Detrital zircon and apatite U-Pb provenance and drainage evolution of the Newark Basin during progressive rifting and continental breakup along the Eastern North American Margin, USA

Our understanding of progressive rifting along continental margins and the processes leading to lithospheric breakup has greatly improved over the past decades. New geophysical data (e.g., Peron-Pinvidic et al., 2013; Lavier et al., 2019), including long-offset reflection and refraction seismic data sets across rifted continental margins, geological studies of exhumed fossil analogs (e.g., Hart et al., 2017), lithospheric-scale numerical modeling (e.g., Lavier and Manatschal, 2006; Huismans and Beaumont, 2011; Brune et al., 2014; Gouiza and Naliboff, 2021), and advances in analytical techniques (e.g., Hart et al., 2017) revolutionized the fundamental concepts of both magma-rich and magma-poor rifting, hyperextension, and continental breakup. While these advances allow for reconstructions of the temporal and thermal evolution of rifted margins, limited progress has been made in understanding the surface topographic evolution during progressive rifting and continental breakup and the sedimentary source-to-sink dynamics of magma-poor or magma-rich continental rifted margins. Topographic surface evolution and sediment routing and dispersal are causally intertwined and offer an opportunity to illuminate and track deeper crustal and lithospheric processes driving dynamic syn-rift topography and drainage evolution in three dimensions.

Recent numerical models for the evolution of rifted and passive continental margins make clear and testable predictions about sediment routing and dispersal during progressive rifting and lithospheric breakup (e.g., Haupert et al., 2016; Pérez-Gussinyé et al., 2020; Neuharth et al., 2022). In the earliest rifting phase, normal faulting in the upper crust forms a series of grabens and half grabens filled with sediment locally sourced from both hanging wall and footwall (e.g., Schlische, 1993; Gupta et al., 1999; Lavier and Manatschal, 2006; Manatschal et al., 2015; Haupert et al., 2016). Normal-fault geometries become more asymmetric and laterally connected as rifting progresses in response to the thermal weakening of the crust caused by asthenospheric upwelling. Progressive hyperextension leads to mechanical coupling and the structural dismemberment of a central keystone rift block, driving rapid basin subsidence and a provenance shift to more regional sourcing (Unternehr et al., 2010; Haupert et al., 2016; Decarlis et a., 2017a). Overall, the transition from the stretching and localized extension to margin-wide hyperextension marks the transition from locally dominated to regional integrated sediment dispersal (Hart et al., 2017). Mechanical coupling can lead to widespread mantle exhumation, but voluminous magmatism can abort this progressive rifting and trigger lithospheric breakup and seafloor spreading (Tugend et al., 2020; Harkin et al., 2020; Pérez-Gussinyé et al., 2020; Sapin et al., 2021). Voluminous magmatism associated with the widespread Central Atlantic magmatic province (CAMP) along the Eastern North American Margin (ENAM) likely drove regional doming, leading to a regional margin-wide paleodrainage reorganization and a shift in provenance signature prior to final lithospheric breakup (Geoffroy, 2005; Saunders et al., 2007).

These conceptual and numerical models advanced our understanding of the two-dimensional rift evolution, but numerous geological studies show that three-dimensional (3-D) complications (e.g., rift segmentation, rift polarity changes, and accommodation zones) strongly control the input of coarse-grained siliciclastics into the early syn-rift basins (Schlische and Anders, 1996; Gupta et al., 1999; Gawthorpe and Leeder, 2000; Kolawole et al., 2021). Many previous studies utilized modern provenance tools in studying convergent and collisional orogens and continental-scale drainage systems, but only a few studies have leveraged them to study the dynamic evolution of rifted margins (e.g., Hart et al., 2017). This methodology, however, is well suited for reconstructing the temporal and spatial response of the syn-rift sedimentary provenance, routing, and dispersal to help elucidate and track dynamic surface evolution and drainage configurations and illuminate crustal and lithospheric evolution during early rifting, hyperextension, and the impact of magmatism.

The early Mesozoic ENAM is an excellent natural laboratory because it is both characterized by a diverse collage of hinterland tectonic terranes as the result of protracted collision accretion and orogenesis (e.g., Hibbard et al., 2006; Hatcher et al., 2007) and impacted by large-volume syn-rift magmatism associated with the CAMP well after the onset of rifting (e.g., Blackburn et al., 2013). The Triassic and Jurassic syn-rift basins of the ENAM and their stratigraphic archives span the diffuse rifting, initial necking, hyperextension, and CAMP magmatism phases and record syn-rift sedimentary provenance evolution and the interplay between crustal thinning and transition to magmatic breakup. These ENAM rift basins stretch from Maritime Canada to Florida (USA) and record the syn-rift stratigraphic and 3-D structural and geometric evolution of continental rifting, including ~30 m.y. of extensional faulting and syn-rift deposition prior to CAMP magmatism and subsequent continental breakup (e.g., Olsen, 1997; Withjack et al., 2013, 2020). The alluvial, fluvial, and lacustrine lithofacies of the Newark Supergroup comprise the syn-rift strata of the individual rift basins along the proximal central ENAM. Carnian to Sinemurian deposits are preserved in a series of half grabens, delimited predominately by east-dipping basin-bounding faults controlled by the inherited pre-rift Appalachian structural grain (Lindholm, 1978). While previous studies focused on individual basin structures and stratigraphy, no integrated regional tectonic, stratigraphic, and provenance studies of ENAM rift basin evolution exist or are framed within the modern framework of non-magmatic or magmatic rift models. Here we present extensive new detrital zircon (DZ; N = 21; n = 3093) and detrital apatite (DA; N = 4; n = 559) U-Pb provenance results from Carnian to Sinemurian syn-rift sedimentary strata of the Newark Basin and reconstruct the sediment sourcing and dispersal during progressive rifting and CAMP magmatism. We supplement the DZ U-Pb provenance data set with DA U-Pb data because DZ data sets provide a record of source region crystallization ages susceptible to orogenic recycling, generating uncertainties within provenance studies (Morton and Hallsworth, 1999). Apatite is mechanically less stable and provides a record of the upper-crustal thermal evolution of source regions due to the intermediate closure temperature (~400–450 °C), thus yielding a critical additional constraint on the provenance (e.g., Morton and Hallsworth, 1999; Cochrane et al., 2014; O’Sullivan et al., 2018). Based on these new U-Pb data from the syn-rift sedimentary strata in the Newark Basin, marked provenance shifts are readily identifiable and point to two major paleodrainage reconfigurations during progressive rifting along the ENAM, following early diffuse rifting, in response to (1) early Norian rift flank uplift and crustal necking and (2) regional doming and/or uplift associated with CAMP magmatism at the Triassic-Jurassic boundary. Post-CAMP deposits tracking the final breakup are not preserved in the proximal rift basins. These results show the ability of DZ and DA U-Pb geochronology to track dynamic surface responses during progressive rifting and onset of magmatic breakup and the power of augmenting DZ data with DA U-Pb analyses in unraveling and refining provenance shifts due to the ability to recognize both high- and medium-temperature tectono-magmatic events.

The Late Triassic to Early Jurassic ENAM syn-rift basins stretch from Florida (USA) to Nova Scotia (Canada) and formed during progressive continental extension prior to the breakup of Pangea and the development of the Central Atlantic (Fig. 1; e.g., Olsen, 1997; Withjack et al., 1998; Leleu et al., 2016; Withjack et al., 2020). These rift basins are exposed only in the proximal portion of the ENAM but are buried under the coastal plain and offshore by thick post-rift passive-margin sequences. The early Mesozoic Newark Supergroup comprises all the Triassic to Jurassic syn-rift strata from the different individual basins (Olsen, 1980, 1997; Froelich and Olsen, 1984) and includes (1) fluvial conglomerate and sandstone; (2) lacustrine siltstone, mudstone, and minor sandstone; and (3) abundant alluvial conglomerate and sandstone (e.g., Lindholm, 1978; Weems, 1980; Hentz, 1985; Olsen et al., 1990; Smoot, 1991; Olsen, 1997; Faill, 2003; Schlische and Withjack, 2005). Deposition of the syn-rift sedimentary strata occurred within a series of north- to northeast-striking half-graben structures and minor full-graben structures superimposed on the structural grain of the Alleghanian orogen (Schlische, 1993; Hatcher et al., 2007). The normal faults bounding these basins are generally seaward dipping and are the result of the reactivation of inherited thrust faults and suture zones formed during the late Paleozoic Alleghanian and older Paleozoic orogenies (Fig. 1; Lindholm, 1978).

The Newark Supergroup represents the entire sequence of Triassic and Jurassic syn-rift sedimentary strata and intercalated basaltic rocks preserved within the ENAM rift basins (e.g., Glaeser, 1966; Olsen, 1978; Froelich and Olsen, 1984; Smoot, 1991; Olsen et al., 1996; Faill, 2003; Leleu et al., 2016). In the Newark Basin, these early, middle, and late syn-rift strata, consisting of alluvial, fluvial, and lacustrine conglomerates, sandstones, siltstones, and mudstones, have been subdivided into groups, formations, and members based on geographic location and lithologic character (Fig. 2). The following paragraphs provide brief descriptions and characteristics from each tectono-stratigraphic interval (TS) of the identified early (TS 1), middle (TS 2), and late (TS 3) syn-rift packages. Summary descriptions of each tectono-stratigraphic interval including rock descriptions, sandstone petrography, unit thickness, and accumulation rate are from Olsen (1980, 1997), Oshchudlak and Hubert (1988), and Smoot (1991).

The early syn-rift sedimentary rocks of the Newark Basin (TS 1) include the Stockton and Lockatong Formations and consist mainly of gray, purple, red, and black mudstone, siltstone, arkosic sandstone, and minor conglomerate with a maximum thickness of ~1.5 km and a time-averaged accumulation rate of 0.15 mm/yr (Fig. 2; Olsen, 1980, 1997; Smoot, 1991). The lower Stockton Formation consists of poorly sorted, cross-bedded sandstone overlying imbricated cobble to boulder conglomerate with abundant normal grading and scour surfaces. On average, the Stockton Formation contains ~52% rock fragments, 24% quartz, 14% plagioclase, and 10% K-feldspar (Oshchudlak and Hubert, 1988). It is commonly overlain by bioturbated mudstone and siltstone (Smoot, 1991). The lower Stockton Formation has been interpreted as having been deposited in a braided fluvial setting with vegetated flood plains (Smoot, 1991). The upper portion of the Stockton Formation is characterized by rhythmically bedded fine-grained sandstone, siltstone, and mudstone exhibiting cross bedding, normal grading, and abundant bioturbation (Smoot, 1991). In contrast to the lower Stockton Formation, the upper Stockton Formation appears to represent a meandering fluvial depositional environment (Smoot, 1991). Along the Newark Basin border fault, alluvial conglomerate and sandstone as well as debris-flow deposits are inferred based on seismic-reflection data sets (Bally et al., 1991; Withjack et al., 2013). Glaeser (1966) utilized grain size, sorting, and compositional analyses to interpret a sedimentary provenance from the southwestern hanging wall of the basin bounding fault for the Stockton Formation.

The Lockatong Formation consists of cyclical gray and black siltstone and minor red mudstone with a maximum thickness of ~1.2 km and a mean accumulation rate of 0.2 mm/yr (Olsen, 1980, 1997; Smoot, 1991). These thinly bedded, laminated, organic-rich deposits contain fish fossils, aquatic reptile fossils, abundant burrow structures, soft-sediment deformation, and mud cracks. On average, the Lockatong Formation contains ~42% rock fragments, 39% plagioclase, 12% quartz, and 7% K-feldspar (Oshchudlak and Hubert, 1988). Smoot (1991) interpreted the cyclical depositional environment as deep to shallow playa lakes. The Lockatong Formation siltstone and mudstone laterally thin and interfinger with coarser siliciclastic rocks likely deposited in fluvial to fluvial-deltaic systems in the southern and northern portions of the Newark Basin (Olsen, 1997). Paleocurrents collected from two locations indicate axial drainage during the deposition of the Lockatong Formation (Parker et al., 1988).

The middle syn-rift sedimentary rocks of the Passaic Formation (TS 2) consist of cyclic, laterally continuous, red and gray siltstone, mudstone, and minor sandstone and evaporite with mud cracks, root structures, and burrows (Olsen, 1980; Smoot, 1991; Olsen et al., 1996). The mudstone and siltstone of the Passaic Formation contain ~55% quartz, 24% rock fragments, 16% plagioclase, and 5% K-feldspar (Oshchudlak and Hubert, 1988). These rocks were deposited in shallow and fluctuating lacustrine environments (Smoot, 1991; Olsen et al., 1996). Laterally, the Passaic Formation interfingers with fluvial and alluvial conglomerate, sandstone, and siltstone at the northern and southern terminations of the Newark Basin (Fig. 2). The strata of the Passaic Formation reach a thickness of ~3.5 km with a mean accumulation rate of 0.2 mm/yr in the Newark Basin (Olsen, 1997). Paleocurrent measurements show a persistent axial drainage system during TS 2 (Parker et al., 1988). Alluvial and fluvial conglomerate and sandstone near the basin-bounding fault are inferred based on a previous seismic study (Bally et al., 1991; Withjack et al., 2013). While predicted by Schlische (1992), there are currently no documented unconformities at the top of TS 2 (Olsen et al., 2010).

The late syn-rift rocks of the Newark Basin (TS 3) include interbedded siliciclastic rocks of the Feltville, Towaco, and Boonton Formations and intercalated basaltic flows (Orange Mountain, Preakness, and Hook Mountain Basalts; Olsen et al., 2010). The Feltville, Towaco, and Boonton Formations contain red and gray thinly bedded to massive organic-rich mudstone, shale, siltstone, and sandstone deposited in lacustrine settings (Olsen, 1980, 1997; Smoot, 1991). Additionally, conglomeratic intervals are present in the Feltville, Towaco, and Boonton Formations (Olsen, 1980). The Feltville Formation reaches a maximum thickness of 600 m and contains ~49% rock fragments, 30% quartz, 16% plagioclase, and 5% K-feldspar, (Olsen, 1980; Oshchudlak and Hubert, 1988). The Towaco Formation exhibits a maximum thickness of 340 m and contains ~48% rock fragments, 32% quartz, 14% plagioclase, and 6% K-feldspar (Olsen, 1980; Oshchudlak and Hubert, 1988). The Boonton Formation reaches a maximum thickness of 500 m and contains ~30% rock fragments, 31% plagioclase, 26% quartz, and 13% K-feldspar (Olsen, 1980; Oshchudlak and Hubert, 1988). Previous age determinations of the interbedded basaltic flows in the Newark Basin, including the Orange Mountain, Preakness, and Hook Mountain Basalts, indicate a peak magmatic pulse of CAMP magmatism in the region at ca. 201 Ma (Marzoli et al., 2011; Blackburn et al., 2013). A marked increase in accumulation rate from ~0.2 mm/yr to 0.6 mm/yr appears to have occurred at the onset of Jurassic deposition in the Newark Basin and is attributed to an increase in extension rate at the onset of CAMP magmatism (Olsen, 1997; Schlische, 2003).

In North America, it is generally accepted that rifting along the ENAM began in the Late Triassic based on palynomorph ages for the oldest syn-rift strata (Cornet, 1977), but the lack of abundant palynomorph assemblages and syn-depositional volcanic zircon U-Pb ages within the earliest syn-rift sediments to constrain the absolute timing of deposition adds significant uncertainty to the timing and spatial trends of rift initiation. Some regional trends, however, are clear from published regional geochronological data sets. For example, sedimentary rocks south of the Culpeper Basin record Carnian to early to middle Norian early syn-rift deposition (Cornet, 1977). The Culpeper, Gettysburg, Newark, and Hartford Basins contain Jurassic basalt flows in the upper portions of the syn-rift stratigraphy, preserving a more complete record of Carnian to Sinemurian syn-rift sedimentation. While a wealth of magnetostratigraphic data provides an important correlation between these ENAM syn-rift basins, the lack of independent age constraints and the potential for major intra-basinal unconformities and depositional hiatuses add considerable uncertainty to the timing of onset and duration of rifting (Olsen, 1997; Olsen et al., 2011, 2015; Tanner and Lucas, 2015). The occurrence of interbedded CAMP-related basalt flows and post-basalt clastic deposition indicate Sinemurian syn-rift sedimentation in the Hartford, Newark, and Culpeper Basins (Olsen, 1997; Marzoli et al., 2011; Blackburn et al., 2013). However, the duration and cessation of the sedimentation within the proximal ENAM realm are unknown due to significant post-rift erosion during the transition to normal seafloor spreading at ca. 165 Ma (e.g., Greene et al., 2017; Withjack et al., 2020).

Inherited pre-rift crustal architecture, including suture zones, thrust faults, and terrane boundaries, plays a key role in the development, localization, and evolution of rift systems (e.g., Manatschal et al., 2015). Along the U.S. ENAM, the pre-Triassic structural and lithologic framework is the result of a complex multi-stage tectonic history impacted by the Proterozoic Grenvillian orogen, Neoproterozoic Iapetus rifting and passive-margin development, and Paleozoic Taconian, Acadian, and Alleghanian subduction, accretion of peri-Laurentian and peri-Gondwanan terranes, and continental collision. This protracted tectonic history yields a heterogenous collection of terranes, lithologic packages, brittle and ductile structural grains, and varied tectono-magmatic histories (Hibbard et al., 2006). This heterogeneous pre-rift tectono-magmatic collage of the ENAM provides an ideal underpinning for identifying sedimentary sources, track provenance changes, and reconstruct paleodrainage evolution during progressive rifting using both classic and novel provenance tools and allows for the determination of the source-to-sink history during the early Mesozoic continental breakup of the super-continent Pangea.

Figure 3 summarizes the potential source regions of the Triassic and Jurassic syn-rift sedimentary rocks along the central ENAM, with potential source regions categorized based on tectonic histories and the similarity of DZ U-Pb age spectra. Figure 4 shows a compilation of zircon U-Pb age spectra of each of these potential source regions. In the Laurentian realm, the Blue Ridge–Grenville region includes Grenville granitic gneisses and the Neoproterozoic passive-margin sequence dominated by Grenville age zircon (900–1300 Ma; Carter et al., 2006; Macdonald et al., 2014). The Laurentian foreland consists of the foreland basin deposits associated with both the Acadian and Alleghanian collisions and contains dominant Taconian (410–500 Ma), Grenville (900–1300 Ma), and Granite-Rhyolite province (1300–1500 Ma) zircon age peaks (Gray and Zeitler, 1997; Becker et al., 2005, 2006; Park et al., 2010; Thomas et al., 2017).

The axial realm of the U.S. ENAM is generally composed of Neoproterozoic to Ordovician oceanic and arc-related terranes with dominant Acadian (340–410 Ma), Taconian (410–500 Ma), and Grenville (900–1300 Ma) zircon age peaks (Carter et al., 2006; Hughes et al., 2014; Merschat et al., 2017). In the northeastern U.S., the axial realm consists of similar terranes as the southern, with the addition of the enigmatic Hawley and Moretown Formations. The Hawley and Moretown source region contains Cambrian to Early Ordovician metasedimentary belts with dominant Grenville (900–1300 Ma) and Pan-African (500–700 Ma) zircon age peaks (Macdonald et al., 2014).

The Carolina terrane source region includes Neoproterozoic metavolcanic and metasedimentary rocks with a dominant Pan-African (500–700 Ma) zircon age peak (Barker et al., 1998; Dennis and Wright, 1997; Wortman et al., 2000). The Avalonia terrane source region comprises Neoproterozoic metavolcanics and calc-alkaline metagranites, metasedimentary rocks deposited in a forearc setting, and a Cambrian cover sequence deposited in a rift-related platform setting (Murphy et al., 2004). These lithostratigraphic units of the Avalonia terrane source region show a dominant Pan-African (500–700 Ma) zircon age peak (Keppie et al., 1998; Thompson and Bowring, 2000; Barr et al., 2003; Murphy et al., 2004). In contrast, Silurian–Devonian foreland basin strata in the northeastern U.S. contain dominant Taconian (410–500 Ma), Grenville (900–1300 Ma), and Granite-Rhyolite province (1300–1500 Ma) zircon age peaks (Wintsch et al., 2007). The Goochland terrane source region is composed of a complex suite of igneous and metamorphic rocks, including Mesoproterozoic granitic gneisses and meta-anorthosites intruded by Neoproterozoic granitic plutons and a heterogeneous suite of paragneisses. The Goochland terrane source region contains Alleghanian (250–340 Ma), Acadian (340–410 Ma), Taconian (410–500 Ma), Pan-African (500–700 Ma), and Grenville (900–1300 Ma) zircon age peaks (Owens and Tucker, 2003; Owens et al., 2010; Bailey et al., 2020). The peri-Gondwanan arc source region consists of Ordovician metavolcanic, metaplutonic, and metasedimentary rocks accreted to the Ganderia terrane and is dominated by Ordovician zircon U-Pb ages (Valley et al., 2020). The Ganderia terrane source region consists of a Mesoproterozoic basement intruded by Neoproterozoic magmatism and is dominated by a Pan-African (500–700 Ma) zircon age peak (Lin et al., 2007; Fyffe et al., 2009; van Staal et al., 2012; Dorais et al., 2012).

In light of these heterogenous tectono-magmatic hinterland source regions and their distinct U-Pb age spectra of Laurentian, peri-Gondwanan, and accreted island arc terranes, we embarked on an extensive DZ U-Pb provenance study of the Triassic–Jurassic syn-rift strata of the Newark Basin to illuminate the progressive structural and topographic evolution of the proximal central ENAM.

Nine sandstone samples were systematically collected from Late Triassic to Early Jurassic proximal syn-rift strata outcrops of the Newark Basin in Pennsylvania and New Jersey (USA). This outcrop sample suite was augmented with 12 samples from core recovered during the Newark Basin Coring Project (NBCP) and the Passaic River Diversionary Tunnel Project (PRDTP) and housed at the Rutgers University Core Repository (Piscataway, New Jersey). The sampling strategy focused on collecting samples from the three tectono-stratigraphic intervals (TS 1–TS 3) as well as from multiple sampling transects across the Newark Basin to capture both temporal and spatial variability.

Samples (N = 21) were separated following standard mineral separation techniques including crushing and grinding, water table, heavy liquid, and Frantz magnetic separation. All U-Pb analyses were conducted at the UTChron facilities at the University of Texas at Austin. Separated detrital zircon were sprinkle mounted on adhesive tape on 1-inch (2.5-cm-diameter) acrylic discs co-mounted with primary and secondary reference materials. Individual, randomly selected unpolished zircon grains were depth-profile analyzed by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) U-Pb geochronology to resolve rim-core relationships, utilizing Photon Machines Analyte G2 excimer laser with a HelEx sample chamber attached to a Thermo Element2 ICP-MS. For all DZ analyses, a 30 μm laser spot with a nominal energy of 4 mJ, an energy density of ~1.98 J/cm², and a pulse rate of 10 Hz was utilized to ablate to a depth of ~15 μm, calibrated using a Bruker optical profilometer. To obtain a statistically robust provenance data set and resolve all detrital age components comprising >5%, we analyzed >117 zircon grains from each syn-rift sample with sufficient zircon available (Vermeesch, 2004). GJ-1 zircon was utilized as the primary reference material (601.7 ± 1.3 Ma; Jackson et al., 2004), while Plešovice zircon (337.13 ± 0.37 Ma; Sláma et al., 2008) was the secondary reference material to monitor the reproducibility of the analyses. The data for these analyses were reduced using Iolite data reduction software with the VizualAge tool (Paton et al., 2011; Petrus and Kamber, 2012), and propagated 2σ errors are utilized and reported. For U-Pb dates <850 Ma, the ²⁰⁶Pb/²³⁸U dates are taken for the reported U-Pb age, and for dates >850 Ma, the ²⁰⁷Pb/²⁰⁶Pb date is reported. Data filters of ²⁰⁶Pb/²³⁸U error limit of 10%, a ²⁰⁷Pb/²⁰⁶Pb versus ²⁰⁶Pb/²³⁸U discordance limit of 30%, and a ²⁰⁶Pb/²³⁸U versus ²⁰⁷Pb/²³⁵U discordance limit of 30% were utilized to remove zircon with large uncertainties and high levels of discordance. Kernel density estimate (KDE) plots, age peak picks, and maximum depositional age (MDA) calculations are made using DetritalPy (Sharman et al., 2018). Complete detrital zircon U-Pb results can be found in the Supplemental Material¹.

To complement and refine the DZ U-Pb data set, a subset of the samples (N = 4) was analyzed for apatite U-Pb. DZ provenance data sets are based on zircon crystallization ages of basement source regions and zircon are prone to orogenic recycling, leading to potential ambiguities within provenance studies, such as the presence of Grenville zircon in all Laurentian and peri-Laurentian source regions along the ENAM (e.g., Carter et al., 2006; Park et al., 2010). In contrast to zircon, apatite is mechanically less stable than zircon during Earth-surface transport (Morton and Hallsworth, 1999). Furthermore, the intermediate closure temperature of apatite (~400–450 °C) provides a record of the upper-crustal thermal and/or orogenic evolution of source regions, thus yielding an additional constraint on the sedimentary provenance and alleviating ambiguities within detrital zircon provenance data sets (Cochrane et al., 2014; O’Sullivan et al., 2018).

Similar to the DZ U-Pb methodology, apatite grains were tape mounted on 1-inch (2.5-cm-diameter) acrylic discs and randomly selected and analyzed by LA-ICP-MS to obtain DA U-Pb provenance data sets. MAD apatite (Thomson et al., 2012) was utilized as the primary reference material (in-house thermal ionization mass spectrometry [TIMS] age of 472.4 ± 0.7 Ma), and McClure Mountain apatite (523.5 ± 1.5 Ma; Schoene and Bowring, 2006) was used as the secondary reference material to determine age reproducibility. For all apatite analyses, a 40 μm laser spot with an energy of 4 mJ, an energy density of ~1.98 J/cm², and a pulse rate of 10 Hz was utilized to ablate to a depth of ~18 μm. Data reduction was completed using the VizualAge_UcomPbine data reduction scheme in Iolite, accounting for the presence of variable common Pb in the primary standard (Chew et al., 2014). Internal errors were exported and reported in the full apatite U-Pb results (see ^{footnote 1}). A ²⁰⁷Pb-based common-Pb correction was applied to each 30 s integration using the Stacey and Kramers (1975) terrestrial Pb-evolution model to calculate its corresponding initial ²⁰⁷Pb/²⁰⁶Pb ratio following an iterative approach (Chew et al., 2011). Filters on each 30 s integration of 20% uncertainty on the ²⁰⁷Pb/²⁰⁶Pb and ²³⁸U/²⁰⁶Pb ratios were used. Terra-Wasserburg plots and KDEs of the common-Pb (Pbc)- corrected apatite U-Pb ages were plotted using IsoplotR and DetritalPy, respectively (Vermeesch, 2018; Sharman et al., 2018).

Informed by the tectono-magmatic and orogenic evolution of the various crustal domains and terranes of the ENAM, the DZ U-Pb age population of the samples analyzed is subdivided into seven major DZ U-Pb components: 250–340 Ma, 340–410 Ma, 410–500 Ma, 500–700 Ma, 900–1300 Ma, 1300–1500 Ma, and >1500 Ma. In the following sections, the DZ U-Pb ages are reported as percentages of these seven dominant tectono-magmatic components (>5%). Figure 5 shows the sample locations and corresponding DZ results in the form of pie charts for each Newark Basin DZ sample. Figures 6 and 7 present KDEs, set with a bandwidth of 10 of the combined zircon rim and core U-Pb ages and corresponding pie charts for Triassic and Jurassic syn-rift sandstone samples, respectively.

Six samples from the Carnian Stockton Formation and two samples from the early Norian Lockatong Formation characterize the early syn-rift strata (TS 1) of the Newark Basin. These sandstone samples were collected from the NBCP core, outcrops along the Delaware River, and a sample ~4 km north of Franklin, New Jersey. The stratigraphically oldest Stockton Formation sample (28NW; n = 130) contains a major component of 500–700 Ma (53%) and minor components of 250–340 Ma (12%) and 1500–3000 Ma (16%; Fig. 6A). Sample PC-3694 (n = 264) exhibits a major age component of 500–700 Ma (37%) and minor components of 250–340 Ma (14%), 340–410 Ma (12%), 410–500 Ma (19%), 900–1300 Ma (6%), and 1500–3500 Ma (6%; Fig. 6B). Sample PC-2817 (n = 285) comprises a dominant component of 500–700 Ma (60%) and subsidiary components of 250–340 Ma (10%), 340–410 Ma (6%), 900–1300 Ma (6%), and 1500–3500 Ma (8%; Fig. 6C). The remaining early syn-rift Stockton Formation samples PC-1754 (n = 83), PC-965 (n = 312), and 29NW (n = 106) broadly show age components and proportions similar to those of sample PC-2817 (Fig. 6D–6F) with a dominant 500–700 Ma component and minor components of 250–340 Ma, 340–410 Ma, 900–1300 Ma, and 1500–3000 Ma.

Three samples were analyzed from the stratigraphically higher Lockatong Formation. Sample 30NW (n = 120), a sandy siltstone sample collected along the Delaware River, shows a dominant 500–700 Ma component (66%) and minor components of 250–340 Ma (7%), 410–500 Ma (6%), 900–1300 Ma (8%), and 1500–3500 Ma (6%; Fig. 6G). In contrast, sample 15NW (n = 134), a sandstone collected from a fluvial-deltaic deposit, is characterized by major components of 900–1300 Ma (57%) and 1300–1500 Ma (20%) and minor components of 410–500 Ma (7%) and 1500–3500 Ma (10%; Fig. 6H). The stratigraphically youngest sample 8NW (n = 136), a silty sandstone collected 3 km north of Norristown, Pennsylvania, shows major 340–410 Ma (20%), 410–500 Ma (20%), and 500–700 Ma (28%) components and minor 250–340 Ma (12%), 900–1300 Ma (7%), 1300–1500 Ma (6%), and 1500–3500 Ma (5%) components (Fig. 6I).

Four samples from the Passaic Formation, characteristic of the middle syn-rift strata (TS 2) within the central ENAM proximal domain, consist of red and gray sandy siltstones to fine-grained sandstones and were collected from outcrops in Pennsylvania and New Jersey and the Martinsville NBCP core. Sample 86NW (n = 152) from the Passaic contains major age components of 500–700 Ma (24%) and 900–1300 Ma (20%) and minor components of 250–340 Ma (11%), 340–410 Ma (13%), 410–500 Ma (9%), 1300–1500 Ma (9%), and 1500–3500 Ma (9%; Fig. 6J). Sample 50NW (n = 145) exhibits a similar age spectrum as sample 86NW, but with an increase in the 900–1300 Ma component to 34% (Fig. 6K). Sample 9NW (n = 128) from the middle syn-rift Passaic Formation displays similar age components as sample 50NW but a reduced 500–700 Ma component (12%) and an increased 900–1300 Ma age component (39%; Fig. 6L). Sample MT-1332 (n = 122) from the uppermost Passaic Formation directly underlying the Orange Mountain Basalt exhibits major DZ U-Pb age components of 500–700 Ma (30%) and 900–1300 Ma (20%) and minor components at 250–340 Ma (8%), 340–410 Ma (14%), 410–500 Ma (14%), and 1500–3500 Ma (7%; Fig. 6M).

The late syn-rift strata (TS 3) are characterized by intercalated basalt flows, marking the onset of CAMP magmatism. The sandstone samples from these strata were collected from the PRDTP and NBCP cores at the Rutgers University Core Repository. Samples were obtained from the Feltville, Towaco, and Boonton Formations. Sample MT-139 (n = 131) from the Feltville Formation directly overlying the Orange Mountain Basalt shows similar peaks to those of sample MT-1332 but with an increase in the major 500–700 Ma component (53%) and a decrease in the 900–1300 Ma age component (16%; Fig. 7A).

Two sandstone samples were analyzed from the Towaco Formation. Sample PT14-429 (n = 130) yielded major components of 500–700 Ma (37%) and 900–1300 Ma (28%) and minor components of 250–340 Ma (6%), 410–500 Ma (7%), and 1500–3500 Ma (10%; Fig. 7B). The stratigraphically higher sample PT14-59 (n = 134) shows a decrease in the 500–700 Ma component (8%) and an increase in the 410–500 Ma (16%), 900–1300 Ma (48%), and 1300–1500 Ma (14%) components (Fig. 7C).

Five samples were collected from the youngest rocks exposed in the Newark Basin belonging to the Boonton Formation. Sample PT9-387 (n = 73) shows major age components of 410–500 Ma (22%) and 900–1300 Ma (64%) and a minor age component of 1300–1500 Ma (8%; Fig. 7D). The stratigraphically younger samples PT9-210 (n = 113), PT7-229 (n = 130), and PT7-47 (n = 134) are all characterized by similar age spectra to that of sample PT9-387 with the major age components of 410–500 Ma, 900–1300 Ma, and 1300–1500 Ma ranging 14%–18%, 48%–55%, and 12%– 19%, respectively (Figs. 7E–7G). Sample 38NW (n = 131), a pebble conglomerate collected near the basin-bounding fault, contains a dominant 900–1300 Ma (85%) component and a minor component of 1300–1500 Ma (11%; Fig. 7H).

The bulk DA Pbc-corrected U-Pb age populations from syn-rift sandstone samples analyzed can be subdivided into five major DA U-Pb components associated with orogenic or tectono-thermal events affecting the ENAM margin—230–340 Ma, 340–410 Ma, 410–500 Ma, 500–700 Ma, and 900–1300 Ma—and are reported as percentages of these tectono-thermal components. Figure 8 presents the bulk DA Pb_c-corrected U-Pb results as KDEs with a bandwidth of 10. Sample PC-965 from the early syn-rift (TS 1) Stockton Formation exhibits dominant components of 230–340 Ma (34%) and 500–700 Ma (31%) with minor components of 340–410 Ma and 410–500 Ma (13% and 11%, respectively; Fig. 8). Sample 86NW from the middle syn-rift (TS 2) Passaic Formation shows a dominant 230–340 Ma (64%) and minor 340–410 Ma (11%), 410–500 Ma (6%), and 500–700 Ma (6%) components (Fig. 8). Sample MT-139 from the late syn-rift (TS 3) Feltville Formation displays dominant components of 230–340 Ma (38%) and 340–410 Ma (25%) and minor components of 410–500 Ma and 500–700 Ma (13% and 11%, respectively; Fig. 8). Sample PT14-59 from the late syn-rift (TS 3) Towaco Formation contains major components of 230–340 Ma (30%), 340–410 Ma (21%), and 410–500 Ma (29%) and minor components of 500–700 Ma and 900–1300 Ma (5% and 6%, respectively; Fig. 8).

The syn-rift stratigraphy of the Newark Basin preserves an ~30 m.y. record of sediment provenance, paleodrainage, and landscape evolution during Late Triassic to Early Jurassic progressive rifting along the central ENAM. Diagnostic DZ U-Pb age components linked to peri-Gondwanan, Laurentian, Taconian, and Grenvillian sources are leveraged to determine provenance and temporal variations in sediment sourcing of the different syn-rift sedimentary packages in the Newark Basin. Overall, these new DZ U-Pb results from 21 samples (3093 DZ U-Pb analyses) document two major paleodrainage reconfigurations during progressive rifting along the central ENAM prior to final breakup and the onset of seafloor spreading (Fig. 9).

The Carnian Stockton Formation, the oldest exposed syn-rift strata within the Newark Basin, consists of conglomerate, sandstone, and siltstone. DZ U-Pb results from the Stockton Formation show a dominant 500–700 Ma component consistent with derivation and the transport of sediments from peri-Gondwanan source regions located to the east of the Newark Basin (Figs. 9 and 10A). These DZ age spectra show similarities with the those of the Avalonia terrane or other similar peri-Gondwanan terranes buried below the coastal plain along the central ENAM by passive-margin sedimentary deposits (Fig. 4). Additionally, the new DZ results from the Stockton Formation show a minor 900–1300 Ma component linked to the Grenville crust. However, sources with Grenville-age DZ components exist in both the footwall and hanging wall of the Newark Basin. The proportion of the Grenville component appears to be similar to that observed in peri-Gondwanan source regions (e.g., Avalonia terrane) and is likely a result of the recycling of Grenvillian zircon. Our new DZ U-Pb results along with previous sedimentological and stratigraphic observations indicate dominant transverse fluvial systems transporting sediment into the Newark Basin from the hanging wall during the Carnian (e.g., Smoot, 1991; Faill, 2003; Fig. 11A).

In sharp contrast, three samples from the overlying early Norian Lockatong Formation exhibit starkly different DZ components and indicate an abrupt shift in provenance (Figs. 6G–6I). The very fine-grained sandstones (samples 30NW and 8NW) show a DZ spectrum that is dominated by 500–700 Ma ages and is nearly identical to that of the underlying Stockton Formation and still sourced from peri-Gondwanan rocks in the hanging wall of the Newark Basin in the early Norian (Figs. 9, 10A, and 11A). In contrast, the fluvial to fluvial-deltaic conglomerate and sandstone (sample 15NW) is characterized by major Grenville (900–1300 Ma) and Granite-Rhyolite province (1300–1500 Ma) and minor Taconian (410–500 Ma) components, suggesting a Laurentian source region. Rocks of a Laurentian affinity are present in both the footwall and hanging wall of the Newark Basin (Fig. 3); however, the absence of Grenville and Granite-Rhyolite province DZ ages from the fine-grained sample (30NW; Fig. 6G) suggests two spatially separate active fluvial systems during the early Norian. These fluvial systems were likely interfingering systems derived from the hanging wall and the emergent footwall uplift. This scenario is supported by published seismic-reflection imaging that shows coarse-grained alluvial and/or fluvial deposits along the Newark Basin footwall indicative of active faulting and footwall exhumation during the Carnian and early Norian, responsible for the transport of Laurentian-sourced material of the footwall into the early syn-rift basin. The absence of Laurentian components within the transverse fluvial systems of the Stockton and Lockatong Formations, however, indicates that the influx of footwall-derived sediments remained overall limited and restricted to alluvial and fluvial systems proximal to the Newark Basin bounding normal fault. The absence of these Laurentian sediments within the Stockton Formation argues for the presence of a drainage divide close to the ENAM border fault system due to paleotopography either inherited from the Appalachian orogeny and/or related to rift flank uplift. Major Appalachian topography, a regional drainage divide, and a west-flowing late Paleozoic and early Mesozoic transcontinental river system are supported by DZ U-Pb data from Triassic–Permian eolian sandstone on the Colorado Plateau (Dickinson and Gehrels, 2003). Dickinson and Gehrels (2003) argued for nearly half the eolian sand delivered to these early Mesozoic ergs originating from the Appalachian orogenic belt along the eastern Laurentia margin. This Appalachian topography likely developed during the Alleghenian orogen before the onset of ENAM rifting but was accentuated by rift flank uplift in the Carnian.

The Passaic Formation consists of interbedded red to gray siltstones with minor very fine-grained sandstones deposited in playa or shallow lacustrine depositional systems (Olsen, 1980; Smoot, 1991). Along with this change in the depositional environment, middle syn-rift deposition is marked by major changes in DZ U-Pb age spectra and provenance. In contrast with the early syn-rift provenance, the middle syn-rift provenance shows the addition of the Grenville (900–1300 Ma) and Granite-Rhyolite province (1300–1500 Ma) age components. This switch in provenance signature signals a basin-wide paleodrainage reconfiguration and switch from an early syn-rift peri-Gondwanan hanging-wall source to a mix of hanging wall–and footwall-derived Grenville and Laurentian sources located to the west of the Newark Basin (Figs. 9, 10B, and 11B). At the northern and southern ends of the Newark Basin, fluvial conglomerate, sandstone, siltstone, and mudstones of the Hammer Creek Formation and an unnamed unit are interbedded with the Passaic Formation (Olsen, 1980; Smoot, 1991). The presence of coarse-grained lithofacies in the basin tips suggests relay ramps provide the dominant accommodation zone for delivery of sediments from the Grenvillian and Laurentian foreland source regions, comparable to the basin development in the Suez Rift system, Egypt (Gupta et al., 1999). Additionally, the Passaic Formation samples contain a sizable Pan-African (500–700 Ma) component, indicating the peri-Gondwanan source regions on the Newark Basin hanging wall remained paleohighs and sources of the Newark Basin deposition during the mid- to late Norian (Fig. 10B).

Jurassic deposition in the Newark Basin is marked by the Orange Mountain Basalt, the first CAMP-related basalt flow preserved in the Newark Basin (Olsen, 1980; Smoot, 1991). Overlying the Orange Mountain Basalt is the Feltville Formation comprising mudstone, siltstone, and sandstone deposited in a shallow lacustrine environment. The DZ U-Pb spectra of the Feltville Formation are characterized by a dominant Pan-African (500–700 Ma) and minor Paleozoic components related to various Laurentian orogenic cycles (250–500 Ma), Grenville (900–1300 Ma), and Granite-Rhyolite province (1300–1500 Ma) components (Fig. 9). These age spectra are in stark contrast with those of the underlying Passaic Formation and are consistent with a return to major peri-Gondwanan and subordinate Laurentian sourcing from the Newark Basin hanging wall (Figs. 10C and 11C). These results suggest a reconfiguration of paleodrainages contemporaneous with the onset of CAMP magmatism. We interpret this reconfiguration of paleodrainages and apparent transition from large rift-shoulder catchments to smaller catchments draining local relief to be attributable to regional topographic uplift and doming of the greater Newark Basin area during CAMP magmatism and magmatic continental breakup.

In the Newark Basin, the Feltville Formation is overlain by basalt flows of the Preakness Basalt—the second sequence of CAMP basalt flows—and the Towaco Formation comprising Jurassic syn-rift sandstone, siltstone, and mudstone deposited in a shallow lacustrine environment (Smoot, 1991). The oldest Towaco Formation sample collected is characterized by a DZ age spectrum similar to that of the Feltville Formation, including a dominant Pan-African (500–700 Ma) and subordinate Paleozoic DZ age components related to the various Paleozoic Laurentian orogenic cycles (250–500 Ma). In contrast, however, the uppermost Towaco Formation shows a marked decrease in the Pan-African (500–700 Ma) component and an increase in other DZ age components, such as Taconian (410–500 Ma), Grenville (900–1300 Ma), and Granite-Rhyolite province (1300–1500 Ma) provenance components. This suggests a transition from peri-Gondwanan hanging-wall sources to a mix of peri-Gondwanan and Laurentian sources (Figs. 9, 10D, and 11D), suggesting ENAM-wide uplift.

The youngest exposed syn-rift strata (Boonton Formation) in the Newark Basin overlie the Hook Mountain Basalt—the third CAMP volcanic flow preserved within the syn-rift stratigraphic column of the Newark Basin. The Boonton Formation consists of sandstone, siltstone, and mudstone deposited in a shallow lacustrine setting (Smoot, 1991). DZ U-Pb results from the Boonton Formation show a continued and increased multi-modal age spectra, consisting of mixed components of both Laurentian (250–500 Ma, 900–1300 Ma, and 1300–1500 Ma) and Pan-African (500–700 Ma) sources similar to those of the underlying Towaco Formation. Together, the DZ U-Pb results from Jurassic syn-rift strata of the Newark Basin indicate a second paleodrainage reorganization during late syn-rift CAMP magmatism (Figs. 9, 10D, and 11D). This switch to mutli-modal DZ U-Pb age spectra in the late syn-rift units suggests that the emplacement of the CAMP caused regional uplift and reworking of the regional and local paleodrainages of the Newark Basin.

DA U-Pb data provide important additional insights into syn-rift provenance because they record the medium-grade tectono-metamorphic evolution of the source region, given apatite’s intermediate closure temperature of ~400–450 °C (Cochrane et al., 2014; O’Sullivan et al., 2018). In addition, while apatite is mechanically less stable than zircon or rutile during Earth-surface sedimentary transport, it is ubiquitous in felsic and mafic igneous rocks, and thus augments DZ provenance data (Morton and Hallsworth, 1999). Refractory zircon recycling along the central ENAM is common, as illustrated by the presence of dominant Grenville zircon in all Laurentian and axial domains and minor Grenville zircon in peri-Gondwanan sources (e.g., Carter et al., 2006; Park et al., 2010; MacDonald et al., 2014). Along the central ENAM, DA U-Pb analyses provide the ability to fingerprint and differentiate potential sediment sources with similar DZ age spectra but different mid- to upper-crustal Paleozoic orogenic and cooling histories recorded by apatite (O’Sullivan et al., 2018). Four samples from the Newark Basin were analyzed for DA U-Pb, including one sample from the Stockton Formation (TS 1), one sample from the Passaic Formation (TS 2), one sample from the Feltville Formation (TS 3), and one sample from the Towaco Formation (TS 3).

The Stockton Formation shows a DA U-Pb age spectrum with a dominant 230–340 Ma component and minor components of 340–410 Ma and 410–500 Ma, indicative of potential source regions that were affected by Alleghanian, Acadian, and Taconian tectonic, metamorphic, or magmatic events. In addition, the Stockton Formation shows a major Pan-African (500–700 Ma) cooling age component that indicates derivation for the early syn-rift Stockton Formation from a source area not overprinted by Paleozoic deformation and metamorphism. Presently, no medium-temperature thermochronometric studies exist for exposed rocks along the central ENAM that yield Neoproterozoic Gondwanan cooling ages (e.g., Sutter et al., 1985; Spear and Harrison, 1989; Harrison et al., 1989; Wintsch et al., 2003). This lack of a Gondwanan cooling record in the exposed bedrock indicates the source region for the Stockton Formation is either buried under the modern central ENAM passive margin or was located at a higher structural level than the exposed source region and thus eroded into more distal Triassic–Jurassic ENAM basins or the Cretaceous–Cenozoic passive continental margins.

The Passaic Formation, the middle syn-rift strata of the Newark Basin, shows a DA U-Pb age spectrum with a dominant Alleghanian (230–340 Ma) component and minor Acadian (340–410 Ma), Taconian (410–500 Ma), and Pan-African (500–700 Ma) components. The DZ age spectrum of the Passaic Formation exhibits dominant components of Pan-African (500–700 Ma) and Grenville (900–1300 Ma) and minor Alleghanian (250–340 Ma), Acadian (340–410 Ma), Taconian (410–500 Ma), Granite-Rhyolite province (1300–1500 Ma), and recycled craton (1500–3500 Ma) components. The contrasting DZ age components of the Passaic Formation relative to the underlying Stockton Formation in conjunction with the minimal DA age components >340 Ma suggests a different source area for the middle syn-rift strata relative to the early syn-rift strata. The middle syn-rift strata have a mix of peri-Laurentian and peri-Gondwanan source regions. The DA results indicate these source regions were overprinted by medium-temperature heating during the Alleghanian orogeny.

The samples from the early syn-rift Stockton Formation and the late syn-rift Feltville Formation have nearly identical DZ age spectra but exhibit markedly different DA U-Pb age spectra (Fig. 8). The Feltville Formation contains a large 230–340 Ma component associated with Alleghanian cooling and minor 340–410 Ma and 410–500 Ma components associated with Acadian and Taconian cooling, respectively. In comparison to the early syn-rift Stockton Formation, the Feltville Formation comprises only a few 500–700 Ma apatite recording cooling related to Pan-African orogeny (Fig. 8). This contrast between the DZ and DA age spectra of the early and late syn-rift rocks of the Newark Basin points to two different potential source regions with identical Gondwanan DZ age spectra or differential spatial unroofing of a single source region. The different potential source regions would include a peri-Gondwanan terrane with Pan-African crystallization with or without subsequent Alleghenian, Acadian, and Taconian tectono-thermal overprinting. The Avalonia terrane, to the northeast of the Newark Basin, is a potential source for one of these age spectra given that it contains similar zircon spectra to Stockton and Feltville Formations and ⁴⁰Ar/³⁹Ar analyses indicate cooling during the Paleozoic Laurentian orogenies (e.g., Sutter et al., 1985; Spear and Harrison, 1989; Harrison et al., 1989; Wintsch et al., 2003).

The Towaco Formation shows mainly DA U-Pb components related to cooling associated with Alleghanian, Acadian, and Taconian orogenic events (Fig. 8). The major Taconian DA component, coupled with large Grenville and Granite-Rhyolite province DZ components, is indicative of paleodrainages sourced from the Alleghanian orogen and foreland and Grenvillian basement rocks that dominate the western source regions of the Newark Basin. The presence of Alleghanian and Acadian DA and DZ components and only minor Pan-African DZ components is also consistent with the continued transport of sediment from the source regions on the footwall of the Newark Basin. These findings demonstrate that DA U-Pb provenance analyses can provide critical additional insights for unraveling and differentiating potential source regions that may be ambiguous from DZ U-Pb provenance data. DA U-Pb analyses are particularly useful for rifted margins with a complex orogenic past with abundant zircon recycling and multiple orogenic events, such as the central ENAM of the eastern U.S.

Extensional faulting during the earliest rifting—the diffuse stretching phase—is characterized by distributed normal faulting in the brittle upper crust commonly reactivating inherited crustal weakness zones (e.g., Schlische, 1993; Gupta et al., 1999; Lavier and Manatschal, 2006; Manatschal et al., 2015; Fig. 12). Along the central ENAM, numerous Paleozoic suture zones and contractional brittle faults were reactivated to form the basin-bounding normal faults and hard-linking transfer faults during initiation of rifting in the Late Triassic (Fig. 1; Lindholm, 1978; Ratcliffe et al., 1986). The early syn-rift strata (TS 1) of the Newark Basin indicate the bulk of sediment was sourced and transported into the basin from hanging-wall source areas by transverse braided and meandering fluvial systems. This early syn-rift sedimentation is confined to small half and full grabens and sourced locally from the adjacent footwall and hanging-wall areas likely controlled by the pre-rift topography, similar to other fossil rift margins (Gupta et al., 1999; Hart et al., 2016; Domènech et al., 2018).

During progressive thinning and strain localization, the formation of a major border fault system, crustal necking, and asthenospheric upwelling drive major uplift and erosion both along the rift margin and within the central portion of the rift (Haupert et al., 2016; Fig. 12). Sediments may be shed from this central keystone block (“H-block” of Lavier and Manatschal, 2006), forming an emergent and eroding topographic high, into the rapidly subsiding adjacent necking basins and potentially even spill into the proximal basins (e.g., Haupert et al., 2016; Decarlis et al., 2017a, 2017b). During middle syn-rift (TS 2) deposition in the Newark Basin, lithofacies shifted to mudstone, siltstone, and minor sandstone in shallow ephemeral lakes dominating the landscape (Olsen, 1980; Smoot, 1991). Our new DZ U-Pb results demonstrate that a dramatic paleodrainage reconfiguration occurred during the middle syn-rift phase within the Newark Basin. Sedimentary basin fill was predominantly sourced and delivered from the eroding rift shoulder to the west of the Newark Basin via large alluvial and fluvial systems with entry points at the northern and southern tips of the basin (e.g., Hammer Creek Formation). The continued subsidiary input of Pan-African DZ components suggests the persistence of a paleohigh or more distal footwall block crest to the east of the Newark Basin. This together with the buried syn-rift deposits of the New York Basin to the east of the Newark Basin implies the possibility of the more distal rift basins trapping sediment from the central portion of the rift during the necking phase.

A complicating spatial factor during the early syn-rift diffuse stretching and necking phases results from along-strike rift segmentation and the formation of accommodation zones leading to a complex mixing of sediment input from local relay ramps and regional sediment eroded off of paleohighs of both the hanging wall and footwall. The middle syn-rift strata of the Newark Basin show this complexity with samples dominated by a Grenville (900–1300 Ma) age component while other middle syn-rift samples show a large Pan-African (500–700 Ma) age component (Figs. 6J–6M). Furthermore, large fluvial fans of the Hammer Creek Formation interfingering within the finer-grained lacustrine deposits of the Passaic Formation suggest episodic sediment input was transported along a relay ramp located in the “narrow neck” region that connects the Newark Basin to the Gettysburg Basin to the south.

During progressive rifting and hyperextension, normal faulting becomes more asymmetric, and crustal mechanical coupling and thermal weakening caused by asthenospheric upwelling results in the dismemberment of the central rift block that formed during necking. Structural dismemberment of the central rift block results in rapid basin subsidence of this block and a flip in sediment sourcing direction (Unternehr et al., 2010; Decarlis et a., 2017a). During coupling and the transition to the exhumation phase, the rift acquires the upper plate–lower plate configuration as one of the conjugate detachment faults becomes dominant and magmatic underplating is concentrated below the upper-plate margin (e.g., Huismans and Beaumont, 2002; Lavier and Manatschal, 2006; Brune et al., 2014). However, the sudden input of excessive magmatism and associated heating can abort continental and lithospheric thinning and hyperextension and trigger magmatic breakup prior to the mantle exhumation phase, resulting in the formation of a magmatic margin superimposed on an early magma-poor rifted margin (e.g., Reston, 2009; Tugend et al., 2020; Harkin et al., 2020; Sapin et al., 2021). Along the ENAM, voluminous CAMP magmatism resulted in effusive intra-basin lava flows, widespread dike emplacement, and mafic lower-crustal underplating, driving both regional doming and local uplift of the upper plate, as seen along other magma-rich margins, such as the North Atlantic or southern portion of the South Atlantic (Geoffroy, 2005; Saunders et al., 2007). Our Late Jurassic syn-rift DZ data from the Newark Basin and their transition to a multi-modal DZ age spectra characterized by a mix of Laurentian and Gondwanan ages nicely illustrates this syn-CAMP regional uplift and erosion of the proximal ENAM.

Multiple magnetic anomalies are identified along the ENAM (East Coast magnetic anomaly, Blake Spur magnetic anomaly, and M-25 magnetic anomaly and are considered to be post-CAMP and interpreted to represent the magmatic breakup and rift-to-drift transition (e.g., Shuck et al., 2019; Lang et al., 2020). After the initiation of seafloor spreading at ca. 165 Ma, post-tectonic cooling of the lithosphere caused subsidence, and sediments were routed from the rift flanks toward the center of the margin, forming a thick sedimentary wedge of the Central Atlantic passive continental margin (e.g., Steckler et al., 1988; Klitgord et al., 1988).

New detrital zircon and apatite U-Pb results provide an ~30 m.y. record of syn-rift provenance and related shifts in paleodrainages and topography during Triassic to Jurassic progressive rifting within the Newark Basin. Two major paleodrainage reconfigurations occurred in the Newark Basin during progressive rifting. During the Carnian to early Norian, the early syn-rift (TS 1) deposition was dominated by braided and meandering fluvial systems of the Stockton Formation and lacustrine deposition of the Lockatong Formation. Sediment was transported from eastern hanging-wall sources with a peri-Gondwanan affinity. The paleodrainages in the early syn-rift phase are likely a result of the pre-rift architecture and paleotopography of the U.S. ENAM. From the middle to late Norian, the middle syn-rift (TS 2) phase was dominated by shallow lacustrine deposition. The sediment was transported from a mix of sources with Laurentian and peri-Gondwanan affinities. Previous paleocurrent analyses indicate an axial drainage system existed within the Newark Basin during the middle syn-rift phase. This paleodrainage reconfiguration is likely a result of rift flank uplift and necking. During the earliest Jurassic, the late syn-rift (TS 3) phase was dominated by lacustrine deposition and marginal fluvial deposition interbedded with CAMP-related basaltic flows. At the Triassic-Jurassic boundary, deposition was dominated by peri-Gondwanan source affinities, while the Early Jurassic was dominated by a mix of peri-Gondwanan and Laurentian source affinities before transitioning to fully Laurentian source affinities. This is likely a result of local and regional uplift due to the emplacement of CAMP magmatism and the disruption of the local paleodrainages proximal to the Newark Basin.

New DA U-Pb results indicate DZ provenance data sets alone potentially miss key provenance shifts or unroofing signatures due to the prevalence of orogenic recycling of DZ components. These results also demonstrate the power of DA U-Pb analyses in unraveling provenance shifts, potentially missed in DZ data, due to the ability to recognize both high- and medium-temperature tectono-magmatic events.

Surface evolution and sediment dispersal track deeper crustal and lithospheric processes driving dynamic syn-rift topography during progressive rifting. In the early diffuse stretching phase, pre-rift topography mainly controlled sediment dispersal. During continued progressive thinning and necking, the combination of rift flank uplift and the formation and deconstruction of a central keystone drove sediment routing in the proximal basins. During deposition of the latest preserved syn-rift strata, widespread dike emplacement, volcanic flows, and magmatic underplating drove regional and local uplift. This uplift and the blockage of drainages controlled the sediment dispersal in the late syn-rift phase. In addition to these complexities, the 3-D rift evolution likely had a control on the sediment routing of coarse-grained siliciclastics into the proximal basins of the U.S. ENAM.

The project was funded by a Geological Society of America Graduate Research Grant, two American Association of Petroleum Geologists Foundation Grants-in-Aid awards, two Jackson School of Geosciences off-campus research awards, and funds from the Chevron (Gulf) Professorship to Stockli. We thank Chuck Bailey, Hope Duke, Kinsey Wilk, Tyler DeCourt, Austin Riopel, and Charles P. Baril for assistance in the field; Paul Olsen, Dennis Kent, and Jim Browning for their assistance with Newark Basin core acquisition; and Lisa Stockli, Federico Galster, and Rudra Chatterjee for their assistance with data collection and reduction. We thank Amy Weislogel, Paul Olsen, Brian Horton, Harm Van Avendonk, Kelly Thomson, Margo Odlum, Alex Johnson, Doug Barber, Andrew Parent, and Brandon Shuck for their thoughtful discussions on this work. We also thank Christopher Spencer, Brian Romans, and Bill Craddock for their insightful comments and reviews of this work.