Machine learning downscaling of satellite derived land surface temperature in the historic urban heritage area of São João del-Rei Brazil
Open Access
- 7 July 2026
- journal article
- Published by Springer Nature in Discover Cities
- Vol. 3 (1)
Abstract
This study analyzes seasonal land surface temperature (LST) patterns in the historic urban heritage area of São João del-Rei, Brazil, using satellite remote sensing and machine-learning-based spatial downscaling. Landsat 8/9 thermal imagery and Sentinel-2-derived spectral indices were processed for January (wet summer) and July (dry winter) over the 2019–2025 period. Landsat-derived LST composites at 30 m were spatially refined to 10 m using a CatBoost Regressor trained with Sentinel-2 spectral predictors, including NDVI, NDWI, NDBI, albedo, and dSAVI. The resulting maps should be interpreted as model-derived downscaled LST surfaces, not as direct street-scale thermal observations. The results indicate clear seasonal differences in the spectral controls associated with LST variability. In January, vegetation- and moisture-related indices showed greater relative importance, suggesting a greater role of vegetation- and moisture-related surface conditions under wet-summer conditions. In July, the built-up index (NDBI) showed greater relevance in the predictor-removal analysis, indicating that impervious and built surfaces became more important for explaining winter LST contrasts within the adopted model. Spatial patterns in the model-derived downscaled LST maps suggest lower estimated surface temperatures near vegetated areas, water bodies, and some compact street segments, while more open and impervious spaces showed higher estimated surface temperatures. These patterns are interpreted in relation to the known morphology and material characteristics of the protected historic fabric, but they do not constitute direct evidence of thermal inertia, urban canyon effects, or pedestrian thermal comfort. The study provides an exploratory and reproducible framework for assessing seasonal surface thermal patterns in historic urban centers and discusses their potential relevance for heritage-sensitive climate adaptation planning.Keywords
Funding Information
- Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
This publication has 39 references indexed in Scilit:
- Using Cultural Heritage in Climate Adaptation: Fields of Application and FunctionsWires Climate Change, 2025
- Impacts of urban and industrial pollution on functional traits of benthic macroinvertebrates: Are some traits advantageous for survival?Science of The Total Environment, 2022
- Downscaling land surface temperatures at regional scales with random forest regressionRemote Sensing of Environment, 2016
- Generating daily land surface temperature at Landsat resolution by fusing Landsat and MODIS dataRemote Sensing of Environment, 2014
- Remote sensing land surface temperature for meteorology and climatology: a reviewMeteorological Applications, 2011
- Urban greening to cool towns and cities: A systematic review of the empirical evidenceLandscape and Urban Planning, 2010
- Thermal infrared remote sensing for urban climate and environmental studies: Methods, applications, and trendsISPRS Journal of Photogrammetry and Remote Sensing, 2009
- Estimation of land surface temperature–vegetation abundance relationship for urban heat island studiesRemote Sensing of Environment, 2004
- A soil-adjusted vegetation index (SAVI)Remote Sensing of Environment, 1988
- Street design and urban canopy layer climateEnergy and Buildings, 1988