On the baroclinic vortex instability: the scattering problem