Numerical simulations of a large-amplitude mesoscale gravity wave event

Numerical simulations with the NCAR=PSU Mesoscale Model 5 (MM5) were performed to study a large-amplitude gravity wave event that occurred on 4 January 1994 along the East Coast of the United States. Results from the MM5 control simulation using a 12-km mesh resolution compared well with the synoptic and mesoscale observational analysis. The simulated gravity waves displayed timing, location, wavelength, and propagation speed similar to those observed in a synoptic-scale environment described by the Uccellini and Koch (1987) conceptual model. Additional features existing upstream of the wave generation region not contained within their conceptual model were a warm occlusion and tropopause fold prior to and during the gravity wave generation. Wave ducting criteria were nearly satisfied along the path of the gravity waves. Several sensitivity tests were performed. In a simulation in which the Appalachian Mountains were removed, the model still produced similar cyclone development and mesoscale gravity waves. Thus topography was not directly responsible for the gravity wave genesis. Also, three different ‘‘fake dry’’ sensitivity tests were performed with the latent heating related to changes of water substance turned off in the model at different stages of the simulation. The results from these simulations suggest that diabatic heating played an important role in both jet=cyclone development and in gravity wave amplification and maintenance, though not wave generation. The simulation with grid resolution increased to 4 km, which included fully explicit microphysics produced gravity wave characteristics similar to those in the control simulation, though the higher resolution resolved much shorter waves (though unverifiable) closely associated with convection. This 4-km sensitivity experiment with no cumulus parameterization also confirmed that the dominant gravity wave was not an artifact of the particular cumulus parameterization scheme used for the control simulation. The reliability of the simulated gravity waves is further confirmed with another sensitivity experiment initialized 20 hours before the observed wave generation in which qualitatively-similar gravity waves were produced.