Merging framework for estimating daily surface air temperature by integrating observations from multiple...

Area of Science:

• Earth and Environmental Sciences
• Climate Science
• Remote Sensing

Background:

• Estimating spatially continuous surface air temperature (SAT) is crucial for climate and environmental research.
• Previous SAT estimation methods primarily relied on Land Surface Temperature (LST) from NASA's MODIS instrument, limiting broader satellite data utilization.
• The potential of thermal infrared observations from meteorological satellites for SAT estimation has been largely overlooked.

Purpose of the Study:

• To develop and validate a novel merging framework for estimating daily mean SAT.
• To integrate LST data from multiple polar-orbiting satellites, including those from EUMETSAT (EPS) and NOAA (JPSS), alongside NASA's missions.
• To assess the framework's impact on spatial coverage and accuracy of daily SAT estimates.

Main Methods:

• Proposed a merging framework integrating LST datasets from five polar-orbiting satellites: Metop-B (EPS), SNPP and JPSS-1 (JPSS), and Terra and Aqua (EOS).
• Developed 10 estimation models utilizing LST from these diverse satellite sources.
• Generated daily merged SAT by averaging estimates from the integrated models.

Main Results:

• The framework achieved a 39% overall increase in spatial coverage for daily SAT estimates in cloud-free areas compared to individual satellite LST datasets.
• Achieved nearly universal daily coverage (average 91%, >75%) for merged SAT.
• Demonstrated 37%-51% greater coverage compared to SAT estimated solely from MODIS LST.
• Estimation models showed comparable predictive performance with RMSEs of 1.7-1.9 K (sample-based) and 1.9-2.2 K (site-based).

Conclusions:

• The developed merging framework significantly enhances the spatial coverage of daily SAT estimates by integrating multi-satellite LST data.
• This approach effectively utilizes thermal infrared observations from meteorological satellites, expanding beyond previous MODIS-based limitations.
• The framework provides a more comprehensive and accurate dataset for climate and environmental studies.