Socioeconomic status and land cover as predictors of the urban heat island effect in Tempe, Arizona

Abstract

Developed land emits heat more effectively than rural land. This results in an urban heat island effect, where cities have hotter temperatures than surrounding rural areas. Urban heat islands pose a public health risk in many cities and especially affect areas of lower socioeconomic status, where people are more vulnerable to extreme heat conditions. The Phoenix Metropolitan Area in Arizona is one of the fastest growing metropolitan areas in the United States and regularly experiences extreme heat in the summer. Tempe, a city within the metropolitan area, has outlined a plan to decrease the urban heat island effect by increasing tree cover to 25% by 2040. Landsat 8 OLI/TIRS satellite imagery was used to estimate land surface temperature (LST), a measure commonly associated with urban heat island effects. A land cover classification and US Census data were used to predict mean LST in Tempe. Exploratory regression and spatial regression identified a six-variable model with increases in mean household income, college population, grass land cover, and water cover all decreasing mean LST, while increases in urban land use and a spatial lag variable increased mean LST. Although overall estimates of tree cover were 23% of the land surface, estimates were high as the classification model overestimated tree cover due to the spatial resolution of the Landsat 8 sensor. Results suggest that although Tempe has made progress in its goal, there are discrepancies between areas of differing socioeconomic status.