The urban heat Island effect: A review on predictive approaches using artificial intelligence models

Abstract

With the global population now exceeding 8 billion and 4.5 billion of whom residing in urban areas, rapid urbanization has contributed to a range of environmental and ecological challenges, notably the Urban Heat Island (UHI) effect. According to statistical data, the ten hottest years on record occurred between 2013 and 2022, underscoring the urgency of addressing urban heat issues. This study provides a comprehensive review of research on the UHI effect, analysing and classifying studies that utilize a variety of input–output datasets. It also examines predictive methods used to estimate UHI intensity, categorizing them into conventional machine learning (ML) algorithms, deep learning (DL) models, and hybrid approaches. While conventional ML algorithms remain widely used, DL and hybrid models have shown superior performance in predictive accuracy. This review aims to enhance understanding of recent advancements in UHI prediction techniques, identify limitations in current methodologies, and propose directions for future research.