70 George L. Mellor, Sirpa M. Hăkkinen, Tai Ezer and Richard C. Patchen
A
o !I 10 1) 10 CI 30 :tb 010 46 $O 6& 10 6&
.... m' ..
c
Fig.4.8 Same as Fig. 4.5, but for the low viscosity case with AM =10 3 m 2 s -' .
similar to the rms velocity errors. The corresponding errors in the z - level case
were, of course, nil.
After subtraction of the area averaged density, the errors in the sigma models
were also nil. Of course, after the model has run with wind forcing and baroclinic
density structure has formed, there must exist some quite small but hard-to-evaluate sigma errors.
However, it is apparent from our results that, with wind forcing, errors in the z -
level model, may be much larger than in the cr grid, in particular, in shallow regions
A
o !I 10 1) 10 CI 30 :tb 010 46 $O 6& 10 6&
.... m' ..
c
Fig.4.8 Same as Fig. 4.5, but for the low viscosity case with AM =10 3 m 2 s -' .
similar to the rms velocity errors. The corresponding errors in the z - level case
were, of course, nil.
After subtraction of the area averaged density, the errors in the sigma models
were also nil. Of course, after the model has run with wind forcing and baroclinic
density structure has formed, there must exist some quite small but hard-to-evaluate sigma errors.
However, it is apparent from our results that, with wind forcing, errors in the z -
level model, may be much larger than in the cr grid, in particular, in shallow regions