Temporary storage of sediment between source and sink can hinder reconstruction of climate and/or tectonic signals from stratigraphy by mixing of sediment tracers with diagnostic geochemical or geochronological signatures. Constraining the occurrence and timing of intrabasinal sediment recycling has been challenging because widely used detrital geo-thermochronology applications do not record shallow burial and subsequent reworking. Here, we apply strontium isotope stratigraphy techniques to recycled marine shell material in slope deposits of the Upper Cretaceous Tres Pasos Formation, Magallanes Basin, Chile. Detrital 87Sr/86Sr ages from 94 samples show that the majority (>85%) of the shells are >1–12 m.y. older than independently constrained depositional ages. We interpret the gap between mineralization age (87Sr/86Sr age) and depositional age of host strata to represent the intrabasinal residence time of sediment storage at the million-year time scale. We also use specimen type to infer relative position of intrabasinal source material along the depositional profile, where oysters represent shallow-water (i.e., proximal) sources and inoceramids represent deeper-water (i.e., distal) sources. The combined use of detrital strontium isotope ages and specimen types from linked depositional segments provides an opportunity to identify and quantify sediment storage and recycling in ancient source-to-sink systems.

Detrital mineral analysis is commonly used to infer geomorphic characteristics, tectonic history, and exhumation patterns associated with ancient sediment source areas. However, analysis of some Quaternary sediment routing systems shows that an appreciable proportion of sediment can be stored for 102 to 106 yr between erosional source and terminal depositional sink (e.g., Clift et al., 2014; Anderson et al., 2016). Transient storage of detritus can hinder accurate reconstruction of external signals by introducing a lag time between denudation and burial and/or mixing sediment with different and potentially unique compositional, geochemical, or geochronological signatures that are linked to specific forcings or source terranes. Detrital chronology has been used to document recycling, but at very long (108 to 109 yr) time scales in which basins are wholly inverted in association with tectonic regime change (e.g., Hadlari et al., 2015; Andersen et al., 2018). There remains a lack of knowledge about recycling at time scales over which basin stratigraphy develops (∼105 to 107 yr) because neither sandstone composition nor crystallization or cooling age of detrital minerals record transient storage and remobilization within a sediment routing system. This study introduces detrital strontium isotope stratigraphy as an approach for identifying and measuring intrabasinal sediment residence time. We use detrital marine carbonate samples from the Upper Cretaceous Tres Pasos Formation of southern Chile to (1) constrain intrabasinal sediment residence time to 106 to 107 yr prior to recycling, and (2) identify distinct intrabasinal sources defined by specimen type.

Detrital Strontium Isotope Stratigraphy

Marine biogenic carbonates preserve the 87Sr/86Sr isotopic composition of ocean water when they precipitate (Wickman, 1948). The 87Sr/86Sr composition of global seawater varies through time and has been well documented for the Phanerozoic (Burke et al., 1982). Strontium isotope stratigraphy (SIS) compares measured isotopic ratios from marine samples to calibrated global isotopic curves in order to determine numerical ages (McArthur et al., 2012). Samples used for SIS are in situ specimens where values can be regarded as depositional ages and used for stratigraphic correlation. We apply detrital SIS (DSIS) to recycled carbonate shell material (i.e., shells and shell fragments that were reworked). Marine shell fragments are assumed to have precipitated in the basin from which they were recovered. The age difference between DSIS sample and host strata defines a period of residence time prior to recycling. This distinguishes DSIS from methods that use other commonly analyzed detrital grains that can be stored and mixed during transport from an extrabasinal source terrane (Fig. 1). A DSIS sample can be used to infer the intrabasinal sediment storage and recycling history that extrabasinal grains from equivalent strata experienced. Carbonate material in the hinterland is a potentially confounding variable; however, there are no known carbonates in the hinterland for this study (Fig. 2A).

Study Area

The Magallanes Basin is a retroarc foreland linked to Mesozoic–Cenozoic fold-thrust belt evolution in southern Chile and Argentina (Fig. 2A; 48°–56°S; Dalziel et al., 1974; Fildani and Hessler, 2005). Basin subsidence and sedimentation patterns were influenced by inversion of the predecessor Rocas Verdes backarc basin such that Magallanes Basin deposition occurred in a relatively spatially fixed foredeep (Fosdick et al., 2013). Upper Cretaceous Magallanes Basin deposits record a progradational north-to-south basin-axial fill, where northern (proximal) shallow- and non-marine strata transition to southern (distal) deep-marine deposits (Hubbard et al., 2010; Manríquez et al., 2019). The studied transect is an ∼85-km-long depositional-dip-parallel section of outcropping Magallanes Basin deposits (Figs. 2B and 2C). The interval is temporally constrained by volcanic ash U-Pb zircon ages from basal Tres Pasos Formation deposits (80.5 ± 0.3 Ma; Daniels et al., 2018) and detrital zircon maximum depositional ages (DZ-MDAs) from lower Dorotea Formation deposits (70.6 ± 1.5 Ma; Daniels et al., 2018; Fig. 3A). We place DSIS results within the stratigraphic context and four-phase model for Tres Pasos Formation and Dorotea Formation evolution from Daniels et al. (2018) (Fig. 3A).

Samples (n = 208) were collected from multiple positions in each of the stratigraphic phases at 14 locations numbered L1–L14 along the study transect (Figs. 2B and 3A). Shell fragments were derived from medium- and coarse-grained sandstone beds, shell-hash conglomerates, and extrabasinal-clast conglomerates. Samples were selected based on presence of primary shell material and absence of diagenetic alteration, which can impact SIS and DSIS applications (McArthur et al., 1994). Samples were washed with deionized water, dried, and ground into fine powder. Before grinding, a subset was inspected via scanning electron microscopy to evaluate crystal structure and integrity and screen for appropriate elemental composition across the surface of the sample (Ca, C, O; see the Supplemental Material1). For samples with sufficient material (165 samples; 79% of total), trace element analysis via inductively coupled plasma mass spectrometry was used for additional screening to confirm sufficiently high concentrations of Sr and acceptably low concentrations of likely cation replacements for Sr (Ca, Fe, Mg, Mn; see the Supplemental Material). Strontium isotopic values were then measured for 147 samples on a Neptune multicollector inductively coupled plasma mass spectrometer following isolation and purification of Sr from carbonate digests using standard ion exchange techniques (see the Supplemental Material). All 87Sr/86Sr ratios have been normalized to the accepted value of NBS987 = 0.710248 (McArthur et al., 1994). 87Sr/86Sr isotopic ratios were then converted to numerical ages using version 5 (26 March 2013; J.M. McArthur, 2013, personal commun.) of the LOWESS lookup table (Howarth and McArthur, 1997; McArthur et al., 2001).

We use published ages from throughout the basin (Daniels et al., 2018) to evaluate ages determined by measured 87Sr/86Sr ratios. We use a DZ-MDA of 67.1 ± 1.8 Ma (Daniels et al., 2019) to define the youngest potential age and upper limit of acceptable values (87Sr/86Sr = 0.707799). We use 88.2 ± 0.6 Ma (87Sr/86Sr = 0.707352) to define the lower limit of acceptable values, which is the oldest in situ SIS age recorded from the underlying Cerro Toro Formation (Bernhardt et al., 2012). Values lower than 87Sr/86Sr = 0.707352 are not used because while the measurement may be accurate, the value is non-unique on the global curve (McArthur et al., 2012). Ninety-four (94) samples produced acceptable 87Sr/86Sr values (45%), 53 produced non-unique values (26%), and 61 were rejected during screening (29%). No samples from location L14 produced acceptable 87Sr/86Sr values due to diagenetic alteration.

Analytical results, uncertainties, and numerical ages are reported in Table S1 in the Supplemental Material. Approximately 14% of DSIS ages are within DZ-MDA ranges, whereas the remaining 86% are ∼1–12 m.y. older than the depositional age (Fig. 3). There is a prominent DSIS age mode at ca. 79 Ma that diminishes in occurrence in more-southern samples, and a secondary mode at ca. 88–87 Ma observed across the transect that corresponds to in situ SIS ages from the Cerro Toro Formation reported by Bernhardt et al. (2012; Fig. 3C). There is a notable stratigraphic change in specimen type, with oysters representing >90% of samples in younger depositional phases (Tres Pasos Formation phase 4 [TP4] and Dorotea Formation) but only 38% in older phases TP1 and TP2.

Intrabasinal Sediment Residence Time

As a first-order finding, we submit that 87Sr/86Sr measurements from detrital shell material can be used to identify the geologic time at which a sample precipitated. Intrabasinal sediment residence time prior to recycling can be determined by comparing the age of host strata to the DSIS ages. Residence-time calculation errors are therefore linked to the age constraint on host strata. Here we calculate maximum intrabasinal residence times as the difference between oldest available MDA with error and the maximum DSIS age (DSIS-MDA) with error. Maximum residence times for Tres Pasos Formation phases are 9.5 m.y. for TP1, 9.8 m.y. for TP2, and 11.7 m.y. for TP4 (Fig. 3). These residence times suggest that Tres Pasos deposits are at least partially composed of recycled Cerro Toro detritus sourced from within the basin.

Individual samples that have a short intrabasinal residence time can be used to constrain depositional age of host strata. This occurs most notably at locations L1 and L13 where DSIS-MDAs fall within the upper bounds of DZ-MDA error (Fig. 3B). Location L13 samples were collected from inner-shelf sandstones of the Dorotea Formation, suggesting only local reworking of oyster shells within the shelf environment where they originally precipitated. Samples from location L1 were collected from sandstone interbeds of TP1 turbidite fan deposits. The relatively lower-energy flows associated with these more unconfined deposits suggests that inoceramids yielding young DSIS-MDAs were minimally reworked and either precipitated locally or were efficiently transported.

Intrabasinal Source Terrane

As a second-order finding, we submit that specimen type can be used to fingerprint the source of intrabasinal sediment based on paleoenvironmental differences in preferred habitats of key specimen types. In this study, samples are dominantly either oysters (51%) or inoceramids (40%). Oysters represent higher-energy and dominantly shallow-water sources (i.e., inner shelf; Dalrymple and Choi, 2007; Schwartz et al., 2017), and inoceramids represent dominantly quiescent and typically deeper-water sources (i.e., outer shelf to basin floor; Dhondt, 1992; Bernhardt et al., 2012). While there is some overlap in their respective potential origination zones, we interpret specimen type as a proxy for gross longitudinal position of intrabasinal sediment source within the basin. Here, oysters represent a more landward sediment source, and inoceramids represent a more basinward source. Specimen types binned by Tres Pasos phase (Fig. 3B) reveal distinct intrabasinal source signatures: TP1 is inoceramid dominated (63%), TP2 is mixed with 56% inoceramids and 42% oysters, and TP3–TP4 are oyster dominated (95% oysters). We interpret this signature to record a progressively more landward intrabasinal sediment source through Magallanes Basin evolution (Fig. 4).

The DSIS age populations in Tres Pasos samples remain relatively constant, indicating sustained recycling of Cerro Toro– and TP1-aged sediment. The paucity of DSIS-MDAs for TP2–TP4 suggests either that carbonates were not precipitating in the basin during that period or that carbonate shells continued to form but were stored elsewhere in the basin. We suggest that differences between youngest MDAs from upper slope and shelf deposits versus slope deposits across the study area support the latter. The oldest DSIS ages from shelf deposits at location L13 are TP4 age (ca. 74 Ma), indicating precipitation of shell material at least during TP4. No significantly older DSIS grains from TP1 or Cerro Toro are recorded here, suggesting they were not recycled at this proximal position. Similarly, all of the youngest DZ-MDAs for TP2–TP4 come from upper slope or shelf positions, whereas stratigraphically equivalent DZ ages are older (e.g., TP4; Fig. 3A). The implication is that while progressively younging DZ-MDAs from shelf deposits indicate efficient sediment transfer from the hinterland to the basin, the lack of equivalent ages from contemporaneous or even younger slope deposits (e.g., DZ-MDA at location L10; Fig. 3A) suggests that sediment is not efficiently transferred to more distal basin positions. We interpret the lack of younging in DSIS samples through Tres Pasos stratigraphy to indicate updip storage of Tres Pasos–age sediment and recycling of Cerro Toro–age sediment during coarse-grained slope deposition of TP2 and TP4 (Fig. 4).

The Tres Pasos phase-specific distributions of specimen types indicate that progressively younger Tres Pasos strata preserve the same age populations as intrabasinally sourced material derived from a progressively more landward position within the basin (Fig. 4). This geologic scenario of intrabasinal sediment recycling is non-unique because multiple mechanisms could produce similar sediment-recycling signals (e.g., downcutting in response to eustatic, tectonic, and/or autogenic base-level fluctuations).

Detrital SIS records recycling of material that was shallowly buried within a sedimentary basin (i.e., not buried sufficiently for low-temperature thermochronometers to be reset; Fosdick et al., 2015) and subsequently recycled within the same basin. Our results cannot be used to quantitatively determine proportions of hinterland versus intrabasinally sourced material because neither the spatial and stratigraphic distribution nor the amount of the parent material (i.e., in situ shells) are explicitly known. Nonetheless, the implication is that if trace amounts of marine carbonate material were recycled, then it is plausible that a large amount of lithogenic sediment was also reworked. Intrabasinal recycling may result in the mixing of hinterland source-area tracers such as detrital zircons that could confound the interpretation of signal propagation into the deep basin (e.g., Sharman and Johnstone, 2017). While our study reveals the need for sediment storage and mixing processes to be considered in provenance analysis, DSIS offers a potential opportunity to identify and quantify intrabasinal sediment recycling in some settings.

We establish the use of detrital strontium isotope stratigraphy (DSIS) to identify a signal of intrabasinal sediment recycling, provide temporal constraint on sediment storage at million-year time scales, and elucidate the importance of intrabasinal sediment mixing in deep-marine deposits. Results show sediment storage and mixing at time scales of >1–12 m.y. in duration. Additionally, using specimen type as a proxy for location of sediment source along the longitudinal basin profile, we document three distinct intrabasinal recycling signatures: (1) inoceramid dominated, sourced from more distal deep-water sediment (TP1); (2) mixed inoceramids and oysters from deep-water and shallow-water sources, respectively (TP2); and (3) oyster-dominated sediment from shallow-water sources (TP3–TP4; Fig. 3). While these results provide insight into intrabasinal sediment recycling and Magallanes Basin evolution, they also establish the utility of detrital strontium isotope stratigraphy as a proxy for intrabasinal recycling.

The funding for this work was generously provided by the sponsors of the Chile Slope Systems (CSS) Joint Industry Project (BHP Billiton, Chevron, ConocoPhillips, Equinor, Hess, CNOOC, Repsol, and Shell). We are grateful to landowners for granting access to outcrops. We thank the many CSS members and friends for assistance in collecting samples. We are grateful for reviews from Tim Lawton, Tomas Capaldi, Kurt Sundell, and an anonymous reviewer.

1Supplemental Material. Sample screening, trace element analysis, strontium isotope measurements, histogram, and kernel density estimate. Please visit https://doi.org/10.1130/GEOL.S.12425390 to access the supplemental material, and contact editing@geosociety.org with any questions.
Gold Open Access: This paper is published under the terms of the CC-BY license.