A numerical model for lee wave forecasting

A three-dimensional numerical model has been developed in order to provide high-resolution forecasts of internal gravity-wave activity over complex terrain. The model is based on the unsteady linear inviscid Boussinesq equations of motion, discretized in finite-difference form. The equations are linearized about a horizontally-homogeneous basic-state wind and temperature field, obtained from either radiosonde data or, for forecasting purposes, large-scale weather prediction model data. Solutions are obtained by integrating the unsteady equations forward in time from an initial state in which there are no waves present, until a quasi-steady wave field is obtained. The model has been used to provide lee-wave forecasts during two field campaigns in 2000:

(i) SAGE III Ozone Loss and Validation Experiment (SOLVE) / THESEO 2000. One of the aims of this NASA / EU led investigation into Arctic stratospheric ozone depletion was to study the formation of gravity-wave induced polar stratospheric clouds (PSC's). Along with mesoscale model (MM5 and HRM) forecast products and another gravity-wave prediction scheme (MWFM), high resolution gravity-wave forecasts were provided daily (at 12 hour intervals up to 3 days ahead) for limited areas (256 x 256 km) over northern Scandinavia and Spitzbergen in order to aid the planning of aircraft missions.

(ii) A field experiment on the Isle of Arran, SW Scotland. This project, conducted by the University of Leeds in collaboration with the UK Meteorological Office, UMIST and the UK Forestry Commission, aimed at studying the effect of gravity waves on the near-surface flow. Of particular interest was the effect of trapped gravity waves, generated in northerly flows by a mountain range at the northern end of the island, on flow separation which occurred on moderate rolling hills at the south end of the island. Gravity-wave forecasts were provided daily (at 12 hour intervals up to 1.5 days ahead) for an area (64 x 64 km) covering the Isle of Arran and the Kintyre peninsula. The forecasts were used to help plan radiosonde launches during promising gravity-wave events.

For both campaigns the basic-state wind and temperature profiles were obtained from ECMWF 0.5 degree global forecast data.

The accuracy of the gravity-wave forecasts is examined by comparing with measurements obtained in both field campaigns and by comparing the actual forecasts with hindcasts initialized with radiosonde ascents made during specific case studies (rather than weather prediction model data). The feasibility of using such a model for routine gravity-wave forecasts is discussed.