P1: MBL/rsk P2: MBL March 21, 1997 19:22 Annual Reviews AR029-05 AR29-05 Annu. Rev. Earth Planet. Sci. 1997. 25:139���74 Copyright c 1997 by Annual Reviews Inc. All rights reserved THE EVOLUTION OF THE ALTIPLANO-PUNA PLATEAU OF THE CENTRAL ANDES Richard W. Allmendinger, Teresa E. Jordan, Suzanne M. Kay, and Bryan L. Isacks Department of Geological Sciences and Institute for the Study of the Continents, Cornell University, Ithaca, New York 14853-1504 e-mail: rwa1@cornell.edu KEY WORDS: South America, continental plateau, uplift, timing, magmatism ABSTRACT The enigma of continental plateaus formed in the absence of continental collision is embodied by the Altiplano-Puna, which stretches for 1800 km along the Central Andes and attains a width of 350���400 km. The plateau correlates spatially and temporally with Andean arc magmatism, but it was uplifted primarily because of crustal thickening produced by horizontal shortening of a thermally softened lithosphere. Nonetheless, known shortening at the surface accounts for only 70��� 80% of the observed crustal thickening, suggesting that magmatic addition and other processes such as lithospheric thinning, upper mantle hydration, or tectonic underplating may contribute significantly to thickening. Uplift in the region of the Altiplano began around 25 Ma, coincident with increased convergence rate and inferred shallowing of subduction uplift in the Puna commenced 5���10 million years later. INTRODUCTION The Altiplano-Puna plateau of the Central Andes (Figure 1) is the highest plateau in the world associated with abundant arc magmatism, and it is second only to Tibet in height and extent. Yet, this remarkable feature was uplifted in the absence of continental collision or terrane accretion in fact, material has been removed from the continental margin during and prior to plateau uplift. Because of its obvious association with Andean magmatism, the plateau was originally thought to be a product of magmatic processes (James 1971b, 139 0084-6597/97/0515-0139$08.00

P1: MBL/rsk P2: MBL January 9, 1998 19:15 Annual Reviews AR029-05c AR29-05 140 ALLMENDINGER ET AL 150 250 300 100 575 350 50 Santa Cruz La Paz Arica Lima Cuzco Antofagasta Salta Bolivia Argentina 25�� 20�� 15�� 65�� 70�� 30�� 75�� Sierras Pampeanas PC SB Chile Altiplano Per�� Puna S u b a n d Figure 1 Location map showing the extent of the high plateau of the Central Andes. Dark gray shows area above 3 km elevation the plateau is defined by the wide area above 3 km between 13 and 27���S. The light gray provinces east of the high topography are thin-skinned thrust belts in the Subandean ranges of Bolivia and the Precordillera (PC) in Argentina. The Sierras Pampeanas and Santa B�� arbara System (SB) are thick-skinned foreland provinces. Thin curves are contours of depth to the Wadati-Benioff zone in kilometers from Cahill & Isacks (1992). The hachured zone trending NW-SE across the Argentine-Bolivian border corresponds to a variety of lateral change in Andean and pre-Andean features and is taken here to be the boundary between the Altiplano and Puna.

P1: MBL/rsk P2: MBL March 21, 1997 19:22 Annual Reviews AR029-05 AR29-05 ALTIPLANO-PUNA, CENTRAL ANDES 141 Reymer & Schubert 1984, Thorpe et al 1981). However, analyses of the plateau topography and structures on the eastern flank of the plateau carried out during the 1980s resulted in the conclusion that crustal shortening could produce most, if not all, of the required crustal thickening and that thickening, combined with lithospheric thinning, could account for the plateau elevations (Isacks 1988, Roeder 1988, Roeder & Chamberlain 1995, Sheffels 1990). Here, we review these arguments, as well as more recent results that appear to show that shortening may not be able to account for all of the crustal thickening. The central Andean plateau must be viewed not just in terms of volumes and magnitudes, but also in light of its evolution. In this review, we focus on the temporal and spatial evolution of the plateau: when it began to lift up and how it varies laterally, as well as the relative importance of magmatism, crustal shortening, and lithospheric thinning. The plateau is composed of two distinct parts: the Altiplano of Bolivia and the Puna of northwest Argentina and adjoining parts of Chile. These areas differ in topography, magmatism, and lithospheric structure, and illustrate the range of conditions under which a continental plateau can develop in a noncollisional orogen. The data that we review here supports and refines Isacks��� (1988) two-stage model for the development of the plateau. Stage 1 uplift began around 25 Ma in the Altiplano segment and between 15 and 20 Ma in the Puna segment, when an episode of low-angle to, locally, nearly flat subduction (Coira et al 1993, Kay et al 1995) thinned and thermally softened the lithosphere underlying the area that was to become the plateau. Shortening ceased in the Altiplano and shifted eastward (Stage 2) beginning between 12 and 6 Ma, but shortening continued in the Puna until 1���2 Ma. PHYSICAL DESCRIPTION OF THE PLATEAU AND RELATED FEATURES A convenient definition of the high plateau of the Central Andes is provided by the notable broadening of the area above the 3-km elevation contour (Figure 1). Defined this way, the high plateau of the Central Andes stretches 1800 km along the backbone of the range, from southern Peru to northern Argentina, and varies between 350 and 400 km in width. This definition of the plateau, which follows that of Isacks (1988), is considerably broader than the more common association of the plateau with the internally draining basins of the Altiplano and Puna. Plate Geometry The geometry of the Nazca Plate beneath South America is well known (Barazangi & Isacks 1976, Bevis & Isacks 1984, Cahill & Isacks 1992,

P1: MBL/rsk P2: MBL March 21, 1997 19:22 Annual Reviews AR029-05 AR29-05 142 ALLMENDINGER ET AL Hasegawa & Sacks 1981, Stauder 1975). Currently, the plateau correlates with a 30������east-dipping segment of the subducted Nazca Plate (Figure 1). To the north and south, where the mountain belt narrows considerably, the subducted plate shallows and is nearly horizontal. Post-Pliocene volcanism follows this correlation: It is absent where the plate is nearly flat and well developed in the plateau where the plate is steeper. The distribution of Neogene volcanism is virtually identical to the spatial extent of the plateau, both latitudinally and longitudinally. The subducted plate geometry differs markedly beneath the northern and southern ends of the plateau (Figure 1). Beneath southern Peru, there is a marked bend in the subducted plate. To the south beneath the Puna, however, the subducted plate gradually shoals between 24 and 30���S. In this zone of shoaling, there is a notable gap in Wadati-Benioff zone earthquakes between 25 and 27���S (Cahill & Isacks 1992). This gap could be an artifact of the short sampling interval of the instrument record, or it could reflect first order, lithospheric scale processes. Contours of depth to the Wadati-Benioff zone project smoothly across the gap, and ray-path modeling and studies of seismic wave attenuation (Whitman et al 1992) indicate that the subducted Nazca plate is present across this earthquake gap. Morphology The availability of regionally consistent topographic data incorporated into dig- ital elevation models has revolutionized the study of modern mountain belts and provides considerable insight into the tectonics of the Central Andes (Figure 2). In this largely arid region, the effects of late Cenozoic tectonics and magmatism on topography have not been obliterated by erosion. Isacks (1988) showed that the average elevation of the plateau between 13 and 29���S is 3.65 km, and he interpreted the 250���300 km wide area of internal drainage in the plateau be- tween 15 and 27���S as evidence of a young age of uplift. The smooth western flank of the Central Andes contrasts strongly with the rough topography on the eastern flank (Isacks 1988). The Puna has an average elevation nearly a kilometer higher than the Altiplano (Figure 3), which has been attributed to greater thinning of the lithosphere beneath the Puna (Whitman et al 1996). The intimate connection between plate motions, mountain belt topography, and the geometry of the subducted Nazca Plate was demonstrated clearly by Gephart���s (1994) analysis of the Isacks topographic data set. He showed that the topography of the Central Andes and the underlying Wadati-Benioff zone is remarkably symmetric about a vertical plane (approximately at the Arica bend) whose pole is oriented about 63���N 113���W. The symmetry axis coincides with the Nazca���South America finite pole of rotation for the period between 36 and 20 Ma the symmetry plane is closely coincident with the Euler equator

P1: MBL/rsk P2: MBL March 21, 1997 19:22 Annual Reviews AR029-05 AR29-05 ALTIPLANO-PUNA, CENTRAL ANDES 143 Altiplano Basin Atacama Basin Puna Santa B��rbara System Beni Basin Sierras Pampeanas Figure _____ Allmendinger et al. Altiplano Basin Atacama Basin Puna Altiplano Basin Atacama Basin Puna Altiplano Basin Atacama Basin Puna Altiplano Basin Atacama Basin Puna Figure 2 Shaded relief map showing the topography of the Central Andes, based on the 1-km DEM of the Defense Mapping Agency. The Altiplano basin is the extremely flat area in the center of the image between 17 and 21���S. The image highlights the differences between the Altiplano and Puna.