Three-dimensional magnetotelluric (MT) imaging in central Colorado reveals a set of north-striking high-conductivity tracks at lower-crustal (50–20 km) depths, with conductive finger-like structures rising off these tracks into the middle crust (20–5 km depth). We interpret these features to represent saline aqueous fluids and partial melt that are products of active extensional tectonomagmatism. These conductors are distributed over a wider region than the narrow corridor along which Rio Grande rift structures are traditionally mapped at the surface, and they consequently demarcate regions of the lower crust where accommodation of bulk extensional strain has concentrated conductive phases. Our observations reveal limitations in existing models of Rio Grande rift activity and may reflect unrecognized spatiotemporal variations in rift system evolution globally.Although middle to late Cenozoic (post-Laramide) extension has impacted much of the western United States, its regional nature and the degree to which it is ongoing at present remain poorly understood. East-west extension (Levandowski et al., 2018) in Colorado is generally viewed as a component of the Rio Grande rift (Fig. 1; e.g., Hudson and Grauch, 2013). Tweto (1979) originally distinguished between the Rio Grande rift “proper” and the Rio Grande rift “system.” In central Colorado, the former is defined by a narrow series of half grabens that in part comprise the upper Arkansas River Valley and that terminate at the normal-fault-bounded Gore Range. The latter includes a broader region of extensional faulting and Neogene sedimentation that reaches eastward into the Front Range and northward to the Wyoming border.Despite Tweto’s distinction, various data sets spur different interpretations as to whether extension is confined to the rift proper or is distributed across the rift system. For example, thermochronology suggests that middle to late Cenozoic extension has been restricted to the narrow corridor defined by the normal-fault-bounded Sawatch, Mosquito-Tenmile, and Gore Ranges (Fig. 1; Abbey and Niemi, 2020). Young cooling ages (<20 Ma) in these ranges (e.g., Landman and Flowers, 2013), and much older cooling ages (>45 Ma) elsewhere (e.g., Kelley and Chapin, 2004), indicate that significant post-Laramide exhumation has been limited to the rift proper. In contrast, other data sets suggest that extensional deformation has occurred, and is still occurring, across the full rift system (Fig. 1). Neogene extensional faults are found outside of the rift proper (Tweto, 1979; Colman, 1985), and apparent offsets in the Eocene erosional surface (Epis and Chapin, 1975) support minor jostling of crustal blocks to the east of the main geomorphic expression of extension. Additionally, young (<10 Ma) volcanic rocks occur exclusively outside of the rift proper (Fig. 1; e.g., Cosca et al., 2014), warped Miocene marker horizons in the westernmost Great Plains demonstrate far-reaching Neogene deformation (Leonard, 2002), and geodetic observations show broadly distributed extensional strain between the Colorado Plateau and the western Great Plains (e.g., Murray et al., 2019). Consequently, conflicting perceptions of the nature of extension in central Colorado arise when considering surface processes versus tectonomagmatic activity.Previous geophysical imaging has been inadequate for determining how and where the crust of central Colorado is undergoing extension. Seismic studies have revealed anomalously hot mantle (e.g., MacCarthy et al., 2014), but their resolution at crustal depths is insufficient to infer extensional processes. A two-dimensional magnetotelluric (MT) study showed a laterally continuous lower-crustal conductor that is interpreted as a signature of extension beyond the axis of the Rio Grande rift proper (Feucht et al., 2017), although off-line insights are limited.Using three-dimensional MT imaging, we show that the middle to lower crust of central Colorado is anomalously electrically conductive along regionally discrete north-south (N-S) tracks. We conclude that these conductors are a signature of active (i.e., modern, presently evolving) tectonomagmatism associated with extension that is distributed beyond the narrow Rio Grande rift proper, across the Rio Grande rift system more broadly. Our results reveal limitations in models of both the Rio Grande rift and rifting processes in general.We inverted full impedance tensor (Z) and vertical magnetic field transfer function (VTF) data from 81 MT sites across central Colorado (Fig. 1; data available at http://ds.iris.edu/spud/emtf). These data were collected by multiple surveys, and constituent transfer functions (TFs) were originally estimated at different periods; we therefore interpolated the data (treating each TF component separately) to a common set of 37 periods from 10−2 to 104 s. We limited our VTF data to 10−2 to 103 s. We enforced error floors of 5% of the off-diagonal components for Z (treating each row separately) and 0.03 (absolute) for the VTFs. See Item S1 in the Supplemental Material1 for more information.We performed three-dimensional inversion of these data with ModEM (Kelbert et al., 2014). Our model grid had a nominal horizontal cell size of 2.5 km and included topography. Above an elevation of 1 km (with respect to sea level), we used a constant vertical discretization of 100 m, and we mapped topography from ETOPO1 (Amante and Eakins, 2009) onto the grid in order to define fixed 10−10 S/m air cells. Below that elevation, we used a logarithmic vertical discretization that extended to a maximum depth of ~930 km; vertical cell size at 40 km below sea level was ~7 km. Our starting model placed conductive (0.1 S/m) sediments of the western Great Plains (thickness defined as the difference between ETOPO1 surface elevation and Great Unconformity elevation from Marshak et al., 2017) atop a depth-dependent conductivity profile defined by the SEO3 conduction model (Constable, 2006), with the 1330 °C isotherm at 70 km depth and with low conductivity values limited to no less than 0.003 S/m. We used ModEM covariance settings of 0.2 horizontally and 0.4 vertically, with two operator passes; this resulted in model structures being smoothed laterally over ~5 km.Our inverse solution (Fig. 2; see also Item S2 in the Supplemental Material), which fits the data to an overall normalized (to error bars) misfit (nRMSE) of 1.53, reveals a set of N-S–trending, high-conductivity (>0.03 S/m) vertical tabular “tracks” near the base of the crust (~40 km depth), with conductive pipe-like “fingers” rising off these tracks into the middle to upper crust. The western track roughly follows the trace of the normal-fault-bounded Sawatch-Mosquito-Tenmile-Gore Ranges, but the eastern track does not obviously relate to any clear geomorphic expressions of extension.Resolution tests (Item S3) indicate that these conductivity structures are robust. Our data require the spatial pattern of elevated conductivity (>0.01 S/m) along discrete tracks. However, only the northern portions of the tracks are strictly constrained to be highly conductive (>0.03 S/m); our data support high conductivity values (>0.03 S/m) along the entirety of the tracks, but the southern portions are only required to be moderately conductive (>0.014 S/m) with present data coverage. The “finger” structures are also robust, although, generally, these are only constrained to be moderately conductive (>0.014 S/m).We recognize two end-member explanations for our observations: elevated conductivity values are due either to existing lithologic assemblages under modern crustal physicochemical conditions or to introduced phases. Although we cannot uniquely discriminate between these two possibilities with existing data, we contend that a combination of introduced free saline aqueous fluids and partial melt best explains our observed electrical structures.The expected behavior of existing crustal rocks alone does not fully account for our observations. For example, volatile-bearing silicate minerals (e.g., amphibole, biotite) can account for high conductivity values at elevated temperatures (Li et al., 2017; Hu et al., 2018); however, upper-bound estimates on lower-crustal temperatures in central Colorado (Schutt et al., 2018) are such that these phases would only contribute to elevated conductivity near the base of the crust (Fig. 3). Grain-size reduction to <10 µm within large-displacement shear zones can lead to increased bulk conductivity (e.g., Pommier et al., 2018), but major shear zones that could yield such small grain sizes (cf. Young et al., 2022) are not observed at the surface coincident with the eastern conductive track. Highly conductive sulfidic and graphitic metasedimentary units can produce lineaments that follow lithotectonic structure (e.g., Murphy et al., 2023), but our N-S tracks are at a high angle to documented Proterozoic fabrics in central Colorado (cf. Frothingham et al., 2022).As existing crustal components cannot fully account for our observations, we recognize the importance of introduced phases. Hydrothermal graphite precipitated from deep magmatic fluids could account for high conductivity values (Murphy et al., 2022); however, we consider this explanation unlikely, as it is unclear if the necessary conditions for this mechanism (appropriate magmatic redox conditions and sufficient volume of mantle-derived magmatism) are satisfied. Instead, we consider a combination of free saline (>10% NaCl equivalent) aqueous fluids and partial melt to be the most reasonable explanation for our observations. These phases are highly electrically conductive (Fig. 3), and they are both common together in tectonomagmatically active regions (e.g., Wannamaker et al., 2008). Aqueous fluids are generated by devolatilization of the lower crust during heating and by expulsion of volatiles from stalled magmatic intrusions, and melt is due to anatexis of the lower crust and to introduction of mafic magma from the underlying hot mantle. (However, as felsic melts are less conductive than mafic melts, crustal anatexis is likely less important in explaining our observations than mantle-derived melts; see Fig. 3.) These phases are unstable on long geologic time scales (>10 m.y.); fluids tend to be consumed by hydration reactions (e.g., Yardley and Valley, 1997) or to escape upward (e.g., Ague, 2014; Manning, 2018), and melt will either crystallize or erupt. Therefore, the presence of these conductors indicates that these phases are currently being introduced into the crust within an active tectonomagmatic system.We interpret these conductors to represent zones of aqueous fluids and melt that are associated with modern, ongoing crustal extension in central Colorado. Our observation of these conductive zones occurring across the Rio Grande rift system meshes well with geodetic observations (e.g., Murray et al., 2019) that show distributed deformation beyond the limits of the Rio Grande rift proper.Although we consider a combination of saline aqueous free fluids and partial melt to be the best explanation for our observed conductivity values, we cannot uniquely determine with existing data either the physical arrangement of these phases or the manner in which extensional strain is being accommodated. Regardless, we can draw several conclusions here.Because the N-S conductive tracks generally follow variations or gradients in crustal thickness (Fig. 2B), we consider it most likely that fluids and melt are concentrated within broad zones of increased extensional strain, which also correspond to zones of enhanced permeability and porosity. Mid- to lower-crustal permeability (and porosity) is inferred to be very low, but small changes in hydraulic properties driven by plastic extensional deformation may yield large increases in fluid content (Manning and Ingebritsen, 1999) and, consequently, bulk electrical conductivity.Additionally, we view the lower-crustal conductive tracks as representing active formation of a new bulk crustal fabric under a generally E-W extensional stress field, rather than passive reactivation of an existing fabric, defined by a combination of shear zones and rock foliations. Regional NW-striking Proterozoic structural fabrics and overprinting NE-striking shear zones in this portion of Colorado are at high angles to our observed lower-crustal tracks. Consequently, Proterozoic structures, including the inferred “Colorado Mineral Belt shear zone” (e.g., McCoy et al., 2005), are apparently not prominent enough in the lower crust to influence patterns of modern crustal deformation in central Colorado (e.g., Caine et al., 2006). In contrast, NE-striking Proterozoic basement structures appear to control extensional deformation to some degree in the Rio Grande rift to the south (Minor et al., 2013).Due to inversion regularization, our conductivity images only provide an upper bound on the width of the high-conductivity tracks; it is possible that conductive phases are confined to laterally narrower regions in the lower crust. However, as these conductors fall below the brittle-plastic transition, we consider it unlikely that they mark discrete shear zones that steeply cut through the entire crustal column. Instead, these conductors more likely represent zones of bulk plastic deformation, with this extensional strain slightly increasing permeability (and porosity). In this context, it is unclear exactly how our observed geoelectric features structurally relate to mapped surface expressions of extensional deformation (Fig. 4), particularly as their geospatial relationships are ambiguous (Item S4).The nature of our observed mid-crustal “fingers” is at present unclear, as correlations with surface features are ambiguous (cf. Figs. 1 and 2). However, we consider it likely that they are transient features associated with fluid flow. They are particularly similar to transient pipes that arise in the “porosity-wave” model of fluid migration during heating-driven metamorphic devolatilization (e.g., Connolly, 2010).Finally, it is worth emphasizing the overall mismatch between the surface expression of extension and our deeper geophysical observations of the associated crustal structures (Fig. 4). Models of the Rio Grande rift often focus on mapped faults and rock exhumation within the narrow corridor of the rift proper (e.g., Abbey and Niemi, 2020); however, our observations demonstrate that rift-related processes are operating throughout the crustal column over a much wider region than the obvious surface expression of extension. In fact, the full width of our observed lower-crustal geoelectric signature of extension matches the width of rifting in southern segments of the Rio Grande rift, thereby exposing limitations in models built around inferred crustal differences between the northern (Colorado) and southern (New Mexico) segments of the rift (e.g., Abbey and Niemi, 2020). Furthermore, this observed disconnect between the lower crust and the surface, despite being observed in other extensional settings (e.g., Wannamaker et al., 2008), is poorly explained in traditional models of rifting. Our observations may reflect an ongoing transition from early broad rifting to later narrow rifting confined to the upper Arkansas River Valley, a progression that is noted in the North Sea rift and that may be a key characteristic of rift evolution (Fossen et al., 2021). Although rifts are often defined by their surface expressions, our observations nevertheless demonstrate that subsurface information is crucial in evaluating the full crustal impact and spatiotemporal evolution of extension.The electrical conductivity model is available at http://ds.iris.edu/ds/products/emc/. We thank the following organizations for permitting data acquisition on their lands: White River National Forest (Supervisor’s Office); Pike–San Isabel National Forest (South Park, South Platte, Leadville, Salida Ranger Districts); Colorado State Land Board (North Central, Northwestern, South Central Districts); Denver Water; Aurora Water; Parker Water & Sanitation District; Lookout Mountain Water District; Lakewood Parks; Lake County Board of County Commissioners; and Farmers Reservoir and Irrigation Company. We thank Evan Cox, Katrina Zamudio, Danny Piccone, Stephanie James, Ashleigh Miller, and Spencer Wilbur for field assistance. We thank Rob Harris, Tien Grauch, Derek Schutt, and two anonymous reviewers for feedback on this manuscript. B.S. Murphy was supported by a Mendenhall Postdoctoral Fellowship through the U.S. Geological Survey. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

中文翻译：

---

科罗拉多州中部活跃延伸的广泛空间足迹的地电证据

科罗拉多州中部的三维大地电磁（MT）成像揭示了一组位于下地壳（50-20公里）深度的北向高电导率轨迹，导电指状结构从这些轨迹升起进入中地壳（20 –5 公里深度）。我们将这些特征解释为代表活跃的伸展构造岩浆作用产物的含盐水性流体和部分熔体。这些导体分布在比传统上在地表绘制格兰德河裂谷结构的狭窄走廊更广泛的区域，因此它们划分了下地壳的区域，在这些区域中，体拉伸应变的调节具有集中的导电相。我们的观察揭示了格兰德河裂谷活动现有模型的局限性，可能反映了全球裂谷系统演化中未被认识的时空变化。尽管新生代中晚期（后拉拉米德）扩张影响了美国西部的大部分地区，但其区域性质和程度目前人们对其正在进行的情况仍知之甚少。科罗拉多州的东西向延伸（Levandowski 等，2018）通常被视为里奥格兰德裂谷的组成部分（图 1；例如，Hudson 和 Grauch，2013）。 Tweto（1979）最初区分了里奥格兰德裂谷“本身”和里奥格兰德裂谷“系统”。在科罗拉多州中部，前者由一系列狭窄的半地堑界定，这些地堑部分包括阿肯色河上游河谷，终止于正断层边界的戈尔山脉。后者包括更广泛的伸展断层和新近纪沉积区域，向东延伸到前岭，向北延伸到怀俄明州边界。尽管特威托有所不同，但各种数据集对伸展是否仅限于裂谷本身或分布在整个裂谷上提出了不同的解释。裂痕系统。例如，热年代学表明，新生代中晚期的伸展仅限于由正断层边界的 Sawatch、Mosquito-Tenmile 和 Gore Ranges 界定的狭窄走廊（图 1；Abbey 和 Niemi，2020）。这些范围内的年轻冷却年龄（<20 Ma）（例如，Landman 和 Flowers，2013 年），以及其他地方更古老的冷却年龄（> 45 Ma）（例如，Kelley 和 Chapin，2004 年），表明显着的拉酰胺后折返已经仅限于裂谷本身。相比之下，其他数据集表明，整个裂谷系统中已经发生并且仍在发生伸展变形（图 1）。在裂谷本身之外发现了新近纪伸展断层（Tweto，1979；Colman，1985），始新世侵蚀面的明显偏移（Epis 和 Chapin，1975）支持了地壳块体在主要地貌表现以东的轻微推挤。扩大。此外，年轻的（<10 Ma）火山岩仅出现在裂谷之外（图 1；例如，Cosca 等人，2014 年），大平原最西端的中新世标志地平线扭曲表明了影响深远的新近纪变形（Leonard，2002），大地测量显示科罗拉多高原和大平原西部之间广泛分布的伸展应变（例如，Murray等人，2019）。因此，在考虑地表过程与构造岩浆活动时，对科罗拉多州中部的扩张性质产生了相互矛盾的看法。以前的地球物理成像不足以确定科罗拉多州中部地壳如何以及在何处经历扩张。地震研究揭示了异常热的地幔（例如，MacCarthy等人，2014），但它们在地壳深度的分辨率不足以推断伸展过程。二维大地电磁 (MT) 研究显示了横向连续的下地壳导体，该导体被解释为格兰德河裂谷本身轴线之外延伸的标志（Feucht 等人，2017），尽管离线见解有限利用三维 MT 成像，我们发现科罗拉多州中部的中下地壳沿着区域离散的南北 (NS) 轨道存在异常导电性。我们得出的结论是，这些导体是活跃的（即现代的、目前正在演变的）构造岩浆作用的标志，其分布范围超出狭窄的格兰德裂谷本身，更广泛地分布在格兰德裂谷系统中。我们的结果揭示了里奥格兰德裂谷和一般裂谷过程模型的局限性。我们反演了科罗拉多州中部 81 个 MT 站点的全阻抗张量 (Z) 和垂直磁场传递函数 (VTF) 数据（图 1；数据可用）网址：http://ds.iris.edu/spud/emtf）。这些数据是通过多次调查收集的，并且成分传递函数（TF）最初是在不同时期估计的；因此，我们将数据（分别处理每个 TF 分量）插值到从 10−2 到 104 s 的 37 个周期的通用集合中。我们将 VTF 数据限制为 10−2 到 103 s。我们对 Z 的非对角分量执行 5% 的误差下限（分别处理每一行），对 VTF 执行 0.03（绝对）的误差下限。有关详细信息，请参阅补充材料 1 中的项目 S1。我们使用 ModEM 对这些数据进行了三维反演（Kelbert 等人，2014）。我们的模型网格的标称水平单元尺寸为 2.5 公里，并包含地形。在海拔 1 km 以上（相对于海平面），我们使用 100 m 的恒定垂直离散化，并将 ETOPO1（Amante 和 Eakins，2009）中的地形映射到网格上，以定义固定的 10−10 S/米气室。在该海拔以下，我们使用对数垂直离散化，最大深度扩展到约 930 公里；海平面以下 40 公里处的垂直单元尺寸约为 7 公里。我们的起始模型设置为导电 (0.1 S/m）西部大平原的沉积物（厚度定义为 ETOPO1 表面高程与 Marshak 等人，2017 年的大不整合高程之间的差异），位于由 SEO3 传导模型定义的与深度相关的电导率剖面之上（Constable，2006 年） ），等温线为 1330 °C，深度为 70 km，电导率值不低于 0.003 S/m。我们使用 ModEM 协方差设置，水平方向为 0.2，垂直方向为 0.4，有两个操作员通道；这导致模型结构在约 5 公里范围内横向平滑。我们的逆解（图 2；另请参阅补充材料中的项目 S2），将数据拟合为 1.53 的总体归一化（误差条）失配 (nRMSE) ，揭示了一组 NS 走向的高电导率 (>0.03 S/m) 垂直板状“轨道”，靠近地壳底部（约 40 公里深），导电管状“手指”从这些轨道升起进入中地壳到上地壳。西部轨迹大致沿着正断层边界的 Sawatch-Mosquito-Tenmile-Gore 山脉的轨迹，但东部轨迹与任何清晰的伸展地貌表现没有明显的关系。分辨率测试（项目 S3）表明这些传导性结构是稳健的。我们的数据需要沿离散轨道的电导率升高 (>0.01 S/m) 的空间模式。然而，只有轨道的北部部分被严格限制为高导电性（>0.03 S/m）；我们的数据支持整个轨道的高电导率值（>0.03 S/m），但在目前的数据覆盖范围内，仅要求南部部分具有中等电导率（>0.014 S/m）。 “指状”结构也很坚固，尽管一般来说，这些结构仅被限制为中等导电性（>0.014 S/m）。我们对观察结果有两种端元解释：电导率值升高要么是由于现有的岩性组合造成的在现代地壳物理化学条件下或引入相。尽管我们无法利用现有数据唯一区分这两种可能性，但我们认为引入的游离盐水流体和部分熔融的组合最好地解释了我们观察到的电结构。现有地壳岩石的预期行为本身并不能完全解释我们的观察结果。例如，含有挥发物的硅酸盐矿物（例如角闪石、黑云母）可以解释高温下的高电导率值（Li 等人，2017 年；Hu 等人，2018 年）；然而，科罗拉多州中部下地壳温度的上限估计（Schutt 等人，2018）表明这些相只会导致地壳底部附近电导率升高（图 3）。在大位移剪切区内将晶粒尺寸减小至 <10 µm 可导致体积电导率增加（例如，Pommier 等人，2018），但主要剪切区可能会产生如此小的晶粒尺寸（参见 Young 等人，2022）在与东部导电轨道重合的表面没有观察到。高导电性的硫化物和石墨变沉积单元可以产生遵循岩石构造结构的地貌（例如，Murphy 等人，2023），但我们的 NS 轨迹与科罗拉多州中部记录的元古代构造成高角度（参见 Frothingham 等人，2022） ）。由于现有的地壳成分无法完全解释我们的观察结果，我们认识到引入相的重要性。从深层岩浆流体中沉淀出来的热液石墨可以解释高电导率值（Murphy 等人，2022）；然而，我们认为这种解释不太可能，因为尚不清楚这种机制的必要条件（适当的岩浆氧化还原条件和足够量的地幔岩浆作用）是否得到满足。相反，我们认为游离盐水（> 10% NaCl 当量）水性液体和部分熔化的组合是对我们的观察结果最合理的解释。这些相具有高导电性（图 3），并且它们在构造岩浆活跃区域中很常见（例如，Wannamaker 等，2008）。含水流体是由下地壳在加热过程中的脱挥发分以及从停滞的岩浆侵入中排出挥发物而产生的，而熔融则是由于下地壳的深熔作用以及从下面的热地幔引入镁铁质岩浆而产生的。 （然而，由于长英质熔体的导电性低于镁铁质熔体，因此地壳深熔在解释我们的观测结果时可能不如地幔衍生熔体那么重要；见图3。）这些相在较长的地质时间尺度（>10 my）上是不稳定的；流体往往会被水合反应消耗（例如，Yardley and Valley，1997）或向上逸出（例如，Ague，2014；Manning，2018），并且熔化会结晶或喷发。因此，这些导体的存在表明这些相目前正在被引入活跃的构造岩浆系统内的地壳中。我们将这些导体解释为代表与科罗拉多州中部现代持续的地壳扩张相关的水性流体和熔体区域。我们对格兰德河裂谷系统中发生的这些导电区域的观察与大地测量观测（例如，Murray 等人，2019）很好地吻合，后者显示了超出里奥格兰德裂谷本身限制的分布式变形。尽管我们考虑了盐水的组合尽管自由流体和部分熔融是我们观察到的电导率值的最佳解释，但我们无法利用现有数据唯一地确定这些相的物理排列或适应拉伸应变的方式。无论如何，我们可以在这里得出几个结论。由于 NS 传导轨迹通常遵循地壳厚度的变化或梯度（图 2B），因此我们认为流体和熔体最有可能集中在拉伸应变增加的广阔区域内，这也对应于渗透率和孔隙度增强的区域。推断中下地壳渗透率（和孔隙度）非常低，但塑性拉伸变形驱动的水力特性的微小变化可能会导致流体含量大幅增加（Manning 和 Ingebritsen，1999），从而导致体电导率大幅增加此外，我们认为下地壳传导轨迹代表了在总体东西向拉伸应力场下新的大块地壳结构的主动形成，而不是由剪切带和岩石叶状结构的组合定义的现有结构的被动重新激活。科罗拉多州这部分区域的西北走向的元古代结构织物和叠印的东北走向的剪切带与我们观察到的下地壳轨迹呈高角度。因此，元古代结构，包括推断的“科罗拉多矿带剪切带”（例如，McCoy 等，2005），在下地壳中显然不够突出，不足以影响科罗拉多州中部现代地壳变形的模式（例如，Caine 等）等，2006）。相比之下，NE 走向的元古代基底结构似乎在某种程度上控制了南部里奥格兰德裂谷的伸展变形（Minor 等，2013）。由于反演正则化，我们的电导率图像仅提供宽度的上限高电导率轨道；导电相可能仅限于下地壳横向较窄的区域。然而，当这些导体低于脆塑性转变时，我们认为它们不太可能标记出陡峭地切穿整个地壳柱的离散剪切带。相反，这些导体更可能代表体塑性变形区域，这种拉伸应变会稍微增加渗透性（和孔隙率）。在这种情况下，尚不清楚我们观察到的地电特征在结构上与拉伸变形的映射表面表达有何关系（图 4），特别是因为它们的地理空间关系不明确（项目 S4）。我们观察到的中地壳“手指”的性质目前还不清楚，因为与表面特征的相关性不明确（参见图 1 和图 2）。然而，我们认为它们很可能是与流体流动相关的瞬态特征。它们与加热驱动变质脱挥发分过程中流体运移的“孔隙波”模型中出现的瞬态管道特别相似（例如，Connolly，2010）。最后，值得强调的是，延伸和表面表达之间的整体不匹配。我们对相关地壳结构进行更深入的地球物理观测（图4）。里奥格兰德裂谷的模型通常侧重于裂谷本身狭窄走廊内绘制的断层和岩石折返（例如，Abbey 和 Niemi，2020）；然而，我们的观察表明，与裂谷相关的过程在整个地壳柱中运行，其区域比明显的表面延伸表现要宽得多。事实上，我们观察到的下地壳地电延伸特征的全宽度与里奥格兰德裂谷南部部分的裂谷宽度相匹配，从而暴露了围绕北部（科罗拉多州）和南部（新墨西哥州）之间推断的地壳差异建立的模型的局限性。墨西哥）裂谷的各个部分（例如，Abbey 和 Niemi，2020）。此外，尽管在其他伸展环境中也观察到了这种下地壳与地表之间的脱节（例如，Wannamaker 等，2008），但在传统的裂谷模型中却很难解释。我们的观察可能反映了从早期宽裂谷到后来仅限于阿肯色河谷上游的窄裂谷的持续转变，这一过程在北海裂谷中被注意到，这可能是裂谷演化的一个关键特征（Fossen等人，2021） ）。尽管裂谷通常由其表面表达来定义，但我们的观察结果表明，地下信息对于评估完整的地壳影响和伸展的时空演化至关重要。电导率模型可在 http://ds.iris.edu/ds/ 上找到。产品/EMC/。我们感谢以下组织允许在其土地上获取数据：白河国家森林（主管办公室）；派克-圣伊莎贝尔国家森林（南公园、南普拉特、莱德维尔、萨利达护林区）；科罗拉多州土地局（中北区、西北区、中南区）；丹佛水务公司；极光水；帕克水和卫生区；瞭望山水区；莱克伍德公园；莱克县县专员委员会；以及农民水库和灌溉公司。我们感谢 Evan Cox、Katrina Zamudio、Danny Piccone、Stephanie James、Ashleigh Miller 和 Spencer Wilbur 的现场协助。我们感谢 Rob Harris、Tien Grauch、Derek Schutt 和两位匿名审稿人对本文的反馈。 BS Murphy 得到了美国地质调查局门登霍尔博士后奖学金的支持。任何贸易、公司或产品名称的使用仅用于描述目的，并不意味着美国政府的认可。我们的观察可能反映了从早期宽裂谷到后来仅限于阿肯色河谷上游的窄裂谷的持续转变，这一过程在北海裂谷中被注意到，这可能是裂谷演化的一个关键特征（Fossen等人，2021） ）。尽管裂谷通常由其表面表达来定义，但我们的观察结果表明，地下信息对于评估完整的地壳影响和伸展的时空演化至关重要。电导率模型可在 http://ds.iris.edu/ds/ 上找到。产品/EMC/。我们感谢以下组织允许在其土地上获取数据：白河国家森林（主管办公室）；派克-圣伊莎贝尔国家森林（南公园、南普拉特、莱德维尔、萨利达护林区）；科罗拉多州土地局（中北区、西北区、中南区）；丹佛水务公司；极光水；帕克水和卫生区；瞭望山水区；莱克伍德公园；莱克县县专员委员会；以及农民水库和灌溉公司。我们感谢 Evan Cox、Katrina Zamudio、Danny Piccone、Stephanie James、Ashleigh Miller 和 Spencer Wilbur 的现场协助。我们感谢 Rob Harris、Tien Grauch、Derek Schutt 和两位匿名审稿人对本文的反馈。 BS Murphy 得到了美国地质调查局门登霍尔博士后奖学金的支持。任何贸易、公司或产品名称的使用仅用于描述目的，并不意味着美国政府的认可。我们的观察可能反映了从早期宽裂谷到后来仅限于阿肯色河谷上游的窄裂谷的持续转变，这一过程在北海裂谷中被注意到，这可能是裂谷演化的一个关键特征（Fossen等人，2021） ）。尽管裂谷通常由其表面表达来定义，但我们的观察结果表明，地下信息对于评估完整的地壳影响和伸展的时空演化至关重要。电导率模型可在 http://ds.iris.edu/ds/ 上找到。产品/EMC/。我们感谢以下组织允许在其土地上获取数据：白河国家森林（主管办公室）；派克-圣伊莎贝尔国家森林（南公园、南普拉特、莱德维尔、萨利达护林区）；科罗拉多州土地局（中北区、西北区、中南区）；丹佛水务公司；极光水；帕克水和卫生区；瞭望山水区；莱克伍德公园；莱克县县专员委员会；以及农民水库和灌溉公司。我们感谢 Evan Cox、Katrina Zamudio、Danny Piccone、Stephanie James、Ashleigh Miller 和 Spencer Wilbur 的现场协助。我们感谢 Rob Harris、Tien Grauch、Derek Schutt 和两位匿名审稿人对本文的反馈。 BS Murphy 得到了美国地质调查局门登霍尔博士后奖学金的支持。任何贸易、公司或产品名称的使用仅用于描述目的，并不意味着美国政府的认可。或产品名称仅用于描述目的，并不意味着美国政府的认可。或产品名称仅用于描述目的，并不意味着美国政府的认可。