A Dynamical and Statistical Characterization of United States Extreme Precipitation Events and their Associated Large-Scale Meteorological Patterns
The utility of employing a large-scale meteorological pattern (LMP) approach in characterizing the synoptic and dynamic nature of extreme precipitation events is demonstrated. Climate model skill in representing the details of extreme event behavior is found to be positively correlated to model ability in simulating the corresponding LMP structure and frequency.
The study provides a practical new framework for examining the weather-climate nexus. More specifically, extreme precipitation events are related to recurring regional large-scale meteorological patterns (LMPs) that drive the smaller-scale weather events.
A hierarchical cluster analysis of daily precipitation is used to isolate canonical extreme precipitation patterns (EPPs) for the boreal warm and cool seasons in observations and a parallel historical simulation of the CCSM4 for the period 1950-2005. The large-scale meteorological pattern (LMP) that induces each of the observed EPPs is isolated and used as fundamental basis function for testing a climate model’s potential for properly simulating extreme precipitation patterns. Parallel statistical and synoptic analyses of both observed and simulated EPPs and LMPs are carried out. The utility of the LMP methodology in characterizing the synoptic and dynamic nature of extreme events is demonstrated. Model skill in representing EPP behavior details is found to be positively correlated to its ability to simulate the corresponding LMP. For example, the model bias in the occurrence frequency of a given LMP is directly related to an analogous frequency bias in the corresponding EPP induced by the LMP.