Abstract. In order to study aerosol–cloud interactions in cirrus clouds, we apply a new multiple-mode ice microphysical scheme to the general circulation model ECHAM5-HAM. The multiple-mode ice microphysical scheme allows for analysis of the competition between homogeneous freezing of solution droplets, deposition nucleation of pure dust particles, and immersion freezing of coated dust particles and pre-existing ice. We base the freezing efficiencies of coated and pure dust particles on the most recent laboratory data. The effect of pre-existing ice, which has been neglected in previous ice nucleation parameterizations, is to deplete water vapour by depositional growth and thus prevent homogeneous and heterogeneous freezing from occurring. As a first step, we extensively tested the model and validated the results against in situ measurements from various aircraft campaigns. The results compare well with observations; properties such as ice crystal size and number concentration as well as supersaturation are predicted within the observational spread.
We find that heterogeneous nucleation on mineral dust particles and the consideration of pre-existing ice in the nucleation process may lead to significant effects: globally, ice crystal number and mass are reduced by 10 and 5%, whereas the ice crystals' size is increased by 3%. The reductions in ice crystal number are most pronounced in the tropics and mid-latitudes in the Northern Hemisphere. While changes in the microphysical and radiative properties of cirrus clouds in the tropics are mostly driven by considering pre-existing ice, changes in the northern hemispheric mid-latitudes mainly result from heterogeneous nucleation. The so-called negative Twomey effect in cirrus clouds is represented in ECHAM5-HAM. The net change in the radiation budget is −0.94 W m⁻², implying that both heterogeneous nucleation on dust and pre-existing ice have the potential to modulate cirrus properties in climate simulations and thus should be considered in future studies.