A thermochronometric view into an ancient landscape: Tectonic setting, development, and inversion of the Paleozoic eastern Paganzo basin, Argentina

Low-temperature thermochronometric methods like fission-track (FT) and (U-Th)/He dating of apatite and zircon have proven to be powerful tools to investigate rock exhumation in active mountain belts (e.g., Bernet et al., 2004; Blythe et al., 2007; Thiede et al., 2009; Enkelmann et al., 2010). In active tectonic settings, the spatial distribution of thermochronometric ages can be compared with observations of climate, topography, structures, and rock types, and this allows us to recognize correlations between crustal cooling and landscape evolution processes (e.g., Zeitler et al., 2001; Willett et al., 2003; Schuster et al., 2005). The reconstruction of tectonic and surface processes in ancient mountain belts, however, is often much more challenging due to the long-term erosion of topography and related synorogenic strata, and the general modification of upper-crustal rocks over geologic time.

To continue to expand the application of thermochronometric tools to understand the evolution of ancient mountain belts, in this study we present new results from multiple thermochronometric methods (apatite and zircon FT and [U-Th]/He analysis) and zircon U-Pb ages from early Paleozoic bedrock and Upper Carboniferous to Permian detrital samples of the eastern Paganzo basin exposed in the Sierras de Chepes of west-central Argentina. The study area was chosen because it includes an exceptionally well-exposed Carboniferous paleo–glacier valley filled with Upper Carboniferous to Permian strata of the eastern Paganzo basin that were deposited on early Paleozoic basement (Sterren and Martinez, 1996; Socha et al., 2006; Figs. 1 and 2). Our new thermochronometric and geochronometric results are discussed in context with published data from surrounding mountain belts of the Sierras de Pampeanas and allow the reconstruction of the Paleozoic through Cenozoic geological history of the eastern Paganzo basin and nearby basement exposures. Our study demonstrates that in some ancient settings, low-temperature thermochonometric approaches can yield meaningful first-order insights into the development of landscapes over long periods of geologic time.

The Upper Carboniferous to Permian strata of the Paganzo basin crop out over a large part of west-central Argentina (Fig. 1; Fernandez-Seveso and Tankard, 1995). These strata are up to 1500 m thick and represent over 35 m.y. of deposition in marine and continental environments (e.g., Fernandez-Seveso and Tankard, 1995; Limarino et al., 2006). The strata of the Paganzo basin contain a composite record of several different styles of deformation and tectonic settings, including Paleozoic strike-slip deformation, Mesozoic extension and inversion, and late Cenozoic contractional deformation (Fernandez-Seveso and Tankard, 1995; Ramos, 2010). Sedimentation in the eastern Paganzo basin started in the latest early Carboniferous to earliest late Carboniferous with the Serpukhovian–Bashkirian glaciation (Fig. 1; e.g., Limarino et al., 2006; Gulbranson et al., 2010). Glacial and deltaic strata related to this initial depositional stage are preserved in the paleo–glacier valley between the towns of Olta and Malanzan in the Sierra de Chepes (Fig. 1; Fig. DR1¹; Sterren and Martinez, 1996; Socha et al., 2006). We refer to these glacial and glacial-fluvial strata as part of the lower section of the Paganzo Group following the terminology of Azcuy and Morelli (1970). Overlying these Carboniferous strata, fluvial red-bed sandstone and conglomerate characterize the upper section of the Paganzo Group in the study area (Sterren and Martinez, 1996). The red-bed strata occur at the eastern and western flanks of the Sierra de Chepes and in the topographic depression between the eastern and western parts of the paleovalley. These red beds were originally interpreted to be entirely Permian in age, but the start of deposition was recently revised to have been in the late Carboniferous, ca. 306 Ma (Gulbranson et al., 2010). We refer to these strata as the red beds of the upper section of the Paganzo Group in the text (Fig. 3).

The Sierra de Chepes consists of several smaller mountain ranges, including the Sierra de Los Llanos, the Sierra de Malanzan, the Sierra de Los Lujan, and the Sierra del Porongo, which are all located in the west-central part of Argentina (Figs. 1 and 2). The Sierra de Chepes is today part of the basement uplifts of the Sierra de Pampeanas (Fig. 4), which are a product of Miocene to Holocene flat-slab subduction of the Nazca plate and Juan-Fernandez Ridge (e.g., Barazangi and Isacks, 1976, 1979; Ramos et al., 2002). The basement rocks of the Sierra de Pampeanas are a product of series of Paleozoic orogenies that developed through the accretion of different terranes into the proto-Andean margin of Gondwana. Three tectonic phases have been identified: the Early Cambrian Pampean (580–510 Ma), the Late Cambrian–Ordovician Famatinian (500–440 Ma), and the Devonian Achalian orogenic cycles (420–350 Ma) (e.g., Toselli and Aceñolaza, 1978; Aceñolaza and Toselli, 1981; Omarini, 1983; Ramos et al., 1986; Baldo et al.. 1996; Rapela et al., 1998, 2007; Martino, 1999; Siegesmund et al., 2004, 2010; Steenken et al., 2004; López de Luchi et al., 2007; Drobe et al., 2009, 2011). The Sierra de Chepes is mainly composed of granitoid rocks dated at 497–477 Ma, and small (couple of meters thick) discontinuous exposures of metasedimentary rocks (Fig. 3; Pankhurst et al., 1998; Stuart-Smith et al., 1999).

During a reconnaissance trip across the Sierras de Pampeanas, we collected six samples from strata of the eastern Paganzo basin and four samples from the underlying basement rocks in the Sierra de Chepes (Table 1; Figs. 2 and 3). Three of the bedrock samples are part of the paleovalley floor beneath the Carboniferous strata (28TR2, 28TR4, and 29TR3), and one sample represents the basement outside the paleovalley system, located ∼25 km north (28TR6; Fig. 2). Four sedimentary samples are from the lower section of the Paganzo Group that is composed of Upper Carboniferous glacial (28TR7, 29TR4, and 29TR6) and glacial-fluvial (29TR7) strata within the paleovalley (Figs. 2 and 3). Two samples are from the upper section of the Paganzo Group and are composed of red-bed sandstones (28TR1 and 29TR2). The contact between the strata of the lower section of the Paganzo Group and bedrock, as well as the U-shaped valley form that widens toward the west, is observed in the study region. In the Data Repository, we provide field pictures of the Upper and Lower Paganzo Group strata and observations of the contact between bedrock and strata (Fig. DR1 [see ^{footnote 1}]).

The four glacial and glacial-fluvial samples were collected at different sites within the paleo–glacier valley system from the base of the lower Upper Carboniferous strata (samples 29TR4 and 28TR7; Fig. 3). These samples are from the Malanzán Formation (350–320 Ma), which was defined for the Sierra de Los Llanos by Bracaccini (1946) and later included in the lower section of the Paganzo Group (Azcuy, 1975; Andreis et al., 1986), and which is equivalent to the Guandacol Formation of Cuerda (1965) defined for the provinces of La Rioja and San Juan. These formations are equivalent to unit SI-1 of Gulbranson et al. (2010), who constrained the age at 324–318 Ma based on macroflora. Sample 29TR6 is from the top of the Malanzán Formation and was deposited at ca. 318 Ma near the contact with the Tupe Formation; the Malanzán Formation was defined for the La Rioja Province by Frenguelli (1944) and Cuerda (1965) and is equivalent to unit SI-2 of Gulbranson et al. (2010; Fig. 1 herein). Sample 29TR7 is from the Loma Larga Formation (315–305 Ma), defined for the Sierra de Los Llanos by Andreis et al. (1986), which is part of the Tupe Formation (SI-2 unit) that was deposited 318–311 Ma (Gulbranson et al., 2010; Figs. 2 and 3; Table 1). The two red-bed samples (28TR1 and 29TR2; Fig. 3) are from the Solca Formation, defined for the Sierra de Los Llanos by Andreis et al. (1986), and equivalent to the Patquia Formation defined for the provinces of La Rioja and San Juan by Cuerda (1965) (Fig. 3). Their depositional age is less well constrained but is interpreted to be between the late Carboniferous and the Early Permian (<306 Ma; Gulbranson et al., 2010). The red-bed strata in the eastern study area are assumed to be Permian in age (<299 Ma) based on fossilized trees (Crisafulli and Herbst, 2008; Gulbranson, 2011, personal commun.).

All samples were crushed and sieved to grain sizes between 80 and 300 µm. After washing the samples, the apatite and zircon grains were separated using standard magnetic and heavy liquid separation.

Clear apatite and zircon grains without inclusions and other impurities were selected from the bedrock samples using a binocular microscope. After documenting the grain dimensions, single grains were packed in Nb-tubes for (U-Th)/He analysis. In general, we analyzed 3–6 aliquots per sample. The samples were analyzed in the Patterson helium-extraction line at the University of Tübingen, which is equipped with a diode laser to extract the gas. Apatite grains were heated for 5 min at 11 A, and zircon grains were heated for 10 min at 20 A. Each grain was heated and analyzed a second time to make sure that the grain was degassed entirely in the first step. The re-extracts generally showed <1% of the first signal. After helium analysis, the grain packages were sent to the University of Arizona at Tucson for U, Th, and Sm measurements using an inductively coupled plasma–mass spectrometer (ICP-MS).

The analytical errors of mass spectrometer measurements are generally very low (<2%); however, the alpha correction factor and the reproducibility of the sample age represent a much larger uncertainty on the (U-Th)/He age, and thus are a more honest display of the age precision. The scatter between the single-grain ages of the same sample is influenced by many factors, such as zoning of U and Th concentration, the alpha ejection, grain size, radiation damage, micro-inclusions, or any impurities. We report the single-grain ages with their analytical error and the mean (U-Th)/He age and the standard deviation as the sample error (Table 2). Due to the old (U-Th)/He ages, the range between single-grain ages appears large (tens of millions of years), but the sample age reproducibility (7%–18% for zircon and 5%–24% for apatite) is common for the (U-Th)/He method (e.g., Flowers et al., 2008; House et al., 2005; Stock et al., 2006; Hacker et al., 2011; Spotila and Berger, 2010). However, since the method is mostly applied to rocks that experienced Cretaceous–Cenozoic cooling, the absolute errors are usually small (<5 m.y.) and give the impression of a higher precision.

Fission-track analysis on apatite and zircon was carried out in the fission-track laboratory at the University of Tübingen. Apatite grains from both the bedrock and sedimentary samples were embedded in epoxy, grinded, and polished to expose internal surfaces of the apatites. The apatite mounts were etched in 5.5 M HNO₃ for 20 s at 21 °C to reveal the spontaneous tracks in apatite (Carlson et al., 1999). Afterward, the mounts were covered with a uranium-free muscovite external detector and irradiated with thermal neutrons at the research reactor facility at Garching (FRM-II, Germany). After irradiation, the external detectors were etched in 40% HF for 30 min to reveal the induced fission tracks. After etching, the external detector was placed back to its original position for track counting in apatite. Spontaneous and induced fission tracks were counted at a nominal magnification of ×1000 using a Zeiss Axioskope microscope, by focusing first at the apatite surface and then up into the surface of the external detector (Jonckheere et al., 2003). The FT lengths and their angle to the crystallographic c-axis were measured on horizontal confined tracks using the same microscope equipped with a drawing tube and digitizing tablet and the FTstage 4.0 system provided by Trevor Dimitru.

Zircon grains were embedded in Teflon, polished, and etched in a eutectic solution of NaOH and KOH at 227–230 °C to reveal the spontaneous tracks. We prepared several mounts per sample and etched for various times ranging from 6 to 12 h. Between the etching steps, we checked the appearance of the etched tracks under the microscope to make sure that all zircon populations were properly etched. After etching, the zircon mounts were prepared for irradiation in the same way as the apatite mounts. Zircon fission tracks were counted under a nominal magnification of ×1000 (dry objective) using a Zeiss AxioImager M2M microscope equipped with an Autoscan stage system.

We measured the zircon U-Pb age of individual zircon grains using laser-ablation (LA) ICP-MS at the Museum für Mineralogie und Geologie (GeoPlasma Laboratory, Senckenberg Naturhistorische Sammlungen Dresden). We analyzed 120 grains per sample for the two detrital samples (29TR2 and 29TR4) and 30 grains for the bedrock sample (28TR4).

Analytical techniques and settings of the analysis of U-Pb isotopes of magmatic and inherited zircon by LA-ICP-MS are described as follows, and instruments settings are given in Table DR1 (see ^{footnote 1}). Zircons were analyzed for U, Th, and Pb isotopes by LA-ICP-MS techniques using a Thermo-Scientific Element 2 XR sector field ICP-MS coupled to a New Wave UP-193 Excimer laser system. A teardrop-shaped, low-volume laser cell constructed by Ben Jähne (Dresden) and Axel Gerdes (Frankfurt) was used to enable sequential sampling of heterogeneous grains (e.g., growth zones) during time-resolved data acquisition. Each analysis consisted of ∼15 s background acquisition followed by 30 s data acquisition, using laser spot sizes of 25 and 35 µm, respectively. A common-Pb correction, based on the interference- and background-corrected ²⁰⁴Pb signal and a model Pb composition from Stacey and Kramers (1975), was carried out if necessary. The necessity of the correction was judged based on whether the corrected ²⁰⁷Pb/²⁰⁶Pb value was located outside of the internal errors of the measured ratios. Discordant analyses were generally interpreted with care. Raw data were corrected for background signal, common Pb, laser-induced elemental fractionation, instrumental mass discrimination, and time-dependent elemental fractionation of Pb/Th and Pb/U using an Excel® spreadsheet program developed by Axel Gerdes (Institute of Geosciences, Johann Wolfgang Goethe-University Frankfurt, Frankfurt am Main, Germany). Reported uncertainties were propagated by quadratic addition of the external reproducibility obtained from the standard zircon GJ-1 (∼0.6% and 0.5%–1% for the ²⁰⁷Pb/²⁰⁶Pb and ²⁰⁶Pb/²³⁸U, respectively) during individual analytical sessions and the within-run precision of each analysis. Concordia diagrams (2σ error ellipses) and concordia ages (95% confidence level) were produced using Isoplot/Ex 2.49 (Ludwig, 2001). The ²⁰⁷Pb/²⁰⁶Pb age was taken for interpretation of all zircons older than 1.0 Ga, and the ²⁰⁶Pb/²³⁸U ages were used for younger grains. For further details on analytical protocol and data processing, see Frei and Gerdes (2009). The Th/U ratios (Tables DR2 to DR4 [see ^{footnote 1}]) were obtained from the LA-ICP-MS measurements of investigated zircon grains. U and Pb content and Th/U ratio were calculated relative to the GJ-1 zircon standard and are accurate to ∼10%.

This study presents data from four different thermochronometric systems that have been applied to two different kinds of samples, i.e., granitoid “bedrock” and sedimentary “detrital” samples. Each thermochronometric system yields ages of the sample that refer to the cooling below a certain closure temperature range and the thermal history within the temperature window where fission tracks anneal and helium diffusion occurs. In the simple case of continuous rock cooling through the temperature window where helium diffusion and FT annealing occur, the thermochronometric age represents the time of cooling below the closure temperature (Dodson, 1973). The zircon FT and (U-Th)/He system have the highest closure temperatures of 290–210 °C and 200–160 °C, respectively, depending on the cooling rate, grain size, and accumulated radiation damage (e.g., Yamada et al., 1995; Brandon et al., 1998; Reiners et al., 2002, 2004; Garver et al., 2005). The apatite FT and (U-Th)/He systems have lower closure temperatures of 110–100 °C and 80–50 °C, respectively (e.g., Green et al., 1986; Carlson et al., 1999; Schuster et al., 2006).

Because of the thermal sensitivity, subsequent heating due to burial or an increase in geothermal gradient can erase the cooling age of a sample. This property is particularly useful for sedimentary samples, because the thermochronometric results can be used to evaluate the postdepositional heating of sediment/strata in a basin setting. A reset sedimentary sample means that postdepositional burial was sufficient to heat the sample well above the closure temperature, which causes complete resetting of the original cooling signal due to the diffusive loss of helium or annealing of fission tracks. If heating was not high enough, but within the temperatures of partial helium retention or partial FT annealing, the sample is called partially reset and is characterized by apparent ages that are older and younger than the depositional age. An unreset sample is a sedimentary sample that has not been exposed to high temperatures, and the apparent ages, therefore, predate depositional age and provide information about the cooling of the source rock from which the sediment was eroded.

Figure 2 displays the results of the bedrock dating using apatite and zircon (U-Th)/He and apatite FT analysis. Analytical results are given in Tables 2 and 3. The zircon (U-Th)/He ages are oldest, and the mean ages range from 378 ± 39 Ma to 338 ± 43 Ma within the paleo–glacier valley; the sample outside the valley yielded a younger zircon age of 289 ± 51 Ma (28TR6; Table 2). The apatite FT ages range from 259 ± 34 Ma to 225 ± 13 Ma, and the (U-Th)/He ages range from 142 ± 34 Ma to 175 ± 34 Ma for the three samples within the paleo–glacier valley. The bedrock sample located to the north and outside the paleo–glacier valley yielded an apatite (U-Th)/He age of 244 ± 13 Ma, which is similar (within error) to the apatite FT age (232 ± 27 Ma) and suggests rapid cooling in the Triassic (Table 2).

The apatite FT and (U-Th)/He ages of the paleo–glacier valley basement postdate the late Carboniferous and thus indicate that they were heated to temperatures of at least within the apatite FT partial annealing zone after their Carboniferous surface exposure (Figs. 5A and 5B). This heating caused partial or full resetting of the apatite systems, which may be one reason that two samples failed the χ² test (FT ages) and show a larger scatter between single-grain aliquot measurements (apatite [U-Th]/He ages; Tables 2 and 3). Variations in grain size and radiation damage have been suggested as possible reasons for an observed spread between sample aliquots (e.g., Farley, 2002; Shuster et al., 2006). To investigate a possible correlation, we plotted the apatite and zircon (U-Th)/He ages of individual grains versus the effective uranium content (eU) and the diameter of the grain, but we did not find a correlation (Figs. DR2 and DR3 [see ^{footnote 1}]). We measured the diameter of the FT etch pits parallel to the crystallographic c-axis (Dpar) as a kinematic parameter for the apatite chemistry. Apatite chemistry may be responsible for different annealing behavior between grains (Carlson et al., 1999; Donelick and O’Sullivan, 2005). We did not observe variations among the Dpar values of individual grains in the bedrock samples, suggesting that the annealing behavior of the grains is similar.

Confined track length measurements were possible for two apatite samples (28TR2 and 28TR4) and yielded a unimodal length distribution with mean track lengths of 12.5 ± 1.3 µm and 12.6 ± 1.5 µm (Fig. 5). We present the apparent ages and spread in single-grain ages for the three thermochronometric systems in a temperature versus time diagram together with the field observations in Figure 5. Additionally, we analyzed the zircon U-Pb age of sample 28TR4, which yielded a concordia age of 489.3 ± 1.7 Ma (1σ; Fig. DR5 [see ^{footnote 1}]). Sample 28TR6 is located north of the paleo–glacier valley, and the cooling ages differ from those samples collected in the valley. This sample was located at a much deeper crustal level during the early Carboniferous in comparison to those samples in the paleovalley. The mean ages and single-grain ages are presented, revealing the thermal history of the sample, in Figure 5.

For the detrital samples, we analyzed at least 100 grains per sample using the apatite and zircon FT method, and zircon U-Pb dating. The analytical details of single-grain measurements are given in the Data Repository (Tables DR2 to DR6 [see ^{footnote 1}]). The mineral separation process for the red-bed samples of the upper section of the Paganzo Group, however, did not yield any apatites for sample 28TR1 and only 20 apatite grains that could be analyzed for sample 29TR2. For the zircon FT analysis, we counted more than 100 grains for all samples, but one (29TR6), which yielded only 70 grains. Distributions of the grain ages from each sample were analyzed using a grain-age-deconvolution and binomial peak-fitting procedure (Galbraith and Green, 1990; Brandon, 1992, 1996; Galbraith and Laslett, 1993). The peak-fitting results are summarized in Table 4 and displayed in Figure 6. Detailed tables of single-grain ages are provided in the Data Repository (see ^{footnote 1}).

With the exception of one sample (28TR1), the detrital zircon FT age distribution is generally older than the late Carboniferous deposition age, with peak ages ranging from 333 ± 40 Ma to 504 ± 70 Ma (Table 4). The detrital apatite FT ages are mostly younger than the deposition ages and are therefore reset (or partially reset), with mean apatite FT ages ranging from 209 ± 7 Ma to 245 ± 8 Ma (Table 3; Table DR5 [see ^{footnote 1}]). The apatite age distributions show generally two age populations, with the youngest age-population peak ranging between 171 ± 10 Ma and 133 ± 13 Ma and older populations with age peaks between 269 ± 25 Ma and 239 ± 21 Ma (Table 4). Sample 28TR7 yielded three age populations that peak at 142 ± 20 Ma, 215 ± 90 Ma, and 322 ± 200 Ma, thus indicating that some grains are not reset. This sample is therefore interpreted as partially reset (Table 4). The confined track length was measured on each sample, which yielded unimodal length distributions with a mean track length ranging from 11.8 ± 1.6 µm to 13.6 ± 1.3 µm (Table 3; Fig. DR4 [see ^{footnote 1}]). The track length distributions are well constrained, with ∼300 confined tracks measured per sample. We could not find a difference between the track length distribution of grains that yield younger FT ages and the length distribution of grains with older FT ages.

The zircon U-Pb age distributions of one glacial sample (29TR4) from the lower section of the Paganzo Group and one red-bed sample (29TR2) from the upper section of the Paganzo Group were measured and showed different age populations (Fig. 7). All single-grain ages and analytical results are given in the Data Repository (see ^{footnote 1}).

The granitoid basement rocks of the Sierra de Chepes are part of the Famatinian belt and underlie the eastern Paganzo basin. The basement rocks represent a continental magmatic arc and its metasedimentary host rocks that formed during the Ordovician along the proto-Andean margin of Gondwana (Aceñolaza and Toselli, 1981). Recent petrologic and geochronological studies of the basement rocks in the Sierra de Velasco, located north of our study area (Fig. 4), suggested a two-stage evolution for the exhumation of the Famatinian belt (de los Hoyos et al., 2011). Stage I suggests moderate exhumation rates (0.3–0.8 mm/yr) during the Early to Middle Ordovician driven by erosion of overthickened crust, followed by stage II, which is defined by long-lasting slow exhumation (0.01–0.09 mm/yr) from the Late Ordovician to early Carboniferous. This northern part of the Famatinian belt was additionally affected by postorogenic shallow magmatic intrusions in the early Carboniferous (monazite U-Pb: 360 Ma; de los Hoyos et al., 2011), but there is no evidence for magmatic activity during the Carboniferous in the Sierra de Chepes and farther south. The Sierra de Chepes basement predominantly consists of I-type and peraluminous granitoid rocks with Late Cambrian to Ordovician crystallization ages (Fig. 4; zircon U-Pb: 477–497 Ma; Pankhurst et al., 1998; Stuart-Smith et al., 1999; this study). Only small patches of greenschist- to amphibolite-grade metasedimentary rocks occur in the Sierra de Chepes with a suggested maximum depositional age of 510 Ma (Drobe et al., 2011), and an Early to Middle Cambrian age of metamorphism (e.g., Pankhurst et al., 1998; Dahlquist et al., 2005).

The bedrock and detrital samples of our study have been analyzed using four different low-temperature thermochronometric systems to reconstruct their thermal history in the upper crust. According to their low closure temperatures, we not only expect the apatite systems to yield the youngest ages, but also that the Paleozoic thermal record may have been easily erased and reset due to subsequent phases of heating and cooling. This is evident by the Mesozoic apatite cooling ages (Tables 2 and 3), which are discussed later herein (sections on “Late Paleozoic Heating and Triassic Cooling History” and “Jurassic through Cenozoic Cooling History and Basin Inversion”).

The zircon (U-Th)/He ages of the bedrock and detrital samples collected in the Carboniferous paleo–glacier valley provide information about Paleozoic cooling of the upper crust in this area. Additional constraints on the cooling history come from field observations that document a depositional contact between the late Carboniferous glacial strata and the granitoid basement rocks within the Olta-Malanzan paleo–glacier valley (Fig. 2; samples 28TR2, 28TR4, and 29TR3). This relationship implies that these three bedrock samples must have been at surface temperatures before ca. 324–318 Ma, the age of the overlying glacial strata of the Guandacol Formation (SI-1; Gulbranson et al., 2010; Gulbranson, 2011; personal commun.). This is not the case for sample 28TR6, which is located north and outside of the paleo–glacier valley (Fig. 2)—it reveals cooling later in time as discussed later herein (section on “Late Paleozoic Heating and Triassic Cooling History”). The three bedrock samples reveal that the last cooling through ∼180 °C occurred in the Late Devonian through early Carboniferous (ca. 378–338 Ma; Table 2). We interpret this cooling to be related to rock exhumation, rather than lowering of the geothermal gradient, based on the fact that basement rocks were exposed at the surface by the end of the early Carboniferous. With surface exposure by ca. 324–318 Ma, our data suggest cooling from ∼180 °C to the surface in 60–14 m.y. Assuming a geothermal gradient of 25 ± 5 °C/km and a surface temperature of 10 ± 10 °C, this suggests a moderate exhumation rate of 0.1–0.6 mm/yr during the early Carboniferous. Support for such moderate exhumation rates also comes from the detrital zircon FT and U-Pb ages of the Carboniferous strata discussed next. This result suggests that different parts of the Famatinian belt experienced different exhumation histories. While the Sierra de Velasco located to the north reveals exhumation with very slow rates during the early Carboniferous (de los Hoyos, 2011), the Sierra de Chepes were exhumed with rates an order of magnitude faster, possibly due to more surface uplift combined with efficient erosion by fluvial or glacial systems. Rapid rock exhumation is generally expected in regions of high uplift rates, creating local relief and elevation, which would also support the development of alpine glaciers that formed the preserved paleo–glacier valleys.

The combination of FT and U-Pb dating on detrital zircons of the late Paleozoic strata allows both identification of the provenance of the sediment supplied to the eastern Paganzo basin and also information about the cooling/exhumation history of the source terranes that provided sediment to this basin. The spread in the single-grain FT ages within the samples is generally wide (Fig. 6; Table DR6 [see ^{footnote 1}]), which is typical for detrital zircon FT dating due to the varying concentration of U and Th between individual zircon grains. The damage produced by the spontaneous fission of uranium and the alpha decay of uranium and thorium is accumulated at low temperatures and causes the transition from crystalline to amorphous (metamict) zircon. The resulting variation in the crystalline-to-metamict zircon properties causes a wide range of FT preservation between individual grains within a sample (Garver et al., 2005). That means the sample is a mixture containing low-retentive zircons and high-retentive zircons, which accumulate and anneal fission tracks at low or high temperatures, respectively (e.g., Seward and Rhodes, 1986; Carter, 1990; Garver et al., 2002, 2005). This temperature range can be from <200 °C to >350 °C (e.g., Tagami et al., 1998; Rahn et al., 2004; Garver et al., 2005). Despite the large spread in the zircon FT age distributions, we assume that the ages are not reset, consistent with the zircon (U-Th)/He age of the underlying basement, which are also not reset and reveal an early Carboniferous cooling.

The results of the zircon FT age distribution and binomial peak fitting reveal that all four samples of the lower section of the Paganzo Group yield only one age population that peaked in Late Devonian–early Carboniferous times, whereas the younger red-bed sample (29TR2) of the upper section yielded two populations with peaks in Silurian and Cambrian times (Table 4; Fig. 6). The other red-bed sample (28TR1) yielded zircon grains that are partially reset, with age population peaks that are both older and younger than the depositional age (Table 4; Fig. 6). That means the zircon ages of that sample do not provide cooling information from their source area, and this sample is therefore excluded for our discussion here.

The zircon FT age populations found in the sedimentary samples represent the average cooling signal in the catchment within the source region. There is a clear difference between the glacial and glacial-fluvial strata of the Lower Paganzo Group and the overlying red-bed sample of the Upper Paganzo Group. Each of the four samples from the Lower Paganzo Group yields one age population with peak ages that range from 369 to 333 Ma. Taking these age populations and the deposition age, these samples suggest a lag time between 9 and 58 m.y. The lag time is the time the rock needed to cool below the zircon FT closure temperature, to be exhumed to the surface, and then to be eroded, transported, and finally deposited in the basin (e.g., Garver et al., 1999; Bernet and Garver, 2005). Similar to the basement samples, we can use the lag time to estimate the range of exhumation rates for the source area of these late Carboniferous strata. Using the zircon closure temperature of 250 ± 40 °C, an ambient temperature of 10 ± 10 °C, and a geothermal gradient of 25 ± 5 °C/km, the source area for the glacial strata was also exhumed with moderate to rapid rates ranging from 0.1 to 1.6 mm/yr. These exhumation rates overlap with those we estimated for the underlying basement (0.1–0.6 mm/yr) and suggest a proximal source area for the glacial strata. The range of exhumation rates derived from the detrital samples is larger due to the wide range of the zircon FT closure temperatures and the larger uncertainty associated with the zircon FT measurements. A local sediment source is supported by the observation of large granitic clasts and boulders in the diamictite at the base of the glacial strata; these are interpreted to have originated from corestones within local weathering profiles (Socha et al., 2006). Petrographic studies on the Carboniferous strata deposited near Malanzan also support a local source region from the Famatinian belt (Net and Limarino, 2006).

In contrast, much lower exhumation rates are derived for the red-bed sample (29TR2). The two age populations peak at 432 and 504 Ma, suggesting a lag time of at least 133 m.y. and 205 m.y. for each age peak (assuming a maximum deposition age of 299 Ma; Table 1). Using the same parameters for the closure temperature and the geothermal gradient as above, the sediment source area for this upper section of the Paganzo Group was an area with exhumation rates on an order of magnitude lower (0.03–0.1 mm/yr) than the strata of the lower section of the Paganzo Group and underlying basement rocks. Alternatively, the sediment source area may have experienced higher exhumation rates, but the amount of rock exhumation was not enough (>6–14 km) to exhume rocks from below the closure temperature depth. In any case, the zircon FT data suggest that the source area for the red-bed strata was different and located farther away than that for the underlying glacial and glacial-fluvial strata. This is also supported by the fact that the red beds are interpreted as having been deposited in braided river environments (Limarino et al., 2006) and yielded two distinct zircon FT age populations, suggesting that the fluvial catchment was probably large and consisted of source areas with different exhumation histories.

To identify possible source regions, we can compare the zircon FT age populations with published biotite K-Ar cooling ages from the surrounding areas (Fig. 4). Biotite K-Ar ages refer to cooling below ∼350 °C and are commonly used to study metamorphic processes (McDougall and Harrison, 1999, and references therein). Devonian and Carboniferous biotite cooling ages (418–315 Ma; Steenken et al., 2010) have been reported from gneiss, mylonite, and pegmatite of the Sierra de San Luis, which is located south of our study area and is part of the Famatinian belt (Fig. 4; Steenken et al., 2008). Carboniferous apatite U-Pb and biotite Rb/Sr ages have also been reported from meta-granodiorite and mylonite of the Tinogasta-Pituil-Antinaco shear zone (Höckenreiner et al., 2003). This NNW-trending shear zone separates the Pampean and the Famatinian belt and is exposed today north of the Sierra de Chepes. This suggests that sediment in the glacial and glacial-fluvial strata may have originated from the nearby Famatinian belt (Sierra de Chepes) or the Tinogasta-Pituil-Antinaco shear zone. In contrast, Ordovician and Silurian biotite cooling ages (480–400 Ma) have been reported from gneiss and mylonite of the Sierras de Cordoba located east of the study area, which are part of the Pampean orogenic belt (Fig. 4; Steenken et al., 2010). Slow rates of exhumation, comparable to those determined for the red-bed strata (29TR2) in our study area, have been suggested for the northern part of the Famatinian belt in the Sierra de Velasco during Late Ordovician to Carboniferous time (de los Hoyos et al., 2011).

The interpretation of two different source regions for the lower and upper sections of the Paganzo Group of the eastern Paganzo basin based on zircon cooling ages is also supported by the difference in the zircon crystallization ages between sample 29TR4 (glacial) and 29TR2 (red bed). The diagram in Figure 7 displays the probability density plot of the zircon U-Pb ages and shows that the glacial sample has a main population at 440–550 Ma, whereas the red-bed sample is composed of older zircons with main populations at 530–700 Ma and 980–1160 Ma (Fig. 7). The glacial sample (29TR4) reveals three grains out of 120 grains (2.5%) that yielded ages between 330 and 340 Ma, which indicate that early Carboniferous magmatic sources were not dominant. Comparing these zircon U-Pb age distributions with crystallization ages reported from Paleozoic granitoids and metamorphic belts in NW Argentina, we find that the Famatinian orogenic belt and the protolith of the Tinogasta-Pituil-Antinaco shear zone mylonite have ages that are similar to the ages of zircons found in the glacial strata (29TR4; e.g., Pankhurst et al., 1998; Sims et al., 1998; Stuart-Smith et al., 1999; Höckenreiner et al., 2003; González et al., 2004; Steenken et al., 2006). A proximal source area is further supported by the U-Pb ages we determined for the bedrock sample 28TR4, which yielded an intrusion age of 489 Ma (Fig. 2; Data Repository [see ^{footnote 1}]), and other U-Pb ages reported from the Sierra de Chepes granitoids (477–497 Ma) and metasedimentary strata (500–650 Ma; Pankhurst et al., 1998; Drobe et al., 2011). In contrast, the zircon U-Pb ages of the red-bed sample suggest a source area located farther east of the Paganzo basin in the Sierras de Cordoba within the Pampean orogenic belt (Fig. 4; e.g., Rapela et al., 1998; Sims et al., 1998; Schwartz and Gromet, 2004; Schwartz et al., 2008; Siegesmund et al., 2010). In summary, a rapidly exhuming source located near the eastern Paganzo basin supplied sediment to the basin during Upper Carboniferous deposition, whereas a different source, located farther east of the basin, supplied sediment to the basin during the latest Carboniferous and Permian.

As expected, the apatite FT and (U-Th)/He cooling ages are younger than the zircon ages due to their lower closure temperatures. The apatite ages of the basement samples and the overlying strata, however, are also younger than the depositional age of the late Carboniferous to Permian strata (apatite FT: 259–225 Ma and apatite [U-Th]/He: 175–142 Ma; Fig. 2; Tables 1 and 2). This suggests that the basement and the overlying Paleozoic strata were heated after their surface exposure and deposition, respectively, to temperatures <160 °C, the lower temperature limit for the zircon (U-Th)/He closure temperature. This heating caused a partial and/or total loss of the previously accumulated daughter products in apatite (i.e., diffusion of helium and annealing of fission tracks), but not in zircon.

The unimodal length distributions of the five detrital samples (11.8 µm to 13.6 µm; Table 3; Fig. DR3 [see ^{footnote 1}]) generally support the interpretation of full resetting of the apatite FT system; however, the large spread observed in the age distribution suggests that partial resetting of the apatite is also a likely interpretation. A partially reset sample is indicated by a bimodal and/or very broad length distribution with shorter tracks (e.g., Gleadow et al., 1986; Green et al., 1989). The binomial peak fitting found generally two age populations with ages postdating deposition. Only one sample 28TR7 (westernmost sample, Fig. 2) yielded a third population that predates deposition and thus is clearly a partially reset sample. The unimodal length distribution of the underlying bedrock suggests maximum heating during Permian time to temperatures in the lower part of the apatite partial annealing zone, probably between 80 °C and 140 °C, followed by cooling (Figs. 5A and 5B).

Previous studies of the Upper Paleozoic strata in the Paganzo basin suggested sedimentation with a general thinning of the strata toward the east. Based on the interpretation of seismic profiles in the La Rioja Basin located to the north (Fig. 4), Fisher et al. (2002) suggested a maximum thickness of 800 m for the Upper Paleozoic strata. Fernandez-Seveso and Tankard (1995) estimated up to 2000 m thickness for the Upper Carboniferous Guandacol Formation and up to 730 m of Permian strata. In total, the estimated thickness of Upper Paleozoic strata appears insufficient to bury the basement to depths with temperatures more than 80 °C for a normal geothermal field. It is possible that estimates of the late Paleozoic strata thickness are too low due to subsequent erosion. We, however, suggest a combination of basement burial by Upper Paleozoic strata and an increase in the geothermal gradient as the probable reasons for the heating of the basement rocks. The southwestern margin of Gondwana was affected by widespread magmatic activity throughout the late Carboniferous to Permian due to major geodynamic reorganization from convergent plate margin to extensional regimes (e.g., Uliana and Biddle, 1988; Mpodozis and Kay, 1992; Kleiman and Japas, 2009; Ramos, 2010). This magmatic event is represented by the Choiyoi igneous province, which covers an area of ∼500,000 km² of western Argentina and eastern Chile, with a thickness of up to 2000 m (e.g., López-Gamundí, 2006; Strazzere et al., 2006; Kleiman and Japas, 2009; Rocha-Campos et al., 2011). This large magmatic event must have, undoubtedly, influenced the geothermal structure of the Sierra de Chepes region.

Based on this reasoning, we suggest that the detrital apatite FT ages are reset (except sample 28TR7). The older of the two apatite FT age populations found in the reset samples ranges from 269 ± 25 Ma to 239 ± 21 Ma and is similar to the underlying bedrock apatite FT ages, which range from 259 ± 34 Ma to 225 ± 13 Ma. All these ages indicate that cooling occurred in the latest Permian through Triassic after the late Paleozoic burial/heating event (Tables 2 and 3). The Triassic cooling is supported by the bedrock sample north of the paleo–glacier valley (28TR6; Fig. 2), which yielded similar apatite FT and (U-Th)/He ages of 232 ± 27 Ma and 244 ± 13 Ma, respectively, suggesting a phase of rapid cooling in the Triassic (Tables 2 and 3). The zircon (U-Th)/He age of the same sample is 289 ± 51 Ma (Table 2). The entire temperature-time path of this sample suggests rapid cooling in the Permian through Triassic from temperatures >180 °C to below ∼45 °C (Fig. 5C).

The Triassic Ischigualasto Basin (Fig. 4) could be a possible sink for the eroded “missing” strata that may have buried the currently exposed Upper Carboniferous strata. This rift basin formed during regional extension at the beginning of Gondwana breakup and contains up to 6 km of clastic strata (Milana and Alcober, 1994; Stipanicic, 2002). The Ischigualasto Basin formed just west of the Sierra de Chepes and is characterized by a thick section of clastic strata and basalt flows (Gonzalez and Toselli, 1975; Valencio et al., 1975; Milana and Alcober, 1994; Stipanicic, 2002; Zerfass et al., 2004). The Sierra de Chepes itself is considered to have been a highland at this time, which is supported by our temperature-time paths of the basement samples outside and inside the paleo–glacier valley (Fig. 5). Evidence for late Carboniferous–Permian heating also comes from a granitoid sample of the Sierra de Cordoba, located east of our study (Fig. 4). The K-Ar analysis of muscovite and potassium feldspar suggests cooling well below 120 °C at 320 Ma, followed by a slight heating event that did not exceed 120 °C at 265 Ma (Jordan et al., 1989; Richardson et al., 2013).

The cooling paths incorporating all three thermochronometric systems per sample suggest that since Triassic time, the bedrock has been below 60–50 °C and has experienced very slow cooling (0.07–0.02 °C/m.y.) to ambient temperatures (Fig. 5). A long-lasting very slow cooling trend from the Triassic through Cretaceous time is also indicated by the wide spread in the bedrock apatite (U-Th)/He ages (244–142 Ma; Table 2) and the youngest age population found in the detrital apatite FT samples, which cluster between 171 and 133 Ma (Table 4).

A slightly different result is indicated by the sample collected outside the paleo–glacier valley (28TR6). The smaller spread in the apatite (U-Th)/He ages suggests that cooling to temperatures below the partial helium retention zone (∼45 °C) occurred already in the Triassic, followed by steady-state conditions with no significant heating or cooling. Evidence for surface exposure of the basement near the location of sample 28TR6 comes from field observations that show Upper Cretaceous strata overlapping the basement. These strata were deposited in Campanian times, based on dinosaur, crocodile, and turtle biostratigraphy (80–70 Ma; Fiorelli et al., 2011).

As stated already, the Sierra de Chepes is today part of the Sierras de Pampeanas, which consists of uplifted basement blocks with high topography up to 3000 m elevation (e.g., Sierra de Velasco) and 1500 m in the study area. Low-temperature thermochronometers have been used in the past to quantify the amount and timing of Cenozoic deformation within the Sierra de Pampeanas (e.g., Jordan et al., 1989; Coughlin et al., 1998; Richardson et al., 2013). Jordan et al. (1989) dated samples from the Sierras de Famatina, Velasco, and Ancasti, located north of our study area, and the Sierra de Cordoba, located to the east. Apatite FT ages range between 237 and 115 Ma and are interpreted to reflect a long steady state in the upper part of the partial annealing zone before late Cenozoic cooling (Jordan et al., 1989). Apatite FT dating of basement rocks collected from elevation profiles in the western ranges of the Sierra de Pampeanas and west of the Sierra de Chepes (e.g., Sierra de Velle Fertil, Sierra de Famatina, Sierra de Maz; Fig. 4) records a wide range of ages from 300 to 7 Ma (Coughlin et al., 1998). Based on these data, it was suggested that Andean deformation started during the late Paleocene–middle Eocene, which was followed by burial/heating due to sedimentation in the developing Andean foreland basin (Coughlin et al., 1998). The Juan-Fernandez Ridge flat-slab subduction caused Neogene basement-involved deformation of the Sierras de Pampeanas, creating topography of >1500 m elevation in the eastern, and >3000 m elevation in the western and northern ranges. Due to the lack of Neogene cooling ages in the eastern ranges, and only a few Neogene ages exposed in the western ranges, it is generally assumed that basement uplift did not result in rock exhumation exceeding 2–3 km (Jordan et al., 1989; Coughlin et al., 1998). Recent studies of illitization processes in the Neogene Vinchina Basin, located farther northwest of the Sierra de Chepes (Fig. 4), have suggested a very low geothermal gradient of 15–18 °C/km (Collo et al., 2011), which would allow a much larger amount of rock exhumation without exhumation of rocks from below the apatite (U-Th)/He or FT closure temperature. Davila and Carter (2013) revisited published apatite FT data from the western ranges of the Sierra de Pampeanas (located west and north of our study) and suggested substantial rock exhumation across the foreland region due to flat-slab subduction and lowering of the geothermal gradient. In summary, the new thermochronometric results together with those published from the eastern ranges reveal a time of very slow cooling since the Mesozoic.

By employing multiple thermochronometric systems to analyses of a suite of bedrock and sedimentary samples in west-central Argentina, we are able to reconstruct the thermal evolution of the basement rocks of the eastern Paganzo basin, identify two distinct sediment source areas that supplied sediment to the basin, and establish the timing of basin inversion. These data provide a thermochronometric view into an ancient landscape and record its evolution from Paleozoic time to the present. A summary of our observations and interpretations is given in Figure 8.

The zircon cooling ages provide estimates of Paleozoic exhumation rates for both the basement rocks of the Sierra de Chepes and the sediment source areas of the Carboniferous glacial and glacial-fluvial strata in the eastern Paganzo basin. For the Sierra de Chepes and the nearby surrounding terranes that provided sediment to the late Carboniferous basin, we derived long-term (>10⁶ yr) exhumation rates of 0.1–1.6 mm/yr for Late Devonian–Carboniferous times. These exhumation rates, derived from zircon thermochronometric systems, are comparable to those reported from tectonically active mountain ranges like the European Alps (0.2–0.3 mm/yr; Bernet et al., 2009), the Himalayas (0.3–1.4 mm/yr; e.g., Bernet et al., 2006; Blythe et al., 2007; Thiede et al., 2009; Enkelmann et al., 2011), and the Andes (0.2–1.3 mm/yr; e.g., Garver et al., 2005). In contrast, exhumation rates are an order of magnitude lower in ancient, currently not tectonically active mountain belts, such as the Appalachians (0.07–0.08 mm/yr; Roden-Tice and Tice, 2005), Scandinavia (<0.02 mm/yr; e.g., Hendriks et al., 2007, and references therein), or Northern (0.02–0.05 mm/yr; Lisker et al., 2006) and Southern Victoria Land in Antarctica (0.08–0.1 mm/yr; Fitzgerald et al., 2006). All of these active and ancient orogens constitute mountain belts with >1500 m peak elevations, and all of them are glaciated or have been glaciated throughout the Quaternary. However, the average long-term exhumation rates differ by an order of magnitude between the tectonically active and ancient “inactive” mountain belts. Carving deep glacial valleys into a tectonically inactive mountain belt often produces impressive topographic relief and may also result in young cooling ages of the lower-temperature systems such as apatite (U-Th)/He, but it does not result in the rapid exhumation of rocks originating from greater depths below the closure temperatures of zircon FT or the (U-Th)/He system. However, this is the case when efficient fluvial or glacial erosion removes the actively uplifting surface in tectonically active mountain belts. In these types of settings, active mountainous regions show higher relief with peak elevations of >4000 m. We thus could reasonably speculate that this region of the Sierra de Chepes and parts of the orogenic Famatinian belt must have been significant mountain ranges that were tectonically active in the Carboniferous. Based on the results of our study, we suggest that at least parts of the orogenic belts in west-central Argentina were characterized by surface uplift and rapid rock exhumation with amounts that were sufficient to expose rocks from depths below the zircon FT closure temperature. In contrast, the area located farther east of the Paganzo basin (Pampean orogenic belt), which delivered sediment to the red-bed strata of the upper section of the Paganzo Group, was tectonically inactive in the late Carboniferous through Permian and probably was characterized by much lower topography relative to the Famatinian belt. Our findings show that we can reconstruct the first-order late Carboniferous topography of the eastern Paganzo basin and surrounding highlands; our study provides an example of the type of information that can be acquired with thermochronometric data from older landscapes.

The interpreted deposition of at least 2.5 km of strata on top of the currently exposed Upper Carboniferous paleovalley strata also indicates that during the Permian, the area of the Sierra de Chepes was subsiding and part of a sedimentary basin. Inversion of the eastern Paganzo basin, marked by rapid cooling and exhumation, occurred in Triassic–Early Jurassic time with the beginning of Mesozoic rift development associated with opening of the Atlantic Ocean (Uliana and Biddle, 1988; Uliana et al., 1989; Zerfass et al., 2004). Since that time, the strata of the Paganzo basin and related basement rocks have been at near-surface temperatures, and the region has experienced very slow cooling. These findings are consistent with documentation of Mesozoic peneplain surfaces that are preserved in the eastern Sierra Pampeanas (Jordan et al., 1989; Carignano et al., 1999). The Sierras Pampeanas are characterized today by active basement-involved deformation due to flat-slab subduction of the Juan Fernandez Ridge (Richardson et al., 2012). Neogene basement uplift has created topography of >1500 m elevation in the eastern Sierras Pampeanas, but it has not caused significant rock exhumation, which could be detected with low-temperature thermochronometric data (Jordan et al., 1989; Richardson et al., 2013). Thermochronometric studies in the Sierra de San Luis (located south of the study area) also suggest that the rocks were already near the surface in the Late Cretaceous and that the amount of sediments eroded since that time is <450 m (Löbens et al., 2011). In summary, our thermochronometric reconstruction of the geologic development of the Paleozoic Paganzo basin and related basement rocks demonstrates the applicability of multiple thermochronometric techniques to older settings that may have experienced multiple stages of both deformation and provenance over long periods of geologic time.

We thank Roberto Donato Martino, Tonya Richardson, Hersh Gilbert, Mario Gimera Lopez, and Yamila Dubokovic for their help during field work. E. Enkelmann thanks E.L. Gulbranson for helpful discussion about the Paganzo stratigraphy. E. Enkelmann also thanks P. Reiners for U, Th, and Sm measurements. Ridgway’s involvement in this project was supported by a National Science Foundation grant to Hersh Gilbert. The paper benefited from insightful and constructive comments from Martin Danisik and three anonymous reviewers.