Urban Heat Island Variability and Implications for Heat-Health Mitigation Strategies

Extreme heat events are associated with a significant public health burden. Growing evidence of intra-city spatial and temporal variability in population sensitivity to heat makes it possible to further our understanding of the contribution of urban heat island effects to elevated health risks. An important prerequisite to determining the contribution of urban heat island effects to health risks is a comprehensive understanding spatiotemporal variability in the heat island itself.

We sought to assess temporal variability in the magnitude and spatial form of the spatial urban heat island of Philadelphia, Pennsylvania, USA, focusing exclusively on summer days associated with above-average air temperature. A secondary objective was to understand physical drivers of observed variability. We developed a collection of twenty-one maps of the surface urban heat island in Philadelphia based on LANDSAT imagery and investigated how the spatial pattern at the postal code scale varied based on seasonal, land cover, and meteorological characteristics.

There are significant differences in the magnitude and spatial manifestation of the urban heat island during extreme heat events. In the collection of images we examined, there is a clear seasonal component to the variability whereby the range of surface temperatures across the County is greatest early in the season. In April and May postal codes with a high percentage of high density development are approximately 4°C warmer than the citywide average; this difference drops to 2°C by September. In the late summer postal codes with low-density development and open space more closely resemble the urban core in terms of surface temperature. These findings encourage future investigators of human health and other ecological impacts of the urban heat island to choose a larger sample of images to accurately assess the mean pattern and variability. Residents of areas with high building density are exposed to higher surface temperatures and thus may be at greater risk of heat-related mortality throughout the warm season and this effect is strongest in late spring and early summer.