For permission to copy, contact [email protected]© 2005 Geological Society of America 67ABSTRACTThe past ~25 m.y. of geologic history in the northern ~300 km (~33°–36°S) of the Andean Southern Volcanic Zone has seen waxing and waning magmatic production rates and episodic eastward relocation of arc segments accompanied by abrupt chemical changes in the magmas. These changes can be linked to episodes of crustal thickening at times of backarc thrusting and to peaks of subduc-tion erosion of forearc crust and mantle lithosphere at times of frontal-arc migration to the east. The magmatic-tectonic coupling is well seen in the history—enhanced by 28 new K-Ar ages, >160 major and trace ele-ment analyses, and Sr, Nd, and Pb isotope analyses—of a west to east transect through the El Teniente copper district near 34°S. The temporal trends in magmatic chemistry in this transect are like the well-documented south to north trends in Pleistocene to Holo-cene volcanic centers of the Southern Volca-nic Zone, and both can be linked to the same events. The magmatic changes require differ-ences in magma source regions as shown by isotopic data and in depths of crustal magma generation/fractionation as shown by pres-sure-sensitive trace element distributions. “Adakitic” magmas in the region are attrib-uted to a combination of melting the base of thickened lower crust and crust entering the mantle through subduction erosion. Subduc-tion erosion is argued to peak in episodes of frontal-arc migration at ca. 19–16 Ma and ca. 7–4 Ma. The combined effects of crustal shortening and forearc truncation in the past 20 m.y. near 34°S have led to the loss of ~170 km of crustal width. The timing and arc length over which these events occurred show that subduction of the Juan Fernández Ridge on the Nazca plate cannot have been the major driving force. The history of the region shows the importance of non-steady-state processes in arc-magma production and the necessity of studying arc systems over millions, not tens of thousands, of years.Keywords: Andes, Southern Volcanic Zone, magmatism, subduction erosion, thick crust, geochemistry, isotopes, Miocene.INTRODUCTIONThe Andean Southern Volcanic Zone (SVZ, Fig. 1) with its relatively simple subduction geometry, nearly constant arc-trench gap, and systematic along-strike chemical changes has been the subject of a series of classical stud-ies aimed at understanding the relationship of Pleistocene to Holocene SVZ magmas to con-tinental-margin tectonics (e.g., López-Escobar et al., 1977; Hickey et al., 1986; Futa and Stern, 1988; Hildreth and Moorbath, 1988; Stern, 1991; Tormey et al., 1991, 1995; Dungan et al., 2001). The pronounced south to north chemical variations identifi ed in these magmas have been attributed to increasing crustal thickness and changing lithology and age of arc basement (e.g., Hildreth and Moorbath, 1988) and/or to variable contributions from subducted sedi-ment and crust added to the mantle wedge by subduction erosion (Stern, 1991). Differences in percentage of mantle melting have also been argued to play a role (e.g., López-Escobar et al., 1977; Tormey et al., 1991; Stern, 1991). A gen-eral consensus has not been reached on the rela-tive importance of these processes, as sources of continental crustal contaminants are diffi cult to differentiate on geochemical grounds.The role of forearc subduction erosion—the process in which crust is plucked or abraded from the hanging wall above the subducting slab and dragged down under the arc where it could become a component in arc magmas—has long been debated along the Chilean margin (e.g., Rutland, 1971; von Huene and Scholl, 1991; Stern, 1991). Evidence for subduction erosion comes from map patterns of Paleozoic and Mesozoic rocks (Fig. 1) that suggest that over several hundred million years, continental crust has been removed from along the Chilean coast. Quantitatively, the removal rate has been argued to be comparable to that of arc-magma produc-tion and sediment subduction (e.g., von Huene and Scholl, 1991). What has been less clear is when subduction erosion occurred, whether the process is continuous or episodic, and whether crustal and lithospheric mantle components from subduction erosion actually make their way into arc-magma sources.Relevant to debates on the sources of SVZ arc magmas and the roles of subduction ero-sion and crustal thickening are early Miocene to Holocene magmatic changes along the mar-gin. These temporal changes have been little addressed as only scanty chemical data have been available on Tertiary magmatic rocks beneath and west of the SVZ. Geochemical studies have been impeded by widespread moderate to severe secondary hydrothermal alteration and uncertainties about the ages of the magmatic rocks (e.g., Vergara et al., 1988). With the exception of a few, mostly recon-naissance studies (e.g., López-Escobar et al., 1997, Vergara et al., 1999, Yáñez et al., 2002), available chemical and isotopic data have been from host rocks of the giant copper deposits between 32°S and 34°S (Pelambres-Pachon, Episodic arc migration, crustal thickening, subduction erosion, and magmatism in the south-central AndesSuzanne Mahlburg Kay†INSTOC, Snee Hall, Cornell University, Ithaca, New York 14853, USAEstanislao Godoy‡SERNAGEOMIN, Casilla 10465, Santiago, ChileAndrew Kurtz§Department of Earth Sciences, Boston University, Boston, Massachusetts 02215, USA†E-mail: [email protected].‡E-mail: [email protected].§E-mail: [email protected] Bulletin; January/February 2005; v. 117; no. 1/2; p. 67–88; doi: 10.1130/B25431.1; 10 fi gures; 4 tables; Data Repository item 2005021.KAY et al.68 Geological Society of America Bulletin, January/February 2005Figure 1. Map of Andes from north of lat 33°S to lat 38°S showing distribution of late Oligocene/early Miocene and middle to late Miocene magmatic units, the SVZ volcanic centers, the generalized extent of forearc Paleozoic and Mesozoic (mainly Cretaceous) outcrops, the Fron-tal Cordillera where the upper Paleozoic to lower Mesozoic Choiyoi Group is exposed, and the locations of the backarc Aconcagua and Mal-argue fold-thrust belts (FTB). Offsets in the modern volcanic line are marked by the northwest-trending dashed lines at ~36.2°S and 34.5°S that separate the SVZ into the northern SVZ, transitional SVZ (Paloma to Tatara–San Pedro), and southern SVZ segments. The El Teniente region within the heavy