Spatio-temporal interactions of surface urban heat island and its spectral indicators: a case study from Istanbul metropolitan area, Turkey.

A surface urban heat island (SUHI) is a significant meteorological phenomenon of the microclimate and environment in urban territories. Knowledge about the variations of SUHI is critical for urban planning and public welfare. In the current study, the seasonal and spatial changes of the Istanbul SUHI and its interactions with spectral indicators of the urban heat phenomenon including the normalized difference vegetation index (NDVI), tasseled cap wetness (TCW), and surface albedo were analyzed. The National Aeronautics and Space Administration (NASA) L2 thermal products (brightness temperature) of Landsat 8 imageries were used to calculate land surface temperature (LST) values. The thermal islands of the study area were detected based on the Urban Thermal Field Variation Index method. The retrieved LST values showed acceptable agreement with in situ observations of mean daily temperature for all the seasons. Monthly precipitation, however, demonstrated good correlation with summer and autumn LSTs. It is found that the central parts of the metropolitan area were subject to the most intense SUHI in the spring and summer seasons. Outskirt areas showed higher thermal values during cooler seasons of autumn and winter. The results of spatio-temporal interactions of SUHI and the spectral indicators revealed a negative correlation for NDVI and TCW and a positive correlation for surface albedo during different seasons from summer 2017 to spring 2018. The highest and lowest correlations were found between SUHI and TCW (spring) and surface albedo (winter), respectively. The regression results overall suggested that TCW and NDVI were the best indicators of SUHI in Istanbul. Surface albedo was not recommended for seasonal monitoring practices of SUHI in the study area due to the high differences in its seasonal interactions.