Bobáľová et al. (2025) Improving Landsat land surface temperature estimation in Google Earth Engine using NDVI-based emissivity

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-11-24
• Authors: Hana Bobáľová, Šimon Opravil
• DOI: 10.1016/j.asr.2025.11.085

Research Groups

• Comenius University in Bratislava, Faculty of Natural Sciences, Department of Physical Geography and Geoinformatics, Bratislava, Slovakia
• Slovak Academy of Sciences, Institute of Geography, Bratislava, Slovakia

Short Summary

This study developed and validated an improved Google Earth Engine (GEE) approach for Landsat Land Surface Temperature (LST) estimation, utilizing NDVI-based emissivity calculations combined with statistical mono-window and radiative transfer equation methods. The new method demonstrated higher accuracy and precision compared to the standard Landsat ST product and approaches relying on ASTER Global Emissivity Dataset (GED).

Objective

• To develop and validate an improved method for estimating Landsat Land Surface Temperature (LST) in Google Earth Engine (GEE) by integrating various NDVI-based emissivity calculations with statistical mono-window and radiative transfer equation methods, aiming to overcome limitations of existing approaches like ASTER GED.

Study Configuration

• Spatial Scale: Landsat satellite resolution (typically 30 meters for thermal bands), applied to create seamless LST products.
• Temporal Scale: Long-term continuity, consistent across all Landsat missions.

Methodology and Data

• Models used:
  • LST calculation: Statistical mono-window method, Radiative Transfer Equation (RTE) method.
  • Emissivity calculation: Various NDVI-based methods.
• Data sources:
  • Satellite: Landsat satellites (thermal infrared data), ASTER Global Emissivity Dataset (ASTER GED) for comparison.
  • Observation: In situ measurements from the SURFRAD network for validation.
  • Platform: Google Earth Engine (GEE).

Main Results

• The developed GEE approach, combining NDVI-based emissivity with the statistical mono-window method, produces seamless LST products without unnatural block artifacts, accurately reflecting current land cover and vegetation conditions.
• Validation against SURFRAD network in situ measurements revealed that the statistical mono-window method was more accurate than the standard Landsat ST product and radiative transfer equation methods, regardless of the emissivity data source.
• The NDVI-based emissivity combined with the statistical mono-window method yielded higher LST precision compared to approaches using ASTER GED emissivity.
• These improved results were consistently observed across all Landsat missions.
• The lowest LST accuracy was found on mixed surfaces, while the highest accuracy was achieved on bare soil.
• Overestimation of satellite LST measurements at high temperatures was apparent on mixed and vegetated surfaces, but this overestimation was more pronounced in the Landsat ST product and other radiative transfer equation methods.

Contributions

• Development of a novel and publicly available Google Earth Engine code for improved Landsat LST estimation using NDVI-based emissivity, addressing limitations of existing methods and ASTER GED.
• Demonstration of enhanced accuracy and precision of the proposed method compared to the standard Landsat ST product and other radiative transfer equation methods.
• Provision of a robust and consistent LST estimation approach applicable across all Landsat missions.
• Identification of LST accuracy variations based on surface type, highlighting challenges in mixed and vegetated areas.

Funding

Not specified in the provided text.

Citation

@article{Bobáľová2025Improving,
  author = {Bobáľová, Hana and Opravil, Šimon},
  title = {Improving Landsat land surface temperature estimation in Google Earth Engine using NDVI-based emissivity},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.11.085},
  url = {https://doi.org/10.1016/j.asr.2025.11.085}
}