AN INVESTIGATION OF REMOTELY SENSED URBAN HEAT ISLAND CLIMATOLOGY

Abstract

Satellite remotely sensed temperatures are widely used for urban heat island (UHI) studies. However, the abilities of satellite surface and atmospheric data to assess the climatology of UHI face many unknowns and challenges. This research addresses the problems and potential for satellite remotely sensed UHI climatology by examining three different issues. The first issue is related to the temporal aggregation of land surface temperature (LST) and the potential biases that are induced on the surface UHI (SUHI) intensity. Composite LST data usually are preferred to avoid the missing values due to clouds for long-term UHI monitoring. The impact of temporal aggregation shows that SUHI intensities are more notably enhanced in the daytime than nighttime with an increasing trend for larger composite periods. The cause of the biases is highly related to the amount and distribution of clouds. The second issue is related to model validation and the appropriate use of LST for comparison to modeled radiometric temperatures in the urban environment. Sensor view angle, cloud distribution, and cloud contaminated pixels can confound comparisons between satellite LST and modeled surface radiometric temperature. Three practical sampling methods to minimize the confounding factors are proposed and evaluated for validating different aspects of model performance. The third issue investigated is to assess to what extent remotely sensed atmospheric profiles collected over the urban environment can be used to examine the UHI. The remotely sensed air and dew-point temperatures are compared with the ground observations, showing an ability to capture the temporal and spatial dynamics of atmospheric UHI at a fine scale. Finally, a new metric for quantifying the urban heat island is proposed. The urban heat island curve (UHIC), is developed to represent UHI intensity by integrating the urban surface heterogeneity in a curve. UHIC illustrates the relationship between the air temperature and the urban fractions, and emphasizes the temperature gradients, consequently decreasing the impact of the data biases. This research illustrates the potential for satellite data to monitor and increase our understanding of UHI climatology.