10th Conference on Mesoscale Processes

Tuesday, 24 June 2003
The New Version of the Canadian Operational GEM Regional Mesoscale Model
Jocelyn Mailhot, MSC, Dorval, QC, Canada; and S. Bélair, A. Tremblay, L. Lefaivre, B. Bilodeau, A. Glazer, A. Patoine, and D. Talbot
A new mesoscale version of the GEM regional model in operation in Canada for numerical weather prediction has been under development at the Canadian Meteorological Centre (CMC) for some time and is now in the process of being implemented at CMC.

The main changes to the modeling system comprise an increased resolution both in the horizontal (15 km instead of 24 km) and the vertical (43 levels instead of 28) and improvements to almost every aspects of the physics package. These include an improved formulation of the boundary layer to represent clouds with an unified moist turbulence approach, the Kain-Fritsch deep convection scheme, and the Tremblay mixed-phase condensation scheme with explicit microphysics, together with revisions to the cloud radiative optical properties. The effect of low-level blocking due to subgrid-scale orography based on the Lott-Miller scheme has also been included.

The performance of the new mesoscale modeling system has been compared against the current operational model for two 6-week cycles during winter and summer 2002 based on the North American radiosondes. Significant improvements are found in the bias and RMS errors of winds and temperatures. Objective precipitation scores also show significant improvements in winter for the bias and threat scores in almost all precipitation categories, while the scores are more similar during summer.

Supplementary URL: