Examining and Modelling the Determinants of the Rising Land Surface Temperatures in Arabian Desert Cities: An Example from Riyadh, Saudi Arabia

Following the oil boom in the 1970s, Riyadh, the capital of the Kingdom of Saudi Arabia has experienced rapid population growth and urban expansions [1], [2]. Rapid population growth has affected the city’s environment, and the socio-economic conditions of its residents while massive construction of urban residential and business infrastructures and transportation networks has resulted in rising air pollution, increased frequency of flooding of the city, and rising of the land surface temperature (hereafter LST), [3], [4], [5]. In a desert environment, rising LST and the formation of urban heat island will have tremendous impacts on the health conditions especially in the case of the children, elderly, and the poor residents of the city [6]. The rapid development of geospatial technologies since the 1990s has allowed researchers to examine the changes and effects of urban expansion on LST in cities around the world [7], [8], [9], [10], [11]. LST data derived from remote sensing imageries have achieved better accuracy than those collected from ground-based weather stations [12], [13]. Yuan and Bauer (2007) examined the effect of the impervious surfaces on the seasonal variation of LST for the City of Twin Cities, Minnesota in 2002 [14]. Using Landsat TM and ASTER data, Liu and Zhang (2011) examined the influences of LST on the formation of urban heat islands for the city of Hong Kong [15]. Wang et al. (2018) used Landsat TM and ETM+ data to understand the impacts of urban expansion on LST in Nanjing City for the period between 1985 and 2009 [16]. For the city of Aksu (China), various landscape metrics were used to The Saudi capital city of Riyadh has experienced rapid population growth and urban expansion over the past 4 decades. One major consequence of such growth is the rising of the city’s land surface temperature (LST). This study used Landsat 7 ETM+ sensor data to map the distribution of Riyadh’s LST and then examined and modelled the impacts of five contributing factors known to increase urban LST. The contributing factors are size/area and population density of each neighbourhood, along with amounts of impervious surfaces, vegetations, and soil/sand measured through remote sensing indices NDBI, NDVI, and NDBsI. The data were analyzed using Pearson’s Product Moment Correlation values, Path Analysis, and Multiple Regression analysis. The result shows that neighbourhood population densities and NDBsI index have strong positive correlations (r= 0.68 and r= 0.60) with LST. Neighbourhood area showed significant but low positive correlation (r= 0.33) and the NDBI and NDVI indices showed strong negative correlations (r= -0.55 and r= -0.64) with the LST. The multiple regression model explained about 77% of the total variation in the LST. The model can be used to predict and simulate future LST distribution for Riyadh as well as other cities in the Kingdom and the region. Centre for Research on Settlements and Urbanism