Detrital zircon typology and U/Pb geochronology for the Miocene Ladrilleros-Juanchaco sedimentary sequence, Equatorial Pacific (Colombia): New constraints on provenance and paleogeography in northwestern South America

This work is licensed under Creative Commons Attribution-ShareAlike 4.0 International.

Provenance studies in terrigenous sedimentary sequences are crucial for paleogeographic and morpho-tectonic reconstructions (e.g., uplift-erosion, paleotopography, paleofluvial networks, etc.) in mountain ranges and associated sedimentary basins (Pettijohn, 1975; Dickinson, 1985; Bernet and Spiegel, 2004; Garzanti et al., 2007). In recent decades, the study of heavy mineral fractions, geochronology of detrital phases (e.g., U/Pb dating of zircon), and geochemical analysis of detrital materials (trace elements and isotopes) have been incorporated into provenance studies to provide valuable data on the configuration of source areas and their proximity to depositional environments (Armstrong-Altrin et al., 2004; Belousova et al., 2005; Boggs, 2009; Dostal and Keppie, 2009; Etemad-Saeed et al., 2011; Grimes et al., 2015).

Although rock fragments (lithics) and individual mineral grains are excellent indicators of source area, several problems can arise during interpretation. These are related primarily to the lithological and/or compositional similarities between different source areas and to the biased distribution of detrital materials in the sedimentary rock due to the different responses of each fragment/mineral to weathering, erosion, transport, and/or diagenesis (Boggs, 2009). One approach to overcoming these limitations is to use zircons found in the sedimentary rocks. Being chemically and physically refractory, zircon is an ultrastable mineral phase that is capable of enduring multiple geological cycles in sedimentary, metamorphic and/or igneous environments (Garzanti et al., 1987; Finch and Hanchar, 2003; Hanchar and Hoskin, 2003; Mange and Wright, 2007; Gehrels, 2012, 2014). At present, zircons are employed as key source indicators, based on age of the parental rock through radiometric dating or age of cooling events via thermochronology (Bernet and Spiegel, 2004; Carrapa, 2010; Thomas, 2011; Garver, 2014; Gehrels, 2014). The relatively recent development of dating techniques (e.g., U/Pb via LA-ICP-MS) that can target specific regions within a single crystal has enabled space-explicit dating of complex multiphase zircon grains, while reducing analytical costs and increasing data throughput as several hundred grains in detrital samples can be analyzed in a short time (Bowring and Schmitz, 2003; Davis et al., 2003; Košler and Sylvester, 2003; Bernet and Spiegel, 2004; Schoene, 2014). Thus, in the last decade, U/Pb age spectra in detrital zircons has been used widely to improve constraint of source areas (Thomas, 2011; Gehrels, 2014). Nonetheless, U/Pb age signatures have been shown to be problematic in regions where there is high litho-structural complexity, such as age overlaps between different geological provinces, multiple litho-tectonic blocks with similar lithologies/ages, etc. (Reiners et al., 2005; Carrapa, 2010). One such region is the northern Andes in Colombia (Restrepo and Toussaint, 1988; Toussaint, 1995; Cediel et al., 2003; Moreno-Sánchez and Pardo-Trujillo, 2003) (Figs. 1; 2), where similar U/Pb datasets have resulted in contradictory interpretations (e.g.,Montes et al., 2015; O’Dea et al., 2016).

Another approach to studying zircons that have been part of the erosive cycle (i.e., eroded, reworked, and redeposited in sedimentary sequences) is the analysis of different typologies, such as crystal form, degree of elongation, and coloration (Fedo et al., 2003; Anani et al., 2011; Zoleikhaei et al., 2016). Typology of zircon populations (morphology and color) is controlled by the physical and chemical conditions during crystallization (Anani et al., 2011). Morphologic and chromatic features can be examined at high resolution under a petrographic microscope (transmitted and reflected light), and can be associated with the various lithologies present in source areas (Dabard et al., 1996).

In addition to the conventional characterization of zircon morphology by visual inspection under an optical microscope, the development of high-resolution Scanning Electron Microscopy and Cathodoluminescence (SEMCL) enables us to characterize the internal structure and zonation patterns for each zircon grain dated in a detrital sample, further facilitating the discrimination of igneous and metamorphic zircons, as well as zircons with complex histories (Hanchar and Rudnick, 1995; Corfu et al., 2003). The combination of different techniques such as U/Pb dating and zircon typological analysis, together with internal texture analysis using SEM-CL, is conducive to more robust studies in terms of the discrimination of source areas (Loi and Dabard, 1997; Willner et al., 2003; Anani et al., 2012); an approach strongly needed in complex lithostructural domains such as the Northern Andes (Restrepo- Moreno et al., 2015).

In this work we carry out a detailed analysis of detrital zircons from the Ladrilleros-Juanchaco Sedimentary Sequence (LJSS) based on external and internal zircon morphology using optical (Gärtner et al., 2013) and SEMCL (Corfu et al., 2003) microscopy. This approach is complemented with zircon chromatic analysis (Pupin and Turco, 1972b; Pupin, 1980; Velez et al., 2005; Shahbazi et al., 2014) and zircon U/Pb dating by LA-ICP-MS (Košler and Sylvester, 2003; Chang et al., 2006; Solari and Tanner, 2011; Schoene, 2014).

The combination of these techniques provides a robust methodological approach to discriminate among zircons on the basis of more than just their radiometric age and external and internal morphology, thus allowing us to determine the age and conditions of zircon formation in the crystalline source (igneous vs. metamorphic) (Poldevaart, 1950; Pupin and Turco, 1972b; Pupin et al., 1978; Pupin, 1988; Corfu et al., 2003). These multi-technique approach is applicable to detrital zircons, enhancing the accuracy in the determination of source areas (Loi and Dabard, 1997; Willner et al., 2003; Anani et al., 2012; Shahbazi et al., 2014).

The LJSS is located in northwestern South America on the Equatorial Pacific coastline, at ~5°N latitude. Strata of this sequence outcrop in coastal cliffs along La Barra-Juanchaco Beach and Malaga Bay, approximately 40 km northwest of the city of Buenaventura (Fig. 1). Physiographically, the LJSS occupies a littoral position on the Pacific lowlands to the west of the Western Cordillera and to the south of the Serranía Baudó. The LJSS is part of the San Juan Sub-basin that, from a stratigraphic and deep marine deposits sedimentary point of view, is correlated with deep marine deltaic activity (ANH-Universidad de Caldas, 2011). The main modern fluvial system in the area, the Río San Juan, drains the western flank of the Western Cordillera to form one of the largest deltaic systems of the South American Equatorial Pacific (González and Correa, 2001).

FIGURE 1

A) Location of the Ladrilleros-Juanchaco Sedimentary Sequence (LJSS) on the Colombian and South American context on the Pacific coast. The general geodynamic setting is also shown with main tectonic plates and crustal blocks marked as CP: Caribbean Plate, NP: Nazca Plate, SAP-Cr: South American Plate and Craton, PCB: Panamá-Chocó Block, NAB: Nor-Andean Block (from Suter et al., 2008; Taboada et al., 2000; Cediel et al., 2003). B) Local geological framework of the LJSS (modified from Montoya, 2003). C) General aspect of outcrops for the LJSS in the studied section, NW-SE.

Litho-structurally, the LJSS (Figs. 1; 2) lies within a tectonic setting known as the Panamá-Choco Block (PCB) and is part of the San Juan Sub-basin, delimited to the south by the Garrapatas fault system and to the north by the San Juan fault system (Cediel et al., 2003, 2009). The regional geodynamic framework is marked by the interaction of major tectonic plates (Nazca, Cocos, Caribbean and South America) in a combined subduction and collision convergent regimen (Kellogg and Bonini, 1982; Kellogg, 1984; Taboada et al., 2000; Spikings et al., 2001; Pindell et al., 2005; Bayona et al., 2011; Pindell and Kennan, 2013). Smaller lithospheric/crustal blocks like the PCB and the Nor-Andean Block (NAB) (Fig.1) also play an important role in the litho-tectonic configuration of the area. The PCB, for instance, is associated with the last collisional episode in the region from Oligocene to present times (Duque-Caro, 1990; Farris et al., 2011).

Previous studies indicate that the LJSS consists of a continuous, well-exposed sequence of clastic materials, such as mudstones with sandstone intercalations, deposited in a marine environment (ANH-Universidad de Caldas, 2011). The sequence is approximately 750m thick (Fig. 3), although Miocene beds can reach thicknesses of >3000m (Cediel et al., 2003). The sedimentary fill of the basin rests on oceanic rocks typical of Colombia’s western crustal domain, west of the Cauca-Romeral fault, i.e., the Provincia Litosférica Oceánica Cretácico Occidental (PLOCO) (Gómez et al., 2015). According to Muñoz and Gómez (2014), the LJSS sandstones correspond to lithic arkoses, feldspathic litharenites, and litharenites that are texturally (clay matrix, poor selection, and grain angularity) and compositionally immature. These sandstones are also characterized by a significant input of volcaniclastic materials (volcanic fragments, horblende + pyroxene associations, long and euhedral zircon grains) and other detrital components (quartz, metamorphic fragments(?), etc.). The presence of volcanic elements affords the opportunity not only to constrain the age of the sequence but to discriminate among source areas associated with either volcano-magmatic arcs (ages varying between the Cretaceous and the Miocene) or metamorphic elements that can be related to mountain ranges with metamorphic basement rocks such as the Central Cordillera (Muñoz and Gómez, 2014). A Burdigalian-Tortonian age (~16-10Ma) has been established for this sedimentary unit based on calcareous microfossils (Vallejo et al., 2016). To date, no geochronological dataset has been used to constrain the age of the LJSS.

Based on a combined analysis of zircon typology, internal textures/zonation patterns and U/Pb dating, our work improves the dating resolution of the sequence, establishes the origin of detrital components, and identifies the paleogeographic conditions that lead to the development of the LJSS.

FIGURE 2

Main physiographic, structural, and chrono-stratigraphic framework (igneous and metamorphic units), adjacent to the Ladrilleros- Juanchaco Sedimentary Sequence (LJSS). Non-colored areas represent sedimentary rocks and Quaternary deposits. Modified from Gómez et al. (2015).

Eleven sandstone samples were collected along various stratigraphic levels of the LJSS to separate zircon grains for typological analysis (Fig. 3). An additional sample was collected from the top of the sequence for LA-ICP-MS U/ Pb dating and the generation of age spectra, as well as to determine internal zircon zonation patterns through SEMCL imagery. All zircon grains were separated by gravimetric methods combining Wilfley® table/gold pan and heavy liquid runs (TBE: 2.96g/cm³). Mineral concentrates were further purified through additional heavy liquid (MeI: 3.32g/cm3) and Frantz® Isodynamic magnetic separations. Zircon grains were mounted in cylindrical resin plugs (2cm diameter, 1cm thick) for their subsequent preparation (grinding and polishing to expose internal areas) for U/Pb dating. To ensure a random, unbiased assembly of zircon grains from detrital samples, we emptied ~200-300 grains directly from the vial. Finally, we obtained petrographic and SEM-CL microscopy images of the zircons to reveal the external morphology and chromatic features and to examine internal zonation patterns, respectively.

FIGURE 3

Stratigraphic column for the Ladrilleros-Juanchaco sedimentary sequence (adapted from Muñoz and Gómez, 2014). Zircon abundance relative to other heavy minerals is shown. Black dots correspond to the stratigraphic position of samples analyzed. Blue dot marks the position of the sample for detrital zircon U/Pb analysis.

To assess zircon typology, we implemented the methodology of Pupin (1980) and Gärtner et al. (2013) to correlate the analyzed zircons with source area lithologies (e.g., plutonic, volcanic and/or metamorphic), while providing information in terms of multi-cycle zircons. To characterize zircon crystals based on morphology/ typology, the following steps were undertaken after Pupin and Minéraux (1976), Pupin (1980), and Gärtner et al. (2013): i) Scanning of grain mounts under the optical microscope to identify ~150 grains of detrital zircon; ii) zircon grain characterization by specific morphometric parameters, such as elongation, roundness, length/width ratios, and degree of face development; iii) zircon grain characterization by color tones and intensity (e.g., hue, saturation, and brightness) and degree of transparency; iv) zircon classification (Fig. 4); v) statistical calculations to quantify variability along the morphometric correlation matrices for each sample; vi) generate specific typological classifications. Microscopic analyses under transmitted and reflected light were carried out at the Instituto de Investigaciones en Estratigrafía (IIES) of the Universidad de Caldas (Colombia) using a Nikon LV100® Tri-Polar Pol Universal and PET SMZ1500® stereomicroscope, coupled with a high-resolution Nikon DS-F11® camera. NIS Elements 3.2® software was used for measuring morphometric features. SEM-CL images were obtained using a SEM-Quanta 250® coupled with Chroma CL Gatan® cathodoluminescence device.

FIGURE 4

Zircon typological classification proposed by Pupin (1980). Index A reflects the Al/alkali ratio, controlling the development of pyramids in the crystals. Index T reflects the effects of temperature on the development of prisms.

All U/Pb measurements were performed by LAICPMS (Chang et al., 2006; Schoene, 2014), using a 213nm New Wave Nd: YAG UV laser, coupled to a Thermo-Finnigan Element-2 dual-focus monocollector ICP-MS mass spectrometer. Isotopic measurements were carried out at the University of Arizona’s Laser-Chron Laboratory (USA) using the approach proposed by Chang et al. (2006).

Applied as integrated provenance analyses, typology and U/Pb geochronology in detrital zircons enables us to constrain the sources of sedimentary material that filled the San Juan Sub-basin during the Miocene. In terms of roundness, the prevalence (91%) of well-faceted to poorly rounded grains indicates that the analyzed zircons were slightly affected by abrasion, suggesting JLSS zircons underwent minimal transport from proximal sources and/or that the kinetic energy during transport was low (Caspers, 1967; Dietz, 1973).

Elongation analysis is a good indicator of potential source rocks (Poldervaart, 1955; Poldervaart, 1956; Hoppe, 1963; List, 1966). In our study, the more elongated zircons are associated with granitic and/or medium- to high-grade metamorphic origins, while short and rounded zircons are linked to reworking or low-grade metamorphism (Poldevaart, 1950; Hoppe, 1963; Finger and Haunschmid, 1988). Elongation is also an indicator of cooling rates, where rapid cooling (i.e., volcanic) generates mostly longer crystals (elongation ratio >3). In this sense, our results show that LJSS zircons do not exhibit a single tendency of elongation. Instead, the majority of crystals analyzed present moderate to high ratios (between ~2 and 3), while two samples give elongation ratios of 5, suggesting that most zircons have an igneous origin characterized by rapid cooling (e.g., shallow intrusions and/or volcanics).

FIGURE 9

U/Pb results for 114 U/Pb dates reported in the detrital zircons for the LJSS sample. A) Concordia Diagram. B) Probability Density Plot (PDP). Larger PDP covers the entire age range reported. Smaller PDP covers ages between 0 and 300Ma where main peaks are marked with the names of some intrusive bodies of the Western Cordillera that fall in these intervals and that are considered as potential sources for detrital material.

SEM-CL images allow identification of internal textural differences within zircon populations. Precambrian zircons present weak and wide zonation with overlapping domains. Mesozoic zircons display superposition of domains exhibiting non-zoned cores with overgrowth and strong zonation. Such patterns are characteristic of granitic igneous rocks but may also occur in some metamorphic rocks (Corfu et al., 2003). Finally, Miocene zircons are characterized by internal textures exhibiting reabsorbed cores with overgrowth, superposition of zonation domains, and fine bimodal oscillatory zonation (Fig. 10). These textural features are typical of zircons formed in acid to intermediate igneous rocks and develop during zircon crystallization, reflecting processes of reabsorption, precipitation, and crystal growth inside magmatic chambers with compositional variations (Corfu et al., 2003). Zircon typologies in our study indicate that the majority originated from granitic rocks of subalkaline affinity, in addition to alkaline granitic rocks (Figs. 7; 8). However, an association of zircon morphology/typology with metamorphic rocks cannot be ruled out.

FIGURE 10

Cathodoluminescence (SEM-CL) images of selected LJSS zircons discriminated by age ranges in the Precambrian, Mesozoic and Miocene. Black circles correspond to ablation pits from U/Pb dating.

Detrital zircon populations analyzed for U/Pb ages show age spectra with significant peaks at ca. 10Ma, 12-13Ma, and 18-21Ma (Fig. 9). Less prominent peaks are occur in the intervals ~23-31Ma, ~38-40Ma, ~75-82Ma, and ~230- 240Ma. A small proportion of zircons with ages between 500Ma and 2800Ma are also present. Zircons with U/Pb ages in the Burdigalian-Aquitanian interval confirm that material was sourced primarily from plutonic and volcanic rocks of known Miocene age (ca. 18-20Ma), such as the Aguaclara Stock, Danubio Stock, Anchicayá Stock, and Torra Pluton (Brook, 1984; ANH-Universidad de Caldas, 2011). In addition, the younger plutonic Miocene bodies, Serravallian to Tortonian (ca. 10-14Ma), may indicate detrital signatures from northern intrusive sources including the Farallones Batholith, the Horqueta Stock, and other bodies reported by Calle et al. (1980), García and García (2012) and Restrepo-Moreno et al. (2013). Oligocene U/Pb signatures are attributable to plutons such as the Piedrancha Batholith (Brook, 1984). The contribution of Cretaceous granites can be linked to the Buga Batholith (Villagómez et al., 2011) and the Mistrató Stock (Calle and González, 1982; Moreno-Sanchez and Pardo-Trujillo, 2003). However, the presence of volcanic lithics and hornblende + pyroxene associations (Muñoz and Gomez, 2014), in conjunction with the euhedral and elongated morphology of some zircon grains (elongation ratios >3), with U/ Pb ages predominantly between 10 and 21Ma, are clear indications of volcanic activity during the Miocene depositional stages of the LJSS in western Colombia.

Our results indicate that the primary sources of detrital material for the LJSS correspond to the Late Mesozoic (i.e., Cretaceous) and Cenozoic litho-stratigraphic provinces (Fig. 2), which can be correlated with the present configuration of basement rocks of the Western Cordillera. The age ranges reported above are typical of the Western Cordillera’s plutonic (intermediate to mafic) and volcanic (intermediate) rocks, suggesting that this mountain range was already a significant topographic barrier, with exposed and geomorphically active crystalline massifs. Conversely, Triassic to Precambrian zircons are often associated with crystalline rocks of the Central Cordillera (Horton et al., 2015; Nie et al., 2010; Vinasco et al., 2006). The presence of a few zircon grains in this U/Pb age-range, with typical peaks at ~230-250Ma, 600Ma, 900-1000Ma, etc., does not necessarily imply a direct fluvial connection between the Central Cordillera and sedimentary formations along the Pacific coast. We interpret such zircons as being derived from reworking of Cretaceous sedimentary covers positioned over the Western Cordillera (e.g., Penderisco Formation, Urrao Member) where populations with these Paleozoic and Precambrian ages have been reported (i.e., polycyclic zircons; ANH-Universidad de Caldas, 2011).

Based on the presence of Cretaceous zircons in the LJSS, it could be argued that, since the Central Cordillera also exhibits considerable pulses of Cretaceous magmatism (Restrepo-Moreno et al., 2009; Villagómez and Spikings, 2013; Cochrane et al., 2014), the possibility of a fluvial connection between the Central Cordillera and the LJSS should not be discounted. Yet, the absence of Jurassic U/ Pb ages from the SSJL is notable, despite of the fact that there are plutonic masses of considerable size and Jurassic age in the Central Cordillera, such as the Ibague Batholith (Gómez et al., 2015), positioned at almost the same latitude as the SSJL.

Relative to the PCB, the virtual absence of Eocene zircons in the interval ~38-50Ma, corresponding to some intrusives typical of the Panamanian Arc (e.g., Mandé Batholith (Botero, 1981; Villagómez and Spikings, 2013) and/or the Acandí Batholith (Álvarez and Parra, 1971; Sillitoe et al., 1982), would indicate that: i) the PCB was still in a distal position with respect to the LJSS at 10Ma, ii) Mandé-like plutonic units were still unexposed massifs (i.e., in subsurficial positions and in the process of exhumation) and thus did not contribute sediments to the basin and/or iii) there was no fluvial network connecting geomorphically active Panamanian elements with the LJSS. However, recent studies indicate that the plutonic masses of the Panamanian Arc were already exhumed and served as source areas for sedimentary formations in Panamá (Ramírez et al., 2016), making options i) and iii) above more reasonable.

The prevalence of Western Cordillera-like zircon detrital signatures (based on U/Pb age and typologies, this study), the dominant heavy mineral fractions for these sediments (Muñoz and Gómez, 2014) and the absence of significant zircon U/Pb age signals older than the Cretaceous can be considered as preliminary evidence for the topographic control exerted by the Western Cordillera hindering fluvial connections between the Central Cordillera and the LJSS. In addition, the presence of zircons potentially derived from Miocene plutonic bodies (e.g., Farallones Stock, Tatamá Pluton, etc.) located north and east of the LJSS, and on an axial portion on the Western Cordillera, indicates that such plutons were exhumed and geomorphically active and thereby formed areas of detrital contribution within the paleofluvial network. Further to this detritalfluvial contribution, we suggest there may have been a significant direct contribution from the volcanic products of a Miocene magmatic arc.

FIGURE 11

Paleogeographic model for the LJSS, Proto-Cauca fluvial system and western cordillera during the Miocene (Burdigalian-Tortonian). Colors are only a reference to differentiate environments and lithologies and do not represent geologic time.

Based on our results, a feasible paleogeographic reconstruction shows that during the time of LJSS sedimentation (Burdigalian-Tortonian) there was a volcanic-magmatic arc in the Western Cordillera of sufficient topographic relief to separate Central Cordillera source areas from Pacific basins. Marked pulses of uplift and erosional exhumation are also reported for the Western Cordillera for the interval 15-9Ma (Restrepo-Moreno et al., 2015). As in present times, this topographic barrier played a pivotal role for a river network restricted to the western flank of the Western Cordillera that controlled the extent of fluvial dissection, erosion, and sediment production-routing into the Pacific plains (Fig. 11). The proposition of an already existing Western Cordillera as a considerable topographic barrier since the middle Miocene is consistent with determinations, based on evidence from borehole and seismic data, on the extent of Pacific basins in Colombia (i.e., those situated to the west of the Western Cordillera) that show significant reduction in thickness as one moves from west to east (i.e., from the Pacific Ocean to the Western Cordillera).

Integration of techniques such as detrital zircon typology/morphometry, U/Pb dating, and SEM-CL imaging in samples from the LJSS arenites affords improved constraint of the depositional age and provenance of the terrigenous materials deposited in the basin during the Burdigalian-Tortonian. The high affinity of the zircons (typology/age/internal zonation patterns) with Miocene igneous rocks (plutonic and volcanic) potentially implies that an active volcanic-magmatic arc with significant topography existed in the Western Cordillera and was the main sediment source. Detrital sources on the Central Cordillera were separated from the Pacific realm by wellestablished Western Cordillera topographic barriers.

Our methodological approach to evaluate provenance in a complex litho-structural regime represents a viable development of the standard methodology, considering the quantity of information provided in provenance studies and paleogeographic reconstruction, and should be systematically used in other sedimentary sequences in the northern Andes.

This work was carried out with support from the Instituto de Investigaciones en Esratigrafía (IIES) and the Research Group in Stratigraphy and Volcanology (GIEV-Cumanday) at Universidad de Caldas (Colombia). Institutional support from the National Hydrocarbons Agency (ANH, Colombia), Universidad Nacional de Colombia Sede Medellín. We are grateful to the personnel at the LaserCrhon Laboratory of the University of Arizona. We greatly appreciate the comments and suggestions from Yamirka Rojas-Agramonte and Javier Fernández-Suárez, whose input significantly improved the quality of the manuscript.