Daytime Land Surface Temperature Extraction from MODIS Thermal Infrared Data under Cirrus Clouds

Area of Science:

• Earth Science
• Atmospheric Science
• Remote Sensing

Background:

• Cirrus clouds significantly impact land surface temperature (LST) retrieval accuracy.
• The generalized split-window (GSW) algorithm is widely used but sensitive to cirrus cloud contamination.
• Existing methods struggle to accurately retrieve LST under cirrus cloud conditions.

Purpose of the Study:

• To develop and validate an extended GSW algorithm for accurate LST retrieval under cirrus cloud conditions.
• To quantify the LST retrieval error reduction achieved by the proposed algorithm.
• To assess the algorithm's performance using simulated and real-world surface temperature data.

Main Methods:

• A correction term based on cirrus optical depth (COD) was integrated into the GSW algorithm.
• COD was determined using a lookup table of cirrus bidirectional reflectance at 0.55 μm.
• The algorithm's slope parameter was modeled using MODIS brightness temperatures and split-window channel emissivities.

Main Results:

• The extended GSW algorithm reduced maximum LST retrieval errors from 11.0 K to 2.2 K in simulated nadir views.
• Sensitivity analysis indicated total errors remained below 2.5 K under various uncertainties.
• Validation with Great Lakes buoy data showed at least a 1.5 K improvement in retrieval accuracy for cirrus skies.

Conclusions:

• The proposed extension significantly enhances the GSW algorithm's capability for LST retrieval in the presence of cirrus clouds.
• The method provides a robust approach to mitigate cirrus-induced errors in LST estimations.
• This advancement is crucial for accurate climate monitoring and surface energy balance studies.