The parameter f in Psi_SO is really -f. That is, the absolute value of the Coriolis parameter in the S.O. is expected. This should probably be changed; but we are working with this assumption in a number of examples, so changing this will require fixing the examples.