Abstract. There are two fundamental mechanisms through which cirrus clouds form; homo- and heterogeneous ice nucleation (henceforth hom and het). The relative contribution of each mechanism to ice crystal production often determines the microphysical and radiative properties of a cirrus cloud. This study attempts to estimate the radiative contribution of hom relative to het by constraining the cloud microphysics in a climate model to conform with satellite retrievals of cirrus cloud effective diameter De, where the sampled cirrus cloud base had a temperature T < 235 K (−38 °C). The CALIPSO (Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation) satellite retrievals for cirrus clouds are compared against an updated in situ cirrus cloud property climatology to evaluate similarities and differences. In this climate modeling study, we ask how the cloud radiative effect (CRE) based on retrieved cirrus cloud properties compares with the CRE predicted by a model configuration representing cirrus clouds formed only through het and also with the CRE predicted by the standard configuration of the model. To answer this question, we constrained version 2 of the Morrison-Gettelman cloud microphysics scheme (MG2), which is used in several climate models, using effective diameter (D_e) retrievals from the CALIPSO satellite. A new subroutine within the MG2 scheme provides retrieved D_e as a function of temperature (T), latitude, season and land fraction, while ice particle mass and area relationships are used to relate D_e to the ice particle size distribution (PSD) slope and to produce new relationships for the number- and mass-weighted ice fall speeds. These and other modifications rendered the MG2 microphysics consistent with the D_e constraint. Using 40-year simulations of the Whole Atmosphere Community Climate Model version 6 (WACCM6), the CRE outside the tropics from the De-constrained WACCM6 was greater than standard WACCM6 by 1.63 W m⁻² in the Northern Hemisphere (NH) and 2.59 W m⁻² in the Southern Hemisphere (SH). Using the version of WACCM6 designed to represent cirrus clouds formed only by het (instead of using standard WACCM6), this difference was 2.37 W m⁻² in the NH and 2.55 W m⁻² in the SH. These differences are larger when only non-summer months are considered.