Reconstructing the tectono-sedimentary evolution of the Early–Middle Jurassic Tlaxiaco Basin in southern Mexico: New insights into the crustal attenuation history of southern North America during Pangea breakup

Recognizing that the Earth’s history is punctuated by the cyclic assembly and breakup of supercontinent landmasses is one of the most outstanding developments in the past six decades of research in Earth Science (Hawkesworth et al., 2013; Spencer et al., 2017; Wang et al., 2021). Although the concept of the supercontinent cycle is widely accepted (e.g., Nance and Murphy, 2013), the geodynamic processes that lead to the assembly and breakup of supercontinents are still poorly understood, and indeed, their understanding represents one of the most ambitious frontiers in Earth Science. Pangea is the most recent supercontinent. It formed on Earth at the end of Paleozoic time (e.g., Rogers and Santosh, 2003), and it is probably the best case of study to understand the tectonic processes that produced the breakup and separation of supercontinental masses. However, the kinematics and dynamics of Pangea breakup are still not completely understood in some places, such as continental Mexico. During the early Mesozoic Pangea breakup, Paleozoic and Proterozoic terranes of Mexico were located along the North America–South America divergent plate boundary (e.g., Dickinson and Lawton, 2001). Due to such a paleogeographic position, the Jurassic tectonic evolution of Mexico was influenced by the activity of major normal to lateral faults that represent an invaluable record of the continental attenuation related to Pangea breakup (e.g., Martini and Ortega-Gutiérrez, 2018). However, difficulties arise in recognizing the major faults that accommodated Jurassic continental extension in Mexico because of the deformational overprinting of Cretaceous and Cenozoic shortening, transpressional, and transtensional episodes that largely obliterate the evidence of previous tectonic events. A sedimentological approach can help to identify major faults that accommodated continental extension during the breakup of Pangea. In effect, the extension associated with Pangea breakup produced the development of several rift basins, some of which are superbly exposed in Mexico (Goldhammer, 1999; Campos-Madrigal et al., 2013; Martini and Ortega-Gutiérrez, 2018). The stratigraphic record of these basins undoubtedly represents a primary archive of information on the tectonic evolution related to Pangea breakup. This is because the internal architecture of an extensional rift basin and its sediment routing are in large part controlled by the exhumation of basement highs along major faults that are the direct expression of the ongoing tectonic process (Gawthorpe and Leeder, 2000; Allen and Allen, 2013). Based on this assumption, the reconstruction of the internal architecture of Mexican Jurassic rift basins, integrated with the provenance analysis of the different architectural elements, will allow the tracking of major active faults acting as main basin boundaries, and consequently, will permit the identification of major structures that accommodated extension during the development of the North America–South America boundary. In this paper, we present new sedimentologic, petrographic, and U-Pb geochronologic data that allow the reconstruction of the internal architecture and sediment routing of the Jurassic Tlaxiaco Basin, which is one of the largest Jurassic rift basins in southern Mexico related with Pangea breakup. The integration of our data shows that the Tlaxiaco Basin was limited to the north and east by two major basement highs that were bounded by two regional-scale faults with WNW and NNW main trends, respectively. The activity of these faults profoundly controlled the sedimentary architecture and evolution of the Tlaxiaco Basin. Based on our data, we discuss previous reconstructions of Jurassic continental extension proposed for southern Mexico and offer a new perspective for future works that aim to reconstruct the breakup of western equatorial Pangea.

In southern Mexico, the stratigraphic record associated with continental rifting during Pangea breakup is represented by Lower–Middle Jurassic, continental to shallow-marine successions that are discontinuously exposed in the surroundings of Tezoatlán, Tlaxiaco, Tecomatlán, and Olinalá (Erben, 1956; Morán-Zenteno et al., 1993; García-Díaz, 2004; Figs. 1 and 2). Based on stratigraphic similarities, López-Ticha (1985) suggested that these Jurassic successions were deposited in a single, 200-km-wide, extensional rift basin named Tlaxiaco Basin (Fig. 1). However, the available data on these successions are few, and to date, no detailed sedimentologic and provenance study has demonstrated a physical connection between these depositional areas. The following briefly summarizes the available data on the Lower–Middle Jurassic successions in the different areas of the Tlaxiaco Basin.

The Lower–Middle Jurassic stratigraphic record exposed in the Tezoatlán area consists of a volcano-sedimentary succession bounded to the north by the Salado River fault (Fig. 1). The Salado River fault is an Early–Middle Jurassic, WNW-trending, normal sinistral fault, along which the Tezoatlán Jurassic succession is juxtaposed with the Paleozoic, greenschist-facies metamorphic rocks of the Acatlán Complex (Martiny et al., 2012; Fig. 1). The lower part of the Lower–Middle Jurassic succession is composed of mafic and felsitic, volcanic to very low-grade metavolcanic rocks and volcaniclastic fluvial deposits (Diquiyú and Rosario formations; Morán-Zenteno et al., 1993; Zepeda-Martínez et al., 2018; Fig. 2). Zircon grains in felsitic volcanic rocks return U-Pb ages between ca. 197 and ca. 184 Ma, which are interpreted as the age of volcanic activity in the Tezoatlán area (Durán-Aguilar, 2014). U-Pb ages on detrital-zircon grains from fluvial volcanoclastic deposits show that the magmatic activity in the Tezoatlán area lasted until ca. 176 Ma (Zepeda-Martínez et al., 2018). Volcanoclastic fluvial deposits are conformably overlain by the Cualac formation (Fig. 2), which is an informal unit composed of conglomerate, sandstone, and minor mudstone (Erben, 1956) ubiquitously containing fossil leaves (e.g., Guzmán and Velasco-de León, 2014). The Cualac formation is interpreted as a set of alluvial fans draining, to the south and southwest, the Paleozoic, greenschist-facies metasedimentary rocks of the Acatlán Complex exposed directly to the north of the Salado River fault (Zepeda-Martínez et al., 2018). A few authors have tentatively interpreted the Cualac formation as the stratigraphic record of an estuarine environment (De Anda-García, 2008; Vite-del Ángel, 2014). However, beds with a clear marine origin have not been reported within the Cualac formation. The Cualac formation is conformably overlain by the informal Tecocoyunca group (Erben, 1956; Fig. 2). The lower part of the Tecocoyunca group is made up of sandstone, coaly mudstone, and coal, with abundant fossil flora and scarce shallow-marine limestone beds (Erben, 1956; Morán-Zenteno et al., 1993). These rocks have been interpreted as a flood-plain area that was occasionally invaded by shallow-marine waters (Erben, 1956; Morán-Zenteno et al., 1993). Durán-Aguilar (2014) interpreted the lower Tecocoyunca group as a lateral facies change of the Cualac formation. However, sedimentological and petrological evidence indicates that the sandstone and mudstone, which were tentatively interpreted by Durán-Aguilar (2014) as flood-plain deposits of the lower Tecocoyunca group interbedded within the Cualac formation, are rather alluvial deposits of the Cualac formation deposited during low-water stages (Zepeda-Martínez et al., 2018). The upper part of the Tecocoyunca group is composed of marine sandstone, mudstone, and limestone with pelecypods and upper Bajocian (ca. 169 Ma using the Gradstein et al., 2012, time scale) ammonites (Sandoval and Westermann, 1986; Cantú-Chapa, 1998).

Based on lithological, stratigraphical, and paleontological similarities, Lower–Middle Jurassic sedimentary rocks of the Tlaxiaco area have been assigned to the Cualac formation and overlying Tecocoyunca group originally defined in the Tezoatlán area (Erben, 1956; Carrasco-Ramírez et al., 2016; Fig. 2). As in the Tezoatlán area, the Cualac formation in the Tlaxiaco area has been tentatively interpreted as the stratigraphic record of coalesced alluvial fans (Corro-Ortiz and Ruíz-González, 2011), whereas the overlaying Tecocoyunca group has been preliminarily interpreted as the deposit of a meandering river and adjacent flood-plain areas that evolved into a shallow-marine environment during a major transgressive event (Erben, 1956; Carrasco-Ramírez et al., 2016). Almost all previously published works on these units are focused on their paleontological content, which includes a well-preserved Jurassic flora dominated by Bennettitales, Cycadales, and Podozamitales (Lozano-Carmona and Velasco-de León, 2016), as well as dinosaur footprints (Rodríguez-de la Rosa et al., 2018). No detailed sedimentological study has been published on these Jurassic units. The provenance of these clastic rocks is also poorly understood.

Clastic deposits in the Tecomatlán area have been tentatively correlated with the Cualac formation and Tecocoyunca group based on similarities in lithology and the Jurassic floral content (Silva-Pineda, 1969; Hernández-Vulpes and Rodríguez-Calderón, 2012; Fig. 2). However, to date, no sedimentological and provenance study has been carried out on these clastic rocks.

The Jurassic stratigraphic record exposed in the Olinalá area consists of a volcano-sedimentary succession similar to the one described in the Tezoatlán area (Fig. 2). The oldest rocks consist of volcanic deposits (Las Lluvias Ignimbrite; Corona-Esquivel, 1981) with zircon grains that yield U-Pb ages between ca. 179 and ca. 177 Ma (Campa-Uranga et al., 2004). Volcanic rocks are overlain by fluvial and shallow-marine clastic deposits that, based on lithological, stratigraphic, and paleontological similarities, have been assigned to the Cualac formation and the Tecocoyunca group (Marshall, 1986; García-Díaz, 2004). A detailed sedimentological and provenance analysis of these clastic deposits in the Olinalá area is presently not available. Differing from the Tezoatlán area in which available paleontological data bracket the beginning of marine transgression to upper Bajocian time (ca. 169 Ma using the Gradstein et al., 2012, time scale; Sandoval and Westermann, 1986), the age of ammonites from the marine upper part of the Tecocoyunca group in the Olinalá area vary from Bathonian to Callovian (ca. 168 to ca. 166 Ma using the Gradstein et al., 2012, time scale; Westermann et al., 1984).

The integration of our field observations with previously published data allowed the construction of four geologic maps of the Lower–Middle Jurassic successions exposed in the Tezoatlán, Tlaxiaco, Tecomatlán, and Olinalá areas (Figs. 3A, 3B, 4A, and 4B). We measured representative stratigraphic columns (Figs. 5A–5I), determined lithofacies assemblages, and identified architectural elements and main stratigraphic surfaces according to the hierarchy scheme of Miall (2006). This scheme upgrades the original scheme of DeCelles et al. (1991), which was specifically designed for alluvial fan deposits. A summary of the surface hierarchy observed in the study area is given in Table 1. A synthesis of the main lithofacies and architectural elements is presented in Tables 2 and 3. Additionally, we measured paleocurrent directions to reconstruct the fluvial drainage pattern and improve the provenance analysis. Paleocurrent data were corrected for bedding dip according to Collinson et al. (2006) and for clockwise tectonic rotation according to the paleomagnetic data of Böhnel (1999).

Along the measured stratigraphic columns, we selected 58 samples of medium- to coarse-grained sandstone for whole-rock compositional analysis (Figs. 5A–5I). We counted between 371 and 493 points for each sample by using the Gazzi-Dickinson method (Gazzi, 1966; Dickinson, 1970). Whole-rock petrographic thin sections were stained for easy recognition of potassium feldspar. Following Garzanti and Vezzoli (2003), metamorphic lithic grains were classified based on their protolith composition and metamorphic rank. Grain parameters (Zuffa, 1985; Garzanti and Vezzoli, 2003) are defined in Table 4, and sandstone modal compositions are presented in File A in the Supplemental Material.¹

We integrated the sandstone compositional analysis with detrital-zircon U-Pb geochronology to reinforce our provenance analysis. Zircons were dated by U-Pb laser ablation–inductively coupled plasma mass spectrometry (LA-ICPMS) at Laboratorio de Estudios Isotópicos, Centro de Geociencias, Universidad Nacional Autónoma de México, following the methodology and data treatment reported by Solari et al. (2018). Individual zircon ages were obtained with a Resolution M-50 Excimer laser, operating at a 193 nm wavelength and coupled to a Thermo ICap Qc quadrupole ICPMS. We selected the analyzed zircons randomly by handpicking to avoid a possible age bias toward certain groups (e.g., shape, color, and dimension). Standard zircon 91500 (ca. 1065 Ma, Wiedenbeck et al., 1995) was used as the primary standard, whereas Plešovice zircon (ca. 337 Ma, Sláma et al., 2008) was employed as the control standard, yielding, in the course of the present analytical sessions, a mean ²⁰⁶Pb/²³⁸U age of 339.8 ± 1.3 Ma (MSWD 2.4, n = 47) in agreement with published values, especially for non-annealed standard zircons (e.g., Solari et al., 2015). We consider the ²⁰⁷Pb/²⁰⁶Pb age for zircon grains older than 1.5 Ga and the ²⁰⁶Pb/²³⁸U age for younger zircons (Spencer et al., 2016). Raw data were reduced by employing Iolite software v. 3.6 (Paton et al., 2011) and the VizualAge data reduction scheme of Petrus and Kamber (2012). Errors are quoted at the 2σ level. A representation of the statistical distribution for the visualization of data was made by employing a Kernel density estimator (KDE; e.g., Vermeesch, 2018), which is a more robust alternative to the probability density plot. This is because a KDE does not explicitly take into account the analytical uncertainties. The bandwidth used for the construction of a statistical distribution of data varies according to the local density (adaptive Kernel density estimation). Sample locations are given in Figures 3–5. The maximum depositional age (MDA) of samples is constrained by the weighted mean of the youngest cluster defined by at least three zircon grains overlapping in age at 2σ (Dickinson and Gehrels, 2009). If a cluster is not defined, the MDA is determined on the basis of the youngest concordant zircon grain, according to Spencer et al. (2016) and Copeland (2020). We use multidimensional scaling (MDS) maps based on the Kolmogorov-Smirnov statistics (Vermeesch, 2013) to compare the detrital-zircon U-Pb age signature of the analyzed samples from the Cualac formation and lower Tecocoyunca group with the available data from Paleozoic and Proterozoic complexes, which represent potential source terranes. Clusters in MDS maps represent samples with similar age spectra, whereas samples with different age spectra plot far apart. To help identify the first- and second-order similarities, each sample is connected to its most similar sample by a solid line and its second most similar sample by a dashed line.

The base of the Cualac formation is a seventh-order surface. Along this surface, the Cualac formation unconformably overlies Paleozoic metamorphic rocks of the Acatlán Complex and conformably overlies volcanic and volcaniclastic fluvial deposits of the Las Lluvias Ignimbrite and the Rosario formation (Fig. 2). The Cualac formation is composed of pebble to cobble conglomerate, conglomeratic to coarse-grained sandstone, minor medium- to fine-grained sandstone, and mudstone that are arranged in a fining- and thinning-upward succession (Figs. 5A–5C and 5F–5H). The thickness of the Cualac formation is on the order of ~100 m in the eastern Tlaxiaco Basin (Tezoatlán and Tlaxiaco areas). In its western part (Olinalá area), we estimate a thickness of ~1300 m by using trigonometry. In the Cualac formation, we recognized ten lithofacies that are arranged into two different architectural elements. For a detailed description of lithofacies, the reader is referred to Table 2.

Most of the Cualac formation consists of decimeter- to meter-thick, lens-shaped conglomerate deposits that are bounded at the base by third-order erosional surfaces and cut into each other both laterally and vertically (Figs. 6A and 6B). Conglomerate deposits are clast-supported, poorly sorted, and typically display trough and planar cross-bedding (lithofacies Gt and Gp), imbricate clasts, and locally, crude horizontal bedding (Gh; Figs. 5F–5H). Conglomeratic facies locally grade upward into cross-bedded, conglomeratic to coarse-grained sandstone (St and Sp) and eventually, medium- to fine-grained sandstone with ripple cross-lamination (lithofacies Sr) draped by horizontally laminated siltstone and mudstone (lithofacies Fl). According to Miall (2006), these deposits correspond to the Gravel Bars and Bedforms (GB) architectural element, which represents in-channel longitudinal and transverse gravel bars. The base of the GB element is a fourth-order surface, locally showing conspicuous basal scour (Figs. 6A and 6B). The top of element GB is represented by a fourth-order surface draped by a few decimeters to meters of horizontally to ripple cross-laminated siltstone and mudstone (Fl and Sr) with ubiquitous fossil leaves. Paleocurrent directions measured on cross-bedded deposits (Gt, Gp, St, and Sp) are dominantly directed to the SW and SE (Figs. 3–5), with a few data pointing to the W and E directions.

Deposits of the GB element are locally interbedded with decimeter- to meter-thick, matrix- to clast-supported, poorly sorted, massive to inverse grading conglomeratic deposits (Gmm, Gcm, and Gci; Figs. 5B, 5C, 5E–5H, and 6A). Often, these deposits contain tabular clasts aligned parallel to bedding and display a sharp non-erosional base. According to Miall (2006), these characteristics are typical of the architectural element Sediment-Gravity-Flow Deposits (SG), which represents deposits emplaced by high- to low-strength debris flows.

The lower Tecocoyunca group is composed of fine to conglomeratic sandstone, mudstone, and minor granule conglomerate that conformably overlie the Cualac formation in transitional contact (Figs. 5A, 5E, and 5H). The precise age of this transitional contact in the diverse studied areas is presently unknown. Centimeter- to meter-scale, normal syn-sedimentary faults and associated growth strata can be observed at different outcrops. A systematic kinematic analysis of these faults is presently not available. Bed-by-bed measurement and sampling of the lower Tecocoyunca group are possible only at a few localities because of deformation related to post-sedimentary faulting. Based on the lihofacies association and three-dimensional architecture of deposits, we recognized four main elements in the lower Tecocoyunca group.

The lower Tecocoyunca group is mostly composed of alternating and horizontally laminated coaly mudstone, coal, and siltstone (Fl and Fsm) and fine-grained sandstone with ripple cross-lamination (Sr), which form meter- to decameter-thick, sheet-like deposits that are laterally continuous for several hundreds of meters (Fig. 6C). Trunk molds, fossil leaves, and rootlet traces (Fig. 6D) are ubiquitous in these rocks. Dinosaur footprints are locally present. According to Miall (2006), these deposits correspond to the Floodplain Fines (FF) element, which represents deposits of overbank sheet flow, floodplain ponds, and swamps.

Element FF is rhythmically interbedded with ~1–3-m-thick, coarse- to fine-grained, cross-bedded to ripple cross-laminated sandstone deposits (St, Sp, and Sr), which are arranged in a 10° to 15° downstream-dipping clinoform set showing a fining-upward succession. These sandstone deposits display a fourth-order sharp erosional basal surface and toward the top, transitionally grade into the overlying FF element. The sandstone ubiquitously contains mudstone intraclasts and displays convolute lamination, load casts, flame structures, pseudo-nodules, and abundant bioturbation. According to Miall (2006), we interpret these deposits as the architectural element Crevasse-Splay Deposits (CS), which represents fan-shaped sandy deposits introduced into floodplain areas during major crevassing events.

In the Olinalá and Tlaxiaco areas, element FF is interbedded with 9–11-m-thick, conglomeratic to sandstone bodies composed of a set of decimeter- to meter-thick clinoforms bounded by third- to fourth-order accretionary surfaces with a dip of ~20° (Fig. 7A). The base of the clinoform set is a fifth-order, flat erosional surface, whereas its top is transitional to the overlying FF element. In their lower part, clinoforms are composed of clast-supported and well-sorted granule to pebble conglomerates showing trough and planar cross bedding (Gt and Gp). Rip-up mud clasts are ubiquitous in these conglomerates. Conglomeratic deposits grade upward into coarse-grained sandstone beds with trough cross bedding (St), which in turn progressively grade into medium-grained sandstones with planar cross bedding (Sp) and fine-grained sandstone to siltstone with ripple cross lamination (Sr; Fig. 5D). Paleocurrent directions obtained from deposits showing cross bedding (St and Sp) point toward the NE, NW, SE, and SW quadrants and are suborthogonal to the dip of clinoforms. According to Miall (2006), we interpret these deposits as the element Lateral-Accretion Deposits (LA), which is the typical point bar described for meandering rivers (e.g., Allen, 1970).

In the Olinalá and Tlaxiaco areas, a few conglomeratic to sandstone bodies interbedded with element FF show a lithofacies assemblage and geometry that are similar to those of element LA. However, in these few conglomeratic to sandstone bodies, paleocurrent directions form an angle of 0°–60° with the dip of clinoforms (Fig. 7B). According to Miall (2006), this three-dimensional architecture is typical of the element Downstream-Accretion Macroforms (DA), which represents in-channel, downstream-accretion sandy bars.

We collected 19 medium- to coarse-grained sandstone samples representative of the Cualac formation. Results from these 19 samples were integrated with detrital modes previously reported by Zepeda-Martínez et al. (2018) for the Tezoatlán area. Selected samples are moderately to poorly sorted sandstones, with angular to subangular grains surrounded by a very thin film of clay and opaque minerals. The analyzed samples vary in composition from quartzose to quartzo-lithic sandstone (Fig. 8A). In order of decreasing abundance, they are composed of monocrystalline quartz (97.8%–31.5% of the total framework grains), metamorphic lithic grains (68.3%–1.9%; Fig. 8B), polycrystalline quartz (0.7%–0%), and heavy minerals such as tourmaline, rutile, epidote, zircon, and prehnite (0.7%–0%). Solid-state deformation and recrystallization structures in quartz such as undulatory extinction, subgrain domains, and shape-preferred orientation suggest a low-grade metamorphic origin for this component. Metamorphic lithic grains are rank 2–4 metapelitic and metapsammitic/metafelsitic fragments (Figs. 8C and 8D) that, according to Garzanti and Vezzoli (2003), correspond to the subgreenschist and greenschist metamorphic facies. Metapelitic grains are composed of micaschist and display a penetrative continuous foliation at the submillimeter scale defined by tiny flakes to well-developed crystals of aligned white mica (Fig. 9A). Metapsammitic/metafelsitic grains vary from quartz-rich metasandstone to quartz-micaschist with a main foliation that is moderately to highly penetrative at the submillimeter scale and is expressed by the alignment of elongated quartz, flakes of clay minerals in grains of rank 2 and white mica in grains of ranks 3 and 4 (Figs. 10B–10D).

Analyzed samples from the Lower Tecocoyunca group are moderately to very well sorted sandstones, with subrounded to well-rounded grains in a matrix made up of clay and opaque minerals. The analyzed samples are quartzose and litho-quartzose sandstones (Fig. 8A) and are composed of monocrystalline quartz (96.8%–63.7%); metamorphic lithic grains (24.1%–0%); volcanic lithic grains (17.5%–0.2%; Fig. 8B); plagioclase (4.3%–0%); K-feldspar showing mesoperthitic texture (2.5%–0%; Fig. 9E); heavy minerals such as zircon, rutile, tourmaline, epidote, apatite, and orthopyroxene (1.8%–0%; Fig. 9F); and polycrystalline quartz (0.8%–0%). Quartz in the lower Tecocoyunca group is mostly included in polycrystalline phaneritic grains that display a polygonal granoblastic texture with triple junctions (Fig. 9G), which is typical of high-grade metamorphic rocks (Garzanti and Vezzoli, 2003; Passchier and Trouw, 2005). Quartz in these high-grade phaneritic fragments ubiquitously contains several needle-shaped rutile inclusions (Fig. 9H). Metamorphic lithic grains are metapsammitic/metafelsitic and metapelitic fragments of ranks 2 and 3 of Garzanti and Vezzoli (2003; Fig. 8D), corresponding to subgreenschist- and greenschist-facies conditions. Metapsammitic/metafelsitic grains are composed of quartz-sericite and quartz-white mica schists with a moderately to highly penetrative foliation at the submillimeter scale. Quartz in these metapsammitic/metafelsitic metamorphic lithic grains does not contain rutile inclusions. Metapelitic grains are mostly composed of sericite and white mica schist with a penetrative continuous foliation at the submillimeter scale. Volcanic lithic grains are felsitic and are characterized by a porphyritic texture, with quartz, plagioclase, and minor K-feldspar phenocrysts in a quartzo-feldspathic microcrystalline groundmass (Fig. 9I).

We selected five samples from the Cualac formation exposed in the Tlaxiaco (MI-0318-3 and ÑU-0318-1; Figs. 5B and 5C), Tecomatlán (Tmt-0219-16; Fig. 4A), and Olinalá (T9 and CU-05b; Fig. 5H) areas for detrital-zircon U-Pb geochronology. Results from these five samples are integrated with the detrital-zircon U-Pb data (samples 13C and 23C; Fig. 5A) previously reported by Zepeda-Martínez et al. (2018) for the Tezoatlán area. In the Tecomatlán area, 0.5 km to the northwest of Peña Colorada (Fig. 4A), some conglomeratic deposits were tentatively correlated with the Cualac formation by previous works (Hernández-Vulpes and Rodríguez-Calderón, 2012). However, conglomeratic deposits at Peña Colorada have volcanic clasts that are absent in the Cualac formation. Therefore, we collected a sample from the matrix of conglomeratic deposits at Peña Colorada (Tmt-0219-19, Fig. 4A) to verify its possible correlation with the Cualac formation. We performed between 87 and 97 point-ablation analyses for each sample and obtained concordant to slightly discordant ages (percentages of discordia vary from −4.9–11.7; Supplemental File B [^{footnote 1}]). Zircons from samples 13C, 23C, MI-0318-3, ÑU-0318-1, Tmt-0219-16, T9, and CU-05b return similar age distributions, characterized by two main age groups of ca. 1380–870 and ca. 790–330 Ma (Figs. 10A–10G). A minor amount of zircon grains return ages in the ranges of ca. 1640–1430 and ca. 2770–1770 Ma. The MDA for samples 13C, 23C, MI-0318-3, ÑU-0318-1, Tmt-0219-16, T-9, and CU-05b are ca. 183.9, ca. 252.0, ca. 332.5, ca. 267.6, ca. 407.5, ca. 478.6, and ca. 463.0 Ma, respectively (Figs. 10A–10G). Sample Tmt-0219-19 from the Peña Colorada conglomerate displays three main age groups of ca. 1400–880, ca. 500–450 Ma, and ca. 60–50 Ma and a MDA of ca. 49.6 Ma (Fig. 10K). The MDS map shows that the Cualac formation samples have first-order distances with greenschist-facies metasedimentary samples from the Acatlán Complex, whereas they do not share any similarity with granulite-facies rocks of the Oaxacan Complex (Galaz et al., 2013; Zepeda-Martínez et al., 2018; Martini et al., 2020; Fig. 11).

We selected six samples from the lower Tecocoyunca group exposed in the Tezoatlán (Tec-0916–3; Fig. 5A), Tlaxiaco (TB-0817-2, TLA-013; Figs. 5B and 3B), Tecomatlán (Tmt-0219-13; Fig. 4A), and Olinalá (OL-0618-T3, OL-1018-1; Figs. 4B and 5I) areas for detrital-zircon U-Pb geochronology. We performed between 83 and 97 point-ablation analyses for each sample. The analyzed grains return ages that vary from concordant to slightly discordant (percentages of discordia vary between −4.8%–11.9%; Supplemental File B). Zircon grains from the analyzed samples yield similar age distributions that are characterized by two main age groups of ca. 1400–880 and ca. 300–240 Ma (Figs. 10L–10Q). Subordinate grains return ages in the ranges of ca. 1700–1470, ca. 670–350, and ca. 191–173 Ma. The MDA for samples Tec-0916-3, TB-0817-2, TLA-013, Tmt-0219-13, OL-0618-T3, and OL-1018-1 are ca. 179.3, ca. 260.0, ca. 176.5, ca. 189.7, ca. 257.5, and ca. 173.2 Ma, respectively (Figs. 10L–10Q). The MDS map shows first-order distances between the analyzed lower Tecocoyunca group sandstone samples and the granulite-facies metamorphic rock of the Oaxacan Complex (Solari et al., 2014; Fig. 11).

In the Tecomatlán area, the contact relationship between the Jurassic succession and Paleozoic metamorphic rocks of the Acatlán Complex is represented by a major fault zone that we describe here for the first time and name the Axutla fault. This structure can be recognized in the field and in satellite images. It has a trend of N290–315° and a dip of 90°–70° to the southwest (Fig. 12). The distribution of the exposures of this structure allows us to follow the Axutla fault trace for ~25 km from Tecomatlán to Axutla (Fig. 12). Generally, horizontal to moderately dipping Jurassic strata become progressively more inclined as they get closer to the fault (Fig. 4A). Along the fault trace, Jurassic strata are subvertical and trend parallel to the main structure. The core of the fault zone is represented by a ~30–100-m-thick breccia composed of blocks of metamorphic rocks from the Acatlán Complex up to some meters in size. Between Tecomatlán and Axutla, two rhyolitic domes are aligned along the trace of the Axutla fault (Fig. 12). Establishing the kinematics of the Axutla fault is difficult because two, and in some cases, even three, superposed kinematic indicators can be observed at each outcrop of this structure, suggesting a complex history of reactivation at different times. Some kinematic indicators suggest normal to normal sinistral movements, although structures indicating a dextral displacement are also observed. Therefore, the kinematics of the Axutla fault remains unresolved. Directly to the east of Tecomatlán, the Axutla fault is cut by the N-S–trending, dextral Tetla fault (Ortega-Gutiérrez et al., 2018; Fig. 12).

The Cualac formation is dominated by the architectural element GB, which is locally interbedded with element SG. These elements are typical of high-energy and steep-gradient fluvial environments characterized by unstable channels such as braided rivers and alluvial fans (Miall, 2006). The scattered exposure of the Cualac formation does not permit an exhaustive reconstruction of the paleocurrent pattern at the regional scale. However, based on the ~230° dispersion angle of the available paleocurrent data in the Tezoatlán and Olinalá areas and the ~150° dispersion angle in the Tlaxiaco area, we tentatively interpret the Cualac formation as the stratigraphic record of a coalescing alluvial fan system, which was originally proposed by Morán-Zenteno et al. (1993). We do not exclude the possibility that at least the distal part of some of these alluvial fans transitionally changed into a transverse braided river.

The lower Tecocoyunca group is dominated by the architectural elements FF and CS, which are interbedded with the elements LA and DA. The element LA represents the classical point bar of Allen (1970), which is typical of meandering rivers (Miall, 2006). The element DA represents mid-channel bars formed by downstream accretion in sand-bed braided rivers (Miall, 2006). The coexistence of LA and DA elements is typical of meandering river belts that locally display minor mid-channel bars (the wandering fluvial style of Church, 1983, and Miall, 2006). The elements FF and CS are typical of overbank areas characterized by flood plain and crevasse events, with coal layers representing raised peat swamps that underwent rapid plant accumulation under humid climatic conditions (e.g., Makaske, 2001; Miall, 2006). According to these considerations, the lower Tecocoyunca group represents the stratigraphic record of a meandering river, which we call the Tlaxiaco River, and its adjacent overbank areas.

Our data indicate that the Cualac formation was mostly sourced by low-grade metapelitic and metapsamitic/metafelsitic rocks. Paleocurrent directions indicate that this low-grade source was located to the NE and NW of the study areas. Paleozoic greenschist- to subgreenschist-facies metapelites and metapsammites of the Acatlán Complex are largely exposed directly to the north of Tezoatlán and Tecomatlán (Ortega-Gutiérrez et al., 2018; Fig. 1). These metasedimentary rocks contain zircons with U-Pb ages defining two main groups of ca. 1370–870 and ca. 800–340 Ma (Galaz et al., 2013; Zepeda-Martínez et al., 2018; Martini et al., 2020; Figs. 10H–10J), which match well with the two U-Pb zircon groups of ca. 1351–868 and ca. 794–336 Ma obtained for the Cualac formation (Figs. 10A–10G). The MDS map shown in Figure 11 supports this U-Pb zircon age similarity, showing that samples from the Cualac formation plot close to samples from the Acatlán Complex. Therefore, we suggest that the metasedimentary rocks of the Acatlán Complex exposed to the north of the study areas are the source of the Cualac formation. Conglomeratic deposits of Peña Colorada (Fig. 4A), which were tentatively correlated with the Cualac formation, display three main age groups of ca. 1400–880, ca. 500–450 Ma, and ca. 60–50 Ma (Fig. 10K). The MDA of these deposits is ca. 49.6 Ma. Therefore, we exclude any possible correlation of the Peña Colorada conglomerate with the Cualac formation.

Samples from the lower Tecocoyunca group ubiquitously contain quartz with rutile inclusions and a polygonal granoblastic texture with triple junctions, as well as minor mesoperthitic K-feldspar, suggesting their derivation from a high-temperature (>700 °C) metamorphic source (Passchier and Trouw, 2005; Cherniak et al., 2007; Winter, 2014). In southern Mexico, rutilated quartz and mesoperthitic K-feldspar have been reported exclusively in Proterozoic granulites of the Oaxacan Complex (e.g., Ortega-Gutiérrez et al., 2018; Fig. 1). The occurrence of detrital orthopyroxene in samples from the lower Tecocoyunca group supports such a provenance interpretation, as this mineral phase is ubiquitous in granulites from the Oaxacan Complex (Ortega-Gutiérrez et al., 2018). Derivation from the Oaxacan Complex is also supported by detrital-zircon grains from the lower Tecocoyunca group, which yield U-Pb ages defining a main group of ca. 1395–900 Ma (Figs. 10L–10Q), matching well with the ca. 1400–880 Ma U-Pb zircon age range from the Oaxacan Complex (Solari et al., 2014; Fig. 10R). This U-Pb zircon age similarity is also shown in the MDS map (Fig. 11), where samples from the lower Tecocoyunca group show close distances with samples from the Oaxacan Complex. A second group of zircons with ages between ca. 290 and ca. 250 Ma are likely derived from Carboniferous–Permian (ca. 290–255 Ma) plutonic bodies that cut Proterozoic rocks of the Oaxacan Complex (Ortega-Obregón et al., 2014; Figs. 1 and 10S). Minor felsitic volcanic and greenschist- to subgreenschist-facies metasedimentary lithic grains are present in samples from the lower Tecocoyunca group. As suggested by the occurrence of few zircons with U-Pb ages between ca. 191 and ca. 173 Ma (Figs. 10L, 10N, 10O, and 10Q), volcanic lithic grains were likely derived from the Jurassic synrift magmatism that accompanied the lithospheric attenuation of Pangea (Martini and Ortega-Gutiérrez, 2018). Low-grade metasedimentary lithic grains are tentatively derived from the metasedimentary rocks of the Acatlán Complex. This interpretation is supported by the occurrence of a few tourmaline grains, as this mineral phase is absent in the Oaxacan Complex and ubiquitous in metamorphic rocks of the Acatlán Complex (Ortega-Gutiérrez et al., 2018).

Based on our sedimentological data and provenance analysis, we interpret the Cualac formation as a set of alluvial fans that drained fans to the SE and SW from the adjacent metasedimentary rocks of the Acatlán Complex exposed to the north (Fig. 13A). It is well known that alluvial fans necessarily require high-relief source areas to form; therefore, they occur in tectonically active regions (Dade and Verdeyen, 2007; Meek et al., 2020) and form major accumulations along major active faults bounding sedimentary basins (e.g., Gawthorpe and Leeder, 2000; Miall 2006). In the study area, alluvial fan deposits of the Cualac formation are exposed directly to the south of the Axutla fault that we document in this work and the Salado River fault that was previously documented by Martiny et al. (2012). No exposure of the Cualac formation has been documented north of these faults. Moreover, alluvial fan deposits are distributed along a main WNW-trending line that coincides with the trend of these two major fault segments (Fig. 14). All these aspects suggest that the Axutla and Salado River faults likely represent the northern boundary of the Tlaxiaco Basin (Figs. 1 and 13A). At present, the Axutla and Salado River faults are two different segments separated by the N-trending Tetla fault (Fig. 1). However, the development of alluvial fan deposits with the same sedimentological characteristics, provenance, and age directly to the south of the Salado River and Axutla faults suggests that these structures are two segments of a single regional-scale fault system that is presently offset by the Tetla fault (Fig. 1). The sum of the Salado River and Axutla segments indicates a minimum length of ~80 km for this structure (Fig. 1), which we name the Salado River–Axulta fault. The Cualac formation represents the sedimentary response to the Salado River–Axutla fault activity; therefore, the age of the Cualac formation roughly constrains the activity of the Salado River–Axutla fault between ca. 176 Ma, the age of the youngest zircon grains in volcanoclastic deposits underlying the Cualac formation (Zepeda-Martínez et al., 2018), and the Bajocian–Bathonian age (ca. 170–168 Ma, Gradstein et al. 2012) of marine deposits that transgressed fluvial successions of the Tlaxiaco Basin (Sandoval and Westermann, 1986).

The stratigraphic transitional relationship of the Cualac formation with the overlying lower Tecocoyunca group indicates that the alluvial fans that developed along the northern boundary of the Tlaxiaco Basin were abandoned, probably because the Salado River–Axutla fault became inactive, and were progressively buried by overbank deposits of the Tlaxiaco River between ca. 173 Ma, the age of the youngest zircon grains in the lower Tecocoyunca group (this work), and the Bajocian–Bathonian age (ca. 170–168 Ma, Gradstein et al., 2012) of the overlying transgressive marine deposits (Westermann et al., 1984; Marshall, 1986; Sandoval and Westermann, 1986; Figs. 13B and 13C). Available sedimentologic and provenance data suggest that the Tlaxiaco River was a major meandering fluvial system that drained high-grade metamorphic rocks of the Oaxacan Complex to the west into the Tlaxiaco Basin (Fig. 13C). This indicates that, at least during the development of the Tlaxiaco River, the Oaxacan Complex was a major topographic high that bounded the Tlaxiaco Basin to the east (Fig. 13C). Such a scenario is supported by the apatite and titanite fission-track data of Abdullin et al. (2020), which indicate that the Oaxacan Complex was episodically exhumed between Late Triassic and Middle Jurassic time along the Caltepec fault (Figs. 1 and 13C) as a result of crustal extension associated with Pangea breakup. Based on these data, we interpret the Caltepec fault as the eastern boundary of the Tlaxiaco Basin (Fig. 13C). According to these considerations, the stratigraphy and internal architecture of the Tlaxiaco Basin, at least in the areas explored in this work, were largely controlled by two major faults, the Salado River–Axutla and Caltepec faults, which represent the northern and eastern boundaries of the Tlaxiaco Basin, respectively. These faults produced the exhumation of different crustal blocks at different times. The Salado River–Axutla fault activated first and produced the exhumation of the Acatlán Complex along the northern boundary of the Tlaxiaco Basin, resulting in the deposit of the Cualac formation fan system (Fig. 13A). Subsequently, the Salado River–Axutla fault progressively became inactive, and the Caltepec fault was activated, producing the exhumation of the Oaxacan Complex along the eastern boundary of the Tlaxiaco Basin and the development of the Tlaxiaco River and associated overbank areas (Figs. 13B and 13C). For a certain time, the Cualac fan system and the Tlaxiaco River interacted, as suggested by the transitional stratigraphic contact between the Cualac formation and lower Tecocoyunca group (Fig. 13B).

Our data indicate that the Salado River–Axutla and Caltepec faults are major structures that controlled the geometry and internal depositional architecture of the Tlaxiaco Basin during Early–Middle Jurassic time. This indicates that the Salado River–Axutla and Caltepec faults took part in the process of continental attenuation related to Pangea breakup. The Caltepec fault has a NNW trend that is similar to the trend of other major normal faults (e.g., Oaxaca, Texcalapa, and El Sabino faults; Alaniz-Álvarez et al., 1996; Campos-Madrigal et al., 2013; Fig. 1) that developed between eastern Mexico and the Yucatán block during Early–Middle Jurassic time (Fig. 14A). These faults are the early manifestation of the development of a major transform boundary, the Tamaulipas-Chiapas transform, which by the end of Middle Jurassic time produced the anticlockwise rotation of the Yucatán block and the opening of the Gulf of Mexico (e.g., Pindell and Kennan, 2009; Figs. 14A and 14B). Therefore, our data support previous reconstructions that suggest that the development of the Tamaulipas-Chiapas transform fundamentally influenced the tectono-sedimentary evolution of Mexico during Early and Middle Jurassic time (Goldhammer, 1999; Padilla and Sánchez, 2007; Pindell and Kennan, 2009; Nova et al., 2019). On the other hand, recognizing the Salado River–Axutla fault as a major Lower–Middle Jurassic structure controlling the evolution of adjacent sedimentary basins shows that continental extension during Pangea breakup was also accommodated by WNW-trending faults. Therefore, our work offers a new perspective for researchers that aim to reconstruct the kinematics of Pangea breakup in Mexico. The existence of major Jurassic faults in Mexico with a WNW trend and a normal sinistral displacement was tentatively suggested by some authors (Anderson and Schmidt, 1983; Dickinson and Lawton, 2001; Pindell and Kennan, 2009). According to them, these faults would have placed south and central Mexico to the NW of its present location during Early and Middle Jurassic time, avoiding the large overlap between North and South America in the Pangea reconstruction (Fig. 14A). However, the existence of these Jurassic WNW-trending faults has been challenged by several authors during the past two decades, and the idea of sinistral block motion in Mexico during Pangea breakup has been largely downplayed (e.g., Iriondo et al., 2005). The kinematics of the Salado River–Axutla fault during Early–Middle Jurassic time is difficult to establish because of the superposition of kinematic indicators suggesting different motions (Martiny et al., 2012; this work). The kinematic indicators superposition advocates for a complex history with multiple reactivation episodes overprinted. Considering that the Tlaxiaco Basin is a rift basin, as indicated by the regional tectonic setting under which it developed, at least a normal displacement is required to accommodate the up to ~1300-m-thick alluvial fan deposits of the Cualac formation along the northern boundary of the basin. At present, a possible lateral kinematic component cannot be excluded for the Salado River–Axutla fault, making this structure the subject of interest of future studies that aim to explore the potential of sinistral displacements in Mexico as a valid solution to the North America–South America overlap in the reconstruction of Pangea.

Our results indicate that the geometry and tectono-sedimentary evolution of the Early–Middle Jurassic Tlaxiaco Basin was influenced by the activity of two major faults. The WNW-trending Salado River–Axutla fault bounded the basin to the north and produced the exhumation of the Paleozoic Acatlán Complex, which was drained to the southeast and southwest by a set of alluvial fans. Subsequently, the NNW-trending Caltepec fault produced the exhumation of the Proterozoic Oaxacan Complex, which formed a prominent topographic relief bounding the Tlaxiaco Basin to the east. The Tlaxiaco River drained the Oaxacan Complex to the west into the Tlaxiaco Basin and progressively buried the previously formed alluvial fan set. U-Pb ages and biostratigraphic data bracket the activity of the Salado River–Axutla and Caltepec faults between ca. 176 and ca. 168 Ma. Therefore, our work shows that the continental attenuation produced during Pangea breakup was accommodated not only by NNW-trending faults associated with the development of the Tamaulipas-Chiapas transform, but also by a WNW-trending fault. Due to the difficulty in determining its kinematics, the potential of the WNW-trending Salado River–Axutla fault in producing sinistral block displacements that can solve the North America–South America overlap problem in the Pangea reconstruction is unknown and must be clarified in future works.

This research was funded by Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica (PAPIIT) grant IN104018 to Michelangelo Martini. The first author acknowledges the Consejo Nacional de Ciencia y Tecnología de México (CONACYT) for the Ph.D. scholarship. We thank Fernando Ortega-Gutiérrez for helpful discussions on basement complexes of southern Mexico; Joaquín Aparicio (Instituto de Geología, Universidad Nacional Autónoma de México [UNAM]) for preparing whole-rock thin sections; Juan Tomás Vazquez (Centro de Geociencias, UNAM) for staining thin sections and preparing polished heavy-minerals mounts; Carlos Ortega Obregón (Centro de Geociencias, UNAM) for providing technical assistance during isotopic dating; Sandra E. Guerrero Moreno, Ismael Luna Osorno, Elisa Vianey Malpica Osorio, Berlaine Ortega Flores, Rodrigo Gutiérrez Navarro, Mariana Peña Guerrero, and Urenia Navarro Sánchez, for helping during fieldwork. We are grateful to Kurt Sundell and Cecilia del Papa, whose constructive comments and suggestions improved the manuscript