Well-mixed atmospheric boundary layers in the MM5 and WRF models

Mesoscale atmospheric simulation models play an increasingly important role in operational forecasting as well as meteorological research. The Taunton, Massachusetts National Weather Service Forecast Office runs a workstation version of the North American Mesoscale model (formerly known as the Eta model), and is in the process of implementing version 5 of the Penn State/NCAR Mesoscale Model (MM5) alongside the new Weather Research and Forecasting (WRF) model. The list of locations around the world who run the MM5 in real-time numbers more than 30, and many of these locations also run the newer WRF model.

Correctly simulating atmospheric boundary layer (ABL) structure remains crucial to successfully model many mesoscale phenomena, including ABL height, sea breeze circulations, and coastal fronts. Both the MM5 and the WRF model have the ability to use a variety of ABL parameterizations, including the MRF boundary layer, YSU scheme, Blackadar, Burk-Thompson, Eta M-Y (Mellor and Yamada, ) and Gayno-Seaman. Preliminary work with the MM5 showed that the ability to correctly simulate a daytime, convective ABL, varied significantly, depending upon the parameterization used. For instance, mean ABL potential temperature could differ by as much as 4 – 5 K in a 24-hour forecast between the MRF and Burk-Thompson schemes.

To gather updated data on the WRF model, both models have been run using identical nested horizontal and vertical grids, and using the same initialization data. Nine days from 2001, which had already been run with the MM5, were re-run with the WRF model, and the results from all the tests will be shown. The outer grid used 60 km spacing, and the single nested grid used a 20 km spacing. All of the runs use 23 vertical levels, with 14 below 500 hPa. The simulated ABLs are compared with soundings at three locations at 00 UTC. The three locations used are Chatham, MA, Albany, NY, and Brookhaven, NY. The Albany location was used as an inland location, Chatham as a coastal location on the elbow of Cape Cod, and the Brookhaven location as something in between, located on Long Island, NY.

Results include differences in surface sensible and latent heat fluxes, mean ABL potential temperature and mixing ratios, and ABL heights. These simulations have also shown upper-level differences as well, most significantly in 500 mb moisture. Understanding the connection between upper-level moisture and ABL parameterization is an on-going project.