We present a multidisciplinary study that constrains the development history of the southern part of the Central Andean Plateau, a prototypical noncollisional orogenic system. In the Antofagasta de la Sierra region of NW Argentina, data from sedimentary geology, sandstone modal composition, detrital zircon U-Pb geochronology, and apatite fission-track and (U-Th-Sm)/He thermochronology indicate that sediments accumulated in the late Eocene to early Oligocene, with a maximum depositional age of ca. 39–38 Ma provided by the youngest detrital zircon U-Pb dates. Provenance data, including paleocurrent indicators, sandstone modal composition, and detrital zircon U-Pb ages, point to prevailing western sources, including the Sierra de Quebrada Honda (a proximal source), the Ordovician to Late Cambrian Famatinian magmatic arc in western Argentina and Chile (a distal source), and the Permian–Triassic plutonic and volcanic rocks in coastal Chile (a distal source). Along the western basin margin, these strata were deformed by a basement-involved thrust fault that was active at ca. 25–20 Ma, as constrained by apatite fission-track and (U-Th-Sm)/He data. Analysis of new and existing U-Pb geochronologic data from both detrital and basement samples across the Puna suggests that the Sierra Laguna Blanca, a major mountain range in the southern Puna, remained buried during the late Eocene to early Oligocene. Our multidisciplinary data indicate that the southern Central Andean Plateau may have hosted a regional basin primarily formed by lithospheric flexure during the late Eocene to early Oligocene. Furthermore, this study refines the history of basin compartmentalization and exhumation of the major mountain ranges in the southern Puna, revealing propagation of deformation from the west to east, starting as early as the late Eocene and continuing to the mid-late Miocene.
Orogenic systems are inherently complex, and unravelling the many possible modes of deformation is challenging even when abundant geologic data are available. For example, various models have been proposed to explain the Cenozoic evolution of the Central Andean Plateau (Fig. 1), an important example of an orogenic system on a noncollisional plate boundary, including: distributed deformation along inherited crustal heterogeneities (e.g., Carrera et al., 2006; Grier et al., 1991; Hongn et al., 2007; Iaffa et al., 2013; Kley et al., 1999, 2005; Monaldi et al., 2008); range uplift modulated by deposition and erosion (e.g., Sobel and Strecker, 2003; Sobel et al., 2003; Hilley and Strecker, 2005); surface uplift, subsidence, contraction, or extension caused by lower lithosphere foundering (e.g., DeCelles et al., 2015a; Garzione et al., 2008; Schoenbohm and Carrapa, 2015; Zhou and Schoenbohm, 2015); and formation of a foreland basin through lithospheric flexure (e.g., Jordan and Alonso, 1987; Horton, 2005; DeCelles et al., 2011). These models are often posed as exclusive end members, whereas in reality, multiple modes of deformation may operate simultaneously, overlap partially or fully spatially, or dominate at different times in the history of orogenic development.
Our study presents an opportunity to resolve the apparent conflict between at least two major models for lithospheric deformation in the Central Andean Plateau. The first class of models emphasizes the reactivation of preexisting crustal heterogeneities, such as inherited deep-seated faults created during Cretaceous rifting. In the model, as the region underwent Cenozoic shortening, the availability of these weaknesses led to range uplift distributed across a broad region (e.g., Hongn et al., 2007; Sobel et al., 2003; Strecker et al., 2012). This “broken foreland” style of deformation has been documented in the Neogene Eastern Cordillera of NW Argentina, where former rift structures were reactivated, and deformation has occurred irregularly, rather than progressing sequentially across the region over time (e.g., del Papa et al., 2013; Hain et al., 2011; Hongn et al., 2007; Pingel et al., 2014; Sobel et al., 2003; Strecker et al., 2012). The alternative class of models argues that the Central Andean Plateau developed as an orogenic wedge, as evidenced by a parallel N-S–trending mountain front and a regional foreland basin that migrated east over time (e.g., Carrapa et al., 2011; DeCelles et al., 2011, 2015b; Horton, 2005). This wedge style of deformation has been proposed in the northern Puna, where sedimentary records reveal a foreland basin in the Paleogene that subsequently propagated to the east, riding over the Eastern Cordillera as early as the Eocene (e.g., DeCelles et al., 2011).
The southern Puna (∼26°S–27°S; Fig. 1) provides the key to resolving the apparent conflict between these models; it is the southern projection of the proposed orogenic wedge and is the southwestern boundary of the broken foreland of the Eastern Cordillera, and thus it may have been influenced by both modes of deformation. The southern Central Andean Plateau may have hosted a single regional basin during the late Eocene to early Oligocene, implying the existence of a regional flexural basin as early as the late Eocene (Zhou et al., 2016). Yet, the modern southern Central Andean Plateau is characterized by a series of high-relief, ∼N-S–trending ranges situated among small-scale salars (modern basins), resembling the tectonic style of the Eastern Cordillera where inversion tectonics has been invoked (e.g., del Papa et al., 2013; Hongn et al., 2007).
In this study, we provide new constraints on the timing and mode of deformation in the southern Puna in order to explore the viability of these models. We present structural mapping, logged sedimentary sections, zircon U-Pb geochronology, and apatite fission-track (AFT) and (U-Th-Sm)/He thermochronology from the Antofagasta de la Sierra region, constraining the provenance and the timing of deposition and deformation. We also present new zircon U-Pb ages from the Sierra Laguna Blanca, a basement range to the east. Our data show that the region was dominated by a propagating orogenic strain front associated with a flexural basin starting as early as the late Eocene, consistent with the wedge class of models. We also find evidence for reactivation of relict landscape or structures on the cratonic side of the orogen starting at the same time. Basins were compartmentalized by broken foreland–style range uplift from the late Oligocene to the present, but deformation proceeded from west to east in the manner of a propagating wedge. Thus, both modes of deformation may have influenced the region, with spatial and temporal overlap. Our reconciliation of two end-member models is useful for future studies that seek to understand the nature of orogenic deformation in other geologically complex regions.
The Central Andean Plateau hosts the world’s second largest continental plateau after the Tibetan Plateau, with an average elevation of over 3 km. It is divided at ∼22.5°S latitude into the Altiplano to the north and the Puna to the south (Fig. 1; Isacks, 1988; Allmendinger et al., 1997; Strecker et al., 2007).
Although volcanic rocks are widely distributed throughout the Central Andean Plateau, they are mostly concentrated along the modern active volcanic arc in the Western Cordillera. Volcanic rocks on the Puna range from large ignimbrite complexes to intermediate lavas to back-arc monogenetic mafic cinder cones (e.g., Kay and Coira, 2009). Cenozoic volcanic activity varies spatially and temporally (e.g., Trumbull et al., 2006; Kay and Coira, 2009), which may reflect the changing angle of the subducting Nazca plate (e.g., Trumbull et al., 2006; Guzmán et al., 2014; Kay and Coira, 2009) and varying fluxes of magma during the proposed Cordilleran cycle (DeCelles et al., 2009; DeCelles et al., 2015b). An important volcanic hiatus ca. 38–27 Ma in the southern Puna (∼26°S–28°S) has been identified (e.g., Trumbull et al., 2006; Guzmán et al., 2014; Kay and Coira, 2009). Volcanic activity resumed in the late Oligocene to early Miocene, reflected by mafic lava flows at the Segerstrom volcanic belt on the Chilean side of the southern Puna (e.g., Baker et al., 1987; González-Ferrán et al., 1985; Kay et al., 1994a).
The present-day morphology of the southern Central Andean Plateau, the adjacent Eastern Cordillera, and the Sierra Pampeanas to the south is characterized by sedimentary basins or modern salt flats (salars) bounded by ∼N-S–oriented ranges composed mainly of Paleozoic metasedimentary and intrusive rocks and Precambrian basement (Fig. 1; e.g., Schnurr et al., 2006). In the northern Puna, including the low-relief plateau region and the adjacent Eastern Cordillera, this tectonic style is linked to inversion of Cretaceous Salta rift structures (e.g., Grier et al., 1991; Hongn et al., 2007). The extent of rifting is defined by deposits observed in the Santa Barbara system, parts of Eastern Cordillera, and the western margin of the plateau (Fig. 1; Marquillas et al., 2005; Salfity and Marquillas, 1994).
Outcrops of Cenozoic sedimentary strata are largely obscured by recent volcanic and colluvial materials, especially in the southern Puna. However, existing data suggest that the oldest Cenozoic sedimentary rocks in the Puna are Eocene (ca. 38 Ma) and continental facies (Fig. 2; e.g., Carrapa and DeCelles, 2008; Jordan and Alonso, 1987; Kraemer et al., 1999; Zhou et al. 2016). For example, in the northern Puna, the Geste Formation has a maximum depositional age of ca. 37–35 Ma (e.g., Alonso et al., 1991; DeCelles et al., 2007; Pascual, 1983). To the east of the Puna, in the Angastaco region, the Quebrada de los Colorados Formation has a maximum depositional age of ca. 37.6 ± 2.0 Ma (Carrapa et al., 2011). In the Salar de Antofalla and Salina del Fraile regions of the southern Puna, an ash from the basal part of the lowest Cenozoic unit, the Quiñoas Formation (e.g., Adelmann, 2001; Canavan et al., 2014; Carrapa et al., 2005; Kraemer et al., 1999; Voss, 2002), yields a 40Ar/39Ar age of 37.6 ± 0.3 Ma (Kraemer et al., 1999). Also in the southern Puna, the lowest Cenozoic unit in the Pasto Ventura region yields a youngest detrital AFT age population of 38.1 +4.1/–4.0 Ma (1σ; Zhou et al., 2016). By at least the Miocene, possibly the late Oligocene (ca. 29–24 Ma), bedrock ranges in the Puna had started to form, leading to compartmentalization of sedimentary deposition (Adelmann, 2001; Alonso et al., 1991; Carrapa et al., 2005; Jordan and Alonso, 1987; Kraemer et al., 1999; Vandervoort et al., 1995; Voss, 2002).
We focus on the Antofagasta de la Sierra region (∼26°29′S, ∼67°56′W). Here, thick sedimentary units are exposed immediately to the east of the Sierra de Calalaste, a major ∼N-S–oriented bedrock range in the southern Puna (Fig. 1). The Sierra de Calalaste is mostly composed of low-grade metamorphosed Ordovician sedimentary and volcanic rocks (e.g., Martínez, 1995). Other rock types include metabasites, granitoids, and migmatitic gneisses, with the latter dated to ca. 390–420 Ma (Kraemer et al., 1999).
In addition, we present bedrock zircon U-Pb geochronologic data from the Sierra Laguna Blanca, an ∼N-S–trending bedrock range located to the east of the Antofagasta de la Sierra region, in the SE Puna (Fig. 1). The Sierra Laguna Blanca is composed of orthogneiss and granitoids with an inferred Ordovician age (e.g., Lucassen et al., 2000), but no zircon geochronologic data were available before this study.
Finally, we analyzed published detrital zircon U-Pb geochronologic data from the Pasto Ventura region adjacent to the Sierra Laguna Blanca, which hosts sedimentary strata as old as late Eocene (Zhou et al., 2016).
Geologic Mapping and Sedimentary Logging in the Antofagasta de la Sierra Region
We mapped exposed sedimentary units in the field and remotely (Fig. 3A) on Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) satellite imagery, aided by aerial photographs (Instituto Geografico Militar, Argentina) and additional satellite images and digital elevation models (DEMs) including Landsat, Google Earth, and Shuttle Radar Topography Mission (SRTM) data. In the western part of the mapping area, our interpretations are supplemented by published geologic maps (Martínez, 1995; Seggiaro et al., 2006) and a recent compilation (Schnurr et al., 2006). Because of the lack of geochronologic data, previous studies assigned various ages to the Antofagasta strata, including mid-Miocene (Martínez, 1995), Tertiary (Schnurr et al., 2006), and a combination of Permian and Eocene (Seggiaro et al., 2006). Three sedimentary sections were measured in the field, recording imbricated pebbles as paleocurrent indicators (Figs. 3B and 4).
Sandstone Modal Composition
Sixteen standard sandstone petrographic thin sections from the Antofagasta de la Sierra sections were stained for Ca- and K-feldspar and point counted (∼450 counts per slide) according to the Gazzi-Dickinson method (Fig. 5; Ingersoll et al., 1984). We follow the grain classification used by DeCelles et al. (2011; see GSA Data Repository1) and plot the results in Dickinson diagrams (Fig. 5; Dickinson and Suczek, 1979; Dickinson et al., 1983).
Zircon U-Pb Geochronology
Zircons from eight detrital samples from the Antofagasta de la Sierra region and four bedrock samples from the Sierra Laguna Blanca were dated using laser ablation–inductively coupled plasma–mass spectrometry (ICP-MS; a VG Series 2 Plasmaquad ICP-MS and a 213 nm New Wave laser system) at the Jack Satterly Geochronology Laboratory at the University of Toronto. Cathodoluminescence (CL) and backscattered electron (BSE) images were used to avoid cracks and target specific crystal domains when conducting laser ablation. Details are included in the Data Repository (see footnote 1).
We focus on analyses that yielded less than ±15% discordance, but we also report analyses with discordance ranging from ±15% to 40% (Fig. 6). Because 206Pb/207Pb ages are most reliable for older zircons, we report the 206Pb/238U age when younger than 1000 Ma and the 206Pb/207Pb age if the 206Pb/238U age is older than 1000 Ma in histograms and probability density function plots (Fig. 6; e.g., Gehrels, 2014). For ages with <±15% discordance in detrital samples, we also report peak ages (labeled in Fig. 6; Data Repository [see footnote 1]), each of which was calculated from several analyses. Peak ages were calculated using the AgePick program developed at the Arizona LaserChron Center.
AFT analysis was conducted on six detrital and two bedrock samples from the Antofagasta de la Sierra region at the Universität Potsdam using the external detector method (Table 1; Gleadow, 1981; Hurford and Green, 1983). Ages were calculated using the ζ calibration method (Hurford and Green, 1983) with a ζ value of 370.1 ± 12.6 (R. Zhou). Horizontal confined fission tracks, only track-in-tracks, were measured from as many c-axis-parallel grains as possible (Donelick et al., 2005). Dpar values (the etch figure length parallel to the c-axis; Donelick et al., 2005) were used to parameterize the kinetic properties of grains that were either counted or contained track lengths. We averaged at least four measured Dpar values from each grain and report corrected values following Sobel and Seward (2010) using a factor of 0.88 (R. Zhou). The BinomFit program (Brandon, 2002) was used under the auto-mode to deconvolve the component ages for the detrital samples (Fig. 7).
Single-Grain Apatite (U-Th-Sm)/He Thermochronology
We analyzed a bedrock sample (ABD22) with apatite (U-Th-Sm)/He thermochronology (AHe) from the Antofagasta de la Sierra region (Table 2). Seven single-grain aliquots were analyzed, and their grain dimensions and numbers of terminations were used to calculate the FT correction factor (Farley et al., 1996). Laboratory work was conducted at the Universität Potsdam and Deutsches GeoForschungsZentrum (GFZ) Potsdam; details are included in the Data Repository (see footnote 1).
ANTOFAGASTA DE LA SIERRA REGION
We divided the sedimentary rocks into five units (from oldest to youngest, N-5 to N-1) based on distinct color and facies variations. Three sections, S3 (552 m, oldest), S2 (1263 m), and S1 (1062 m, youngest), were then logged (Fig. 3). S1 covers the majority of N-2 and the lowest part of N-1; S2 covers the upper part of N-4 and lower part of N-3; S3 covers the upper part of N-5 and the lowest part of N-4 (Fig. 3B). A 500 m separation between the top of S2 and the base of S1 was estimated from dip of the strata and the measured surface separation from satellite images; sections S2 and S3 overlap. Therefore, the Antofagasta de la Sierra region hosts at least an ∼3.4-km-thick sedimentary succession. Contacts between units N-5 through N-2 are conformable, and paleocurrent indicators from these units reveal northeastward (n = 7), southeastward (n = 1), and eastward (n = 4) flow directions (Fig. 4). N-2 and N-1 are separated by an unconformity, and in contrast to the lower strata, paleocurrent indicators from N-1 indicate westward flow. The top of N-1 is overthrust by bedrock along a major basin-bounding fault, and only several tens of meters of N-1 are preserved.
N-5 is the lowest member and is deposited directly onto basement rock. It is characterized by fine-grained red sandstone and mudstone. It contains structureless, well-sorted, 10–20-cm-thick conglomerates (Fig. 3G) with ∼2–3-cm-diameter rounded to well-rounded clasts. The sandstone layers are composed of massive, laminated, and large-scale cross-stratified sandstones (Fig. 3G). The well-sorted conglomerate horizons are interpreted as gravel bar deposits, and the large-scale cross-stratified sandstones are interpreted as eolian dunes. N-4 is dominantly composed of thick (>20–50 cm) conglomerate layers that are matrix supported and planar bedded. Conglomerate layers are mostly laterally continuous and host thin (<10 cm), discontinuous layers of sandstone. Clasts in the conglomerate layers are ∼5–10 cm in diameter, poorly sorted, and angular to subangular (Fig. 3F). N-4 also less commonly contains ∼30–50-cm-thick sandstone layers composed of structureless medium-grained sandstones. N-4 is interpreted to have been mostly deposited by high-density flows during high-discharge events, as evident from the planar-bedded conglomerate. Accumulation and discharge may have varied, allowing the formation of interbedded sandstones. N-3 contains alternating dark-red mud-siltstone to medium sandstone layers, typically resulting in ∼3–5-m-thick sandstone and mudstone bed couplets. The sandstone layers lack cross-stratification, are dominantly massive, and occasionally contain discontinuous conglomerate (∼2–5 cm thick) horizons. N-3 is interpreted to reflect a fluvio-lacustrine environment, with shallow lacustrine being the dominant depositional environment. N-2 is characterized by massive, stratified and imbricated, lenticular conglomerates (Fig. 3D). Clasts within the conglomerates are typically 2–5 cm in diameter and angular to subangular. Boulders are rarely present. These conglomerates are interbedded with medium- to coarse-grained, cross-laminated and planar-laminated sandstones (Fig. 3D), forming ∼1 m conglomerate-sandstone pairs (Fig. 3D). N-2 is interpreted as fluvial-alluvial deposits, with the majority deposited in small-scale channels. N-1 is mainly composed of trough cross-stratified conglomerate to medium-grained sandstones that developed rough cross-bedding (Fig. 3C). It also contains massive conglomerate and massive medium-grained sandstone layers, with thin (<50 cm thick) mudstone that contains round carbonate nodules that are ∼1 cm in diameter. N-1 is interpreted as fluvial-alluvial deposits, with some subaerial exposure suggested by the presence of carbonate nodules.
Sandstone Modal Composition
The 16 sandstones examined in this study were lithic and feldspathic arenites, containing abundant monocrystalline quartz (Qm) grains and feldspar (F, including both K-feldspar and plagioclase; Figs. 5C, 5D, and 5E). K-feldspar (K) was more abundant than plagioclase (P). A large population of metamorphic (phyllite and schist) lithic grains (Lm) was observed (Fig. 5C). Polycrystalline quartz (Qp) and foliated polycrystalline quartz (Qpt) were also present in lesser amounts (Figs. 5D and 5E; Data Repository [see footnote 1]). Few volcanic grains were observed. Overall, the sandstone samples from the Antofagasta de la Sierra sections show some compositional variation, but the differences are insignificant for the majority of the sections (Fig. 5B), with an average Qm/F/Lt (Lt-total lithic grains) of 44/29/27 and an average Qt/F/L (Qt-total quartzose grains, L-total non-quartzose lithic grains) of 56/29/15. They contain more abundant polycrystalline quartz (Qp) and metamorphic grains (Lm) than sedimentary grains (Ls; Fig. 5B), with an average Qp/Lm/Ls of 32/56/11. These samples plot into the mixed zone among the recycled orogenic, cratonic interior, and basement uplift sources in the Dickinson diagrams (Fig. 5A). Abundant metamorphic grains, polycrystalline quartz, and feldspar imply that the sediments were derived from upper-crustal terranes composed of plutonic and medium- to high-grade metamorphic rocks. Therefore, the source region may consist of older orogenic crust (recycled orogenic setting) and continental blocks (Dickinson et al., 1983). Notably, the studied samples do not indicate a volcanic arc source (Fig. 5A).
The sandstone modal composition of the Antofagasta de la Sierra strata reveals similarity with that of the Quiñoas Formation in the Salar de Antofalla region to the west and with the lowest unit in the Pasto Ventura region to the east, both of which are late Eocene to early Oligocene (Adelmann, 2001; Carrapa et al., 2005; Kraemer et al., 1999; Zhou et al., 2016; Fig. 5A). When compositional data from all three locations are plotted together, including published data from younger strata, they define an evolving trend toward the volcanic arc region in the Dickinson diagrams (Fig. 5A), consistent with the increasing presence of air-fall volcanic materials, such as volcanic ashes, observed in younger sedimentary strata. Since the late Oligocene, the southern Central Andean Plateau and the associated Western and Eastern Cordilleras have been volcanically active, as recorded in widely distributed volcanic rocks (e.g., Trumbull et al., 2006; Kay and Coira, 2009; Kay et al., 2010). Therefore, we argue that the studied sedimentary sections in the Antofagasta de la Sierra region are no younger than late Oligocene based on the lack of volcanic input, and they are time-equivalent with the Upper Eocene to Lower Oligocene Quiñoas Formation in the Salar de Antofalla region to the west (Fig. 1).
Basin Inversion and Basement-Involved Deformation
Deformation in the Antofagasta de la Sierra region is dominantly controlled by a series of ∼N-S–trending structures (Fig. 3A). The study region is bounded to the west by a major W-dipping reverse fault, which carries bedrock of the eastern Sierra de Calalaste over the sedimentary section. The fault trace on the surface is evident by a distinct color change in satellite images that enables us to measure an ∼50°–60° dip based on the mapped fault trace and topography (Fig. 3A, lowest stereonet). Field measurements of associated minor faults suggest that the fault curves locally but overall strikes ∼N-S and dips steeply (Fig. 3A). A major N-S–striking syncline lies in the footwall of the thrust, affecting the majority of the mapped strata. The syncline has a subvertical axial surface, plunges ∼S, and is truncated by the basin-bounding reverse fault (Fig. 3A). To the north, the reverse fault cuts into bedrock of the eastern Sierra de Calalaste, and the major syncline dies out, replaced by a basement-cored anticline in the immediate footwall of the fault (Fig. 3A). Further east, sedimentary units have been folded into an ∼N-S–striking, open anticline-syncline pair, segmented by local reverse faults that strike NW-SE (Fig. 3A). Additionally, Quaternary terraces in the north are displaced by several ∼N-S–striking normal faults that have been inferred to be related to recent plateau-scale extension (e.g., Allmendinger et al., 1989), lower lithospheric foundering (e.g., Kay et al., 1994b; Schoenbohm and Strecker, 2009; Zhou et al., 2013), activation of regional-scale lineaments (e.g., Acocella et al., 2011), and/or subduction of the Copiapo Ridge (e.g., Álvarez et al., 2015), but these are beyond the scope of this study.
Magmatic Intrusion into Sedimentary Strata
The sedimentary strata in the Antofagasta de la Sierra region have been intruded by undated mafic sills (Fig. 8). The intrusions run parallel to sedimentary bedding, but can locally cut across sedimentary bedding at a low angle, and they are traceable for hundreds of meters on the surface (Fig. 8). The outcropping sills are 2–6 m thick and consist of medium- to fine-grained, dark-green rocks that display concentric weathering (Fig. 8J). Grain size in the sills decreases from the center to the ∼10–20-cm-thick, dark-purple chilled margins (Fig. 8K). Thin-section observations reveal a mafic composition and diabasic texture, with fine-grained, needle-like plagioclase crystals and visible olivine grains in a highly altered matrix. These sills suggest that postdepositional magmatic intrusions may have played a critical role in perturbing the geothermal field in three dimensions after basin formation. This inference is supported by the apatite thermochronology results (described later herein); samples may have been reset due to heating during (multiphase) sill intrusion despite being sampled at least ∼5–6 km (map distance) from the observed intrusions (Fig. 3A).
Detrital Zircon U-Pb Geochronology
Eight detrital samples yielded similar zircon U-Pb ages, despite being sampled across the entire ∼3.4-km-thick section (Figs. 3 and 6). Three major age groups were observed: ca. 200–300 Ma (Permian–Triassic), ca. 450–650 Ma (late Neoproterozoic, Cambrian to middle Ordovician) and ca. 900–1300 Ma (late Proterozoic; Fig. 6). The ca. 200–300 Ma ages were present in all samples, although only samples ADR1 and ADR25 yielded enough analyses to calculate peak ages at 257 Ma and 284 Ma, respectively (Fig. 6). The 450–650 Ma age group showed strong clustering at ca. 450–500 Ma (latest Cambrian to middle Ordovician) and weaker clustering at ca. 500–550 Ma (Cambrian; Fig. 6). All samples yielded a wide range of late Proterozoic ages at ca. 900–1300 Ma (late Proterozoic). Some early Proterozoic ages, ca. 1800–1900 Ma, were also present, with peak ages of 1848 Ma and 1900 Ma defined in samples ADR4 and ADR1, respectively (Fig. 6).
CL images yielded additional insight into provenance. Core-rim structures were observed on zircons with ca. 900–1300 Ma core ages and 450–650 Ma rim ages (Fig. 9). This is consistent with the material having been sourced in the Precambrian Antofalla terrane, which makes up the basement for most of the western-central Puna (Ramos, 2008, 2010). The Precambrian basement experienced multiphase metamorphism during late Proterozoic to early Ordovician time, resulting in the series of ages found in the zircon overgrowth (rims, Fig. 9). We also found a number of ca. 450–500 Ma zircons with distinct oscillatory zoning (Figs. 9C and 9D), implying a magmatic origin (Corfu et al., 2003), such as the plutonic rocks of the Ordovician to the latest Cambrian Famatinian arc in western Argentina and Chile (since ca. 495 Ma; e.g., Ducea et al., 2015). The ca. 200–300 Ma ages were exclusively found in zircons with distinct oscillatory zoning (Fig. 9E and 9F), pointing to a source in the Permian–Triassic plutonic and volcanic rocks in coastal Chile on the western margin of the Puna (e.g., López-Gamundí and Breitkreuz, 1997; Breitkreuz and Zeil, 1994).
Detrital samples also yielded a limited number of Cenozoic ages; we measured four grains in three samples that had ages of 39–38 Ma (Fig. 6). In sample ADR6, two Cenozoic ages were 38.3 ± 0.5 Ma and 39.4 ± 0.6 Ma, enabling us to calculate a weighted mean age of 38.69 ± 0.74 Ma. Th/U ratios for all four grains were 0.45 (ADR1), 0.65 (ADR7), 0.60, and 0.14 (ADR6), which are typical of igneous zircons. The ca. 39–38 Ma detrital zircon ages may offer a maximum deposition age of the hosting strata (e.g., Dickinson and Gehrels, 2009). They are consistent with our Upper Eocene to Lower Oligocene age inference for the Antofagasta de la Sierra sections based on structural geology, sandstone modal composition, and detrital apatite thermochronology.
Bedrock Apatite Thermochronology
We performed low-temperature thermochronology on two bedrock samples from the eastern edge of the Sierra de Calalaste in order to constrain the timing of exhumation, and therefore the timing of deformation, of the eastern range.
Sample ABD22 is a metamorphosed sandstone collected from the hinge zone of the basement-cored anticline in the north of the Antofagasta de la Sierra region (Fig. 3). It yielded an AFT central age of 19.9 ± 1.5 Ma with relatively long mean track length (MTL: 12.63 ± 2.38 µm; Fig. 10). The track-length histogram has a peak around 14–16 µm but notably also contains distributed short, ∼4–12 µm tracks (Fig. 10), implying relatively fast exhumation after a protracted period spent within the partial annealing zone (PAZ; e.g., Armstrong, 2005; Gleadow et al., 1986). Seven apatite grains were analyzed from the same sample, ABD22, using AHe thermochronology (Table 2; Fig. 10). Because the AHe ages cluster despite a wide range of eU values, we argue that the sample experienced rapid cooling from a temperature higher than the partial retention zone (PRZ) and calculate a mean AHe age of 19.9 ± 2.0 Ma (Fig. 10).
Sample ABD2 is a gneiss and was sampled in the hanging wall of the W-dipping, basin-bounding reverse fault (Fig. 3A). It yielded an AFT central age of 88.8 ± 7.6 Ma with significantly shortened mean track length (MTL: 9.72 ± 2.22 µm; Fig. 10). The track-length histogram contains a single peak around 10 µm (Fig. 10), suggesting that the sample has experienced partial resetting or spent a prolonged period of time within the PAZ prior to final exhumation (e.g., Armstrong, 2005; Gleadow et al., 1986).
We used the QTQt program (Gallagher, 2012) to invert the possible time-temperature (t-T) history for both samples (Fig. 11). Models were started at a high temperature of 160–180 °C, in order to anneal the AFT system (Fig. 11). For sample ABD22 (Fig. 11A), both AFT and AHe data were used in the inverse model, which only constrained the post–ca. 45 Ma history of the sample. Model results indicate that the data are best reproduced by a heating event at ca. 45–35 Ma, bringing the sample to ∼90 °C (the middle part of the PAZ; Fig. 11A). Holding at this subsurface temperature is required until ca. 35–25 Ma, followed by final cooling to the surface temperature at ca. 25–20 Ma (Fig. 11A). In sample ABD2 (Fig. 11B), the inverse model with AFT data provides constraints for its thermal history since the Cretaceous. The model favors cooling from the initial input temperature (160–180 Ma) to ∼60–70 °C, the upper part of the PAZ, by ca. 140 Ma (Fig. 11B). The temperature of the sample may have stayed unchanged from ca. 140 Ma to ca. 20 Ma, with cooling to surface temperature allowed, but not required, in the early Cenozoic (ca. 60–50 Ma; Fig. 11B). The sample is required to have resided at a subsurface temperature of ∼60 °C at ca. 35–25 Ma, before its final cooling to the surface at ca. 20 Ma (Fig. 11B).
The t-T histories of samples ABD2 and ABD22 both reveal a cooling event that brought these samples from at least ∼60 °C to surface temperature at ca. 25–20 Ma (Fig. 11). The dissimilarities in precooling temperatures between the two samples imply that they were located at different structural depths prior to the most recent cooling/exhumation episode, with ABD22 likely being deeper (higher temperature). In both cases, the samples may have held at a subsurface temperature of ∼60–90 °C at ca. 35–25 Ma (Fig. 11), in line with the timing of proposed late Eocene to early Oligocene (ca. 38–28 Ma) basin formation in the Antofagasta de la Sierra region.
Detrital Apatite Thermochronology
For detrital apatites, low-temperature thermochronometers, such as the fission-track system, record exhumation of their source regions if they are not thermally reset in the host basin after deposition. From the six detrital samples studied by AFT thermochronology, four yielded a range of AFT ages from ca. 20 to 200 Ma (Fig. 7). Notably, two detrital samples (ADR5 and ADR6) from the middle-lower part of the total sedimentary column yielded notably young ages at ca. 13–14 Ma, younger than age populations of samples at both lower and higher stratigraphic positions (Fig. 7). The single-grain AFT ages within the same samples are statistically consistent and result in similar pooled ages of 14.2 ± 1.0 Ma and 13.6 ± 1.2 Ma in ADR5 and ADR6, respectively. The mean track lengths from ADR5 and ADR6 are 13.41 ± 1.89 µm and 13.01 ± 2.70 µm, respectively; these are longer than from the rest of the detrital samples, which have mean track lengths of ∼10–11 μm (Fig. 7). We rule out burial heating as a cause for the ca. 13–14 Ma ages in ADR5 and ADR6 because they are located in the middle-lower part of the section, and they yielded younger ages than the lowest sample, ADR21 (Fig. 7). Instead, we favor the interpretation that these two young ages reflect postdeposition heating by magmatic intrusion. We did not observe sills nearby in the field, but they could be more proximal in the subsurface. Because the effect of postdepositional thermal event(s) on the other samples higher or lower in the section cannot readily be constrained, further interpretation of the detrital AFT population, such as a lag-time analysis (e.g., Coutand et al., 2006), is not advisable. Indeed, the lack of long (∼14 µm) track lengths in the rest of the detrital AFT samples supports the inference that they have been affected by postdepositional thermal event(s).
SIERRA LAGUNA BLANCA AND PASTO VENTURA REGION
Bedrock Zircon U-Pb Geochronology for the Sierra Laguna Blanca
U-Pb ages for zircon from the slightly strained, granitic basement samples of the Sierra Laguna Blanca yielded ages ranging from ca. 450 to over 2000 Ma. We calculated concordia ages (Ludwig, 1998) from the youngest data clusters for all samples, yielding ages of 510.4 ± 4.1 Ma (LB5369), 517.1 ± 3.3 Ma (LB4992), 510.7 ± 5.2 Ma (LB4621), and 474.4 ± 4.9 Ma (LB4025; Fig. 12; Data Repository [see footnote 1]), indicating the timing of partial melting events. Scattered ages as old as 2000 Ma indicate that rocks of the Sierra Laguna Blanca are a result of partial melting of Precambrian basement during the mid-late Cambrian to the earliest Ordovician, predating the Famatinian magmatic arc (ca. 495 Ma; e.g., Ducea et al., 2015).
Detrital Zircon U-Pb Ages for the Southern Puna
Because bedrock in the Puna consists mostly of Precambrian terranes that experienced multiple phases of metamorphism during the Proterozoic and Paleozoic, detrital zircon U-Pb ages obtained in the Puna commonly yield similar-looking age plots with peaks at ca. 400–600 Ma and ca. 900–1300 Ma (e.g., DeCelles et al., 2011; Zhou et al., 2016; this study). Here, we employ the K-S test (the Kolmogorov-Smirnov test; e.g., Berry et al., 2001; DeGraaff-Surpless et al., 2003; Guynn and Gehrels, 2010; Press et al., 1986) for a number of detrital zircon U-Pb data sets from the Puna in order to better distinguish possible sources. The K-S test functions to reject, rather than to realize, a null hypothesis that the two distributions came from the same parent population (e.g., Guynn and Gehrels, 2010). Two samples that fail the K-S test (p value < 0.05) were likely not derived from the same source, but passing the test does not prove correlation. The individual K-S test is found to too frequently reject the null hypothesis that samples come from a common source when they are identical (Saylor and Sundell, 2016). Therefore, we used the mean p value of a large number of K-S tests (n = 999) between randomly generated subsamples from each sample pair (Saylor and Sundell, 2016). The K-S test was conducted using age fractions of <600 Ma and >600 Ma, in addition to the full age range (Fig. 13). We chose a subsampling size of 20 (for age fractions <600 Ma and >600 Ma) and 50 (for full age ranges), limited by the size of smallest samples (Data Repository [see footnote 1]).
Two candidate sources were tested, the Sierra Laguna Blanca (Fig. 1B) and the Puncoviscana Formation, which outcrops in the eastern Puna. Because the four samples were taken across the range of the Sierra Laguna Blanca, they represent the spectrum of derived detrital zircon ages. For the Puncoviscana source, we used detrital zircon U-Pb ages from the Puncoviscana Formation reported by DeCelles et al. (2011), which was sampled from the eastern Puna.
We find that the tests for ages older than 600 Ma do not reject any sample pairs (mean p values > 0.05; Fig. 13), highlighting the fact that the ca. 900–1300 Ma ages were inherited from the Proterozoic terranes and carried through the Cenozoic sedimentary rocks. The mean p values from K-S tests younger than 600 Ma reject the possibility of the Sierra Laguna Blanca as a source for the lowest sample from the Pasto Ventura strata and all but one samples from the Antofagasta de la Sierra region, which were all deposited during the late Eocene to early Oligocene (Zhou et al., 2016; this study). Based on ages obtained from the eastern Puna, our analysis also rejects the Proterozoic–Lower Cambrian Puncoviscana Formation as a major sedimentary source for the Antofagasta de la Sierra and Pasto Ventura regions.
DEVELOPMENT OF THE REGIONAL LANDSCAPE OF THE SOUTHERN CENTRAL ANDEAN PLATEAU
With the data obtained in this study, we can now constrain the depositional age of the majority of strata and the exhumational age of the majority of bedrock ranges in the southern Puna, allowing a better understanding of the development history of the region (Fig. 14).
Several lines of evidence suggest that the strata in the Antofagasta de la Sierra region began to accumulate in the late Eocene to early Oligocene. First, field observations and sandstone modal composition analysis reveal little volcanic material in the strata, implying deposition during a volcanic lull in the late Eocene to early Oligocene (e.g., Trumbull et al., 2006). Second, the strata yield similar sandstone modal compositions to that of the Quiñoas Formation to the west and the lowest unit in the Pasto Ventura region to the east, both of which are argued to be late Eocene to early Oligocene (Fig. 5; Adelmann, 2001; Carrapa et al., 2005; Kraemer et al., 1999; Zhou et al., 2016). Third, the youngest detrital zircon U-Pb ages are clustered at ca. 39–38 Ma (Fig. 6), providing a maximum depositional age. Fourth, apatite thermochronology indicates cooling in the Sierra de Calalaste at ca. 25–20 Ma (Figs. 10 and 11), suggesting overthrusting of the section occurred at this time, constraining a minimum depositional age.
Our data demonstrate the existence of an Eocene–Oligocene regional basin connecting the Quiñoas Formation, Antofagasta de la Sierra strata, and Pasto Ventura strata. First, thermochronologic data demonstrate the two mountain ranges that compartmentalize the region today, the Sierra de Calalaste and the Sierra Laguna Blanca, had not yet been exhumed, and therefore could not have formed barriers to an integrated regional basin in the Eocene–Oligocene. Furthermore, paleocurrent indicators in both the Quiñoas Formation and Antofagasta de la Sierra strata indicate eastward paleoflow. Previous studies have documented a proximal source for the Quiñoas Formation in the Sierra Quebrada Honda (Fig. 1; Adelmann, 2001; Carrapa et al., 2005; Kraemer et al., 1999; Voss, 2002), which is composed of migmatitic gneisses and metabasites, intruded by several generations of granitoids and aplites (e.g., Kraemer et al., 1999). This is in line with sandstone modal composition data from the Antofagasta de la Sierra strata, which contain a high fraction of monocrystalline quartz (Qm), polycrystalline quartz (Qp), foliated polycrystalline quartz (Qpt), and metamorphic grains (Lm; see Data Repository [footnote 1]). Therefore, the data support a regional-scale sedimentary basin in the southern Puna (∼26°S–27°S) during the late Eocene to early Oligocene that covered the present-day Salar de Antofalla region, the Sierra de Calalaste, the Antofagasta de la Sierra region, the Sierra Laguna Blanca, and the Pasto Ventura region, extending at least ∼150 km from west to east (Zhou et al., 2016). The maximum thickness of the Quiñoas Formation to the west is ∼1.4 km, consisting of playa mudflat, alluvial fan and sheetflood, and distal braided river deposits (e.g., Adelmann, 2001; Carrapa et al., 2005). To the east in the Pasto Ventura region, only an ∼400 m section of the lowest unit is exposed, but it contains coarse cross-bedded red sandstone and siltstone, pedogenic carbonate nodules, and paleosol horizons, indicating slow accumulation and prolonged subaerial exposure in the late Eocene to early Oligocene, with possible periods of erosion or at least sedimentary hiatuses (Zhou and Schoenbohm, 2015; Zhou et al., 2016). Therefore, this early southern Puna basin was asymmetric, with its depocenter located in the Antofagasta de la Sierra region where the thickest (>3.4 km) Upper Eocene to Lower Oligocene units are found.
After its formation in the late Eocene to early Oligocene, this basin in the southern Puna was compartmentalized by several major basement-cored mountain ranges (Figs. 1 and 14), forming small-scale basins, some of which continued to accumulate sediment, while others experienced sediment starvation (e.g., Adelmann, 2001; Carrapa et al., 2005; Zhou et al., 2016). Carrapa et al. (2005) recorded exhumation of the western Sierra de Calalaste at 29–24 Ma using AFT thermochronology, which they related to rise of the western Sierra de Calalaste. This interpretation is supported by accumulation of the Chacras Formation at 24.3 ± 0.9 Ma (Adelmann, 2001; Carrapa et al., 2005). Exhumation of the eastern range of the Calalaste is younger, at ca. ∼25–20 Ma, as constrained by our modeled apatite thermochronological data (Figs. 11 and 14). To the east, AFT and (U-Th-Sm)/He data reveal that the Sierra Laguna Blanca experienced its most recent exhumation in the middle-late Miocene (ca. 15–10 Ma; Zhou et al., 2014). Active shortening is also documented through contractional syndepositional deformation in the ca. 11–10 Ma Pasto Ventura strata (Schoenbohm and Carrapa, 2015; Zhou and Schoenbohm, 2015). Last, bedrock along the southern margin of the Puna Plateau (north of the Fiambalá Basin; Figs. 1 and 14A) underwent rapid exhumation at ca. 25–15 Ma, as revealed by AFT data (Fig. 14; Carrapa et al., 2006).
DYNAMIC MODELS FOR THE SOUTHERN CENTRAL ANDEAN PLATEAU
In refining the development history of the southern Puna, we find that (1) this region was occupied by a foreland basin, with a western depocenter, during the late Eocene to early Oligocene; and (2) in the southern Puna, a wave of deformation initiated as early as late Eocene in the west and progressed to the east, as evidenced by documented cooling ages of major bedrock mountain ranges (Fig. 14).
We argue that, during the late Eocene to early Oligocene, lithospheric flexure was the dominant deformation mechanism in the southern Puna, based on the geometry of the basin (this study; Zhou et al., 2016). The proximal source region of this basin, the Sierra Quebrada Honda, marked the location of the orogenic strain front during the late Eocene to early Oligocene (Fig. 14B). Although superficially supportive of the wedge class of models for orogenic dynamics, there are two problems with extending this model into the southern Puna. First, studies in the northern Puna and the Altiplano place the orogenic strain front in the western plateau during the late Eocene (e.g., Carrapa et al., 2011; DeCelles and Horton, 2003; McQuarrie et al., 2005; Uba et al., 2006), requiring a map-view deflection of orogenic strain front from the northern to the southern Puna (Figs. 14A and 14C). Second, the Sierra Chango Real, the proposed eastern boundary of the early basin (Zhou et al., 2016), yields late Eocene AFT ages (Coutand et al., 2001), suggesting exhumation during the formation of this basin. This location was active, yet it was ∼180 km east of the proposed orogenic strain front.
We suggest that these two apparent conflicts may be resolved by invoking a role for inherited cratonic structures in modulating the propagation of the Cenozoic strain front. The Cretaceous rift developed mostly in the northern Puna Plateau and its Eastern Cordillera, covering both regions with thick Cretaceous rift deposits (Marquillas et al., 2005). Rift structures are prone to reactivation during shortening (e.g., Carrera et al., 2006; Hongn et al., 2007). In the Eastern Cordillera of the northern Puna, shortening may have preferentially activated the inherited Cretaceous rift-related faults (e.g., Hongn et al., 2007; Carrera et al., 2006); syndepositional deformation in the Quebrada de los Colorados Formation is in line with this inference (del Papa et al., 2013). Although Cretaceous rift deposits are not present in the southern Puna (e.g., Marquillas et al., 2005), thermochronologic data document Cretaceous cooling of several ranges on the eastern margin (e.g., La Quebrada; Carrapa et al., 2014), suggesting that the rift shoulders may have extended onto the plateau. Rapid exhumation of the Sierra Chango Real, as suggested by 38–29 Ma AFT ages and relatively long ∼13–14 µm mean track lengths (Coutand et al., 2001), may be a result of reactivation of rift-related faults. Alternatively, the southern Puna may indeed lack significant rift structures, but may instead dominantly host relict Paleozoic topography, supported by the observed Paleozoic zircon (U-Th-Sm)/He ages from the bedrock ranges in the southern Puna (Carrapa and DeCelles, 2015; Reiners et al., 2015; Zhou et al., 2014). The lack of rift-related faults may position the southern Puna less favorably for accommodating large amounts of Cenozoic shortening, impeding the ∼W-E propagation of Cenozoic deformation through the southern Central Andean Plateau (Fig. 14).
Sandstone modal composition, detrital zircon U-Pb ages, and modeled bedrock AFT data are consistent with the onset of sedimentary accumulation in multiple locations in the Southern Central Andean Plateau in the late Eocene to early Oligocene. These new data support the inference that the entire southern Puna was occupied by a regional, asymmetric sedimentary basin during the late Eocene to early Oligocene, bounded by topographic highs to the west in the Sierra Quebrada Honda, and to the east in the Sierra Chango Real (Zhou et al., 2016). This basin was compartmentalized by basement-cored range uplift starting in the late Oligocene, propagating from west to east during the late Eocene to mid-late Miocene, with the exhumation of the eastern Sierra de Calalaste being refined to ca. 25–20 Ma in this study. We also document magmatic intrusion in the mid-late Miocene that, although it does not outcrop extensively at the surface, may nonetheless have reset the detrital AFT record in the Antofagasta de la Sierra strata; we suggest caution in interpreting similar data in other arc basins. Our results imply that the southern Puna could have been dynamically connected to the regional central Andean foreland basin that initiated as early as late Eocene, highlighting the potential for the flexural behavior of the lithosphere despite inherited crustal heterogeneities, such as preexisting deep-seated faults. Along-strike comparison leads us to hypothesize that the different types of inherited structures in the northern and southern Puna may explain the observed map-view deflection of the orogenic strain front for the late Eocene to early Oligocene Puna (Fig. 14).
This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant to L.M. Schoenbohm. R. Zhou acknowledges support from DAAD (Deutscher Akademischer Austauschdienst), the Geological Society of America Foundation, and American Association of Petroleum Geologists Foundation. We thank the Materials Research Department of the GSI Helmholtzzentrum for heavy ion irradiation. We thank Barbara Carrapa, Peter DeCelles, Andres Folguera, Nicholas Perez, Associate Editor Stefano Mazzoli, Associate Editor Stephen Johnston, and Editor David Schofield for discussion, reviews, and comments that significantly improved this manuscript.
↵† Present address: The University of Queensland, St. Lucia, QLD 4072, Australia; .
Science Editor: David I. Schofield
Associate Editor: Stephen T. Johnston
- Received 19 July 2015.
- Revision received 14 June 2016.
- Accepted 19 July 2016.
- © 2016 Geological Society of America