I've looked at lots of EM results for spiral inductors from many different tools and I have never seen a result like that. I will suggest that this might be a numerical artifact, possibly because you are using a volume mesher (I am guessing this is the case). You should try with a another EM tool, preferably several of them, and compare results. Be sure to use a planar tool (I recommend Sonnet for shielded and Momentum for unshielded, I work for Sonnet). Using a volume mesher (if that is what you are doing) for a planar circuit is not a good idea. Do not analyze with a ground ring unless you are going to use the same ground ring when you include it in the final circuit (probably a good idea).