Many cumulus parameterization schemes use resolvable grid-scale variables to predict the cloud base and cloud top with some closure assumptions. It was hoped that by doing so, cumulus parameterization could provide heating profiles that bear physical variability. However, such a practice often causes energy accumulation in smaller scales, and produces marginally-resolvable-scale features in numerical weather simulations.
In this study, analyses and a vertical normal mode decomposition of the heating profiles are conducted. These analyses may aid us to gain some understanding of the response of large-scale dynamics to various depths of cloud forcings.