Determination of CERES TOA Fluxes Using Machine-Learning Algorithms

Objective
- It is observed that 5.6% of all CERES Terra/Aqua data contains missing cloud cover information or insufficient imager data for a reliable scene identification (some times it can reach up to 50% of data for a specific scene type)
- The unavailability of imager data lead to gaps in global radiation budget dataset.
- In this study, our objective is to develop a Machine learning methodology for the improved determination of CERES scene type and subsequent clear-sky TOA flux estimation using standalone CERES TOA radiance measurements (without any MODIS/Imager data).