Influence of pre-Andean crustal structure on Cenozoic thrust belt kinematics and shortening magnitude: Northwestern Argentina

Cordilleran-style orogens form during convergence of oceanic and continental plates and are characterized by long belts of continental magmatism and shortening. An active example of such an orogenic system is in South America, where shortening of the overriding plate results in continued growth of the Andes. In spite of the documentation of major along-strike variations in the style and magnitude of Cenozoic shortening in the Andes (e.g., Allmendinger et al., 1983; Isacks, 1988; Kley and Monaldi, 1998; Kley et al., 1999), there is not a considerable along-strike difference in the relative convergence velocity of the oceanic and continental plates nor in the age of the subducting oceanic Nazca plate (Oncken et al., 2006). In contrast, some of the observed spatial variations in the style and magnitude of Cenozoic retroarc shortening match with changes in pre-Andean stratigraphy and structure of the South American plate (e.g., Mpodozis and Ramos, 1989; Allmendinger and Gubbels, 1996; Kley et al., 1999). For example, in northernmost Argentina and Bolivia (Fig. 1; 17–23°S), Cenozoic thin-skinned shortening within a thick Paleozoic basin exceeds 300 km (Fig. 2; e.g., McQuarrie, 2002). Southwest of Salta, Argentina (Fig. 1; ∼25°S), where this thick Paleozoic basin was absent prior to Cenozoic time, steeply dipping reverse faults that are locally inverted normal faults are suggested to have accommodated <100 km of shortening (Fig. 2; e.g., Allmendinger et al., 1983; Grier et al., 1991). Despite this large along-strike variation in shortening magnitude and structural style, a corresponding major southward decrease in elevation and crustal thickness does not accompany this transition in the Central Andes (e.g., Isacks, 1988), prompting speculation that magmatic addition, tectonic underplating, and/or crustal flow may have contributed significantly to the crustal volume south of the thin-skinned Bolivian fold-and-thrust belt (Kley and Monaldi, 1998; Husson and Sempere, 2003; Gerbault et al., 2005).

Although inversion of rift faults and the distribution of pre-orogenic basins have long been suggested to influence the style of deformation in the Central Andes (e.g., Allmendinger et al., 1983), only recently have workers integrated geo-thermochronological results with structural analysis in southern Bolivia to show that the spatial extent of the Altiplano Plateau was largely controlled by the distribution of Mesozoic rift faults and was established by ca. 25 Ma (Sempere et al., 2002; Elger et al., 2005; Ege et al., 2007; Barnes et al., 2008). However, the influence of these pre-Cenozoic heterogeneities in influencing the kinematics of the thrust belt has not been sufficiently investigated in northwestern Argentina, despite the observation that early Andean deformation spatially correlates with Cretaceous rift basins (Kley and Monaldi, 2002; Carrera et al., 2006; Hongn et al., 2007; Insel et al., 2012). One study at ∼25.75°S, utilizing apatite thermochronometry of Cenozoic basin strata that spatially correlate to a Cretaceous rift basin, suggests that a lack of influence of preexisting structures on thrust belt propagation is reflected by a progressive eastward migration of Cenozoic exhumation (Carrapa et al., 2011). Likewise, stratigraphic and detrital provenance analyses within a fault-bounded basin in the Eastern Cordillera at ∼23.25°S record progressive eastward migration of the thrust belt and coupled foreland basin system and imply a lack of influence of older structures on Cenozoic thrust belt propagation (Siks and Horton, 2011). These and similar studies are focused upon Cenozoic strata that reflect regional deposystems and evolving sediment source areas. In contrast, the approach taken here involves (U-Th)/He apatite and zircon analysis of reverse fault hanging walls that were uplifted and exhumed during fault displacement. In addition to resolving the potential spatial heterogeneity of thrust belt kinematics implied by these prior studies, evaluating the importance of pre-orogenic crustal architecture (e.g., Allmendinger et al., 1983; Allmendinger and Gubbels, 1996; Kley et al., 1999) is critical for understanding the main factors influencing structural style relative to other models that largely neglect pre existing upper-plate heterogeneities and instead implicate climate (e.g., Lamb and Davis, 2003; Strecker et al., 2007), mantle dynamics (e.g., Russo and Silver, 1994; Sobolev and Babeyko, 2005; Schellart et al., 2007; Husson et al., 2012), or buoyant anomalies within the downgoing plate (e.g., Jordan et al., 1983; Isacks, 1988; Ramos, 2009). Also, in spite of the hypothesized importance of shallow subduction beneath the Central Andes during Miocene time (e.g., Ramos, 2009), few workers have evaluated the spatio-temporal correlation between the kinematic history of the thrust belt and an eastward migration of retroarc magmatism thought to indicate shallow subduction.

This paper focuses on an E-W transect across the Eastern Cordillera tectonomorphic province of the Andean retroarc thrust belt of northwestern Argentina (Fig. 1). Results presented here provide new constraints on the style, timing, kinematics, and magnitude of shortening of the fold-and-thrust belt at ∼24.75°S latitude. These results: (1) indicate that the northwestern Argentine thrust belt was deformed above a W-dipping décollement that transferred slip to a system of E-dipping back thrusts; (2) constrain the timing of eastward deformation propagation within the Eastern Cordillera and suggest that the Cretaceous rift architecture influenced the evolution of the thrust belt at this latitude; and (3) increase the estimate of the magnitude of shortening at this latitude, but they still suggest that significantly less shortening was accommodated south of the thin-skinned Bolivian fold-and-thrust belt. This work complements existing work and underscores the importance of the preexisting tectonic framework in controlling the spatial distribution of shortening, particularly during the nascent stages of thrust belt development. This, in turn, strongly influenced the evolution of the orogenic system.

Tectonomorphic provinces of the central Andean retroarc include, from west to east (Fig. 1A; Jordan et al., 1983): (1) the Puna Plateau, a relatively low-relief, topographically high (average elevation ∼4 km) region of internal drainage, where Paleogene thrust belt structures are mostly buried by Cenozoic sedimentary and volcanic rocks (this province is the southern continuation of the broad, lower-relief Altiplano Plateau of Bolivia); (2) the Eastern Cordillera, a high-relief, topographically high (peak elevations >6000 m), externally drained Cenozoic thrust belt with predominantly west-vergent structures in Argentina that transition northward into a bivergent system in Bolivia; and (3) the Santa Bárbara Ranges, a region near the modern deformation front that consists of mainly east-dipping reverse faults, transitioning along strike northward to the Subandes, a W-dipping thin-skinned thrust belt in northernmost Argentina and southern Bolivia.

The Paleozoic geology of western Bolivia consisted of >10 km of sedimentary rocks deposited in a back-arc setting during Cambrian to Carboniferous time (Fig. 2; e.g., Sempere, 1995). In northwestern Argentina, this Paleozoic basin was shallower and more limited in extent (Fig. 2; e.g., Starck, 1995; Egenhoff, 2007) and formed on the northeastern flank of a NNW-trending Paleozoic basement high (Transpampean arch; Figs. 1B and 2A; Tankard et al., 1995), thought to reflect a remnant Ordovician (“Ocloyic”) mountain belt (Mon and Salfity, 1995; Starck, 1995). Poorly constrained Late Devonian to Mississippian orogenesis from central Argentina to Peru (“Eohercynian/Chañic” orogenesis) eroded the original margins of the Ordovician to Carboniferous basin (Fig. 2A; Starck, 1995). Although Ordovician rocks are prevalent on the Puna Plateau west of this regional transect, Ordovician to Devonian rocks are not exposed in the Eastern Cordillera southwest of Quebrada del Toro (Fig. 1B), indicating that this locality may approximate the northeastern boundary of major Paleozoic deformation.

Widespread but low-magnitude Mesozoic extension affected much of western South America and has been variously attributed to collapse of a Carboniferous to Permian mountain belt (e.g., Kay et al., 1989), back-arc extension linked to subduction along the western margin of South America (e.g., Welsink et al., 1995), or failed rifting related to opening of the Atlantic Ocean (e.g., Grier et al., 1991). In northwestern Argentina, up to 5.5 km of Cretaceous sediment, the Salta Group, were deposited in concomitant rift basins (Salfity and Marquillas, 1994; Monaldi et al., 2008). Across much of the transect, Cretaceous strata unconformably overlie the Puncoviscana Formation, demonstrating that thick overlying Paleozoic strata present in southern Bolivia were absent in northern Argentina prior to Andean orogenesis (Salfity and Marquillas, 1994).

Major crustal shortening in the Central Andes began after the South American plate overrode the subduction zone during opening of the South Atlantic Ocean (e.g., Coney and Evenchick, 1994). Deformation propagation has been sporadic through time, but most authors agree that shortening in the central Andean thrust belt began in Late Cretaceous to early Eocene time in northern Chile (Sempere et al., 1997; Arriagada et al., 2006; Jordan et al., 2007) and propagated in an overall eastward direction. In northwestern Argentina, growth strata and apatite fission-track data reflect 40–30 Ma deformation in the eastern Puna Plateau and western Eastern Cordillera, followed by an enhanced period of exhumation from 20 to 15 Ma (Andriessen and Reutter, 1994; Coutand et al., 2001; Deeken et al., 2006; Hongn et al., 2007; Carrapa and DeCelles, 2008; Bosio et al., 2009; Carrapa et al., 2011). A pre–15 Ma (Reynolds et al., 2000), possibly Eocene angular unconformity across the Santa Bárbara Ranges (Salfity et al., 1993) may reflect an early phase of shortening or the eastward migration of a flexural forebulge (DeCelles et al., 2011).

In contrast to the 40–15 Ma evolution of the thrust belt, interpretations of the 15–0 Ma deformation history in northwestern Argentina vary significantly (Carrapa et al., 2011; Hain et al., 2011). Some authors use regional correlations of sedimentary rocks interpreted in the context of a flexural foreland basin to infer a progressive eastward migration of the thrust belt (e.g., DeCelles et al., 2011). Others suggest that an initially intact flexural depositional system was “broken” in mid- to late Miocene time as basement-involved reverse faults were initiated or reactivated away from what was once a continuous, along-strike thrust front and foreland basin (e.g., Hain et al., 2011).

The oldest unit exposed in the retroarc of the Central Andes is the Puncoviscana Formation, which consists of variably metamorphosed siltstone, argillite, and turbiditic sandstone, and it constitutes the majority of outcrop in the mapped area (Fig. 3). In the Cachi Range, the westernmost mountain range in this transect, the Puncoviscana Formation exhibits a gradational contact with the higher-metamorphic-grade La Paya Formation (Galliski, 1983). These rocks have Neoproterozoic to Cambrian protoliths and are correlative (Pearson et al., 2012); for this reason and for simplicity, these rocks are collectively referred to as the Puncoviscana Formation. In general, the Puncoviscana Formation exhibits finer grain size toward the west, including outcrops of chert south of La Poma (Fig. 3). Locally, however, metamorphic recrystallization has increased grain size. In contrast, in the eastern Lampasillos Range and Quebrada de las Capillas to the east (Fig. 3), the Puncoviscana Formation consists of 10–30-cm-thick fine-grained quartzites that are interbedded with 5–100-cm-thick slate beds. Closer to Quebrada del Toro to the east, these rocks alternate on the kilometer scale with low-grade metapelitic rocks characterized by 3–20-m-thick siltite and fine-grained quartzite interspersed with slate. In the eastern Lampasillos Range and Quebrada de las Capillas, primary depositional features are common, including flute casts and ripple marks (Fig. 4A), and some rocks are volcaniclastic. In the Lesser and Mojotoro Ranges (Fig. 3), the Puncoviscana Formation mainly consists of low-grade, fine-grained quartzite and metapelite.

Northwest of the Quebrada del Toro (Figs. 1 and 3), 526–517 Ma plutons (Hongn et al., 2010) intruded the Puncoviscana Formation. In the Cachi Range, 485–483 Ma granitoids (Fig. 2; Pearson et al., 2012) also intruded the Puncoviscana Formation and are among the northernmost outcrops of the Famatinian magmatic arc.

East of Quebrada del Toro (Fig. 3), the Middle to Upper Cambrian angular unconformity between highly deformed rocks of the Puncoviscana Formation and overlying quartzite and shale of the Upper Cambrian Mesón Group (Adams et al., 2011) is exposed. Lower Ordovician shale and quartzite of the Santa Victoria Group overlie the Cambrian Mesón Group. Together, these rocks constitute the majority of the Pascha, Lesser, and Mojotoro Ranges (Fig. 3).

Up to 2 km of synrift Cretaceous nonmarine conglomerate and sandstone of the Pirgua Subgroup disconformably overlie Paleozoic rocks in the Santa Bárbara Ranges (Salfity and Marquillas, 1994; Kley and Monaldi, 2002). In several localities, synrift depocenters correspond to Cenozoic reverse fault hanging walls (Kley and Monaldi, 2002). Much of the mapped region west of the Santa Bárbara Ranges represents the Salta-Jujuy High, considered to be a horst block in the central part of the Salta rift (Salfity and Marquillas, 1994). However, a minor remnant of Pirgua Subgroup is exposed near the Pascha Range (Salfity and Monaldi, 1998), and up to 3 km of strata are exposed in the southern portion of the Cachi Range (Fig. 3; Carrera et al., 2006). Overlying the Pirgua Subgroup and adjacent structural highs, there is a thinner but more regionally contiguous package of postrift Upper Cretaceous to Lower Eocene sandstone, limestone, and shale of the Balbuena and Santa Bárbara Subgroups (Salfity and Marquillas, 1994). In the Quebrada de las Capillas, the depositional contact between previously deformed Puncoviscana Formation and overlying Balbuena Subgroup is exposed (Figs. 3 and 4B). Here, the angular discordance is 45°–90°, and overlying sandstone contains angular clasts of Puncoviscana Formation quartzite; minor fault slip has also occurred along the primarily depositional contact.

Most workers attribute the accommodation space within which Balbuena rocks were deposited to postrift thermal relaxation and associated subsidence, an interpretation that is corroborated by the spatial coincidence of Paleogene depocenters with Cretaceous grabens (e.g., Starck, 2011). Elsewhere, workers have interpreted the Paleocene to Miocene succession as part of an eastward-advancing flexural foreland basin system (e.g., DeCelles et al., 2011). It is likely that the foreland basin related to growth of the Andes interacted in a complex way with waning thermal subsidence following Cretaceous extension.

The Paleocene–Lower Eocene fluvial and lacustrine deposits are overlain regionally by Middle-Upper Eocene paleosols that transition across strike to a disconformity in the eastern part of the Eastern Cordillera (Salfity et al., 1993; DeCelles et al., 2011). In turn, the paleosols are overlain by 2–6 km of Upper Eocene to Lower Miocene upward-coarsening fluvial and eolian deposits, preserved within the current transect at the eastern side of Quebrada de las Capillas (Fig. 3), and capped locally by middle Miocene to Pliocene fluvial, lacustrine, and alluvial-fan deposits (Hernandez et al., 1996; Starck, 1996; Reynolds et al., 2000, 2001; Echavarria et al., 2003; DeCelles et al., 2011).

In mid- to late Miocene time, retroarc magmatism migrated well east of the magmatic arc, with associated volcanic centers defining the NW-trending Calama–Olacapato–El Toro lineament that crosses the current transect near the Quebrada del Toro (Figs. 1 and 3; e.g., Allmendinger et al., 1983). Although no volcanic centers are exposed along the current transect, tuff, volcaniclastic, and flow deposits of intermediate composition (Mazzuoli et al., 2008) occur as interbeds in a dominantly clastic sedimentary succession near the northern Quebrada del Toro; the Las Burras (Hongn et al., 2010) and Acay (Petrinovic et al., 1999) plutons likely represent the intrusive equivalents of these rocks and are exposed just north of the current transect. Miocene magmatism was followed by eruption of shoshonites of likely Pleistocene age near the northern part and eastern margin of the Cachi Range (Fig. 3; Kay et al., 1994; Ducea et al., 2013).

In the higher-elevation regions along the transect, Pleistocene glacial deposits and landslides are abundant. Dark alluvial and landslide deposits, also of likely Pleistocene age (Trauth et al., 2000), unconformably blanket Cenozoic rocks and range-bounding reverse faults in many places at range margins. Terrace deposits of likely late Pleistocene age, in turn, overlie these sediments and are locally covered by Holocene alluvium. For simplicity, all Cenozoic sedimentary rocks are considered as one map unit (Fig. 3).

Work presented here was conducted along an ∼130-km-long E-W transect across the Eastern Cordillera at 24.5–25°S latitude (Figs. 1 and 3). Field work involved geological mapping and structural analysis, coupled with sample collection for U-Pb and (U-Th)/He analysis of detrital and igneous zircon and apatite. Much of the field work was accomplished by multiday foot traverses across high-elevation mountain ranges for which minimal published data exist. This regional transect was then linked with a balanced cross section across the Santa Bárbara Ranges to the east (Kley and Monaldi, 2002) as well as thermochronological results and a balanced cross section across the Puna Plateau to the west (Coutand et al., 2001). This work also builds upon regional-scale mapping across the Salta province (1:500,000 scale; Salfity and Monaldi, 1998) and west of Quebrada de las Capillas (1:250,000 scale; Blasco et al., 1996), and more detailed mapping in the northern Calchaquí Valley (Hongn et al., 2007), Quebrada del Toro (Marrett and Allmendinger, 1990), and northern Cachi Range (Pearson et al., 2012).

Within the study area, cleavage intensity and the grade of metamorphism increase to a maximum toward the west in the Cachi Range. Here, bedding is transposed, and rootless isoclinal folds prohibit assessment of stratigraphic facing direction beyond the outcrop scale. Although open to tight chevron folds are common across the transect (Fig. 4C), rocks generally are less tightly folded toward the east. Bedding, bedding-parallel cleavage, and primary cleavage within slate, phyllite, and quartzite generally dip moderately to steeply to the northwest or southeast across the transect (Figs. 3 and 5), with the exception of within the Quebrada de las Capillas and southern Lesser Range, where foliations dip toward the northeast or southwest. Within the Cachi Range, fine-grained metamorphic minerals include ∼1-mm-diameter anhedral cordierite that increases in size toward the core of the range; farther south, correlative rocks were subjected to granulite-facies metamorphism and anatexis (Pearson et al., 2012). Much of this deformation and metamorphism is thought to be Cambrian and Ordovician in age (Mon and Hongn, 1991; Mon and Salfity, 1995). N-S–striking ductile shear zones, also likely Ordovician in age, have also been documented in the Cachi Range (Pearson et al., 2012). East of the Cachi Range, rocks underwent lower-grade peak metamorphic conditions, and cordierite porphyroblasts are rare to absent. Although the orientations of bedding, bedding parallel cleavages, and primary cleavages are variable in orientation, on the scale of the transect, the Puncoviscana Formation usually dips more steeply than younger strata and is generally NE-SW striking and moderately to steeply NW-SE dipping (Fig. 5), an observation consistent with regional measurements of Puncoviscana Formation rocks (Piñán-Llamas and Simpson, 2006).

Prominent Cenozoic faults consist of two types: (1) N-S–striking, mainly E-dipping reverse faults that are expressed in the modern topography, juxtaposed rocks of markedly different age, and are locally demonstrably inverted Cretaceous normal faults; and (2) mainly NW-striking sinistral faults with minimal displacement that are discontinuous and often en echelon. These near-vertical faults have likely been active during Holocene time.

Major N-S–striking reverse faults generally dip 45°–60° toward the east and are characterized by up to 200-m-wide zones of intense fracturing, with rare through-going fault surfaces or fault gouge. In the footwalls of reverse faults where Upper Cretaceous and Paleogene rocks are preserved, overturned synclines are common, with steeply dipping axial surfaces that are subparallel to superjacent reverse faults (Fig. 3 and 4D). Corresponding hanging-wall anticlines were also observed in Cambrian–Ordovician, Cretaceous, and Cenozoic rocks and are consistent with fault-propagation folding being a dominant structural style in the Eastern Cordillera at this latitude (Fig. 3E). The (U-Th)/He zircon data obtained from the Cachi Range suggest that major antiforms in the hanging walls of reverse faults, also interpreted to be fault-propagation folds but at a more regional scale, accommodated the formation of up to 15 km of structural relief (Pearson et al., 2012). See notes in Figure 3 for descriptions of individual Cenozoic structures.

NW-SE-striking lineaments are visible in aerial imagery across the southern Central Andes (Allmendinger et al., 1983) and are associated with some of the main retroarc magmatic complexes (e.g., Riller et al., 2001). In some cases, these lineaments are pre-Cenozoic in age (Monaldi et al., 2008) and appear to segment Cenozoic contractional structures (Coutand et al., 2001). At the latitude of this study, these NW-SE–trending lineaments were locally confirmed as sinistral strike-slip faults based upon meter-scale offset of beds, brittle fault fabrics and kinematic indicators, tool marks, and NE-striking, antithetic dextral faults. In the Quebrada del Toro, portions of the Solá and Gólgota reverse faults (Fig. 3) that are coincident with the El Toro lineament (Salfity et al., 1976) are locally NW striking with horizontal slickensides, demonstrating that late strike-slip faulting exploited preexisting contractional structures (Marrett et al., 1994). Despite the prevalence of these faults, none are through-going at the regional scale, and the more significant of these faults are only characterized by tens of meters of displacement (Acocella et al., 2011).

U-Pb zircon geochronology by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS), following methods described by Gehrels et al. (2008), was applied to eight detrital samples to better constrain provenance, ages of deposition, and deformation. A tuff was also collected for U-Pb zircon analysis to constrain the age of growth strata in the Quebrada del Toro in the footwall of the Solá fault. Analytical details are available in the Supplemental File.¹

Detrital zircon ages are shown on relative age-probability diagrams (Fig. 6). In accord with Dickinson and Gehrels (2009), we constrain maximum depositional ages using the youngest age cluster in a sample defined by three or more overlapping analyses. For the igneous sample, we attempt to minimize errors resulting from inclusion of inappropriate analytical data in age calculations by reporting a weighted mean age (Ludwig, 2001) of concordant and overlapping ²⁰⁶Pb/²³⁸U ages, with final uncertainties that include all random and systematic errors (Fig. 7).

Apatite (U-Th)/He thermochronology is used here to constrain the timing and magnitude of rock exhumation, which we infer resulted from rock deformation. Our results supplement earlier work in the region (e.g., Coutand et al., 2001; Deeken et al., 2006; Carrapa et al., 2011) by placing ages of low-temperature thermochronometers in a structural context. (U-Th)/He thermochronometry of apatite generally reflects the time since cooling of the apatite below ∼70 °C (assuming an effective grain radius of 60 µm and a cooling rate of 10 °C/m.y.; Farley, 2000). Using apatite fission-track ages from vertical transects in the Cumbres de Luracatao (Fig. 1), Deeken et al. (2006) obtained a Miocene geothermal gradient of ∼18 °C/km. Using stratigraphic exhumation depths and lack of complete closure of the apatite fission-track system, Coutand et al. (2006) calculated a similar value of <18 °C/km for the Angastaco Basin ∼100 km to the south. This low geothermal gradient and a mean annual surface temperature of 10 ± 5 °C yield a closure depth of the (U-Th)/He system in apatite of 3–4 km; a more conservative gradient for a foreland basin (∼22 °C/km; Allen and Allen, 1990) yields closure depths of 2–3 km.

Rock samples collected for (U-Th)/He apatite thermochronometry consist of quartzites of the Puncoviscana Formation and Santa Victoria Group, and small Cambrian and Ordovician granitoids (two plutons in the Cachi and Mojotoro Ranges, and one dike in the Lampasillos Range; Fig. 3) that are exposed in the hanging walls of major reverse faults. Forty-nine individual apatites were dated from 11 samples (eight metasedimentary and three igneous); five grains were dated for seven of the samples, whereas four grains were dated from the other four samples (Table 1).

We constructed a restorable, area-balanced cross section at ∼24.75°S (Fig. 3) using the software LithoTect®. The main goal was to better constrain shortening estimates at this latitude and appraise along-strike heterogeneity in the magnitude of retroarc shortening in the Central Andes. Additionally, the results better constrain the subsurface structure within the basement-involved, locally inverted thrust system. The method here utilized forward modeling and iterative restorations. The thrust belt at this latitude involved previously deformed, strain-hardened rocks that were subjected to pre-Cenozoic metamorphism, behave as mechanical basement, and commonly deform into pop-up structures. For these reasons, hanging-wall strain during fault displacement was modeled using inclined shear in an orientation antithetic to faults (e.g., Groshong, 1989).

The cross section is constrained by regional mapping and structural analysis. Geometries of thrust sheets where Cambrian, Ordovician, Cretaceous, and Cenozoic rocks are exposed are better constrained than in the Lampasillos Range and intermittently along the Quebrada de las Capillas where exposed rocks are dominantly the Puncoviscana Formation (Fig. 3). Where possible, we used along-strike constraints (i.e., down-plunge viewing) to reconstruct the geometry and style of deformation. Eroded hanging walls were also drawn with the minimal displacement required to satisfy available observations, yielding “minimum” shortening estimates. However, along-strike preservation of likely breached fault-propagation folds east of Quebrada de las Capillas suggests that current shortening estimates there are not greatly underestimated (Fig. 3). Although rocks younger than the Puncoviscana Formation are generally not exposed in the Cachi Range, (U-Th)/He zircon and apatite thermochronological results constrain the geometries of eroded rocks (Pearson et al., 2012).

Marine-influenced carbonates of the Cretaceous Balbuena Subgroup were used as a regional reference horizon and provide a pre-Andean datum with which to estimate the minimum Cenozoic structural relief. The undeformed regional elevation of these rocks is constrained by interpretations of subsurface seismic lines at the flanks of the Santa Bárbara Ranges published by Kley and Monaldi (2002). Folds and tilted strata in fault hanging walls suggest shallowing fault dips in the subsurface (e.g., Grier et al., 1991; Kley and Monaldi, 2002; this study). This requires that slip accommodated on multiple faults at shallower crustal levels is transferred at depth to fewer structures that accommodate greater slip. This observation, coupled with the lack of major structural relief of basement rocks above their regional elevation, precludes the presence of deep décollements or whole-scale crustal faulting at this latitude. Although a two-décollement model, such as Kley and Monaldi’s (2002) for the Santa Bárbara Ranges, could provide a better explanation for local structures and deep seismicity (up to ∼25 km; Cahill et al., 1992), our work suggests that structural relief was accommodated above a regional, shallowly W-dipping décollement.

The back limb of the hanging-wall anticline in the Mojotoro Range is roughly concordant with the Mojotoro fault, providing a good constraint on the décollement at 9 ± 1 km depth there (Fig. 3). The shallower and deeper uncertainty limits represent, respectively, the décollement depth if the hanging wall were to deform by flexural slip and the uncertainty in dip of the Mojotoro fault, which may dip more steeply than hanging-wall strata (Fig. 2). A deeper décollement is incompatible with thrust sheet thicknesses within the Lesser, Pascha, and Mojotoro Ranges. Planar back limbs of these thrust sheets also suggest a décollement depth of ∼9 km. This décollement is comparable to Kley and Monaldi’s (2002) shallower décollement determined independently for the Santa Bárbara Ranges to the east. Structures west of the Quebrada del Toro generally involve thicker thrust sheets and require a deeper décollement at ∼11 km. Given that Balbuena Subgroup rocks at lower structural levels (e.g., in the Calchaquí Valley) appear minimally deformed, their post-Cretaceous structural relief constrains the displaced thickness of supra-décollement rocks, yielding a regional décollement dip of ∼2° west of the Quebrada de las Capillas (Fig. 3).

U-Pb analyses of detrital zircons and a U-Pb zircon age on a tuff help to constrain the provenance, timing of sediment source emergence, and the age of deposition of sedimentary and low-grade metasedimentary rocks in the region. U-Pb zircon results are available in the Supplemental Table.² For Puncoviscana Formation rocks, the youngest zircon populations dominate and vary slightly in age (Fig. 6), indicating a continuous supply of young zircons during deposition, and suggesting that maximum depositional ages obtained from these rocks likely approximate depositional ages (Fig. 6). Cambrian zircons also dominate detrital zircon populations from Ordovician rocks of the Santa Victoria Group; these ages are comparable to those obtained from the Puncoviscana Formation west of the Quebrada del Toro (Figs. 3 and 6; Pearson et al., 2012; this study). These Ordovician rocks also contain a prominent Neoproterozoic population of grains similar to the youngest age peak from a sample of Puncoviscana Formation collected by Adams et al. (2008; sample QT4 in Fig. 3) in Quebrada del Toro (Fig. 5). One detrital zircon sample collected from an outcrop of turbidites mapped as Puncoviscana Formation in the Quebrada de las Capillas (09DP47; Fig. 3; Salfity and Monaldi, 1998) yielded several grains with Ordovician ages. Two of these Ordovician analyses are reasonably concordant and have acceptable errors but do not define a robust population (Dickinson and Gehrels, 2009). We tentatively maintain that these rocks are Puncoviscana Formation but suggest that additional age evaluation of turbidites in this region is warranted.

Two detrital zircon samples and a tuff were collected from the Miocene Agujas conglomerate (Marrett and Strecker, 2000) exposed in the western part of Quebrada del Toro. These strata occur in the core of a tight ∼N-S–trending syncline in the footwall of the Solá and Gólgota faults, and their deposition reflects structural growth (Fig. 3; DeCelles et al., 2011). Consistent with the results of DeCelles et al. (2011), a U-Pb zircon age of a tuff from this section of 9.4 ± 1.6 Ma (Fig. 7) indicates late Miocene deposition (time scale used throughout this paper is that of Ogg et al., 2008). This is >5 m.y. after major exhumation in the Cachi Range (Pearson et al., 2012) that was concurrent with the establishment of a topographic high and internal drainage at the modern eastern margin of the Puna Plateau (Vandervoort et al., 1995; Coutand et al., 2006). Upper Cambrian and Ordovician zircons dominate detrital zircon populations of samples collected from the Agujas Conglomerate near this tuff (Fig. 6). Although recycling cannot be ruled out, this suggests that their sediment source during deposition was from the east, given that the main source of Ordovician grains on the Puna Plateau to the west was already hydrographically isolated.

Forty-seven individual apatite grains analyzed by (U-Th)/He thermochronometry yielded mostly Miocene and early Pliocene ages, with a lesser number of Paleocene to Eocene dates (Table 1; Fig. 7). Two other grains yielded Proterozoic and Paleozoic ages, with low effective U (eU = U + 0.235Th) and Th concentrations, making it unlikely that these anomalously old ages result from radiation damage that enhanced He retentivity; instead, these old ages may have been compromised by He implantation (Spiegel et al., 2009). Multiple grains that generally form well-defined Upper Miocene to Lower Pliocene age clusters are considered here to represent recent exhumation and cooling of rock samples (Table 1; Fig. 8).

Four Eocene (U-Th)/He apatite ages from the Lampasillos Range may represent an early signal of Cenozoic exhumation and cooling and are consistent with subtle Eocene growth strata documented in the Luracatao and Calchaquí Valleys (Bosio et al., 2009; Hongn et al., 2007). Unfortunately, the quality of these ages is questionable, given their very low eU concentrations; these and eight other analyses with eU concentrations of <5 ppm are not considered in the following age evaluations. Nonetheless, several (U-Th)/He zircon grains in the northern Cachi Range also yielded Eocene ages (Fig. 8; Pearson et al., 2012), which may attest to significant Eocene deformation at this time.

A (U-Th)/He apatite sample collected within the core of the Cachi Range yielded a grain age of 13.8 ± 0.5 Ma (Fig. 3; weighted mean ages henceforth), supplementing ca. 15 Ma (U-Th)/He zircon (Pearson et al., 2012) and apatite fission-track ages (Deeken et al., 2006) collected farther south in the same range. After this time, (U-Th)/He apatite results suggest that the location of exhumation jumped ∼75 km toward the east to the Lesser, Mojotoro, Pascha, and Zamanca Ranges, which record an ∼60 km westward-younging progression of cooling between 12.8 and 4.4 Ma toward the Quebrada de las Capillas (Figs. 3 and 8). Across strike ∼15 km west of the Zamanca fault, samples collected from the immediate footwall of the Mesada fault (this study) and ∼40 km along strike of there in the hanging wall of the Tin-Tin fault (Carrapa et al., 2011) disrupt the westward-younging trend, yielding (U-Th)/He apatite ages of ca. 7 Ma. From a regional perspective, these results build upon and modify preexisting results in the region and suggest a pulse of exhumation and deformation at or since Eocene time within the Luracatao Valley, Cachi Range, and Calchaquí Valley, followed by a second pulse in exhumation at 15–10 Ma that occurred mainly in the Cachi, Lesser, and Mojotoro Ranges. Exhumation then progressed ∼60 km westward from the Mojotoro and Lesser Ranges, with additional widespread unroofing occurring at ca. 7 Ma in the Quebrada de las Capillas and Lampasillos Range (Fig. 3).

Total Cenozoic shortening across the Eastern Cordillera from the area-balanced cross section is 95 km (45% over an E-W distance of 211 km; Fig. 3). The magnitude of shortening is not grossly underestimated within the Mojotoro, Lesser, Pascha, Zamanca, and Cachi Ranges (Fig. 3). However, this estimate is likely to be a minimum in the Quebrada de las Capillas and Lampasillos Ranges given the lack of hanging-wall strata to constrain deformed thrust sheet geometries. Adding 95 km of shortening to area-balanced shortening estimates accommodated by the Santa Bárbara Ranges toward the east (21 km; Kley and Monaldi, 2002) and a rough, line-length balanced shortening estimate of the Puna Plateau southwest of the current transect at ∼25°S (26 km; Coutand et al., 2001) results in a total shortening magnitude of 142 km (26%). This total encompasses the entire retroarc thrust belt east of the modern magmatic arc in northern Chile. Additional Cretaceous to Paleogene shortening in northern Chile (Arriagada et al., 2006; Jordan et al., 2007) would increase this estimate.

Although this work was focused in the Eastern Cordillera, field observations and limited mapping in the eastern Puna Plateau also suggest that the existing estimate of shortening accommodated there may be greatly underestimated. Neogene volcanic and sedimentary rocks buried many structures that are likely Cenozoic in age; where exposed, Ordovician rocks clearly accommodated major shortening. Although some of this shortening likely occurred during Paleozoic time, apatite fission-track results and subsurface seismic data from the eastern margin of the Puna Plateau demonstrate that significant Cenozoic shortening and exhumation occurred locally (Coutand et al., 2001; Carrapa et al., 2005). If the Puna Plateau accommodated shortening equivalent to the Eastern Cordillera (45%), the predicted total shortening across the plateau is 90 km, which would increase the retroarc estimate to 206 km. However, this is still ∼85 km less than predicted by mass balance estimates that assume an initially 40-km-thick crust and local isostatic compensation (Fig. 2; Isacks, 1988; Kley and Monaldi, 1998).

Time-averaged shortening rates using existing estimates (Coutand et al., 2001) and the balanced cross section yield a shortening rate of 1.9 mm/yr from 40 to 12 Ma for the entire thrust belt at 24–25°S (Fig. 9), which is probably a minimum value given that shortening within the Puna Plateau is likely underestimated. This was followed by shortening at a rate of 6.5 mm/yr in the Eastern Cordillera from 12 to 4 Ma, a sharp increase that occurred after the location of shortening jumped ∼75 km eastward to the Mojotoro, Lesser, Pascha, and Zamanca Ranges and Quebrada de las Capillas. A final episode of shortening at a rate of 5.3 mm/yr occurred within the Santa Bárbara Ranges from 4 to 0 Ma (Kley and Monaldi, 2002). The resultant time-averaged, long-term shortening rate since ca. 40 Ma is 3.6 mm/yr.

Coupled with previously published constraints, results presented here (Figs. 9 and 10) corroborate earlier work suggesting an overall eastward migration of the fold-and-thrust belt during Cenozoic time, with additional local westward propagation into lesser-deformed regions during times of inferred subcritical orogenic wedge taper. Cretaceous to Eocene growth structures and exhumation in what is now the forearc of northern Chile record early stages of Cenozoic shortening near the latitude of this study (Maksaev and Zentilli, 1999; Arriagada et al., 2006; Jordan et al., 2007). In northwestern Argentina, however, Cretaceous and Paleocene time marks a period of thermal subsidence that reflects waning Cretaceous rifting (e.g., Starck, 2011). By late Eocene time, contractional deformation and exhumation had begun in the eastern Puna Plateau (Coutand et al., 2001; Carrapa and DeCelles, 2008) and westernmost Eastern Cordillera (e.g., Deeken et al., 2006). Constraints on the eastern limit of observed Eocene exhumation are limited by data quality, but Eocene exhumation may be recorded by samples collected from the Luracatao and Cachi Ranges (Figs. 1 and 3; Deeken et al., 2006; Pearson et al., 2012; this study), which are prominent topographic features that mark the western boundary of the Cretaceous Salta rift. Eocene growth strata in intervening valleys may also reflect deformation in the western Eastern Cordillera at this time (Hongn et al., 2007; Bosio et al., 2009). Deformation and exhumation within the eastern Puna Plateau continued during late Eocene to early Oligocene time (Coutand et al., 2001), progressing westward into the interior of the Puna Plateau into the late Oligocene (Carrapa et al., 2005), as occurred in the Altiplano Plateau to the north (Fig. 10; e.g., Elger et al., 2005). From 20 to 15 Ma, rocks in the Luracatao and Cachi Ranges in the western Eastern Cordillera record another period of major exhumation (≥8 km locally; Deeken et al., 2006; Pearson et al., 2012); by this time, the modern eastern extent of the Puna Plateau was established (Vandervoort et al., 1995). Exhumation continued within the Cachi Range until <13.8 Ma (this study).

At 12–10 Ma, results presented here suggest that the deformation front shifted ∼75 km eastward to the E-dipping Lesser and W-dipping Mojotoro faults (Figs. 3, 9, and 10), which are exposed within the eastern portion of the Salta-Jujuy High of the Cretaceous rift system (Salfity and Marquillas, 1994). The Mojotoro Range is the northern continuation of the Metán Range, which is also bounded by a major W-dipping fault and spatially correlates with a >3-km-thick Cretaceous synrift depocenter (Salfity and Marquillas, 1994). Thus, ca. 10 Ma (U-Th)/He apatite ages from the Mojotoro Range (this study) support a proposed early phase of deformation of the same age in the Metán Range (Cristallini et al., 1997; Hain et al., 2011). Structural growth during deposition of the Agujas Conglomerate (10.5–9 Ma) also occurred within the Quebrada del Toro to the west (Fig. 3; DeCelles et al., 2011; this study), indicating that faulting occurred over a >50 km width during this time.

After initiation of significant exhumation above the Lesser and Mojotoro faults, deformation propagated progressively westward into the Salta-Jujuy High, within the E-dipping fault subsystem beneath the Lesser, Pascha, and Zamanca Ranges. A westward migration of deformation is supported by westward-younging (U-Th)/He apatite ages (Fig. 8) and westward shallowing of thrust sheet dips toward the Quebrada del Toro that record rotation of previously deformed rocks in the hanging walls of the younger, western faults. A >12.8 Ma (andesite K-Ar age; Mazzuoli et al., 2008) angular unconformity below the Barres sandstone in the footwall syncline of the Gólgota fault (Fig. 4F) and ca. 10 Ma growth strata above the W-dipping Solá fault (Fig. 7) attest to an early phase of contractional deformation here. However, (U-Th)/He apatite results from hanging-wall rocks structurally above these localities and abundant Ordovician zircon grains likely originating from the east indicate that most exhumation did not occur until after 6–4 Ma, which coincides with the timing of exhumation associated with the Mesada and Tin-Tin faults to the west (Carrapa et al., 2011; this study).

Poor exposure and a lack of suitable rocks for thermochronometry inhibit assessment of the age of deformation in the Santa Bárbara Ranges, which are the easternmost portion of the thrust belt at this latitude. Existing constraints suggest that a pre–15 Ma unconformity in the Santa Bárbara Ranges may represent an earlier phase of deformation (Salfity et al., 1993) or the passage of a flexural forebulge (DeCelles et al., 2011), followed by the main period of post–9 Ma (Reynolds et al., 2000), likely Pliocene uplift (e.g., Kley and Monaldi, 2002). This is indicative of another (>100 km) eastward jump in the location of deformation. The structural similarity between the Santa Bárbara and Zapla Ranges and the Mojotoro, Lesser, Pascha, and Zamanca Ranges to the west is intriguing: Both areas are bounded on the east by a major, W-dipping structure, and the latter region records a westward migration of exhumation toward the lesser-deformed Quebrada del Toro, where recent deformation has been documented (Hilley and Strecker, 2005). Progressive rotation of faults depicted by Kley and Monaldi (2002) also hints at a westward migration of deformation from the Piquete and Centinela synrift depocenters in the Santa Bárbara Ranges westward toward the minimally deformed Lavayén Valley, where active seismicity is focused within the growing Zapla Range (Cahill et al., 1992). This pattern of local migration of fault subsystems away from the foreland may be common during inversion of rift systems, as preexisting bivergent faults are reactivated, followed by new faults formed in footwalls of inverted structures as the sub critical orogenic wedge gains taper (Fig. 11).

A sporadic eastward migration of deformation interspersed with local, in-sequence westward-migrating faulting is a scenario that differs markedly from that encountered ∼100 km to the south (Carrapa et al., 2011). There, Carrapa et al. (2011) documented a progressive eastward migration of deformation. The discrepancy in results may reflect the influence of preexisting Salta rift architecture. To the south, mainly E-dipping, preexisting Cretaceous faults (Grier et al., 1991; Cristallini et al., 1997) were progressively inverted in front of the orogenic wedge in a roughly continuous E-W rift basin. In contrast, at 24–25°S, a lack of inversion-prone Cretaceous rift faults within the central portion of the horst block may have promoted an eastward jump in the location of deformation (Figs. 10 and 11). The original configuration of the rift also likely influenced the topographic expression of mountains across the Eastern Cordillera. South of the present transect, where the rift basin is better developed, W-dipping, antithetic pop-up structures are less common; instead, rocks were deformed within a major eastward-propagating back-thrust belt, with subdued topography east of Cachi reflecting strain accommodation by primarily E-dipping structures. In contrast, at the latitude of the Salta-Jujuy High, several W-dipping pop-up structures form sharp topographic boundaries on the eastern margins of ranges (Fig. 3).

Some workers have suggested that the segmented, “broken” nature of the Laramide and Sierras Pampeanas forelands reflects basement deformation that occurs during shallow subduction (e.g., Dickinson and Snyder, 1978; Jordan and Allmendinger, 1986). An eastward sweep of magmatism across the Altiplano-Puna Plateau from 25 to 15 Ma (e.g., Allmendinger et al., 1997) has led some researchers to suggest that southward-migrating shallow subduction occurred beneath much of the Central Andes during this time, presumably associated with oblique subduction of the Juan Fernández Ridge (Yañez et al., 2001). In the literature, shallow subduction is generally thought to cause magmatic lulls; the conventional model suggests that retroarc magmatism signals steepening of the subducting slab that follows shallow subduction (e.g., Dickinson and Snyder, 1978). However, retroarc magmatism occurred northwest of the Quebrada del Toro at ca. 15 Ma (Hongn et al., 2010), which predates by ∼5 m.y. the interpreted location of the Juan Fernández Ridge beneath 24–25°S (Yañez et al., 2001) and the ∼75 km eastward jump in the deformation front by the Eastern Cordillera. Similarly, in the southern Altiplano, enhanced retroarc deformation at 19–7 Ma followed the onset of retroarc magmatism by up to 8 m.y. (e.g., Allmendinger et al., 1997; Elger et al., 2005).

Timing constraints suggest that the inferred interval of slab shallowing corresponds closely with enhanced deformation and thrust belt propagation in Bolivia and northwestern Argentina. One possibility is that lithospheric delamination could result in enhanced magmatism due to an influx of asthenosphere, which may also create space beneath the upper plate for a shallowing slab, in turn promoting further foreland-ward propagation of the thrust belt. This model may explain some aspects of the Miocene kinematic history of northwestern Argentina whereby magmatism is followed by slab shallowing and an enhanced eastward propagation of shortening.

Several workers have documented that the modern spatial extent of the Altiplano Plateau was already in place before 25 Ma (e.g., Horton et al., 2001), following rapid Eocene and Oligocene advancement of the thrust front to regions of preexisting rift depocenters (Sempere et al., 2002; Elger et al., 2005; Oncken et al., 2006; Ege et al., 2007). Additional constraints presented here refine the timing and kinematics of the retroarc thrust belt in northwestern Argentina. During shallow subduction, the upper plate accommodates increasing strain. As retroarc thrust belts commonly involve a craton-ward thinning wedge of sedimentary rocks, enhanced foreland-ward propagation of the deformation front during shallow subduction would thus be likely to encounter older basement rocks with less overlying strata and greater preexisting hetero geneities. Plateau formation may be enhanced in regions of pre-orogenic foreland heterogeneities because distal uplifts increase orography and the formation of internally drained basins, in turn providing a positive feedback for formation of an orogenic plateau (Sobel et al., 2003). With continued deformation in the Sierras Pampeanas, the Puna Plateau may grow southward as intramontane basins accommodate additional strain that follows initial reactivation of preexisting heterogeneities, much like within the Salta rift of northwestern Argentina.

A primary observation by tectonicists working in the Andes is that the maximum magnitude of crustal shortening coincides with southern Bolivia, with shortening decreasing significantly along strike (Fig. 2B; Isacks, 1988; Kley and Monaldi, 1998). Hypotheses that seek to explain the along-strike change include the orientation of the relative convergence (Gephart, 1994), mantle flow beneath the long subducting slab (Schellart et al., 2007), and variations in the pre-orogenic stratigraphic architecture of the overriding plate (Fig. 2A; Allmendinger and Gubbels, 1996; Kley et al., 1999).

Results from the present study suggest that the pre-Cenozoic structural and stratigraphic architecture strongly influenced the spatio-temporal evolution of the thrust belt (Figs. 2 and 10). There is a striking correlation between the magnitude of shortening and distribution of Paleozoic strata in the retroarc of the Central Andes, suggesting that it is not mantle flow or convergence parameters that control the ability of the upper plate to deform, but rather the pre-orogenic architecture of the upper plate (Fig. 2; Allmendinger and Gubbels, 1996; Kley et al., 1999). Due to the apparent decreased ability of the upper plate to accommodate a portion of the convergence between the South American and Nazca plates, the relative convergence along strike at the subduction interface would be predicted to increase away from the Central Andes, which may explain the formation of the Bolivian orocline (e.g., Isacks, 1988; Allmendinger and Gubbels, 1996; Kley et al., 1999; Arriagada et al., 2008).

Coupled with existing estimates for the Puna Plateau (Coutand et al., 2001) and Santa Bárbara Ranges (Kley and Monaldi, 2002) near the latitude of the current transect, results presented here constrain a minimum estimate of 142 km for the total magnitude of shortening at 24–25°S (Figs. 2A and 3). These results also suggest that at least in the Eastern Cordillera at this latitude, this shortening estimate is not greatly underestimated. For comparison, the ∼95 km of shortening within this domain is <50% of that accommodated in the Eastern Cordillera of Bolivia (McQuarrie et al., 2008). If the kinematics of shortening within the Altiplano and Eastern Cordillera in Bolivia were largely controlled by the distribution of Mesozoic rift basins, then the northern and southern segments of the thrust belt are very similar, but differ in their magnitude of shortening. One possibility is that the Cretaceous rift basin in Bolivia was wider (e.g., Cominguez and Ramos, 1995) and more favorably oriented for inversion than in northwestern Argentina (Fig. 10), which allowed for a greater magnitude of distributed shortening during Cenozoic time. At the latitude of northwestern Argentina, the basement-involved Santa Bárbara Ranges also could not accommodate the large-magnitude (>100 km), thin-skinned shortening absorbed by the Bolivian Subandes because such a thick, pre-orogenic Paleozoic basin did not exist at this latitude (Fig. 2).

The shortening estimate calculated here is greater than existing approximations in this region (e.g., Grier et al., 1991; Coutand et al., 2001), but it is still ∼150 km less than predicted (Fig. 2; Isacks, 1988; Kley and Monaldi, 1998). Recent studies have suggested that local existence of anomalously thin crust beneath the Puna Plateau (∼42 km; e.g., Yuan et al., 2002) may indicate Cenozoic crustal loss. However, at ∼25°S, most geophysical studies suggest an ∼55-km-thick crust across much of the Andes (Yuan et al., 2002; Tassara et al., 2006; Wölbern et al., 2009). The initial crustal thickness in the Central Andes is poorly constrained. Assuming that the Andes are underlain by a 55-km-thick crust and had an initial crustal thickness of 35 km, ∼190 km of shortening would be required at this latitude. This is 48 km more than the ∼142 km of shortening documented here; given that shortening in the Puna and western Cordillera may be underestimated, we suggest that crustal addition (e.g., by crustal flow, magmatic underplating, etc.) may not be necessary to explain the observed crustal thickness at this latitude.

Regional geological mapping, structural analysis, and geo- and thermochronological results indicate that the northwestern Argentine thrust belt at 24–25°S was deformed above a W-dipping décollement that transferred slip to primarily E-dipping reverse faults in a major back-thrust belt that propagated in an overall eastward direction during Cenozoic time. Following rapid eastward propagation of the thrust belt at ca. 40 Ma, (U-Th)/He and U-Pb age dating of apatite and zircon constrains an ∼75 km eastward propagation event at 12–10 Ma, when the thrust front bypassed the central portion of a horst block in the Cretaceous rift system, followed by subsequent initiation of new faults in a subsystem that propagated toward the west into this region. Subsequently, deformation again migrated >100 km eastward to a Cretaceous synrift depocenter in the Santa Bárbara Ranges, likely followed by westward-migrating deformation to its current location in the Lavayén Valley. Approximately 100 km to the south, deformation migrated progressively toward the east through time with no local westward migration documented. This suggests that the architecture of the thrust belt was strongly influenced by the configuration of the Cretaceous rift, which likely influenced the discontinuous nature of deformation propagation in an overall foreland direction since Eocene time; these results are in accord with recent work in southern Bolivia.

A regional balanced cross section across the Eastern Cordillera, coupled with existing shortening magnitude estimates for the Santa Bárbara Ranges and Puna Plateau, increases the estimate of the magnitude of shortening at this latitude to ∼142 km, but it confirms that significantly less shortening was accommodated south of the thin-skinned Bolivian fold-and-thrust belt. We suggest that greater shortening than has been previously documented was probably accommodated within the Puna Plateau and ancient retroarc of northern Chile, and that crustal shortening alone may explain the observed thick crust at 24–25°S. The overall along-strike decrease in shortening magnitude is well explained by the distribution of pre-Cenozoic basins that are able to accommodate large-magnitude thin-skinned shortening. Coupled with a likely correlation of Cenozoic thrust belt kinematics with the spatial distribution of the Cretaceous rift, this suggests that the pre-orogenic architecture strongly influenced the style, kinematics, and magnitude of shortening, which, in turn, influenced the geodynamic evolution of Andean orogenesis.

This research was conducted as part of the Convergent Orogenic Systems Analysis (COSA) project, in collaboration with and funded by ExxonMobil. National Science Foundation grant EAR-0732436 supported data acquisition at the Arizona LaserChron Center. This work benefited from discussions with many people, including M. McGroder, F. Fuentes, R. Waldrip, J. Kendall, G. Gray, R. Bennett, S. Lingrey, T. Hersum, T. Becker, and R.N. Alonso. B. Ratliff provided assistance with LithoTect® software. F. Shazanee, C. Hollenbeck, A. Abbey, I. Nurmaya, and M. Hearn helped with mineral separations. Constructive reviews by J. Barnes and G. Hilley improved the manuscript.