Magnetotelluric insights into the formation and reactivation of trans-crustal shear zones in Precambrian basement of the eastern U.S. Midcontinent

Due to thick, pervasive Phanerozoic cover, the tectonic evolution of the Precambrian crystalline basement in the U.S. Midcontinent (Fig. 1), a portion of the North American cratonic platform that is bounded by Phanerozoic cordillera of the Rocky Mountains in the west and the Appalachian Mountains in the east, remains poorly understood. What little is known about the petrology, geochemistry, and geochronology of the basement in this region comes from studies of sparse outcrop exposures (e.g., Day et al., 2016), from extrapolation based on distant basement exposures in the North American cordillera and the Canadian shield (e.g., Whitmeyer and Karlstrom, 2007), and from analyses of samples from sparse basement-penetrating boreholes (e.g., Bickford et al., 2015). Borehole data also constrain the depth to the Great Unconformity (the boundary between the Precambrian basement and Phanerozoic cover), which serves as a marker horizon for discriminating the timing of tectonic events and the relationships between Phanerozoic and Precambrian structures (e.g., Marshak et al., 2017).

Geophysical analyses provide additional information about Precambrian basement characteristics. For example, seismic reflection profiles produce valuable insights into local crustal structure (e.g., Baranoski et al., 2009; McBride et al., 2016), but data coverage is sparse, and active-source seismology is inherently limited in the amount of information that it can yield about the crystalline basement due to ubiquitous diffractions and the general lack of major continuous reflective horizons. Passive-source seismic techniques (e.g., Schmandt and Lin, 2014; Shen and Ritzwoller, 2016; Xiao et al., 2022) can constrain crustal thickness and delineate large-scale variations in petrophysical properties, although they generally provide poor resolution at crustal depths. Potential field (gravity and magnetic) methods (e.g., Atekwana, 1996) are valuable in identifying structural lineaments and overall basement fabrics above the Curie depth; however, lacking corroborative evidence from other geological and geophysical data sets, conclusions from such analyses are often non-unique. In this context, three-dimensional images of electrical conductivity, obtained by inverting magnetotelluric (MT) data, can provide a unique and valuable complementary source of information about basement tectonics in the Midcontinent, with good resolution through much of the lithospheric column over a large region with existing data sets.

Previous MT studies of this region have generally focused on the northern Midcontinent (Yang et al., 2015; Wunderman et al., 2018) and the Midcontinent Rift System (Bedrosian, 2016), although a recent analysis from Missouri (DeLucia et al., 2019) has begun to provide valuable insights into a portion of the eastern Midcontinent. Here, we expand upon the work of DeLucia et al. (2019) by presenting new three-dimensional electrical conductivity images for a region encompassing several U.S. states, from Missouri in the west to West Virginia in the east, and from Michigan in the north to Tennessee in the south (Figs. 1 and 2). Significantly, our electrical conductivity images highlight key structural features of the Precambrian crystalline basement that may represent trans-crustal (and, perhaps, even trans-lithospheric) shear zones that have played a role in localizing Phanerozoic deformation.

The eastern U.S. Midcontinent (Fig. 2) comprises a portion of the North American cratonic platform, a broad region of relatively stable continental crust composed of Precambrian crystalline basement covered by largely undeformed Phanerozoic sedimentary strata of variable thickness (e.g., Sloss, 1988; Marshak and van der Pluijm, 2021). The nature of the Precambrian basement of this region is largely obscured by Phanerozoic overburden; in the eastern Midcontinent, basement rocks are only exposed in the St. Francois Mountains of southeastern Missouri (e.g., Day et al., 2016; Pevehouse et al., 2020; Fig. 1). Consequently, the limited availability of geological data has resulted in longstanding debates about the Precambrian evolution of this region (cf. Whitmeyer and Karlstrom, 2007; Daniel et al., 2023a; Frost and Frost, 2023).

Based on sparse age dates from borehole samples and on extrapolation from basement exposures in the southwestern United States, the eastern Midcontinent is viewed as being underlain by northeast-trending packages of Paleoproterozoic to Mesoproterozoic rocks that young to the southeast (e.g., Whitmeyer and Karlstrom, 2007). The basement in the northwestern portion of our study area is divided into two Paleoproterozoic domains of ca. 1.8–1.7 Ga crust and ca. 1.7–1.6 Ga crust (Figs. 1 and 2), the latter of which is generally viewed as comprising a juvenile continental marginal arc that was incorporated into Laurentia during ca. 1.65–1.6 Ga orogenesis (e.g., Whitmeyer and Karlstrom, 2007). To the north and southwest of our study area, these Paleoproterozoic domains were variably overprinted by ca. 1.4 Ga metamorphism associated with the Picuris and Baraboo orogenies (Daniel et al., 2013, 2023b; Aronoff et al., 2016; Medaris et al., 2021; Fig. 1); the signature of this crustal reworking may extend into our study area, although direct evidence for this is, at present, lacking.

The basement in the southeastern portion of our study area is generally characterized by rocks associated with the ca. 1.3–1.0 Ga Grenville orogenic cycle (Figs. 1 and 2). This crustal domain and the adjacent region to the northwest experienced spatially variable deformation and metamorphism associated with long-lived Grenvillian tectonism. The westward limit of this deformation is generally assumed to coincide with the Grenville Front magnetic lineament (Fig. 2; Baranoski et al., 2009), which itself is inferred to represent a reverse-sense shear zone that juxtaposes crustal domains with differing magnetic properties, possibly due to different levels of crustal exhumation or to different degrees of deformation between crustal blocks. This view is largely based on the extrapolation of exposed structures in Canada (Baranoski et al., 2009), although seismic reflection data sets provide some support for the notion that this potential field lineament tracks the Grenville deformation front (Green et al., 1988; Pratt et al., 1989). The main Grenville suture is not exposed in our study area, but it has nevertheless been mapped through this region with a combination of radiogenic isotope (Fisher et al., 2010) and geophysical (Steltenpohl et al., 2010; Murphy and Egbert, 2017; Murphy et al., 2023) data. Splays of this suture may underlie the northeasternmost portion of the Midcontinent (Fig. 2) and may connect to structures in southern Ontario (e.g., Whitmeyer and Karlstrom, 2007).

The nature of the basement between the ca. 1.7–1.6 Ga crustal province and the Grenville-aged crustal domain is poorly understood. Most conceptual models assume southeastward (in present coordinates) crustal growth at 1.5–1.3 Ga via cryptic juvenile terrane accretion and/or continental marginal arc evolution within a long-lived convergent margin, with additional widespread crustal magmatic overprinting, as discussed below (e.g., Whitmeyer and Karlstrom, 2007; Bickford et al., 2015). However, recent zircon U-Pb ages suggest that older (>1.5 Ga) crustal material may, in fact, extend into or otherwise exist within this enigmatic domain (Petersson et al., 2015; Fig. 2).

Data from the Nd-Sm and Lu-Hf radiogenic isotope systems, obtained from borehole samples, provide an additional means of inferring the evolution of the basement in the eastern Midcontinent region. These data can be used to infer the date (T_DM) at which the constituent petrologic material was originally extracted from an inferred model depleted-mantle source and, therefore, to track crustal growth and recycling (Fig. 2; Bickford et al., 2015; Petersson et al., 2015). Based on whole-rock Nd-Sm model ages, crust of the eastern U.S. Midcontinent has been divided into two domains along the “Nd line”: a region where T_DM is generally >1.55 Ga, and a region where T_DM is generally <1.55 Ga (Fig. 2; Van Schmus et al., 1996; Bickford et al., 2015). The former domain, with zircon U-Pb ages that are much younger (>100 Ma) than the associated T_DM dates, represents recycling of the ca. 1.7–1.6 Ga crustal province. In the latter domain, the small (<100 Ma) difference between T_DM dates and zircon U-Pb ages (Fig. 2) has often been taken as evidence that this region comprises the 1.5–1.3 Ga juvenile southeastern margin of Laurentia (e.g., Daniel et al., 2023a). However, zircon Lu-Hf model ages from Ohio, east of the Nd line, that are generally >1.6 Ga (Fig. 2; Petersson et al., 2015) again potentially indicate that older (>1.5 Ga) crust extends or otherwise exists within this enigmatic domain.

Much of the >1.3 Ga crust in southern Laurentia was extensively overprinted by widespread voluminous felsic volcanism and plutonism at ca. 1.5–1.3 Ga. In the eastern Midcontinent, this magmatic episode produced the ca. 1.4 Ga Eastern Granite-Rhyolite province (Fig. 2; e.g., Bickford et al., 2015); many of the dates shown in Figure 2 are from borehole samples of magmatic rocks associated with this event, with isotopic model ages interpreted as reflecting the age of the assimilated crustal component, as discussed above. This felsic magmatism has historically been attributed to crustal anatexis associated with extension-driven lithospheric heating and basaltic underplating within an overall compressional or transpressional tectonic regime (Whitmeyer and Karlstrom, 2007; Bickford et al., 2015). In this context, given the apparently continuous collisional events on the southeastern margin of Laurentia through the Paleoproterozoic and Mesoproterozoic, Eastern Granite-Rhyolite province magmatism could be a result of back-arc extension (e.g., Menuge et al., 2002; Slagstad et al., 2009) or post-orogenic extensional collapse (e.g., Walker et al., 2002). However, the geochemistry of Eastern Granite-Rhyolite province rocks proves challenging for such interpretations. The uniquely ferroan, alkalic to alkali-calcic, metaluminous composition of these rocks seems to require underplating and fractionation of asthenosphere-derived tholeiitic melts, possibly with some component of crustal assimilation, but with essentially no input from any sort of magmatic-arc or post-subduction source (Frost and Frost, 2013, 2023). By comparison with modern analogs, Frost and Frost (2023) show that typical back-arc extension alone is unlikely to explain the unique geochemical signature of the Eastern Granite-Rhyolite province rocks; instead, asthenospheric upwelling and associated partial melting apparently needs to be the key source of melt to produce such rocks, and those authors favor asthenospheric upwelling associated either with a mantle plume or with piecemeal delamination or fragmented slab foundering. (Delamination in particular may be supported by H-κ-c receiver-function observations presented by Xiao et al., 2022.) Furthermore, the recent recognition of contemporaneous (ca. 1.4 Ga) orogenic episodes to the north (Baraboo orogeny; Medaris et al., 2021) and southwest (Picuris orogeny; Daniel et al., 2013; Aronoff et al., 2016) of the Eastern Granite-Rhyolite province (Fig. 1) poses a major problem for formulating a coherent picture of the early to mid Mesoproterozoic southern margin of Laurentia that can explain both voluminous ferroan magmatism and simultaneous orogenic metamorphism.

At least three episodes of rifting affected Laurentia during the Proterozoic; these events are variably represented within the eastern Midcontinent. First, Laurentia experienced at least localized rifting at ca. 1.47–1.38 Ga, as indicated by rift-related sedimentary rocks of the Belt-Purcell Supergroup in the northwestern United States and southwestern Canada and by similar rocks in the southwestern United States (Jones et al., 2015). Whether extensional faulting of this age affected the eastern Midcontinent remains unknown, although it is possible that the East Continent Rift Basin (Baranoski et al., 2009; Fig. 1) initially formed during this time. The eastern Midcontinent region subsequently experienced intracontinental rifting with the development of the Midcontinent (Keweenawan) Rift System at ca. 1.1 Ga (e.g., Hinze and Chandler, 2020). The main rift cuts to the northwest of the eastern Midcontinent, through Iowa, although an arm of the rift extends under Michigan and possibly farther south (e.g., Baranoski et al., 2009; Stein et al., 2014; Hinze and Chandler, 2020). Historically, the southeastern extent of Midcontinent Rift structures has been assumed to be either the segment in Michigan or the poorly understood Fort Wayne “rift” anomaly in Indiana–Ohio (Hinze and Chandler, 2020), with the inferred Grenville Front truncating extensional structures (Fig. 2). Recently, potential field data and limited age dates have been reinterpreted as supporting a continuation of the Midcontinent Rift System southward into Kentucky (Stein et al., 2018). In this alternative interpretation, potential field structures that were previously interpreted as part of the Grenville Front have been reinterpreted as part of the Midcontinent Rift System. However, strong support for this model from subsurface geologic data is lacking, and interpretations that link potential field structures to the Grenville Front may be equally plausible (Hinze and Chandler, 2020). Finally, rifting further affected the eastern Midcontinent at the end of the Proterozoic and into the early Cambrian, in association with the breakup of Rodinia and the opening of the Iapetus Ocean. Successful Iapetan rifts occurred at the eastern end of our focus area, through Virginia–West Virginia (Fig. 2), but several other significant extensional corridors, including the Reelfoot rift along the Missouri–Illinois–Kentucky–Tennessee border, the Rough Creek graben in western Kentucky, and the Rome Trough in eastern Kentucky and West Virginia, were active during this time (e.g., Thomas, 1991, 2006).

In the Paleozoic, with deformation shifting eastward due to development of the Appalachian orogenic system, distinctly different patterns of sedimentation and deformation developed in the Midcontinent. Widespread sedimentary cover, which in places reaches thicknesses of 7 km, was deposited during several cycles of transgression and regression (e.g., Sloss, 1963; Burgess, 2008), and large-scale epeirogenic structures (regional intracratonic basins, domes, and arches) formed in response to differential subsidence and flexure induced by orogenesis to the east (Sloss, 1988; Kolata and Nelson, 1990; Howell and van der Pluijm, 1990, 1999). In addition, numerous Midcontinent fault-and-fold zones (MFFZs in Fig. 2) cut across and deformed Phanerozoic strata (Marshak and Paulsen, 1996; McBride and Nelson, 1999; Cox, 2009). Many, if not most, of these zones root in basement, have fairly steep dips in the upper crust, and splay up-dip into flower structures near the surface (Marshak and van der Pluijm, 2021). Some fault strands die out up-dip in monoclinal folds. The largest transpressional and transtensional displacements occurred in these zones coevally with orogenic events of the Appalachian system (Marshak and Paulsen, 1996; Cox, 2009). Some of the Midcontinent fault-and-fold zones delineate boundaries between intracratonic basins and domes, but others cut across epeirogenic structures. Notably, the dominant trends of Midcontinent fault-and-fold zones are parallel to those of documented Proterozoic rifts, and in some cases the fold-and-fault zones can be directly linked to deep-seated crustal structures (e.g., Craddock et al., 2017; Marshak and van der Pluijm, 2021). Marshak and Paulsen (1997) and Thomas (2006) interpreted these associations to mean that fault-and-fold zones initiated during Proterozoic extensional tectonism, either as normal faults or strike-slip faults, but without great enough displacements to result in formation of a rift trough deep enough to be visible after Phanerozoic inversion and erosion. Overall, this basement-controlled deformation has imparted a distinct structural pattern to much of the U.S. Midcontinent, with roughly perpendicular sets (both southwest–northeast-striking and southeast–northwest-striking) of structures partitioning the upper crust into rectilinear blocks (e.g., Marshak and Paulsen, 1997).

We use ModEM (Kelbert et al., 2014), an isotropic three-dimensional MT inversion code, to develop our new inverse solution for the eastern U.S. Midcontinent. We use a model grid that has a nominal horizontal cell size of 10 km and that is oriented 40° E of N (Fig. S1 in the Supplemental Material¹). This grid orientation was chosen to compromise between domain size, station layout, and region of interest, and it has the additional benefit of roughly aligning with the trend of regional structures (as described in the discussion of data errors below). We use a Lambert conformal conic projection (with standard parallels at 36.5°N and 42.5°N) for the coordinate system in our inversion domain, and our coordinate system rotation is around the projection origin at 39.5°N, 86.85°W. Our model grid uses a logarithmic vertical discretization and extends to a maximum depth of ~1000 km; vertical cell size is ~200 m at the surface and ~10 km at 60 km depth (Fig. S2). We use electromagnetic fields calculated from a national-scale electrical conductivity synthesis model that covers the contiguous United States (specifically CONUS-MT-2019, https://doi.org/10.17611/dp/emc.2019.conusmt.1; Kelbert et al., 2019) for computational boundary conditions in the inversion. (Refer to the Supplemental Material for more information about our inversion methodology.)

For the inversion, we use a starting model that includes a laterally variable distribution of 0.03 S/m (30 Ωm) Phanerozoic sedimentary overburden overlying a constant 0.01 S/m (100 Ωm) crust and a mantle with a one-dimensional, depth-dependent conductivity distribution. Where possible, sedimentary unit thickness is calculated as the difference between surface elevation (from ETOPO1; Amante and Eakins, 2009) and the elevation of the Great Unconformity (Marshak et al., 2017); in the southern and eastern portions of our model domain, we rely on sediment thickness from CRUST1.0 (Laske et al., 2013). We also impose the Great Lakes as 0.002 S/m (500 Ωm) freshwater bodies in the uppermost cells of our grid. The mantle conductivity profile is calculated with the SEO3 model (Constable, 2006) assuming the 1330 °C isotherm at 180 km depth, and low conductivity values are limited to no less than 0.01 S/m (100 Ωm) within the lithospheric column. (Fig. S3 shows plots of our starting model.)

The ModEM algorithm is a regularized inversion, so it attempts to find the “smoothest” model that fits the data. In practice, the parameters that control the regularization are set by the user. Our preferred inverse solution uses a uniform ModEM model covariance setting of 0.4 with two covariance operator passes. Given our grid discretization, this configuration results in structures being smoothed laterally over a length scale of ~35 km. (Figs. S4 and S5 provide a comparison of inverse solutions with different model covariance parameters.)

We use both full impedance (Z) and vertical magnetic field transfer function (VTF) data from 215 long-period MT sites (Fig. S1) that were collected by the EarthScope USArray program (Schultz et al., 2006–2018), the MAGIC project (Evans et al., 2016; Long et al., 2020), and the U.S. Geological Survey (Bedrosian et al., 2021). These data are available through the IRIS (EarthScope) Searchable Product Depository (SPUD) Electromagnetic Transfer Function (EMTF) database (http://ds.iris.edu/spud/emtf; Kelbert et al., 2011, 2018). As the data from these three different projects are provided at different sets of periods, we interpolate these data to a common set of 22 periods from 10 s to 10,000 s. We perform this interpolation for each transfer function component separately. Due to potential source biases in long-period VTFs, we limit our use of those data to the 10 s–1000 s period band. We rotate the data to align with the inversion grid (~40° E of N, although the exact rotation varies by site location, since absolute angular relationships change when moving away from the center of a map projection). In performing this data rotation, we also rotate the full error covariance where possible in order to correctly specify data errors in our chosen coordinate system. Although we are only able to use the diagonal elements of the full error covariance (the variances of individual transfer function components) for specifying data errors in ModEM, our chosen grid orientation likely is a natural coordinate system for expressing data errors since our grid axes roughly align with the regional geoelectric structural grain in the eastern Midcontinent (cf. DeLucia et al., 2019; Kelbert et al., 2019; Murphy et al., 2023). We generally enforce error floors of 5% of the off-diagonal components for Z, treating each row separately and applying the error floor to both entries in the row, and 0.03 (absolute) for the VTFs. However, because source biases associated with geomagnetic pulsations (e.g., Murphy and Egbert, 2018) are common in our data as narrow-band “glitches” in otherwise smooth transfer function curves, we enforce error floors of 10% for Z and 0.1 (absolute) for the VTFs within a latitude-dependent band spanning roughly a third of a decade between 10 s and 100 s. The exact limits of this band depend upon the location of the site in question, with higher latitude sites having the enlarged error floors at the upper end of the 10–100 s decade, and lower latitude sites having the enlarged error floors at the lower end of that decade (refer to Murphy and Egbert, 2018, for a discussion of this latitude dependence).

To test the requirements that our data place upon structures within our preferred inverse solution, we modify the conductivity model to remove or adjust the feature(s) in question, restart the inversion, and assess the changes in the model and in the data fit once the inversion has reconverged. We use the modified conductivity model as both the prior model and the starting model when restarting the inversion. Because the ModEM regularization also penalizes deviations from the prior model, the inversion algorithm will attempt to refit the data as best as possible with a model that deviates as little as possible from the modified model. If this process yields a model in which the resulting structures match those originally recovered in our preferred inverse solution, or if the resulting reconverged inverse solution does not fit the data as well as the original preferred inverse solution, then we can conclude that the structural feature being tested is required by the data.

In order to highlight the regional geoelectric basement structural grains of the eastern Midcontinent in our conductivity images, we perform post-inversion image processing on individual two-dimensional depth slices from our preferred three-dimensional inverse solution. We use a ridge-filtering operation, implemented with the scikit-image Python package (van der Walt et al., 2014), to highlight lineaments within our conductivity images. For every point in the supplied image (in our case, specifically log10 of a given depth slice), this operator evaluates the local eigenvalue decomposition of the second spatial derivative of the image over some length scale and assigns a value based on the ratio of the eigenvalues to that spatial position; high values are associated with more ridge-like structures in the image (Meijering et al., 2004). (The exact value assignment in the algorithm also takes into account the sign on the eigenvalue ratio in order to specifically select ridge “peaks” and ignore adjacent “troughs.” Note that the values are unitless and are normalized to be between zero and one, with high values reflecting ridge crests.) Here, we use a length scale of three cell widths for this filtering operation; this value is appropriate for our conductivity images, as our chosen model covariance settings would transversely smooth linear conductors over roughly that many cells.

Figures 3 and 4 show slices through our preferred inverse solution. (Fig. S6 also shows additional depth slices.) This solution fits our data set to an overall normalized root-mean-square error (nRMSE) of 1.19. Our reported misfit is normalized to the data errors, so on average our solution fits the data overall to within a factor of 1.19 of the error bars. (Fig. S7 provides plots of site-by-site misfit in different period bands.) Figure 5 shows the ridge-filtered versions of the depth slices shown in Figure 3; this filtering operation highlights the conductive lineaments within these depth slices. Our inverse solution reveals a rich geoelectric structure beneath the eastern Midcontinent region. Rather than interpret individual structural anomalies, we focus on the trends in crustal conductivity lineaments.

As Figure 5 demonstrates, conductivity lineaments in the lower crust beneath Missouri (MO), Illinois (IL), Indiana (IN), Michigan (MI), and western Ohio (OH) are dominantly oriented northwest–southeast (NW–SE). This set of linear model features includes a major NW–SE-striking conductor in Missouri that has been previously named the Missouri high-conductivity belt by DeLucia et al. (2019) as well as a series of subparallel weak- to moderate-amplitude (0.01–0.03 S/m) conductors to the east and northeast. Although of low amplitude (only about an order of magnitude more conductive than the surroundings), these conductors are nevertheless required by our data (Resolution Test 1, Figs. S9 and S10). As these structures have also been observed (but not closely examined) to varying degrees in previously published inversions (Bedrosian, 2016; DeLucia et al., 2019), we consider them to be robust geoelectric features. We also discount the possibility that the parallel conductive lineaments are inversion artifacts due to geoelectric anisotropy at length scales less than those established by our inversion methodology. In three-dimensional MT inversions, intrinsic subsurface anisotropy tends to result in alternating conductive and resistive stripes with a wavelength comparable to the depth at which they appear, which is reflective of the depth to the anisotropic structure (Bedrosian et al., 2019). In our conductivity images (Figs. 3 and 4), the conductive lineaments are spaced at a length scale of ~100 km, but they appear at crustal depths (<50 km); consequently, we consider it unlikely that they are inversion artifacts due to intrinsic anisotropy.

Significantly, these NW-trending high-conductivity belts roughly correlate spatially (cf. Figs. 2 and 5) with known upper-crustal fault-and-fold zones, such as the Lincoln–Cap au Gras structure (Harrison and Schultz, 2002) and the Sandwich fault zone (Kolata et al., 1978; Nelson, 1995). These structures are part of an array of NW–SE-oriented fault-and-fold zones that cut across the Midcontinent (Marshak and Paulsen, 1996, 1997). Some of these fault-and-fold zones can be traced into states to the west of our study area, where they align with offsets in the Midcontinent Rift axis (Fig. 5; refer also to the discussion in DeLucia et al., 2019).

These NW–SE-striking structures display a range of dips in Figure 4; however, resolution tests indicate that, with the exception of the Missouri high-conductivity belt, we generally cannot resolve the dips of these conductors with the available data (Resolution Tests 2 and 3, Figs. S11–S14). Similarly, resolution tests suggest that these structures extend through the mid and lower crust (15 km to the Moho), but, with the exception of the Missouri high-conductivity belt, this full depth extent is only weakly constrained by our data at present (Resolution Tests 4 and 5, Figs. S15–S18). Additionally, although these structures appear to extend beneath the Moho in some locations in our preferred inverse solution (Fig. 4), our data generally do not require a continuation of these structures into the mantle lithosphere (Resolution Test 6, Figs. S19 and S20).

In Kentucky (KY), West Virginia (WV), western Virginia (VA), and eastern Ohio (OH), lower-crustal conductivity lineaments are dominantly oriented northeast–southwest (NE–SW). These conductors are generally of high amplitude. The two dominant subparallel conductors beneath western VA and WV, which we identify as C1 in Figures 3 and 4, have been observed in previous MT imaging studies of this portion of the United States (Murphy and Egbert, 2017; DeLucia et al., 2019). In Figure 4, these NE–SW-striking conductors appear to extend into the uppermost mantle; this model feature is required by our data (Resolution Test 6, Figs. S19 and S20). Several of the high-conductivity belts in this region are subparallel to Iapetan rift zones (cf. Figs. 2, 3, and 5).

We highlight a weak NE–SW-striking throughgoing conductor that cuts from southern Illinois, through central Indiana, into northwest Ohio (Figs. 3F, 4, and 5D) and that is the northwesternmost of the NE–SW-striking set of conductive lineaments; we identify this conductive lineament as L1. Our data seem to require a continuous, weak (~0.003–0.01 S/m) conductor through this region (Resolution Test 7, Figs. S21 and S22), although it is, at present, weakly resolved with available data. Regardless, as previously published MT inversions (Bedrosian, 2016; DeLucia et al., 2019) have also shown this structure L1, we consider it to be robust to some extent.

We also identify two key, high-amplitude (~1 S/m) anomalies that fall roughly within the boundary zone between the NW-trending set of conductive lineaments and the NE-trending set of lineaments (Fig. 3). The first of these, which we identify as C2, is spatially associated with the Tennessee–Illinois–Kentucky lineament (TIKL), also called the South-Central magnetic lineament (SCML), which is a major linear feature in crustal magnetic field maps (Fig. 3B; Hildenbrand and Ravat, 1997). The second of these, which we identify as C3, is a steeply plunging, pipe-like anomaly in northwest Ohio, at the northeastern end of lineament L1, that has been observed in many previous MT studies (Yang et al., 2015; Bedrosian, 2016; Gribenko and Zhdanov, 2017; DeLucia et al., 2019). Both C2 and C3 robustly extend into the mantle lithosphere (Resolution Test 6, Figs. S19 and S20).

The Missouri high-conductivity belt has been interpreted as a Proterozoic signature of lithosphere-scale transtensional shearing, possibly along the Mesoproterozoic (ca. 1.5–1.3 Ga) southern margin of Laurentia (DeLucia et al., 2019). This interpretation is primarily based on a structural attitude that is nearly perpendicular to the inferred trend of the Proterozoic margin in this portion of Laurentia (cf. Whitmeyer and Karlstrom, 2007; Lund et al., 2015; Daniel et al., 2023a) and on spatial relationships with potential field structures, including particularly the Missouri gravity low (Hildenbrand et al., 1996; Figs. 3A and 4), which has been interpreted as a transtensional rift structure (as discussed in Lowell et al., 2010). This interpretation for the Missouri high-conductivity belt is strengthened by the spatial relationship between this conductivity structure and ca. 1.4 Ga magmatic rocks in the Eminence–Van Buren volcanic field and the St. Francois Mountains of southeastern Missouri, which may have formed as a result of the transtensional, lithosphere-scale tectonism along the Missouri gravity low structure (Lowell et al., 2010, and references therein). As our NW–SE-striking conductors are generally subparallel to the Missouri high-conductivity belt, we similarly interpret those lineaments as representing additional zones of localized crustal- or lithosphere-scale deformation. However, because we cannot decipher a meaningful sense of deformation for all of these structures, and because the transtensional deformation inferred along the Missouri high-conductivity belt may be a localized expression (e.g., dilational jog) of larger-scale tectonism that may have yielded differing senses of deformation in different locations, we interpret these NW-trending conductors simply as representing transcurrent faults or transforms that likely were “leaky,” with either local or regional components of transtensional deformation. These structures consequently likely facilitated the movement of deeply derived materials.

Significantly, these belts are roughly parallel to other major known shear zones and crustal structures that have played an important role in the tectonic evolution of North America (Marshak and van der Pluijm, 2021; cf. Figs. 1, 3, and 5), such as the Alabama–Oklahoma transform (Thomas, 2006), the Southern Oklahoma Aulacogen (Chase et al., 2023), and the Dakota–Carolina corridor (Marshak and Paulsen, 1996). This structural fabric appears to have subsequently controlled the orientation and location of basement-penetrating faults that formed during Neoproterozoic–Cambrian Iapetan rifting, and it may have, perhaps, even influenced structural patterns in the Mesozoic opening of the North Atlantic (Marshak and Paulsen, 1996; Thomas, 2006, 2011). As we discuss here, we cannot uniquely determine either the age or the nature of these belts with existing data, but our preferred interpretation is that they represent crustal-scale shear zones that were active along and therefore segmented the Mesoproterozoic (ca. 1.5–1.3 Ga) southeastern margin of Laurentia (in present coordinates).

The source of high conductivity values within geoelectric anomalies ideally provides a means of deducing their origin; however, the cause of relatively high conductivity values (0.01–0.03 S/m) in these NW-trending belts is at present not well constrained. Because magnetic minerals such as magnetite are also electrically conductive (Keller, 1966), high concentrations of such phases may contribute to enhanced conductivity. However, the highest amplitude zones within our imaged conductivity anomalies often spatially align with the edges, rather than exactly the local maxima, of magnetic anomalies (Fig. 4). For example, in cross section A–A′ in Figure 4, note that the highest conductivity values within the Missouri high-conductivity belt and the conductive zones in Illinois (IL)–Indiana (IN) are generally offset from magnetic highs. Consequently, magnetic phases are unlikely to be the main cause of the observed high conductivity values. There are, however, notable exceptions to this spatial pattern. For example, note the spatial correlation between the high-amplitude portion of the Missouri high-conductivity belt (Fig. 4, cross section B–B′) and the Bloomfield pluton, an inferred Proterozoic magmatic intrusion based on potential field observations (Ravat et al., 1987). For the Bloomfield pluton, with an anomaly of ~1500 nT, the minimum percentage of magnetite is ~3% for the upper part of the modeled buried body (Ravat et al., 1987). However, for equant minerals, >15% of the conductive phase is required to create an interconnected network that can increase bulk (regional) electrical conductivity (e.g., Nelson and Van Voorhis, 1983), and, for magnetite in particular, laboratory measurements suggest that a volume percentage of at least 20%–50% may be needed to appreciably affect bulk conductivity (e.g., Vella and Emerson, 2012). Although higher concentrations of magnetite are possible locally within rocks associated with the Bloomfield pluton (and within similar magnetic structures that occur along these NW-trending zones), magnetite alone is an unlikely explanation for the observed regional high conductivity values.

Sulfide minerals are highly electrically conductive (Keller, 1966) and could conceivably contribute to our observed high conductivity values in the NW-trending conductive zones. However, as with the magnetic minerals discussed above, large proportions (>15%; Nelson and Van Voorhis, 1983) of sulfide minerals would be required throughout the mid–lower crust to explain these conductors. We consider such high concentrations over large regions of the crust to be unrealistic for magmatically derived sulfides. Although these large sulfide volumes are possible within metasedimentary rocks trapped within suture zones (where they are also associated with electrically conductive graphite-bearing units; e.g., Murphy et al., 2023), these NW-trending conductive zones cut across the inferred orogenic grain in this portion of Laurentia (cf. Daniel et al., 2023a). We consequently consider suture-bound electrically conductive metasedimentary rocks to be an improbable source of these anomalies.

Deep saline fluids are an unlikely cause of elevated conductivity here. Such fluids are unstable on long geologic time scales (>10 Ma), so any fluids associated with original (Precambrian) tectonism would have long ago been consumed by retrograde hydration reactions (Yardley and Valley, 1997; Manning, 2018). Additionally, there is no evidence for modern active tectonomagmatic processes in this region that could introduce such fluids into the mid–lower crust, and we consider it highly implausible that surface fluids could migrate to such great depths in the crust under presumably lithostatic pressure gradients, particularly given the very low permeabilities that exist in tectonically stable lower crust (e.g., Manning and Ingebritsen, 1999). Similarly, crustal temperatures throughout the eastern Midcontinent are generally low enough (<700 °C; Schutt et al., 2018; Boyd, 2019) that both intrinsic and extrinsic volatile-enhanced semiconduction in silicate minerals are improbable explanations for the observed high conductivity values (temperatures >700 °C are required to yield conductivity values of ~0.1 S/m; e.g., Yang et al., 2012; Li et al., 2017; Hu et al., 2018; refer also to Murphy et al., 2022).

Given these considerations, we favor the hypothesis that high conductivity values within these NW-trending zones are due to graphite precipitation from CO₂-rich fluids that were mobilized from within the mid–lower crust in response to tectonomagmatic forcing or that were fluxed through the crust from the mantle in association with extension-driven mantle melting and devolatilization (e.g., Murphy et al., 2022; refer also to Huizenga, 2011, and Luque et al., 2014). As shown by Murphy et al. (2022), the cooling of CO₂-rich magmatic fluids at geochemical conditions near the fayalite-magnetite-quartz (FMQ) oxygen fugacity buffer can lead to the precipitation of sufficient amounts of electrically conductive graphite within structurally controlled fluid pathways to produce a large-scale conductivity anomaly. Mantle-derived magmatism, in particular, would likely mobilize enough CO₂ under the correct geochemical redox conditions to precipitate graphite veins along zones of focused fluid flow (Murphy et al., 2022). This hypothesis is strengthened by gravity and magnetic highs that are generally located laterally adjacent to high-conductivity zones (Fig. 4). These potential field features may represent plutonic rocks that could have been originally emplaced during the crustal-scale deformation that we hypothesize created these NW-trending zones, with the high conductivity values now demarcating the structural corridors that served as channels for the source magmas and that facilitated migration of associated fluids away from these intrusions as they crystallized. Our invoking graphite here is further supported by the association with known (Eminence–Van Buren; Lowell et al., 2010) and inferred (Vincennes; Hildenbrand et al., 2002; Decatur; McBride et al., 2016) Proterozoic igneous magmatic centers, which are located on the edges of the NW-trending zones (Figs. 3 and 5). In particular, the geochemistry of volcanic rocks from the Eminence–Van Buren volcanic field indicates an asthenospheric contribution (Lowell et al., 2010), and such melts often have appropriate redox conditions to promote graphite precipitation from associated magmatic fluids (refer to discussion in Murphy et al., 2022).

This hypothesis is compatible with numerous studies that have shown that fluid flow occurs not only through brittlely fractured crust in upper-crustal fault zones, but also within ductile mid–lower crustal shear zones (e.g., Glazner and Bartley, 1991; Blenkinsop and Kadzviti, 2006; Condit and Mahan, 2018). Ductile shearing additionally promotes interconnectedness of conductive phases to further enhance bulk conductivity (e.g., Jödicke et al., 2004). This explanation for our high conductivity values builds upon the hypotheses presented by DeLucia et al. (2019), who suggested that the observed signatures formed in a transtensional tectonomagmatic system with a stress state that would facilitate fluid migration from depth. This interpretation then carries with it the implication that the observed NW-trending conductors once were focused zones of major tectonomagmatic activity and associated fluid flow, as “leaky” crustal-scale structures. Significantly, graphite-infused shear zones conceivably have remained weak and, therefore, could have localized displacements that propagated into strata above the Great Unconformity during the Phanerozoic, as we discuss below.

As with the cause of high conductivity values, the age of these high-conductivity structures is only loosely constrained. Phanerozoic intracontinental deformation did not produce penetrative deformation throughout this region that would have been capable of creating new or substantially modifying existing crustal-scale conductivity structures (e.g., Marshak et al., 2017), so these conductors most likely originally formed during the Precambrian (Marshak et al., 2000). Furthermore, these NW-trending zones do not parallel well-studied late Mesoproterozoic (Keweenawan) or Neoproterozoic (Iapetan) rift structures (cf. Fig. 2); in fact, they appear to be associated with fault zones that controlled offsets in Keweenawan rift structures to the west of our study area (refer to discussion in DeLucia et al., 2019; Fig. 5) and with faults that influenced structural patterns in the Reelfoot rift (cf. Csontos et al., 2008). Consequently, they most likely formed prior to the late Mesoproterozoic (before ca. 1.1 Ga). These conductors are found within interpreted >1.6 Ga crust (cf. Fig. 2), so they may represent either ca. 1.8–1.6 Ga plate margin geometries, or geometries that formed at ca. 1.5–1.3 Ga, or geometries that originally formed at ca. 1.8–1.7 Ga and that were repeatedly reactivated until 1.3 Ga. It is worth noting that zircon U-Pb and Lu-Hf isotopic data indicate the presence of older (ca. 2 Ga) crustal components in the vicinity of our NW-trending belts (Fig. 2; Petersson et al., 2015), so these structures possibly formed as early as ca. 2 Ga. Regardless, based on these considerations, these NW-trending conductive zones are loosely constrained to be late Paleoproterozoic to mid Mesoproterozoic in age. Based on their orientations, they may have formed during extension of a similar orientation that, along the western margin of Laurentia, yielded the ca. 1.47–1.38 Ga Belt rift (Jones et al., 2015).

Despite the uncertainty regarding timing, we consider it likely that these structures formed at ca. 1.4 Ga, in association with formation of the Eastern Granite-Rhyolite province, as they could have served as conduits for the distinctive ferroan magmatic rocks of this province. Such a scenario would readily explain the association with the ca. 1.4 Ga, Eastern Granite-Rhyolite province-related Eminence–Van Buren volcanic field. It would also mesh well with our invocation of graphite as the cause of high conductivity values, as ferroan magmas are often accompanied by a high mantle-derived CO₂ flux and are generally characterized by redox conditions that are conducive to graphite precipitation (cf. Frost and Frost, 1997; Murphy et al., 2022). However, new dates from a gabbro from the Decatur igneous center, which falls along the edge of one of our NW-trending zones (Fig. 3; McBride et al., 2016), yields an age of ca. 1.07 Ga (Freiburg et al., 2020), which is much younger than the age of the event that formed the Eastern Granite-Rhyolite province. Although it is possible that igneous activity in the Decatur region is unrelated to our NW-trending zones, or that magmatism reused preexisting crustal pathways, we nevertheless lack dates on basement magmatic rocks that would support our proposed link to Eastern Granite-Rhyolite province magmatism here.

Regardless, our NW-trending zones mesh well with the model for early Mesoproterozoic tectonics that has been recently proposed by Daniel et al. (2023a). In an attempt to reconcile the complex and seemingly contradictory set of tectonic requirements for the early Mesoproterozoic, Daniel et al. (2023a) lay out a scenario in which the southeastern (in present coordinates) Laurentian margin was segmented, with dextral strike-slip faulting at ca. 1.53–1.48 Ga in the eastern Midcontinent that transitioned to E–W oblique subduction along a NE–SW margin at ca. 1.48–1.4 Ga and subsequently to slab rollback with margin-perpendicular extension after ca. 1.4 Ga. Dextral shearing and oblique subduction would have provided a conducive stress regime to establish NW–SE-oriented “leaky” transforms or transcurrent structures with either local or regional areas of transtension, exactly the type of structures that can explain our observed NW-trending conductivity belts. We must acknowledge, however, that it is unclear if this tectonic model is compatible with the unique geochemistry of ca. 1.4 Ga magmatism (cf. Frost and Frost, 2023).

We propose that geoelectric structures C2 and C3 are part of these NW-trending zones and represent dominant segmentation boundaries along the Laurentian margin, similar to structures proposed in the model of Daniel et al. (2023a). As major crustal features on the southeastern edge of the NW-trending structural zone, C2 and C3 could have focused substantial crustal-scale magmatism and associated fluid flow, which could explain their very high conductivity values compared to other conductive structures in our images. This interpretation also meshes well with previous interpretations of the Tennessee–Illinois–Kentucky lineament (TIKL), which coincides with C2 (cf. Figs. 3B and 3F), as previous research has argued that this potential field structure was a major crustal boundary that localized the emplacement of magmatic intrusions (e.g., Hildenbrand and Ravat, 1997; Hildenbrand et al., 2002).

Since these NW-trending zones align with Midcontinent fault-and-fold zones that can be traced northwestward to transfer faults along the 1.1 Ga Midcontinent Rift to the west of our immediate study area (Fig. 5; refer also to discussion in DeLucia et al., 2019), it is possible that that they controlled the location and orientation of accommodation zones during formation of the rift. Within our study area, these NW-trending zones could have also influenced the potential southeastward continuation of the Midcontinent Rift (e.g., Stein et al., 2018). Furthermore, our NW-trending zones could themselves have been reactivated to some extent during Midcontinent Rift activity, where rift deformation generally spatially intersects these structures. Renewed motion along these shear zones could have allowed Keweenawan fluids to circulate through the shear zones, further enhancing their conductivity (perhaps particularly within structure C3).

Whereas the NW-trending high-conductivity belts may share a common explanation linked to Mesoproterozoic plate-margin processes, the NE-trending belts likely have more diverse origins, as is reflected in their greater range of azimuths.

Conductive lineaments in Kentucky have been previously suggested to be the signature of crustal metasomatism within the Neoproterozoic Rough Creek graben–Rome Trough system (Bedrosian, 2016), which formed during Iapetan rifting. However, this interpretation has not yet been tested by rigorous demonstration of a spatial correlation between rift-bounding structures and high-conductivity zones. We leave further assessment of such an interpretation to future work.

Major crustal terrane sutures are recognized globally to manifest geophysically as high-conductivity lineaments (refer to discussion in Murphy et al., 2023). Consequently, due to spatial correlations with a major crustal magnetic lineament (the New York–Alabama [NY–AL] lineament; Fig. 3B; Steltenpohl et al., 2010) and with a mapped contrast in radiogenic isotope affinities (Fisher et al., 2010), the two major NE-trending high-conductivity lineaments that we label C1 (Fig. 3) have been previously interpreted as Grenville-aged suture zones (Wannamaker, 2005; Murphy and Egbert, 2017; refer also to discussion in Murphy et al., 2023). This interpretation is further strengthened by roughly collocated, discontinuous linear magnetic lows (Figs. 3 and 4) that could be indicative of reduced magnetic properties in fault-damaged rocks within crustal-scale shear zones. The high conductivity values (>1 S/m) within structures C1 are specifically interpreted to be due to highly electrically conductive, suture-bound graphite- and sulfide-bearing metasedimentary rocks that were trapped within these ancient, closed subduction zones (Murphy and Egbert, 2017; Murphy et al., 2023). Such rocks would have been originally deposited in a euxinic (sulfidic and anoxic) oceanic basin (fairly common in the Proterozoic) between converging continents, and during orogenic metamorphism the constituent biogenic carbon and authigenic sulfidic material would have been converted to highly electrically conductive graphite and pyrite/pyrrhotite (refer to discussion in Murphy et al., 2023). Elevated conductivity values (0.03–0.1 S/m) also extend into the mantle lithosphere; at such depths, the observed values may at least partially be the result of grain-size diminution within mantle shear zones (e.g., Naif et al., 2021).

Notably, the western lineament within C1 follows the Cambridge–Burning Springs fault system, also called the Cambridge Arch–Burning-Mann lineament (cf. Figs. 2 and 3; Root, 1996; Shumaker and Wilson, 1996), a deep Proterozoic crustal boundary interpreted from potential field data and structural contour maps that may be the suture along which the Elzevir block (Fig. 1) was accreted during the ca. 1.25 Ga Elzevir orogeny (e.g., McLelland et al., 2010), an early phase of Grenville tectonism. The eastern lineament within C1 likely formed during a later episode of the greater Grenville orogeny (possibly in the ca. 1.15 Ga Shawinigan orogeny, based on the extent of Grenville deformation in this portion of Laurentia; refer also to discussion in Murphy et al., 2023; cf. McLelland et al., 2010). Structures C1 therefore demonstrate the multiphase nature of the Grenville orogeny in this portion of Laurentia, and they highlight the boundary along which exotic terranes were stitched into Laurentia during that long-lived orogenic event.

The steep attitudes of structures C1 through the crustal column (Fig. 4) indicate that they represent basement-penetrating, high-angle structures that formed during thick-skinned Grenvillian deformation. Studies of the NY–AL lineament have suggested that this major Grenvillian suture zone likely accommodated both dip-slip and strike-slip deformation (Steltenpohl et al., 2010). Major shearing greatly promotes interconnectedness of conductive phases and consequently substantially enhances the bulk conductivity of conductive lithologic units (e.g., Jödicke et al., 2004), so the observed high conductivity values within C1 are consistent with substantial deformation in multiple styles along this structure. Our observations of a pair of high-angle structures C1 also mesh well with seismic reflection results that show bivergent, high-angle (>30°) structures associated with Grenvillian structures (Culotta et al., 1990) where we observe our pair of high-conductivity lineaments. However, recent receiver-function imaging results have emphasized low-angle (<10°) structures in the upper–mid crust that shallow westward over the region where we observe structures C1 (Long et al., 2019). In our cross sections through structures C1 (Fig. 4, cross sections C–C′ and D–D′), we note a thin, highly resistive (~0.0001 S/m) zone that sits on top of structures C1 at depths of <15 km, with a major subhorizontal gradient in conductivity values at ~10–15 km depth. We consider it likely that the receiver-function imaging is sensing this petrophysical structural boundary, and that this feature represents thin-skinned thrusting of an allochthonous sheet over older, thick-skinned Grenvillian structures during the latest phases of the greater Grenville orogeny.

Our conductivity images show two broad domains that are characterized by different overall trends in conductivity belts (Fig. 5). As we have argued above, the northwestern domain of NW-trending high-conductivity belts is likely associated with penetrative crustal-scale shear zones that formed within the older Precambrian cratonic core of Laurentia, west of the limit of Grenvillian deformation, in association with early Mesoproterozoic plate-margin processes. The southeastern domain with NE–SW-striking conductivity structures is likely associated with later Precambrian to early Phanerozoic tectonism along the Laurentian margin, particularly Grenvillian orogenesis and Iapetan rifting. Consequently, fundamentally different processes operating in different geographic regions at different times gave rise to these differing sets of conductivity lineaments.

The boundaries of these broad regions are not well defined in our conductivity images. However, as we interpret in Figure 5E, these two geoelectric structural domains appear to overlap in the vicinity of L1, the weak (0.003–0.01 S/m), NE–SW-striking, throughgoing conductive lineament that is the northwesternmost member of the NE-trending lineament set. It is worth noting that conductive lineament L1 does not coincide with the Nd line and the transition from the ca. 1.7–1.6 Ga crustal province to enigmatic, inferred ca. 1.5–1.3 Ga marginal crust. That isotopically defined crustal boundary is mapped ~100 km northwest of the weak, throughgoing conductor L1 (cf. Figs. 2 and 3). However, the Nd line is only constrained by a few data points in Illinois–Indiana–Ohio, and Lu-Hf data from Ohio indicate that recycled >1.6 Ga crustal material extends eastward past this line (Fig. 2; Petersson et al., 2015). The only geophysical evidence to support the interpretation of the Nd line as a significant, discrete crustal boundary is long-wavelength (low-pass filtered to wavelengths of hundreds of kilometers) crustal magnetic field data, which show a NE–SW-striking boundary in crustal magnetization that roughly coincides with the drawn isotopic boundary (Ravat, 2007).

Despite the imperfect alignment of L1 with these other geochemical boundaries and geophysical lineaments, the subparallel trend with respect to inferred (albeit poorly resolved or understood) geochemical spatial patterns does suggest that this weak throughgoing conductor L1 may represent a component of a broad crustal boundary zone within this portion of Laurentia, which in part defines a diffuse transition between our two geoelectric structural domains. We propose that this boundary zone separates cratonic crust with sufficient strength to withstand Iapetan rifting from weaker crust that was more susceptible to rifting. The crust of the two broad domains may have formed at different times or may be distinguished by compositional differences that influence rheology. In this context, the weak throughgoing conductor L1 could be the signature of moderately conductive graphite- and sulfide-bearing metasedimentary rocks that were trapped within a collapsed intracontinental or oceanic marginal basin (e.g., Murphy et al., 2023) between these crustal domains, although further work is necessary to test this hypothesis.

Notably, large-scale structures within Phanerozoic fold-and-fault zones in the eastern Midcontinent often spatially align with—and, in some cases, directly overlie—the mid–lower crustal conductors that we document here (cf. Figs. 2, 3, and 6). For example, high-conductivity lineaments underlie the NW-trending, seismically active Sandwich fault in northern Illinois (Kolata et al., 1978; Nelson, 1995), as well as the Lincoln–Cap au Gras structure and the seismically active Ste. Genevieve fault zone. A high-conductivity lineament also underlies a major lateral step in the La Salle deformation belt (Fig. 6), a north–south-trending fault-and-fold zone within Phanerozoic sedimentary rocks in Illinois (Marshak and Paulsen, 1997; McBride, 1997). The belt has been interpreted as an inverted rift, with the step as a transfer zone. The Missouri high-conductivity belt aligns with both a lateral step in the portion of the Reelfoot rift that underlies the New Madrid seismic zone and a lateral step in the Midcontinent Rift. Similarly, in Ohio, the Bowling Green–Auglaize fault system (Schwartz and Christensen, 1988; Wickstrom et al., 1992) and the Cambridge–Burning Springs fault system (Root, 1996; Shumaker and Wilson, 1996) follow mid-crustal conductivity lineaments (Fig. 6).

These spatial relationships provide evidence that the Precambrian basement structures that we describe here have played some role in localizing recent deformation in the eastern Midcontinent due to transmitted far-field stresses, as has been proposed previously (e.g., Marshak and Paulsen, 1997; Marshak and van der Pluijm, 2021). These ancient high-angle structures may have been reactivated with dominantly reverse-sense motion during Paleozoic orogenic events to the east and south, so that, at least in the upper crust, they resemble Laramide structures of the Rocky Mountains and the Colorado Plateau (e.g., McBride, 1997; McBride and Nelson, 1999; Marshak et al., 2000). Our observations of trans-crustal (and possibly trans-lithospheric) high-conductivity zones suggest that at least some of the Midcontinent Phanerozoic fault-and-fold zones extend downward into basement-transecting structures that, notably, intersect the Moho, rather than merge into mid-crustal detachments. Furthermore, our observations indicate that these basement high-conductivity shear zones have been, and continue to be, significant weaknesses in the craton (possibly due to graphite content) that are susceptible to reactivation under low differential stress conditions (e.g., Marshak and Paulsen, 1996). These structures, therefore, have controlled the geometry of Phanerozoic upper-crustal intracratonic deformation, and they may still play a role in localizing contemporary seismicity (e.g., Marshak and Paulsen, 1997; Marshak and van der Pluijm, 2021).

Using long-period MT data, we have shown that the cratonic platform of the eastern U.S. Midcontinent is characterized by distinct high-conductivity belts within the Precambrian basement. We recognize two distinct sets of belts with fundamentally different structural orientations: in Missouri, Illinois, Indiana, and western Ohio, the high-conductivity belts are oriented NW–SE, whereas in eastern Ohio, Kentucky, West Virginia, and western Virginia, the high-conductivity belts are dominantly oriented NE–SW.

We propose that the NW-trending conductive zones represent early Mesoproterozoic transforms or transcurrent structures that locally or regionally accommodated some component of transtensional deformation. The high conductivity values along these structures are likely due to graphite precipitation from CO₂-rich magmatic fluids within deformation-damaged rocks in these crustal-scale shear zones. We consider it likely that these NW-trending zones served as conduits for the emplacement of ferroan magmas that comprise the Eastern Granite-Rhyolite province. However, we acknowledge that existing tectonic models are perhaps insufficient to fully reconcile the asthenospheric input required to produce ferroan granites (Frost and Frost, 2023) with the dextral shear and oblique subduction that can best explain both our geophysical observations and regional geologic data sets (Daniel et al., 2023a). As such, our proposed link between the NW-trending zones and ferroan magmatism requires validation through further study, although our observations nevertheless provide information for testing and revising models of early Mesoproterozoic Laurentian tectonics.

The NE-trending conductive lineaments may have a more diverse origin than the NW-trending belts, an observation reflected in the greater variation in azimuths of the NE-trending set. Several of the NE-trending lineaments can be clearly linked to Grenville orogenesis, whereas others likely represent the signature of Neoproterozoic rifting, although further analysis is needed to test this hypothesis.

The boundary between these two fabric-defined domains includes a weak throughgoing conductive lineament that may, at least partly, represent a cryptic crustal boundary zone between areas more susceptible to Iapetan rifting and areas that were less susceptible to Iapetan rifting. The nature of this cryptic boundary remains unclear, particularly given the conflict between datasets that have supported a major crustal boundary along the inferred Nd line (e.g., Ravat, 2007; Bickford et al., 2015) and newer zircon U-Pb and Lu-Hf ages that suggest >1.6 Ga crustal material extends eastward of that proposed boundary (Petersson et al., 2015).

Regardless of the exact origin of the high-conductivity belts in these contrasting domains, our geoelectric structures align with several mapped Midcontinent fault-and-fold zones that involve Phanerozoic strata. This observation lends support to past inferences that some Midcontinent fault-and-fold zones are rooted in basement structures that were reactivated by far-field stresses throughout the Phanerozoic. Our geoelectric images provide insights into deep structures that control modern intraplate deformation and seismicity.

The conductivity model presented herein is available through the IRIS (EarthScope) Earth Model Collaboration (EMC) (http://ds.iris.edu/ds/products/emc/). The model has also been incorporated into the revised CONUS-MT national-scale conductivity model of the contiguous United States (CONUS-MT-2021, https://doi.org/10.17611/dp/emc.2021.conusmt.1; Murphy et al., 2023). We thank Jamey Jones, Brandon Chase, and an anonymous reviewer for helpful feedback on this work, and we thank Krissy Lewis, Brian Shiro, and Janet Carter for facilitating U.S. Geological Survey internal review. B.S. Murphy acknowledges support from the National Science Foundation Graduate Research Fellowship Program under grant 1314109-DGE to Oregon State University for the beginning of this work and support through the U.S. Geological Survey Mendenhall Research Fellowship Program for the conclusion of this work. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government.