Cenozoic Exhumation and Deformation of the Intermontane Pastos Chicos Basin in the Southern Central Andes: Implications for the Tectonic Evolution of the Andean Plateau (Puna) and the Eastern Cordillera Between 23° and 24°S, NW Argentina

The Andean Plateau of north‐western Argentina (Puna) at a mean elevation of ca. 4.2 km constitutes the southern continuation of the Altiplano; it is a compressional basin‐and‐range province comprising fault‐bounded, high‐elevation mountain ranges and largely internally drained basins with often thick sedimentary and volcaniclastic fill. Growing sedimentological and structural evidence supports the notion that the north‐western Argentine Andes between 22° and 26°S developed from an initial extensive broken‐foreland system that extended across the present‐day eastern Andean flank during the early to middle Eocene. However, compelling evidence of the tectonic history of this region is still missing. Here, we present new apatite fission track and zircon (U–Th)/He thermochronological data and U–Pb zircon ages from intercalated volcanic ash deposits from the Pastos Chicos Basin (23.5°S, 66.5°W) to constrain basin formation and the timing of major crustal deformation in the northern Puna. Inverse thermal modeling of the thermochronological data provides further temporal constraints on the late Cenozoic cooling history of the crust in this region and, by inference, on the timing of upper‐crustal shortening, range uplift, and basin formation in the northern sector of the present‐day Puna Plateau. Specifically, we argue for plateau‐wide distributed deformation in the Eocene between 23° and 24°S, followed by spatially disparate and diachronous deformation (Oligocene to Pliocene).

A steady foreland-ward propagation of deformation, typical of thin-skinned foreland fold-and-thrust belts and characterized by an advancing deformation front, is promoted by deep sedimentary basins with mechanically weak layers that are prone to form detachment horizons that drive shortening (e.g., Allmendinger et al., 1983). For example, the Subandean fold-and-thrust belt of the southern Central Andes in Bolivia and north-western Argentina is characterized by detachment surfaces in a thick Cambrian-Carboniferous sedimentary sequence (e.g., Allmendinger & Gubbels, 1996;Echavarria et al., 2003). However, this Paleozoic sequence pinches out south of 22°S and deformation in the fold-and-thrust belt is replaced by isolated basement-cored uplifts in the Santa Bárbara System (Kley & Monaldi, 2002;Mon & Salfity, 1995). In addition, the region of the present-day Puna, Eastern Cordillera, and Santa Bárbara System between 22° and 27°S experienced continental extension in the region of the Salta Rift between the Cretaceous-early Paleogene; this extensional province consisted of several sub-basins surrounding a central horst, the Salta-Jujuy high (e.g., Galliski & Viramonte, 1988;Marquillas et al., 2005;Salfity & Marquillas, 1994). Normal faults of the paleo-rift often follow Paleozoic thrust faults and metamorphic foliations (Hongn et al., 2010;Omarini et al., 1999). Many of these structures were reactivated during Cenozoic shortening of the northern part of the Puna Plateau (e.g., Allmendinger et al., 1983;Grier et al., 1991;Hongn et al., 2010;Kley et al., 2005).
Here, we constrain the onset of crustal deformation and associated range uplift in the northern Puna using apatite fission track and zircon (U-Th)/He thermochronology (AFT and ZHe, respectively) from basement rocks collected in the Pastos Chicos Basin and its bounding ranges in the Jujuy Province of north-western Argentina (ca. 23.5°S, 66.5°W, Figure 1). Thermal modeling of AFT data resolves the cooling, and, by inference, deformation history of the basin-bounding ranges of the Pastos Chicos Basin, revealing a distinct episode of basement deformation during the Oligocene and early Miocene. Combined with published data on regional deformation characteristics, our results provide constraints on the spatio-temporal deformation patterns during the early orogenic processes of Andean evolution that later culminated in the formation of the second largest orogenic plateau on Earth.

Geological Setting
The Pastos Chicos Basin is an intermontane basin located on the northern Puna Plateau between 23° and 24°S at ca. 3,800 m asl (Figure 1). It is bordered by the fault-bounded Sierra de Tanque and Sierra de Cobres to the west and east, respectively. Both ranges comprise Cambro-Ordovician metasedimentary rocks (Bahlburg, 1990) intruded by Ordovician granitoids (Bahlburg et al., 2016 and references therein) and have been uplifted along basinward-verging, NNE-striking, reverse faults ( Figure 2). The basin strata consist mainly of deformed Mio-Pliocene fluvial and alluvial redbeds, conglomerates, and volcaniclastic rocks that unconformably cover the Paleozoic basement rocks (e.g., Schwab, 1973). Whether the basement was at some point overlain by the regionally widespread Eocene Casa Grande Formation (Fernández et al., 1973) is not clear, but cannot be ruled out, as these strata exist in the adjacent Salina de Olaroz/Cauchari area to the west and potentially in the subsurface of the Salinas Grandes Basin to the east (e.g., Gangui & Götze, 1996;López Steinmetz & Galli, 2015;Seggiaro et al., 2015). The Mio-Pliocene basin deposits consist of the ≥10.8 ± 0.3 Ma Trinchera Formation (henceforth referred to as Upper Vizcachera Formation, following Seggiaro et al., 2015), the ca. 10-4 Ma Pastos Chicos Formation (Schwab, 1973;Schwab & Lippolt, 1976), and a suite of 6.8-6.5 Ma ignimbrites originating from the Cerro Coranzulí volcanic complex to the north (Figure 1; Seggiaro, 1994)-recently discussed in light of a complex, but single, caldera formation at about 6.6 ± 0.2 Ma (Seggiaro et al., 2019).
Today, the Pastos Chicos Basin is drained by the Río de las Burras (fed by the Río Coranzulí and Río Pastos Chicos) through a narrow bedrock gorge east of the town of Susques that routes runoff to the Salinas Grandes Basin. As a result, the basin morphology is characterized by dissected alluvial fans (associated with the Pastos Chicos Formation) and deeply incised canyons (e.g., Schwab, 1973), especially near the town of Susques, while, at the eastern flanks of the Sierra de Cobres in the adjacent Salinas Grandes Basin, the Río de las Burras has formed a ca. 900 km 2 alluvial fan. The basin-bounding Sierra de Tanque west of the Pastos Chicos Basin comprises a series of bivergent reverse faults placing Ordovician and Cretaceous strata onto the basin deposits, including the Pastos Chicos Formation (e.g., Seggiaro et al., 2015) (Figure 2). South of the town of Susques, outcrops of the Ordovician basement, unconformably overlain by units of the Upper Vizcachera Formation, show several west-verging folds and thrusts ( Figure 2, Henríquez et al., 2020). Similar basement outcrops exist within the central part and along the western margin of the Pastos Chicos Basin, indicating that the sedimentary fill is relatively thin. In contrast, the neighboring Olaroz/Cauchari and Salinas Grandes basins to the west and east exhibit ca. 4 km of syn-orogenic sedimentary fill (Coutand et al., 2001;Seggiaro et al., 2015). Linked faults along the center of the Sierra de Cobres (herein referred to as the Cobres Fault) juxtapose rocks of the Cobres Granite (478 ± 4 Ma, Insel et al., 2012) with Paleozoic rocks, and locally Cenozoic strata ( Figure 2).
The onset of Cenozoic shortening and mountain building in the northern Puna is recorded by syntectonic strata of the regionally extensive middle Eocene Casa Grande Formation at the eastern Puna Plateau margin (Coutand et al., 2001;Gangui & Götze, 1996;Montero-López et al., 2018). This deformation may have affected a ca. 250-km-wide region that is now represented by the present-day Puna Plateau (del Papa et al., 2013;Montero-López et al., 2021). To the west, the Cordillera de Domeyko in northern Chile records rock uplift and rapid exhumation by the middle Eocene to early Oligocene (e.g., Arriagada et al., 2006;Maksaev, 1990;Maksaev & Zentilli, 1999;Mpodozis et al., 2005). Coeval deformation is also documented along the Bolivian proto-Eastern Cordillera (e.g., Benjamin et al., 1987;Ege et al., 2007;Horton et al., 2002;Lamb & Hoke, 1997) Seggiaro et al. (2015)) showing volcanic ash (squares) and bedrock (numbered circles) sample locations. Numbers denote highest to lowest apatite fission track cooling ages for each vertical transect. The dashed black line (A-Aʹ) marks the centerline of the 20-km-wide topographic swath profile shown below. The profile shows average (black line) and maximum/minimum (gray envelope) elevations, as well as projected sample locations and simplified faults.
C o b re s R a n g e T a n q u e R a n g e  Andriessen & Reutter, 1994), Nevado de Cachi (ca. 40 Ma from deformed strata, , Cumbres de Luracatao (≥25 Ma from AFT, Deeken et al., 2006), and Sierra Chango Real ranges (ca. 30 Ma from AFT, Coutand et al., 2001). Late Oligocene deformation is also reported for the Sierra de Rinconada, which constitutes the northern continuation of the Sierra de Tanque, where provenance indicators in the early Oligocene to early Miocene Cabrería Formation (San Juan de Oro Basin) suggest range formation (Caffe & Coira, 2002;López Steinmetz & Montero-López, 2019). Similar observations associated with ranges bordering the Tres Cruces and Casa Grande areas have also been reported (Adelmann, 2001;Insel et al., 2012). Deformation and uplift that lead to the present-day topographic configuration of the adjacent Eastern Cordillera commenced at ca. 15-10 Ma (Deeken et al., 2005;Lapiana, 2021;Pingel et al., 2013Pingel et al., , 2014Siks & Horton, 2011).

Apatite Fission Track and (U-Th-Sm)/He Zircon Thermochronology
To constrain the cooling history, and by inference, the exhumation and deformation history of mountain ranges bordering the intermontane Pastos Chicos Basin, we collected 12 bedrock samples from four vertical profiles across the Sierra de Tanque and Sierra de Cobres for AFT and selected (U-Th-Sm)/He zircon thermochronology. Two granitic samples from the western basin-bounding Sierra de Tanque (ST profile) were collected from elevations of 4,080 and 4,190 m. Four samples were collected from Ordovician metasedimentary rocks from the western flanks of the Sierra de Cobres (SUS profile) between 3,580 and 3,760 m. Furthermore, samples have been collected from the northern (SC profile) and southern (COB profile) sectors of the Cobres Granite between elevations of 3,580 and 4,280 m. For sample locations see Figure 2 and Table 1. AFT and (U-Th-Sm)/He zircon (ZHe) thermochronology utilizes the retention of radiogenic decay products below a crystal system-and cooling rate-specific closure temperature to estimate the timing of exhumation and crustal deformation. These closure temperatures represent the base of the partial annealing (PAZ for AFT) and partial retention (PRZ for ZHe) zones. Typically, the temperature range of AFT PAZs is ∼120°C to 60°C (e.g., Gleadow & Duddy, 1981) and for ZHe PRZs ∼200°C to 130°C (Wolfe & Stockli, 2010), within and above which decay products are not preserved over geological timescales and thermochronological ages are reduced or set to zero.
Bedrock samples were crushed and sieved, and zircons and apatites were extracted by standard magnetic and heavy liquid methods.
AFT analyses used the external detector method (Hurford & Green, 1983); sample preparation followed the procedure outlined by . Apatites were embedded in epoxy, polished, and then etched for 20 s at 21°C with 5.5 N nitric acid. Following irradiation at the Oregon State University TRIGA reactor, muscovite detectors were etched for 45 min at 21°C with 40% hydrofluoric acid. Samples were analyzed with a Leica DMRM microscope at Potsdam University, Germany by A. Deeken (COB and SUS samples) or with a Zeiss Axioskop microscope at Stanford University, U.S.A. by I. Coutand (ST and SC samples). The microscopes were provided with a drawing tube located above a digitizing tablet and a Kinetek computer-controlled stage driven by the FTStage 3.11 program (Dumitru, 1993). Analysis was performed with reflected and transmitted light with 100× air objectives, a 1.25× tube factor, and 10× eyepieces. For age determinations, up to 20 good-quality grains per sample were selected at random and dated. Raw data were reduced using macTrack X software with ζ (for CN5 glass) = 389.9 ± 7.5 and 369.6 ± 5.1 for analysts Anke Deeken and Isabelle Coutand, respectively. Samples with P(χ 2 ) ≥ 5% represent a single population (Galbraith & Laslett, 1993;Green, 1981). For such samples, the pooled age is reported, otherwise the central age is reported. To assess the kinetic properties of the apatites, four Dpar measurements (etch-pit length parallel to the c-axis) were averaged from each analyzed crystal Ketcham et al., 1999). The Dpar values of the COB and SUS samples were calibrated using a linear correction factor of 0.87, calculated by comparing Dpar measurements of two age standards (Durango and Fish Canyon Tuff) with those obtained by Donelick et al. (1999) (following Sobel & Seward, 2010. All analytical data can be found in Table 1. Zircon (U-Th-Sm)/He analyses were carried out on two samples (COB2 and SUS1) at Potsdam University and Helmholtz Centre Potsdam-GFZ German Research Centre for Geosciences using laboratory procedures described in Galetto et al. (2021). Three single grain aliquots were measured per sample. Individual ages were calculated using standard alpha-ejection corrections assuming a homogeneous U and Th distribution (e.g., Farley et al., 1996;Hourigan et al., 2005). Both uncorrected and alpha-ejection corrected cooling ages and additional analytical data are reported in Table S1 in Supporting Information S2.

Thermal Modeling
Cooling histories were obtained from AFT-length distributions combined with AFT and ZHe ages by inverse modeling using the HeFTy thermal modeling software (version 1.9.3) (Ketcham, 2005).
Time-temperature constraints were set to model a simple cooling path with three constraints (simple cooling models). In cases where ZHe data were available (COB2 and SUS1), the inversion models of all samples from the corresponding profile started within time-temperature conditions of 200°C-300°C, at time intervals well before the obtained ZHe ages. In the case of the SC profile for which no ZHe data were available, we assume starting conditions based on the ZHe data from COB2 and set the initial time-temperature constraint to 160°C-180°C and 75-85 Ma. We consider this to be a viable option to increase the accuracy of the model, as both profiles are from the Sierra de Cobres, whose rocks at this time have probably undergone a similar cooling history. This was not an option for the ST samples. Instead, we chose to set the starting temperature-time constraint between 140°C-200°C and 90-100 Ma, well before the observed AFT ages of 19 and 24 Ma, to allow an unbiased development of time-temperature paths during model runs. Initial model runs identified a relatively broad range of time-temperature histories for all samples but showed clearly defined inflection points associated with distinctive cooling rate variations. To increase the model efficiency, additional model runs were carried out in which the originally identified inflection points were generously enclosed by a second constraint between 60°C and 120°C. The final constraint was set to 15°C ± 5°C at 0 Ma.
To test the hypothesis of burial heating by sedimentary overburden since the middle Eocene, we ran a second set of thermal models (reheating models). During these runs, we applied the same initial T-t conditions as for the simple cooling models, added an additional constraint to "force" an Eocene cooling (25-50 Ma, 30°C-80°C), and altered the following constraints to allow the potential reset of AFT ages (10-25 Ma, 80°C-150°C).
All model constraints are summarized in Table S2 in Supporting Information S2. In general, we ran 30,000 random simulations and derived the oldest tracks and T A s (total annealing temperature at which the oldest track is formed) from the best fitting time-temperature paths. The only exceptions are the simple burial and reburial models of SUS3, where we ran 300,000 random simulations to increase the number of model output results. For the other SUS models, this approach was omitted due to a very small number of track length measurements (N = 8 and 9).

U-Pb Zircon Geochronology
We collected samples of 11 volcanic air-fall deposits from the sedimentary basin fill of the Pastos Chicos Basin for U-Pb zircon dating. Following standard magnetic and heavy liquid mineral separation, zircons were handpicked, mounted in epoxy, and polished. U, Th, and Pb isotope analysis were obtained using a Laser Ablation Multi-Collector Inductively Coupled Plasma Mass Spectrometer at the University of California, Santa Barbara. The zircon (U-Th)/Pb analytical protocol follows that of Cottle et al. (2013) and Cottle (2014) using standard materials, data reduction, and correction methods described in Wiedenbeck et al. (1995), Andersen (2002), Jackson et al. (2004), Crowley et al. (2007), and Paton et al. (2010). Due to potentially significant pre-eruptive residence times and/or post-eruptive reworking, many of the analyzed samples show a non-uniform pattern of U-Pb zircon age distributions. Therefore, we systematically excluded the oldest ages from our calculations of an average zircon crystallization age until near-unity values for the mean square of weighted deviates (MSWD < 2) were achieved. In addition, we applied a generalized Chauvenet Criterion to the remaining grain ages to identify outliers (Vermeesch, 2018). In those cases where no coherent young population (less than three grains) was found, we infer the youngest 206 Pb/ 238 U zircon ages to represent a maximum depositional age. All analytical data, instrumental setup, and data on secondary reference zircons can be found in Text S1 in Supporting Information S1 and Tables S3-S5 in Supporting Information S2.

AFT Analytical Results
Bedrock samples from the basin-bounding ranges yielded dominantly late Oligocene to early Miocene AFT cooling ages (Table 1, Figure 3). Ages of the samples from the northern Sierra de Cobres (SC: 23.3 ± 0.9 and 29.2 ± 1.5 Ma) are slightly older than those from the southern Sierra de Cobres (COB: 20.8 ± 1.1 Ma to 24.7 ± 1.3 Ma) and from the Sierra de Tanque (ST: 19.1 ± 1.2 and 23.8 ± 1.1 Ma). In contrast, the four SUS samples, collected between the Cobres and the Susques faults, have late Eocene to early Oligocene AFT ages (31.4 ± 3.2 to 37.0 ± 2.8 Ma).
All ages from the SC and COB profiles are positively correlated with sample elevation, whereas those of the ST samples show no age-elevation relationship ( Figure 4). However, the age of sample ST3 (23.8 ± 1.1 Ma) is compatible with those of the COB samples. At first glance, ages of the SUS samples show no clear age-elevation relationship due to an inconsistent age distribution of the topmost samples of the vertical profile (SUS 1 and 2). Because the elevation difference (ca. 10 m) and horizontal separation (less than 600 m) of these samples is relatively small, we combined the AFT data from both samples (Sample SUS-1/2 in Table 1) and thus improved the age-elevation relationship. The apparent exhumation rate derived from the age-elevation relationship of the COB samples is 0.14 ± 0.06 mm/yr between ca. 25 and 21 Ma; 0.04 ± 0.01 mm/yr between ca. 29 and 23 Ma for the SC samples; and 0.09 ± 0.18 mm/yr between ca. 35 and 32 Ma for the SUS samples ( Figure 4).
The measured mean track lengths (MTLs) of the SUS1-3 samples and of sample COB3 are short (12.01 ± 0.79 to 12.45 ± 1.81 μm, SD); a characteristic typical of partially annealed samples. In contrast, the MTLs of the remaining samples are relatively long (12.91 ± 2.18 to 13.86 ± 1.54 μm, SD), indicative of faster cooling (Table 1, Figure 3). Also, the fraction of shorter tracks is slightly higher in the SUS samples and in the samples COB3 and COB5. The mean Dpar values (Table 1)  Only samples COB1 and COB5 provided a satisfactory number of track-length measurements (N = 86 and 100), whereas the number of track-length measurements of the remaining samples (N ≤ 41) is low compared to the ∼100 counts considered desirable for track-length inversion modeling. The modeled cooling histories of these samples are therefore not well constrained. The cooling histories of the SUS and COB samples are more robust due to additional information from ZHe data.

ZHe Analytical Results
ZHe analyses were carried out on three single-grain aliquots from two bedrock samples from the Sierra de Cobres. COB2 and SUS1 yielded average weighted, alpha-ejection corrected ZHe ages of 82.5 ± 1.3 Ma (single grain ages: 84.5, 77.3, and 85.7 Ma) and 243.5 ± 3.7 Ma (single grain ages: 244.8, 233.3, and 252.3 Ma), respectively (Table S1 in Supporting Information S2 and Figure 4). Both ages differ significantly from thermal modeling using K-Feldspar 40 Ar/ 39 Ar thermochronology on the COB1 sample reported by Insel et al. (2012), whose thermal model indicates temperatures of ≥250°C ± 20°C prior to 83 Ma ( Figure S1 in Supporting Information S1). Since this approach is controversial (e.g., Popov & Spikings, 2020), we use our new ZHe data to constrain the thermal history of the corresponding samples (COB and SUS series).

Simple Cooling Inversions
The simple-cooling inversions of the COB series yielded oldest track ages that generally increase with sample elevation, from ca. 26-27 Ma at the base to 32 Ma at the top (Table S2 in Supporting Information S2). The modeled T A s range between 100°C and 110°C ( Figure 5 and Figure S2 in Supporting Information S1). For samples SC4 and SC1, ages of the oldest tracks are ca. 28 and 32 Ma, respectively; the modeled T A s for both samples are ca. 110°C ( Figure 5 and Figure S3 in Supporting Information S1). The inversions of the SUS samples yielded ages of the oldest tracks ranging between 39 and 47 Ma; the modeled T A s are ca. 100°C ( Figure 5 and Figure S4 in Supporting Information S1). Finally, the inversions of sample ST3 and ST2 yielded ages of the oldest tracks of 30 and 24 Ma with T A s of ca. 105°C ( Figure 5 and Figure S5 in Supporting Information S1).   The COB models suggest rather steady cooling through the ZHe PRZ at ca. 1.3 ± 0.2°C/Ma, followed by an increase of the cooling rates to ca. 3.7 ± 0.3°C/Ma at about 23-20 Ma towards the present surface temperature (unless stated otherwise all cooling rates and their uncertainties represent average and 2 standard deviations calculated from each transects weighted mean paths). The thermal modeling of SC1 and SC4 reveals a similar break in slope (from 1.4 ± 0.2 to 3.2 ± 1.0°C/Ma), but the change in cooling rate appears to occur earlier, between ca. 28 and 22 Ma. Best-fit solutions for SC1 and SC4 suggest that surface temperatures may have been reached by about 20 and 10 Ma, which would require relatively high cooling rates of roughly 9 and 6°C/Ma, respectively. The simple cooling models of the SUS samples all indicate that following an initial, slow cooling of the samples through the ZHe PRZ and AFT PAZ at average cooling rates of ca. 0.3 ± 0.0°C/ Ma, samples began rapid cooling. By 33-29 Ma all SUS models exhibit a significant rate change towards rapid cooling until near-surface conditions are attained (on average 2.4 ± 0.3°C/Ma). Both ST models imply rapid cooling since at least 20 Ma (an earlier history is not well constrained) to surface temperatures at an average cooling rate of 3.9 ± 1.4°C/Ma. In both samples, the best-fit model suggest near-surface temperatures by as early as ca. 10 Ma, translating to a cooling rate of approximately 7-8°C/Ma for the rapid cooling phase.

Reheating Inversions
All reheating inversions yielded good statistical results (i.e., good-fit solutions), indicating that burial heating is a possible thermal scenario in the Pastos Chicos Basin. In most cases, however, the calculated oldest tracks (Table S2 in Supporting Information S2, Figure 5 and Figures S2-S5 in Supporting Information S1) postdate the potential reheating history and often even the reheating peak, which renders the pre-oldest track cooling history uninterpretable.
Exceptions are reheating inversions of COB5, SUS3, and SUS4, whose oldest tracks predate an initial cooling and subsequent reheating history. COB5 shows an initial cooling to 60°C between ca. 85 Ma (ZHe) and 41 Ma (3.0°C/Ma), followed by reheating to 108°C by 28 Ma (3.7°C/Ma) and a subsequent cooling to surface temperatures since ca. 28 Ma at rates of 3.3°C/Ma. SUS3 records an initial cooling to 60°C between ca. 245 Ma (ZHe) and 44 Ma (0.5°C/Ma), followed by reheating to 101°C by 30 Ma (2.9°C/Ma) and subsequent cooling to surface temperatures since ca. 30 Ma at rates of 2.9°C/Ma. Results from SUS4 are largely consistent with SUS3. Apart from the onset, the final cooling histories of all reheating-model runs are congruent with the respective simple cooling inversions.

U-Pb Zircon Results and Chronostratigraphy
Obtained U-Pb zircon ages from tuffs in the Pastos Chicos Basin range between 7.1 and 3.4 Ma and are summarized in Table 2. These ages partly confirm earlier results and help further refine the chronostratigraphy of tectono-sedimentary processes in the Puna. Based on our radiometric data and geological field observations, we are able to clearly distinguish between the strongly deformed, ca. 7.1-6.5 Ma redbeds of the Upper Vizcachera Formation below the Coranzulí Ignimbrite (6.6 Ma) and the overlying, largely undeformed, ca. 5.6-3.4 Ma deposits of the Pastos Chicos Formation.  Error bars indicate ±1σ. Dashed lines through AFT ages represent error-weighted linear fit through age-elevation data using a Monte-Carlo approach (Browaeys, 2022); the inverse slope of this line is the apparent exhumation rate (mm/yr).

Cooling History of the Pastos Chicos Basin
Most of our thermal inversion models do not allow an interpretation of the thermal history of the mountain ranges bordering the Pastos Chicos Basin in terms of an Eocene reheating scenario (except COB5, SUS3, and SUS4). This seems contradictory given the presence of the regionally widespread Eo-Oligocene Casa Grande Formation in the neighboring basins of the Puna (up to about 1 km thick, e.g., Gangui & Götze, 1996;Henríquez et al., 2020Henríquez et al., , 2022López Steinmetz & Galli, 2015;Seggiaro et al., 2015) and the Eastern Cordillera (ca. 700 and 1,400 m thick in the Casa Grande and Cianzo basins, Montero-López et al., 2018;Siks & Horton, 2011). Yet, there are no documented outcrops of the Casa Grande Formation to suggest that the Pastos Chicos Basin was also buried beneath thick Eo-Oligocene strata (Figure 2). In this context, we refrain here from a detailed discussion of an Eocene reburial episode and instead focus on the results of our simple cooling inversions, whose thermal history largely overlaps with the final cooling history of the reheating inversions ( Figure 5 and Figure S2-S5 in Supporting Information S1).
In the following, we estimate the onset of exhumation considering the range of good-fit model solutions of the highest and/or most reliable samples from each elevation transect, that is, COB5, SC1, SUS3, and ST2 ( Figure 5).   Several of our thermal models suggest that rapid cooling was followed by early attainment of near-surface conditions around 20 Ma (SC1 and SUS3) and 10 Ma (COB5 and ST2) associated with much lower cooling rates. As previous studies have shown, this may be related to the removal of easily erodible sediments overlying resistant bedrock (e.g., Deeken et al., 2006;. As rock uplift continues, more resistant bedrock is exposed, resulting in a decrease in exhumation while surface uplift increases. Attainment of near-surface temperatures at the northern Sierra de Cobres by 20 Ma approximately corresponds with the onset of deformation in the adjacent Sierra de Tanque and with uplift of the Sierra de Lina about 35 km to the west sometime after 22 Ma (Figures 1 and 6, Henríquez et al., 2020). Alternatively, reduced cooling rates may also indicate episodes of tectonic quiescence; however, the fact that we observe deformed basin strata up to 6.5 Ma suggests that shortening, and thus rock uplift, continued in the area of the Pastos Chicos Basin well into the late Miocene.
In summary, our thermochronology combined with published data suggest that final exhumation and deformation in the Pastos Chicos Basin began between ca. 30 and 20 Ma (late Oligocene) in the eastern Sierra de Cobres, propagated westward to the Sierra de Tanque and Sierra de Lina by ca. 20 Ma (early Miocene, Henríquez et al., 2020), and then shifted eastward ( Figure 6). For example, (a) AFT thermal modeling suggests exhumation of the Sierra Aguilar to the east at about 20-16 Ma (Deeken et al., 2005;Henríquez et al., 2020;Lapiana, 2021); followed by (b) the exhumation of the Sierra Alta (15-10 Ma, Deeken et al., 2005;Henríquez et al., 2022); (c) deformation associated with uplift of the Aparzo and Tilcara ranges by ca. 10 Ma (Lapiana, 2021;Siks & Horton, 2011), and finally (d) the deformation of sectors of the Santa Bárbara System after 5 Ma (including the Zapla anticline, Kley & Monaldi, 2002). Thus, although deformation ultimately shifted eastward, its progression toward the foreland was diachronous. Note that this diachronous development is much more pronounced farther south, especially in the Eastern Cordillera south of 24°S (Mio-Pliocene) and during the Oligocene south of 26°S (e.g., Hain et al., 2011;Pearson et al., 2013;Zapata et al., 2020).

Mio-Pliocene Basin Evolution
So far, the chronostratigraphy of the Pastos Chicos Basin has been mainly tied to a limited set of K-Ar dates (and stratigraphic correlation) originating from the adjacent Olaroz/Cauchari Basin to the west and from the southernmost Pastos Chicos Basin (e.g., Schwab, 1973;Schwab & Lippolt, 1976). Using our results from thermal modeling and U-Pb zircon geochronology, we refine the stratigraphy of the Pastos Chicos Basin and reconstruct the tectonic-sedimentary evolution of this region on the Puna Plateau.
Considering the possible lag times between the thermochronologic signal and accelerated surface uplift (discussed in Section 5.1), topographic growth of ranges bounding the Pastos Chicos Basin, and thus development of significant sedimentary accommodation space, likely occurred sometime after the modeled onset of final exhumation (30 and 20 Ma). This may be indicated by the oldest Pastos Chicos Basin strata, the late Miocene redbeds of the Upper Vizcachera Formation that unconformably overlie deformed Ordovician basement rocks (e.g., Figure 7). As older sedimentary strata appear to be missing, these Miocene strata (Upper Vizcachera Formation) could represent an erosional response to surface uplift.
The Upper Vizcachera Formation in the Pastos Chicos Basin is mainly composed of fine-grained, continental, fluvial redbeds and intercalated volcanic ash layers. Most of the outcrops of this unit are strongly deformed and show a basal unconformable contact with the Ordovician basement. Our U-Pb zircon ages constrain deposition of this unit to >7-6.5 ± 0.1 Ma, consistent with the timing of the eruption of the Coranzulí volcanic complex and the regionally widespread emplacement of ignimbrites at about 6.6 ± 0.2 Ma (Seggiaro et al., 2019). Subsequently, the Pastos Chicos Formation was deposited. U-Pb zircon ages from ash deposits indicate that the Pastos Chicos Formation was deposited between at least 5.6 and 3.4 Ma. Field observations indicate that this unit consists primarily of coarse-grained sand and conglomerates deposited in alluvial fans sourced from the Sierra de Tanque to the west. At several locations corresponding to more distal sectors these alluvial fan deposits interfinger with well-stratified and undeformed layers of fine-grained, fluvio-lacustrine redbeds similar to the 10.1029/2022TC007487 14 of 23 Upper Vizcachera Formation. The ignimbrite and overlying Pastos Chicos Formation are mostly undeformed, suggesting that basin-internal deformation largely ceased around 6.6 Ma. Only along the western basin margin did deformation continue until at least 3.4 Ma, as indicated by tilted strata of the Pastos Chicos Formation that are part of a buttress unconformity (Figure 7).

Regional Implications
Although our data do not unequivocally demonstrate Eocene deformation at the location of the present-day Sierra de Cobres and Sierra de Tanque, we consider it highly unlikely that these areas were not deformed at that time, given the large spatial extent of contemporaneous deformation and the position of the two ranges within this extensive deformation zone (Figure 8). For example, during the middle Eocene, the present-day Sierra de Cobres was aligned along a deformation zone that extended 600 km from the Tupiza area in southern Bolivia at 21.5°S to the Sierra Chango Real at 27°S (Figure 8a, Coutand et al., 2001;Deeken et al., 2006;Ege et al., 2007;Payrola et al., 2009). Likewise, the neighboring Sierra de Tanque was aligned with the Sierra de Rinconada and Filo de Copalayo ranges, to the north and south, respectively, which also record Eocene deformation (Carrapa & DeCelles, 2008;López Steinmetz & Montero-López, 2019). Across-strike, Eocene deformation affected a region more than 300 km wide that presently comprises the northern Puna Plateau. From west to east, mountain ranges with documented Eocene deformation at the latitude of the Sierra de Tanque and Sierra de Cobres (ca. 23°-24°S) include the Cordillera de Domeyko in northern Chile at ca. 69°W (e.g., Arriagada et al., 2006;Maksaev, 1990;Maksaev & Zentilli, 1999;Mpodozis et al., 2005); the Sierra de Lina at 66.8°W (Henríquez et al., 2020); and the Sierra Aguilar and Sierra Alta as part of the present-day Eastern Cordillera (Montero-López et al., 2018, 2021. Including the studies listed above, many researchers have documented Eocene deformation throughout north-western Argentina and northern Chile with an area extending from approximately 22° to 27°S and from 69° to 65°W (Figure 8a, e.g., del Papa et al., 2013;Henríquez et al., 2020Henríquez et al., , 2022Montero-López et al., 2021;Payrola et al., 2020). A more comprehensive list of references used to determine the onset of deformation on a more regional scale (incl., Aramayo et al., 2017;del Papa et al., 2021;  is provided in Table S6 in Supporting Information S2 and a graphical representation of the data is shown in Figures 8 and 9. As noted above, deformation was virtually ubiquitous during the Eocene in an area that presently constitutes the north-western Argentine Andes (Figure 8a). After the widespread deposition of early foreland strata and the associated burial of formerly partially exhumed bedrock, deformation appears to have been renewed during the early Oligocene and early Miocene, between 30 and 20 Ma. For example, the Sierra de Rinconada (28 Ma, Ege et al., 2007), Sierra de Cobres (30 Ma, this study), Cumbres de Luracatao (23 Ma, Coutand et al., 2006), and Sierra Chango Real (30 Ma, Coutand et al., 2001) formed a narrow, 600-km-long, and approximately north-south trending belt of deformation. Hence, compared to the previously discussed extent of the Eocene broken foreland, deformation during the early Oligocene and early Miocene appears much more localized (Figure 8b). Importantly, the long sub-meridional belt of localized deformation largely coincides with at least two well-known zones of crustal weakness that partially overlap ( Figure 8b). First, the Ordovician Eastern Puna Eruptive Belt Méndez et al., 1973), a magmatic arc system related to the Famatinian orogeny, which runs from the Sierra de Cochinoca/Escaya Range in the northern Puna southward through the Sierra de Cobres and Sierra de Tanque to the Cumbres de Luracatao and Sierra Chango Real. The second zone corresponds to the western and northern branches of the Salta Rift (e.g., Marquillas et al., 2005;Marquillas & Salfity, 1988). Many Cenozoic structures located within these two zones have been associated with the reactivation of Paleozoic shear zones and planar metamorphic fabrics and inversion of Cretaceous normal faults (e.g., Grier et al., 1991;Hongn et al., 2010;Hongn & Riller, 2007;Kley et al., 2005). Exceptions from the narrow deformation pattern are the Sierra de Calalaste within the present-day Puna Plateau (Carrapa et al., 2005;Kraemer et al., 1999;Zhou et al., 2017), the Tafi del Valle Range in the northern Sierras Pampeanas (Zapata et al., 2020), and the Sierra de Candelaria in the southern Santa Bárbara System (Zapata et al., 2020). However, all these ranges are located south of 26°S, at the latitude of the northern Sierras Pampeanas and are either associated with regions directly bordering the Eastern Puna Eruptive Belt or located within the southern branch of the Salta Rift (Figure 8b).
Between 20 and 15 Ma renewed deformation and exhumation are documented largely west of the aforementioned early Oligo-Miocene deformation belt and affected the Sierra de Tanque (this study) and Sierra de Lina (Henríquez et al., 2020), as well as the Sierra Laguna Blanca of the southern Puna Plateau (Zhou et al., 2017).
Onset of deformation of the intervening Filo de Copalayo Range (≥15 Ma) is inferred from the results of Carrapa et al. (2009), who determined the exhumation of the former sedimentary cover of the range (Geste Formation) using detrital apatite (U-Th)/He (AHe) dating. To the east, thermal modeling suggests rapid exhumation of the Sierra Aguilar between 20 and 16 Ma (Lapiana, 2021). Onset of deformation between 15 and 10 Ma was concentrated in the southern part of the Eastern Cordillera (Andriessen & Reutter, 1994;Carrapa et al., 2011;Deeken et al., 2006;Payrola et al., 2021;Pearson et al., 2012Pearson et al., , 2013Pingel et al., 2019) and has affected the borders of all major intermontane basins in north-western Argentina until the present day, including the Humahuaca and Toro basins, as well as the basins of the greater Calchaquí region. Between 10 and 5 Ma, renewed deformation in the Puna Plateau appears to be absent and tectonic activity shifted eastward to the present-day Eastern Cordillera DeCelles et al., 2011;Lapiana, 2021;Pearson et al., 2013;Pingel et al., 2013Pingel et al., , 2020Rahl et al., 2018;Siks & Horton, 2011) and the Sierras Pampeanas (Löbens et al., 2013;Mortimer et al., 2007;Strecker et al., 1989) and, to some extent, the southern Santa Bárbara System (Arnous et al., 2020;Zapata et al., 2020). Interestingly, the eastern boundary of the Eastern Cordillera deformed synchronously at ca. 10 Ma along its entire length from the Tilcara Ranges (23.5°S, Siks & Horton, 2011;Lapiana, 2021) in the north to the Sierra Mojotorro (25°S, Pearson et al., 2013) and Sierra Metán in the south (25.5°S, Hain et al., 2011). An exception to this pattern appears to be the Sierra Macón in the central Puna, whose final exhumation began around 10 Ma , possibly in response to partial removal of the underlying lithosphere (DeCelles et al., 2015). Finally, the latest initiation of deformation since 5 Ma is mainly observed in the Santa Bárbara System. This is exemplified by the post-5 Ma growth of the Zapla anticline and uplift in the northern Santa Bárbara System at ca. 24°S (Kley & Monaldi, 2002).
Overall, the spatio-temporal pattern of basement deformation in the Pastos Chicos Basin and adjacent regions at 23°-24°S seems incompatible with the notion of a steadily eastward-propagating deformation front as would be expected in a classical orogenic-wedge system (e.g., Dahlen, 1990). Although deformation ultimately migrated eastward during the evolution of this sector of the Andes, propagation of deformation toward the foreland was diachronous (Figures 8 and 9). Importantly, the style of deformation described here records spatially disparate uplifts in a formerly contiguous foreland basin, where maximum horizontal compressive stresses were localized at inherited crustal fault zones and drove them toward failure, resulting in deformation distributed over a broad area extending far into the back-arc of the orogen. This spatio-temporal pattern of range formation and basin fragmentation is akin to the pattern of deformation observed in a broken foreland, analogous to parts of the Colombian and Peruvian Andes (e.g., Mora et al., 2006;Parra et al., 2009;Ramos & Folguera, 2009), the Miocene to Recent basement uplifts in the Sierras Pampeanas and Santa Bárbara System broken-foreland provinces of the Andes (e.g., Allmendinger et al., 1983;Hain et al., 2011;Mortimer et al., 2007), the Cretaceous to Eocene Laramide province of North America (e.g., Jordan & Allmendinger, 1986), the Tian Shan (e.g., Sobel & Dumitru, 1997;, and the Qilian Shan of China (e.g., Tapponnier et al., 1990).
The development of a broken foreland is often related to shallow subduction (e.g., Jordan et al., 1983). For example, in the southern Bolivian Andes at 21°S, middle Eocene to early Oligocene initiation of horizontal crustal shortening and rock uplift in the Bolivian Eastern Cordillera, co-occurred with proposed episodes of shallow subduction (e.g., Anderson et al., 2018). Based on the Eocene deformation pattern observed in the Argentine Andes (Figure 8a), one might be tempted to project the flat slab further south. However, the presence of an Eocene magmatic arc in the Cordillera de Domeyko (e.g., Haschke et al., 2002) precludes flat subduction at that time. Several authors have suggested post-Eocene shallow subduction at different times in the Andean section between 23° and 24°S ( Figure 10). For example, Haschke et al. (2002) observe a magmatic gap between 38 and 26 Ma associated with flat-slab subduction at 20°-26°S. Similarly, Kay et al. (1999) proposed a magmatic gap between 28 and 12 Ma at 22°-24°S. Based on deformation patterns, Horton (2018) proposed flat-slab subduction between 35 and 20 Ma at 15°-25°S. Finally, Ramos and Folguera (2009) proposed a period of flat-slab subduction between 18 and 12 Ma at ca. 20°-24°S. Yet, none of these proposed time periods seem to fit with the observed crustal deformation patterns ( Figure 10). Thus, instead of being attributed to enhanced plate coupling during flat-slab subduction, the observed non-systematic (spatially disparate and diachronous), thick-skinned thrusting may be rather due to heterogeneous deformation of a large orogenic wedge (Armijo et al., 2010) intermittently aided by slip on pre-existing crustal structural discontinuities such as the Paleozoic shear zones and Cretaceous normal faults of the north-west Argentine basement. Figure 10. Probability density plot of apatite fission track (AFT) and apatite (U-Th)/He thermochronology (AHe) cooling ages from 23° to 24°S (Stalder et al., 2020; this study, data shown in Figure 9) illustrating the mismatch between deformation and proposed flat-slab subduction periods: (1) Haschke et al. (2002); (2) Ramos and Folguera (2009); (3) Horton (2018); (4) Kay et al. (1999).

Conclusions
The spatio-temporal information on deformation in the northern Puna and adjacent provinces to the east (Eastern Cordillera and Santa Bárbara System), drawn from new AFT and ZHe thermochronology and inverse modeling combined with published data, allows us to assess the tectonic evolution across the eastern Andean margin since the Eocene. After widespread deformation across almost the entire Puna Plateau during the Eocene, deformation in north-western Argentina was concentrated in a narrow belt of weakened basement (at the position of the Eastern Puna Eruptive Belt) during the Oligocene to early Miocene that more or less follows the present-day eastern margin of the Puna Plateau. Following initial westward migration of tectonism in the early Miocene, deformation shifted more systematically eastward into the present-day Santa Bárbara System, whereas south of 24°S spatio-temporal patterns in deformation show a rather diachronous orogenic development, likely associated with pre-existing crustal heterogeneities.
Zircon U-Pb geochronology of deformed and undeformed ash-bearing strata and field observations are used to specify the tectonic-sedimentary evolution of the Pastos Chicos Basin. Consistent with thermal models, deposition began with the Upper Vizcachera Formation prior to 7 Ma. These units were deformed until about 6.5 Ma, when the regionally extensive Coranzulí Ignimbrite was emplaced (6.6 ± 0.2 Ma). This ignimbrite and subsequent basin fills of the Pastos Chicos Formation (5.6-3.4 Ma) have remained largely undeformed, suggesting that tectonic shortening ceased in this sector of the Andes, consistent with increased deformation in the Eastern Cordillera and Santa Bárbara System.

Data Availability Statement
Data from this study can be found in Supporting Information and on figshare at https://doi.org/10.6084/ m9.figshare.20209865.v1.