Estimation of cloud cover on Cuba

Given the need to find a new way of assessing both numerical and official forecasts of clouds at the Institute of Meteorology (Insmet), the general objective of this work is set to design and validate a methodology to estimate cloud cover on Cuba, based on the relationships between the presence or absence of clouds and gray levels on GOES-E’s satellite images. The specific objectives are to determine the thresholds for the presence of either clouds or surface from images of channel 1 (0.6 μm), 2 (3.9 μm) and 4 (10.7 μm), to obtain a computational tool for the cloud cover determination over Cuba and validate the results from the comparison between visual classifications and those based on non-scalable categorical variables. GOES-13 images (4km resolution) obtained from the Insmet receiving station 15 minutes before synoptic times are used. The algorithm designed is reliable all year but its level of confidence decreases in very cold nights with clear skies of winter.