Physically constrained spatiotemporal modeling: generating clear-sky constructions of land surface temperature from...

Area of Science:

• Earth and Space Sciences
• Geophysics
• Remote Sensing

Background:

• Satellite remote sensing provides crucial global atmospheric and geophysical data.
• Missing data in satellite observations is common due to cloud cover and infrequent satellite passes.
• Existing methods for infilling land surface temperature (LST) data often neglect the diurnal temperature cycle.

Purpose of the Study:

• To develop and demonstrate a novel method for reconstructing continuous spatial and temporal land surface temperature (LST) fields.
• To incorporate the physical constraints of the diurnal cycle into LST data infilling.
• To improve the understanding of spatiotemporal thermal patterns from satellite data.

Main Methods:

• Parameterizing the diurnal cycle into a functional form with spatiotemporal parameters.
• Utilizing multiresolution spatial basis functions to estimate parameters from sparse satellite observations.
• Reconstructing LST fields with continuous spatial and temporal distributions.

Main Results:

• Successfully reconstructed LST fields with continuous spatial and temporal distributions.
• Demonstrated the methodology's effectiveness using MODIS data from NASA's Aqua and Terra satellites.
• Provided a method to better inform scientists about spatiotemporal thermal patterns.

Conclusions:

• The developed method effectively infills missing LST data by accounting for the diurnal cycle.
• This approach enhances the utility of satellite remote sensing data for studying surface and climate variability.
• The methodology shows promise for application over complex geographical domains.