Du et al. (2026) A scene-mapped framework for deriving clear-sky surface longwave downward radiation from remotely sensed measurements

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-10-02
• Authors: Yihan Du, Tianxing Wang, Yi Zheng, Dahui Li, Yuyang Xian, Feiyao Gao, Stefan A. Buehler
• DOI: 10.1016/j.rse.2026.115707

Research Groups

• Guangdong Provincial Key Laboratory of Geodynamics and Geohazards, School of Earth Sciences and Engineering, Sun Yat-Sen University, Zhuhai, China
• School of Geospatial Engineering and Science, Sun Yat-Sen University & Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai, China
• Atmospheric Science, Department of Earth System Science, Earth and Society Research Hub (ESRAH), University of Hamburg, Hamburg, Germany
• Key Laboratory of Comprehensive Observation of Polar Environment (Sun Yat-sen University), Ministry of Education, Zhuhai, China
• Key Laboratory of Natural Resources Monitoring in Tropical and Subtropical Area of South China, Ministry of Natural Resources, China

Short Summary

This study develops a novel scene-mapped lookup-table (LUT) framework to accurately estimate clear-sky surface longwave downward radiation (LWDR) from remotely sensed measurements, demonstrating improved accuracy and transferability compared to conventional methods.

Objective

• To develop a scene-mapped lookup-table (LUT) framework that operates in brightness-temperature (BT) space, using surface temperature (ST) and total column water vapor (TCWV) as physical descriptors, to derive clear-sky surface longwave downward radiation (LWDR) from remotely sensed measurements.
• To overcome regional biases and limited transferability of existing LWDR parameterizations and conventional LUT searches.

Study Configuration

• Spatial Scale: Global/Multi-regional (designed for broad applicability across diverse surface types and climatic regions).
• Temporal Scale: Designed for long-term data production and retrievals, though specific validation period is not detailed in the provided text.

Methodology and Data

• Models used: Scene-mapped LUT framework (S-Poly: polynomial regression, S-LUT: sub-LUT distance search), Radiative Transfer Models (RTMs) for LUT precomputation.
• Data sources: Remotely sensed measurements (MODIS inputs), multi-network ground measurements (for validation), ERA5 reanalysis (for mixed configuration).

Main Results

• The S-LUT method achieved a Root Mean Square Error (RMSE) of 26.36 W m⁻² and a bias of -0.86 W m⁻².
• The S-Poly method achieved an RMSE of 26.89 W m⁻² and a bias of -2.75 W m⁻².
• Both methods showed lower RMSEs than the global-LUT benchmark (29.26 W m⁻²) and the general clear-sky parameterization benchmark (30.03 W m⁻²).
• This represents RMSE reductions of 12.2% (S-LUT) and 10.5% (S-Poly) relative to the general clear-sky parameterization benchmark.
• Both methods maintained stable accuracy across diverse surface types and climatic regions.
• S-LUT demonstrated greater consistency and achieved an RMSE of 22.48 W m⁻² under a mixed MODIS–ERA5 configuration.

Contributions

• Introduction of a novel scene-mapped LUT framework operating in brightness-temperature space, using surface temperature and total column water vapor to define scene-specific subsets.
• Development of two robust implementations (S-Poly and S-LUT) that significantly improve accuracy and transferability of clear-sky LWDR estimation compared to conventional methods and general parameterizations.
• Provides a flexible methodological concept for adapting radiation retrieval to different sensors and exploring future extensions to all-sky conditions.

Funding

• Not specified in the provided text.

Citation

@article{Du2026scenemapped,
  author = {Du, Yihan and Wang, Tianxing and Zheng, Yi and Li, Dahui and Xian, Yuyang and Gao, Feiyao and Buehler, Stefan A.},
  title = {A scene-mapped framework for deriving clear-sky surface longwave downward radiation from remotely sensed measurements},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115707},
  url = {https://doi.org/10.1016/j.rse.2026.115707}
}