Bermuda High Research

 

Abstract of our work

On the simulated and observed connection between the location of the Bermuda High and the summertime distribution of precipitation over the continental United States

Authors: Laura J. Bell, Richard B. Rood, and Derek J. Posselt, University of Michigan, Ann Arbor, MI.

 

Climate models have historically had difficulty predicting summertime precipitation in the United States.  In regions where topography plays an important role in precipitation patterns, finer model resolutions have shown more realistic simulations of precipitation.  Increasing resolution in other locations has had varied impacts based on the season.  For regions such as the Southeast U.S., higher resolution does not necessarily lead to more realistic results, particularly for summertime months.  This identifies a need for evaluating a model's ability to demonstrate the connection between atmospheric dynamics and resulting precipitation patterns.  One dynamical feature of interest is the Bermuda High (BH).  The BH is a high pressure system located in the Atlantic and it is strongest during northern hemisphere summertime months.  Its extent reaches the Southeast U.S. thereby influencing wind patterns, moisture fluxes, and precipitation patterns.  The significance of the BH in the chain of dynamics that result in precipitation is one focus of our study.

Five nonconsecutive years were selected for an investigation into similarities and differences between data sets and different model resolutions for sea level pressure (SLP), moisture flux, and precipitation.  Included is 1988, a year characterized by drought in the Southeast U.S., and 1993, a flood year in the upper Midwest.  Drawing from Community Atmosphere Model (CAM) data and Modern Era Retrospective-Analysis for Research and Applications (MERRA) data, we have compared ½, 1, and 2 degrees spatial resolution in the CAM and the equivalent ½ degree in the MERRA.  Winds in both datasets, particularly those that originate in the Gulf of Mexico, show evidence of a connection to the strength and location of the BH; southeasterly winds in the Gulf tend to be directed northward along the Continental Divide then continue in a clockwise pattern reaching as far east as the Ohio River Valley.  In both CAM and MERRA, these winds import significant quantities of moisture from the Gulf creating a region of large moisture flux in the southern Great Plains region.  As the location of the BH changes, Gulf of Mexico winds also shift, resulting in a change in regions of positive moisture flux and associated changes in simulated precipitation patterns in the CAM.  Comparison of each summer season indicates that the precipitation distribution produced by the CAM is consistently linked to the location of the maximum moisture flux and moisture flux convergence, whereas MERRA often produces precipitation over large portions of the interior of the U.S. that are not associated with large moisture flux or moisture flux convergence. The results reflect the importance of mesoscale precipitation production mechanisms that are represented in the MERRA, but which are not well simulated by GCMs.