EUVIMEDEUVIMED entdecken
Crossref

Performance evaluation of Landsat TIRS/2 and ECOSTRESS Land Surface Temperature Products over Mountainous Snow-Covered Areas

Sara Arioli, Picard Picard, Simon Gascoin, Laurent Arnaud

Recent Advances in Remote Sensing · 2026

DOI: 10.62880/rars26006 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Recent Advances in Remote Sensing, 2026. Verfasst von Sara Arioli und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The limited evaluation of remote sensing Thermal InfraRed (TIR) products over snow-covered surfaces compromises their exploitation in cryospheric sciences. This study assesses for the first time the accuracy of high-resolution Land Surface Temperature (LST) products from Landsat 8, Landsat 9, and ECOSTRESS in a snow-covered Alpine site, using 2 m resolution maps of the snow surface temperature acquired in-situ over two winter seasons. The results show a general warm bias across all LST products compared to in-situ measurements, with varying magnitudes. Landsat Collection 2 LST products, with a Mean Absolute Error (MAE) of +1.06 K, are within the expected 2% accuracy, in agreement with previous studies on warmer surfaces. The MAE is slightly warmer for ECOSTRESS Version 1 LST, with +1.24 K, exceeding its declared accuracy goals (0.3 K precision, 1.0 K accuracy), potentially due to calibration issues previously found for temperatures below 295 K. This performance remains satisfactory considering that ECOSTRESS is designed to operate over warm surfaces. Still, a possible source of error was identified in the low emissivity values used for the LST retrieval, likely due to limitations in the representation of snow emissivity in the ECOSTRESS spectral library. The NDSI-based emissivity estimation approach used by Landsat yields better results than ECOSTRESS, and could be implemented by future TIR missions such as TRISHNA and LSTM, that will also have bands in the optical domain. More problematic, ECOSTRESS Version 2 LST shows a significantly higher bias of +3.91 K, indicating a performance degradation. The likely cause of this issue is the correction of a cold bias observed in ECOSTRESS Version 1 LST for temperatures above 270 K. As these changes compromise the performance of ECOSTRESS LST over snow-covered areas, its correction in upcoming ECOSTRESS LST collections would contribute to extend the possible exploitation of ECOSTRESS LST over snow and other cold surfaces of several years.

Zitieren

APA

Arioli, S., Picard, P., Gascoin, S., & Arnaud, L. (2026). Performance evaluation of Landsat TIRS/2 and ECOSTRESS Land Surface Temperature Products over Mountainous Snow-Covered Areas. Recent Advances in Remote Sensing. https://doi.org/10.62880/rars26006

Vancouver

Arioli S, Picard P, Gascoin S, Arnaud L. Performance evaluation of Landsat TIRS/2 and ECOSTRESS Land Surface Temperature Products over Mountainous Snow-Covered Areas. Recent Advances in Remote Sensing. 2026. doi:10.62880/rars26006

BibTeX
@article{Arioli2026,
  author = {Arioli, Sara and Picard, Picard and Gascoin, Simon and Arnaud, Laurent},
  title = {Performance evaluation of Landsat TIRS/2 and ECOSTRESS Land Surface Temperature Products over Mountainous Snow-Covered Areas},
  journal = {Recent Advances in Remote Sensing},
  year = {2026},
  doi = {10.62880/rars26006}
}
Kostenlos starten →