This paper presents an advancement of the recently developed particle-resolved aerosol model PartMC-MOSAIC (Particle Monte Carlo-Model for Simulating Aerosol Interactions and Chemistry) to investigate the impacts of mixing state on cloud droplet formation and to provide a tool for the quantification of errors in cloud properties introduced by simplifying mixing state assumptions. We coupled PartMC-MOSAIC with a cloud parcel model. We initialized the cloud parcel simulation with hourly PartMC-MOSAIC model output from a 48-hour urban plume simulation. The cloud parcel model then explicitly simulated activation and condensational growth of the particles as the parcel underwent cooling at a specified rate and the particles of the aerosol population competed for water vapor. We used this capability to quantify the relative importance of size information versus composition information for the prediction of the cloud droplet number fraction, mass fraction of black carbon that is nucleation-scavenged, cloud droplet effective radius, and relative dispersion of the droplet size distribution by introducing averaging of particle-resolved information within prescribed bins. For the cloud droplet number fraction, both composition averaging and size-bin averaging individually led to an error of less than 25% for all cloud parcel simulations, while averaging in both size bins and composition resulted in errors of up to 34% for the base case cooling rate of 0.5K/min. In contrast, for the nucleation-scavenged black carbon mass fraction, the results for size-bin averaging tracked the reference case well, while composition averaging, with or without size-bin averaging, led to overestimation of this quantity by up to 600%. Copyright © 2012 American Geophysical Union. All Rights Reserved.