Khairil Mursalin, Parwati Sofan, Ike Sari Astuti
Intense human activity in urban areas plays an important role in increasing temperatures, which is reflected in the high surface temperature over built-up areas in cities. The aim of this study is to analyze the Urban Heat Island (UHI) phenomenon in Jakarta and Surabaya, using Landsat data associated with spectral indices and Land Use Land Cover (LULC), specifically for built-up areas. We used spectral indices for detecting built-up areas such as Index-based Built-Up Index (IBI), New Built-up Index (NBI), Normalized Difference Built-up Index (NDBI), and Urban Index (UI) to analyze UHI over a 10-year period from 2012 to 2022. Machine learning algorithms were employed to map the LULC, achieving an overall accuracy of 84% with the Support Vector Machine algorithm and 83% with the Random Forest. The analysis revealed that IBI has the highest correlation (0.69-0.84) with LST, compared to other built-up indices in Jakarta and Surabaya. The UHI classification based on LULC showed that residential areas had the highest average temperature compared to bare land, industrial areas, vegetation, and water bodies, with a temperature of 43.3°C for Jakarta and 43.°C for Surabaya, due to the high density of residential areas and buildings in the city. The spectral index correlation results show that IBI has the highest value, 0.69 for Jakarta and 0.85 for Surabaya. Further research needs to explore high spatial resolution data to distinguish detailed built-up objects in the city. © 2024 SPIE.
Department of Geography, State University of Malang, Malang, Indonesia; Research Center for Remote Sensing, National Research and Innovation Agency (BRIN), Cibinong, Indonesia
