Investigating the relationship between the urban heat island effect and short-duration extreme rainfall in Kuala Lumpur

Tan, Yan Kai (2025) Investigating the relationship between the urban heat island effect and short-duration extreme rainfall in Kuala Lumpur. Final Year Project, UTAR.

PDF Download (4Mb)

Abstract

Urbanisation significantly alters land surface characteristics, leading to the intensification of the urban heat island (UHI) effect, which may influence the short-duration extreme rainfall. This study investigates the relationship between UHI intensity and short-duration extreme rainfall in Kuala Lumpur through an integrated remote sensing, machine learning and statistical approach. Landsat imagery from 2007, 2015 and 2023 was used to analyse spatiotemporal changes in land use and land cover (LULC) and to estimate land surface temperature (LST). LULC classification was performed using Support Vector Machine (SVM) and Random Forest (RF) algorithms, while LST was estimated using the Single Channel (SC) algorithm and surface urban heat island intensity (SUHII) was subsequently derived from the LST data. Hourly rainfall data exceeding the 99th percentile from 2007 to 2023 were used to assess spatiotemporal variation, diurnal distribution and trends. Statistical relationships between SUHII and hourly extreme rainfall were examined using the coefficient of determination (R²) and Kendall’s Tau (τ). Results show that SVM consistently outperformed RF in terms of overall accuracy and kappa coefficient across all study years. Built-up areas and SUHII both exhibited a net increase, particularly in northern Kuala Lumpur, likely due to intense urbanisation and industrial activities. The number of hourly extreme rainfall events also increased, especially during late afternoon and evening hours. However, the mean intensity of extreme rainfall events remained relatively stable. Correlation analysis identified moderate, statistically significant relationships between the annual SUHII and the annual total number of hourly extreme rainfall events at four of nine stations (R² = 0.2530 - 0.3088; τ = 0.3616 - 0.4593; p < 0.05). These findings suggest that urban-induced heating may contribute to enhanced localised convective rainfall. It is recommended that UHI mitigation measures, such as green infrastructure and climate-sensitive urban planning, be prioritised to manage future rainfall-related flood risks in urban environments. Keywords: Urban Heat Island; Land Surface Temperature; Land Use and Land Cover; Remote Sensing; Machine Learning; Short-Duration Extreme Rainfall; Rainfall Analysis; Relationship Subject Area: TA170-171 Environmental Engineering

Actions (login required)