Modeling the impact of land use/land cover (LULC) factors on diurnal and nocturnal Urban Heat Island (UHI) intensities using spatial regression models

Abstract

This paper models the impact of different land use/land cover (LULC) factors on diurnal and nocturnal Urban Heat Island (UHI) intensities using spatial regression models. This research first developed an extensive dataset of 17 LULC factors as the independent variables and the diurnal and nocturnal UHI intensities as the dependent variables. Second, ordinary least-squared (OLS) regression was implemented to check for spatial autocorrelation. The results showed that spatial autocorrelation exists for diurnal and nocturnal UHI intensities and thus spatial regression models are needed. Hence, the following spatial regression models were implemented: the spatial error model (SEM) and the spatial lag model (SLM). The findings showed that the SEM is the best in modeling the diurnal UHI intensity while the SLM is the best in modeling the nocturnal UHI intensity. The results also showed that the diurnal UHI differs from the nocturnal UHI by its intensity, geographical distribution, and the contributing LULC factors or causes. The outcomes of this research can be used to inform different UHI-related decisions such as identifying and modeling the impacts of the most significant LULC factors on the diurnal and nocturnal UHI intensities as well as determining the most suitable geographical locations for implementing UHI mitigation initiatives.