Cloud fraction recomputed following the RCEMIP protocol (cloud condensate threshold value of 1×10−5 g g−1, or 1 % of the saturation mixing ratio over water, whichever is smaller), but additionally setting a minimum threshold for cloud condensate of 5×10−7 g g−1. With this additional threshold, extremely thin anvil clouds are discarded. The discarded "clouds" have almost zero effect on shortwave fluxes and a less than 1 W m-2 global effect on longwave fluxes.

The reason why there are many of these very thin clouds at the tropopause in ICON is that the microphysics scheme uses a sedimentation velocity of ice particles that is quite small compared to other models (but in line with observations), and that the threshold value for self-collection is set to 10−7, and to 10−6 for ice. This is done to speed up the computation. Smaller threshold values would result in less of the very thin clouds.