In this study, the super droplet-method (SDM) is used in large-eddy simulations of an isolated cumulus congestus observed during the 1995 Small Cumulus Microphysics Study field project in order to investigate the intra-cloud variability associated with entrainment and mixing. The SDM is a Lagrangian particle-based method for cloud microphysics that provides droplet size distributions (DSD) coupled to the simulated cloud-scale dynamics. The authors show that sensitivity to the spatial resolution and the initial number of particles is larger, and sensitivity to the initial conditions is smaller, when the order of the DSD moment is smaller. Through the use of simulations with reliable statistics, microphysical variability is investigated at scales of ∼ 100 m that can be considered well resolved in both the numerical simulations and in-situ aircraft observations. Large spatial variability in cloudy volumes is shown to be strongly affected by entrainment. Mean values of the adiabatic fraction (AF), cloud droplet number concentration, and the cubed ratio of the mean volume radius and the effective radius (k) agree well with observations in the middle and upper cloud layers. Moreover, the AF and k values are found to be positively correlated, and the reduction of the mean volume radius scaled by its adiabatic value with the decrease of the mean droplet concentration scaled by its adiabatic value is found to be smaller than the theoretical prediction of homogeneous mixing. The latter supports the notion of inhomogeneous mixing due to entrainment.