Cloud Occurrence Frequency (COF) of different cloud types from three cloud classification products are evaluated in
this study over six sub-regions of China in 2012, including surface products, Feng-Yun-2E (FY-2E) and CloudSat satellite
products. The spatial distribution and seasonal variation of cloud types are compared on the macro-characteristics. The
main cloud types are cirrus (Ci), cirrostratus (Cs), stratocumulus (Sc), altocumulus (Ac), altostratus (As), cumulus (Cu)
and nimbostratus (Ns). The results of this study show that the surface observation in general overestimates a lot more than
FY-2E and CloudSat; FY-2E provides more accurate high cloud estimates; CloudSat provides better middle cloud type
detection; surface and CloudSat observation capture more consistent distribution of Sc, Ac, Cu and high cloud; all three
datasets provide similar evidence on seasonal cycle, especially on Cu, Cb, Ac, with the pattern of peak in summer, their
COF seasonal variation decrease. Seasonal variations of Sc, As, Ns COF especially from CloudSat are more reasonable.
Cloud type and frequency is also compared among the three datasets by using limited temporal and spatial thresholds. In
general, all three datasets yield more consistent detection of high cloud types than the other cloud types. Although this
paper emphasizes on the comparison of spatial distribution, seasonal variation and coincident samples, it also discusses the
strengths and weakness of different passive/active techniques.