Reply to Keith Hutchison's comment
Andrzej Kotarba · 2020
Thank you for the comment.Assigning 33% to 'probably clear' and 66% to 'probably cloud' would definitively change the final estimation of cloud amount.To get the exact figure all the data for this study would have to be reworked.However the existing results suggest that 66% is well justified (we found 66,6%), and 33% is only ∼5% more that 27.7% found by us.The difference remains for 'confident' classes, which are the most frequent and influence the final results significantly (on a global scale).