Totally desagree. It's more likely this has been listed as Southwest Asian than Mediterranean. It correlates both runs (K=10 and K=12), where this admixture (Southwest Asian) is also measurable in Iberians and other ethnic groups. At least, the Mediterranean component hasn't increased more than 2% in the spanish average in comparison with the previous run, and it's listed even far from non European groups.That means it should even get lower, not higher.
If you see some differences from a run to another it's just due to the number of components (10 before, now 12). None of the clusters mean exactlty the same, and that's more obvious checking East and West Euro. Those two have increased more the averges, so it's plain impossible that Mediterranean include such influences and the others nothing. Not reasonable.
The E-M78 map from Oxford Journals looks quite ilustrative:
http://mbe.oxfordjournals.org/content/24/6/1300/F2.large.jpg