Synthesizing environmental, social, and urban density metrics to predict urban heat island dynamics using remote sensing and support vector regression

Kusumadewi, Tarranita ORCID: https://orcid.org/0000-0001-8290-2451, Surjono, Surjono ORCID: https://orcid.org/0000-0002-3024-5397, Leksono, Amin Setyo ORCID: https://orcid.org/0000-0001-5002-0569, Rohma, Salma Ainur, Husna, Sofia Amalia, Karami, Ahmad Fahmi and Arif, Yunifa Miftachul ORCID: https://orcid.org/0000-0002-2183-0762 (2025) Synthesizing environmental, social, and urban density metrics to predict urban heat island dynamics using remote sensing and support vector regression. Engineering, Technology and Applied Science Research, 15 (3). pp. 23141-23148. ISSN 1792-8036

Preview

Text 23836.pdf - Published Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Complex interactions between environmental, social, and urban density factors drive Urban Heat Island (UHI) dynamics. The current study synthesizes multidimensional metrics, including normalized difference indices, such as vegetation (NDVI), water (NDWI), moisture (NDMI), and social parameters (population density), such as urban morphology metrics (UDI) and built-up index (NDBI) to predict UHI intensity using remote sensing data and Support Vector Regression (SVR). Landsat-8, Sentinel-2A, and NASA-SRTM data from 2014 to 2023 were used to analyze Land Surface Temperature (LST) trends and spatially identify UHI hotspots in Malang City, Indonesia. The SVR model demonstrated robust performance, achieving an R² of 0.78, an RMSE of 0.23, and a MAPE of 0.46%. Results indicate that increasing urban density (UDI and NDBI) and population density significantly amplify LST, while higher NDVI values mitigate UHI effects. Temporal and spatial analyses reveal a steady expansion of UHI hotspots from central districts (e.g. Klojen) to peripheral areas (e.g. Sukun and Kedungkandang), driven by vegetation loss and urban sprawl. These findings underscore the potential of synthesizing multidimensional metrics for UHI prediction and highlight the value of integrating remote sensing data with machine learning models. The study provides actionable insights for urban planners to design targeted interventions, such as urban greening and density management, to mitigate UHI effects and enhance urban sustainability.

Item Type:	Journal Article
Keywords:	UHI; normalized difference indices; environmental indices; multidimensional metrics; remote sensing; SVR; land surface temperature; urban density
Subjects:	05 ENVIRONMENTAL SCIENCES > 0502 Environmental Science and management > 050205 Environmental Management 12 BUILT ENVIRONMENT AND DESIGN > 1205 Urban and Regional Planning > 120504 Land Use and Environmental Planning 12 BUILT ENVIRONMENT AND DESIGN > 1205 Urban and Regional Planning > 120507 Urban Analysis and Development
Divisions:	Faculty of Technology > Department of Architecture
Depositing User:	Tarranita Kusumadewi
Date Deposited:	12 Jun 2025 13:58

Downloads

Origin of downloads

Actions (login required)

View Item