"What the article shows is that for major evolution to take place species become successful and diversify, i.e. natural diversity. Sure natural selection then creates groups and then defines the new species but without natural diversity, major changes in a species doesn't exist."

I have no idea what that means. Natural selection by definition requires diversity in the form of heritable traits--it's hardly news and it's not what this paper is about.. What the paper is saying is essentially that distance alone is sufficient to isolate conspecific populations sufficiently to permit speciation. This is neither new nor particularly controversial (we've known about ring species for donkeys years, and there is plenty of literature on genetic distances equating long linear distances with short distances across barriers like mountains), except that the authors of this paper have a model of the way a single widespread species can undergo a kind of genetic coalescence at the population level that can lead to speciation, and the model looks promising in that it produces output that fits actual observations reasonably well. This speaks to the number of species found in broad geographic areas with few barriers but has nothing whatsoever to do with their morphological or ecological diversity.

"Prior to the Cambrian age there was a mass extinction this opened up many niches and released the pressures of natural selection and allowed the surviving creature to diversify."

No. Ecological release doesn't diminish the role of Natural Selection at all, it just changes the playing field by removing the incumbant advantage in niches. Generalists are competing with other generalists which encourages specialization *BY NATURAL SELECTION*, rather than competing with specialists in occupied niches, which is a game the generalists can't win.