My research at Durham seeks to better understand the connection between magmatism and growth of the central Andean plateau. I will do this by looking at the ages and chemistry of Bolivian lavas and looking for patterns that might reveal why the volcanoes formed in the first place, and what that might mean for plateau development. Here are the volcanoes and some of the main sample localities for the study.
The Serraníá Intersalar, or Intersalar Volcanic Field (IVF).
The lines across the map represent distances above the subducting Nazca slab.
As you can see, the volcanoes extend far enough into the rear arc that they may not be formed by slab dehydration. The fact that there are many more volcanoes further east, really suggests something else is going on. Here are all of my sample sites (red and blue circles) and more easterly volcanoes studied previously (by Davidson, de Silva, Hoke and Lamb)
In the .gif below, three proposed models for rear arc volcanism are illustrated. The first shows dehydration of the slab, which adds volatiles (mainly water) into the mantle wedge (asthenosphere) allowing melts to form. The second is an idea by Hoke and Lamb (2007) that is based on the unusual lithospheric structure beneath the region. In this model, the convecting asthenosphere swells up into the lithospheric gap, causing decompression and melting. This model also requires lateral migration of melt through the thickened crust. The third model was made by the folks who discovered the strange lithospheric structure. Myers, Beck, and Zandt and their colleagues suggested that the removal of the lithosphere that created the gap may have triggered the volcanism in the process.
The whole-rock geochemical data from WSU GeoAnalytical lab is now in. The data are really cool. For starters, we can test the slab dehydration model by looking at incompatible elements such as potassium (K). One would expect that slab contribution would lesson with distance into the continent, resulting in lower degrees of partial melting and higher concentrations of K. This is not apparent from the data.
It seems most likely that the volcanism is related to foundering in some way. The geochemical differences between the arc front and the rear arc are subtle, although Nb shows the most systematic variation with vent location and could be an important indicator of source characteristics.