Towards Routine Radiance-Based Validation of VIIRS LST Using GDAS Profiles

Land Surface Temperature (LST) plays a pivotal role in various Earth science applications, including climate monitoring, urban heat island analysis, water management, and agricultural assessments. Accurate validation of LST is crucial to ensure the reliability of remotely data products. In this pursuit, radiance-based validation method has emerged as an important component. The advantage of such method lies in their ability to directly relate satellite measured radiance to atmospheric and surface temperature profiles, enabling more robust and comprehensive validation procedures.

This study aims to advance the validation of VIIRS (Visible Infrared Imaging Radiometer Suite) LST products through the implementation of a routine radiance-based validation approach. By leveraging GDAS (Global Data Assimilation System) profiles from cloud platforms, such as the Amazon Web Services (AWS) used in this study, we proposed a robust framework that integrates radiative transfer modeling, optimization methods, with atmospheric and surface temperature profiles to access the accuracy of VIIRS LST retrievals. This framework demonstrates the potential of using cloud-based data sources like AWS to enhance the validation process, offering a pathway towards improved LST retrievals and a deep understanding of Earth’s surface temperature dynamics.

The framework begins with the data matchup pairs obtained from temperature-based LST validation. Utilizing the time information, the GDAS time stamp is calculated and used as a query to automatically retrieve data from cloud. Subsequently, profile information is extracted from GDAS and the simulation cards are prepared for forward radiative transfer simulation using MODerate resolution atmospheric TRANsmission (MODTRAN). An optimization algorithm then adjusts the satellite LST value by optimizing a cost function defined using the VIIRS band M15/M16 brightness temperature difference between the satellite observations and simulations. The solution with the lowest cost is adopted as the satellite LST obtained from the radiance-based simulation. In this study, ground matchups from Surface Radiation Budget (SURFRAD), Baseline Surface Radiation Network (BSRN), Atmospheric Radiation Measurement (ARM), and National Data Buoy Center (NDBC) have been utilized in the r-based validation and the results have been compared with the t-based LST validation. The comparison results indicate a generally smaller standard deviation compared to the temperature based LST validation. Moreover, the initial experimental results demonstrate the successful routine execution of this framework for providing radiance-based validation results. Additional ground network will be incorporated in the future.