New age constraints on the break-up of Rodinia and amalgamation of southwestern Gondwana from the Choquequirao Formation in southwestern Peru

Supplementary material: Detailed U–Pb LA-ICP-MS and CA-ID-TIMS methods, and zircon, titanite and rutile LA-ICP-MS U–Pb, CA-ID-TIMS U–Pb, ID-TIMS U–Pb and trace element data are available at https://doi.org/10.6084/m9.figshare.c.6266972

Precambrian basement inliers from the Central Andes are scarce, limiting opportunities to gain insights into the tectonic history of the underlying terranes. With so few constraints, it is uncertain where the boundary is located between the Arequipa Terrane and the easternmost extent of the Amazon Craton (Cárdenas et al. 1997; Chew et al. 2008; Mišković et al. 2009; Reimann Zumsprekel et al. 2015; Hodgin et al. 2021a). The poorly studied Choquequirao Formation, which is an extensive amphibolite-grade igneous and metasedimentary succession exposed in the Eastern Cordillera of southern Peru (Fig. 1), can thus shed light on the boundary between the Arequipa Terrane and the Amazon Craton, as well as the timing of their amalgamation. In addition, the Choquequirao Formation can also provide an insight into earlier plate tectonic processes, such as rifting. The age, provenance and metamorphic history of the Choquequirao Formation can also provide insights into the tectonic history of supercontinent break-up and reassembly during the Neoproterozoic. Most palaeogeographical models suggest that the break-up of Rodinia (McMenamin and McMenamin 2001) occurred primarily during the Tonian–Ediacaran (Li et al. 2008; Merdith et al. 2017a, b; Zhao et al. 2018), whereas other models, suggest that break-up was not completed until the early–middle Cambrian (Karlstrom et al. 2018; Busch et al. 2021). Most palaeogeographical models invoke a multistage break-up of Rodinia (Merdith et al. 2017a; Robert et al. 2020; Evans 2021), yet geological evidence for multistage break-up from critical continental margins is challenging to identify (e.g. Thomas 1991; Eyster et al. 2018; Hodgin et al. 2022). In particular, ribbon terranes that formed by two rifting events (Lister et al. 1986; Péron-Pinvidic and Manatschal 2010) represent a unique opportunity to evaluate missing links within palaeogeographical models (e.g. Dalziel 1993, 1994; Li et al. 1995).

The Arequipa Terrane in southern Peru is commonly invoked as the conjugate rifted margin of eastern to southeastern Laurentia (e.g. Escayola et al. 2011; Casquet et al. 2012; van Staal et al. 2013; Ramacciotti et al. 2015; Rapela et al. 2016; Robert et al. 2020; Evans 2021); yet there is still a lack of clear geological evidence for rifting associated with a multistage break-up of Rodinia. Geological evidence related to a hypothesized accretion of the Arequipa Terrane to the Amazon Craton during the assembly of the supercontinent Gondwana is also sparse (Chew et al. 2008; Reimann Zumsprekel et al. 2015; Hodgin et al. 2021a). As a result, many tectonic models have suggested that the Arequipa Terrane was amalgamated to Amazonia in the Mesoproterozoic (Loewy et al. 2004; Chew et al. 2007a, b; Ramos 2008; Martin et al. 2020), implying that there may not be a Neoproterozoic Wilson Cycle preserved on the eastern margin of the Arequipa Terrane related to Rodinia dispersal and Gondwana amalgamation. Due to a lack of definitive geological evidence, a wide range of tectonic and palaeogeographical models have been put forward for the Arequipa Terrane that can be tested by investigating the age and provenance of the Choquequirao Formation.

The Choquequirao Formation crops out in the Eastern Cordillera of the Central Andes in southern Peru (Fig. 1). It occurs within the Machu Picchu Inlier, which is itself part of the Abancay Deflection structural zone (Fig. 1) (Marocco 1978; Carlotto 2002; Carlotto et al. 2009). The name of the succession is derived from the important Inca archaeological site, Choquequirao, located centrally within the map area and whose metamorphic rocks were used to build the site (Carlotto et al. 2011). Only a handful of geological studies have been carried out within the remote, mountainous region containing the succession. The earliest studies by Heim (1948), Egeler and De Booy (1957, 1961) and Fricker and Weibel (1960) attributed the garnet-, sillimanite- and staurolite-bearing amphibolite-grade metamorphic rocks of the Choquequirao Formation to the Precambrian, whereas subsequent studies have generally assigned the succession to the early Paleozoic (Marocco 1978; Cárdenas et al. 1997; Carlotto et al. 1999, 2011).

The internal stratigraphy of the estimated 3.5 km-thick Choquequirao Formation can be divided into five mappable stratigraphic units: (1) gneiss, amphibolite and diamictite; (2) massive quartzite and marble; (3) quartzo-feldspathic paragneiss; (4) micaceous schist; and (5) quartzite and micaceous schist with marble (Cárdenas et al. 1997; Carlotto et al. 1999). The lowermost stratigraphic unit consists of gneiss, amphibolite and diamictite occurring within the cores of east–west-orientated anticlines and in the hanging walls of major thrust faults (Figs 2 & 3) (Marocco 1978; Cárdenas et al. 1997; Carlotto et al. 1999, 2011). The gneiss within the basal unit has been divided into orthogneiss and paragneiss lithologies, and it is associated with up to c. 1000 m of amphibolite, which itself may be derived from a combination of sedimentary and igneous protoliths (Cárdenas et al. 1997). The age of the basal gneiss is unknown but basal orthogneiss has been tentatively correlated with dated inliers at Río Pichari, c. 200 km to the NW, and at Amparaes Dome, c. 100 km to the NE (Fig. 1). A crystalline basement sample of charnockite from the Río Pichari Inlier, similar to the orthogneiss lithologies in the basal Choquequirao Formation (Cárdenas et al. 1997), yielded an upper intercept age of 1140 ± 30 Ma (Dalmayrac et al. 1977, 1988). Hornblende-rich orthogneiss in the basal unit has also been lithologically correlated to hornblende granite at Amparaes Dome (Fig. 1) (Cárdenas et al. 1997), which has been dated subsequently at 479 ± 2.3 Ma (Reitsma 2012). However, the presence of contact metamorphism, as indicated by garbenschiefer in the metasedimentary rocks surrounding the granite at Amparaes Dome, suggests an intrusive relationship. Granites of similar age have now also been found to intrude metasedimentary rocks of the Choquequirao Formation (Reitsma 2012; INGEMMET 2020). As a result, the basal stratigraphic unit of the Choquequirao Formation is conservatively considered to consist primarily of amphibolite and associated paragneiss. Within the basal stratigraphic unit, we also report diamictite, which is particularly well exposed in the core of an anticline at Rio Aobamba (Figs 2 & 3). The diamictite, which consists primarily of granitoid clasts similar to the Cryogenian Chiquerío Formation in coastal southwestern Peru (Chew et al. 2007b), is likely to sit stratigraphically between the amphibolite of unit 1 and the overlying marble of unit 2.

The second lithostratigraphic map unit is estimated to be 500 m thick, and consists predominantly of massive quartzite and marble (Fig. 3) (Carlotto et al. 1999). The marble intervals generally occur near the base and top of the unit. The basal marble, which is best exposed in the core of an anticline between Nevado Padreyoc and Nevado Salkantay (Fig. 2), reveals that the estimated 100 m carbonate interval (Carlotto et al. 2009) should be considered a minimum thickness.

The overlying lithostratigraphic unit consists of an estimated 700 m of predominantly quartzo-feldspathic paragneiss. Upsection, the fourth stratigraphic unit consists of an estimated 700 m of micaceous schist. The fifth and uppermost lithostratigraphic unit consists of an estimated 500 m of quartzite and micaceous schist containing interbedded marble. There are no direct age constraints on any of the stratigraphic units.

Overlying the Choquequirao Formation in angular unconformity (Carlotto et al. 1999) is a generally less metamorphosed, fossil-bearing succession of slates, quartzites and conglomerate belonging to the lower Ordovician Verónica and San José formations (Egeler and De Booy 1961; Carlotto et al. 1999). There are substantial thickness and facies changes in the conglomeratic Verónica Formation (Egeler and De Booy 1961; Carlotto et al. 1996), which consists of quartzite clasts that are likely to have been derived from the underlying Choquequirao Formation, and which also implies an intervening metamorphic event. The Verónica Formation grades up into the overlying San José Formation (Egeler and De Booy 1961; Hodgin et al. 2021a), which consists predominantly of fossiliferous slate and phyllite (Fig. 3). The oldest biostratigraphic age constraints from strata overlying the Choquequirao Formation come from c. 479–478 Ma graptolites in the lower San José Formation (Gutiérrez-Marco et al. 2019). Deposition of the Choquequirao Formation is thus considered to predate the c. 480–445 Ma Famatinian Orogeny (Pankhurst et al. 1998).

Given the age constraints, high metamorphic grade and lithology of the Choquequirao Formation, the succession has been tentatively correlated with other pre-Ordovician inliers in the Eastern Cordillera (Cárdenas et al. 1997; Carlotto et al. 1999, 2011; Hodgin et al. 2021a). In southern Peru, the Choquequirao Formation may be equivalent to the undated Iscaybamba Complex, which consists of orthogneiss, amphibolite, andesite, quartzite and schist (Fig. 1) (Laubacher 1978; Palacios et al. 1996; Chávez et al. 1997; Sánchez and Zapata 2003; Carlotto et al. 2009); to the Amparaes Dome, which consists of amphibolites, quartzites, sandstones, schists and marbles; to the upper Cambrian Llallahue Formation, which consists of quartzite and arkosic sandstone (Hodgin et al. 2021a); or to the Neoproterozoic Chiquerío and San Juan formations overlying the Arequipa basement in coastal southwestern Peru (Chew et al. 2007b). In central and northern Peru, the Choquequirao Formation could be equivalent to high-grade metasedimentary rocks of the Marañón and Huaytapallana complexes that were intruded and deformed at c. 480 Ma (Fig. 1) (Chew et al. 2007a, 2016; Cardona et al. 2009).

Zircon crystals were separated from geochronological samples using standard mineral separation procedures, then imaged with a cathodoluminescence (CL) detector and dated by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at Boise State University (BSU) (see Supplementary Section S1 for details of the U–Pb zircon LA-ICP-MS geochronology methods). Metamorphic zircons were identified based on CL zonation, the Th/U ratio and Ti-in-Zr thermometry (Ferry and Watson 2007; Rubatto 2017). Zircon domains with Th/U values of less than 0.1 were generally identified as metamorphic (Rubatto 2017). Due to the prevalence of Pb loss in zircons from these high-grade and poly-deformed rocks, a younger cutoff of 900 Ma was used in reporting ²⁰⁶Pb/²³⁸U v. ²⁰⁷Pb/²⁰⁶Pb dates (e.g. Hodgin et al. 2022), and a concordance filter was set at −10 to +15%. The tectonomagmatic environment of the melt from which specific age populations of zircons crystallized was assessed using LA-ICP-MS zircon trace element indicators, such as U/Yb v. Nb/Yb (Grimes et al. 2015). The youngest detrital zircons in each sample identified by LA-ICP-MS were dated more precisely by chemical abrasion isotope dilution thermal ionization mass spectrometry (CA-ID-TIMS) at BSU (see Supplementary Section S1for details of the U–Pb zircon geochronology methods). Due to the prevalence of Pb loss in these high-grade poly-deformed rocks, most zircon crystals dated by CA-ID-TIMS were broken into multiple fragments in order to determine precisely the upper intercepts corresponding to the crystallization age and lower intercepts corresponding to the timing of Pb loss during peak metamorphism. Metamorphic rutile and titanite crystals were also analysed by LA-ICP-MS from sample B1425. A selection of titanite crystals were subsequently dated by ID-TIMS.

Comparative analysis of detrital zircon spectra was carried out using the IsoplotR software package (Vermeesch 2018). In addition to generating comparative kernel density (KDE) and cumulative age distribution (CAD) plots, datasets were quantitatively compared using the Kolmogorov–Smirnov (K–S) statistical difference. This method takes the maximum vertical distance between two CADs, and multiple values of which can then be plotted using multidimensional scaling (Vermeesch 2013). Age spectra from the Choquequirao Formation detrital zircon samples (this study) were first compared to compilations of detrital zircon data representative of the Arequipa Terrane in coastal southern Peru (Chew et al. 2007b), from the Altiplano of southern Peru (this study) and from the southwestern rifted margin of the Amazon Craton (Babinski et al. 2013; Harris 2020; Harris et al. 2023). Comparative analysis also included samples from high-grade metasedimentary rocks of the Huaytapallana and Marañón complexes in the Eastern Cordillera of central and northern Peru (Chew et al. 2007a, b, 2008, 2016; Cardona et al. 2009), which were divided into samples having either a pre-Pampean rift-related affinity or a post-Pampean orogenic affinity following the detrital zircon tectonostratigraphic methodology developed by Cawood et al. (2012). Samples from the Huaytapallana and Marañón complexes were included for comparative analysis if the youngest detrital zircon was within error of the Early Ordovician, the minimum age of the Choquequirao Formation. Samples were characterized as having an orogenic affinity if they were dominated by late Neoproterozoic–early Paleozoic age peaks and as having a rift-related affinity if the dominant age peaks instead consisted of older regional basement sources (e.g. >1 Ga). Southwestern Kalahari was included in our analysis as it has been identified as the most likely conjugate rifted margin to eastern Arequipa and related terranes (Rapela et al. 2016; Casquet et al. 2018). Specifically, we incorporated a compilation from Neoproterozoic rift-related detrital zircon of the Gariep Belt in southwestern Kalahari (Basei et al. 2005; Hofmann et al. 2015; Thomas et al. 2016) and from the Saldania Belt in southern Kalahari (Andersen et al. 2018). The same ²⁰⁶Pb/²³⁸U v. ²⁰⁷Pb/²⁰⁶Pb age cutoffs and concordance filtering described above were applied to all datasets.

One U–Pb igneous sample (B1429) and four U–Pb detrital zircon samples were analysed from the Choquequirao Formation. The four detrital zircon samples are described in stratigraphic order, followed by a detrital zircon sample (B1797) from the Llallahue Formation in the Altiplano of southern Peru, which was included for comparative analysis (Fig. 3b).

B1429 is a fine-grained biotite-rich quartzo-feldspathic gneiss within an amphibolite map unit near Santa Teresa (Fig. 2: 13.1417° S, 72.5954° W). Millimetre-scale foliation is defined by biotite and the sample displays deformation-induced myrmekite structures, which are typically associated with amphibolite-grade metamorphism in K-feldspar-bearing granitoids (Simpson and Wintsch 1989; Menegon et al. 2006). Abundant angular oxides and interstitial grains of potassium feldspar and quartz further support our interpretation of an igneous protolith. The sample yielded four very small zircons, only two of which could be analysed by LA-ICP-MS. The ²⁰⁷Pb/²⁰⁶Pb dates were c. 1900 and c. 1100 Ma, which are consistent with Precambrian detrital zircon dates recovered from the other samples in this study, and may be explained as inherited or entrained grains within a zircon-poor protolith.

B1428 is a sample of coarse-grained quartzite that comes from the basal quartzite map unit collected near the core of an anticline near Abra San Juan on the hiking trail from Maizal to Yanama (Figs 2 & 3: 13.3329° S, 72.8658° W). One hundred and eight analyses yielded a broad dominant population ranging from 1500 to 950 Ma with a primary peak at c. 1200 Ma, and secondary peaks at c. 1450 and c. 1000 Ma (see Supplementary Table S1). Minor populations occur at c. 2250, c. 800 and c. 600 Ma. The youngest detrital zircon was split into two fragments, which were each dated by CA-ID-TIMS. The analyses overlap with concordia but they do not overlap with each other and are interpreted as slightly discordant. A weighted mean ²⁰⁷Pb/²⁰⁶Pb date is 635.1 ± 1.4 Ma. Assuming a lower intercept of 460 Ma yields an upper intercept date of 638.7 ± 4.2 Ma (Fig. 4; see Supplementary Table S2).

B1427 is a sample of coarse-grained leucosome-bearing quartzo-feldspathic schistose gneiss collected from the basal quartzite of unit 2 along the trail from Maizal to Abra San Juan at 13.3434° S, 72.8817° W (Figs 2 & 3). Ninety-two analyses yielded a broad, dominant population ranging from 1500 to 900 Ma with peaks at c. 1450, c. 1200, c. 1100 and c. 950 Ma (see Supplementary Table S1). Notable minor populations occur at c. 600 and c. 2700 Ma. The youngest detrital zircon was split into two fragments, which were each dated by CA-ID-TIMS. The analyses do not overlap and are each slightly discordant. A weighted mean ²⁰⁷Pb/²⁰⁶Pb date is 627.9 ± 1.8 Ma. Assuming a lower intercept of 460 Ma, an upper intercept date is 639.5 ± 5.6 Ma (Fig. 4; see Supplementary Table S2).

B1505 is a sample of micaceous quartzo-feldspathic gneiss collected from the paragneiss map unit that overlies the basal quartzite map unit (Figs 2 & 3) (Carlotto et al. 1999). It was collected at Quebrada Victoria, also referred to as Río Blanco, at 13.3636° S, 72.8840° W. One analysis on a metamorphic zircon yielded a Th/U ratio of 0.01 and a date of 458.1 ± 23.2 Ma. The remaining 110 detrital zircon analyses yielded a broad dominant population ranging from 1500 to 900 Ma with a primary age peak at c. 1100 Ma, and secondary age peaks at c. 1450, c. 1300 and c. 950 Ma (see Supplementary Table S1). Minor populations occur at c. 1800, c. 1550, c. 850, c. 700 and c. 600 Ma. The youngest detrital zircon was split into three fragments dated by CA-ID-TIMS, resulting in discordant analyses that did not agree with each other. A weighted mean ²⁰⁷Pb/²⁰⁶Pb date is 635.1 ± 1.4 Ma. The three analyses resulted in an unanchored lower intercept date of 458.7 ± 13.8 Ma and an upper intercept date of 636.1 ± 8.6 Ma (mean squared weighted deviation (MSWD) = 1.2; probability of fit (POF) = 0.27). Assuming a lower intercept of 460 Ma yielded a similar upper intercept date of 636.8 ± 4.2 Ma (Fig. 4; see Supplementary Table S2).

B1425 is a sample of thin-bedded semi-pelitic quartzo-feldspathic quartzite collected near the 29 km marker along the trail from Marampata to the Choquequirao archeological site (Figs 2 & 3: 13.4011° S, 72.8573° W). Zircons from this sample were predominantly small and complex. Two metamorphic zircon domains with Th/U ratios of 0.07 and 0.08 yielded dates of 542 ± 41 and 496 ± 45 Ma, respectively. The analysed detrital zircon yielded a broad, dominant age population ranging from c. 1350 to c. 1100 Ma with a peak at c. 1250 Ma (see Supplementary Table S1). Secondary age populations occur at 2050–1700 Ma, c. 1450, c. 1000 and 900–700 Ma. Three of the youngest detrital zircon grains were dated by CA-ID-TIMS. Two of the grains (z1 and z2) fall on a discordia line with a lower intercept of 467.9 ± 17.6 Ma that is consistent with the age of Pb loss resolved by CA-ID-TIMS analyses in other samples, lending confidence to these two grains potentially sharing a common crystallization age defined by an upper intercept of 791.0 ± 4.5 Ma (Fig. 4; Supplementary Table S2). The remaining grain (z3) is located farther from concordia, and is likely to fall on a discordia line between peak metamorphism at c. 460 Ma and an early Neoproterozoic upper intercept age.

B1797 is a sample of arkosic sandstone from the upper Cambrian Llallahue Formation in the Altiplano of southern Peru (Figs 1,, 2,, 3) (Hodgin et al. 2021a). Due to its biostratigraphically well-constrained depositional age and its inferred tectonostratigraphic setting within an upper Cambrian back-arc basin (Hodgin et al. 2021a), this sample was included for comparison with detrital zircon samples from the Choquequirao Formation. LA-ICP-MS analyses resulted in 14 discordant dates and 76 concordant dates, of which the latter are included in our provenance analysis. The sample yielded a broad dominant population ranging from 1450 to 850 Ma with prominent age peaks at c. 1100 and c. 900 Ma (see Supplementary Table S1). Minor populations occur at c. 1650, 750–700 and c. 500 Ma. The three youngest detrital zircons from the c. 500 Ma population were dated by CA-ID-TIMS yielding ²⁰⁶Pb/²³⁸U dates of 496.75 ± 0.82, 500.11 ± 1.00 and 506.99 ± 3.01 Ma (see Supplementary Table S2). The youngest of these detrital zircon analyses from the Llallahue Formation (c. 497 Ma) constrains its maximum depositional age and is consistent with the trilobite fossils extracted from the same sample, which are early Furongian (Hodgin et al. 2021a), given that the base of the Furongian Epoch has a revised maximum depositional age of c. 494 Ma (Cothren et al. 2022). These strata were thus deposited during approximately coeval magmatism.

Sample B1425 (13.4011° S, 72.8573° W) contained lenticular-shaped titanite crystals characteristic of metamorphic titanite (Essex and Gromet 2000; Hodgin et al. 2021b). Twelve titanite crystals were analysed by LA-ICP-MS (see Supplementary Table S3), and nine of these titanite crystals, which were split into 14 total titanite mineral fractions, were subsequently analysed by ID-TIMS (Fig. 5; Supplementary Table S4). A concordia-constrained 3D isochron using all analyses resulted in a well-resolved initial ²⁰⁷Pb/²⁰⁶Pb isotopic composition of 0.85219 and a lower-intercept age of 447.08 ± 0.32 Ma, approximately corresponding to the timing of peak metamorphism. We note that there is significant scatter in the ID-TIMS data, whose ²⁰⁶Pb/²³⁸U dates range from 465 to 430 Ma. The overdispersion and systematic discordance within the data suggest that the metamorphic titanite may record at least two Pb-loss pathways associated with earlier and later metamorphic events. Two apparent Pb-loss pathways displayed by the metamorphic titanite are both poorly constrained, yet the lower intercept overlaps with c. 325 Ma metamorphic rutile, which is likely to correspond to Hercynian orogenesis. A poorly constrained, upper intercept age of 508 ± 70 Ma overlaps with c. 540 and c. 490 Ma metamorphic zircon (Fig. 5a), and may be related to Pampean and/or early Famatinian orogenesis.

Zr-in-titanite thermometry (Hayden et al. 2008; Kapp et al. 2009) on metamorphic titanite crystals resulted in temperatures of c. 700°C, which is consistent with amphibolite-grade peak metamorphism associated with leucosome generation and extensive migmatization in Choquequirao Formation metasedimentary rocks.

Sample B1425 (13.4011° S, 72.8573° W) contained a population of rounded to prismatic rutile crystals ranging in diameter from 120 to 350 µm and hosting a variety of 10–20 µm-scale inclusions, with the most abundant identified optically and by CL as apatite, titanite and quartz. Five spot analyses were conducted from the centres of homogeneous and unfractured crystals (Fig. 5), yielding common Pb-corrected ²⁰⁶Pb/²³⁸U LA-ICP-MS dates with a weighted mean of 327.4 ± 11.5 Ma (Fig. 5; Supplementary Table S5), approximately corresponding to the age of high-grade 325–310 Ma metamorphism along the western Gondwana margin preserved at other localities in the Eastern Cordillera of central and northern Peru (Chew et al. 2007a, 2016; Cardona et al. 2009).

From previous studies it is known that the Choquequirao Formation predates cross-cutting Early Ordovician intrusions (Reitsma 2012; INGEMMET 2020) and also predates Early Ordovician deposition of the Verónica and San Jose formations, which unconformably overlie it (Egeler and De Booy 1961; Carlotto et al. 2011; Hodgin et al. 2021a). Previous mapping and reconnaissance of the Choquequirao Formation led to hypotheses that the depositional age of the succession may be Precambrian (Heim 1948; Fricker and Weibel 1960; Egeler and De Booy 1961) or early Paleozoic in age (Marocco 1978; Cárdenas et al. 1997; Carlotto et al. 1999, 2011; Hodgin et al. 2021a). Recent studies on potentially equivalent high-grade sedimentary complexes in central and northern Peru have revealed the presence of poorly differentiated Precambrian, early Paleozoic and late Paleozoic successions (Chew et al. 2007a, 2008, 2016; Cardona et al. 2009). To address uncertainty in the age of the Choquequirao Formation, we developed new ages from detrital zircon samples to constrain its maximum age, and new metamorphic ages to better constrain its minimum age and tectonic history. In addition, zircon trace element and provenance analyses were used to interpret the tectonostratigraphic setting of the Choquequirao Formation, and the identity of the underlying crust. Finally, these data were synthesized within a larger tectonic context to address the timing of the Rodinia supercontinent break-up and subsequent Gondwana supercontinent amalgamation.

Prior to putting forward our interpretations of these high-grade, poly-deformed rocks, we lay out five plausible tectonostratigraphic settings of the Choquequirao Formation: (1) a >800 Ma basin that predates the break-up of Rodinia; (2) a c. 800–550 Ma rift basin that formed on the margin of eastern Amazonia or western Arequipa during the break-up of Rodinia; (3) a collisional basin associated with the c. 550–525 Ma Pampean Orogeny postulated to extend into the Eastern Cordillera of Peru (Aceñolaza and Toselli 2009; Escayola et al. 2011); (4) a 510–490 Ma back-arc basin that formed during a magmatic lull between the Pampean and Famatinian orogenies (Hodgin et al. 2021a); and (5) a c. 492–480 Ma extensional basin that formed during an early pulse of the Famatinian Orogeny (e.g. Astini 2008; Weinberg et al. 2018). Finally, we acknowledge that other ages and basin types could be present and that more than one tectonostratigraphic interval may be captured within the thick and poorly defined Choquequirao Formation.

All four U–Pb detrital zircon samples in this study yielded minor populations of Neoproterozoic in situ ages ranging from 750 to 550 Ma, which were the youngest detrital zircon analyses in the samples. Due to analytical imprecision of the in situ analyses and a lack of control for Pb loss, the youngest detrital zircons were subsequently dated by CA-ID-TIMS to more precisely and accurately constrain the maximum age of the Choquequirao Formation. Three of the four samples each contained a single zircon that was split into multiple fragments, which resulted in non-overlapping dates that displayed Pb loss along a discordia line, and yielded upper intercept ages overlapping in uncertainty between 640 and 635 Ma. By assuming that these detrital zircon grains could represent a single age population, the CA-ID-TIMS analyses were combined to yield an upper intercept of 638.7 ± 3.4 Ma and a lower intercept of 462.2 ± 8.1 Ma (2σ, MSWD = 0.59, POF = 0.71). Combined, the statistical metrics (MSWD and POF) indicate an increased probability of derivation from a single age population. The zircons also displayed consistent crystal size, crystal morphology, CL response and zonation, and trace element profiles (Fig. 4; Supplementary Table S1). For example, the U/Yb v. Nb/Yb tectonomagmatic fingerprinting proxy (Grimes et al. 2015) applied to all three zircons suggests a potentially common mantle-derived source (Fig. 6). On the basis of these factors, we interpret the upper intercept of 638.7 ± 3.4 Ma and the lower intercept of 462.2 ± 8.1 Ma as indicating that these grains from different samples may have shared a common igneous source and a common timing of peak metamorphism. The combined analyses and resulting intercept dates are thus interpreted as our most robust constraint on the maximum age of deposition and timing of peak metamorphism, respectively. The maximum age constraint of 638.7 ± 3.4 Ma pertaining to stratigraphic units 2 and 3 may be close to the age of deposition. This interpretation is supported by the presence of a glacial diamictite at Rio Aobamba in the basal stratigraphic unit that was likely to have been deposited during the c. 651–635 Ma Marinoan Snowball Earth glaciation (Hoffman et al. 1998; Nelson et al. 2020). An unconformity in the basal unit also means that the diamictite could have been deposited entirely or in part during the c. 717–660 Ma Sturtian Snowball Earth glaciation (Macdonald et al. 2010a; Rooney et al. 2015). From the combined geochronological and stratigraphic data, we infer that basal deposition of the Choquequirao Formation began during the 717–635 Ma Cryogenian Period. Based on the tectonomagmatic affinity of the c. 640 Ma detrital zircons, one potential source of mantle-derived detritus may be the basal amphibolite unit. However, geochemical characterization and direct dating of the amphibolites will be required to test this linkage and characterize the presence and timing of rift-related magmatism.

In addition to information provided regarding the maximum age of deposition, the provenance of the four detrital zircon samples can be used to investigate the tectonic setting in which the succession was deposited (e.g. Cawood et al. 2012). To that end, observed trends within the four detrital zircon samples from the Choquequirao Formation are summarized and then compared to reference datasets from known tectonic setting.

Age spectra from all four samples are very similar, with the greatest similarity observed between the three basal samples (units 2 and 3 of the Choquequirao Formation). The three basal samples contained a broad, dominant age population ranging from 1500 to 900 Ma with peaks at c. 1450, c. 1200, c. 1100 and c. 950 Ma, and minor populations at c. 2700, c. 2250, c. 1800, c. 1550, c. 850, c. 700 and c. 600 Ma. The uppermost sample from unit 4 (B1425) had a somewhat more restricted dominant age population from c. 1350 to c. 1100 Ma with a well-defined peak at c. 1250 Ma, and secondary age populations at 2050–1700, c. 1450, c. 1000 and 900–700 Ma. The upper sample can be differentiated by having a more prominent 2050–1700 Ma age population, reduced abundance of c. 1100–950 Ma ages and an absence of the youngest c. 600 Ma population. We note that sample B1425 was finer grained than the three basal samples, which could have contributed to a minor redistribution of age population density according to grain size and transport distance within an overall common provenance (e.g. Leary et al. 2020). While there are minor differences, all four samples are interpreted as having a coherent provenance related to a strong likelihood of sharing a common crustal affinity and tectonic setting.

The age spectra of the four detrital zircon samples from the Choquequirao Formation were combined and compared with reference datasets of known tectonic setting (e.g. Cawood et al. 2012). Such a comparison indicates that the Choquequirao Formation follows detrital zircon age patterns associated with specific tectonic settings, and was thus likely to have been deposited in an extensional setting such as a rift basin or a passive margin, although an intraplate or back-arc setting cannot be ruled out. Due to the presence, but low percentage, of young detrital zircons in the three basal samples, a syndepositional magmatic source with low zircon fertility, as predicted in a rift setting, is most consistent with a rift basin (Cawood et al. 2012). Again, we note that this could also be consistent with a back-arc setting. In contrast, the uppermost sample, whose youngest detrital zircon is at least 150 myrs older than the time of deposition, may have more affinity with a passive margin setting or an intraplate setting. Due to the association of rift-related provenance signatures from the underlying strata, a passive margin setting is most probable, even though a back-arc setting is possible. The tectonic setting inferences are further supported by the lithostratigraphy of the Choquequirao Formation. The presence of basal amphibolites potentially related to rift-related volcanism and the apparent facies changes in the lowest stratigraphic units may be consistent with a rift–drift transition near the base of the Choquequirao Formation, possibly in unit 2. We also note that the age spectra of the Choquequirao Formation contrast with collisional to convergent basin settings, including back-arc settings, that are found regionally in slightly younger successions (Chew et al. 2007a, 2008, 2016; Cardona et al. 2009; Reimann Zumsprekel et al. 2015; Hodgin et al. 2021a). An enigmatic c. 640 Ma back-arc setting has been described by Escayola et al. (2007) in a potentially correlative tectonic position adjacent to the eastern margin of the Pampia Terrane in northwestern Argentina. We note the absence of late Neoproterozoic orogenic detritus in the Choquequirao Formation, which contrasts with abundant Neoproterozoic detritus in the succession described by Escayola et al. (2007). Thus, we interpret the provenance of the Choquequirao Formation as rift-related and predating latest Neoproterozoic–early Paleozoic orogenesis and associated basin development along the Gondwanan margin (Aceñolaza and Toselli 2009; Escayola et al. 2011). The uppermost unit of the Choquequirao Formation was not sampled, and it may represent a transition to another tectonostratigraphic setting that post-dates rift-related deposition.

The interpretation of the tectonic setting of the Choquequirao Formation as a rift to passive margin succession does not necessarily clarify the identity of the underlying crust. This is due to the uncertainty of the underlying crustal affinity in this region. The boundary between the easternmost extent of the Amazon Craton and the westernmost extent of the Arequipa Terrane in the Eastern Cordillera of Peru has been extensively debated (Cárdenas et al. 1997; Chew et al. 2007a, b, 2008; Mišković et al. 2009; Reimann et al. 2010; Reimann Zumsprekel et al. 2015; Hodgin et al. 2021a). By revealing age signatures that are potentially characteristic of basement sources from southeastern Amazonia or Arequipa, the provenance of the Choquequirao Formation can be used to determine the tectonic affinity of the underlying crust upon which the succession was deposited. One way to conduct such a test is to compare detrital zircon ages from the Choquequirao Formation to reference detrital zircon datasets derived from successions overlying basement of southwestern Amazonia and western Arequipa in coastal southern Peru. We also compare detrital zircon age spectra from the Choquequirao Formation to reference datasets from other successions in Peru.

As a result of our statistical analysis implementing the K–S distance between different detrital zircon datasets, the Choquequirao Formation is most similar to detrital zircon age spectra that are representative of the Arequipa Terrane in coastal southwestern Peru (Fig. 7) (Chew et al. 2007b). In particular, both detrital zircon compilations have a dominant population from 1300 to 950 Ma, minor 2000–1600 and 800–600 Ma populations, and a notable absence of detrital zircons from c. 1600 to 1450 Ma. In contrast, the Choquequirao Formation is least similar to age spectra from the compilations from southwestern Amazonia (Babinski et al. 2013; Harris 2020; Harris et al. 2023). The Amazonian compilations tend to have a greater abundance of >2000 Ma ages, a much more dominant 2000–1600 Ma population, significant 1600–1450 Ma populations that are absent from Arequipa, and subdued to minor Grenvillian age populations that tend to be the most dominant from Arequipa. The Choquequirao Formation also displays a high degree of similarity with the upper Cambrian Llallahue Formation, which has been interpreted as having been deposited in a back-arc setting on Arequipa crust in the Altiplano of southern Peru (Hodgin et al. 2021a).

An additional linkage to an Arequipa provenance comes from the presence of 800–750 Ma mantle-derived zircons from the Choquequirao Formation (Fig. 6) and other early Paleozoic successions deposited on Arequipa (Reimann et al. 2010; Hodgin et al. 2021a). The common occurrence of mantle-derived 800–750 Ma detrital zircons on the Arequipa Terrane can be explained by their derivation from regional A-type Tonian–Cryogenian granitoid rocks in the Eastern and Western Cordillera of the Central Andes in Peru (Mišković et al. 2009) and on the southern continuation of the Arequipa Terrane (e.g. MARA Terrane) in northwestern Argentina (Colombo et al. 2009; Casquet et al. 2012). The most parsimonious interpretation of the detrital zircon comparative analysis is that the Choquequirao Formation, which is statistically similar to age spectra from the Arequipa Terrane, most probably had Arequipa basement as its primary source and was deposited on or near the eastern margin of the Arequipa Terrane.

Assuming that the Arequipa Terrane developed into a ribbon continent generated by two separate Neoproterozoic rifts during the break-up of Rodinia, it has generally been proposed that the eastern conjugate rifted margin is either southwestern Amazonia (Escayola et al. 2011; van Staal et al. 2013) or southwestern Kalahari (Rapela et al. 2016; Casquet et al. 2018). As discussed above, detrital zircon compilations from the rifted margin of southwestern Amazonia are dissimilar to detrital zircon compilations from the Choquequirao Formation, which does not support the model of southwestern Amazonia as a likely conjugate rifted margin of eastern Arequipa. Our analyses indicate a much greater statistical similarity between the Choquequirao Formation and the detrital zircon records from the rifted margins of southwestern Kalahari (Fig. 7). While there is a greater degree of statistical similarity between age spectra of the Choquequirao Formation and the Gariep Belt of southwestern Kalahari (Basei et al. 2005; Hofmann et al. 2015; Thomas et al. 2016), age spectra between the Saldania Belt (Andersen et al. 2018) and the Choquequirao Formation are so similar that the two populations cannot be statistically differentiated (Fig. 7). This suggests that southwestern Kalahari, and in particular the Saldania Belt, could be the conjugate rifted margin of eastern Arequipa. This interpretation has been developed from other lines of geological evidence related to the Pampean Orogeny (Rapela et al. 2016; Casquet et al. 2018), including documentation of dextral shearing from c. 535 to 530 Ma (Iannizzotto et al. 2013; von Gosen et al. 2014) that could help to explain the dextral translation of Arequipa and related terranes. An earlier rift-related link is supported by additional geological evidence. Tonian–Cryogenian mantle-derived detrital zircons and regional occurrences of Tonian–Cryogenian intraplate magmatic sources on Arequipa and related terranes are similar to a large number of c. 880–750 Ma intraplate magmatic sources in southwestern Kalahari (Frimmel et al. 2001; Bartholomew 2008; Hanson et al. 2011; Will et al. 2020). These lines of evidence suggest a possible shared history of intraplate magmatism that may have developed earlier in southwestern Kalahari. We infer that protracted intraplate magmatism eventually led to successful Cryogenian–early Ediacaran rifting on the eastern margin of Arequipa, followed by successful late Ediacaran rifting on its western margin (Fig. 8) (Busch et al. 2022).

Detrital zircon samples from the high-grade metasedimentary rocks of the Huaytapallana and Marañón complexes in central and northern Peru contain contrasting age spectra (Chew et al. 2007a, 2008, 2016; Cardona et al. 2009) that are divided into samples having either a pre-Pampean rift-related affinity or a post-Pampean orogenic affinity (Chew et al. 2008; Cawood et al. 2012). The samples with an orogenic affinity contain prominent Paleozoic ages, tend to be dominated by latest Neoproterozoic–early Paleozoic age peaks, have comparatively reduced Mesoproterozoic age populations and contain a significant number of >2 Ga ages. In contrast, samples with a pre-orogenic or rift-related affinity from the Huaytapallana and Marañón complexes can be characterized as having pre-Pampean (e.g. >560 Ma) youngest detrital zircons, minor Neoproterozoic age peaks, a dominant late Mesoproterozoic age peak and rare occurrences of ages >2 Ga. Samples with an orogenic affinity incorporated into our comparative analysis include CM-228 and CM-112 (Cardona et al. 2009), AM076 (Chew et al. 2007a), and FW2-007 (Chew et al. 2016). Samples with a pre-orogenic rift-related affinity incorporated in our comparative analysis include CM-116 (Cardona et al. 2009), DC-05-5-4 (Chew et al. 2007a, 2008) and FW2-004 (Chew et al. 2016).

The Huaytapallana and Marañón samples with a rift affinity agree very well statistically with samples from both the Arequipa Terrane (Marcona, Llallahue) and the Choquequirao Formation. In contrast, the age spectra of the rift-related samples are dissimilar to compilations of rift-related detrital zircon ages from southern and southwestern Amazonia (Fig. 7) (Babinski et al. 2013; Harris 2020; Harris et al. 2023). The simplest interpretation of the high-grade metasedimentary rock samples with a rift-related affinity in central and northern Peru is that they are most likely from rift-related strata deposited on the Arequipa Terrane. This suggests that the Arequipa Terrane may extend farther north into the Central Andes, even though the northern boundary of the Arequipa Terrane is generally identified between southern and central Peru (e.g. Ramos 2008). We acknowledge that this suggestion is not consistent with mapped differences in whole-rock Pb isotopic signatures in Peru (Macfarlane et al. 1990), yet it is possible that Pb mobility in the mantle wedge associated with subduction throughout the Phanerozoic (Macfarlane et al. 1990 and references therein) may have played a larger role in resetting Pb crustal signatures in central and northern Peru. Further investigation from the Huaytapallana and Marañón complexes, and other Precambrian units, may help to better clarify the extent of the Arequipa Terrane, as well as the occurrence of early Paleozoic orogenic events.

The late Mesoproterozoic supercontinent of Rodinia was in the process of amalgamation and consolidation until c. 880 Ma and its break-up did not commence until c. 800 Ma (Li et al. 2008). The break-up of Rodinia was protracted and diachronous (Li et al. 2008) with the opening of interior oceans, such as the 650–550 Ma opening of the Iapetus Ocean (Robert et al. 2020) following earlier ocean-opening events from 800 to 650 Ma (Li et al. 2008). The diachronous fragmentation of Rodinia and delayed opening of the Iapetus Ocean are global developments in which the tectonic history of the Arequipa ribbon terrane can be situated. Increasingly, it has been put forward that the Iapetus Ocean opened in multiple events that may have generated rift-related ribbon continents such as the Arequipa Terrane (van Staal et al. 2013; Robert et al. 2020).

The timing of the initial opening of the Iapetus Ocean remains somewhat poorly constrained. It has been proposed that the eastern margin of Arequipa and related terranes underwent rifting as early as c. 925 Ma based on the youngest detrital zircon recovered from the Chilla Beds in Bolivia (Bahlburg et al. 2020) but this is inconsistent with the onset of Rodinia rifting at 800–750 Ma (Li et al. 2008; Merdith et al. 2017a), as well as the opening of the proposed conjugate rifted margin in southwestern Kalahari at 750–700 Ma (Macdonald et al. 2010b; Hofmann et al. 2014). The Chilla Beds may represent the eastern rifted margin of the Arequipa Terrane, except that the maximum depositional age is likely to be older than the true depositional age. As shown in detrital zircon studies of modern sediment, maximum depositional ages can be >100 myr older than the true depositional age, even in proximity to volcanic sources (Sharman and Malkowski 2020 and references therein).

Our detrital zircon analyses and revised lithostratigraphy from the Choquequirao Formation suggest that the succession represents rift-related deposits during the Cryogenian–Ediacaran. Affinity of the detrital zircon spectra with detrital zircon records from the Arequipa Terrane (see the previous subsection) support the interpretation that the Cryogenian–Ediacaran rift-related succession developed on the eastern margin of the Arequipa Terrane. Our comparative detrital zircon analysis further suggests that the Choquequirao Formation is most similar to rift-related detrital zircon records from the Gariep and Saldania belts of southwestern Kalahari, which may have been the conjugate rifted margin of eastern Arequipa. This reconstruction is supported by the presence of 850–750 Ma intraplate magmatism on Arequipa and its southern continuation (Colombo et al. 2009; Mišković et al. 2009; Casquet et al. 2012), and c. 800 Ma mantle-derived detrital zircons across the Arequipa Terrane (Hodgin et al. 2021a). Our precise maximum age constraints at c. 640 Ma represent significant new minimum age constraints on rift-related strata associated with the opening of an oceanic tract, which may be the northern extension of the Puncoviscana (Escayola et al. 2011) and Clymene oceans (Casquet et al. 2018). Our new age constraints overlap with other age constraints from the southern continuation of the Arequipa Terrane in northwestern Argentina, where early Ediacaran marbles dated at 635–620 Ma (Murra et al. 2016) and obducted oceanic crust dated at 647 ± 77 Ma (Escayola et al. 2007) constrain the opening of a Cryogenian–early Ediacaran oceanic tract.

By dating metamorphic zircon, titanite and rutile, we identified multiple episodes of metamorphism within the poly-deformed, leucosome-bearing (formed by partial melting) Choquequirao Formation. Metamorphic zircon was dated by in situ spot analyses at c. 550, c. 490 and c. 460 Ma. Detrital zircons fragments dated by CA-ID-TIMS were used to precisely identify a primary Pb-loss event associated with peak metamorphism at c. 460 Ma. Owing, perhaps, to its lower closure temperature compared to zircon (Cherniak 1993), peak metamorphism in metamorphic titanite was identified by ID-TIMS at c. 445 Ma. The precise ID-TIMS titanite analyses exhibited significant overdispersion between 475 and 430 Ma, which could represent either a span of ages corresponding to distinct episodes of metamorphic crystallization during Famatinian orogenesis or Pb-loss pathways, which could be related to distinct metamorphic events prior to and following Famatinian metamorphism. Given that some titanite ID-TIMS analyses that were generated from fragments of the same unzoned titanite crystals contributed to overdispersion, we favour the hypothesis that overdispersion is related primarily to Pb loss. This interpretation is supported by nearly all analyses being discordant and falling upon apparent Pb-loss pathways. Older dates on metamorphic zircon and younger dates on metamorphic rutile from the same sample (B1425) further suggest that overdispersion in our ID-TIMS titanite dataset is likely to be related to Pb loss.

First, we discuss the tectonic implications of well-resolved 460–440 Ma peak metamorphism in metamorphic zircon and titanite from the Choquequirao Formation. The timing of peak metamorphism recorded from in situ and isotope dilution zircon analyses is c. 460 Ma but peak metamorphic conditions arguably persisted for c. 20 myr, corresponding broadly to the final Oclóyic phase of Famatinian orogenesis (Pankhurst et al. 1998). We note that the initial timing of peak metamorphism may be associated with ophiolite emplacement in the central and northern Eastern Cordillera at 465 ± 24 Ma (Castroviejo et al. 2009, 2010; Rodrigues et al. 2010; Tassinari et al. 2011; Willner et al. 2014). The timing may correspond to the suturing of the Paracas Terrane (Ramos 2008). Alternatively, the ophiolitic crust in central and northern Peru could have originated in a back-arc basin that closed during the Famatinian Orogeny, which is the preferred tectonic model in the Eastern Cordillera of southern Peru (Bahlburg et al. 2006, 2011; Ramos 2008; Carlotto et al. 2009). The opening and closing of a back-arc basin is also consistent with our provenance results, which suggest the continuation of the Arequipa Terrane into central and northern Peru. Regardless, the highest temperature metamorphic conditions were likely to have been reached at the time of ophiolite emplacement at c. 465–460 Ma, followed by continued convergence associated with peak metamorphic conditions that persisted for c. 20 myr.

A separate upper Cambrian back-arc basin has recently been described in the Altiplano of southern Peru (Hodgin et al. 2021a). However, the upper Cambrian basin is capped by a significant Early Ordovician unconformity, and thus it appears to be unrelated to subsequent Middle–Late Ordovician back-arc formation and closure. Alternatively, in the better-studied type locality of the Famatinian Orogeny in northwestern Argentina, a separate pulse of intense back-arc extension has been documented (Wolfram et al. 2017) involving tholeiitic back-arc volcanism (Hauser et al. 2008; Coira et al. 2009; Bahlburg et al. 2016), leading to subsequent Middle–Late Ordovician back-arc closure and peak metamorphism during the Oclóyic phase of Famatinian orogenesis (Weinberg et al. 2018). Thus, we infer that protracted peak metamorphism at 460–440 Ma was probably caused by continued convergence associated with and following closure of a short-lived Famatinian back-arc basin.

Turning to pre-Famatinian metamorphism, we first discuss our results in the context of the limited geological evidence available for such metamorphic events. The deposition of quartzite clasts in the Lower Ordovician Verónica Formation (Hodgin et al. 2021a) implies that the underlying quartzites of the Choquequirao Formation, which are the most likely source of the clasts, had been metamorphosed prior to the Early Ordovician. The apparent Pb-loss discordia path in our metamorphic titanite dataset suggests that recrystallization and Pb loss followed an earlier metamorphic event. The poorly constrained upper intercept of c. 510 Ma overlaps with c. 550–490 Ma metamorphic zircon dates from the same sample (B1425). While we have some confidence in identifying at least one pre-Famatinian deformation event in the Choquequirao Formation rocks, we are unable at this time to disentangle the earlier metamorphic events with precision, which could be attributed to metamorphic overprinting, and in particular Famatinian peak metamorphism. As demonstrated from the Huaytapallana and Marañón complexes, the age of the metasedimentary rocks does not necessarily correlate with metamorphic grade, as amphibolite-grade metamorphic events associated with deposition of orogenic strata appear to have occurred during the Ordovician, Carboniferous and Permian (Chew et al. 2007a, 2016; Cardona et al. 2009). Further investigation will be required to tease apart evidence for pre-Famatinian tectonic events recorded within the poly-deformed rocks of the Choquequirao Formation.

The Choquequirao Formation also preserves a record of post-Famatinian deformation. Five U–Pb common Pb-corrected rutile in situ analyses yielded a weighted mean of 327.4 ± 11.5 Ma, which can be interpreted as recording a pulse of metamorphism during Hercynian metamorphism that is recorded elsewhere in the Eastern Cordillera at 325–310 Ma (Chew et al. 2007a, b, 2016; Cardona et al. 2009). Interestingly, this pulse of Carboniferous metamorphism is better expressed in the Marañón Complex of northern Peru than it is in the more proximal Huaytapallana Complex of central Peru, where Permian deformation is better preserved (Chew et al. 2007a, b, 2016). There is significant dispersion within our metamorphic rutile dataset, masked to a degree by relatively large analytical error. The oldest rutile crystal may thus record late-stage Famatinian metamorphism and the youngest rutile crystal may record Permian metamorphism (Fig. 5). These results highlight the preservation of multiple metamorphic events within poly-deformed rocks, and in individual samples and crystals, of the Choquequirao Formation.

The Choquequirao Formation (Formation) is a succession of amphibolite-grade igneous and metasedimentary rocks that crop out in a remote area of the Eastern Cordillera of the Central Andes in southern Peru. Three detrital zircon samples contain mantle-derived 640–635 Ma detrital zircons that constrain the maximum age of the lower part of the succession. The recognition of glacial diamictite in the basal strata of the Choquequirao Formation supports a Cryogenian basal age. The absence of early Paleozoic detrital zircons, which are by comparison common in the upper Cambrian–lower Ordovician Llallahue Formation and the Marañón and Huaytapallana complexes, suggests that deposition of the Choquequirao Formation predates the early Paleozoic orogenesis associated with the final amalgamation of Gondwana. The Choquequirao Formation is most likely to represent a Cryogenian–Ediacaran rift succession related to the break-up of Rodinia. The similarity of the Choquequirao Formation detrital zircon age spectra to the age of the Arequipa basement and to detrital zircons records from Neoproterozoic sedimentary rocks overlying the Arequipa Massif suggests that the Choquequirao Formation was deposited on the Arequipa Terrane, which we propose was its eastern rifted margin. Our detrital zircon compilation analysis indicates a high degree of similarity with pre-orogenic metasedimentary samples identified as rift-related from the Marañón and Huaytapallana complexes in central and northern Peru. This suggests that vestiges of the poorly exposed basement rock of the Arequipa Terrane may extend into northern Peru. Our analysis shows that southwestern Amazonia is an unlikely conjugate rifted margin to eastern Arequipa compared to southwestern Kalahari. The strong similarity between Choquequirao Formation age spectra and compiled data from the Saldania Belt in southern Kalahari supports a rift-related linkage within Rodinia reconstructions. Similar linkages between the Saldania Belt and the southern continuation of the Arequipa Terrane in NW Argentina (e.g. MARA Terrane) have been proposed previously (Rapela et al. 2016; Casquet et al. 2018).

Metamorphic titanite ID-TIMS dates display significant overdispersion along apparent Pb-loss pathways, yet a cluster of c. 470–430 Ma dates generally overlaps with c. 460 Ma peak metamorphism recorded by metamorphic zircon. The younger of two apparent Pb-loss pathways displayed by the metamorphic titanite has a lower intercept that overlaps with c. 325 Ma metamorphic rutile, probably corresponding to Hercynian orogenesis, which is recorded throughout the Eastern Cordillera (Chew et al. 2007a, 2016). The poorly constrained older upper intercept age of c. 510 Ma overlaps with 550–490 Ma metamorphic zircon and may be related to Pampean and/or early Famatinian orogenesis. An earlier phase of metamorphism is also indicated by a change in metamorphic grade across an angular unconformity between the quartzite-bearing Neoproterozoic Choquequirao Formation and the overlying quartzite-clast-bearing conglomerate of the lower Ordovician Verónica Formation (Hodgin et al. 2021a). These new data from the Choquequirao Formation may thus constrain the timing of the opening of the Palaeo-Iapetus (Puncoviscana–Clymene) Ocean in the Cryogenian–lower Ediacaran between the eastern margin of Arequipa and a conjugate rifted margin in southern Kalahari. Subsequently, Arequipa and related terranes collided with the Rio de la Plata and Amazon cratons during the Pampean Orogeny, which was followed by a series of Paleozoic orogenic events that are recorded in the poly-deformed Choquequirao Formation.

We are grateful to Tyler Barringer (SunPower Corporation) and Sam LoBianco (UC Santa Barbara) for field assistance; Janeth Cáceres for logistical support; Marion Lytle and Darin Schwartz (Boise State University) for analytical support; and Nicholas Swanson-Hysell for academic support to E.B. Hodgin. Constructive reviews from Victor Ramos and Cees van Staal helped to improve this paper.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

EBH: conceptualization (lead), investigation (lead), methodology (lead), formal analysis (lead), writing – original draft (lead), writing – review and editing (lead); VC: conceptualization (supporting), investigation (supporting), project administration (supporting), logistical support (lead), writing – original draft (supporting), writing – review and editing (supporting); FAM: funding acquisition (lead), resources (supporting), conceptualization (supporting), project administration (supporting), writing – review and editing (supporting); MDS: formal analysis (supporting), resources (lead), data curation (lead), methodology (supporting), software (supporting); JLC: formal analysis (supporting), methodology (supporting), data curation (supporting).

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

All data generated or analysed during this study are included in the published article and in its supplementary information files.

The publisher apologizes for including incorrect versions of Figures 2, 6 and 8. The correct figures have now been inserted.