It does seem logical to ask the question as to whether the Canadian could possibly be correct when you frame it with the "this is a northern stream issue, northern stream vorts originate in the main domain of the canadian model, so shouldn't it be able to resolve those better" thinking.
Is there any evidence to suggest that that model handles the northern stream better?
And how are 500 verification scores created for global models? Can a model handle certain regions better, some worse and that affect its scores? Or no?