A simple parametric relationship is established between factors describing the shape of cloud condensation nuclei (CCN) activation spectra and observable properties of the aerosol population they grow on (size distribution and solubility). This is done independently for maritime and continental aerosol types because of their very different characteristics. The data used for the multiple statistical adjustments in the procedure described in this paper are generated by running a numerical model of aerosol growth coupled to a simple cloud droplet activation scheme. Each aerosol population (maritime and continental) is assumed to be of homogeneous chemical composition, lognormally distributed and with variable solubility. The parameterization is then evaluated using a large set of aerosol populations with randomized properties. Finally, the study presents a preliminary analysis of the most important aerosol properties that influence the shape of the CCN spectra. An idealized scenario of a clean maritime boundary layer cloud perturbed by anthropogenic emissions (such as the ship track problem) illustrates the capability of the parameterization to selectively increase the cloud droplet concentration in a partially polluted cloud. The calibration results presented in this paper are not meant to be the definitive activation spectra produced by any lognormally distributed aerosols. These results are indeed a step toward an objective initialization of CCN spectra and hence toward the computation of cloud droplet concentrations based on measurable multimodal aerosol features, as required by three-dimensional numerical models with a coupled interactive aerosol module. ¿ 2000 American Geophysical Union