There is insufficient data to identify areas most at risk to the impacts of extreme heat

Mapping extreme heat risk requires insight into a given area’s built environment, the underlying demographics and their activity patterns, and heat exposure over time; the associated data may be missing or unincorporated.

Some considerations for the difficulties involved:

We explored the similarities/differences in existing heat vulnerability mapping approaches. Previous models (1) failed to characterize extreme heat variations (e.g. days with different temperature extremes, difference between single hot days and prolonged heat events); (2) generally lacked a validation of associations between mapping results and various health outcomes (e.g. only relying on mortality data for validation but no validation of morbidities or clinical visits); (3) had a low ability to incorporate spatiotemporal variability of demographic patterns (e.g. difference in daytime/nighttime population patterns due to daily mobility); (4) were unable to adopt perceived heat exposure (e.g. Universal Thermal Climate Index); and (5) did not measure heat vulnerability at street-/building-levels even though 3-dimensional urban forms exist across major cities. Additionally, most applications (including those from local governments) were developed from a top-down perspective without a participatory design¹

Census tracts are commonly used for assessing heat vulnerability (e.g., Hsu et al. 2021). They offer several advantages, such as providing a general picture of where disadvantaged populations reside in hot areas. Census tracts also align better with other datasets, particularly health data, making them more convenient for comprehensive analyses. However, Census tracts do not capture localized factors that contribute to heat risk, such as specific land-use patterns or the presence of impervious surfaces and tree cover. Indeed, within a tract, there can be significant localized variability in the capacity of communities and individuals to take action. In addition, while Census tracts are useful for obtaining funding, most scales of governance are not tied to these boundaries, limiting the actionability of the data generated at this scale. In cases where there are few tracts or exceptionally large tracts, Census blocks are used to more precisely identify vulnerable groups and the factors influencing heat risk. Analyzing data at the block level can result in more targeted interventions. However, even at this scale, pockets of vulnerability can exist beneath the block level, where most socioeconomic and health indicators are not available or discernible… The temporal scale is also critical in heat risk mapping. Many heat mapping campaigns are conducted on just 1 day and usually not on the hottest days, providing an incomplete picture of exposure. The information and tools needed for responding to acute heat events differ from those needed for long-term resiliency planning. Collecting data over time, rather than through “one-off” campaigns, is essential for developing baselines for heat exposure and quantifying potential benefits from mitigation programs like tree planting and cool roofs.²