Influence of depositional load on the development of a shortcut fault system during the inversion of an extensional basin: The Eocene-Oligocene Abanico Basin case, central Chile Andes (33°-35°S)

The current paper analyzes the evolution of the Abanico Basin, in the Chilean Principal Cordillera in central Chile (33o-35oS). According to previous studies, the basin has been affected by two main deformational episodes: the first, related to extension, and the second, to partial basin inversion. Deposits of the Abanico Formation, EoceneOligocene, and the Farellones Formation, Early-Middle Miocene, represent these two deformational episodes, respectively. Studies of the basin deposits and the structural features show that the basin was asymmetric and that it developed two main depocenters. The eastern depocenter is deeper and was controlled by a west-dipping fault system (El Diablo Fault System) that formed its eastern border. We propose that the geometry of the basin and the magnitude of inversion were strongly influenced by the load generated by the thick volcanic and volcaniclastic pile that accumulated in the eastern depocenter. Through analogue modeling, we interpret that the Aconcagua fold-thrust belt was triggered at ~16 Ma by a shortcut thrust rooted in the El Diablo Fault System, which during inversion was blocked because of the high load exerted by the deposits of the Abanico and Farellones formations. In addition, the stratified Mesozoic succession that formed the eastern border of the basin allowed the propagation of the shortcut thrust along the ductile Oxfordian gypsum layer. This would have determined the east vergence of the fold-thrust belt. On the western side of the basin, the absence of weak layers prevented development of a west-vergent fold-thrust belt.


General outline
The Andean Cordillera extends for more than 8,000 km along the entire South American western continental margin (Fig. 1a) (e.g., Dewey and Bird, 1970). The morphological features, as well as the tectonic and paleogeographical evolution of the Andean Range, can be used to subdivide it into three major segments: Northern Andes (8ºN-3ºS), Central Andes (3°-47°S) and Southern Andes (47°-56°S) (Gansser, 1973). The Andean Cordillera derives from crustal shortening, thickening and magmatic addition along the continental margin subducted by the Nazca plate (e.g., Isacks, 1988;Ramos, 1988;Allmendinger et al., 1997). In general, the structural style of the southern Central Andes, between 18º and 40ºS, is considered to be direct evidence of subduction activity. Magmatic evidence indicates that subduction has been continuous from the Early Jurassic to the present (Andean Tectonic Cycle) (e.g., Coira et al., 1982;Jordan et al., 1997;Malumián and Ramos, 1984;Ramos, 1988;Mpodozis and Ramos, 1989;Charrier et al., 2007 and references therein). Although this evidence suggests a generalized compressive stress regime on the continental margin, a geochemical analysis of the magmatic units and field observations indicate variations in the stress regime showing that the Andean Tectonic Cycle in this part of the Central Andes consists of alternations of short lived pulses of increased compression followed by longer periods during which extension concentrated in the arc domain (e.g., Ramos, 1988;Mpodozis and Ramos, 1989;Jordan et al., 1997;Hartley et al., 2000;Charrier et al., 2002Charrier et al., , 2007 and references therein).
Tectonic events along the continental margin on the western side of Gondwana and South America during the Andean Cycle occurred coevally with the continental breakup process and subsequent drift of South America. During the first stage of the Andean Cycle (Early Jurassic to late Early Cretaceous), a magmatic arc with a back-arc basin on its eastern side developed with an orientation almost parallel to the western Gondwana margin. In contrast, the last two stages of the cycle (Late Cretaceous to Middle Eocene, and Middle Eocene to present) are characterized by gradual eastward migrations of the magmatic arcs (e.g., Coira et al., 1982) and the development of a retroarc foreland basin on the eastern side of the arc. The length of these stages and sub-stages was controlled by changes in the geometry, the convergence rates of the intervening plates (Nazca and South American), and the absolute velocity of the South American plate (Fig. 1b).
In Middle Eocene times, an episode of high convergence rate (Charrier et al., 2002 and high absolute western motion of South America (Ramos, 2010) caused increased compression on FIG. 1. a. Topographic image of the Andean segmentation (according to Gregory-Wodzicki, 2000). The study area is indicated; b. Nazca Plate movement reconstructions (Pardo-Casas and Molnar, 1987); c. Average convergence rate; and average obliquity angle of the Nazca Plate (Farallón) in relation with the continental margin. The black line shows the results of Somoza (1998) while the dashed line illustrates the Pardo-Casas and Molnar data (1987).
the continental margin of South America: the Incaic orogenic phase. This phase caused deformation of the previous units and uplift, which resulted in a NNE-SSW oriented relief that can be traced from southern Perú to southern central Chile (e.g., Charrier et al., 2007Charrier et al., , 2009. In central Chile, after the Incaic orogenic phase, an episode of intra-arc extension occurred between >37 and ~23 Ma, coinciding with a period of decreasing convergence rate (Pardo-Casas and Molnar, 1987;Somoza, 1998) (Fig. 1b) and caused the development of the Abanico Basin (e.g., Godoy and Lara, 1994;Charrier et al., 2002Charrier et al., , 2007Charrier et al., , 2009 (Fig. 2a). A new event of increased compression, that coincided with an episode of increasing convergence rate occurred at the end of the Oligocene (ca. 23 Ma) and the beginning of the Miocene, caused the inversion of the basin (Fig. 1c).
In this context, the evolution of the Abanico Basin can be subdivided into two stages. The first corresponds to the Middle Eocene to Late Oligocene extension (>37-~23 Ma). The second, during the Early Miocene (~23-16 Ma), corresponds to the tectonic inversion (e.g., Charrier et al., 2002Charrier et al., , 2005Charrier et al., , 2007Fock et al., 2006), which only caused partial inversion (in the sense of Williams et al., 1989). In both stages, basin infill consists essentially of volcanic and volcaniclastic deposits and a small amount of sedimentary deposits. For the extensional stage, the volcanic rocks have petrologic and geochemical affinities related to a thin crust, whereas the volcanic rocks in the second stage provide evidence for a considerably thicker crust (Nyström et al., 1993;Kay and Kurtz, 1995 1 ;Kurtz et al., 1997;Charrier et al., 2002). Late Miocene shortening has been recognized at the northern end of the Fierro Thrust (sensu Godoy et al., 1999). However, there is no evidence that this deformation was related to tectonic inversion.
From ~16 Ma, a continuous compressive regime caused further deformation and uplift of the mountain region. The thin-skinned Aconcagua fold-thrust belt was developed to the east of the Abanico Basin affecting the latest Jurassic to Early Cretaceous backarc deposits.
In this paper, the 'basin inversion' will refer primarily to the reactivation of structures occurred approximately between 23 and 16 Ma, the period when most of the Farellones Formation was deposited.

Regional problem
The extensional stage of the Abanico Basin occurred along three main fault systems, from W to E: the Infiernillo-Cerro Renca-Portezuelo de Chada faults, the San Ramón fault and the El Diablo-Las Leñas faults (see below) (Fig. 2). These fault systems defined two main depocenters or sub-basins and also controlled their tectonic inversion (Fig. 2). We named the sub-basin defined by the central and easternmost fault systems the 'eastern depocenter' (Fig. 2), while the 'western depocenter' is defined by the central and the westernmost east-dipping fault systems. The eastern depocenter was wider and more subsiding than the western depocenter ( Fig. 2) (e.g., Fock, 2005;Farías et al., 2008). Moreover, the eastern depocenter hosted a thick succession of the Abanico Formation ( Fig. 2) that reached a thickness of >3,000 m (Fock, 2005). During the inversion stage, this same depocenter accumulated more than 2,500 m of syntectonic succession of the Farellones Formation ( Fig. 2) (Fock, 2005). The idea of both formations developing under different tectonic environments is supported by structural evidence of extension during the deposition of the Abanico Formation, the existence of growth strata between both formations, and the geochemistry of these formations (Jordan et al., 2001;Charrier et al., 2002Charrier et al., , 2007Fock, 2005).
However, there are points that still remain poorly understood: 1. The relationship between basin asymmetry, thickness of the basin unfill, and a higher subsidence rate toward the eastern side; 2. The influence of the magmatic processes during the extensional and inversion stages; 3. The effect of the accumulation of the Farellones Formation during the basin inversion; 4. The relationship between inversion and other major structural features in the evolution of the Andean Cordillera, like the fold-thrust belt systems; and 5. The relationship between the structures, magmatism, erosion and sedimentation and consequently, the influence on the control of the evolution of the basin and the evolution of the mountain range.
Thus, the objective of this article is to analyze the published and newly recovered information about the tectonic inversion of the Abanico Basin, between 33º and 35ºS, in order to understand the influence of two aspects: 1. The load accumulated during extension and contraction in the basin on the inversion process; FIG. 2. a. Simplified geological map illustrating the main structures and units recognized in the study zone (based on Charrier et al., 1996Charrier et al., , 2002Charrier et al., , 2005Sernageomin, 2002;Fock, 2005;Fock et al., 2006). The outcrops of the Abanico Basin are delimited by the Abanico and Farellones formations. The image in the upper-left border in grayscale shows the morph-structural segmentation of this zone, from W to E: Coast Range, Central Depression and Principal Cordillera. The location of generalized section A-A' in 'b' is indicated; b. Generalized section showing the main structures and units related to the Abanico Formation and its deformation at 34ºS (Fock, 2005 and 2.The inversion process on the development of the associated fold-thrust belt. These are essential items that may establish the cause-effect relationship between the amount of deposits accumulated during extension and contraction and the reactivation and generation of new structures during tectonic inversion. On this basis, it will be possible to improve the understanding of the relationships between basin development and inversion, and its influence on the subsequent Andean evolution and uplift.

The Abanico Basin
The Abanico Basin formed a major tectonic and paleogeographic feature along the southern part of the Central Andes. It has been proposed that it extended from ~29º to ~39ºS in Chilean territory; further south, it extends into Argentinean territory (Charrier et al., 2005). In the region under study, the basin extended from the present day eastern border of the Coastal Cordillera to almost the axis of the Principal Cordillera (Fig. 2). It is essentially NS orientated and has a width of 70 to 80 km (Fock, 2005) (Fig. 2).
The Abanico Basin deposits are well exposed on the western slope of the Principal Cordillera, between 32º and 35ºS, and consist of intermediate and felsic lavas, different kinds of volcaniclastic deposits, and locally thick sedimentary, mostly lacustrine intercalations (Charrier et al., 2002;Fock, 2005;Jara, 2014;Jara and Charrier, 2014, this volume). The overlying and less deformed Farellones Formation consists mostly of lavas and volcaniclastic deposits.
The maximum thickness observed of the Abanico Basin in the western depocenter reached 1,300 m, while the thickness in the eastern depocenter reached >3,000 m. In the eastern compartment, 3,000 m syntectonic deposits of the Farellones Formation were added on top of the Abanico Formation, resulting in an accumulation of at least 6,000 m of deposits ( Fig. 2) (e.g., Zurita et al., 2000;Charrier et al., 2002Charrier et al., , 2005Charrier et al., , 2007Jara et al., 2009a). These data are consistent with thermal maturity studies (Zurita et al., 2000) and structural profiles (Fock, 2005;Farías et al., 2008;Jara et al., 2009b;Piquer et al., 2010), supporting the hypothesis that the Abanico Basin was almost clogged during the inversion stage. The greatest accumulation of both syn-rift and syn-inversion deposits has been found next to the main faults. More than 3,000 m of the Abanico Formation is located near the El Diablo Fault System (see below), and approximately 2,500 m of the Farellones Formation is located close to the San Ramón Fault System (Fock, 2005) (see below). The most used criterion to differentiate between the Abanico and Farellones formations is the greater amount of deformation of the first relative to the second; the Farellones Formation is generally flat-lying or only slightly deformed. Some authors have also considered the low-grade metamorphism affecting the Abanico Formation, which is almost absent in the Farellones Formation (e.g., Levi et al., 1989;Godoy et al., 1996). Local growth strata in the uppermost Abanico Formation and lowermost Farellones Formation (Charrier et al., 2002;Fock et al., 2006) indicate that the transition between both formations was associated with coeval deformation. Those growth strata are related to the reactivation of pre-existent normal faults that participated in the extensional stage (Charrier et al., 2002;Fock et al., 2006).
The geochemistry of the volcanic deposits of the Abanico Formation has been abundantly analyzed and discussed and most authors agree on its tholeiitic character, suggesting considerable crustal thinning during the Eocene-Oligocene (Nyström et al., 2003;Charrier et al., 2002;Fuentes et al., 2002;Fuentes, 2004;Muñoz et al., 2006). In contrast to the Abanico Formation, the calcalkaline character of the Farellones Formation indicates crustal thickening (e.g., Nyström et al., 1993Nyström et al., , 2003 during the Early Miocene. Stratigraphic columns of the Abanico and Farellones formations show a great variation in thickness and lithology. Lithologic subdivisions defined in one stratigraphic column are almost impossible to follow from one subdivision to the next, even though they are close to each other (Fig. 3). The correlations across and along the basin are precluded with regards to the present level of stratigraphic understanding of these formations. This complexity is probably related to: 1. the existence of abundant faults affecting both formations; and 2. rapid facies variations along and across the basin, mostly determined by the existence of volcanic centers developed within the basin. Stratigraphic columns in the Abanico Formation (Fig. 3) show the existence of thick packages of lavas intercalated with fine-grained sedimentary deposits suggesting that the volcanic vents were located nearby fluvial and lacustrine environments. In the Maipo River Valley (east of Santiago), the Farellones Formation depositional cycle consists of a thick succession of mostly andesitic lavas at the bottom, fining-upward volcaniclastic deposits, and fine-grained detritic sediments with minor tuff levels. These levels are followed by a thickening-upward succession of volcaniclastic deposits and ends with a thick succession of lavas (Fig. 3, columns G and H). However, this type of cycle is not evident in the Cachapoal River Valley (east of Rancagua) suggesting that volcanic vents, temporally and spatially, had a more random distribution than in the eastern Maipo River Valley (Fig. 3). Lacustrine intercalations are quite frequent in the Abanico Formation in the eastern Cachapoal River Valley, particularly in the upper Cachapoal and Las Leñas River Valleys. Thick lacustrine deposits (up to 500 m of grained conglomerates, sandstones, mudstones, and limestones) alternate with volcanic deposits. In some localities, up to 2-3 m thick limestone layers alternate repeatedly with red, massive, up to 4 m thick volcaniclastic intercalations over a thickness of more than 300 m (Fig. 4a). The thick calcareous layers indicate abundant lime production. This, together with the volcaniclastic intercalations, suggests that: 1. the lakes in which these volcaniclastic deposits (likely volcaniclastic flows) should have been big enough to allow abundant and long-lasting lime production in spite of having been repeatedly covered by volcaniclastic material; and 2. that subsidence in the basin was fast enough to host the abundant volcanic material supplied and to maintain the appropriate depth in the lake that allowed continued limestone deposition.
The rate of deposition (RD) in the Abanico Basin is difficult to assess. The beginning of the deposition of the Abanico Formation occurred in the Middle Eocene, according to the age of the Tapado Fauna (Tinguiririca River Valley) in the easternmost Abanico deposits (>37 Ma, Flynn et al., 2003) (see below). To the north, data in the Maipo River Valley section indicate that the youngest ages of the Abanico Formation would be ~23 Ma (see below). According to these ages, the deposition in the Abanico Basin lasted up to 14 My. If the total thickness of the Abanico Formation is ~3,000 m, the RD is ~214 m/My. Other RDs in this basin can be obtained comparing different periods and their associated thicknesses, as indicated in columns A, C, G, H, J and K ( Fig. 3 and Table 1). We obtained a great variety of RD, between 50 and 900 m/My (Table 1), with an average of 300 m/My. This wide range of RD could be associated with: 1. a lack of precision in the ages obtained; 2. inappropriate dating methods not suited to date rocks with a strong alteration; or 3. differences in the rate of accumulation due to the different kinds of deposits that coexist in the basin (Fig. 4a, b). Table 1 shows that the higher RD is associated with low error margins in the ages considered for this analysis ('e' in Table 1). Although it seems tempting to relate the high RD with periods with dominant volcanic activity, it appears that the opposite occurs if we analyze the type of deposits ('td' in Table 1). In this case, the periods of high RD appear to be associated with dominant clastic deposition. Therefore, we suggest that the high depositional rates estimated for the Abanico Basin could be mostly related to inaccuracies in the obtained ages due to the alteration caused by low-grade metamorphism.

Fault Systems
The three main fault systems that controlled the extension and contraction of the Abanico Basin ( Fig. 2) follow the regional N-S orientation in the northern part of the study region (33º-34ºS) and NNE-SSW in the southern part (34º-35ºS) (Fig. 2a). Their particular features are described below, from W to E: a. The east-dipping Infiernillo fault is a regional structure located along the eastern border of the Coastal Cordillera, between 33º and 34ºS. It is well exposed in the Cuesta de Chacabuco, Cerro Renca and Portezuelo de Chada (Fock, 2005). All three faults define the western basin border (Fock, 2005) Villarroel and Vergara (1988), Palma (1991), Aguirre (1999), Zurita (1999), Baeza (1999), Medina (2001), Muñoz (2005), Fock (2005) and Farías (2007). the evolution of the western depocenter between the Late Oligocene and Middle Miocene (Fuentes, 2004;Rodríguez et al., 2012). This system will be called IFS (Infiernillo Fault System) in this article. b. The east-dipping San Ramón fault is located on the western border of the Principal Cordillera, at the western Cerro San Ramón (33º-34ºS) (Rauld, 2002;Charrier et al., 2005;Rauld et al., 2009), and is one of the main faults that controlled the evolution of the eastern depocenter of the Abanico Basin during the Oligo-Miocene (see below) (e.g., Charrier et al., 2002Charrier et al., , 2005Campbell, 2005). This system will be named SRFS (San Ramón Fault System) in this article. Campbell (2005) proposed that its northern prolongation is the Pocuro fault, exposed north of 32º45'S (Aguirre, 1960;Rivano et al., 1993;Rivano and Sepúlveda, 1986;Jara, 2014;Jara and Charrier, 2014, this volume).
c. The steep west-dipping El Diablo fault is the easternmost fault associated with the evolution of the Abanico Basin (Fig. 2b), located immediately to the west of the AFTB. This structural system, named EDFS (El Diablo Fault System), separates the basin deposits from the Mesozoic units exposed to the east, and defines the eastern border of the basin along a major portion (ca. 100 km) of the Principal Cordillera in central Chile. It would continue from the Aconcagua (33°S) (Charrier et al., 2005) until the Maule River (Astaburuaga, 2012) through well aligned faults. Its prolongation south of the El Volcán River Valley (33°45'S) has been given names such as: 1. the Las Leñas fault, in the Las Leñas River Valley, a tributary of the Cachapoal River, at 34º15'S; and 2. the El Fierro fault in the Tinguiririca River Valley, at 35ºS (e.g., Charrier et al., 1996Charrier et al., , 2002Charrier et al., , 2005Farías et al., 2010) (Fig. 2a). It was active at least from the Middle Eocene (see below) and is still presently active with shallow seismicity reaching up to, at least, 10 km in depth (Charrier et al., 2005;Comte et al., 2008;Farías et al., 2010). The present-day displacement along the faults is dextral strike-slip. This fault differs from the one described by Godoy et al. (1999), the thin-skinned flatter El Fierro fault, which is located several km west of the here proposed El Diablo Fault System and corresponds to a low angle out-of-sequence thrust.

Regional significance of the fault systems
The greater thickness of the material accumulated in the basin next to the faults that bound its eastern compartment (SRFS and EDFS, Fig. 2) indicates that their first movement would be normal and their second movement would be reverse (e.g., Charrier et al., 2002Charrier et al., , 2005Fock, 2005). During inversion, large-scale symmetric and asymmetric folds developed along these fault systems and essentially affected the Abanico Formation (Figs. 2a and 4c, Fock, 2005;Fock et al., 2006). Small-scale folds and growth strata strike parallel to the main faults, but show different vergences ( Fig. 2b) (e.g., González, 1963;Thiele, 1980;Thiele et al., 1991;Charrier et al., 2002Charrier et al., , 2005Fuentes et al., 2002;Rauld, 2002;Fock, 2005;Fock et al., 2006), suggesting that they were formed by the reactivation (inverse reactivation) of minor blind faults associated with the development of the extensional basin (Fock, 2005;Fock et al., 2006). These structures have been observed in the following river valleys (Figs. 2b and 5): Rocín, Maipo (at El Toyo), El Volcán, Las Leñas and Upeo (Charrier et al., 2002;Campbell, 2005;Fock, 2005). Some authors have proposed a secondary control on the Abanico Basin defined by oblique, NNE-SSW oriented faults (Rivera and Cembrano, 2000;Rivera and Falcón, 2000). These faults could represent transfer structures separating the depocenters in the basin. However, these structures are badly exposed and can only be inferred using indirect evidence. Recent analogue models have shown that basement structures oblique to the main structural systems controlling the basin development strongly influence the geometry of the surface deformation (Pinto et al., 2009;González et al., 2012).
At the basin scale, the main structural systems described affected the brittle crust, reaching depths of at least 10 km (e.g., Fock, 2005;Farías et al., 2010). In addition, some authors (Rauld et al., 2009;Armijo et al., 2010) have recently proposed that the SRFS provides evidence on the surface for an active, west-vergent, mega crustal detachment that played a major role in the uplift of the Principal Cordillera, thereby dismissing the importance of the structural systems on the eastern edge of the Andes. However, these authors disregarded solid evidence supporting a generalized east-vergence of the orogen and the existence of a main crustal detachment underneath the Andes connecting the subduction zone and the Andean front in this region (e.g., Giambiagi et al., 2003a, b;Farías et al., 2010).

Ages of deformation
The radioisotopic ages obtained in the deposits associated with the growth strata and associated folds allow us to constrain the age of these structures and of the inversion stage of the basin. Additionally, if we separate the Abanico Formation from the Farellones Formation based on the greater deformation of the first relative to the second, then the age of the transition between the Abanico Formation and the overlying Farellones Formation can be determined by dating the deformed and undeformed layers associated with the growth strata and associated folds. The available ages that constrain the transition between these formations are: 1. At El Toyo, in the Maipo River Valley next to San José de Maipo (Fig. 2b), the upper levels of a fold associated with growth strata have been dated at 21.74±0.04 Ma ( 40 Ar/ 39 Ar plateau age on plagioclase) and 21.3±0.3 Ma (U-Pb SHRIMP on zircon crystals) (Zurita et al., 2000;Fock, 2005;Fock et al., 2006). This indicates that the deformation occurred before 21.74±0.04 Ma and that this is the approximate age of the base of the Farellones Formation; 2. Further south, in the same valley at the Vuelta del Padre (Fig. 2b), in the lower undeformed levels of the Farellones Formation, Vergara and Drake (1978) obtained two whole rock K-Ar ages of 22.9±0.5 Ma and 22.4±5.0 Ma, and a K-Ar age on plagioclase of 24.1±1.0 Ma, which are somewhat older than the age obtained at El Toyo (~23 Ma) for the lowermost Farellones Formation; 3. Further southeast, in the El Volcán River Valley, an E-W tributary of the Maipo River (Fig. 2b), folded layers that participate in a growth structure were dated at 21.4±1.0 Ma (Fock, 2005;, indicating that deformation was active at least after that age; 4. One hundred kilometers further south, in the Las Leñas River Valley (Charrier et al., 2002), growth strata are constrained in age between 20.09±0.69 Ma ( 40 Ar/ 39 Ar laser fusion of single plagioclase crystals) (Flynn et al., 1995) and 16.1±0.5 Ma (Kay and Kurtz, 1995), indicating a somewhat younger deformation age in this region. The geometry of the syntectonic wedge does not suggest that ~20 myr old strata were deposited during compression, but it is clear that the ~16 Ma old strata are almost undeformed, indicating that the inversion process was quenched during this time at this location; 5. The oldest rocks of the Farellones Formation in its type locality (Fig. 2b) were dated at 21.6±0.2 Ma ( 40 Ar/ 39 Ar, plagioclase; Aguirre et al., 2000) and the youngest at 16.6±0.7 Ma (K-Ar, whole-rock) (Beccar et al., 1986); 6. At Lagunillas, a sky station located ~1,500 m above the Maipo thalweg and the El Toyo growth strata, Vergara and Drake (1978) dated the upper levels of the Farellones Formation at 17.3±0.2 Ma (K-Ar on plagioclase); 7. The K-Ar ages for younger and non-deformed strata of the Farellones Formation span a range between 14.1±0.6 and 8.3±0.1 Ma (Charrier and Munizaga, 1979). Therefore, these ages and deformational FIG. 5. El Volcán fold and associated to it, growth strata at the El Volcán locality, view to the south. The black lines underline the geometry of the strata. The age of an andesitic lava is shown. This structure is interpreted as a backthrust associated with the EDFS.
characteristics of the deposits constrain the inversion process for the study region between ~23 and ~16 Ma (Early Miocene). Cenozoic plutons emplaced in the basin form parallel alignments to the main faults that control the Abanico Basin (Fig. 2b). In general, the ages of these plutons decrease eastward, ranging from ~20-18 Ma in the western alignment (SRFS) to ~13-8 Ma in the next eastern alignment (EDFS) (Vergara and Drake, 1978;Munizaga and Vicente, 1982;Kurtz et al., 1997;Gana and Wall, 1997;Cornejo and Mahood, 1997;Sellés and Gana, 2001;Fock, 2005;Muñoz et al., 2009). However, there are older intrusions on the eastern border of the basin such as the Juncal pluton of ~16 Ma (Montecinos, 2008;Montecinos et al., 2008), apparently within the EDFS and also a volcanic rock of ~18-16 Ma (Aguirre et al., 2009) that would correspond to a sill within this fault system. These ages coincide with the age of the Farellones Formation, suggesting that the structural systems controlled the location of the magmatic activity during the partial inversion stage.
According to the available ages from the Abanico Formation (>37 to ~23 Ma), the extensional stage would have lasted for >14 Ma, after the Incaic orogenic phase. According to the ages of the Farellones Formation related to the inversion (~23 to ~16 Ma), the duration of the partial inversion stage has been estimated at ~7 Ma, which agrees with previous work in other regions of the basin (e.g., Charrier et al., 2002). On the other hand, in regard to the thickness of the deposits of the Abanico and Farellones formations, the minimum basin depths considered are 3,000 m and 2,500 m during the extensional and partial inversion stages, respectively. The minimum basin width is estimated to be between 70 and 80 km, considering the current outcrops (e.g., Fock, 2005). The convergence rates in extension and contraction can be associated with the plate velocities in the course of these periods; the Nazca plate had a rate movement of 15 cm/year in the extensional stage, whereas it was close to 10 cm/year in the partial inversion stage (Somoza, 1998). Based on the geochemical similarities with volcanism in the Central Andes, the crustal thickness during extension is expected to be ~30-35 km, while during contraction this is estimated to be ~35-40 km (Kay and Kurtz, 1995;Kay et al., 1999;Nyström et al., 2003).

Analysis of the basin geometry and chronology of the extensional structures
The eastern side of the Abanico Basin was associated with the development of the first main fault, and the accumulation of material would have started next to it, configuring the basin depocenter (Fig. 2). The data from the eastern depocenter are the following: in the eastern border of the Abanico Formation, at Termas Baños del Flaco (Tinguiririca River Valley, 35º00'S, Fig. 2 Charrier et al., 1996). In this area, the thickness is estimated at 3,000 m. Twelve to 14 km west of the Termas Baños del Flaco, middle Eocene mammalian fossils have been found in deposits of this formation, indicating that deposition in this region began before deposition at Termas Baños del Flaco (El Tapado fauna: Flynn et al., 2003Flynn et al., , 2005Croft et al., 2008a, b). Further north in the same eastern depocenter, in the El Volcán River Valley at 33º45'S ( Fig. 2b), the thickness of the Abanico Formation is greater than 3,000 m (Fig. 2c, Fock, 2005), with 40 Ar/ 39 Ar ages of 34.3±0.3 Ma at the base, and 21.4±1.0 Ma at the top   (Fig. 3, column H). Additionally, in the western margin of the Principal Cordillera, in the Cerro El Abanico (Fig. 2c), which corresponds to the western margin of the eastern depocenter, an 40 Ar/ 39 Ar age of 30.93±1.90 Ma was obtained from the lower portion of the Abanico Formation ( 40 Ar/ 39 Ar total fusion of plagioclase crystals; Vergara et al., 1999Vergara et al., , 2004. In this region the measured thickness is also ~3,000 m. In the Río Las Leñas region, the burial of the basal and eastern outcrops of the Abanico Formation (~34°25'S, Fig. 2b) at a depth of 6,000 m (Zurita et al., 2000) supports the asymmetry of the eastern depocenter and its deepening toward the east.
In the western depocenter, in the Cuesta de Chacabuco region, north of Santiago and east of the Coastal Cordillera, the Abanico Formation has been described as resting unconformably on the Late Cretaceous Lo Valle Formation, with a maximum thickness of 1,300 m, and a whole rock 40 Ar/ 39 Ar age of 29.2±0.2 Ma at its lower levels (Fuentes et al., 2002). In addition, in the Santiago region, radiometric ages from the Abanico Formation are comprised between 28 and 20 Ma (Wall et al., 1999). Further south, at Angostura de Paine (Fig. 2b), the age obtained close to the base of the Abanico Formation is 26.3±0.9 Ma (K-Ar, whole rock, Sellés and Gana, 2001). The given ages for the oldest deposits on the western side of the western depocenter, in the Cuesta de Chacabuco, Santiago and Angostura de Paine areas, are relatively similar to each other. The age from Cuesta de Chacabuco coincides well with the oldest age from the lowest exposed portion of the Abanico Formation in the western border of the eastern depocenter. These ages indicate that the oldest deposits of the Abanico Formation in the western depocenter have Late Oligocene ages. We must note that a 34.3±2.2 Ma age has been reported for a dyke that intrudes the lower part of the Abanico Formation in the Cuesta de Chacabuco; however, the geochronological data have a high degree of uncertainty (Gana and Wall, 1997). In light of this, we should mention that there is a recent study showing errors in the 40 Ar/ 39 Ar radiometric determinations in igneous rocks from central Chile caused by an excess of 40 Ar (Montecinos, 2008;Montecinos et al., 2008). However, there is no systematic study refuting the ages cited in this article.
Thus, based on the given ages, it is possible to conclude that the thickness and age of the deposits show a direct relationship with the basin's asymmetry. After considering the geometry, thickness of the succession and structural control during extension, we conclude that the Farellones Formation deposits indicate that there was space enough to accumulate them during the inversion stage.
Fock (2005), Fock et al. (2006) and  suggested that the western depocenter has been uniformly filled, beginning in the Middle Eocene, with the deposition of the Cordón de los Ratones Beds in the Angostura de Paine area (Sellés and Gana, 2001). Because of the rather old age of the latter and the proposed angular unconformity separating this unit from the Abanico Formation (Sellés and Gana, 2001), it is not possible to consider with certainty that these beds are the oldest deposits accumulated in the basin. Furthermore, the folds and felsic intrusions in the Cordón de los Ratones Beds have no continuity in the Abanico Formation (Sellés and Gana, 2001). At that locality, the Cordón de los Ratones Beds have ages of ~43 Ma (43.0±0.4 Ma, U-Pb SHRIMP; Fock, 2005), while the Abanico Formation yields ages of ~26 (26.3±0.9 Ma, K-Ar whole rock; Sellés and Gana, 2001), providing evidence for a ca. 17 Ma hiatus (Fock, 2005). Therefore, even though the Cordón de los Ratones Beds are located next to the western basin bounding fault system (IFS), this unit is possibly older than the Abanico Basin at this locality. Furthermore, the Cordón de los Ratones Beds are apparently not genetically related with the IFS, because its succession and the Abanico Formation are affected by the Portezuelo de Chada Fault (see figure 5.2 in Fock, 2005), and the outcrops to the west suggest that this fault did not control the western border of the basin. If in fact there was a structural control on this border, the main fault would be located further west. More studies are needed to better understand this part of the Abanico Basin.

Analogy between the basin and the experimental generation of faults in the extensional stage
The evolution of the Abanico Basin during extension can be compared to the results of analogue modeling designed to simulate this geological phenomenon at the first order (Fig. 6a) (Pinto et al., 2010). In these experiments, the modeling techniques were similar to those used for brittle ductile systems at the Laboratory of Experimental Tectonics of the Department of Geosciences, Université de Rennes 1, and described in detail in the literature (e.g., Balé, 1986, for the use of velocity discontinuity (VD); Faugère and Brun, 1984, for the use of silicone and sand; Davy and Cobbold, 1991, for scaling). Brittle behavior was represented by sand putty SGM 36 (Dow Corning, USA) with a viscosity (µ) around 10 4 Pa·s at 20°C and a density (ρ) close to 1,000 kg/m 3 , and ductile behavior was represented by a mass of transparent silicone putty SGM 36 (Dow Corning, USA) with a viscosity (µ) around 10 4 Pa·s at 20°C and a density (ρ) close to 1,000 kg/m 3 . The experimental device consisted of a fixed and rigid base, on which a thin plate moving at a constant rate was pulled (extension) or pushed (compression). The limit of the mobile plate induced an asymmetric velocity discontinuity (VD) at in the base of the model, which localizes the deformation (Malavielle, 1984;Balé, 1986;Allemand et al., 1989;Ballard, 1989). The angle of obliquity used in the compressive experiments was 30° with respect to the VD (Nalpas et al., 1995;Brun and Nalpas, 1996). Above the VD a thin layer of silicone allowed the localization of the deformation in a larger zone than the zone with only a VD (Figs. 6  and 7). The model was set in a 60x44 cm sandbox, wide enough to achieve a relatively large amount of movement without border effects and to be able to cut the experiment several times after deformation. The prekinematic pile of the models was made, from bottom to top, of around 1.5 cm of transparent silicone (just above the VD) and black and white horizontal sand layers (Fig. 6). Brittle sand layers were made of various colors in order to reveal the structures and to observe them on the photographs. The color of the sand does not modify its behavior. Two groups of experiments were performed. The first group consisted of models with only extension orthogonal to the VD in order to create a basin (Fig. 6a). In the second group, oblique compression (30°) was applied after the extension, with an amount of 'orthogonal shortening' equal to the previously obtained extension (Fig. 6b to f). In both extension and oblique compression, sedimentation of fresh sand (blue and white and black and white) was continuously sprinkled manually onto the model in order to simulate synkinematic mass transfer (e.g., Barrier et al., 2002). Blue and white horizontal sand layers were deposited inside the basin during extension, black and white horizontal sand layers were deposited inside and/or outside the basin during compression, with 1 cm representing around 3 km in nature. This flatness of the layers does not simulate the variability of the thickness of the layer in nature; this sedimentation in the experiment is a standard procedure for analyzing the geometry of the deformation. For the extensional series, three modes of sedimentation were tested: 1. without sedimentation; 2. with partial sedimentation; and 3. with complete sedimentation (Fig. 6a). The compressive models were based on the same extensional series, i.e.: 1. without sedimentation (Fig. 6b); 2. with partial sedimentation (Fig. 6c, f); and 3. with complete sedimentation (internal and external to the basin, Fig. 6d, and external to the basin, Fig. 6e). Partial sedimentation means that there were topographic heights around the sedimentation zones, and complete sedimentation means that the entire surface of the sedimentation reached the highest topography created by deformation. Photographs of the model surface were taken at regular time intervals in order to observe the development of the structures. After deformation, the internal structure and lateral evolution were observed on a series of cross-sections cut parallel to the compression direction (perpendicular to the VD).
Only the more representative experiments (from Pinto et al., 2010) are presented and discussed with regard to the geological model of the Abanico and Farellones formations. We only present a qualitative interpretation without a mechanical analysis, but our suggestions are coherent with previous research in inversion analogue modeling (e.g., Nalpas et al., 1995;Brun and Nalpas, 1996;Dubois et al., 2002). The internal geometry of the deposits is not a variable that we considered in our modeling. In any case, the previous data presented on the deposits of the Abanico Basin (item 2.1) suggest that the paleotopography of the border of the basin during the Oligo-Miocene was quite smooth. The grabens were filled horizontally, because the device and method of sedimentation has to be simple and systematic. Further modeling is required, including factors such as regional erosion and magmatic fluids in the faults during inversion. The latter would affect the reactivation of the structures involved in the inversion process. However, we should clarify that it is a standard procedure in analogue modeling to isolate factors in order to systematically analyze their influence on the deformation. Natural processes are obviously very complex and several factors can not be analyzed simultaneously, therefore only one factor is analyzed at a time.
A schematic section summarizes the main structures generated after extension and inversion of the basin during the oblique compression phase (Fig. 7). The main characteristics related to extension are the generation of two grabens, which were called G1 and FIG. 7. Schematic section that summarizes the main structures generated in compression after the extensional phase. This generic case represents the sedimentation contemporaneous to the extension and contraction/inversion phases.
G2 in relation with their temporality (Fig. 5). The first graben, G1, has associated faults N1 and n1, while the second, G2, is determined by structures N2 and n2 (Fig. 5). The spatial distribution of the faults from E to W is: N1, n1, n2 and N2. The main principal fault with higher movement and associated subsidence is N1, whereas n1 and N2 are intermediate in terms of movement and fault n2 has negligible movement. This difference in movement is directly related to the chronology of the faults; the fault that began in the eastern edge was affected for a longer period of time by the sediment load, which helped to increase the activity of the fault (Pinto et al., 2010). This produced an asymmetric basin during extension with a maximum of subsidence and sedimentation in its eastern side (Fig. 6a). Setting up the analogy between the structural systems of the Abanico Basin and the faults in the analogue model, from W to E: the IFS could be associated with fault N2, the SRFS with n1 (Fig. 8a), and the EDFS with N1. However, n2 does not have an analogue in the real system of the Abanico Basin. The asymmetry of the Abanico Basin is evidenced by a variation in the thickness of the deposits, and this asymmetry is comparable to the analogue experiments. The deepest graben that hosts the greatest amount of sediments, G1, is produced by N1 and n1, and is comparable with the eastern basin depocenter, i.e., the zone between the SRFS and EDFS. This depocenter was configured from ~37 Ma on its eastern border in the EDFS.
In addition, according to the analogue experiments, the very thick deposits of the Abanico Formation in this region enhanced the great displacement and strong subsidence along the EDFS at the basin scale. Furthermore, the detachment of the Mesozoic units to the east of the basin was facilitated by the thick Oxfordian gypsum layers (Auquilco or 'Yeso Principal') (e.g., Dellapé et al., 1979;Thiele, 1980;Legarreta and Uliana, 1996) (Fig. 9). Additionally, in the basin floor, this level might have enhanced subsidence in the basin during the extension. This influence from the evaporitic intercalations has been shown in analogue experiments on basins in the southern North Sea, where layers of this type acted as detachment levels during extension, facilitating the greater subsidence of the involved blocks and which may have facilitated the thrusting and inversion during compression (Nalpas et al., 1995). Additionally, it is possible, with the aid of analogue modeling, to infer that the eastern depocenter progressively increased its width toward the west. The western border of this depocenter was developed at ~30 Ma, and was defined for the activity of the SRFS; fault n1 in the experiments will be analogue to this structure.
G2 is similar to the western depocenter of the Abanico Basin, which began to develop at ~30 Ma. It is built up as a half-graben associated with the activity of the IFS; the N2 fault in the experiments would be a similar feature (Figs. 2b  and 7). As well, it is possible to infer that the accumulation of the Abanico Formation deposits progressed toward the west at ~30 Ma, configuring the western depocenter.

Analogy between the partial inversion of the Abanico Basin and the reactivation of the faults in the experiments
As mentioned previously, several authors have proposed that the extensional deposits of the Abanico Basin were extruded by a reactivation of the main normal fault systems described above (Fig. 2c). Its evolution during the partial inversion can be compared to the results of the analogue modeling (Figs. 6d-f and 7) (Pinto et al., 2010). The main characteristics related to inversion are that the normal structures, which generated grabens during the extensional stage, were reactivated as high-angle reverse faults during oblique compression (R1, R2 and r1, in Figs. 6d-f and 7). These faults define the inverted grabens, from E to W: faults R1 and r1 define the inverted graben Gi1, and faults R2 define the inverted graben Gi2 (Figs. 6d-f and 7). The shortening was absorbed by these reactivated faults, as well as by two new reverse faults, which were generated toward the exterior of the graben, I1 and I2 (Figs. 6d-f and 7).
Then, it is possible to establish that the structural systems cropping-out in the field: IFS, SRFS and EDFS are analogous to R2, r1 and R1, respectively. The hypothesis of the almost clogging of the Abanico Basin during its inversion stage and the previous analysis of the extensional stage suggest that the basin development was very similar to the experiments with partial sedimentation, during both extension and contraction ( Fig. 6f; Pinto et al., 2010).
Previous studies on basin inversion, established that the reactivation of faults is selective. This conclusion was based on the me-chanical characteristics of the rock, such as friction, cohesion coefficients and pore fluid pressure (e.g., Byerlee, 1978;Sibson, 1995), and on the rheology of the entire crust (Ranalli, 2001), but did not consider the relationship with the load of sediments in the basin. The analogue experiments by Pinto et al. (2010) indicate that the sedimentary load in the basin during extension and oblique compression inhibits inversion. Therefore, we propose that the load of both formations, Abanico and Farellones, hindered the complete reactivation of the structures in the Abanico Basin in the Early Miocene (~16 Ma).  Fock, 2005;Fock et al., 2006;Giambiagi et al., 2003b). Section (a) shows the inversion phase, and (b) shows the stage when the generalized inversion is stopped and the shortcut is produced over the Oxfordian gypsum layer; this acts as a detachment level of the AFTB. Point A in (b) shows the intersection of the top of the reactivated fault and the shortcut thrust. It will be noted that the reactivation between points A and B has finished. The base level of the generalized outcrops is shown with a segmented line. The main reactivated El Diablo Fault System is labeled as EDFS. The labels for the analogue faults R2, N2 and I2 are the same as those in figure 7.

Geometry of the inverted basin
Contrary to what the analogue models show, the thickest deposits of the Farellones Formation in the Maipo River Valley (Fock, 2005) are not associated with the oldest fault system, EDFS. Instead, these deposits are located near the SRFS, at the Cerro Los Azules locality in the Cerro San Ramón (Fock, 2005). However, in light of the high deformation observed at the easternmost side of the basin, it is possible to suggest that the uplift and erosion near the EDFS after ~16 Ma was huge, and that a great part of the eastern Farellones Formation had been eroded and that the presently exposed thickness of the Farellones Formation on this side is not its original thickness. This hypothesis could be verified by analyzing the history of the erosion after the Early Miocene ~16 Ma in the Principal Cordillera at this latitude. In addition, it is necessary to emphasize that erosion during and after inversion has not been considered in the analogue experiments, and that we cannot compare the original thickness with the residual thickness of the Abanico and Farellones formations. Moreover, there is a strip of ~20-18 Ma plutons along the SRFS which would support increased volcanic activity on this side of the basin and would possibly explain the increased thickness of the Farellones Formation on this side. However, we have not analyzed the influence of magmatism in the inversion processes so we cannot further expand upon this second hypothesis.

Generation of new structures associated with
the migration of the deformation According to Fock (2005) (Fig. 2b), reverse movements along the EDFS would have induced the development of the AFTB in the Early Miocene; however, this author did not establish how this occurred. The analogue experiments show the development of shortcut faults, I2, departing from the basin bounding faults in the regions where sedimentation was greater and the movement on the normal faults was higher (Fig. 6d). These structures can be compared to the detachment level of the AFTB related to the evolution of the EDFS during the reactivation. When the reactivation movement on the main normal fault becomes significant and there is a thick load over it, then it difficult to completely reactivate this fault and it is easier to create a shortcut fault. Thus, we propose that displacement along the detachment level of the AFTB was triggered by the merging of the thin-skinned shortcut thrusts formed in the upper part of the thick-skinned EDFS (Fig.  8b). Furthermore, we postulate that the great amount of load on this basin margin enhanced the development of the shortcut faults and, therefore, was the key factor that trigger the AFTB (Fig. 8b), and not simply the presence of the Oxfordian gypsum layer (Thiele, 1980;Ramos et al., 1996). It is important to note that when the AFTB was generated, in the Early Miocene, the Abanico deposits were progressively uplifted with almost no internal deformation. This uplift would be the result of a deep structure in the basement, which could explain the great shortening documented for the AFTB (L. Giambiagi, personal communication, 2011). We propose that the EDFS would be the deep structure that would confine the basin and remained active during the main uplift of the Andes at this latitude during the development of the AFTB.
We consider that the displacement associated with basin inversion along the uppermost part of the EDFS (the portion of the fault above the point where the shortcut roots) would have ended at ~16 Ma, considering that this is the oldest available age for the almost not deformed strata of the Farellones Formation at this latitude. After this moment, the deformation would have concentrated on the AFTB. The mentioned ~16 Ma age is consistent with the <18 Ma age given for the initiation of activity in the AFTB, based on the age of the volcanic Contreras Formation that underlies the syntectonic Tunuyán Conglomerate (Giambiagi et al., 2001(Giambiagi et al., , 2003b. According to the analysis presented here, the Abanico Basin was completely filled before the deformation progressed eastward. This assertion is confirmed by the existence of volcanic clasts in the lower levels of the Tunuyán Conglomerate that have been interpreted as coming from the formations of the Chilean Principal Cordillera, although there are no detailed studies to exactly determine the origin of the clasts. Moreover, according to Giambiagi et al. (2003b), the erosion of the Mesozoic rocks only began at 15 Ma, based on the presence of Mesozoic clasts in the middle portion of the Tunuyán Conglomerate, indicating that these rocks were affected by the AFTB after inversion of the Abanico Basin. Therefore, the rocks outcropping in the Principal Cordillera at ~16 Ma would have corresponded to the Farellones Formation. After that, the Mesozoic rocks immediately east of the EDFS would have been uplifted and the volcanic cover would have eroded until the exposure of the Mesozoic succession at 15 Ma. From this, it follows that during the final depositional stages of the Farellones Formation, these deposits had covered the Mesozoic succession beyond the EDFS, which corresponded to the eastern edge of the Abanico Basin.
Another interesting aspect of the development of the deformation is the depth of the detachment levels of the different fold-thrust belts in the Central Andes in central Argentina and Chile, and ultimately to their thin-or thick-skinned character. The experiments show a direct relationship between the amount of load where the shortcut thrust is developed and the depth at which the shortcut inception is located (Fig. 6b-f) (Pinto et al., 2010). The inception of the shortcut thrust is shallower when the amount of sediment is greater (compare eastern reverse fault (I2, Fig. 7) in experiments c, d and f in Fig. 6). North and south of the AFTB, thick-skinned fold and thrust belts are developed: La Ramada (RFTB), to the north, and Malargüe (MFTB), to the south (e.g., Ramos et al., 1996Ramos et al., , 2002Ramos et al., , 2004Giambiagi et al., 2003aGiambiagi et al., , 2003bGiambiagi et al., , 2005Turienzo, 2010). In these cases, preexisting structures to the east have been reactivated allowing the uplift of the basement blocks, developing a complicated structural pattern (Giambiagi et al., 2003a). However, considering the analysis in this study and that the Abanico Basin extends to the north and south of the study area, we suggest that the RFTB and MFTB would have also been triggered by the inversion of the Abanico Basin and the shortcut thrusts departing from the reactivated EDFS or equivalent basin bounding fault systems.
It is also important to note that the high westward dip of the Mesozoic layers located immediately to the east of the EDFS, such as in the El Volcán region to the east of Santiago (Fig. 9), has most likely been caused by rotation along the shortcut departing from the EDFS (Fig. 9).
The difference in rheology between the eastern and western sides of the basin enables us to propose a different organization and configuration of the deformation on both sides of the Abanico Basin. Typically, in areas of brittle rheology, the contractional deformation is characterized by reverse faults with few folds. The layers remain roughly horizontal during the deformation and displacement. Conversely, when there are detachment levels or stratified rock units that allow movement along the detachment or stratification planes, the contractional deformation can generate folds of different wavelengths, and reverse ramp-flat faults. Furthermore, the accommodation of the deformation is easier when there is a weaker rheology with a detachment level. The analysis of the deformation associated with the partial inversion of the Abanico Basin indicates that there is a relationship between the rheology and the overall geometry with the structures and layers between the western and eastern blocks of the system. The western block has a more brittle rheology (Coastal Cordillera, Fig. 2b), and only developed reverse faults with steep dips and without significant folds. In this area, the amount of shortening is very low. On the contrary, in the eastern block, it is observed that the folds are associated with reverse faults with high or low angles, the latter with ramps and flats (Chilean Principal Cordillera, Fig. 2b). The deformed layers show a strong rotation (see above), growth strata and flexures of long wavelengths (Figs. 2b, 5 and 9). There is also a strong shortening compared with the western side. This is to be expected considering the presence of abundant sedimentary rocks in the eastern side of the Principal Cordillera at this latitude, and in particular the presence of a thick layer of Oxfordian gypsum (Fig. 9b). Thus, to be consistent with the rheology of the upper crust, we propose that, on the eastern side of the basin, there was a larger amount of shortening, and the vergency of the inversional deformation was organized toward the east.
There are folds, such as the Río Maipo anticline, that could correspond to backthrust systems associated with the SRFS in the western border of the Abanico Basin ( Fig. 2c) (Fock, 2005). The analogue experiments also show the development of a minor thrust (Fig. 6c, f) and folds associated with the main faults during inversion (Fig. 6b). Additionally, the Río Maipo anticline, next to El Toyo (Fig. 2b), presents a growing geometry of the basal levels of the Farellones Formation in the Early Miocene (~23 Ma) (e.g., Fock, 2005). It is possible to make the same analogy with the El Volcán fold (Fig. 5) (Fock, 2005), which developed from ~21 Ma, because of the propagation of a backthrust associated with the EDFS (Figs. 2 and 5) (Fock, 2005). For the other folds and associated growth strata identified to the north and south of the study area (Río Rocín and Río Upeo (Charrier et al., 2002;Campbell, 2005;Jara and Charrier, 2014, this volume), it is claimed that these are backthrust structures based on their location and vergency. However, there are still few detailed studies of them (e.g., Baeza et al., 2000). The Río Rocín structure would be associated with the SRFS, and the Río Upeo structure with the EDFS, respectively.

Evolution of the Abanico Basin
Based on Fock (2005), the previous discussion, and the adaptations of the tectono-sedimentary analysis from this study, we propose the following paleogeographic and structural evolution of the Abanico Basin, between 33º and 35ºS ( Fig. 10): Late Eocene (>37-30 Ma): After the contractional Incaic orogenic phase in the Middle Eocene , an extensional event generated the EDFS. This normal fault system controlled the beginning of deposition on the eastern side of the Abanico Basin in a half-graben (>37 Ma). According to this configuration, a large amount of volcanic and volcaniclastic deposits accumulated in this side of the basin and, therefore, caused greater subsidence along the EDFS.
Early Oligocene (~30 Ma): The IFS and SRFS were generated as normal faults associated with extension during this period creating the half-graben sedimentation space for the Abanico Formation on the western side (western depocenter) of the basin, which was mainly controlled by the IFS. Furthermore, the eastern half-graben was controlled by two faults: the SRFS and EDFS (eastern depocenter). Crustal thinning is associated at this moment.
Late Oligocene (~30-23 Ma): Extension continued and there was an increase in the load in the basin associated with magmatic activity. Thick volcanic and volcanic-sedimentary successions were deposited. The basin deepened on the eastern side of the depocenter as a result of the increasing load, causing it to become asymmetrical.
Early Miocene (~23-16 Ma): Compression and crustal thickness started in this period. The normal structures were reactivated as high-angle reverse structures. Deposition of the Farellones Formation began in this period and concentrated in the eastern depocenter. The accumulation space in the western depocenter was smaller, and the deposits of the Farellones Formation were considerably thinner.
Middle Miocene (~16 Ma): Compression and crustal thickenning followed during this period. The AFTB began to be generated by a shortcut thrust rooted in the EDFS, along the eastern border of the basin, and caused the high load of deposits that inhibited inversion along this fault. After the development of the shortcut thrust, the reactivation of the upper part of the EDFS came to an end.

Contributions of the basin analysis to the global knowledge of Andean evolution
The main motivation of this study was to understand the documented inversion of the Abanico Basin, particularly the synrift and inversion processes of this basin. This would be an atypical case of basin inversion, because previously known cases and models correspond to basins associated with marine environments, which underwent a complete inversion process and therefore, their syntectonic deposits accumulated outside the basin (e.g., McClay, 1995). As a consequence, our analysis was based on the correlation with analogue models (Muñoz et al., 2007;Pinto et al., 2010) to understand the influence of sedimentation and load on the style of deformation during the extension and contraction/ inversion processes. We conclude that: 1. there may be a partial inversion and accumulation space within the basin, permitting the deposition of the Farellones Formation; 2. on the eastern border of the Abanico Basin a major displacement may have taken place in the EDFS during extension associated with a greater load on this edge; and 3. during the compressive stage, a greater load would have produced a new reverse fault that developed as a shortcut structure departing from the inverted structures. Until now it was not possible to understand the connection between the high-angle eastern basinbounding fault (EDFS) and the low-angle structures that formed to the east of the fault (AFTB) in the same contractional process. In this study, we established how these fault systems would be connected and we propose that the detachment level of the AFTB was initiated by a shortcut structure generated from the basin-bounding fault. Also, we propose that one of the fundamental factors in the generation of the shortcut faults is the great thickness of the deposits accumulated on this side of the basin, which reduces the reactivation of the structures.
In addition, in the particular case of the Abanico Basin, it was also possible to explain its geometric asymmetry. This geometry had not been understood before Fock (2005), who defined it and registered the contractional processes with consistent evidence. This author defined the syntectonic character of the Farellones Formation through the identification, analysis, and dating of the growth strata. As indicated in this work, the growth strata in this region are consistent with the inversion process.
This study confirms the crustal character of the Abanico Basin's eastern fault systems, as has been already proposed by Fock (2005), Fock et al. (2006), Farías et al. (2008Farías et al. ( , 2010 and others. From the beginning, the EDFS was one of the main faults that controlled the basin and its deformational characteristics at least since ~30 Ma. This fault system should have its root at >10 km of depth. Therefore, the model suggested by Rauld et al. (2009) and Armijo et al. (2010) (see the Introduction), which considers the SRFS as the only important fault in the Chilean Principal Cordillera, is untenable. These authors did not consider the extensional development of the Abanico Basin in their model. This extensional stage was fundamental to the evolution of the Andes because it was during this period that the main structures were generated, and these structures were only reactivated later. Their model is based only on the western part of the Chilean Andes and rejects solid evidence from the eastern part, not only in Chile, but also further east in Argentina, such as the EDFS and many structures associated with the AFTB, the Frontal Cordillera and Precordillera and the foreland basin in Argentina.

Conclusions
The study of the Abanico Basin and analogue models of the inversion processes revealed the importance of infill controlling the style and evolution of the deformation of inverted basins. Making the analogy between real data and modeling, we propose that the geometric and temporal asymmetries of the Abanico Basin are associated with load differences in its borders. The development of the basin would have started in the eastern margin with the activation of the EDFS during the extensional period in the Middle Eocene. Deposition started in this side of the basin, which became more loaded and deeper. This feedback between loading and deepening would explain the great thickness of the deposits in the eastern depocenter during the extension of the Abanico Basin.
After that, the basin developed toward the west; the development of the SRFS configured the western part of the eastern depocenter. The SRFS and the EDFS absorbed most of the deformation associated with the deepening of the basin and promoted a large amount of subsidence.
In addition, it is established that the western border of the basin was controlled by the IFS. This fault was created later; consequently, the western depocenter received less material than the east-ern one. As a result, the western side of the basin was less loaded and shallower than the eastern side. This is consistent with the ages of the rocks in the Abanico Formation and the asymmetry of the basin.
Likewise, analogue modeling predicts that during the basin inversion, the deepest depocenter still has enough space for accumulation because it was generated in the earliest stage of the basin. This prediction is consistent with the distribution of the Farellones deposits, which are more abundant in the eastern depocenter of the basin.
Finally, we propose that the AFTB began in the Middle Miocene (~16 Ma), when the additional load provided by the Farellones Formation almost clogged inversion of the basin. Its first devel-opment would be favored by the generation of a shortcut structure associated with the basin inversion, specifically from the reactivated EDFS. The shortcut thrust was generated, among other causes, because of the high load represented by the Abanico and Farellones deposits in the eastern depocenter of the Abanico Basin. We suggest that the EDFS absorbed some part of the deformation, and that with the appearance of the shortcut, the reactivation of the upper part of the EDFS came to an end and only its lower part remained active. High load in the eastern side of the basin, during extension and contraction, would have been an important factor in the development of the shortcut thrust toward the east. The rheological characteristics influenced a brittle deformation zone on the western side of the basin, versus a more ductile deformation on the eastern side, which explains how the shortcut thrust was formed, allowing the convergence of the whole system toward the east.
The role of magmatism and erosion in the basin evolution remains poorly understood. More studies should focus on this, which could shed light on the problems that still exist in our understanding of the Andean uplift and its relationship with basin inversion.