These metrics combine temperature with other variables like humidity, wind, and solar radiation. For health studies focusing on extreme heat events rather than average conditions, additional metrics and approaches are available.
Because people acclimate to their local climate, some analyses may choose to use a e.g., percentiles of temperature relative to the past 10 years. Relative metrics can be useful for cross-site and cross-country comparisons. For example, an extreme heat day in Phoenix, AZ may have a much higher absolute temperature than an extreme heat day in Seattle, WA, but both represent an unusually hot day for each city's residents. instead of absolute temperature value. For more information, read "Relative and absolute exposures in multi-site and multi-cohort research" (opens in new tab) led by Emma Nichols and authored by the GECC team.
Jump to:
Dry-bulb temperature
Air temperature.
Air temperature.
Dry-bulb temperature is how most people measure temperatures in their daily lives, so it's easily conveyed to the public. It's also easy to measure, its data is widely available, and it is appropriate for health studies assessing the full spectrum of cold to hot temperatures.
Heat Index (HI)
Incorporates temperature and humidity.
Humidity slows down perspiration, which makes it feel hotter than it actually is. HI estimates what the temperature "feels like" if you're resting in the shade.
Developing warning systems (opens in new tab) for heat-related health consequences, such as fatigue or heat stroke.
Wet-Bulb Globe Temperature (WBGT)
Incorporates temperature, humidity, wind, and solar radiation.
Measures The impact of excess heat on the body by estimating what the temperature "feels like" if you're doing physical activity in the sun.
Establishing safety thresholds for outdoor workers or school-age children who spend a lot of time outside, especially during periods of extreme heat.
Universal Thermal Climate Index (UTCI)
Measures The impact of excess heat on the body based on the human body's Body core temperature, sweat rate, skin wettedness, etc. to temperature, humidity, and solar radiation (opens in new tab).
Accounts for responses like body core temperature, sweat rate, and skin wetness. Works for both hot and cold conditions across climates and seasons.
Studying the impact of weather on human health and comfort to inform daily forecasts (opens in new tab), epidemiological studies, and climate impact research, especially across various geographical regions.
Extreme heat
A single day is typically labelled extreme if it exceeds a specific threshold of intensity and duration of heat.
May be defined on an Fixed values of a chosen metric, such as daily maximum temperature or Relative to local climate scale.
One day crosses the threshold
Heat waves
A period of consecutive, extreme days characterized according to duration, severity, and timing in the season.
Designated using locally relevant thresholds.
Several consecutive days cross the threshold
Additional approaches for characterizing extreme heat
Jump to:
For studies in a single location, researchers may choose to use data from a local weather station. However, for modern health studies at the population-level, heat is best assessed using large-scale gridded weather datasets, since these are modeled to be spatially and temporally complete over large areas. For example, ERA5-Land is a satellite reanalysis product that covers the entire globe back in time to 1950 with an hourly temporal resolution and a spatial resolution of ~9km.
Gridded data are the most comprehensive option, but keep in mind that they are estimated with uncertainty, oftentimes interpolating values between weather stations. Differences across time and broad space are captured in these datasets but not intra-urban heat variation from urban heat island effects.
Gridded weather datasets are useful for researchers across disciplines because they are broad in coverage, publicly available, and standardized, but researchers should always defer to their specific study aims in selecting a data source. There are three main types of weather datasets available:
Direct observation
Weather data recorded either by a fixed weather station or using a personal monitoring device. Raw data that has been directly captured and not subject to any modelling or estimation.
Ideal for validating models, understanding short-term variation, evaluating human exposures in specific settings, and analyzing small samples.
Dynamically modelled estimates
Apply Models derived from first principles of physics, requiring fewer data points and creating models with coarser time and space scales to weather station and satellite data to interpolate conditions between observations — for areas that fall between weather stations.
Statistically modelled estimates
Identify patterns within existing observed data using mathematical models.
Ideal for studies that require very fine spatial resolution in a single country or region context; less ideal for international comparison. Keep in mind that even fine resolution statistical models are not always appropriate for measuring intra-urban variation because these models leave out hyper-local urban heat island mechanisms.
Ex: PRISM, gridMET, Daymet, TopoWx
Explore available datasets
Gridded datasets commonly used for ambient air temperature metrics in environmental health studies
Click a tile below to learn more — the "GECC-recommended" tile flips over to show a recommendation, and the others link out to the relevant tables.
Gridded climate datasets
Most commonly used (opens in new tab)
Most commonly used weather data products, especially in the US
From Table 1, Data Considerations for Estimating Ambient Heat Exposure for Environmental Epidemiological Studies, Gause et al. (2026)
Full catalog (opens in new tab)
Exhaustive list of weather data products available for researchers
From eTable 1, Data Considerations for Estimating Ambient Heat Exposure for Environmental Epidemiological Studies, Gause et al. (2026)
What time scale of weather exposure are you studying? Weather research typically involves calculating window averages or aggregating data in order to identify patterns. However, this analysis first requires selecting an appropriate temporal scale for your health study. When selecting a temporal scale, researchers should be sure to consider their overall study design, as well as the particular health effects they would like to observe.
Hours to days
Compare health outcomes on hot days with those on comparable cooler days.
What would the health outcome have been if the day had not been unusually hot?
Weeks to months
Aggregate exposure over longer windows and compare them with relative periods of etiologic interest.
What are the health effects of a particularly warm vs. less warm period, accounting for both space and time?
Years to decades
Use annual, multi-year, or decadal exposure summaries, cohort health studies, and ecological comparisons across regions to understand long-term exposures.
What are the health consequences of prolonged exposures to higher temperatures or repeated exposures to extreme temperatures?
Researchers must consider the spatial extent and resolution of available weather data when designing their health studies.
See the GECC GitHub repository (opens in new tab) for a tutorial on aggregating gridded ERA5-Land exposure data for use in ecological health studies.
The following tool provides specific guidance, relevant data sets, and suggested reading across common heat-related research scenarios.
Select one or more applicable situations to receive expert recommendations to guide your research.
Your recommendations
Select a situation above to see recommendations here.
Interested in reading more about heat metrics? Check out this resource from the World Resources Institute: Beyond the Thermometer: 5 Heat Metrics That Drive Better Decision-Making (opens in new tab)
The GECC is funded by the National Institute on Aging (NIA) U24AG088894.