Los Andes del Sur fueron creados a través de dos ciclos principales en los cuales la zona de subducción se subhorizontalizó y posteriormente se empinó en diferentes segmentos comenzando en el Cretácico superior. La primer fase contraccional en los Andes del Sur migró desde el borde de placas hacia el interior continental en relación al desarrollo de dos zonas de subducción subhorizontal, una desarrollada entre los 36° y los 39°S y la otra entre los 40° y los 46°S, ambas relacionadas a la expansión oriental del magmatismo del arco volcánico. En el Cretácico más alto hasta el Eoceno, la sección subhorizontal más septentrional siguió acentuándose asociándose así al levantamiento de relieves montañosos lejanos al límite de placas, al tiempo que el sector meridional se empinó rápidamente lo que provocó el colapso extensional de los relieves previos y la efusión de productos volcánicos de intraplaca. Posteriormente, el área entre los 36° y los 44°S se empinó en forma conjunta en tiempos Oligocenos superiores provocando la inyección de astenósfera por debajo de la placa Sudamericana y consecuentemente la formación de grandes plateaus volcánicos y desarrollo de cuencas extensionales superpuestos a los relieves andinos previamente ascendidos. El Mioceno superior está caracterizado por el desarrollo de tres zonas de subducción subhorizontal localizadas entre los 34°30′y los 50°S. En forma similar a los casos anteriores, estas zonas se asociaron a la expansión oriental del magmatismo de arco y a la construcción lateral del relieve andino. Su desarrollo finalizó en el Plioceno al Cuaternario al tiempo del emplazamiento de grandes plateaus volcánicos asociados a extensión que afectó y aun afecta grandes sectores de la Patagonia y Andes Centrales Australes. Finalmente, un segmento entre los 35° y los 39°S podría estar relacionado al desarrollo incipiente de una zona de subducción somera en los últimos 3 Ma. Esta ciclicidad en el desarrollo de zonas de subducción someras en los Andes del Sur, que ha determinado un comportamiento en el cual se han sucedido fases contraccionales seguidas de fases extensionales asociadas a importante volcanismo de intraplaca podría ser la consecuencia de la colisión de dorsales sísmicas contra el margen, inicialmente en el Cretácico superior al Eoceno y posteriormente en el Mioceno superior. Otros factores tales como la colisión de plateaus oceánicos asociados a zonas de fractura altamente serpentinizadas y consecuentemente con alta flotabilidad isostática son también considerados como gatilladores de procesos de subducción someros.
It is generally accepted that the Andes are linked to the subduction (see Glossary) of the plates that conform today, and have conformed in the past, to the Pacific oceanic lithosphere, beneath the South American plate (Fig. 1). This process started some 115 Ma, and accelerated after 90 Ma, when South America started to drift to the west as a consequence of the opening of the South Atlantic Ocean (Somoza & Zaffarana, 2008). The southern Andes were totally created in the absence of collision, contrary to the northern Andes, where deformation and consequent uplift could have had a component derived from accretion of oceanic terranes along the margin (Pindell et al., 2005). The chronology of deformation along the margin shows that subduction of oceanic crust does not, by itself, result in mountain uplift. Other factors are generally involved in the continental shortening that produces mountain relief and consequent exhumation processes; these include an absolute trenchward movement of the South American plate and a relatively small angle of subduction with respect to the horizontal, among other proposed variables (Oncken et al., 2006; Schellart et al., 2007; DeCelles et al., 2009; Ramos, 2009a,b).
Box 1. Glossary
Apatite fission track data technique: Based on the analysis of radiation damage trails (‘fission tracks’) (each fission track is created by spontaneous fission of a single atom of 238U), within apatite grains in a rock, the thermal history of a region can be unravelled. Temperature–time conditions can be correlated with the exhumation history of an area: rocks are transported through different temperature windows during mountain-building processes and consequent denudation, allowing the creation of fission tracks below 120 °C.
Arc-related: Rocks that have been formed from mantle melts derived from the dehydration of a subducted oceanic slab beneath another plate. Arc-related rocks form parallel-to-the-collision (subduction)-zone volcanic chains (volcanic arcs).
Asthenosphere: Section of the mantle that is not taking part in the upper lying tectonic plates. It is placed under the 1250 °C isotherm and has the capability to experience flux over long periods of time (Myr) because of the content of small amounts of melted material.
Bernárdides system: Mountain system sited between 42 and 46°S, separated from the main Andes, disrupting in the middle of the foreland zone.
Denudation: Process by which deep-sited rocks are surface exposed. Upper lying material is mainly removed by fluvial and/or glacial erosion usually during mountain-building stages.
Divisadero Group: Arc-related rocks that were emplaced in the retroarc zone between 41 and 46°S between 130 and 100 Ma.
Granitoids: Rocks derived from a magma (melt) that have been hosted several kilometres beneath the surface, allowing relatively long periods of cooling that lead to classical recognizable textures in which centimetre minerals constitute the whole rock volume. Moreover, they are enriched in SiO2 in comparison with other similarly deep-sited rocks.
Liquiñe-Ofqui fault zone: Series of faults running through the present arc zone for almost 1000 km that accommodate displacements parallel to the plate boundary, imposed by an oblique convergent vector between the two plates (Nazca and South American plates).
Patagonian batholith: 2000-km-long stripe of arc-related intrusive rocks along the western Patagonian margin. Its origin is related to the last 155 Ma of subduction along the margin. Its intrusive character is a result of the high erosion rates produced by Pliocene to Quaternary glacial activity that eroded upper contemporaneous volcanic structures.
Retroarc zone: The area that is located next to the arc zone in the direction of subducted slab deepening (in the direction opposite to the trench).
Slab (oceanic slab): A tectonic plate or part of a tectonic plate composed of oceanic crust, created at a mid-ocean ridge.
Strike–slip displacements: A vertical, near-vertical or dipping fault that separates two blocks that slide horizontally, one respect to the other.
Subduction: Process by which oceanic crust, forming part of a plate, sinks in the asthenosphere because of its higher density, beneath an upper lying plate.
Synextensional: Sedimentary/volcanic sequences that were accumulated at the time and at the site at which the crust was stretching. Faults generated during crustal extension allow the accumulation of thick sedimentary columns (1000–3000 m) in relatively short time periods (3–5 Myr).
Synorogenic depocentres: Sedimentary/volcanic sequences that were accumulated co-evally to mountain-building processes.
Terrane: Crustal piece that is attached to a plate margin after the closure of an oceanic basin. This attachment implies a collision that leads to mountain building.
Trench: Linear depression beneath sea water up to 11 000 m deep where the subducted oceanic plate bends beneath an overriding/upper plate.
The development of the Andes was not a steady process; it started at 115 Ma and has experienced, since then, discrete phases in latest Cretaceous, late Eocene, latest Oligocene to Miocene and late Quaternary times, as inferred in pioneer works (Groeber, 1946, among others). These phases have led to the final shape of the mountain chain (Cobbold & Rossello, 2003; Charrier, Pinto & Rodríguez, 2007) through the stacking of crustal sheets that have provoked a progressive widening of the deformational belt. Moreover, a series of crises to this general model of mountain growth have been proposed, in which wide sectors of the deformational belt could have collapsed, resulting in a retreat of the Andean front (Cazau et al., 1989; Hervéet al., 1995; Jordan et al., 2001; Folguera et al., 2009; Rojas Vera et al., 2010). Widespread volcanism has accompanied these stages of mountain inhibition, producing broad volcanic plateaux from the Pacific coast to hundreds of kilometres in the continental interior (Bermúdez et al., 1993; Muñoz et al., 2000).
Thus, it is clear that the history of the growth of the Andean mountains is more intricate than previously assumed. The mountain relief resulting from the stacking of crustal sheets and volcanism, which can influence the pattern of biotic distribution through changes in topography and patterns of precipitation, has varied, expanding and retreating through time. When was the definitive elevation of the southern Andes achieved? Did a definitive uplift of the mountain chain exist? Or has the chain reached successive configurations which have been eventually destroyed through time? Have there been any differences regarding mountain amplitude between northern and southern areas along the southern Andes? What was the final chronology of uplift? This article explores such questions through a synoptic description of the processes that have dominated the Andean scene in the last 100 Myr, proposing a model for common genesis linked to cyclical variations of the subducted slab geometry through time.
1Subducted slab shallowings (Fig. 2A), which were associated with the eastward redefinition of the slab dehydration zone and, consequently, magma-producing location; this has controlled the stacking of crustal sheets through the thermal weakening of the continental crust that favoured horizontal collapse (Ramos, Cristallini & Pérez, 2002). Consequently, arc-related magmas emplaced over the area of subducted slab dehydration coexist in time and space with mountain-building processes taking place hundreds of kilometres from the oceanic trench.
2Subducted slab steepenings (Fig. 2B) after shallow subduction configurations, in which hot mantle is injected in a broadened asthenospheric wedge. This mantle interacts with water budgets hosted in the base of the lithosphere derived from the dehydration of subducted oceanic lithosphere during the previous shallow subduction setting. This water reservoir lowers mantle melting point, provoking magma production and its emplacement in the continental retroarc zone, at the time that arc-related products associated with ongoing subducted slab dehydration are produced near the oceanic trench as a result of a higher angle of subduction (Fig. 2B).
Fossil shallow to flat subduction episodes have been recognized along the Andes on the basis of the general similitude to present settings (James & Sacks, 1999; Kay, Burns & Copeland, 2006; Kay & Coira, 2009). The Andean system is constructed over three flat subduction zones: the Bucaramanga, Peruvian and Pampean flat slab zones (Fig. 1A). The three share some characteristics, such as the uplift of basement blocks hundreds of kilometres from the trench, emerging as independent mountain systems from the Andean axial zone, and the expansion of arc-related volcanics to the foreland area, producing arc gaps at the hinterland zone (Fig. 2A). These indicators are tracked through the subduction margin, identifying suspected fossil shallow subduction settings (Fig. 1B). Therefore, on this basis, a late Eocene to early Oligocene shallow subduction setting is proposed to have developed beneath the southern Peru and Bolivian regions, whereas, to the south an early Miocene shallow subduction setting would have affected the Puna area (James & Sacks, 1999; Kay & Coira, 2009). Another shallow subduction zone was proposed south of the Pampean flat subduction zone, where arc-related rocks expanded to the east over the foreland area more than 550 km (Fig. 2B) (Kay et al., 2006; Litvak, Forguera & Ramos, 2008). These phenomena have been inferred to have contributed to the construction of broad mountain segments in relatively short periods of 20–30 Myr. These were followed by emplacement of voluminous within-plate volcanic rocks in the foreland area, retreat of arc activity and partial collapse of the emerged mountain structure, leading to basin formation over the fold and thrust belt. The southern Central Andes and Patagonian regions have been the exception in these analyses (Fig. 1B). Their evolution, with the remarkable exception of three proposals explored and discussed in a regional context in this article (Rapela et al., 1988; Suárez and De La Cruz, 2001; Espinoza, Morata & Lagabrielle, 2007), has not been clearly integrated into this general model.
The southern Andes are produced in a deformational belt trapped among four plates (Fig. 3): the South American, Nazca, Antarctica and Scotia plates. The Nazca and Antarctic plates are separated by the Chilean ridge, a mid-ocean ridge that is being subducted at 46°30′S, and that at 14 Ma was subducted at the southernmost tip of South America (Cande & Leslie, 1986). Oblique collision of the Chilean ridge against the subduction border (Chilean trench) produced oblique convergence between the Nazca and South American plates that decoupled the South American plate through the Liquiñe Ofqui fault zone (Lavenu & Cembrano, 1999), running along the present arc front and concentrating strike–slip displacements parallel to the subduction border.
The Southern Andes are characterized by a 2000-km-long belt of granitoids hosted along their axial part, named as the Patagonian batholith. Their ages range from middle Jurassic to Neogene, although they have clear predominance of Late Cretaceous magmatic components (Ramos et al., 1982; Bruce et al., 1991). Most of these magmas are related to the roots of a volcanic arc generated from the dehydration of the successive plates that subducted beneath the South American plate in the last 160 Myr (Bruce et al., 1991; Cingolani et al., 1991). Part of their exhumation has been linked to more than 14 km of denudation during late Miocene to Quaternary times as a result of strike–slip displacements along the arc front, and particularly through the Liquiñe-Ofqui fault zone (Hervéet al., 1993), and strong glacial activity (Thomson et al., 2010).
However, the Patagonian batholith and hosting rocks have an older exhumation story that started in Late Cretaceous times (Thomson et al., 2010). Fission track data reveal a complex pattern of exhumation of the fold and thrust belt (Fig. 3): south of 46°S, where the Chilean ridge is presently colliding against the South American plate, a relatively simple pattern of exhumation appears, in which late Oligocene to late Miocene ages are progressively younger to the east, following the expected mountain lateral growth; north of this latitude, Late Cretaceous to Eocene ages are present at the Pacific Ocean coast, whereas, to the east, neither single pattern nor trend can be clearly isolated. In the following sections, the complexity of Andean growth is discussed on the basis of previous studies, showing that this orographic barrier was uplifted, eroded and extensionally destroyed more than once, constituting a highly variable topographical feature associated with variable volumes of volcanic activity.
THE LATE CRETACEOUS TO EARLY MIOCENE TECTONIC EVOLUTION OF THE SOUTHERN ANDEAN REGION (35–52°S)
As indicated in the Introduction, arc-related products of Late Cretaceous age are represented by the Patagonian batholith that runs parallel to the subduction margin for more than 2000 km (Fig. 4). Between 40 and 48°S, Early to Late Cretaceous magmatic rocks are eastwardly expanded, passing to partly equivalent volcanic rocks of the Divisadero Group (Ploszkiewicz & Ramos, 1977; Ramos et al., 1982). Pankhurst et al. (1999) have shown that the Patagonian batholith at these latitudes grew laterally through the eastward migration of the magmatic activity at 140 Ma, when magma emplacement was at the present Chilean coast, to the present foothills at around 100 Ma. The area of arc migration coincides with the remnants of a broad fold and thrust belt that incorporated sedimentary sequences from the drainage divide area to the far retroarc zone (Fig. 4). The main phase of activity of this deformational belt ended by 90 Ma in the hinterland zone (main Andes) (Baker et al., 1981; Ramos et al., 1982; Suárez and De La Cruz, 2000; Folguera & Iannizzotto, 2004). Then, contractional deformation expanded to the east, uplifting the Bernárdides system previous to 80 Ma, the age of the latest Cretaceous sequences that are covering unconformably these previously deformed rocks (Ferello, 1969; Franchi & Page, 1980).
Equivalently, between 36 and 39°S, Late Cretaceous arc rocks expanded over the eastern slope of the Andes (Llambías & Rapela, 1989; Franchini et al., 2003). This eastward arc migration advanced from the Central Valley in Chile, where the late Early Cretaceous arc front was located (Munizaga, Huete & Hervé, 1985), to the foreland area, where latest Cretaceous to Palaeocene volcanic rocks were hosted (Fig. 4) (Ramos & Folguera, 2005). Then, arc-related activity remained at the foreland area until Eocene times, when it extinguished and retreated (Llambías & Rapela, 1989; Kay et al., 2006; Zamora Valcarce et al., 2009). Mountain-building processes in the foreland zone at these latitudes were associated with the development of the Agrio fold and thrust belt, where fission track ages revealed a 90-Ma exhumation episode previous to the emplacement of latest Cretaceous arc-related products at these longitudes (Zamora Valcarce et al., 2009). Afterwards and similarly, between 39 and 43°30′S, latest Cretaceous to Eocene arc-related series expanded to the east, forming the Pilcaniyeu volcanic belt (González Díaz, 1979; Rapela et al., 1988; Mazzoni et al., 1991). These series represent a more distal position of the arc rear than the described Late Cretaceous expansion represented by the Divisadero Group (Fig. 4).
During an equivalent time period, south of 43°30′S, the foreland area was intruded and partly covered by within-plate volcanic rocks, whereas mountain-building and arc processes in the region stopped (Fig. 4) (Pezzutti & Villar, 1979; Chelotti, 2004; Foix, Paredes & Giacosa, 2008). Contemporaneously and subsequently, Palaeocene to Oligocene extensional depocentres, interfingered with volcanic activity, were developed at the foreland zone (Mazzoni, 1994; Madden et al., 2005; Re et al., 2005; Paredes et al., 2008). The two identified foreland arc incursions in latest Cretaceous times coincide with the two proposed foreland basins in this region (Fig. 4). The northern arc migration zone was developed between 90 and 60 Ma. This process overlapped in time with the development of the Neuquén basin in the foreland area (Fig. 4). Similarly, the eastward arc migration that has been registered between 41 and 48°S coincides with the latitudinal extent of the Chubut Group and mountain-building processes located up to the far foreland area in the western Deseado massif (Homovc & Constantini, 2001) (Fig. 4).
After the latest Cretaceous to Eocene eastward arc migration between 36 and 39°S, voluminous mantle-derived flows erupted from the coastal zone to the far foreland area (Fig. 4). Arc activity remained at the eastern slope of the Andes at around 44 Ma (Franchini et al., 2003; Zamora Valcarce et al., 2009), when it retreated to the present drainage divide area (Suárez & Emparán, 1995; Kay et al., 2006). There, arc-volcanics interfingered with a basin fill accumulated between 27 and 19 Ma in the extensional Cura Mallín basin (Fig. 4) (Jordan et al., 2001; Rojas Vera et al., 2010). Since c. 35 Ma, within-plate eruptions, at the arc-migration zone, have occurred at the coastal area (Muñoz et al., 2000). A volcanic plateau of 300 km in diameter, whose components were gathered in the Palaoco Formation, erupted at the foreland area in a rather restricted time span from approximately 27 to 23 Ma (Ramos & Barbieri, 1989; Kay et al., 2006). Its lifetime, as well as its latitudinal extent, coincides both with the Cura Mallín basin in the hinterland zone and the segment of eastward arc migration until Eocene times in this region (Fig. 4).
To the south, between 40 and 43°30′S, an arc migrated to the east from Late Cretaceous times to Eocene times, forming the Pilcaniyeu magmatic belt (Fig. 4). The El Maitén belt represents the westward retirement of the arc in the region from 35 to 25 Ma (Rapela et al., 1988), at the time when the Somuncura within-plate plateau and the coastal magmatic belt started to grow (Fig. 4B) (Muñoz et al., 2000; Mahlburg Kay et al., 2007).
Between 43°30′ and 44°30′S, arc activity had extinguished in the region with the eruption of 80-Ma volcanics of the Tres Picos Formation and other later Palaeocene to Eocene within-plate sequences (Pezzutti & Villar, 1979; Franchi & Page, 1980; Marshall et al., 1981). This Palaeocene to Eocene non-arc activity expanded to the east over the foreland area and to the south of 44°S along the Andean foothills (Ramos et al., 1982; Ramos & Kay, 1992). Late Oligocene to early Miocene within-plate eruptions took place in a similar longitudinal belt to the Somuncura plateau volcanics and through the same time period, and conformed the Meseta Cuadrada and Kanquel volcanic plateaux (Baker et al., 1981) (Fig. 4). Late Oligocene to early Miocene within-plate eruptions coexisted with foreland extension interfingered with volcanic activity at the Sarmiento basin (Fig. 4) (Mazzoni, 1994; Chelotti, 2004; Paredes et al., 2008). The hinterland zone was also the locus of extensional processes that led to marine sedimentation interfingered with volcanic activity in the Traigúen basin (Hervéet al., 1995), whose latitudinal amplitude coincides with the area of contemporaneous foreland eruptions (Fig. 4).
THE LATE MIOCENE TO QUATERNARY TECTONIC EVOLUTION OF THE SOUTHERN ANDEAN REGION (35–52°S)
Kay et al. (2006) proposed a slab shallowing setting for the 8–4-Ma time period at this latitude on the basis of chemical similarities between arc-related volcanic rocks of the Present Pampean flat slab to the north (27°–33°S) and of the Chachahuén volcanic centre located 500 km from the Chilean trench at 38°S. More recently, Litvak et al. (2008) extended this analysis, finding another seven volcanic centres aligned longitudinally in a stripe, 400 km in length, along the foreland area (34°30′–38°S) (Fig. 5A). Dating revealed that arc expansion started at 17 Ma at the highest Andes, drifting to the San Rafael block at around 14–4 Ma (Fig. 5) (Litvak et al., 2008). A foreland basin (Rio Grande) developed east of the advancing mountain front between 18 and 8 Ma (Silvestro et al., 2005). This was lately cannibalized by the advancing orogenic front between 8 and 6 Ma (Silvestro & Atencio, 2010) that uplifted the San Rafael block at the time of eastward arc expansion at these latitudes (Fig. 5) (Folguera et al., 2009).
After the late Miocene arc expansion that took place in the region and the mountain-building processes that fragmented the foreland area, the Payenia volcanic field and other neighbouring volcanic plateaux erupted (Fig. 5B). These volcanic sequences were linked to extensional phenomena and collapse of the mountain topography (Rojas Vera et al., 2010). This extension started in the southern region after 5.5 Ma and propagated to the Payenia volcanic plateau situated to the north in the last 2 Ma (Fig. 5B). This extension is dated at between 1.8 and 0.01 Ma, and its latitudinal extent coincides with the late Miocene segment of westward arc expansion between 34°30′ and 38°S (Fig. 5) (Folguera et al., 2009). The chemistry of these volcanic sequences shows a mantle within-plate origin at the easternmost sections (Bermúdez et al., 1993; Kay et al., 2006). Eruption of the several volcanic centres that fed this volcanic field took place from normal faults described in early geological works of the area (Fig. 6) (Polanski, 1963; González Díaz, 1964, 1972; Fidalgo, 1973).
To the south, between 40 and 44°S, late Miocene magmatism expanded to the foreland zone with minor intensity relative to the north (Fig. 7). In the hinterland area, between 40 and 42°S, 16–13-Ma granitoids are eastwardly displaced with respect to the present arc front (Fig. 7) (González Díaz, 1982). This magmatic belt broadens to the foreland area, represented by volcanic facies occupying at similar times a latitudinal band between 40 and 44°S (Mazzoni & Benvenuto, 1990; Ré, Geuna & López Martínez; 2000). This eastward arc expansion coexists with mountain-building processes in the area and the development of a foreland basin represented by the Ñirihuau and Collón Cura Formations from 21 to 13 Ma (Fig. 7) (Giacosa & Heredia, 2004). Even further to the south, between 45 and 50°S, other eastward arc expansion has been registered (Fig. 8). Hervéet al. (2007) determined a series of arc-fronts, represented by arc-related granitoid belts, that determined a clear pattern that becomes progressively younger to the east. The last members of this migration are emplaced at the eastern foothills with a maximum expansion at around 47°S (Fig. 7) (Nullo, 1978; Ramos, 1982; Espinoza et al., 2007). This eastward arc migration coexisted with the development of a foreland basin between 19 and 15 Ma (Marshall & Salinas, 1990) and with the contractional reactivation of the Bernárdides system and the Deseado massif at the foreland area (Fig. 7) (Ramos, 2002).
These two other late Miocene eastward arc expansions were followed by within-plate eruptions in the foreland area (Fig. 7). Lagabrielle et al. (2007) described synextensional emplacement of these products at 47°S, potentially linked to the development of a slab window derived from the collision of the Chilean ridge.
Although Quaternary times at the retroarc area are dominated by extensional processes associated with within-plate volcanic activity (Fig. 7), mountain-building processes have affected differentially the northern described sector. Krawczyk et al. (2006), using seismic refraction, identified a shallower sector of the Nazca plate at 37°S on the order of 8° with respect to neighbouring sectors. This shallower setting coincides with the collision of the Mocha oceanic plateau, formed by highly serpentinized oceanic rocks sheared at the Mocha fracture zone (Fig. 8) (Contreras-Reyes, 2008). This plateau collided against the Chilean trench between 36 and 39°S in the last 4 Ma, producing mountain-building processes that acted initially at the Chilean coast (Glodny et al., 2008) and propagated into the retroarc area (Fig. 8) (Folguera & Ramos, 2009). Moreover, the arc front has eastwardly migrated in this segment by some 40 km, indicating that slab shallowing was the cause of the renewed wave of contractional deformation at these latitudes in late Pliocene to Quaternary times (Folguera & Ramos, 2009).
We have discussed the occurrence of slab shallowings associated with arc expansions and mountain-building processes, and their posterior steepening triggering asthenospheric injection and extensional settings. These occurrences can be partly explained by the available configurations of plate interactions through time along the Chilean subduction zone. Cande & Leslie (1986) have suggested that interaction between the mid-ocean ridge that separated the Aluk and Farallones plates and the subduction border occurred at 52 Ma at around 42°S (Fig. 9). At 42 Ma, based on these reconstructions, the mid-ocean ridge would have abandoned the triple junction point with the South American plate (Cande & Leslie, 1986). This process implies the subduction of progressively younger oceanic crust during latest Cretaceous to Palaeocene times at the time of Late Cretaceous to Eocene magmatic expansion in the southern Andes (Fig. 9). Slab shallowings could have been produced by hot, and therefore buoyant, oceanic crust attached to the Aluk/Farallones ridge. Direct interaction of this ridge south of 42°S would have produced the early abandonment of the slab shallowing at these latitudes that triggered the injection of within-plate materials and extension in Palaeocene times (Fig. 9). After subduction of the Aluk/Farallones ridge, oceanic crust became progressively older and colder, sinking more intensely and triggering a slab steepening setting in the foreland area, as depicted in Figure 9.
Similarly, the southernmost proposed late Miocene shallow subduction zone (Espinoza et al., 2007) could have been produced by the progressive approach of the Chilean ridge that separates the Nazca and Antarctic plates (Cande & Leslie, 1986) (Fig. 9C–E). As indicated, arc expansion occurred previous to ridge docking when young and buoyant crust subducted beneath the South American plate. As in the previous case, final subduction of the ridge and progressively older and colder oceanic crust would have produced a slab steepening setting that triggered asthenospheric injection, within-plate eruptions and localized foreland extension. A substantial part of these eruptions could have been produced by the slab window, itself created by the Chilean ridge subduction (Ramos & Kay, 1992) (Fig. 9E).
The Payenia shallow subduction segment (34°30′–38°S) shares its timing with the present Pampean flat subduction zone to the north (Ramos et al., 2002). In both zones, eastward arc expansions started at around 17 Ma, and reached the foreland area at around 10 Ma. They differentiated from each other over the last 5–4 Ma, when the Payenia shallow subduction zone became a steepening subduction setting, whereas, to the north, arc-related products advanced another 150–200 km to the foreland area (Ramos et al., 2002). One explanation for the Payenia shallow subduction zone is that both zones between 27 and 38°S constituted one single shallow subduction zone from approximately 17 to 5 Ma, and then the southern segment steepened. These two shallowings differ, however, in one important way: although the Pampean shallowing constitutes a magmatic expansion in which the arc front (westernmost arc magmatism) redefined eastwardly, the Payenia zone is linked to an arc front rather stationary at the Chilean slope of the Andes at the time of eastward arc expansion. In contrast, incipient and mild shallowing of the Nazca plate between 35 and 39°S in the last 4 Ma has been linked to subduction of highly serpentinized oceanic crust. Because the collision of fracture zones and, consequently, subduction of serpentinized oceanic crust have been common processes along the Andean subduction zone, these shallowings could be more frequent than previously assumed. Nevertheless, the shallowings produced in the previous cases were associated with the approximation of an oceanic ridge and, consequently, their deformational and magmatic effects are larger, leaving a clearer imprint in the geological record.
Therefore, a variety of processes seem to have induced shallowing subduction, a mechanism that created favourable conditions for lateral mountain growth at different times. In order of importance, comparing deformational effects produced at the plate margin, those linked to the subduction of nonseismic ridges are the most important in relation to mountain-building processes. The three Present flat configurations would be linked to this phenomenon. These are followed by shallowings associated with approaching seismic ridges. The duration and size of both shallow subduction configurations are highly variable, depending on the geometry of collision (basically the angle between the oceanic ridge and continental margin) and the length of the ridge. The relative importance of shallowings in association with the collision of fracture zones, and the consequent deformational effects, are still poorly understood, but seem to be minor.
The chronology of Andean uplift and the associated volcanism described above have important consequences for palaeogeography. One important inference for phylogeography is that no definitive elevation of the Andes existed along the margin in the last 100 Myr. The present topography is a feature inherited from the 17–5-Ma slab shallowings that affected the region, partly destroyed by 5–0-Ma extensional phenomena. Late Cretaceous to Eocene slab shallowings also produced a topography of similar amplitude to that seen in the Neogene. Late Oligocene to early Miocene extensional relaxation of the fold and thrust belts could have led to substantial losses of mountain relief, although the real magnitude of this topographic collapse is somehow speculative. The differences in mountain amplitude between northern and southern segments depend strongly on the time period considered. During Late Cretaceous times, the segments between 36–39°S and 42–46°S developed probably the highest and widest mountains across the margin. This follows from the fact that arc expansion was at a maximum and therefore deformational features advanced differentially to the east, creating a broad and consequently high mountain barrier. This pattern changed during the Neogene, as reflected by the current topographic variations in the Andes, which show two areas nucleated at 35–38°S and 45–50°S, where arc expansion was at a maximum and, consequently, the Andes grew exceptionally wide and therefore high.
Finally, mountain systems east of the Andean axis, emergent at the foreland area, have been constructed through the specific processes described corresponding to shallow subduction configurations. From south to north, the western Deseado massif is part of the frontal fold and thrust belt first uplifted in Late Cretaceous times and, more recently, in late Miocene times, in front of two shallow subduction zones. Similarly, to the north, the Bernárdides system, the western north Patagonian massif and the Neuquén mountain foothills have been constructed through these two stages and face two shallow subduction settings as well. To the north, the San Rafael block seems to be a younger feature because of the late Miocene contractional stage in the region. This stage was associated with a shallow subduction zone. Late Oligocene–early Miocene (first) and Pliocene to Quaternary (later) extensional stages have probably lowered these topographic barriers, as clearly demonstrated for the San Rafael block.
We thank Editors of this volume, Daniel Ruzzante and Jorge Rabassa. Besides, the authors kindly acknowledge many researchers with whom these topics were discussed during the last years: Pancho Hervé, Constantino Mpodozis, Reynaldo Charrier, Pirzio Godoy, José Antonio Naranjo, Thierry Sempere, Peter Cobbold, César Arriagada, Rubén Somoza, Olivier Galland, Daniel Melnick, Alfonso Encinas and Andrés Tassara. This is the R 31 work of the Instituto de Estudios Andinos Don Pablo Groeber.