Systematic Errors that are Due to the Monochromatic-Equivalent Radiative Transfer Approximation in Thermal Emission...

Area of Science:

• Atmospheric science
• Radiative transfer modeling
• Satellite remote sensing

Background:

• Data assimilation models rely on accurate radiative transfer models (RTMs) for simulating satellite radiances.
• Fast, parameterized RTMs are used for efficiency but often simplify spectral integration, introducing potential errors.
• The monochromatic-equivalent approach in fast RTMs replaces spectral integration with spectrally averaged values, an approximation assumed to be negligible.

Purpose of the Study:

• To quantify the error introduced by the monochromatic-equivalent approximation in fast RTMs.
• To demonstrate an improved fast RTM using Planck-weighted mean transmittances to mitigate these errors.
• To investigate the impact on satellite data assimilation, focusing on a specific channel of the High-Resolution Infrared Radiation Sounder (HIRS).

Main Methods:

• Analyzed the error magnitude arising from the monochromatic-equivalent approach in parameterized RTMs.
• Developed and tested a fast RTM incorporating Planck-weighted mean transmittances.
• Focused on Channel 12 of the NOAA-14 HIRS instrument, known for exhibiting significant errors.

Main Results:

• The monochromatic-equivalent approach introduces a systematic, height-dependent bias in data assimilation.
• The proposed fast RTM with Planck-weighted mean transmittances significantly reduces or eliminates this bias at the source.
• Channel 12 of the NOAA-14 HIRS instrument shows the largest error due to this approximation.

Conclusions:

• The assumption of negligible error in fast RTMs is invalid and leads to systematic bias in data assimilation.
• Correcting the source of the error within the RTM is more effective than post-assimilation bias correction.
• Implementing Planck-weighted mean transmittances offers a viable solution for improving the accuracy of fast RTMs in satellite data assimilation.