The structure of urban environments is known to alter local climate, in part due to changes in land cover. A growing subset of research focuses specifically on the UHI in terms of land surface temperature by using data from remote sensing platforms. Past research has established a clear relationship between land surface temperature and the proportional area of land covers, but less research has specifically examined the effects of the spatial patterns of these covers. This research considers the rapidly growing City of Phoenix, Arizona in the United States. To better understand how landscape structure affects local climate, we explored the relationship between land surface temperature and spatial pattern for three different land uses: mesic residential, xeric residential, and industrial/commercial. We used high-resolution (2.4 m) land cover data and an ASTER temperature product to examine 90 randomly selected sample sites of 240 square-meters. We (1) quantify several landscape-level and class-level landscape metrics for the sample sites, (2) measure the Pearson correlation coefficients between land surface temperature and each landscape metric, (3) conduct an analysis of variance among the three land uses, and (4) model the determinants of land surface temperature using ordinary least squares linear regression. The Pearson's correlation coefficients reveal significant relationships between several measures of spatial configuration and LST, but these relationships differ among the land uses. The ANOVA confirmed that mean land surface temperature and spatial patterns differed among the three land uses. Although a relationship was apparent between surface temperatures and spatial pattern, the results of the linear regression indicate that proportional land cover of grass and impervious surfaces alone best explains temperature in mesic residential areas. In contrast, temperatures in industrial/commercial areas are explained by changes in the configuration of grass and impervious surfaces.
