Exploring the Influence of Various Land Use Land Cover on Land Surface Temperature of Coastal Tourism Areas in Bali Using Landsat 9


  • I Wayan Diara Department of Environmental Science, Udayana University, Indonesia
  • Ketut Dharma Susila Faculty of Agriculture, Udayana University, Indonesia
  • Wiyanti Wiyanti Faculty of Agriculture, Udayana University, Indonesia
  • I Nyoman Sunarta Faculty of Tourism, Udayana University, Indonesia
  • Tati Budi Kusmiyarti Spatial Data Infrastructure Development Center, Udayana University, Indonesia
  • Moh Saifulloh Spatial Data Infrastructure Development Center, Udayana University, Indonesia; Faculty of Marine Science and Fisheries, Udayana University, Indonesia https://orcid.org/0000-0001-5762-5755




land use land cover (LULC), land surface temperature (LST), remote sensing, environmental, urban tourism


The Bali Tourism area represents complex environments where human activities intersect with natural landscapes, resulting in diverse land use land cover (LULC) patterns. However, understanding the dynamics of LULC in these areas and its interaction with land surface temperature (LST) remains a challenge. This study addresses this gap by investigating LULC mapping in urban tourist destinations and its influence on LST variations. The research problem focuses on exploring the relationship between various land cover types and LST variations. The main objective is to assess the interaction of LULC variations with LST in urban tourist environments. To achieve this goal, an integrated approach combining remote sensing techniques and machine learning will be employed. LULC mapping will utilize support vector machine (SVM) techniques with datasets sourced from multi-channel data, and spectral indices such as enhanced built-up and bareness index (EBBI) and normalized differences vegetation index (NDVI) derived from Landsat 9. The findings present a vivid overview of the research area, where built-up areas dominate, and spanning 108.61 km². Other land cover classifications include rice fields/grasslands, plantation/perennial plants, barren land, mangrove forests, shrublands, and water bodies, accurately mapped with high precision (overall accuracy = 88.52% and Kappa = 81%). Maximum LST values peak in built-up and barren areas, reaching 29.89°C and 29.28°C, respectively, while other land cover types exhibit comparatively lower values. Our analysis of the spectral index used in LULC classification uncovers a positive correlation with EBBI (R2 = 37.78%) and a negative correlation with NDVI (R2 = 10.69%, based on a substantial sample size of 67 869 pixels. We strongly urge future researchers to leverage high-resolution data for localized urban studies and stress the critical importance of enforcing stringent spatial planning regulations to safeguard green spaces, thus ensuring ecological equilibrium for future generations.