Improvements in MAIAC EPIC cloud detection for aerosol retrievals and atmospheric correction

Abstract:

Quality of aerosol retrievals and atmospheric correction depends strongly on accuracy of the cloud mask (CM) algorithm. Based on the heritage of developed MAIAC CM algorithm for MODIS, VIIRS, AHI, we improved EPIC CM algorithm of MAIAC. EPIC onboard DSCOVR observe Earth 1 million miles away from Earth toward the sun with 88 km of spatial resolution near equator from 13 (in winter) to 22 (in summer) per day. To fit into unique geometry and spatial resolution of EPIC, we refined bright cloud test for high airmass region, and refined spatial variability test based on difference between latest image and reference clear-skies image of the surface constructed by time-series analysis of pixel. Further improvements of EPIC CM algorithm include algorithm dealing with sub-pixel cloud contamination, algorithm using O2-A band over ocean and land, algorithm based on the spectral properties of aerosol and surface over land at UV-visible channel. Especially, the preliminary results showed that additional usage of O2-A band and UV-visible channel of EPIC for CM provide better quality of aerosol retrievals and atmospheric correction results. The study will also provide the time-series analysis of atmospheric correction results before and after applying the improved EPIC CM algorithm.