Previous: Introduction Next: Ozone Profile Tests Up: Ext. Abst.

 

RegCM2 - Regional Climate Model

The use of regional climate model as a tool for so-called dynamical downscaling became quite popular in 90â. It means to run a limited area model nested in GCM for long-period simulations. There are some advantages of such an access, i.e. finer resolution of results, possibility of more detailed parameterization etc., but it is more demanding on computer resources and computer time. After couple year of experience with the first generation of NCAR RegCM (Regional Climate Model) (Dickinson et al., 1989; Giorgi, 1990) based on NCAR-PSU (National Center for Atmospheric Research ö Pennsylvania State University) MM4 compressible hydrostatic grid point model but with modified physics for use in climate studies Giorgi et al. (1993a,b) upgraded the model resulting in the second generation RegCM, hereafter reffered to as RegCM2. The physics of the model moved to MM5 (Grell et al., 1991) and NCAR CCM2, during the last few years some other modifications appeared. Radiation package has been replaced by that of CCM3 (Kiehl et al., 1996), an option concerning parameterization of deep convection has been added, changes in moisture scheme adopted, lake effects included etc. The model has a lot of optional parameters, it is possible to run it with different selections of geometry parameters, domain size and position, gridsize (B type of Arakawa-Lamb classification), projections, orography resolution, number of sigma levels, top of the atmosphere etc. For numerical integration, split-explicit time scheme is used, boundary conditions being interpolated from global data with selected step, but the minimum in our sample data for June 1994 is 6 hours. 13 land-use climatologic categories are involved with seasonal resolution for parameterization of surface processes of land ö atmosphere interactions.

For the purposes of this study we run the model with number of levels set on 20, one month experiment takes about 4.5 hours for gridsize of 180 km on the domain of Europe using Sun Ultra 60 workstation with two 360MHz processors. We have run several sensitivity tests on the influence of ozone profile parameterization in the model for June 1994 as well as some real month study for June, March and October 1994.


Previous: Introduction Next: Ozone Profile Tests Up: Ext. Abst.