Variability in annual temperature cycle in the urban areas of the United States as revealed by MODIS imagery

Due to its large spatial coverage and frequent revisit, satellite-derived land surface temperature (LST) has been recently used to explore annual temperature cycle (ATC) variations at regional and global scales. However, variability in seasonality of LSTs has not been examined in detail, particularly in urban areas where elevated temperatures are normally observed. By assuming repetitive temperature cycles, this study aims to reveal differences in ATC parameters between urban and rural areas and the impacts of surface urban heat island (UHI) on the ATC range over the continental United States. To this end, urban areas of larger than 10 km2 (a total of 1856 urban polygons) in the continental United States were identified from the map of urban extents produced by Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light data from 2012. The corresponding rural polygons of the same size were generated by using a buffer method. The ATC parameters were optimized using a sinusoidal function fitted with the 8-day MODIS LST composite data. Results showed that urban and rural areas exhibited a significant difference, with a p-value <0.01, in the ATC parameters, including mean annual surface temperature (MAST), yearly amplitude of surface temperature (YAST), and the revised phase shift parameter. The higher MAST and YAST, but lower phase shift values, were mainly associated with the urban areas. This finding indicated that urban areas contributed to changes in extreme temperatures (the minimum and maximum temperatures) as well as to an overall warming. The regression analysis suggested that surface UHI intensities were positively correlated with the differences in MAST (R2 = 0.9) and YAST (R2 = 0.5) between urban and rural areas, but negatively correlated (R2 = 0.2) with the differences in the revised phase shift parameter. In addition, highest surface UHIs (∼3 k) and largest differences in ATC parameters were observed in tropical regions, followed by temperature zones, continental (cold) zones, and arid zones. Overall, this study revealed that urbanization-induced land cover changes could influence urban systems by enhancing temperature variations. However, it should be cautioned that uncertainty in the analysis may arise from the characterization of surface UHIs for the selected ∼2000 cities with varying configuration and morphological dimensions.