Four previously identified patterns of meso-scale cloud organization in
the trades -- called Sugar, Gravel, Flowers and Fish -- are studied
using long-term records of ground-based measurements, satellite
observations and reanalyses. A neural network trained to detect these
patterns is applied to satellite imagery to identify periods during
which a particular pattern is over the Barbados Cloud Observatory.
Surface based remote sensing at the observatory is composited and shows
that the patterns can be distinguished by differences in cloud
macro-physical structures. Variations in total cloudiness among the
patterns are dominated by variations in cloud-top cloudiness. Cloud
amount near cloud base varies little. Each pattern is associated with a
distinct atmospheric environment whose characteristics are traced back
to origins that are not solely within the trades. Sugar air-masses are
characterized by weak winds and of tropical origin. Fish are driven by
convergence lines originating from synoptical disturbances. Gravel and
Flowers are most native to the trades, but distinguish themselves with
slightly stronger winds and stronger subsidence in the first case and
greater stability in the latter. These results suggest that due to the
tight bound of the patterns to wind and air-mass origin, the patterns
with the higher cloud fraction, Flowers and Fish, will be disfavoured in
a warming climate with more equable sea-surface temperatures and fewer
mid-latitudinal disturbances.