Cloud-radiative heating (CRH) within the atmosphere and its changes with warming affect the large-scale atmospheric wind patterns in a myriad of ways, such that reliable predictions and projections of circulation require reliable calculations of CRH. In order to assess sensitivities of upper-tropospheric midlatitude CRH to model settings, we perform a series of simulations with the Icosahedral Nonhydrostatic Model (ICON) over the North Atlantic using six different grid spacings, parameterized and explicit convection, and one- versus two-moment cloud microphysics. While sensitivity to grid spacing is limited, CRH profiles change dramatically with microphysics and convection schemes. These dependencies are interpreted via decomposition into cloud classes and examination of cloud properties and cloud-controlling factors within these different classes. We trace the model dependencies back to differences in the mass mixing ratios and number concentrations of cloud ice and snow, as well as vertical velocities. Which frozen species are radiatively active and the coupling of microphysics and convection schemes turn out to be crucial factors in altering the modeled CRH profiles.
