Afternoon heat islands reflect areas of Durham which are most intensely impacted by high summer temperatures and global warming. During a given summer afternoon one can expect temperatures to range from cooler to warmer based on the level of tree coverage and other factors. Heat stress, heat stroke and even death can result from prolonged exposure to high temperatures without relief.

Heat islands are defined by the Environmental Protection Agency (EPA) as “urban areas characterized by temperatures higher than those of the surrounding non-urban area. As urban areas develop, buildings, roads, and other infrastructure replace open land and vegetation. These surfaces absorb more solar energy, which can create higher temperatures in urban areas.”

These “islands” of persistently higher-temperatures make it harder to get relief from a summer day and more likely to be a place of poor air quality and respiratory stress. Knowing what neighborhoods in our community face these distinct challenges can help us mitigate the effects of heat islands by adding green space and trees, while offering other support for households experiencing potentially harmful exposures during peak temperatures.

This metric reports the percentage of area in each census blockgroup or tract which was in the top quintile for temperature according to the CAPA heat watch study model. In this case, that equated to temperatures at or above 84.5 degrees Fahrenheit. The CAPA study modeled temperature between 3-4pm on a typical summer day in degrees Fahrenheit using a random forest machine learning model, trained on data points collected in July 2021 by Durham volunteers and the Museum of Life and Science as part of the CAPA Heat Watch Program.

Volunteers collected temperature samples throughout Durham which were then used to model temperature variations across the city. More about the project as a whole and data produced by Durham and Raleigh volunteers can be found here.

The CAPA study focused on downtown Durham and surrounding neighborhoods; in the Compass this metric omits census areas which had less than 50% of their area included in the study.