All-Day Cloud Classification via a Random Forest Algorithm Based on Satellite Data from CloudSat and Himawari-8

It remains challenging to accurately classify complicated clouds owing the various types of and their distribution on multiple layers. In this paper, multi-band radiation information from geostationary satellite Himawari-8 cloud classification product polar orbit CloudSat June September 2018 are investigated. Based sample sets matched by two data, a random forest (RF) algorithm was applied train model, retrieval method developed for classification. With use method, were inverted classified as clear sky, low clouds, middle thin cirrus, thick multi-layer deep convection (cumulonimbus) clouds. The results indicate that average accuracy all during day is 88.4%, misclassifications mainly occur between cirrus cumulonimbus 79.1% at night, with more occurring Moreover, Typhoon Muifa 2022 selected case, type (CLT) an FY-4A used examine method. system Muifa, area using corresponded well mesoscale convective (MCS). Compared CLT product, classifications ice-type (thick cirrus) effective, location, shape size these varieties similar.