Application of the non-hydrostatic MC2 LAM to the Brig 1993 flash flood event and the PYREX IOP 3 foehn case

R. Benoit and M. Desgagné, Recherche en Prévision Numerique, Environment Canada 2121, Dorval, Canada H9P 1J3

Over the past three years, the dynamical framework of the semi-implicit semi-Lagrangian discretization of the Euler compressible fluid equations of Tanguay, Robert and Laprise (1990) has been developed further into a complete non-hydrostatic 3D weather forecasting LAM known as the MC2 model (Benoit et al., 1995). In various 2D orogra-phic flow regimes over idealized terrain, the MC2 delivers simulations comparable to the state-of-the-art (Pinty et al., 1995); the SL scheme maintains very sharp small scale features and does not distort the solution as long as the Courant number is kept below1.

The phase I of the Mesoscale Alpine Programme (MAP) organized by Alpine countries of Europe is being dedicated to modeling activities, and we are applying the MC2 to the September 1993 flash flood case which hit Brig, a city in Switzerland along the Rhone River valley. This falls under the first objective of MAP - topographically influenced precipitation - for which MC2 has already been used extensively, down to the 2 km resolution over the Canadian Rockies. In comparison, Mediterranean fall time events have much greater intensity, very dense data coverage and dramatic societal impact. For the Brig case, MC2 is nested on a set of 50, 10 and 2 km grids for periods of 36, 24 and 12 h respectively, focusing on September 23 1993. The MC2 experiments complement well the hydrostatic simulations of the similar Piedmont flood case of November 1994 made by the SMR group (Italy) at 40/20 km and the MeteoSwiss group (Switzerland) at 14 km. Deep convection for our 10 km grid is computed with the Fritsch-Chappell (1980) scheme including moist downdrafts while for the 2-km run convection is completely explicit. The sample results shown on next page are from runs at 10 km and 2 km which are initialized from the T213L31 ECMWF analysis directly on its terrain following coordinate (courtesy of ECMWF Res. Dept.). Canadian Meteorological Centre T119L21 (1.5° ) global analysis has also been tested but with significantly worse results. Preliminary runs show that the location and amplitude of the precipitation cores are much affected by the resolution of initial conditions and topography. Terrain elevation is from the newly released GLOBE dataset (global, currently 30 arc second resolution over Europe, N. America and elsewhere) from NOAA/NGDC. Directional roughness on the 2 km grid was computed from the 6 arc second (185 m) terrain model of Switzerland (Volkert, 1990). The lower two panels are from the 10 km simulation: on the left is the surface wind (kt) arrows and isotachs (10 kt increments), after 15 h, valid at 15 UTC September 23 1993; on the right is the 24 h accumulated precipitation at the end of the same day (10 mm increments). Note the complexity of the wind field over the Mediterranea as well as the currents over Genova and the Po valley merging over the Piedmont foothills. The three largest rain maxima have been highlighted: 223 mm off Barcelona, 232 on the lower Rhone and 176 near Monte Rosa. The top two panels are from the 2 km run, after 4h, valid at 20 UTC September 23 1993: on the left is a full domain view of orography (gray shades), accumulated precipitation (contours) and 900 hPa flow; on the right is a zoom of the first two fields over the storm area. The three largest rain cores are indicated, with a 131 mm center approaching the Toce River valley, where the raingauges maxima occurred (Cavalli and Pallanzeno stations) (courtesy of P. Binder, MeteoSwiss).

In parallel to the simulation of the Brig case, the MC2 is also tested on the PYREX IOP 3 following the protocol of experiments setup by Bougeault for the international COMPARE II model intercomparison. As our initial contribution to the second objective of MAP - Gravity waves and Foehn - the simulation of the IOP 3 at the kilometric scale has been achieved (not shown here) with emphasis on the non-hydrostatic effects (such as lee wave), 3 dimensionality and PV diagnostics in relation with flow splitting.

References

Benoit, R., M. Desgagné, P. Pellerin, S. Pellerin, Y. Chartier and S. Desjardins, 1995: The Canadian MC2: a semi-Lagrangian, semi-implicit wide-band atmospheric model suited for fine-scale process studies and simulation. In preparation for submission to Mon. Wea. Rev..

Fritsch, J.M., and C.F. Chappell, 1980: Numerical prediction of convectively driven mesoscale pressure systems. Part I: Convective parameterization. J. Atmos. Sci., 37, 1722-1733.

Laprise, R. 1995: The formulation of André Robert MC2 (Mesoscale Compressible Community) model. Accepted for publication in Atmos.-Ocean, for the A.J. Robert Symposium special issue.

Pinty, J.P., R. Benoit, E. Richard, R. Laprise, 1995: Simple tests of a semi-implicit semi-Lagrangian model on 2D mountain wave problems. Accepted for publication in Mon. Wea. Rev. Tanguay ,

Robert, M.A., and R. Laprise, 1990: A semi-implicit semi-lagrangian fully compressible regional forecast model. Mon. Wea. Rev., 118, 1970-1980.

Volkert, H., 1990: An Alpine orography resolving major valleys and massifs. Meteorol. Atmos. Phys., 43, 231-234.



MAP Data Centre - April '05 - MAP WebMaster