Modeling of mountain waves in T-REX

The main objective of the Terrain-induced Rotor Experiment (T-REX) is to understand the nature of coupling of mountain-induced rotor circulations to the structure and evolution of overlying mountain waves and to the underlying boundary layer. T-REX field activities took place in March and April 2006 directly east of the southern Sierra Nevada Mountains. Several different high-resolution modeling efforts occurred in support of real-time and research objectives of the experiment. At the Global Systems Division (GSD) of the NOAA Earth Systems Research Laboratory, we ran the ARW and NMM cores of the Weather Research and Forecasting (WRF) community model twice per day at 2-km grid spacing with 51 levels. The two WRF cores were initialized with analyses produced by the operational RUC (Rapid Update Cycle) model. The ARW model configuration included a 1.5-order TKE closure scheme and a diffusive damping sponge in the upper 5 km of the model, whereas the NMM utilized “internal divergence damping” throughout to reduce the vertical propagation of gravity wave energy. The two WRF cores used an identical set of interoperable physics that became available just prior to the startup of the T-REX field phase, so that the sensitivity of the structure of mountain waves and wave breaking to the model numerics and their predictability could be assessed.

Results obtained from the real-time forecsts show the NMM divergence-damping effect reduced the amplitude of vertically propagating waves considerably more so than did the ARW upper-layer sponge scheme. Gravity waves in the lower stratosphere were much smoother and lower in amplitude in both WRF models than in the COAMPS model run by the Naval Research Laboratory. This was surprising, in light of recent idealized flow simulations performed by Doyle and Jiang (2005), which discussed how comparatively less dispersion in the WRF model resulted in greater gravity wave amplitudes downstream of bell-shaped mountains. Nonetheless, cases did occur of upper-level wave breaking in our T-REX WRF model simulations. The perturbations were at times very large, and corresponded well with aircraft measurements. One case in particular stands out, wherein horizontal wind speed perturbations of 50 m/s and potential temperature perturbations of 15 K over a distance of just 30 km occurred.

It has been established previously that a well-posed non-reflective upper boundary condition should be used to obtain accurate simulations of mountain waves. We will discuss the results of sensitivity experiments being run for selected IOP events after the field phase to investigate such issues as the method used for reducing spurious gravity wave reflection off of the top boundary, the NMM divergence damping method, and the handling of the terrain insofar as its smoothness and effects on the surface pressure gradient force are concerned.