The western U.S. receives precipitation predominantly during the cold season when storms approach from the Pacific Ocean. The snowpack that accumulates during winter storms provides about 70-90% of water supply for the region. Associated with the warm sector of extratropical cyclones over the Pacific Ocean, atmospheric rivers (ARs) provide enhanced water vapor transport from the tropics to produce heavy precipitation upon landfall in the mountainous terrain of the western North America. Using a suite of idealized aqua-planet simulations and AMIP simulations with the Model for Prediction Across Scales (MPAS) coupled to the Community Atmosphere Model (CAM) physics parameterizations at multiple resolutions ranging from 30km to 240km, we investigate the sensitivity of simulated AR frequency to model resolution. The impacts of global warming on ARs and heavy precipitation are investigated using model outputs from the Community Earth System Model Large Ensemble Project (CESM-LE) and the multi-model ensemble of the Coupled Model Intercomparison Project Phase 5 (CMIP5) to evaluate the thermodynamical and dynamical contributions to changes in extreme precipitation. Analysis of the hierarchy of climate simulations highlights uncertainty in model simulation of the jet position as a major source of uncertainty in simulating AR frequency and a possible dynamical convergence as model resolutions approach roughly 50km resolution. To extend the modeling hierarchy for improved representation of the interactions between ARs and the complex terrain that generate high intensity precipitation and floods, MPAS global variable resolution simulations with a regional refinement down to sub-10km resolution will be discussed.