The Model for Prediction Across Scales (MPAS) is a nonhydrostatic global variable resolution modeling framework. Using MPAS coupled with the Community Atmosphere Model (CAM5) physics suites, simulations are being performed at grid spacing ranging from 4 km to 50 km using regional refinement over North America. These simulations provide an opportunity to systematically evaluate the impacts of model resolution on simulation of extreme events. We will present analysis of mesoscale convective systems, which are responsible for most of the extreme precipitation events during the warm season in the central and eastern U.S., and atmospheric rivers, which are associated with cold season extreme precipitation in the western U.S. Biases in simulating mesoscale convective systems and atmospheric rivers will be discussed in the context of model biases in representing the large-scale circulation and that vary with model resolution and the implications to model extreme precipitation.