Recent work has demonstrated the importance of anthropogenic landscape heterogeneity in Amazonia and the central U.S. in forcing mesoscale atmospheric circulations and generating clouds and precipitation. It has been shown that failing to account for this surface heterogeneity may have important implications for the simulation of land--atmosphere interactions in large-scale atmospheric models. The work presented here shows that, in addition to correctly specifying the surface forcing, accurate simulation of landscape-induced mesoscale effects with a regional model (e.g., Regional Atmospheric Modeling System, RAMS) depends critically on informed choices of aspects of model configuration. These include horizontal resolution, strength of nudging, and atmospheric initialization. Sensitivity experiments designed to illustrate the impact of varying these factors on RAMS-simulated, landscape-induced mesoscale circulations are presented here. These results, supported by dynamical and scaling arguments, suggest that model configuration can have at least as great an impact on the simulated mesoscale circulations, fluxes, and PBL structure as correctly accounting for mesoscale landscape heterogeneity. Such experiments therefore provide crucial guidance for optimally constructing simulations of mesoscale land--atmosphere interaction important for a variety of applications from field campaigns to parameterization development.