Convective Cloud Fraction

Convective cloud fraction is the fraction of a model grid cell that a cumulus scheme reports as covered by convective cloud. It exists because the scheme’s cloud is sub-grid by construction: the convection is not resolved, so its areal extent has to be diagnosed rather than observed from the model state.

The reason this small diagnostic matters more than its obscurity suggests is that it is the handoff point between the convection scheme and the radiation scheme. Radiation does not see the convection scheme’s internal plume; it sees a cloud fraction and a condensate amount. Whatever the convection scheme does internally reaches the energy budget only through those numbers.

The coupling that causes trouble

Convection schemes and radiation schemes are developed separately, often by different groups, and combined by the model user. Nothing enforces that the cloud fraction one produces means what the other assumes.

Two failure modes recur:

• Double counting. If the cumulus scheme reports condensate that the microphysics scheme also carries, radiation sees the cloud twice. The result is too much reflected shortwave and a cold bias that is hard to attribute, because neither scheme is individually wrong.
• The overlap assumption. Cloud fraction at each level has to be combined into a total, and the assumption used — maximum, random, or maximum-random overlap — changes the total substantially for deep convective columns. This choice usually lives in the radiation code rather than the convection code, so it is easy to miss when changing convection schemes.

This is the concrete reason for advice that otherwise sounds like superstition: in WRF, do not change the cumulus scheme without checking the radiation and microphysics configuration. The schemes are not independent, and the coupling runs through cloud fraction. Energy-budget surprises after a scheme swap are usually this.

What sets it

The fraction is typically diagnosed from the scheme’s detrainment — detrained condensate is what becomes visible cloud — sometimes with an updraft-area assumption. Since detrainment is among the least constrained parts of a convection scheme, the cloud fraction inherits that uncertainty and passes it directly to the radiation.

In convection-permitting configurations the quantity largely disappears, because resolved convection produces condensate the microphysics handles and radiation sees directly. This removes a genuine source of inconsistency and is an underrated argument for the approach, separate from the more commonly cited improvements in rainfall intensity and timing.

See also: cumulus convection schemes, radiation schemes, and cloud detrainment.