A 20-year study of the surface urban heat island effect in Phoenix, AZ using a spatiotemporal modeling framework

Keywords: Urban heat island, spatiotemporal modeling, remote sensing
Session Type: Virtual Paper Abstract
Day: Saturday
Session Start / End Time: 2/26/2022 05:20 PM (Eastern Time (US & Canada)) - 2/26/2022 06:40 PM (Eastern Time (US & Canada))
Room: Virtual 1

Abstract

Previous studies have focused on either spatial or temporal variations of the surface urban heat island effect for many cities in different climate zones across the globe. A real coupling of the spatial and temporal domain is rarely implemented in part due to the lack of appropriate tools to work with. This study presents a first investigation of the long-term spatiotemporal changes in the surface UHI in relation to the spatial and temporal variations in the land cover patterns over Phoenix, AZ, a US city well-known for its UHI effect. This study takes advantage of a newly developed spatiotemporal weighted regression framework (STWR), the performance of which was tested against conventional models including the OLS and GWR through a series of simulation and empirical studies. Results show that the STWR better captures the spatiotemporal relationship of the surface temperature with the land cover changes in Phoenix over the last 20 years. It achieves better parameter estimation and prediction with significantly lower errors than the OLS and GWR. Therefore, for analysis involving spatiotemporal patterns, the STWR is an effective alternative to conventional regression models because all available spatial and temporal information prior to the time of estimate will be used collectively to make sound statistical estimates and predictions.