Two alternative models have been proposed for subduction of the Nazca plate beneath central Peru. The first, based principally on seismicity, postulates shallow subduction with a plate dip of 10° to 15°. The second, based on the analysis of converted seismic waves, anelasticity structure, and seismicity at depths less than 90 km, postulates steeper (that is, normal) subduction with a plate dip near 30°. A detailed study of the spatial distribution of earthquakes beneath central Peru shows that seismicity beneath the continent tends to be confined to small, active clusters. One such cluster occurs at intermediate depth about 700 km from the trench and forms a primary basis for the shallow-subduction model. Other subcontinental clusters that occur throughout central Peru at shallower depths can not be related directly to the downgoing plate. Apart from the intermediate cluster, there is little seismic evidence for a shallow seismic plane beneath central Peru; this suggests that the location of that cluster is fortuitous.
Other geophysical evidence, including the pattern of seismicity at depths <90 km and observations of seismic waves converted or reflected at the top of the descending plate, indicates a dip of about 30°. Consonant with that evidence are geologic and geochemical results which show that undersaturated alkalic rocks of eastern Peru were probably related to a subduction zone of 300- to 400-km depth (∼30° dip) as recently as 5 m.y. ago. In the model that presumes steep subduction, the descending plate is nearly aseismic below about 90 km, and the intermediate-depth earthquakes that are observed hundreds of kilometres east of the trench lie within the thick lithospheric block of the South American continent.