Thermal Analysis for Resilient Transport Infrastructure: A Downscaling Approach Using Satellite EO Data and UAV

In the context of the ongoing climate crisis, the Urban Heat Island (UHI) phenomenon represents one of the most critical challenges for the sustainability of the built environment. Critical transport infrastructures, such as highways, railways, and airports, play a pivotal role in this process. Due to their extension and to the thermophysical properties of the construction materials employed (e.g., asphalt, concrete), they act as significant thermal collectors. On the other hand, in urban areas transport infrastructures negatively affect the local microclimate, thereby reducing the resilience of urban areas.

 

In this context, satellite Earth Observation (EO) has emerged as a promising tool for monitoring temperature variations [1]. However, temperature measurement in the context of transport infrastructure remains challenging due to the limitations imposed by the spatial resolution of satellite sensors. The primary issue concerns the limited spatial resolution of currently available thermal satellite sensors (with a native resolution of 100 m, resampled to 30 m), such as the TIRS instrument on Landsat 8/9 [2]. While these sensors provide accurate radiometric data, they lack the geometric detail required to analyze specific transportation assets. To overcome this limitation, this research proposes an innovative methodology based on a multi‑scale thermal downscaling procedure, implemented within the Google Earth Engine platform and applied on a real‑scale parking area scenario.

 

The adopted methodology relies on the synergistic integration of multi‑scale satellite data, following the approach implemented by [3]. This study exploits the well‑established relationship between Land Surface Temperature (LST) and land‑cover metrics, including the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Built‑up Index (NDBI), and the Normalized Difference Water Index (NDWI). A multiple linear regression model was first defined using spectral indices and LST derived from Landsat 8 data; this model was then applied using the corresponding spectral indices extracted from Sentinel‑2 imagery [4] to predict LST at a spatial resolution of 10 m. Subsequently, a second downscaling step was performed by applying a multi‑regression approach based on RGB bands from Sentinel‑2 and UAV imagery, enabling the estimation of surface temperature at sub‑meter resolution. Through this two‑stage procedure, the resolution of LST maps was significantly enhanced, achieving a resolution commensurate with the scale of transport infrastructure. This approach was applied to a parking area in Rome, demonstrating the potential of a sequential thermal downscaling procedure that progressively refines satellite‑derived temperatures using higher‑resolution Sentinel‑2 data and UAV imagery. The results confirm that thermal analysis based on satellite EO data and downscaling techniques is a promising, effective, and cost‑efficient method for assessing infrastructure resilience.

 

References

[1] Almeida CR, et al,. Study of the Urban Heat Island (UHI) Using Remote Sensing Data/Techniques: A Systematic Review. Environments 2021;8(10)

[2] Landsat Official Website. Accessed 01-2025. https://landsat.gsfc.nasa.gov/satellites/landsat-9/landsat-9-instruments/landsat-9-spectral-specifications/

[3] Onačillová, K. et al.. Combining Landsat 8 and Sentinel-2 Data in Google Earth Engine to Derive Higher Resolution Land Surface Temperature Maps in Urban Environment. Remote Sens. 2022, 14, 4076.

[4] European Space Agency - Sentinel-2 User Handbook, (2015)