Evaluation of CMIP5 Model Fidelity in Producing Large-Scale Meteorological Patterns Conducive to Extremes over the Northwestern United States

Climate change is driving demand for climate information at impacts-relevant scales. However, projecting future climate at such scales is challenging given current modeling capabilities, especially in regions of complex terrain and high climatic diversity such as the northwestern United States (NWUS). Here we use reanalysis data to identify and characterize large-scale meteorological patterns (LSMPs), which can be readily resolved in contemporary climate models and relate them to local scale extremes in temperature and precipitation over the NWUS. We define daily LSMPs using circulation fields from the lower, middle, and upper troposphere. First, the Self-Organizing Maps (SOMs) approach is employed to characterize LSMPs for the wet (November through March) and dry (June through September) seasons over a 35-year climatology by sorting days with similar patterns into clusters or “nodes”. Next, we identify which nodes are most conducive to daily temperature and precipitation extremes at regional and local scales across the domain. These archetypal patterns are then used as an observational basis to systematically evaluate climate model fidelity in producing the frequency and structure of synoptic states favorable to such conditions. Results elucidate the extent to which the suite of models, from the Fifth Phase of the Coupled Model Intercomparison Project, physically capture key circulation patterns over the NWUS in historical simulations and further provide a means for identifying which models reproduce LSMPs. Such dynamically based understanding of synoptic scale variability within a climate context can be leveraged to interpret projected future changes in LSMPs and associated extremes under anthropogenic warming.