This is fixing the issue reported in #108 that the barotropic component of zonal wind in southern hemisphere was wrongly computed using the data in the northern hemisphere. Other barotropic quantities computed were not affected.
Thanks @chpolste for figuring out the fix.
