Publications

Nanjam, R; Ahmadi, FF (2021). A New Spatiotemporal Model for Analyzing the Variations of Urban Heat Islands Using Remotely Sensed Multi Spectral Images: The Case of Mashhad City, Iran. JOURNAL OF THE INDIAN SOCIETY OF REMOTE SENSING, 49(10), 2489-2502.

Abstract
Urban heat islands (UHIs) refer to areas of the earth's surface with higher temperatures than neighboring regions. The spread of this phenomenon varies according to spatial and temporal dimensions, and these variations are affected mainly by changes in land cover and heat emissions from human activities. Due to the high variations in the unit of time, the spatial continuity, and expansion of this phenomenon, remote sensing is a suitable tool for studying it. The purpose of this study is to present a fixed model with general thresholds for studying UHIs. Therefore, the fixed model was adjusted based on the features that exhibit similar behaviors to respond to the temporal and spatial variations of the urban heat and cool islands. To achieve this objective, one of the cities of Iran with high UHIs intensity and vastity was assigned as the study area. Therefore, the UHIs of Iranian cities were investigated from the Moderate Resolution Imaging Spectroradiometer (MODIS) production by applying the Getis-ord algorithm on the land surface temperature (LST) map obtained by the split-window method. Then, the UHIs of Mashhad were studied by applying clustering on the LST map of Landsat images. In this study, the UHIs of different seasons of 2018 for Mashhad were interpreted based on the LST, and the results showed that the UHIs show different behavior in different seasons in terms of location and intensity. Therefore, it is necessary to extract features that exhibit similar behavior for UHIs. Features or the emergence of UHIs were analyzed by statistical methods, principal component analysis (PCA), and regression. The results of the principal component analysis showed that the four indices (Normalized Difference Vegetation Index (NDVI), Normalized Difference Build-Up Index (NDBI), Normalized Difference Bareness Index (NDBaI), and Normalized Difference Water Index (NDWI) are useful features for studying UHIs. On the other hand, these features were suitable for presenting a fixed model for examining the UHIs of different seasons. The results of the fixed multivariate regression model with different thresholds had an average overall accuracy (OA) of 82.15% for different seasons, and the results for the fixed model with fixed thresholds had an average overall accuracy of 83.54% for summer and winter and 59.1% for spring and autumn.

DOI:
10.1007/s12524-021-01404-8

ISSN:
0255-660X