CGCM2 predictors: A2 1 and B2 1 experiments

See CGCM2 Window Limits and Box Numbering for correspondence between BOX numbering and longitude-latitude coordinates. The latitude and longitude coordinates correspond approximately to the centres of the grid boxes. This data is available for North America only.

Click on the box you wish to retrieve

Computing the derived variables

The normalized predictor variables derived from the CGCM2 A2 1 and B2 1 experiments (the number 1 corresponds to the first member of the three ensemble runs) have been calculated in the same way as the CGCM1 predictors, except that the calculation is performed directly on the original CGCM2 gaussian grid (Flato and Boer, 2001) instead of a regular latitude-longitude grid. The numerical division of the derived variables was then modified to match the gaussian grid. While modifying the code, an error in the computation of the geostrophic wind divergence derived variable was found (see Choux, 2005). The appropriate correction has been done (see CGCM2 Divergence-Predictor Correction). The same correction was applied to the NCEP divergence-predictors that you will need to calibrate SDSM. We recommend that you do not use the divergence-predictors from HadCM3 and CGCM1 because of this error. Instead, you can use the Y component of the geostrophic wind (v component). Indeed, we can demonstrate that the divergence of the geostrophic wind is reduced or proportional to the v component of the geostrophic wind.

Divergence_Wind_Geostrophic.pdf
Alpert_et_al_QJRMS_1995.pdf

CGCM2 Predictors files

In order to reduce data volume, the global window has been divided into seven smaller windows, with each window encompassing a major land area, with the land-sea boundaries defined according to the CGCM2 land-sea mask. The dimensions of the window are quite similar to those chosen for the HadCM3 predictors but not exactly the same because grid dimensions differ.

The predictor variables are supplied on a grid box by grid box basis in a zip file. Three subdirectories are created during the unzipping process and these subdirectories contain the predictors needed to run SDSM:

NCEP_1961-2001:

This directory contains 41 years of daily predictor data, derived from the NCEP reanalyses, normalized over the 1961-1990 period. These data were interpolated to the same grid as CGCM2 (gaussian grid) before the normalization.

CGCM2A2_1961-2100:

This directory contains 140 years of daily GCM predictor data, derived from the CGCM2 A2(1) experiment, normalized over the 1961-1990 period.

CGCM2B2_1961-2100:

This directory contains 140 years of daily GCM predictor data, derived from the CGCM2 B2(1) experiment, normalized over the 1961-1990 period.

The zipped filename refer to the location of the grid BOX on the global CGCM2 grid, while the BOX number in the X direction refer to longitude, and the BOX number in the Y direction refer to latitude (see CGCM2 Window Limits and Box Numbering for correspondence between BOX numbering and longitude-latitude coordinates). The model grid is uniform along the longitude with grid box size of 3.75° and nearly uniform along the gaussian latitude (approx. 3.75°).

[Note that CGCM2 has also 365 days in each year, i.e. leap years are not included. You will need to make sure that the SDSM settings reflect this.]

For more information on predictors, see Predictors Help.

References

Choux M., (2005): Development of new predictor variables for the statistical downscaling of precipitation. Degree Master of Engineering, Department of Civil Engineering and Applied Mechanics, McGill University. (Dec. 2005).

Flato, G., Boer, G.J., (2001): Warming asymmetry in climate change simulations. Geophysical Research Letters, 28: 195-198.