The “Passive” Margin of Eastern North America: Rifting and the Influence of Prerift Orogenic Activity on Postrift Development


 We have analyzed and synthesized geologic and geophysical data from the onshore Newark rift basin and adjacent onshore and offshore basins to better understand the Mesozoic development of the eastern North American rift system and passive margin. Our work indicates that rifting had three phases: (1) an initial, prolonged phase of extension and subsidence; (2) a short-lived phase with higher rates of extension and subsidence, intrabasin faulting, and intense magmatism; and (3) a final phase with limited subsidence and deposition. Additionally, our work shows that anomalous uplift and erosion, associated with crustal-scale arching/warping subparallel to the prerift and syn-rift crustal fabric not the continent-ocean boundary, affected a region landward of the basement hinge zone. Uplift and erosion began during the final rifting phase and continued into early drifting with erosion locally exceeding 6 km. Subsequent subsidence was minimal. We propose that denudation unloading related to relic, prerift orogenic crustal thickness and elevated topography triggered the anomalous uplift and erosion. After the Paleozoic orogenies, postorogenic denudation unloading (cyclic erosion and isostatic rebound/uplift) significantly thinned the thickened crust and reduced topographic elevation. During rifting, extension stretched and tectonically thinned the crust, promoting widespread subsidence and deposition that dampened the postorogenic cycle of erosion and isostatic rebound/uplift. During the rift-drift transition, with extension focused near the breakup site, denudation unloading resumed landward of the basement hinge zone, producing significant erosion and uplift (related to isostatic rebound), crustal thinning, and topographic decay that left behind only eroded remnants of the once massive rift basins.


Introduction
The classic model of passive-margin development (e.g., [1]) has two distinct stages. During the initial rifting stage, regional extension thins the continental lithosphere, producing subsiding rift basins. Breakup occurs and the drifting stage begins, characterized by focused extension at seafloorspreading centers and broad thermal subsidence on the conjugate passive (rifted) margins that surround the site of breakup. On many passive margins, a significant episode of uplift, not predicted by simple lithospheric stretching, develops during the transition from rifting to drifting (e.g., [2]). The resultant unconformity, referred to as the breakup unconformity by Falvey [3] or the postrift unconformity by Klitgord et al. [4], separates eroded prerift and syn-rift rocks from overlying postrift rocks. Researchers have proposed a variety of mechanisms to explain the development of a breakup/postrift unconformity, including contributions from depth-dependent stretching with excess thinning of the lithospheric mantle near the site of breakup (e.g., [5]), lateral conductive heat transfer from the site of breakup (e.g., [6]), a dynamic response to secondary small-scale convection triggered by rifting and breakup (e.g., [7,8]), deep lithospheric necking during rifting and breakup (e.g., [9]), magmatic underplating during rifting and breakup (e.g., [10,11]), flexural uplift caused by mechanical unloading during rifting and breakup [12], peripheral bulging that accompanied offshore postrift subsidence and deposition [13], phase transitions associated with density reductions within the lithospheric mantle during rifting and breakup (e.g., [14]), and delamination and/or dripping of the lower lithospheric mantle near the site of breakup [2].
(3) Its development was followed by substantial subsidence, locally exceeding 10 km near the continent-ocean boundary. Several of the proposed mechanisms for the development of a breakup/postrift unconformity, listed above, likely contributed to the development of the PRU on the ENAM passive margin. For example, depth-dependent stretching or delamination and/or dripping would produce a more transient phase of uplift and erosion surrounding the site of breakup followed by significant subsidence, whereas deep lithospheric necking, flexural uplift caused by mechanical unloading, or peripheral bulging would produce a prolonged phase of erosion near the basement hinge zone flanking the site of breakup.
To further understand the characteristics and causes of uplift on the ENAM margin, as well as other passive margins, we have analyzed and synthesized geologic and geophysical data from the onshore Newark rift basin (the largest and most thoroughly studied basin in the central segment of the ENAM rift system), the adjacent onshore and offshore rift basins, and the northern end of the postrift Baltimore Canyon trough (Figures 1-5) As discussed below, our work shows that a distinctive style of uplift and erosion affected the study area landward of the basement hinge zone. The uplift and erosion were associated with broad, crustal-scale arching/warping, extending well landward of the basement hinge zone and subparallel to the prerift and syn-rift crustal fabric, not the continent-ocean boundary. Erosion was long-lived, beginning during the waning phase of rifting and continuing into the early drifting stage, and locally exceeded 6 km. Subsequently, postrift subsidence was minimal. Based on these characteristics, we suggest that the cause(s) of the uplift and erosion landward of the basement hinge zone differed from those of the PRU. Specifically, we propose that denudation unloading associated with relic crustal thickening and elevated topography created during prerift Paleozoic orogenic activity contributed to the late syn-rift and early postrift uplift and erosion observed landward of the basement hinge zone.

ENAM Margin
The structure and composition of the ENAM lithosphere reflects a long and complex tectonic history including the assembly, rifting, and breakup of Rodinia; the development of the Laurentian passive margin during the early Paleozoic; and the subsequent Paleozoic orogenic activity that culminated in the assembly of Pangea (e.g., [20][21][22][23][24][25][26][27]). The final orogenic event in the study area, the Alleghanian orogeny, was completed by the Permian (~270 to 260 Ma) (e.g., [22,27]). Several postorogenic processes affected the ENAM lithosphere before the onset of rifting (~270 to 230 Ma). Gravitational collapse triggered by lithospheric delamination may have altered the structure and composition of the ENAM lithosphere (e.g., [28]). Also, a postorogenic cycle of erosion and isostatic rebound (denudation unloading) initiated by erosion-related transfer of mass away from the Appalachian Mountains significantly thinned the crust and reduced topographic elevation before rifting (e.g., [29,30]).
By the Late Triassic, 30 to 40 million years after the assembly of Pangea, a massive rift zone associated with NW-SE extension had developed within the former orogenic belt (e.g., [15][16][17][18][19]). The rift system preserved on the ENAM margin is hundreds of kilometers wide and composed of numerous parallel and en-echelon rift basins ( Figure 1). Prerift deformation affected the geometry, location, and distribution of the ENAM rift basins. For example, the strike and dip of the border-fault zones of most ENAM rift basins mimic the attitude of the prerift crustal fabric (e.g., [19,[31][32][33][34]) ( Figure 1). Additionally, most ENAM rift basins are offset from the continent-ocean boundary by tens to hundreds of kilometers ( Figure 1). Numerical modeling studies [35,36] suggest that offset rift basins, like the ENAM rift basins, are more likely to develop when weak crustal layers decouple laterally offset, preexisting zones of weakness within the upper crust from those within the upper mantle. The resulting distribution of extension is depth-dependent. Extension is broadly distributed within the continental crust, producing a wide zone of offset rift basins, whereas extension is more focused within the lithospheric mantle near the site of breakup.
The cessation of rifting (and presumably the rift-drift transition) was diachronous, occurring first in the southern segment of the ENAM rift system (latest Triassic), then in the central segment (Early Jurassic), and finally in the northern segment (Early Cretaceous) (e.g., [19]) (Figure 1), agreeing with the quantitative plate reconstructions of Schettino and Turco [41]. The exact timing of breakup and the early seafloor-spreading history, however, remain controversial (e.g., compare the plate reconstructions of [41][42][43][44]). The ENAM margin, from Florida to southern Nova Scotia, is magma-rich, characterized by a wedge of seaward-dipping reflectors (SDRs) near the continent-ocean boundary (Figure 3(c)). The SDRs, presumably of volcanic or volcaniclastic origin, formed during the rift-drift transition and 2 Lithosphere are associated with the East Coast Magnetic Anomaly (ECMA) (e.g., [45][46][47]) ( Figures 1 and 3(c)). The remainder of the margin, from northern Nova Scotia to the Grand Banks, lacks SDRs and is, thus, considered magma-poor (e.g., [48,49]). A series of broad postrift basins (e.g., the Baltimore Canyon trough, Figure 3(c)), associated with tectonic subsidence and sediment loading, developed after rifting along the ENAM margin mostly seaward of the basement hinge zone. As mentioned previously, a regional, time-transgressive unconformity, termed the postrift unconformity (PRU), separates the eroded prerift and syn-rift rocks from the overlying postrift strata along the ENAM passive margin (e.g., [4]) (Figures 3(b) and 3(c)).

Study Area
The study area encompasses the onshore and offshore parts of the central segment of the ENAM margin (Figures 1-3). It includes the onshore Newark rift basin, other onshore rifts basins (i.e., the Connecticut Valley and Pomperaug basins), offshore rift basins on the western end of the Long Island platform (i.e., the New York Bight and Long Island basins), and the northern end of the postrift Baltimore Canyon trough. Additionally, potential-field studies and limited seismic-reflection data suggest that additional ENAM rift basins exist beneath the coastal-plain sediments of the study area (e.g., [4,50,51]).
The continent-ocean boundary, as indicated by the ECMA, has a distinct bend in the study area, trending approximately NE-SW in the south and ENE-WSW in the north ( Figure 2). The basement hinge zone largely parallels the continent-ocean boundary throughout the study area (e.g., [52]) ( Figure 2). The onshore Newark rift basin, located about 200 km to the northwest of the continent-ocean boundary, generally trends NE-SW, parallel to preexisting Paleozoic zones of weakness ( Figure 2). The onshore Connecticut Valley rift basin and the offshore New York Bight and Long Island rift basins trend N-S, parallel to the preexisting Paleozoic fabric (e.g., [53,54]) but highly oblique (~70°) to the basement hinge zone and continent-ocean boundary to the south and southeast ( Figure 2). Thus, the trends of prerift and syn-rift structures differ substantially from those of the major breakup-related crustal features (i.e., the basement hinge zone and continent-ocean boundary) in the northeastern part of the study area.  [60]). Box shows the study area. The cessation of rifting (and presumably the onset of breakup) was diachronous, occurring first in the southern segment (latest Triassic), then in the central segment (early Early Jurassic), and finally in the northern segment (Early Cretaceous) (e.g., [19,60]).

Newark Rift Basin
As mentioned previously, the Newark rift basin is the largest and most thoroughly studied rift basin in the central segment of the ENAM rift system (Figures 1-5). Seismic-reflection, core, borehole, and vitrinite-reflectance data, as well as abundant field exposures, provide critical information about its geologic history and the tectonic evolution of the central segment of the ENAM margin (e.g., [17,[55][56][57][58][59][60]).
The Newark rift basin is asymmetric, bounded on the northwest by a series of NE-striking, SE-dipping, rightstepping faults with normal separation that together form the basin's border-fault system (Figures 4(a), 4(b), and 5(a)). Many of these faults are reactivated mylonite zones (e.g., [33,61]) that developed during the various Appalachian (and possibly older) orogenies that preceded rifting. Generally, the syn-rift strata dip 10 to 20°NW toward the borderfault system (Figures 4(a) and 4(b)). Several intrabasin faults,  [50,52,53,60,104,125,126]). Potential-field studies and limited seismic-reflection data suggest that additional ENAM rift basins might exist beneath the coastal-plain sediments in the southwestern part of the study area (e.g., [4,50,51]). A well (S34, black dot) encountered interpreted syn-rift Triassic strata between the Connecticut Valley and New York Bight rift basins [111]. Bold blue lines are regional transects shown in Figure 3. Thin, light blue lines show locations of sections in Figure 4. 4 Lithosphere with normal separation in cross-sectional view, cut the synrift strata and link with the border-fault system ( Figure 5). Structural analyses using fault/fracture orientations, slip measurements, CAMP-related dike orientations, and early seafloor-spreading directions (e.g., [16,17,19,32,62]) suggest that the extension direction was approximately NW-SE during rifting. A series of anticlines and synclines, commonly called transverse folds, also deform the syn-rift strata in the hanging walls of the border-fault system and intrabasin faults (e.g., [59,[63][64][65]) ( Figure 5(a)). Although subtle thickness changes suggest that some of this folding occurred during syn-rift deposition, possibly as a result of along-strike variations in fault displacement (e.g., [59,64]), structural studies indicate that much of this folding formed in response to left-lateral, strike-slip deformation during the early stages of drifting (e.g., [60,[66][67][68]). The deposition of the nonmarine (predominantly fluvial and lacustrine) Stockton, Lockatong, and Passaic formations occurred during the main rifting phase, about 230 to 201.5 Ma (Late Triassic) (e.g., [56,57,69,70]). During the main rifting phase, the width of the Newark rift basin steadily increased from less than~25 km during the deposition of the  [53,115]. (c) Southern transect crossing southern Newark rift basin and northern Baltimore Canyon trough. Based on seismic line NB-1 [60] and interpretation of USGS seismic line 25 [52,127] and data from LASE Study Group [131]. Note the highvelocity lower crust interpreted as intruded/underplated continental crust and/or initial oceanic crust [132]. Prerift fabric and lower crustal laminations between seismic lines NB-1 and USGS 25 are based on seismic-reflection data from Sheridan et al. [51]. SDRs are seaward-dipping reflectors, and ECMA is the East Coast Magnetic Anomaly.

Lithosphere
Stockton Formation to more than~100 km during the deposition of the Passaic Formation [60]. By the end of the main rifting phase, a through-going, border-fault system had developed with locally more than 8 km of syn-rift sediment/sedimentary rock preserved in its hanging wall [60].
CAMP-related magmatic activity signaled the onset of the late rifting phase, about 201.5 Ma (latest Triassic to earliest Jurassic). The intrusion of CAMP-related igneous sheets and dikes, the extrusion of CAMP-related basalt flows (i.e., the Orange Mountain, Preakness, and Hook Mountain basalts), and the deposition of the predominantly fluvial and lacustrine sediments between and above the flows (i.e., the Feltville, Towaco, and Boonton formations) characterize the late rifting phase. Deposition during the late rifting phase likely ended by~198 Ma, based on published sedimentation rates and chronostratigraphy of the youngest extant syn-rift Jurassic formations [71], age dates of the CAMP-related lava flows [38], and estimates of missing syn-rift Jurassic overburden [55,72]. By the end of the late rifting phase, the northern end of the Newark rift basin likely connected with the   Previous analyses of vitrinite-reflectance data indicate that, after syn-rift deposition, the Newark rift basin underwent differential erosion, ranging from 1 km in the northwest to 6 km in the east and southeast [55,60]. Restorations using these erosion estimates together with seismic-reflection, core, borehole, and field data show that the NW tilting, intrabasin faulting, and transverse folding that affect the syn-rift strata in the Newark rift basin occurred both during and after syn-rift deposition [60]. Based on the age of the oldest, relatively flat-lying postrift strata that overlie the southeastern edge of the Newark rift basin (i.e., earliest Late Cretaceous; [73]), most of the northwest tilting, intrabasin faulting, and transverse folding had ceased before the earliest Late Cretaceous. How much of this deformation occurred during the late rifting phase and CAMP-related magmatism? How much occurred later during breakup or the early stages of drifting? Did the magnitudes and rates of tilting, extension, and subsidence vary spatially and/or temporally during and after rifting? What mechanism(s) produced the tilting and erosion? To address these critical questions, we have constructed and restored three representative cross sections from the Newark rift basin, incorporating new seismic-reflection data and sonic transit-time analyses.

Construction and Restoration of Cross Sections
To construct and restore the three cross sections through the Newark rift basin, we used the seismic, field, core, borehole, and vitrinite-reflectance data presented in Withjack et al. [60]. Additionally, we performed compaction-based, sonictransit-time analyses for 11 wells with wireline-log data from the Newark rift basin (locations in Figure 5(b) and Table 1). This included seven wells from the Newark Basin Coring Project (NBCP) (e.g., [57,74]), two wells associated with the TriCarb Consortium for Carbon Sequestration in the northern Newark basin (New York State Thruway Authority Tandem Lot#1, TriCarb Well#4) (e.g., [58,75]), and two industry wells in the southern Newark basin (Cabot KB#1, Parestis#1). The wireline-log data for these wells are available from Lamont Doherty Earth Observatory, the NY State Geological Survey, and the Pennsylvania Geological Survey, respectively. The results of the sonic transit-time analyses allowed us to independently validate and constrain the results of the prior vitrinite-reflectance analyses and to estimate the magnitude of exhumation in locations without vitrinitereflectance data. Finally, we included recently published seismic-reflection data from the northern end of the Newark rift basin associated with the TriCarb Consortium for Carbon Sequestration (e.g., [58]) (Figure 4(a)). Below, we briefly describe the vitrinite-reflectance analyses, the sonic-transittime analyses, the interpretation of the seismic-reflection data, and the restoration approach.

Vitrinite-Reflectance Analyses.
With a constant, uppercrustal geothermal gradient, temperature increases linearly with depth; similarly, the logarithm of the percent vitrinite reflectance (% R o ), a diagenetic to very-low-grade organic metamorphic indicator, will increase linearly with depth [76]. We determined the amount of eroded syn-rift section for the Newark rift basin using the method of Dow [76], which extrapolates the regressed line of a semi-logarithmic borehole vitrinite-reflectance-vs.-depth profile back to % R o = 0:2, the percent reflectance of recently deposited woodyplant matter or low-grade peat [77]. The difference between the depth intercept at 0.2% reflectance and the current ground surface, or erosional unconformity in question, is the estimate of the missing section. Recently, Nielsen et al. [78] suggested that the increase in vitrinite reflectance with depth may not be log-linear but bend or decrease in slope around 0.7% R o . In the Newark rift basin and other ENAM rift basins (i.e., the Taylorsville, Richmond basins), however, vitrinite-reflectance profiles, not affected by transient heat flow or contact metamorphism, are log-linear with depth through 0.7% R o [55,[79][80][81]. Therefore, the application of the Dow method is appropriate for this study. An important assumption of the Dow method is that the reflectance-depth gradient is the same in the missing section as in the extant section. A major advantage of the Dow method is that the missing-section estimate depends only on the reflectance and depth data, not on assumptions of paleogeothermal gradient or thermal modeling which can introduce additional uncertainty. Malinconico [55,72] and Withjack et al. [60] provide downhole and surface mean-random vitrinite-reflectance (% R o ) data for the Newark rift basin, descriptions of sample locations, details of sample preparation, reflectance-measurement methods, and thermal history. First, we calculated the missing-overburden estimate for the Newark basin with the Dow method using the downhole, vitrinite-reflectance data in the Lockatong Formation of the Nursery core (location in Figures 5(b) and 6(a), and Table 1), the longest log-linear vitrinite-reflectance vs. depth gradient interval in any of the seven NBCP cores. Other NBCP cores had either shorter linear-reflectance runs, few if any vitrinite-bearing strata, or anomalous high-reflectance spikes due to transient fluid flow. The calculated, reflectance-based, paleogeothermal gradient from the Nursery core is similar to those from a short, log-linear section of the Princeton core and from a surface dip-section through the Jurassic strata of the northern Newark basin. Thus, a basin-wide application of the vitrinite-reflectance-vs.-depth gradient of the Nursery core is reasonable [55,72]. To calculate the amount of eroded syn-rift strata throughout the Newark rift basin, we applied the reflectance-depth gradient of the Nursery core to a reflectance point in four other NBCP wells and in the Cabot KB#1 well (locations in Figure 5(b) and Table 1) and to surface reflectance points scattered throughout the basin (locations in Figure 5(b)). We evaluated only reflectance points considered unaffected by contact metamorphism or transient fluid flow.
We calculated the uncertainty associated with the linear regression of the downhole, vitrinite-reflectance data using the parametric bootstrap [82], a statistical resampling technique [79]. The calculated missing syn-rift section at the Nursery site was 4.9 km, with an uncertainty of ±1.4 km, 8 Lithosphere approximately ±30% of the estimated missing section ( Figure 6(a)). Because all reflectance-based erosion estimates for the Newark basin are based on the same Nursery core reflectance data and bootstrap calculations, the uncertainty at any one location is also ±30%, and a positive uncertainty at any one location correlates to positive uncertainty at all other sites. The ±30% uncertainty compares favorably with the percent uncertainty in previous studies of other basins using maturity data for estimated erosion calculations: 18%-60% (Dow method, [83]); 6%-37% (Dow method, [84]); and 32%-83% (thermal history reconstruction using vitrinite-reflectance and apatite fission-track data, [85]).

Sonic Transit-Time Analyses.
The sonic transit-time technique is a compaction-based approach for estimating exhumation magnitudes with many successful applications worldwide (e.g., [86][87][88][89][90]). Corcoran and Doré [91] provide a detailed overview of the technique, including a discussion of its strengths and limitations. The sonic transit-time technique relies on the following assumptions: (1) sonic transit time is a reasonable proxy for porosity, (2) porosity of sediment/sedimentary rock decreases exponentially with depth, (3) mechanical and thermochemical compaction during burial produces the porosity reduction, and (4) porosity reduction is irreversible. The technique has two key steps. The first involves defining a lithology-specific relationship between sonic transit-time values and depth (called a normal-compaction trend, NCT) for the basin. The second involves calculating the depth at which observed, presentday sonic transit-time values would overlie the NCT, reflecting the maximum burial depth. The difference between the calculated maximum burial depth and the current depth yields an estimate for the net exhumation. Shales and mudstones are the preferred lithologies for sonic transit-time analyses, because their grain size and mineralogy are generally more homogeneous and their NCTs are generally simpler than those of coarser-grained sediments/sedimentary rocks (e.g., [89]). Fluvial and lacustrine shales occur at all stratigraphic levels throughout the Newark rift basin (e.g., [57]). They are most commonly illitic in composition, although the exact clay-mineral assemblage can vary depending on the depositional and diagenetic environment (e.g., [92][93][94]). To identify the shales in the 11 wells with wireline-log data from the Newark rift basin (locations of borehole and NBCP sites in Figures 5(b) and Table 1), we conducted a standard petrophysical evaluation (i.e., gamma-ray, density, neutron-porosity, and deep-resistivity) using the Paradigm® Geolog® Formation Evaluation suite. To minimize the potential of mistaking highly radioactive feldspathic sandstones for shales, we used gamma-ray logs and density-neutron logs, together, to calculate shale volume (VSH). Additionally, the calculations of VSH allowed us to normalize wireline-log data acquired at different times by different vendors. To isolate the shales in the 11 wells, our analyses only included sonic transit-time values where the calculated shale volume exceeded 90%. To remove logging artifacts (e.g., cycle skipping, noise triggering), we applied a smooth median filter, and then we resampled the log data at 3.05 m-step (10 ft-step) increments.
Generally, the exponential function of a normal compaction trend (Δt NCT ) has the form: where z is depth, Δt 0 is the sonic transit time at deposition, Δt MA is the sonic transit time of the mineral matrix (i.e., the sonic transit time where porosity goes to zero as depth approaches infinity), and b is the exponential decay constant (e.g., [89,90]). Using relationships between neutron porosity and sonic transit time, the petrophysical analyses of the wireline data from the 11 wells in the Newark rift basin indicated that Δt MA ranges from 144 to 154 μs m -1 . Because all wells in the Newark rift basin exhibit abnormal compaction, we were unable to directly calculate the initial transit time (Δt 0 ). Observations from other basins show that Δt 0 typically ranges from 640 to 673 μs m -1 (e.g., [89,90]). We, therefore, created four possible scenarios for the NCT for the Newark  Figure 5(b) and Table 1 for well-site locations.
rift basin using the high and low reported values for Δt 0 (640 and 673 μs m -1 ) and the minimum and maximum calculated values of Δt MA (144 and 154 μs m -1 ). We used the sonic transit times from the Martinsville #1 well, which has the slowest sonic transit times and the longest depth span of the sonic transit-time data (i.e., nearly 1 km), to determine the value of b for the four scenarios. Based on a log-linear transformation of Eq. (1), the value of b for the four scenarios ranged from 0.0004298 to 0.0004828 m -1 , similar to reported values from other basins (e.g., [87,89]). Together, this information yielded four potential NCTs for the Newark rift basin ( Figure 6(b)). Comparisons of the present-day sonic transittime values in the 11 wells with those of the potential NCTs allowed us to calculate the potential range of the maximum burial depths and, thus, the exhumation magnitudes for the 11 wells (Figures 6(c) and 6(d)). The total uncertainty related to the uncertainties in the values of Δt 0 and Δt MA and in the spread in the sonic transit-time vs. depth data is less than ±10% (Figures 6(c) and 6(d)).

Exhumation Estimates from Combined Vitrinite-
Reflectance and Sonic Transit-Time Analyses. Together, the vitrinite-reflectance and sonic transit-time analyses provide an independent check on the exhumation estimates for the Newark rift basin. Comparisons of the results from six wells from the basin show that exhumation estimates are highly consistent for the two techniques, differing by 1 km or less and having overlapping uncertainties ( Figure 6(d)). For example, the vitrinite-reflectance and sonic transit-time analyses predict about 4.9 km and 5.3 km of missing syn-rift section, respectively, for the Nursery well ( Figure 6(a)). This overall agreement between the results of the vitrinitereflectance and sonic transit-time analyses validates the initial use of the Dow method to estimate exhumation magnitudes in the Newark rift basin [55,72]. Field relationships (i.e., thicknesses of stratal units based on their outcrop dips and widths) suggest that the statistically derived uncertainty of ±30% for the vitrinite-reflectance analysis is likely too large. For example, field relationships indicate that at least 4 km of the syn-rift section once covered the Lockatong Formation at the locations of the Nursery and the Princeton wells. Thus, the exhumation magnitude must be greater than 4 km at these locations ( Figure 6(d)). Also, at least 3.5 km of syn-rift section once covered the extant syn-rift section at the locations of the Rutgers, Somerset, and Cabot wells. Thus, the exhumation magnitude must be greater than 3.5 km at these locations ( Figure 6(d)). With a reduced uncertainty of ±15% for the vitrinite-reflectance analysis, exhumation estimates from the vitrinite-reflectance and sonic transit-time techniques are consistent with each other (i.e., they still have overlapping uncertainties) and with field relationships from the Newark rift basin.
The new erosion map for the Newark rift basin ( Figure 5(b)) is comparable to that presented in Withjack et al. [60], differing primarily in the northern part of the basin. The maximum erosion (>5 km) occurs in the southern part and on the southeastern side of the Newark rift basin. The minimum erosion (<1 km) occurs near the border-fault system in the central part of the basin. The magnitude of ero-sion is about 3 to 4 km in the northernmost part of the Newark basin. Contour lines, showing equal magnitudes of erosion, generally follow the strike of bedding in the southern and central parts of the basin, cut across the strike of bedding in the northernmost part of the basin, and are offset by the intrabasin faults. The thicknesses and geometries of the postrift Jurassic sedimentary rocks (based on offshore seismic data, e.g., [52,53]) and the younger coastal-plain deposits (e.g., [95]) to the southeast of the Newark rift basin suggest that any postrift section above the Newark rift basin was thin (<1 km). Thus, most, if not all, of the estimated eroded material was syn-rift strata.

Seismic-Reflection Data.
Seismic line NB-1, published by Bally et al. [96] and Withjack et al. [60], was acquired and processed by NORPAC Exploration Services in 1983. The time-migrated line trends NW-SE, across the southern part of the Newark rift basin (Figures 2 and 4(b)). It is located to the northeast of a major right-step in the border-fault system and crosses a transverse fold (the Ferndale structure) and a major intrabasin fault (the Flemington/Furlong fault). Our interpretation of the line (Figure 4(b)), displayed without vertical exaggeration assuming a velocity of 5 km/s based on seismic-velocity analyses, honors the seismic data and all available surface geology (e.g., locations and dips of formation contacts and major faults) and drill-hole data (e.g., [33,57]). The line shows that a major SE-dipping fault zone with normal separation bounds the basin on the northwest. The fault zone, characterized by a series of high-amplitude reflections, is relatively planar and dips~30°to the southeast. Using core data, Ratcliffe et al. [33] demonstrated that this fault zone is a mylonitic Paleozoic thrust fault reactivated during rifting. The syn-rift strata dip~10°to 15°toward the northwest. Near the basin-bounding fault zone, however, the strata are nearly flat-lying. This change in dip is associated with the transverse fold (the Ferndale structure) whose axial trace is parallel to the seismic line ( Figure 5(a)). The seismic data suggest that the Stockton Formation and an unexposed older unit (which laps onto Paleozoic prerift strata) gradually thicken toward the northwest (i.e., toward the border-fault zone). The change in bedding dip is~3°from the top to the bottom of the Stockton Formation.
Seismic line Sandia 101, acquired in 2011 as part of the project for the TriCarb Consortium for Carbon Sequestration [58], trends E-W, across the northern end of the Newark rift basin (Figures 2 and 4(a)). The seismic profile, processed by Conrad Geoscience Corporation [97], provides a depthmigrated image of the subsurface geometry of the basin. Our interpretation of the line honors the seismic data, the available surface geology, and drill-hole data from the New York State Thruway Authority (NTSTA) Tandem Lot #1 test well. The most prominent feature on the seismic line is the Palisades sill, appearing as a pair of subparallel, highamplitude reflections associated with the top and bottom of the intrusion. The sill, a CAMP-related diabase intrusion, has an irregular bowl-like shape, cuts across stratal reflections, has a lateral extent of more than 15 km, and a thickness of up to 500 m. The seismic data show that the syn-rift section thins subtly toward the east. 11 Lithosphere 5.5. Restoration Approach. The first step of the restoration involved the construction of the cross sections (Figure 7(a)). Field, well, and seismic data (i.e., our interpretations of seismic lines Sandia 101 and NB-1, Figures 4(a) and 4(b)) constrained the structural geometries for the northern (A-A ′ ) and southern (C-C′) cross sections. Field data alone defined the structural geometries for the central section (B-B′). Thus, the subsurface geometries for this cross section were less certain.
The next step of the restoration involved the replacement of the eroded syn-rift section. Using the exhumation estimates from the vitrinite-reflectance and sonic transit-time analyses (Figure 7(a)), together with the locations and dips of formation contacts, calculated formation thicknesses, and the presence of igneous intrusions, we replaced the eroded syn-rift section (Figure 7(b)). To better constrain the geometry of the eroded youngest syn-rift section near the border-fault system on the central and southern cross sections, we needed to remove the effects of the transverse folding. As mentioned previously, structural analyses suggest that most transverse folding formed in response to left-lateral strike-slip deformation during the early stages of drifting (e.g., [60,[66][67][68]). Many transverse folds, including those on the central and southern cross sections ( Figure 5), are found near right-steps or bends in the border-fault system that would have acted as restraining bends during the strike-slip deformation. To remove the effects of the transverse folding on these sections, we straightened the folded beds, extending them directly toward the projected location of the border-fault zone across the restraining step/bend. This represented the geometry of the syn-rift strata before the transverse folding (Figure 7(c)). To define the geometry of the eroded syn-rift strata today, we then refolded the syn-rift strata near the border-fault system, honoring the available field and seismic data and conserving crosssectional area with vertical shear (Figures 7(d) and 8(e)).
The remaining steps of the restoration involved the systematic removal of deformation, backwards through time, for four key events in the geologic history of the Newark rift basin: (1) an early stage of drifting (Figure 8(d)), (2) the final phase of rifting and/or breakup (Figure 8(c)), (3) the end of the late rifting phase (Figure 8(b)), and (4) the end of the main rifting phase (Figure 8(a)). Landscape-evolution models by Pazzaglia and Brandon [98] suggest that, at the end of rifting, the mean elevation of the study area was between~0.5 and 2 km above sea level. Thus, based on these estimates, we assumed that the depositional surface of the Newark rift basin was flat and about 1 km above sea level at the end of the late and main rifting phases. We also assumed that the footwall of the border-fault system was about 1 km above the depositional surfaces (e.g., [99]). Today, the restored depositional surfaces are tilted and locally more than 6 km above sea level, reflecting postdepositional northwest tilting and uplift (Figure 8(e)). The tilting, uplift, and erosion occurred before the deposition of the relatively flat-lying, postrift strata of earliest Late Cretaceous age that overlie the southeastern part of the Newark rift basin [73].
For the Early to Middle Jurassic restoration (i.e., an early stage of drifting, Figure 8(d)), we removed about a third of the northwest tilting and uplift, rotating the cross sections clockwise. Fission-track (sphene and zircon) studies from the southern Newark rift basin and the surrounding crystalline basement [100] suggest that the northwest tilting and uplift were regional, involving not only the Newark basin but also its border-fault system and footwall. For the Early Jurassic restoration (i.e., the final stages of rifting and/or breakup, Figure 8(c)), we removed an additional third of the northwest tilting and uplift. We also removed the transverse folding because, as stated previously, structural analyses suggest that most transverse folding occurred during the early stages of drifting (e.g., [67,68]). For the early Early Jurassic restoration (i.e., the end of the late rifting phase, Figure 8(b)), we flattened the top surface of the youngest syn-rift unit (to 1 km above sea level) by rotating the sections clockwise, removing the final third of the northwest tilting and uplift. For the southern cross section, we removed the postdepositional offset of the intrabasin fault by translating the hanging wall upward along the fault surface. Because the intrabasin faults link with the border-fault system of the Newark rift basin, the border-fault system was also likely active after syn-rift deposition. We assumed that the offset on the border-fault system was similar to that on the intrabasin faults, and we removed this offset from the border-fault zones on all sections. For the late Late Triassic restoration (i.e., the end of the main rifting phase, Figure 8(a)), we corrected for compaction (using OSX Backstrip, v. 2.9, [101]) and flattened the top surface of the Passaic Formation (to 1 km above sea level) by rotating the hanging wall of the border-fault system clockwise. We removed the offset on the border faults and intrabasin faults by translating the hanging wall upward along the fault surfaces.
We emphasize that these restorations, like most restorations, are approximate because of uncertainties in the current geometries of the syn-rift strata, the magnitude of exhumation from the vitrinite-reflectance and sonic transit-time analyses, the exact timing of northwest tilting and uplift, the exact magnitude and sense of slip on the faults, and the paleoelevation of the depositional surfaces. Despite these limitations, the restorations provide critical information and useful insights about the evolution of the Newark rift basin and the surrounding region.

Evolution of Newark Rift Basin
As discussed previously in Withjack et al. [60], the Newark rift basin widened through time during the main rifting phase (~230 to 201.5 Ma). By the end of the main rifting phase, the Newark rift basin was very wide (locally >100 km) with little, if any, intrabasin faulting (Figure 8(a)). It was asymmetric, bounded on the northwest by a major border-fault system with steeper dips in the north and more moderate dips in the south. The cumulative heave on the border-fault system (a proxy for extension) ranged from about 4 to 19 km, with significantly greater heave in the south (Figures 8(a) and 9(a)). Heave/extension rates during the main rifting phase averaged from about 0.1 km/My in the north to about 0.7 km/My in the south (Figure 9(a)). The thickness of basin fill near the border-fault system increased southward, 12 Lithosphere     14 Lithosphere ranging from about 4 km in the north to 9 km in the south (Figures 8(a) and 9(b)). Consequently, the compacted accumulation rates increased southward, ranging from 0.1 km/My in the north to 0.3 km/My in the south (Figure 9(b)), similar to values presented by Olsen and Kent [102]. The dips of the strata deposited during the main rifting phase gradually increase downward, reflecting northwest tilting toward the border-fault system (i.e., growth faulting) during the main rifting phase, reaching values of 5°NW near the base of the syn-rift section (Figures 8(a) and 9(c)). Thus, the overall rate of northwest tilting of the syn-rift strata was~0.1 to 0.2°/My during the main rifting phase (Figure 9(c)).
Tectonic activity in the Newark rift basin changed profoundly during the late rifting phase (201.5 to~198 Ma) (Figure 8(b)). For the first time, significant magmatic activity affected the Newark rift basin. CAMP-related igneous dikes, sills, and sheets intruded the basin and the surrounding basement rocks, and basalt flows episodically filled the basin (Figure 8(b)). The border-fault system remained active and, for the first time, major intrabasin faults developed, linking with the border-fault system and dissecting the basin. Heave/extension rates increased, reaching values of 1 km/My or more in the central and southern parts of the basin (Figure 9(a)). Syn-rift strata accumulated within the hanging walls of the border-fault system and the intrabasin faults (Figure 8(b)). Compacted accumulation rates increased dramatically during the late rifting phase with the highest rates in the central part of the basin (~1 km/My), reflecting a significant northward shift in the basin depocenter (Figures 8(b) and 9(b)). These compacted accumulation rates are similar to those determined by Olsen et al. [71] for the

Lithosphere
Early Jurassic Feltville, Towaco, and Boonton formations. The average rate of northwest tilting increased throughout the basin during the late rifting phase, reaching~1°/My in the central and southern parts of the basin (Figure 9(c)). In the central and northern parts of the Newark rift basin, basin width increased substantially (Figure 8(b)). By the end of the late rifting phase, the northern end of the Newark rift basin likely connected with the onshore Connecticut Valley rift basin and the offshore New York Bight rift basin [60] (Figure 2).
During the final phase of rifting, the linked border and intrabasin faults remained active, but little sediment accumulated within their hanging walls (Figure 8(c)). This signaled a change from widespread deposition during the main and late rifting phases to limited deposition and/or erosion during the final phase of rifting. Widespread northwest tilting, uplift, and erosion affected the Newark rift basin during the transition from rifting to drifting (Figures 8(c), 8(d), and 8(e)). About 3 to 7°of northwest tilting occurring after syn-rift deposition, during and/or after the late rifting phase, and before the deposition of the postrift strata of earliest Late Cretaceous strata that overlie the southeastern edge of the Newark rift basin (Figure 9(c)). Transverse folding associated with left-lateral strike-slip deformation affected the syn-rift strata in the hanging walls of the border faults and intrabasin faults after rifting. Based on the preservation of the late synrift strata in the troughs of many transverse synclines, folding likely ceased before significant erosion had occurred.

Onshore and Offshore Basins in the Study Area
The restorations show that tectonic activity in the Newark rift basin changed profoundly during the transition from rifting to drifting with: (1) the onset of CAMP activity, the development of intrabasin faults, and the increased rates of extension and subsidence during the late rifting phase, and (2) postdepositional northwest tilting, uplift, and erosion during the final rifting phase into the early drifting stage. As mentioned previously, the study area includes several onshore rifts basins in addition to the Newark rift basin, several offshore rift basins, and the northern end of the postrift Baltimore Canyon trough (Figure 2). The onshore and offshore rift basins, like the Newark rift basin, are offset from the continent-ocean boundary by tens to hundreds of kilometers. These rift basins, and the postrift Baltimore Canyon trough, provide additional information about the tectonic events during the transition from rifting to drifting in the central segment of the ENAM margin.  Figure 4(c)). The ages of the three basalt flows in the Connecticut Valley rift basin match those of the Newark rift basin (e.g., [37,70]). Also, like the Newark rift basin, the compacted sediment accumulation rates in the Connecticut Valley rift basin increased substantially during the late rifting phase (i.e.,~1 km/My; [70]). As mentioned previously, by the end of the late rifting phase, the northern end of the Newark rift basin likely connected with the Connecticut Valley rift basin. The Connecticut Valley rift basin has a N-S trend (parallel to the Paleozoic crustal fabric) and a major, W-dipping border-fault system that bounds the basin on the east (e.g., [54,103]) (Figures 2, 3(a), and 4(c)). Like the Newark rift basin, numerous, NE-striking intrabasin faults with normal separation and hanging-wall anticlines and synclines (transverse folds) affect the syn-rift strata of the Connecticut Valley rift basin (e.g., [54,65,103,105]). The syn-rift strata have highly variable dips, but overall dip~10 to 30°E toward the border-fault system. Dip analyses by Wise [54] suggest that significant tilting (~15°E or more) occurred after syn-rift deposition. Like the Newark rift basin, the Connecticut Valley rift basin also underwent significant postdepositional erosion, ranging from~1.5 km in the eastern part of the basin to more than 5 km near the western margin of the basin based on field relationships, maturation indices of organic matter [106], and fission-track (apatite and zircon) analyses [107]. If the depositional surface of the Connecticut Valley rift basin was roughly 1 km above sea level at the end of the late rifting phase, as it was in the interconnected Newark rift basin, then the Connecticut Valley rift basin must have undergone significant eastward tilting and uplift after synrift deposition. The timing of the eastward tilting, uplift, and erosion is poorly constrained. However, if the New York Bight rift basin is the offshore continuation of the Connecticut Valley rift basin, as proposed by Hutchinson et al. [53] and discussed below, then much of the eastward tilting, uplift, and erosion in the Connecticut Valley rift basin occurred before the deposition of the Late Jurassic and younger postrift strata that overlie the tilted syn-rift beds of the New York Bight rift basin.
The small Pomperaug rift basin lies about 20 km to the west and about 60 km to the northeast of the larger Connecticut Valley and Newark rift basins, respectively (Figures 2  and 3(a)). Like its larger counterparts, it has a lower/older sedimentary package deposited during the main rifting phase and an upper/younger package composed of CAMP-related basalt flows and interbedded sedimentary strata associated with the late rifting phase [108,109]. The stratigraphic packages associated with the main and late rifting phases, however, are significantly thinner than the equivalent packages in the Connecticut Valley and Newark rift basins [108]. A W-dipping border-fault zone with normal separation bounds the basin on the east (e.g., [104,108,110]). Generally, the syn-rift strata dip toward the border-fault zone (i.e.,~10 to 20°E) [110]. Burton et al. [108] propose that the Pomperaug rift basin was a small rift basin with both local and regional sedimentary and igneous sources. As discussed below, significant normal faulting (down to the west) after syn-rift deposition lead to its preservation today. 16 Lithosphere

Offshore Rift Basins on Long Island Platform.
Hutchinson et al. [53], using stacked seismic-reflection profiles, identified and described several buried ENAM rift basins on the Long Island platform mostly landward of the basement hinge zone, including the New York Bight and Long Island rift basins in the study area (Figures 2 and 3). To supplement the seismic studies by Hutchinson et al. [53], we interpreted reprocessed, time-migrated versions of key seismic lines from the Long Island platform (generously provided by Spectrum Geo). These time-migrated lines better image the structural and stratigraphic features of the New York Bight and Long Island rift basins (Figures 2, 3(b), 4(d), and 4(e)). The New York Bight rift basin, on the western end of the Long Island platform, is cut by several N-striking, W-dipping faults with normal separation (Figures 2, 3(b), and 4(d)). Generally, the imaged syn-rift beds are subparallel to each other and dip about 10 to 20°E toward the eastern borderfault system (Figure 4(d)). Their subparallel geometry (i.e., lack of fanning toward the border-fault system) suggests that most of the eastward tilting occurred after their deposition. The imaged syn-rift beds are erosionally truncated by the PRU, and stratal geometries indicate that several kilometers of erosion occurred near the western margin of the basin before the deposition of the Late Jurassic and younger postrift strata that overlie the basin (Figure 4(d)). These postrift strata dip very gently seaward (i.e., SE to SSE) toward the continent-ocean boundary.
As mentioned previously, Hutchinson et al. [53] proposed that the New York Bight rift basin is the offshore continuation of the Connecticut Valley rift basin. Evidence supporting their proposal includes the close proximity of the basins (Figure 2), the presence of Triassic strata in a well (S34, Figure 2) drilled between the basins [111], and the general alignment of the basins with each other and with potential field data. Additionally, both basins have similar structural geometries with a W-dipping border fault on the eastern side and syn-rift beds dipping toward the east (Figures 4(c) and 4(d)). If the New York Bight rift basin is the offshore continuation of the Connecticut Valley rift basin, as supported by all available geological and geophysical data, then much of the eastward tilting, uplift, and erosion of the syn-rift strata in the Connecticut Valley rift basin likely occurred before the deposition of the Late Jurassic and younger postrift strata that overlie the New York Bight rift basin.
The Long Island rift basin trends N-S. An E-dipping fault system (~30°E) with normal separation bounds the basin on the west. Syn-rift strata within the basin generally strike N-S and dip toward the border-fault system (~20°W) (Figures 2,  3(b), and 4(e)). The subparallel geometry of bedding indicates that most of the westward tilting toward the borderfault system occurred after the deposition of the imaged syn-rift strata [53]. The imaged syn-rift beds are erosionally truncated by the PRU, and stratal geometries indicate that several kilometers of erosion occurred near the eastern margin of the basin before the deposition of the Late Jurassic and younger postrift strata that overlie the basin (Figure 4(e)). These postrift strata dip very gently seaward (i.e., SSE) toward the continent-ocean boundary. 7.3. Baltimore Canyon Trough. The postrift Baltimore Canyon trough is more than 100 km wide and up to about 13 km deep in the study area (Figures 2(a) and 3(c)) (e.g., [52]). The basin, associated with tectonic subsidence and sediment loading, developed after rifting during the drifting phase (i.e., from Early Jurassic to present) primarily offshore and seaward of the basement hinge zone. During the early stages of basin development, sediments rapidly accumulated within the deepening trough (e.g., [13]). By the Late Jurassic, however, basin subsidence had slowed substantially. Approximately 70% of the postrift sedimentary rocks in the Baltimore Canyon trough (i.e., 9 km) are Jurassic in age [52]. Grow et al. [52] suggest that erosion of the youthful Appalachian Mountains produced the large sediment influx needed to fill the basin during the Early and Middle Jurassic, whereas others attribute the large sediment influx to either a peripheral bulge that accompanied the offshore subsidence [13] or residual onshore elevation associated with rifting [98].  (Figures 10(a) and 10(b)). The axis is also subparallel to a major gradient in the Bouguer gravity field, with lower values in the west and higher values in the east (Figure 10(b)). The gradient purportedly marks the location of the ancient Laurentian passive margin, separating the Grenville basement on the west from accreted Paleozoic terranes on the east (e.g., [112]). The location of the axis of the crustal-scale arch is generally coincident with a pronounced change in crustal thickness (i.e., thicker in the west and thinner in the east) [112] (Figure 10(b)). In the southern part of the study area, the postdepositional northwest tilting and uplift of the Newark rift basin define a broad, faulted, crustal monocline with a NW-dipping limb (Figure 10(e)). Based on the postdepositional westward tilting of the syn-rift strata of the Long Island rift basin, additional crustal-scale warps may exist on the Long Island platform (Figure 10(b)).
The exhumation estimates and restorations for the Newark rift basin suggest that the crustal-scale arching/warping observed in the study area began as early as the final rifting phase in the Early Jurassic, several millions after CAMPrelated magmatic activity. It continued during the early stages of drifting as sediments filled the rapidly subsiding postrift Baltimore Canyon trough to the south and east (e.g., [13,98]). Based on the age of the oldest, relatively flatlying postrift strata that overlie the southeastern edge of the Newark rift basin (i.e., earliest Late Cretaceous; [73]) and 17 Lithosphere the age of the oldest, relatively flat-lying postrift strata that overlie the offshore rift basins in the study area (i.e., Late Jurassic; [53]), the crustal-scale arching/warping and erosion ceased during the Late Jurassic for the offshore basins and by the earliest Late Cretaceous for the Newark rift basin. Thus, the episode of uplift, crustal-scale arching/warping, and erosion was prolonged, lasting for more than 40 million years. Based on our exhumation estimates for the Newark rift basin and those reported for the Connecticut Valley rift basin, erosion was significant, locally exceeding 6 km.  [134]. Estimates of crustal thickness based on teleseismic receiver-function analyses by Li et al. [112]. (c-e) Northern, central, and southern regional transects showed with the restoration of the estimated eroded section. References in the caption of Figure 3.

Lithosphere
We are not the first to propose that crustal-scale arching/warping developed in the study area after syn-rift deposition. Based on stratal geometries alone, Wheeler [113] theorized that a broad arch formed between the Newark and Connecticut Valley rift basins after rifting. He concluded that the small Pomperaug rift basin, located between the larger rift basins, was faulted downward, leading to its preservation. Sanders [63] proposed that postrift arching elevated the center of a large, interconnected Newark and Connecticut Valley rift basin. Our analysis, using information unavailable to Wheeler and Sanders (e.g., vitrinitereflectance data, sonic transit-time analyses, onshore and offshore seismic-reflection data, crustal thickness estimates) has allowed us to better define and understand the characteristics, timing, and causes of this anomalous crustal-scale arching/warping observed on the ENAM "passive" margin.

Uplift and Erosion in the Study Area and Potential Causes
As described below, the study area has two distinct styles of uplift and erosion related to the rift-drift transition. The first style, located in the southeast primarily between the basement hinge zone and the continent-ocean boundary, is associated with the formation of the well-documented, postrift unconformity (PRU) (e.g., [4,52,53]). The second style, characterized in this paper, is located in the north and west primarily landward of the basement hinge zone and affects the offset ENAM rift basins and the surrounding rocks.

Uplift and Erosion Primarily Seaward of the Basement
Hinge Zone. In the Baltimore Canyon trough, the PRU separates the prerift and syn-rift rocks from the overlying postrift strata. Grow et al. [52] place the PRU above the seaward-dipping reflectors. As mentioned previously, the PRU affects a broad region aligned with the continentocean boundary, stretching from the continent-ocean boundary landward across the basement hinge zone. Substantial subsidence and deposition followed the uplift and erosion associated with the development of the PRU with sediments rapidly accumulating within the deepening Baltimore Canyon trough throughout the Jurassic (e.g., [13]). The postrift strata that directly overlie the PRU are oldest in the east near the continent-ocean boundary (i.e., Early Jurassic in age) and become progressively younger toward the west (Figure 3(c)). Thus, the erosion associated with the PRU was diachronous, (i.e., short-lived near the continentocean boundary and more persistent near the basement hinge zone). As discussed previously, researchers have proposed a variety of mechanisms to explain the style of uplift and erosion associated with the development of a breakup/postrift unconformity, like the PRU. Several mechanisms would produce a more transient phase of uplift and erosion surrounding the site of breakup followed by significant subsidence (e.g., depth-dependent stretching with excess thinning of the lithospheric mantle near the site of breakup (e.g., [5]); delamination and/or dripping of the lower lithospheric mantle near the site of breakup [2])). Some mechanisms would produce uplift and a more prolonged phase of erosion flanking the site of breakup (e.g., deep lithospheric necking during rifting and breakup (e.g., [9]); flexural uplift caused by mechanical unloading during rifting and breakup [12]; peripheral bulging that accompanied offshore postrift subsidence and deposition [13])). As mentioned previously, several of these mechanisms likely contributed to the development of the PRU on the ENAM margin.

Uplift and Erosion
Primarily Landward of the Basement Hinge Zone. The second style of uplift and erosion in the study area has distinctly different characteristics than those associated with the PRU. The uplift and erosion are associated with faulted, crustal-scale arching/warping that occurred well landward of the basement hinge zone and continent-ocean boundary, affecting the offset ENAM rift basins and the surrounding rocks. The axis of the crustalscale arching/warping is subparallel to the prerift and synrift crustal fabric, not the continent-ocean boundary. In fact, it veers sharply northward away from the continent-ocean boundary in the northern part of the study area. Uplift and erosion likely began during the waning phase of rifting, several million years after CAMP activity, and were long-lived (>40 My), continuing well into the drifting stage (i.e., Late Jurassic to earliest Early Cretaceous). Postrift subsidence was minimal. For example, the thickness and geometry of the postrift Jurassic sedimentary rocks (inferred from seismic data, e.g., [52]) and the younger coastal-plain deposits (e.g., [95]) to the east of the Newark rift basin suggest that any postrift section above the Newark rift basin was thin (<1 km).
What mechanism(s) produced the uplift and erosion landward of the basement hinge zone during the rift-drift transition? Mechanisms like lateral conductive heat transfer from the site of breakup to the rift flanks (e.g., [6]), a dynamic response to secondary small-scale convection triggered by rifting and breakup (e.g., [7,8]), deep lithospheric necking during rifting and breakup (e.g., [9]), flexural uplift caused by mechanical unloading during rifting and breakup [12], and peripheral bulging that accompanied offshore postrift subsidence and deposition [13] would produce uplift and erosion aligned with and flanking the site of breakup. The axis of the crustal-scale arching/warping associated with the uplift and erosion landward of the basement hinge zone, however, is subparallel to the prerift and syn-rift fabric, veering sharply away from the basement hinge zone and continent-ocean boundary in the northeastern part of the study area. As discussed below, two possible contributing mechanisms are (1) crustal intrusion and/or magmatic underplating associated with CAMP, and (2) relic crustal thickening and elevated topography associated with prerift orogenic activity.
The first proposed mechanism involves the emplacement of CAMP-related mafic intrusions within or directly below the extending lower crust beneath the offset ENAM rift basins. Simple calculations assuming local Airy isostasy and typical densities for the crust (2800 kg/m 3 ), mantle (3300 kg/m 3 ), and mafic intrusions (3000 to 3100 kg/m 3 ) indicate that a 10-to 15-km cumulative thickness of crustal intrusions and/or underplated material would produce~1 km of 19 Lithosphere uplift and, with denudation unloading over time,~6 km of total denudation (e.g., [114]). Supporting evidence for this mechanism includes the abundance of CAMP-related sills, sheets, and dikes in upper crustal rocks near and within the ENAM rift basins, the presence of subhorizontal reflections/laminations (i.e., possible sills) within the lower crust beneath the ENAM rift basins (Figures 2(b) and 2(c); [51,53,115]), and the anomalously high-velocity rocks detected within the lower crust beneath the Connecticut Valley rift basin [116]. One significant problem with this proposed mechanism is that, thus far, geophysical studies on the ENAM margin (e.g., [117][118][119]) support only a limited amount of mafic intrusion into the lower crust landward of the basement hinge zone (i.e., considerably less than the 10 to 15 km required to produce a total denudation of 6 km).
The second proposed mechanism involves denudation unloading related to residual crustal thickening and elevated topography inherited from prerift orogenic activity. As discussed previously, multiple Paleozoic orogenies had produced a thickened, amalgamated crust in the study area, leading to the formation of the central and northern Appalachian Mountains. Two processes likely thinned the crust and reduced the topographic elevation before rift onset in the Late Triassic: (1) postorogenic gravitational collapse triggered by lithospheric delamination and/or ductile flow (e.g., [28]), and (2) a postorogenic cycle of erosion and isostatic rebound/uplift (denudation unloading) initiated by erosionrelated transfer of mass away from the mountains (e.g., [29,30]). This latter process likely played an increasingly important role in reducing the crustal thickness and topographic elevation as the ENAM lithosphere cooled through time (e.g., [120]) (Figure 11(a)).
Geomorphological studies suggest that postorogenic cycles of erosion and isostatic rebound can persist for hundreds of millions of years (e.g., [121,122]). Thus, it is not unreasonable to assume that the ENAM postorogenic cycle was incomplete when rifting began. During the main and late rifting phases, the crust stretched and thinned, promoting subsidence and deposition within the wide ENAM rift basins, thus, limiting the erosion-related transfer of mass from the study area and dampening the postorogenic cycle of erosion and isostatic rebound (e.g., [29]) (Figures 11(b) and 11(c)). During the final rifting phase, extension and subsidence waned landward of the basement hinge zone as extension became increasingly focused near the site of breakup. With reduced subsidence, the rift basins landward of the basement hinge zone filled to their spill points, causing river systems to cut back into the rift-basin fill (e.g., [123]). Additionally, breakup-related pathways likely formed, connecting the rift basins landward of the basement hinge zone to the embryonic Atlantic Ocean to the south and east and dropping base level. If residual crustal thickening and elevated topography associated with the prerift orogenic activity remained during the final rifting phase, then the increased erosion-related transfer of mass away from the study area would have revived the postorogenic cycle of erosion and isostatic rebound/uplift landward of the basement hinge zone (Figure 11(d)). This postorogenic denudation unloading would have continued into the early stages of drifting until the landscape was reduced to base level, leaving behind isolated, eroded remnants of the once massive, interconnected rift basins (Figures 11(e)). Thus, denudation unloading related to relic, prerift, orogenic crustal thickening and elevated topography could have caused much of the uplift and erosion during the rift-drift transition landward of the basement hinge zone on the ENAM margin.
Simple calculations assuming local Airy isostasy (e.g., [114], equation 8b) and typical values for the densities for the crust (2800 kg/m 3 ) and mantle (3300 kg/m 3 ) show that, if the residual topographic elevation associated with prerift orogenic activity was~1 km above sea level at the end of rifting, then landscape reduction to sea level would cause~6 km of total denudation and crustal thinning. These values are consistent with the Early Jurassic paleoelevation predictions of Pazzaglia and Brandon [98] and with our observed exhumation estimates for the Newark rift basin, respectively. The exact isostatic response would vary, depending on the value of the effective elastic thickness (presumably low, e.g., [124]) and on spatial variations in the crustal thickness and the densities of the crust and lithospheric mantle. Thus, prerift orogenic activity (e.g., crustal thickening and amalgamation), postorogenic events (e.g., gravitational collapse and denudation unloading), and syn-rift processes (e.g., crustal thinning beneath the rift basins and CAMPrelated crustal intrusion and/or underplating) would have influenced the magnitude and distribution of the postrift denudation unloading with greater uplift and erosion occurring in regions with thicker crust and/or less dense crust or mantle.
The second proposed mechanism, likely augmented by rift-related processes, can explain the distinctive characteristics of the onshore uplift and erosion in the study area. Namely, it can account for its parallelism to the prerift and syn-rift fabric not the continent-ocean boundary, its timing beginning during the final, waning phase of rifting and continuing into the early drifting stage, and the minimal onshore postrift subsidence after uplift and erosion had ceased.

Geologic Evolution of Study Area during Rift-Drift Transition
By the Late Triassic, postorogenic gravitational collapse and denudation unloading had significantly reduced the crustal thickness and topographic elevation associated with prerift Paleozoic mountain building. During the main rifting phase (Late Triassic), extension tectonically thinned the crust, promoting subsidence and deposition and, thus, dampening the postorogenic cycle of erosion and isostatic rebound. Preexisting zones of weakness associated with the Paleozoic orogenies strongly influenced the location and distribution of the rift basins and the attitude of their border-fault zones. The rift basins widened and deepened through time and, by the end of the main rifting phase (latest Triassic), the northern Newark, New York Bight, and southern Connecticut Valley rift basins were likely interconnected (Figure 12(a)).  During the late rifting phase (latest Late Triassic to early Early Jurassic), the rift basins continued to widen and deepen (Figure 12(b)). Igneous sheets and dikes, related to shortlived CAMP-related igneous activity, intruded the basin fill,   22 Lithosphere surrounding basement, and deep crust, and lava flows intermittently filled the composite rift basin. Intrabasin faults became active, dissecting the basin and linking with the border-fault systems on the southwest and northeast sides of the composite rift basin. Extension rates and sediment accumulation rates increased substantially. The rift-basin border-fault zones and intrabasin faults remained active during the final rifting phase. During the rift-drift transition, extension became increasingly focused near the site of breakup where a massive volcaniclastic wedge (i.e., SDRs) developed (Figure 12(c)). Far from the site of breakup, with extension and subsidence waning, the postorogenic cycle of erosion and isostatic rebound reintensified, causing tilting, uplift, and erosion. Together, the northwest tilting and uplift of the Newark rift basin and the eastward tilting and uplift of the Connecticut Valley and New York Bight rift basins defined a broad, faulted, crustal arch in the northern part of the study area ( Figure 12(c)). In the southern part of the study area, northwest tilting and uplift of the Newark rift basin produced a broad, faulted, crustal monocline. Erosion and uplift associated with the postorogenic denudation unloading greatly reduced the depth and width of the composite rift basins, isolating their depocenters, as rifting transitioned to drifting. Sediments derived from the eroding rift basins and the surrounding rocks filled the rapidly subsiding Baltimore Canyon trough (Figure 12(d)). Erosion and uplift associated with the postorogenic denudation continued until the Late Jurassic (offshore) to earliest Late Cretaceous (onshore). Transverse folding associated with left-lateral, strike-slip deformation affected the synrift strata in the hanging walls of the border faults and intrabasin faults and likely started soon after rifting had ceased but before significant erosion had occurred (i.e., in the Early to Middle Jurassic).

Conclusions
To better understand the evolution of the "passive" margin of eastern North America during the rift-drift transition, we have constructed and restored three representative cross sections from the Newark rift basin using seismic, field, core, and borehole data, and the results of vitrinitereflectance and sonic transit-time analyses. The restorations indicate that the Newark rift basin had three distinct rifting phases: (1) Main rifting phase (Late Triassic,~230 to 201.5 Ma).
The basin widened significantly through time, eventually exceeding 100 km in width. The cumulative heave on the border-fault system (a proxy for extension) and the thickness of the basin fill increased southward.
(2) Late rifting phase (latest Triassic to earliest Jurassic, 201.5 to~198 Ma). Significant magmatic activity affected the basin with CAMP-related dikes, sills, and sheets intruding the basin fill, the surrounding rocks, and the deep crust. CAMP-related basalt flows episodically filled the basin. The border-fault system remained active and, for the first time, major intrabasin faults developed, linking with the border-fault system. Heave/extension rates and accumulation rates increased markedly. The central and northern parts of the basin widened significantly.
(3) Final rifting phase (early Early Jurassic). The borderfault system and intrabasin faults remained active, but little deposition occurred in their hanging walls, signaling a transition from regional subsidence and deposition to regional uplift and erosion.
In addition to the Newark rift basin, the study area includes several onshore and offshore ENAM rifts basins and the northern end of the postrift Baltimore Canyon trough. Information from these basins suggest that postdepositional northwest tilting and uplift of the Newark rift basin in the west and postdepositional eastward tilting and uplift of the New York Bight and Connecticut Valley rift basins in the east formed a faulted, crustal-scale arch, 150 km wide, in the northern part of the study area. The axis of the crustal-scale arch is subparallel to the prerift and syn-rift fabric, veering sharply northward away from the basement hinge zone and continent-ocean boundary. In the southern part of the study area, the postdepositional northwest tilting and uplift of the Newark rift basin define a broad, faulted, crustal monocline. The crustal-scale arching/warping and erosion in the study area began during the final rifting phase in the Early Jurassic, several million years after CAMP-related magmatic activity. It continued during early drifting with the eroded material filling the rapidly subsiding postrift Baltimore Canyon trough to the south and east and ceased by the Late Jurassic (offshore) to earliest Late Cretaceous (onshore). Thus, the crustal-scale arching/warping and erosion were prolonged, lasting for more than 40 million years. As a result, the rift basins in the study area underwent significant exhumation, locally exceeded 6 km.
We propose that relic crustal thickening and elevated topography associated with prerift Paleozoic orogenic activity, likely augmented by rift-related processes, was the primary cause of the crustal-scale arching/warping and erosion landward of the basement hinge zone during the rift-drift transition. Multiple Paleozoic orogenies had thickened the crust in the study area, creating the Appalachian Mountains. Postorogenic gravitational collapse and postorogenic denudation unloading (cyclic erosion and isostatic rebound) had significantly reduced crustal thickness and topographic elevation by the Late Triassic onset of rifting. During the main and late rifting phases, extension thinned the crust, promoting subsidence and deposition within wide rift basins and dampening the postorogenic cycle of erosion and isostatic rebound. As rifting waned and subsidence slowed during the final rifting phase, the postorogenic cycle of erosion and isostatic rebound reintensified landward of the basement hinge zone, reducing crustal thickness and topographic elevation. The amount of the erosion and uplift produced by denudation unloading varied spatially, depending, in part, on crustal thickness and crustal/mantle density. Prolonged differential uplift and erosion associated with denudation unloading, produced the crustal-scale arching/warping 23 Lithosphere in the onshore part of the study area and left behind only tilted, isolated, eroded remnants of the once massive, interconnected rift basins.
Our work shows that prerift orogenies can profoundly affect subsequent syn-rift and postrift development. Preexisting zones of weakness can influence the location and distribution of rift basins (i.e., produce offset rift basins like the ENAM rift basins) and the attitude of their border-fault zones. Additionally, relic crustal thickening and elevated topography inherited from prerift orogenic activity can modify postrift evolution by producing significant postrift uplift and erosion, leading to rift-basin exhumation and increased sediment influx into adjacent offshore postrift basins.

Conflicts of Interest
The authors declare that they have no conflicts of interest.