A scale-aware parameterization of precipitation formation in climate models - Integral constraints on autoconversion and accretion
Cloud schemes in regional and global climate models have to be highly parameterized to allow for long integration times. Because of the non-linearity of many cloud microphysical processes and the unresolved subgrid scale variability of cloud properties and other meteorological variables, cloud schemes are often empirically tuned to obtain agreement with observed top of the atmosphere radiative fluxes. This tuning approach significantly limits the use of these schemes in climate change scenarios in two ways. The parameterizations become scale-dependent and it remains unclear whether this empirical and unphysical tuning to current-day conditions is still valid for global warming scenarios.