More tectonics of the Italian earthquake

The estimated death toll from last night's M 6.3 earthquake in Italy is currently 150, with 10's of thousands of people left homeless. My thoughts are with the people there, especially those still searching for their loved ones.

As you can see from the USGS map and moment tensor (and from Highly Allochthonous, who posted a great explanation while I was on kid duty), the earthquake occurred on a normal fault associated with the collision of Africa and Europe...

At this point, if you've taken an intro geology class, you're probably shaking your head, wondering if I've made a mistake reading those moment tensor beachball diagrams, or interpreting the tectonic setting. (And if you're a plate tectonic skeptic. you're probably jumping up and down, convinced that you've proved, once and for all, that subduction doesn't exist.)

For everyone else, let me take a step back and explain.

The cartoon version of plate tectonics goes like this. Plates pull apart at mid-ocean ridges and come together at subduction zones. (And then there's California, reminding us that there's a third dimension, and plates can also slide past one another.) The cartoon version of faults is similar: normal faults let the crust stretch horizontally, thrust (or reverse) faults let the crust shorten horizontally, and strike-slip faults let rocks slide past one another (such as in California). So one might expect that mid-ocean ridges (and continental rifts) would have normal faults, and subduction zones (and continent-continent collisions) would have thrust faults.

Fig. 1. In intro geology classes, I usually talk about the subducting plate moving toward (and underneath) the overriding plate. I lie. Kind of.

There's a subduction zone in Italy. It's responsible for the Appenine Mountains.

Except... well, the Appenines are currently stretching, in the direction in which they originally shortened.

It's actually not that uncommon to see normal faults around subduction zones. A lot of subduction zones have spreading near or behind the chain of volcanoes - "back-arc" spreading. (The best example is probably in the Mariana Islands in the Pacific.) If the subducting plate sinks faster than the top plate moves towards it, the top plate tends to pull apart in the hottest, weakest part, near the chain of volcanoes.

Fig. 2. Slab rollback, the more common way that subduction zones work. The subducting plate falls into the mantle, and drags part of the overriding plate with it. But the overriding plate doesn't move fast enough, so it gets pulled apart.

As Chris shows, there's evidence for that process (slab rollback) in Italy, in all directions.

Fig. 3. This map belongs to Chris Rowan. Go to his blog and see his explanations, because they're good.


There are other possible causes for normal faulting around continental collisions. Amongst geologists who work in the Himalayas, there are ongoing debates over the role of "channel flow" (in which part of the deep crust gets squeezed out along two shear zones), extensional collapse (from building mountains too high to be supported by a hot, weak lower crust - I've heard this referred to as "Camembert tectonics"), and lithospheric delamination (in which the mantle lithosphere beneath the collision either peels off or comes off in some kind of blob that sinks into the mantle). The ideas that were inspired by Himalayan tectonics deal specifically with things that happen in continental collisions. They could be applied to the Alps, but the Appenines are smaller and not as hot or thick.

One possibility is that the subduction zone east of Italy, in the Adriatic Sea, has rolled back, as in the model that Chris described. But then things got more complicated (in part, because there's a big piece of continental lithosphere sitting just on the other side of the Adriatic Sea). Parts of the subducting plate tore off and sunk into the mantle, leaving hotter rock to rise up in their place. In some places, however, the subducting plate is still attached, and still falls, and drags on the overriding crust. In those places, the whole rollback process gets concentrated, tearing apart the overriding crust.

Fig. 4. Parts of the subducting slab get detached and sink into the mantle, letting hot rock rise in its place. Rollback becomes concentrated in the areas where the slab is still attached.


And that might explain the earthquake that devastated L'Aquila.

Maybe.

Other posts in the geoblogosphere (in reverse chronological order):

Arrowsmith Blog
Highly Allochthonous
cryology & co.
Ontario geo-fish
Seismo Blog
goodSchist
olelog
Dave's Landslide Blog

Selected sources on extension in collisional mountain belts:

Dewey, J. F., 1988, Extensional collapse of orogens: Tectonics, v. 7, p. 1123-1139.

Beaumont, C., Jamieson, R.A., Nguyen, M.H., and Lee, B, 2001, Himalayan tectonics explained by extrusion of a low-viscosity crustal channel coupled to focused surface denudation: Nature, v. 414, p. 738-743.
Vanderhaeghe, Olivier, and Teyssier, Christian, 2001, Partial melting and flow of orogens: Tectonophysics, v. 342, p. 451-472.
Wortel, M.J.R., and Spakman, W., , 2000, Subduction and slab detachment in the Mediterranean-Carpathian region: Science, v. 290, p. 1910-1917.