Evaluation of spectral indices for characterizing Urban Cool Island during the dry season from Landsat8; a hot semi-arid climate case study of the Gaza Strip.

Abstract

Several Spectral Remote Sensing Indices have been widely investigated for characterizing the spatial patterns of urban heat islands (UHI) phenomenon commonly found in temperate and subtropical climate regions. However, those studies were and still are few in evaluating these indicators in hot and semi-arid climatic regions such as the Gaza Strip. A few studies have shown that this phenomenon occurs in reverse in hot and semi-arid regions, where it is called the urban cool island (UCI) phenomenon. The spectral characteristics of land covers in the hot arid and semi-arid climate regions have a different response to land surface temperatures (LST) from other climatic regions. Therefore, the goal of this paper is to identify and characterize the UCI using spectral indices (SIs) during a dry season over the hot arid, and semi-arid climate area of the Gaza Strip.

SIs have become readily available and are considered promising for characterizing the spectral and thermal patterns of urban and non-urban areas, whereas the thermal spatial pattern can be used later as inputs in many urban studies and climatic modeling applications. The aims of this study are; (1) to identify the presence of the UCI phenomenon over the hot arid and semi-arid climate area of the Gaza Strip; (2) to evaluate 21 SIs (SIs' full names mentioned in table 2) including; ABEI, BAEI, BI, BRBA, BSI, BUI, DBSI, ENBI, IBI, I_imp, MSAVI2, NBI, NBUI, NDBI, NDSI, NDVI, SAVI, UI, VgNIR-BI, VrNIR-BI, and VIBI for characterizing the UCI pattern over the Gaza Strip. The spatial analysis was applied in two land cover scenarios; including the “All” classes scenario (urban, all vegetation types, bare soil, and sand), and the “urban” class. The LST (100 m) and spectral indices (30 m) are derived from Landsat 8 images of summer.

The results of this research provided the existence of the UCI phenomenon in the hot arid and semi-arid climate area of the Gaza Strip. The statistical relationship between SIs and UCI is best-described using the second-order polynomial fit, within the "All" class are the following SIs; DBSI (R² = 0.58), ABEI (R² = 0.52), BSI (R² = 0.50), and BAEI (R² = 0.44), whilst within the "urban" class are the following SIs; MSAVI2 (R² = 0.50), P_imp (R² = 0.40), ABEI (R² = 0.33), and NBI (R²= 0.33).