Provenance evolution during progressive rifting and hyperextension using bedrock and detrital zircon U-Pb geochronology, Mauléon Basin, western Pyrenees Open Access

Stratigraphy at rifted and passive continental margins is an important recorder of continental extension and breakup. The stratigraphic record can be used to reconstruct the temporal variations in the interplay between extensional tectonics and sedimentation, and to provide a more complete understanding of the tectonic, climatic, thermal, and geomorphic evolution of continental rifted margins and their transition to passive margins. For magma-poor margins in particular, the processes that occur during continental extension, break-up, and margin development have long been debated and modeled on the basis of observations from rifted margins such as the Iberia-Newfoundland or northwest Australian margins, and exhumed fossil rifted margins such as in the eastern Alps, as well as numerical models (e.g., Froitzheim and Eberli, 1990; Driscoll and Karner, 1998; Whitmarsh et al., 2001; Pérez-Gussinyé and Reston, 2001; Huismans and Beaumont, 2002; Lavier and Manatschal, 2006; Osmundsen and Ebbing, 2008; Péron-Pinvidic and Manatschal, 2009; Unternehr et al., 2010). These geological and numerical models have focused on the processes accommodating lithospheric break-up as well as structural evolution during progressive strain localization from diffusive rifting, crustal necking (extension and thinning that leads to a zone of crustal thinning from ∼30 km to ≤10 km and an inflection point in the seismic Moho), hyperextension (crustal thinning to ≤10 km), mantle exhumation, and eventual lithospheric separation to seafloor spreading (Péron-Pinvidic and Manatschal, 2009; Mohn et al., 2010). While this structural and geometric evolution has been discussed (e.g., Whitmarsh et al., 2001), the spatial and temporal complexities of tectonically controlled sedimentation, such as the relative amounts of mixing between proximal and distal sources, recycling between subbasins, spatial and temporal basin compartmentalization, and subbasin reintegration during progressive rifting remain unknown. Such sedimentary routing variations could have implications for stratigraphic models of basin evolution, structural controls on subsidence, and the progressive geometric evolution of rifted margins.

Where preserved and exposed, hyperextended systems offer a window into the synrift and early postrift sedimentary records of rifted continental margins that are generally inaccessible due to their submarine locations and burial by thick passive margin sediments (e.g., Iberian-Newfoundland margin and the Gulf of Mexico). This has restricted tectonic sedimentation studies to reflection and refraction seismic surveys and/or sparse boreholes that generally do not penetrate the synrift sedimentary sections within the necking domain or the distal rifted margin. In contrast, fossil rifted margins preserved in orogens offer the opportunity to explore and characterize the complexities of the tectonic and stratigraphic evolution of hyperextended margins. Such fossil margins, such as in the Eastern Alps, tend to be variably tectonized and overprinted by later orogenic deformation, and often preserve only a fragmented synrift basin record (e.g., Masini et al., 2012). The Early Cretaceous Mauléon rift basin in the western Pyrenees is therefore noteworthy because it offers access to a relatively complete synrift to postrift sedimentary record in a setting where the postrift inversion and shortening-related structural and thermal overprinting seem to be relatively modest (e.g., Teixell, 1998; Lagabrielle and Bodinier, 2008; Jammes et al., 2009; Masini et al., 2014; Tugend et al., 2014; Vacherat et al., 2014), making it an ideal locality for the reconstruction of synrift and postrift sedimentation and for the understanding of rift-related basin evolution at hyperextended rift margins.

To better understand the evolution of the Mauléon Basin during rifting, we apply a combination of bedrock and detrital zircon (DZ) U-Pb dating. DZ U-Pb analyses have evolved over the past decade into a powerful tool in process-oriented provenance analyses (e.g., Gehrels, 2014, and references therein) and studies have employed these techniques to focus on source-to-sink problems at passive margins and in intercontinental rifts (e.g., Cawood and Nemchin, 2000; Lamminen and Köykkä, 2010; Craddock et al., 2013; Lamminen et al., 2015); however, no systematic study has focused on a high-resolution interpretation of sedimentary provenance at hyperextended continental rifted margins with the emphasis on source-to-sink changes during progressive rifting. This study aims to fill this gap by presenting zircon U-Pb data from the western Pyrenean tectonic hinterland and prerift strata, as well as a systematic detrital provenance analysis for synrift and postrift strata from the south Mauléon subbasin (SMB) and north Mauléon subbasin (NMB). By defining the major DZ age components of Paleozoic to Mesozoic prerift strata, it is possible to fingerprint and deconvolve the DZ age distributions of basin-fill deposits and reconstruct the lateral and temporal variations in sedimentation within and between structural domains and subbasins during rifting. Note that previous studies have applied zircon dating to define geochronological and thermochronological constraints for the development of the Pyrenees (e.g., Vacherat et al., 2014), but many of these studies mainly focused on the evolution of the central and eastern Pyrenees from extension to early continental convergence (e.g., Whitchurch et al., 2011; Filleaudeau et al., 2011; Mouthereau et al., 2014). In contrast, the data presented here serve to illuminate the spatial and temporal variations in sedimentation in the western Pyrenees during synrift and postrift time and provide constraints on sedimentary processes during progressive extension and rifting.

Western Pyrenean moderate-grade metamorphic thrust sheets and lower amphibolite to granulite facies basement resulted from the collision of Gondwana and Baltica-Laurentia during the late Paleozoic Variscan orogeny (e.g., Matte, 1986; Ziegler, 1990). Permian extension and orogenic collapse was followed first by the northward propagation of North Atlantic rifting, resulting in Jurassic to Early Cretaceous exhumation and marine incursion, and then later by the initiation of Cretaceous rifting and seafloor spreading in the Bay of Biscay (Jammes et al., 2009; Filleaudeau et al., 2011; Vacherat et al., 2014). While diffuse extensional faulting has been suggested in the northern and southern Pyrenees during Early Cretaceous time from 145 to 132 Ma, which may represent early stages of Cretaceous extension (Vergés and García-Senz, 2001; Jammes et al., 2009; Vacherat et al., 2014), the future Mauléon Basin area was characterized by stable carbonate platforms with some evidence for limited Early Cretaceous salt mobility. The onset of dramatic synrift subsidence, however, clearly postdates the deposition of lower Aptian carbonates (Ducasse and Velasque, 1988; Canérot et al., 2001; Masini et al., 2014).

Continental extension in the Bay of Biscay, leading to break-up and seafloor spreading, propagated eastward and resulted in large-magnitude regional crustal extension (∼120%) in southwestern France and northern Spain, local thinning of the crust to 0–10 km, and local exhumation of sublithospheric mantle rocks (Lagabrielle and Bodinier, 2008; Jammes et al., 2009; Masini et al., 2014). Such dramatic crustal thinning is now widely referred to as hyperextension (e.g., Sutra and Manatschal, 2012). One of these areas of western Pyrenean hyperextension is the Mauléon-Arzacq rift system, which is characterized north to south by four separate domains. The northernmost domain is the Arzacq Basin, where the European continental crust beneath the basin shows southward thinning, and the sedimentary sequences show southward thickening, approaching the hyperextended region (Teixell, 1990; Daignières et al., 1994; Biteau et al., 2006). To the south, the Grand Rieu high was a barrier between the Arzacq Basin and the Mauléon Basin until the end of active hyperextension. The third domain is the Cretaceous Mauléon Basin (Fig. 1), which formed over hyperextended crust and can be subdivided into the NMB and SMB. The Mauléon Basin is bounded by fault zones to the east and west, by the Grand Rieu high to the north, and by the Axial domain (western equivalent of the Axial Zone of the central and eastern Pyrenees or the Pyrenean hinterland) to the south.

The stratigraphy of the Mauléon Basin can be subdivided into two groups, western Pyrenean prerift strata and basin-fill strata (Fig. 2). The western Pyrenean prerift units include lower crustal granulite, Paleozoic metasedimentary strata, and late Paleozoic to early Mesozoic sedimentary strata. The basin-fill units include synrift and postrift sedimentary basin deposits. The following is a brief description of these units; see Masini et al. (2014) and references therein for a detailed discussion of Mauléon Basin stratigraphy.

Vielzeuf (1984) identified two lower crustal granulite units: the lower unit is a metabasic granulite and the upper unit is a quartzofeldspathic metasedimentary granulite with a Cambrian to Ordovician protolith (Boissonnas et al., 1974). Both of these units underwent granulite facies metamorphism during the Variscan orogeny, with peak temperatures and pressures of 775 ± 50 °C and 6 ± 0.5 kbar (Masini et al., 2014, and references therein). Reported biotite ⁴⁰Ar/³⁹Ar ages indicate that these strata had been exhumed to middle crustal depths of ∼10 km by Late Triassic to Early Jurassic time (Masini et al., 2014). Stratigraphically above the granulites are lower Paleozoic sedimentary strata (Fig. 2) that underwent low-grade anchizonal to lower greenschist facies metamorphism during the Variscan orogeny (Heddebaut, 1973).

The deformed and metamorphosed lower Paleozoic strata are overlain by undeformed Permian conglomerate, sandstone, silt, and pelite. As is the case across much of central and western Europe, the Permian section is overlain by Buntsandstein, a group of Triassic continental deposits consisting of shale, sandstone, and conglomerate (e.g., Germanic facies; Curnelle, 1983; Fréchengues, 1993; Masini et al., 2014), which are in turn overlain by Late Triassic transgressive Muschelkalk platform carbonate, Keuper evaporites, and a second carbonate platform (see Fig. 2). These deposits are cut by a major erosional unconformity due to a Late Jurassic regression (Curnelle, 1983; Fréchengues, 1993; Masini et al., 2014). The final prerift deposits are Barremian to lower Aptian carbonate and marl (Masini et al., 2014).

Mauléon Basin synrift deposition (Fig. 2) began with upper Aptian and Albian carbonate and marl (Masini et al., 2014). Toward the Axial domain, these grade into Albian and Cenomanian delta-derived siliciclastic turbidite and conglomerate (Boirie and Souquet, 1982). These megasequences and flysch were deposited concurrently and diachronously in the SMB and NMB as rifting occurred (Souquet et al., 1985; Masini et al., 2014). Postrift deposition (Fig. 2) began during Cenomanian time as siliciclastic sedimentation ended with another transgression that covered the Axial domain, which led to the deposition of platform carbonate in a calc-turbidite and hemipelagic system (Masini et al., 2014).

The opening of the SMB was accommodated by the south Mauléon detachment (SMD), which exhumed upper crustal Paleozoic section, as shown in Figure 1 (Masini et al., 2014). From Albian to Cenomanian time, the Axial domain and the Jara-Arbailles ridge were the southern and northern extents, respectively, of the SMB (Boirie and Souquet, 1982; Souquet et al., 1985; Claude, 1990; Masini et al., 2014). The current model envisions that from late Aptian to early Albian time, progressive strain associated with crustal thinning was transferred from the SMD along a mid-crustal detachment to the lower crust of the Arzacq Basin, leading to an asymmetric rift geometry (Masini et al., 2014).

During mid-Albian time, the locus of extension shifted northward with the inception of the north Mauléon detachment (NMD), forming the NMB (Jammes et al., 2009; Masini et al., 2014). The Jara-Arbailles ridge became a breakaway block separating the SMB and NMB (Fig. 1). On the basis of crosscutting relations, the NMD was active during the deposition of the second megasequence and exhumed the already thinned lower crust and mantle (Boirie and Souquet, 1982; Jammes et al., 2009; Masini et al., 2014). Due to lack of exposure of the stratigraphic record, Masini et al. (2014) noted that it is difficult to determine the timing of cessation of activity along the NMD.

Estimates of shortening that occurred in the west-central Pyrenees during Pyrenean orogenesis between Late Cretaceous and early Miocene time range from ∼75–165 km (∼30%), on the basis of north-south balanced cross sections, to ∼180 km of total convergence from plate reconstructions (Teixell, 1998; Beaumont et al., 2000; Rosenbaum et al., 2002; Mouthereau et al., 2014; Teixell et al., 2016). This variability may stem from balanced cross-section assessments underestimating shortening by not taking into account the role of hyperextended domains at the early stages of convergence (Mouthereau et al., 2014). While the shortening values vary, balanced cross sections and surface structure restorations show that the amount of shortening decreases from east to west across the Pyrenees (Seguret and Daignieres, 1986; Teixell, 1998).

In the western Pyrenees, shortening caused the Mauléon Basin to be partially inverted as a tectonic pop-up block. The North Pyrenean frontal thrust system to the north of the Mauléon Basin and the Igountze-Mendibelza or Lakora thrust to the south of the Mauléon Basin accommodated most of this shortening as the basin was thrust northward over the former Grand Rieu high and Arzacq domains and southward onto the Axial domain (Casteras, 1969; Teixell, 1990, 1998; Muñoz, 1992; Daignières et al., 1994). Masini et al. (2014) noted that within the Mauléon Basin these thrust systems cut through the basement while east of the basin the thrusts cut the sedimentary cover. They suggested that this lateral change in structural level of thrusting may have caused more of the deformation to be accommodated at a deeper crustal level in the Mauléon Basin, allowing for greater preservation of prerift, synrift, and postrift structures as compared to the central and eastern Pyrenees.

In this study we seek to systematically define the DZ U-Pb age distributions of Paleozoic to Mesozoic prerift strata and thereby to fingerprint and deconvolve the DZ U-Pb age distributions of basin-fill deposits, with the objective of reconstructing the lateral and temporal variations in sedimentation within and between structural domains and subbasins during continental extension and rifting. To pursue this goal, samples were taken from various stratigraphic intervals within the necking, hyperthinned, and exhumed mantle domains (as mapped by Tugend et al., 2014). Zircon was then separated from a total of 47 samples: 30 Paleozoic and Mesozoic metasedimentary prerift samples, 2 felsic granulite samples from the Labourd Massif, and 15 sedimentary synrift and postrift samples from the Mauléon subbasins (Fig. 1).

Standard mineral separation techniques were applied, including crushing, grinding, water table concentration, heavy-liquid density, and magnetic susceptibility separation techniques. Zircon separates were then sprinkled onto double-sided tape on 1 in (∼2.54 cm) epoxy resin mounts, and at least 120 zircon grains were randomly chosen to be analyzed using laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) U-Pb geochronology to obtain provenance data sets that resolve any age component >5% (Vermeesch, 2004). The analyses were completed using a PhotonMachine Analyte G.2 excimer laser with a large-volume Helex sample cell and a Thermo Element2 ICP-MS. GJ1 was used as the primary reference standard (Jackson et al., 2004) and Pak1 (in-house zircon standard, 42 Ma, from Pakistan) was used as secondary reference standard. A 30 μm laser spot was used to ablate ∼16 μm deep pits on the flat prism plane of the nonpolished, tape-mounted zircons, providing a depth profile of each analyzed grain. This technique enables the resolution of multiple zircon growth zones evident from core and rim ages (Stockli and Stockli, 2013). The data from the analyses were then reduced using Iolite data reduction software and VizualAge (Paton et al., 2011; Petrus and Kamber, 2012). For increased precision the ages presented are ²⁰⁶Pb/²³⁸U ages for zircons younger than 1200 Ma, and ²⁰⁷Pb/²⁰⁶Pb ages for zircons older than 1200 Ma (Gehrels et al., 2008). All ages reported use 2σ absolute propagated uncertainties, ²⁰⁷Pb/²⁰⁶Pb ages are <30% discordant, and ²⁰⁶Pb/²³⁸U ages are <10% discordant (Gehrels, 2011). The discordance reported is calculated with the ²⁰⁶Pb/²³⁸U and ²⁰⁷Pb/²³⁵U ages if younger than 1200 Ma and the ²⁰⁶Pb/²³⁸U and ²⁰⁷Pb/²⁰⁶Pb ages if older than 1200 Ma. Mineral separation and LA-ICP-MS analyses were completed at the UTChron facilities at the Jackson School of Geosciences at the University of Texas at Austin.

For the presentation and discussion of the detrital zircon U-Pb results, we adhere to the following terminology in accordance with common use in published literature (e.g., Davis et al., 2003; Vermeesch, 2004; Gehrels et al., 2008; Castiñeiras et al., 2008; Filleaudeau et al., 2011; Whitchurch et al., 2011; Malusà et al., 2013; Mouthereau et al., 2014; Pullen et al., 2014). The entire spectrum of U-Pb ages is referred to as an age distribution. Specific groups of age modes for samples or stratigraphic intervals that reflect a particular provenance are referred to as signatures, and the individual modes of the age distribution are discussed as age components.

All DZ U-Pb analyses are displayed as kernel density estimation plots (KDE; Vermeesch, 2012) and tabulated in Supplemental File S1¹. The subsequent presentation of the data focuses primarily on DZ U-Pb ages younger than 1200 Ma because that is the age range where the most diagnostic variations in ages are observed, and the aim of the bedrock DZ analyses is to identify age components in each of the prerift strata to ultimately fingerprint the synrift and postrift strata. Residual DZ ages older than 1200 Ma, which make up only ∼15%–25% of the ages obtained, generally do not show distinct age variations that would be useful in discerning differences in sedimentary provenance or routing between synrift and postrift basin samples.

We analyzed 32 samples to identify the western Pyrenean prerift strata signatures and characterize the bedrock source terranes. These include 9 granulite samples from the hyperthinned domain that yielded 1123 individual U-Pb ages (Figs. 1 and 3A). A total of 19 samples (15 Ordovician, 3 Devonian, and 1 Carboniferous) from the Paleozoic metasedimentary units (metamorphosed during the Variscan orogeny) yielded 1843, 378, and 139 individual DZ U-Pb ages, respectively (Figs. 1 and 4). A total of 4 samples (1 Permian and 3 Triassic) from unmetamorphosed (post-Variscan) late Paleozoic to early Mesozoic prerift strata yielded 117 and 368 individual DZ U-Pb ages, respectively (Figs. 1 and 4). These results are presented in stratigraphic order and displayed as KDE plots (Figs. 3A and 4). The KDE plots for each individual sample (Fig. 4) are combined into a single KDE to illustrate the DZ U-Pb signature for each stratigraphic interval (Fig. 5).

Nine granulite samples from the Labourd Massif were analyzed to define the DZ signature of the lower crustal units exposed in the distal hyperthinned domain of the NMB (Figs. 1 and 3A). The granulites of the Labourd Massif can be subdivided in terms of their protolith composition into orthogneissic and paragneissic granulites. We obtained ∼200 zircon U-Pb ages from orthogneissic granulite samples 12WPY08 and 12WPY48, which only display a single magmatic age mode (Fig. 3B); these samples yielded concordant U-Pb ages of 279 ± 2 Ma and 274 ± 2 Ma, respectively (Fig. 3C).

We obtained 923 U-Pb ages from 7 paragneissic granulite samples, which display multimodal DZ age distributions (Fig. 3B). All of these samples (except 12WPY12) show distinct zircon rim ages similar to the orthogneissic magmatic zircons (Fig. 3D). These zircon overgrowth ages (n = 378) cluster ca. 295 Ma. Given their paragneissic detrital nature, samples 12WPY12, 12WPY18, and 13WPY05 show significant DZ core ages that range from 778 to 453 Ma (n = 325, 35% of all U-Pb ages obtained from these samples) with a distinct mode of ca. 590 Ma. In addition, sample 12WPY12 shows a significant secondary DZ core age that ranges from 1071 to 806 Ma (n = 60, 7%) and has a mode of ca. 955 Ma.

All 15 Ordovician samples were collected from greenschist to amphibolite facies metasedimentary rocks: 5 from Mount Monoa in the distal necking domain, 1 near the base of Mount Jara in the proximal hyperthinned domain, 7 from Mount Baygoura in the hyperthinned domain, and 2 from Pic de Garralda in the distal hyperthinned domain (Fig. 1). These samples exhibit two distinct DZ U-Pb age components (Fig. 5), one ranging from 748 to 521 Ma (n = 678 ages, 37%) with a mode of ca. 626 Ma and the second ranging from 1189 to 850 Ma (n = 480 ages, 26%) with a mode of ca. 994 Ma. Only one sample, 12WPY42 from Mount Baygoura, shows a significant deviation in DZ age spectrum from the other samples with an additional DZ age component that ranges from 497 to 423 Ma (n = 29 ages, <1%) with a mode of ca. 470 Ma (Fig. 4).

Two Devonian samples were collected from greenschist to amphibolite facies metasedimentary strata, one from Mount Monoa in the distal necking domain and one near the base of Mount Jara in the proximal hyperthinned domain (Fig. 1). Both samples yielded two distinct DZ age components (Fig. 5). One ranges from 731 to 525 Ma (n = 125 ages, 33%) and has a mode of ca. 675 Ma. The second ranges from 1187 to 846 Ma (n = 100 ages, 27%) and has a mode of ca. 999 Ma. These samples also show a less prominent DZ age component that ranges from 834 to 740 Ma (n = 38 ages, 10%), with a mode of ca. 768 Ma.

The single Carboniferous sample was collected from a Namuro-Westphalian greenschist facies metasedimentary rock in the proximal necking domain of Mendibelza (Fig. 1). This sample shows three distinct DZ age components (Fig. 5). The first ranges from 364 to 267 Ma (n = 22 ages, 16%) and has a mode of ca. 334 Ma, the second ranges from 509 to 399 Ma (n = 26 ages, 19%) and has a mode of ca. 472 Ma, and the third ranges from 734 to 526 Ma (n = 49 ages, 35%) and has a mode of ca. 610 Ma. This sample also has two less prominent DZ age components. The first ranges from 847 to 767 Ma (n = 9 ages, 7%) and has a mode of ca. 801 Ma. The second ranges from 1134 to 904 Ma (n = 19 ages, 14%) and has a mode of ca. 998 Ma.

The single Permian sample was collected from sandstone interbedded with conglomerate near the top of Mount Hautza (Fig. 1). This Permian sample shows three distinct DZ age components, one ranging from 373 to 275 Ma (n = 13 ages, 11%) with a mode of ca. 300 Ma, a second ranging from 727 to 531 Ma (n = 43 ages, 37%) with a mode of ca. 606 Ma, and a third ranging from 883 to 739 Ma (n = 19 ages, 16%) with a mode of ca. 780 Ma. This sample also has another less prominent DZ age component that ranges from 1027 to 940 Ma (n = 11 ages, 9%) with a mode of ca. 1002 Ma (Fig. 5).

Two Triassic samples were collected from sandstone, one near the base and one near the top of Mount Jara, in the proximal hyperthinned domain. A third sample was collected from sandstone near the top of Mount Hautza (Fig. 1). These samples show one significant DZ age component (Fig. 5), which ranges from 719 to 534 Ma (n = 168 ages, 46%) and has a mode of ca. 597 Ma. These samples also display three minor DZ age components (Fig. 5). The first ranges from 388 to 263 Ma (n = 15 ages, 4%) and has a mode of ca. 328 Ma, the second ranges from 512 to 437 (n = 22 ages, 6%) and has a mode of ca. 475 Ma, and the third ranges from 1039 to 931 Ma (n = 39 ages, 13%) and has a mode of ca. 948 Ma.

We analyzed 15 samples from synrift and postrift strata across the Mauléon Basin and its different rift domains to constrain the spatial and temporal variations in sediment provenance during continental extension and rifting. The DZ U-Pb results are presented from the proximal to distal domains of the Mauléon hyperextended region. All ages obtained are displayed in KDE plots (Fig. 6).

Seven samples from the Mendibelza area characterize the synrift to postrift stratigraphic evolution of the proximal necking domain (Fig. 1). These samples stratigraphically transition upsection from Albian Spicula Marls (n = 1) to Albian to Cenomanian turbidite (n = 1), conglomerate (n = 4), and sandstone (n = 1). The DZ analyses show five distinct age components (see Fig. 7 for graphical presentation of the distributions): Carboniferous–Permian (mode ca. 311 Ma, n = 142 ages, 16%), early Paleozoic (mode ca. 469 Ma, n = 89 ages, 10%), Cryogenian–Ediacaran (mode ca. 606 Ma, n = 326 ages, 37%), Tonian (mode ca. 801 Ma, n = 46 ages, 5%), and Stenian–Tonian (mode ca. 961 Ma, n = 128 ages, 15%).

In the proximal necking domain, DZ provenance signatures exhibit a systematic provenance shift from Albian through Cenomanian time (Fig. 7). The Carboniferous–Permian DZ age component (mode ca. 311 Ma) increased in abundance from ∼5% to ∼37% with minor fluctuations and then decreased to ∼30% in the stratigraphically highest sample (Fig. 7). Across the stratigraphic section, the early Paleozoic DZ age component (mode ca. 469 Ma) had a relatively constant abundance (∼7%) until it increased to ∼24% toward the top of the section. The Cryogenian–Ediacaran DZ age component (mode ca. 606 Ma) showed the highest abundance of any of the age components but decreased in abundance from ∼50% to ∼33% upsection. The Tonian DZ age component (mode ca. 801 Ma) was consistently the least abundant, remaining relatively constant at <10%. The Stenian–Tonian DZ age component (mode ca. 961 Ma) showed a decrease in abundance from 27% to <10% with minor fluctuations. Overall the samples from the Mendibelza area indicate a general increased amount of younger zircon ages and a decrease in the older zircon ages upsection through the Albian to Cenomanian stratigraphy.

Due to limited exposure and preservation of the SMB, only one sample was collected and analyzed from sandstone at the base of Mount Monoa. This sample characterizes synrift deposition in the distal necking domain (Fig. 1). Here there are only three main DZ age components (Fig. 7): Cryogenian–Ediacaran (mode ca. 645 Ma, n = 48 ages, 40%), Tonian (mode ca. 786 Ma, n = 11 ages, 9%), and Stenian–Tonian (mode ca. 995 Ma, n = 23 ages, 19%).

Seven samples from the NMB characterize the hyperthinned to exhumed mantle domain (Fig. 1). These samples were collected from Albian to Cenomanian sedimentary rock near the Abarratia quarry (n = 2), Cenomanian to Turonian turbidite (n = 3) and sandstone (n = 1), and Coniacian to Santonian conglomerate (n = 1). Similar to samples from the Mendibelza area and the SMB, DZ analysis indicates five main age components, as shown in Figure 7: Carboniferous–Permian (mode ca. 325 Ma, n = 195 ages, 23%), early Paleozoic (mode ca. 485 Ma, n = 90 ages, 11%), Cryogenian–Ediacaran (mode ca. 617 Ma, n = 259 ages, 31%); Tonian (mode ca. 776 Ma, n = 40 ages, 5%), and Stenian–Tonian (mode ca. 980 Ma, n = 121 ages, 15%).

During Albian to Cenomanian time, the Carboniferous–Permian DZ age component (mode ca. 325 Ma) dramatically increased in abundance from ∼34% to ∼70%. In Cenomanian time, this DZ age component decreased to ∼15%, before it increased to ∼29% during Turonian time, and finally decreased to ∼4% during Coniacian to Santonian time. The early Paleozoic DZ age component (mode ca. 485 Ma) indicated a decrease in abundance from ∼12% to ∼7% during Albian to Cenomanian time, increased to ∼21% during Cenomanian to Turonian time, and finally decreased to ∼8% during Coniacian to Santonian time. The Cryogenian–Ediacaran DZ age component (mode ca. 617 Ma) decreased from ∼35% to ∼17% during Albian to Cenomanian time, and increased to ∼44% during Cenomanian time. During Turonian time this DZ age component decreased again to ∼30%, before it increased to ∼51% during Coniacian to Santonian time. The Tonian DZ age component (mode ca. 776 Ma) abundance was consistently <∼10%. The Stenian–Tonian DZ age component (mode ca. 980 Ma) decreased in abundance from ∼12% to ∼4% during Albian to Cenomanian time, and later during Cenomanian time the abundance increased to ∼27%. This DZ age component then decreased during Cenomanian to Turonian time to ∼14% and increased again to ∼33% during Coniacian to Santonian time.

Through comparing the prerift DZ U-Pb signatures with the synrift and postrift DZ age distributions, we are able to interpret and reconstruct the basin sedimentary evolution during progressive rifting. These findings are then used to construct a basin evolution model for hyperextended rift margins. In order to do this, DZ U-Pb age components are separated into those that are diagnostic and those that are nondiagnostic in defining the provenance signature of the samples. Diagnostic DZ age components are defined as those that were not found in all of the analyzed samples and therefore can help to distinguish between stratigraphic units. In contrast, nondiagnostic DZ age components are defined as those that are present in all of the prerift intervals.

Although the age range and mode exhibit some minor variations between samples and between stratigraphic intervals, all of the prerift samples share nondiagnostic age components of ca. 615 Ma, ca. 780 Ma, and ca. 1000 Ma, predominantly the ca. 615 Ma and ca. 1000 Ma components (Fig. 5). Thus, we suggest that zircons with these nondiagnostic DZ ages were continually recycled into Paleozoic metasedimentary and prerift sedimentary units, producing these common provenance signatures.

While these nondiagnostic DZ age components are inadequate when characterizing differences between prerift strata, they illustrate a signature that is vital in determining the provenance of the prerift units. This DZ signature is similar to those that are found in Ediacaran to Carboniferous age deposits throughout parts of Iberia (Fernández-Suárez et al., 2002; Bea et al., 2010; Fernández-Suárez et al., 2013; Gutiérrez-Alonso et al., 2015). Zircon grains corresponding to this signature likely originated from the East African orogen, northern Egypt, and the Sinai Peninsula, and thus Iberia, and consequently the western Pyrenees, were much closer to eastern Gondwana than previously expected (Fernández-Suárez et al., 2002, 2013; Bea et al., 2010; Gutiérrez-Alonso et al., 2015).

The key diagnostic DZ ages collected from the orthogneissic and paragneissic granulites from the Labourd Massif are ca. 274, ca. 279 Ma, and ca. 295 Ma (Fig. 3). For the orthogneissic granulites, this magmatic component dates early to middle Permian zircon crystallization, which is within the 304 to 266 Ma age range, and more specifically resulted from a magmatic event ranging from 276 to 266 Ma, established elsewhere by Pereira et al. (2014) on the basis of zircon U-Pb analyses from the southern Pyrenees. These plutons are therefore related to Variscan magmatism; similar early to middle Permian magmatic events are also preserved in the southern, western, and northern Pyrenees. In addition, the ages of the orthogneissic granulites are similar to the defined age (278 Ma) of felsic lavas from the Ossau massif (Innocent et al., 1994). In the paragneissic granulites the DZ age of ca. 295 Ma represents a zircon rim age (Fig. 3B). Given that all but one paragneiss sample contained this zircon rim age, it is clear that granulite facies metamorphism was widespread throughout the Labourd Massif during the Variscan orogeny.

The paragneissic granulites yielded two separate trends in terms of inherited zircon core ages. The first trend is observed in the paragneiss sample that does not display the ca. 295 Ma rim age, 12WPY12 (Fig. 3B). This sample has significant ca. 480 Ma, ca. 600 Ma, and ca. 955 Ma age components, which are similar to the DZ ages found in the Ordovician samples (Fig. 5). This implies that the protolith for this paragneiss sample may have been deposited during Ordovician time and later underwent metamorphism during the Variscan orogeny. The other six paragneiss samples in Figure 3B show a less abundant ca. 615 Ma DZ age component and are lacking the ca. 1000 Ma DZ age component. It is possible that the U-Pb depth profiling approach might not have penetrated all of the inherited cores, leading to a decrease in the abundance of these inherited zircon core ages. Alternatively, the protoliths of these six samples may not have been derived from Ordovician, but rather from Cambrian units (Boissonnas et al., 1974). Fernández-Suárez et al. (2013) noted that there is a general a lack of the ca. 1000 Ma DZ age component in some Cambrian age units in northern Iberia, and because no Cambrian samples were analyzed from the western Pyrenean prerift strata, it is possible that the Cambrian samples in this area may also lack this ca. 1000 Ma DZ age component; additional analysis from Cambrian samples from the western Pyrenees would be needed to verify this.

Most of the samples from Ordovician and Devonian metasedimentary units show little variation in the DZ signature between samples and all display two (nondiagnostic) age components of ca. 625 Ma and ca. 1000 Ma; only one Ordovician metasedimentary sample shows a diagnostic DZ age component with an age mode of ca. 470 Ma (Fig. 4). These zircons are interpreted to be from magmatism that occurred during the transition from the Cadomian orogeny to the Variscan orogeny (Deloule et al., 2002; Cocherie et al., 2005; Castiñeiras et al., 2008; Denèle et al., 2009; Gutiérrez-Alonso et al., 2015). Exposures of these igneous bodies can currently be found in the eastern Axial zone and also north of the North Pyrenean fault (Whitchurch et al., 2011; Filleaudeau et al., 2011; Mouthereau et al., 2014).

Many of the same nondiagnostic DZ age components (ca. 610 Ma, ca. 800 Ma, ca. 1000 Ma) are present in the Carboniferous units, likely due to the recycling of sediments during the Variscan orogeny. These units also display two diagnostic DZ age components. The first, one of the most abundant, has an age mode of ca. 472 Ma. The only other unit bearing this DZ age component is a single sample from the Ordovician metasedimentary unit (Fig. 4). One possible explanation for these grains being deposited in the Carboniferous unit would be exhumation of Ordovician units, with this age component, during the Variscan orogeny and recycling of these grains into Carboniferous units. The second diagnostic DZ age component has an age mode of ca. 330 Ma (Fig. 5), which is likely associated with synorogenic Variscan magmatism between 350 and 315 Ma (Schaltegger et al., 1996; Castro et al., 2002; Gutiérrez-Alonso et al., 2011). If these zircons are magmatic, the similarity in the age mode (ca. 330 Ma) and the depositional age of the unit (326–304 Ma) could possibly indicate rapid denudation during late Carboniferous time. Alternatively, this similarity could indicate that these zircons are volcanic in origin, but late Variscan volcanism appears to be sparse until the latest Carboniferous time and only well documented for Permian time (Bixel, 1988; Innocent et al., 1994; Pereira et al., 2014). By the end of Carboniferous time the Variscan orogeny had come to an apex in the western Pyrenees with magmatism as well as the low-grade anchizonal to lower greenschist facies metamorphism of Ordovician–Carboniferous strata (Heddebaut, 1973).

While the Permian sample shares many of the same nondiagnostic characteristics with the Carboniferous sample (ca. 606 Ma, ca. 780 Ma, ca. 1002 Ma), it completely lacks the ca. 472 Ma DZ age component, which indicates that the Carboniferous units were not recycled into Permian units (Fig. 5); instead, the source terrane must have contained nondiagnostic DZ age components, such as in the Ordovician and Devonian strata. The ca. 300 Ma age component is the only diagnostic component in this unit (Fig. 5). These zircons are interpreted to be derived from late Variscan volcanic rocks, which crop out in the southern Pyrenees (Lago et al., 2004, and references therein).

One of the diagnostic aspects of the Triassic samples is the significant decrease in abundance of ca. 300 Ma and ca. 330 Ma DZ age components observed in Permian and Carboniferous units, respectively. This indicates minimal recycling of these units into the Triassic units (Fig. 5). The second diagnostic DZ age component is a minor ca. 475 Ma age component that is only found in the Carboniferous and Ordovician units, which may indicate minimal recycling from these units as well.

All of the western Pyrenean prerift samples were plotted using a standard statistical technique, multidimensional scaling (MDS), to highlight provenance trends during prerift time (Vermeesch, 2013). The resulting MDS plot can be separated into two patterns. The first is illustrated by granulite samples that trend nearly perpendicular to the other prerift samples (Fig. 8). In addition, the paragneissic granulites plot closer to the other prerift samples; the paragneissic samples share inherited DZ core ages similar to age components found in the other prerift strata. The second pattern shows nearly identical Ordovician and Devonian bedrock families that evolve in a diagonally linear trend to Carboniferous, Permian, and Triassic bedrock families, and nearly intersect the paragneissic granulites (Fig. 8). This linear trend observed in Paleozoic to early Mesozoic prerift strata does not represent a significant shift in detrital provenance, but simply reflects the addition of younger detrital U-Pb age components in progressively younger stratigraphic intervals.

With the signatures of the western Pyrenean prerift strata characterized, it is now possible to use the diagnostic DZ age components from the prerift units to interpret the Mauléon Basin DZ data. The significant variations in the DZ signatures of synrift and postrift sedimentary rocks from the proximal necking domain to the exhumed mantle domain of the NMB and SMB are used to determine the synrift to postrift basin evolution. We discuss the data from proximal to distal in the following.

The Mendibelza transect crosses the Albian to Cenomanian proximal necking domain and the synrift to postrift deposits exhibit a classic unroofing sequence (Fig. 7). Due to the presence of significant ca. 600 Ma and ca. 960 Ma, and lack of ca. 300 Ma DZ age components in the stratigraphically lowest Albian synrift samples, the source terrane for these samples is likely Paleozoic to early Mesozoic strata such as Ordovician, Devonian, and/or Triassic units (Figs. 5 and 7). Upsection, these samples show a significant increase in abundance of Carboniferous–Permian detrital zircons as the Cryogenian–Ediacaran and Stenian–Tonian detrital zircons generally decrease in abundance (Figs. 7 and 9). This upsection variation in basin deposits as rifting progressed during Albian to Cenomanian time can be accounted for by the addition of a source terrane with ample ca. 300 Ma zircons, while the Paleozoic source terrane became buried or depleted. In the proximal and necking domains, only Variscan plutons that are present throughout the Pyrenean Axial Zone in the central and eastern Pyrenees have been dated as ca. 300 Ma (e.g., Denèle et al., 2014). Therefore, Carboniferous–Permian detrital zircons in Albian to Cenomanian deposits in the Mendibelza area may have originated from similar Variscan-aged plutons that continued west from the Axial Zone into the Axial domain of the western Pyrenees, but are no longer exposed or have been completely eroded. This implies that from Albian to Cenomanian time, during the evolution from synrift to postrift, the proximal necking domain may have been sourced regionally from Axial domain Variscan plutons rather than locally from Paleozoic to early Mesozoic strata.

Due to limited exposure and preservation of the SMB or distal necking domain, only one Albian sample was collected and analyzed from the SMD hanging wall in Irouleguy Valley (Fig. 7). While SMB sediments were deposited synchronously with Albian deposits from the Mendibelza transect, the sample from the SMB lacks Carboniferous–Permian detrital zircons that are prevalent throughout the Mendibelza transect (Fig. 7). This suggests that during Albian hyperextension, the SMB was either isolated and disconnected from the Mendibelza basin, strongly implying locally sourced fault-controlled subbasins (SMB was sourced from the metamorphosed Paleozoic units exposed in the immediate footwall of the SMD), or the subbasins were not isolated, indicating that sediments eroded from Axial domain plutons would have accumulated in Albian–Cenomanian strata in the NMB (Mendibelza and the SMB would show distinct provenance signatures due to significant sediment bypassing).

The NMB is located within the hyperthinned to exhumed mantle domain that is characterized by synrift and postrift sedimentation from Albian to Coniacian–Santonian time (Fig. 7). Similar to the Mendibelza transect, these samples show a progressive tectonic unroofing sequence (Figs. 7 and 9). During Albian extension, Carboniferous–Permian (ca. 300 Ma) detrital zircons dominated the DZ signature of the NMB, but appear to be absent in the intervening SMB. Unless early synrift sediment routing completely bypassed the SMB, it appears that ca. 300 Ma Variscan-aged detrital zircons from Mendibelza and the NMB were derived from different sources in Albian time (Figs. 7 and 9). Consequently, this requires an additional, non–Axial Zone–derived ca. 300 Ma DZ source in the NMB. During early hyperextension, NMB strata were deposited in immediate proximity of fault-bound lower crustal granulites of the Labourd Massif; it is therefore likely that ca. 300 Ma detrital zircons from the NMB were derived from these granulites as they were exhumed during Albian to Cenomanian time. The MDS analysis of the samples from the NMB (Fig. 10) shows that the stratigraphically lowest sample (13WPY04) plots between the bedrock and granulites trends and then clearly evolves into the granulite field (12WPY54), indicating a potential unroofing sequence and that sample 12WPY54 is likely derived directly from granulites. This trend is significantly different than Albian–Cenomanian samples from Mendibelza (12WPY24–12WPY28) that are generally clustered on the MDS plot (Fig. 10). While 13WPY04 is not as clearly differentiated from the Mendibelza samples on the MDS plot as 12WPY54, the rim KDE plot (Fig. 11) clearly indicates that 13WPY04 lacks the multimodal rim age distribution that is present in Albian to Cenomanian Mendibelza samples, indicating that 13WPY04 and the Mendibelza samples were likely derived from a different source. Instead, 13WPY04 shows rim and core distributions that are more similar to the rim and core distributions from the Labourd Massif granulites. While trace element analyses could be used in the future to further characterize the similarities and differences between sedimentary strata at Mendibelza, SMB, NMB, and granulites, the differences illustrated through MDS and core and rim age analyses strongly suggest that the ca. 300 Ma DZ age component in the NMB is significantly different from Mendibelza, and was likely derived from the granulitic basement of the underlying distal rift margin rather than from sediment bypass and Variscan plutons in the Axial domain.

In the NMB, the abundance of the granulite-derived zircons dramatically decreased from a maximum of 70% to 15% in Cenomanian to early Turonian synrift to postrift strata (Figs. 7 and 9). The MDS plot also indicated that by Cenomanian time the subbasins became increasingly integrated, because 12WPY57 and 12WPY59 from the NMB are not distinguishable from Mendibelza samples 12WPY24–12WPY28, and therefore have identical provenance (Fig. 10). These samples evolve in a linear fashion that represents a mixing line between Paleozoic strata and ca. 300 Ma Variscan metamorphosed strata and granulite plutons. As rifting progressed, fault-bound granulites of the distal hyperextended margin were buried and no longer contributed detritus to fault-controlled subbasins, which became increasingly sourced from Ordovician, Devonian, and/or Triassic units that are characterized by Cryogenian–Ediacaran detrital zircons (Figs. 5 and 7). As the basin continued to fill, the input from the Axial domain increased and the Mendibelza samples (12WPY21 and 12WPY23) and the NMB samples (12WPY58 and 12WPY49) evolved similarly, again indicating similar provenance (Fig. 10). In the NMB, the Carboniferous–Permian DZ age component increased again to ∼30% in late Turonian time, indicating that sediments may have been increasingly regionally sourced from the Axial domain Variscan plutons in the hinterland (Figs. 7 and 9).

The stratigraphically highest samples from the proximal Mendibelza (late Cenomanian) and the distal NMB (Coniacian–Santonian) exhibited a marked decrease in ca. 300 Ma detrital zircons, by 7% and 25%, respectively (Fig. 7 and 9). This abrupt change appears to signal a diachronous shift in postrift provenance across the hyperextended rift margin and is characterized by decreased Carboniferous–Permian detrital zircons and increased contribution of Paleozoic to Mesozoic prerift-related detrital zircons from Ordovician, Devonian, and/or Triassic strata (Figs. 5 and 7). This marked shift in postrift provenance could potentially be explained by the formation of salt-walled basins during Cenomanian time, such as the Chaînons bérnais (e.g., Canérot, 1989; Jammes et al., 2009). Alternatively, late Santonian inversion tectonics could have triggered this provenance switch due to promotion of axial drainage along the northern Pyrenees (e.g., Rosenbaum et al., 2002; Vissers and Meijer, 2012; Metcalf et al., 2009; López-Mir et al., 2014). Whitchurch et al. (2011) pointed to regional orogen-parallel, axial sediment routing from east to west during initial orogenesis in the south-central Pyrenees from Late Cretaceous to Paleogene time. If the western Pyrenees were affected by this initial orogenesis, it is possible that the Mauléon Basin area could have been affected by similar orogen-parallel east-west sediment routing, providing sediments potentially recycled from early Paleozoic strata and deposited into Late Cretaceous strata.

These new constraints on the stratigraphic and sedimentary provenance evolution of the Mauléon Basin can be integrated with models for the development of magma-poor hyperextended continental margin structural domains to reconstruct the evolution and architecture of the region. The DZ U-Pb signatures of Paleozoic and Mesozoic strata deposited prior to Cretaceous extension documented continual recycling of metasedimentary rocks of Gondwanan affinity. While the timing of diffuse and protracted extension is difficult to determine, the DZ U-Pb signatures indicated a dominance of regional sourcing through Triassic time without any major provenance shifts or significant fault-controlled local sourcing. During strain localization and the progressive onset of crustal necking, sedimentary provenance underwent an abrupt transition to locally derived sediment within individual structurally controlled subbasins and an upsection transition back to regional hinterland-derived sedimentation in the proximal necking domain. In the Mauléon Basin this is apparent in the proximal necking domain of the Mendibelza area, where Albian early synrift sediments were derived locally from Paleozoic to early Mesozoic strata, with increasing input of regionally sourced sediments derived from Variscan plutons of the Axial domain of the western Pyrenean hinterland (Fig. 12, gray arrow). As strain continued to localize and the necking and hyperthinned domains developed, sedimentation in fault-controlled subbasins was locally sourced from adjacent exhumed footwall units and appeared to be isolated from hinterland-derived sediment supply. This effect manifested in the distal necking domain in the SMB in the sourcing of Albian units from metamorphosed Paleozoic units from the immediate footwall of the SMD and, in the hyperthinned domain of the NMB, by local sourcing of Albian to Cenomanian strata from exhumed granulite facies lower crust (Fig. 12, white arrows). As rifting progressed, these distal, isolated subbasins were filled by progressively more regionally sourced synrift and postrift sediments until these subbasins became reintegrated and hinterland-derived sediments spilled into the distal necking domain, and eventually the hyperthinned and exhumed mantle domains. This fill-and-spill integration of subbasins occurred throughout the diachronous evolution of the Mendibelza area and the NMB as illustrated by Albian to Cenomanian deposits in the Mendibelza area that were isolated and did not reach the distal hyperextended domain of the NMB until Cenomanian to Turonian time (Fig. 12, black arrows). On the basis of provenance data, it is not possible to differentiate late synrift from postrift strata, because the transition was gradual and not abrupt. During late synrift to postrift sedimentation, provenance returned to a regionally sourced, hinterland-dominated sediment supply until the region became influenced by subsequent events.

The generalities of this sediment dispersal model from the western Pyrenees give insight into observations that are transferable to other hyperextended basins. The model indicates that during the early stages of rifting, while extension is diffuse, there is little to no change in provenance, but once crustal necking begins an abrupt change occurs. At this stage of rifting, the proximal and distal parts of the margin are isolated from hinterland sedimentary supply and become locally sourced. During the progression from synrift to postrift, the proximal margin has an increase in hinterland sedimentary supply, and as the subbasins fill-and-spill they become reintegrated as the margin transitions to a passive continental margin (Nottvedt et al., 1995; Anderson et al., 2000; Mohn et al., 2010; Masini et al., 2013).

The Mauléon Basin offers a unique opportunity to apply DZ U-Pb provenance techniques to an exhumed, fossil, magma-poor, hyperextended margin to gain insight into the interplay between tectonics and sedimentation during the progression of rifting. New bedrock U-Pb age data from lower crustal granulites from the Labourd Massif in the western Pyrenees indicate that the orthogneissic granulite plutons crystallized at 279 ± 2 Ma and 274 ± 2 Ma during late Variscan magmatism. Paragneissic granulites yielded zircon rim ages of ca. 295 Ma that indicated widespread Variscan granulite facies metamorphism of the Labourd Massif. Inherited zircon core ages from these paragneissic granulites point to a Cambrian or Ordovician aged metasedimentary protolith.

DZ U-Pb signatures of the western Pyrenees prerift Paleozoic and Mesozoic strata indicate a Gondwanan signature, typical for Precambrian to Paleozoic strata from Iberia, that was likely sourced from the East Africa orogen, northern Egypt, and the Sinai Peninsula of Gondwana. The western Pyrenean Paleozoic and Mesozoic prerift strata all exhibit common age components with modes of ca. 615 Ma, ca. 780 Ma, and ca. 1000 Ma, suggesting either continual recycling and/or a supply of well-mixed Gondwanan-sourced zircons throughout the Paleozoic and early Mesozoic. Ordovician, Carboniferous, and Permian strata also contain major diagnostic DZ U-Pb age components that differentiate their unit signatures, with age modes of ca. 300 Ma in Carboniferous and Permian strata and ca. 480 Ma in Ordovician and Carboniferous strata, which are interpreted to be related to Variscan and Cadomian magmatism, respectively.

DZ U-Pb analyses have confirmed a diachronous multibasin rifting evolution of the Mauléon-Arzacq rift system in the western Pyrenees, and provided a general model for sediment dispersal processes at hyperextended continental margins. DZ provenance analyses of synrift and postrift sedimentary rocks from the Mauléon Basin indicate that during the early stages of diffuse extension, there was little to no change in provenance. However, an abrupt change occurred once crustal necking began, as the proximal and distal parts of the margin were isolated from the hinterland sedimentary supply and became locally sourced. During the late synrift to postrift time, the proximal margin had an increase in hinterland sedimentary supply, and as the subbasins filled and spilled, they became reintegrated and the hinterland sediments reached the distal hyperthinned and exhumed mantle domains. At that point, sedimentation returned to a regionally sourced hinterland drainage system until the region became influenced by subsequent events. In addition, tracking the presence of the ca. 300 Ma age component spatially and temporally across the basin in synrift and postrift strata showed that the lower crustal granulites were exhumed during Albian to Cenomanian time. This is a key constraint in defining the activity along the NMD, as the detachment must have been active and the hyperthinned domain must have been formed prior to these granulites being exhumed.

This project was financially supported by American Association of Petroleum Geologists Foundation Arthur A. Meyerhoff Memorial Grants-In-Aid, a Jackson School of Geosciences off-campus research award, and Petrobras. We thank Emily Cooperdock, Gavin Wagoner, and Jacqueline Reber for their assistance in the field, and Lisa Stockli and Spencer Seman for their assistance with data acquisition, data reduction, and interpretation. We also acknowledge invaluable insights and discussions with Luc Lavier, Edgardo Pujols-Vazquez, Anna Eliza Svartman Dias, Gianreto Manatschal, Victor Hugo, Emmanuel Masini, Suzon Jammes, and Antonio Teixell. We thank one anonymous reviewer and M.E. Bickford and F. Mouthereau, editor S. de Silva, and associate editor Derek Keir for their thorough and thoughtful comments that clarified and improved the manuscript.