Mapping the Rivera and Cocos subduction zone

The crust and upper mantle seismic structure beneath southwestern Mexico was investigated using several techniques including teleseismic tomography using 3D raytracing, a joint tomographic inversion of teleseismic and regional data that included relocation of regional seismicity, and a P to S converted wave study. The data used in these studies came from a broadband seismic deployment called MARS. The seismic deployment lasted 1.5 years from January 2006 to June 2007 and the stations covered much of Jalisco and Colima states as well as the western part of Michoacan states. At depth less than 50 km, P-wave receiver function images show a clear dipping slow velocity anomaly above a fast velocity layer. The slow anomaly convertor seen in receiver functions is directly above a fast dipping seismic anomaly seen in regional tomography results. The slow velocity with high Vp/Vs ratio is interpreted as a high pore fluid pressure zone within the upper layer of subducting oceanic crust. Regional seismicity was located using the double difference technique and then relocated in a tomography inversion. The seismicity is located very close to the slow dipping boundary to depths of 30-35 km and thus along the plate interface between the subducted and overlying plate. Deeper events are below the slow layer and thus are intraplate. Receiver function results also show a weaker continental Moho signal above the dipping slab that I interpret as a region of mantle serpentinization in the mantle wedge. Inland of the subduction zone, a clear Moho is observed with a maximum thickness of near 42 km although it thins to near 36 km depth towards the north approaching the Tepic-Zacoalco Rift. Using H-K analysis to examine Vp/Vs ratios in the crust, I find a band of very high Vp/Vs along the Jalisco Volcanic lineament as well as beneath the Michoacan-Guanajuato volcanic field. These observations suggest the continental crust is warm and possibly partially molten over broad areas associated with these two magmatic regions and not just locally beneath the volcanoes. I also found seismicity associated with the Jalisco Volcanic Lineament but it was trenchward of the volcanoes. This may indicate extension in this region is part of the explanation for this magmatic activity. At depths below 100 km, the tomography results show clear fast anomalies, about 0.3 km/s faster than the reference model, dipping to the northeast that I interpret as the subducting Rivera and Cocos plates. Tomography models show that the Rivera slab is dipping much steeper than the Cocos plate at depth. Below 150 km depth, the Rivera plate shows an almost vertical dip supporting the interpretation that the slab has steepened through time beneath Jalisco leading to a coastward migration of young volcanism with mixed geochemical signatures. The location of the young volcanism of the Jalisco Volcanic Lineament is just at the edge of the steeply dipping slab seen in the tomography. The magmatism is thus likely a nascent arc. The models also display evidence of a gap between the Rivera and Cocos plates that increases in width with depth marking the boundary between the two plates. The gap lies just to the west of Colima graben and allows asthenosphere to rise above the plates feeding Colima volcano. Another interesting finding from this study is a possibility of a slab tear along the western edge of the Cocos plate at a depth of about 50 km extending 60 km horizontally. The tear is coincident with a lack of seismicity in this region although there are events below and above the tear.