Dust Machine Learning Probability and Assessment

- NASA SPoRT introduced the "Dust RGB" via NASA satellites to demonstrate GOES-R ABI capabilities and then evaluated the impact in operations (Fuell et al. 2016)

- The Dust RGB allows for continued dust detection at night, but the cooling ground surface limits the effectiveness as night progresses.

- SPoRT has developed a 'Machine Learning' (ML) model using a physically-based approach which can correctly label 85% of dust pixels and 99% of no-dust pixels