Hey Dale!
For most turbulent cases the range is around 1e3 to 1e4 (shear layers, fully-developed duct flows etc.) but from what I have read is that usually at free-stream boundaries of most external flow the EVR is between 1 and 10. However you are correct here Dale that this ratio can be used to see where the mesh has to be resolved better. I would have to read this up again and do a small research - forgot the exact procedure on such cases to be honest 
By the way, for k and epsilon, have you considered this post from @1318980 (Darren): Defining Turbulent Boundary Conditions ?
Best,
Jousef