Intercomparison of gravity wave parameterizations: Hines Doppler-spread and Warner and McIntyre ultra-simple schemes

The "three-part ultra-simple" gravity wave parameterization of Warner and McIntyre is compared with the "Doppler-spread" parameterization of Hines. The two parameterizations are tested on a background state at rest with constant buoyancy frequency, as well as on background states defined by the CIRA86 data at 70N in July and January. To achieve as clean a comparison as possible between the two parameterizations, two approaches are taken. The first approach is to adjust the free parameters to obtain the same source level momentum fluxes, and as similar a source spectrum shape, as is possible. The second approach is to adjust the source level momentum fluxes to obtain the same momentum fluxes at mesospheric altitudes. The resulting vertical profiles of the momentum fluxes, of the wave-induced forces, and of the energy dissipation rates produced by the two parameterizations are compared. When a similar gravity wave source spectrum is used, specifically the source spectrum recommended by Hines, momentum deposition generally tends to occur lower in the atmosphere for the Warner and McIntyre parameterization than for the Hines Doppler-spread parameterization. In order to obtain similar wave-induced forces, and dissipation rates in the mesosphere from the two parameterizations, it has been found that the Warner and McIntyre parameterization requires the source spectrum to be scaled so that the net momentum flux in the lower stratosphere is an order of magnitude higher than the Hines Doppler-spread parameterization.