Continental-Scale Investigation of Underlying Electrical Conductivity Structure in Mainland China Using Geomagnetic Data

Abstract

The magnetotelluric method has been used to fully study regional electrical conductivity structures in different areas in mainland China; however, there is a lack of overall understanding of the electrical structure distribution. A novel insight for the study of continental-scale underlying conductivity structures was proposed in this work via geomagnetic data recorded by permanent stations. To study the underlying electrical structure distribution in mainland China, we mapped the conductors and resistors at a depth range of 4–100 km beneath mainland China using Parkinson vectors through magnetic transfer function. Three-component geomagnetic data within a low artificial disturbance period (local time 23:00–05:00) from 98 stations in 2019 were collected and processed to derive Parkinson vectors in the frequency band of 0.001–0.5 Hz. The distribution of subsurface electrical structures at distinct depths was constructed using corresponding frequency through the skin depth. We compare the consistent results herein with previous magnetotelluric studies, which indicated the reliability of our method. Combining previous multiple geophysical inversion results, we found that large-scale plastic bodies are distributed along the east of the Qinghai-Tibet Plateau and extend to the west of Yunnan. In central mainland China, the areas are mainly highly resistive, indicating that the structures are overall rigid. In north China, there exist high-low-high-low conductive structures from west to east. The separate high- and low-conductive electrical bodies in the North China Craton provide geophysical evidence that the Craton is composed of multiple blocks. The distributions of the underlying electrical structures in this work can provide an overall perspective for studying tectonic evolution and geodynamics in mainland China.

1. Introduction

The magnetotelluric (MT) method is an effective method for investigating the electrical conductivity of the Earth’s interior. In mainland China, the MT method has been widely applied to survey the distribution of regional underlying electrical conductors and resistors [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. Xiao et al. [8] found a relatively high-conductivity layer in the middle-lower crust at depths of approximately 10–100 km in the northeastern part of the Qinghai-Tibet Plateau (QTP). Bai et al. [15] constructed four MT profiles utilizing 325 sites within a total length of 2110 km and found high-conductivity zones at depths of 10–100 km in the eastern QTP. Ye et al. [10] revealed an upper and mid-crustal low resistive body above a depth of 10–30 km in the southeastern QTP and western Yunnan by inversing a crustal resistivity model using MT data collected from 170 sites. These studies demonstrate that, from north to south, high-conductivity layers exist in different areas of the eastern QTP. In southwest China, a high-resistivity belt [16] and a resistive block [12] were found based on MT profiles over 1000 km. In addition, fault zone conductors and lower crust conductivity structures were revealed in southwest China utilizing two profiles of approximately 250 km [17]. In northern China, the complicated North China Craton has become a focal area of current research on the underlying electrical layers using the MT method [18], where resistors [19,20] and conductors [11,21,22] have been studied. The locations of underlying electrical structures by the MT method mentioned above are shown in Figure 1. MT can provide lateral and vertical inversion slices to indicate electrical bodies and is mainly utilized to study tectonic evolution [23], geodynamic processes [15], and earthquake triggers [16] of the interior Earth. The above studies have reported many electrical conductivity structures in different regions, but there is a lack of continental-scale electrical structure distribution in mainland China [24]. Furthermore, for data collection, the MT sites can only be sited within a broad corridor of hundreds or even thousands of kilometers in length of the profile lines, which requires substantial field work [25]. Alternatively, geomagnetic data can be another option to study underlying conductivity structures.

Typically, geomagnetic fields observed by ground-based magnetometers contain multiple factors besides the Earth’s main field [26], including solar activity [27,28,29], ground vibrations [30,31,32], and underground conductive structures effect [33,34,35]. The short-term changes in the three geomagnetic components X (north-south), Y (east-west), and Z components (or H (horizontal), D (declination), and Z (vertical) components) tend to be confined to a plane called the preferred plane [36]. The plane will be almost horizontal if strata are the horizontal layers, but it is sometimes tilted due to the asymmetrically induced currents owing to oceans or regional underlying conductivity structures [36]. The reverse inclination of the plane reflects the direction of the underlying conductivity structure, and a larger dip of the plane indicates a greater difference in conductivity [37]. To indicate the inclination of the plane, the Parkinson vector was proposed to point toward high-conductivity electrical structures [36,37]. The direction of the vectors indicates highly conductive areas and their magnitude reflects the electrical difference [37]. The vectors depend on the frequency and can be calculated at specific frequency bands through the magnetic transfer function [36,38]. Generally, Parkinson vectors are utilized to study two aspects. On the one hand, they are utilized to study the geomagnetic coast effect [39]. The Parkinson vectors at the coastal observatories point toward the ocean and are almost perpendicular to the coastline, which indicates that the conductivity of the ocean is higher than that of the inland continent. Typically, the further the station is from the coastline to the inland, the weaker the impact of the geomagnetic coast effect [39]. This effect has been observed in many countries worldwide [40,41,42,43,44]. On the other hand, the Parkinson vectors are utilized to study the distribution of underlying electrical conductivity structure in inland areas, such as sharp lateral discontinuities in central Argentina Alicia [45], a high-conductivity layer in northern China [46], and two conductors beneath western Junggar and southwestern Chinese Altaids [47]. In addition, the temporal variations of Parkinson vectors have drawn the attention of scientists in recent years [33,48,49].

The Parkinson vector is an alternative indicator for mapping the distribution of conductors or resistors. Its advantage is that instead of deploying thousands of kilometers of MT profiles, we can use permanent geomagnetic observatories to investigate underlying electrical bodies. In mainland China, three-axis fluxgate magnetometers with a sampling interval of one second are installed at almost 100 permanent geomagnetic stations to monitor daily geomagnetic fields (the north-south (X), east-west (Y), and vertical components (Z)). The geomagnetic station network provides a good opportunity to study the continental-scale underlying electrical structures. In this study, we utilized geomagnetic data collected from 98 permanent stations in the Geomagnetic Network of China in 2019 (Figure 1) to investigate the large-scale spatial distribution of underlying conductivity structures by Parkinson vectors.

2. Methodology and Results

The magnetic transfer function [36,38] describes the relationship between the horizontal and vertical components in the geomagnetic field, which can be written as,

$$Z\left(f\right)=A\left(f\right)·X\left(f\right)+B\left(f\right)·Y\left(f\right)$$

(1)

X(f), Y(f), and Z(f) are the north-south, east-west, and vertical components of the geomagnetic data at a particular frequency band f, respectively. A(f) and B(f) are coefficients of the magnetic transfer function. We computed X(f), Y(f), and Z(f) utilizing a moving window of 180 min in step of one minute. A(f) and B(f) are calculated from X(f), Y(f), and Z(f) utilizing least square method. Note that A(f) and B(f) comprise the real parts Ar(f) and Br(f) and the imaginary parts of Au(f) and Bu(f), which can be written as follows,

$$A\left(f\right)=Ar\left(f\right)+iAu\left(f\right)$$

(2)

$$B\left(f\right)=Br\left(f\right)+iBu\left(f\right)$$

(3)

The three times the standard deviations of Ar(f) and Br(f) are computed and determined as the thresholds. Ar(f) and Br(f) are utilized to define and compute the Parkinson vectors; their absolute values are smaller than the thresholds for the mitigation of unwanted interference owing to sudden perturbations. The Parkinson vector is defined by the azimuth PA(f) and magnitude PM(f), and they can be calculated by Ar(f) and Br(f), respectively, using the following formulas,

$$PA\left(f\right)=arctan \left(\frac{Br\left(f\right)}{Ar\left(f\right)}\right)+180°$$

(4)

$$PM\left(f\right)={\left(Ar{\left(f\right)}^2+Br{\left(f\right)}^2\right)}^{1/2}$$

(5)

The azimuth distribution during the study period is constructed by utilizing the entire PA(f) with 36 angle bins at 360° with an interval of 10° [33]. The major distribution of the azimuths (PA_major) is considered to be the long-term direction of the Parkinson vectors at one station which reflect the electrical conductivity structures or geomagnetic coast effect because the underlying structures are generally considered to be unchanged within one year. PM(f) with PA(f), which is distributed within the bin of PA_major, is selected, and the major distribution value of the selected PM(f) is determined as PM_major, which reflects the electrical difference at deep depths. We used data from the QZ station in Figure 1 as an example to demonstrate the processes for retrieving PA_major and PM_major. Geomagnetic data between 23:00 and 5:00, local time, were chosen to mitigate the influence of artificial noise. Figure 2a shows the one-day (i.e., 1 January 2019) raw three-component geomagnetic data at QZ station. The red line indicates the low artificial noise period (i.e., local time (LT) 23:00–05:00) data that were utilized in this study. Figure 2b shows the sounding curves using data by LT 23:00–05:00 and the other period. We can see the errors are smaller in the time period of LT 23:00–05:00 than in the other period.

To analyze the distribution of the electrical structure at different depths, we preliminarily utilized the 1-D electrical resistivity mean model in mainland China [24], (Figure 3a) to compute the average resistivity ρ(h) at a depth of h, as shown in Figure 3b. The depth range was between 0 and 300 km from the Earth’s surface in the model (Figure 3a,b). The relationship between depth and frequency can be calculated by skin depth [50], which can be written as,

$$Ds\approx 0.503{\left(\rho \left(h\right)/f\right)}^{1/2}$$

(6)

where Ds is the skin depth (km), ρ(h) is the average value of electrical resistivity (Ω·m) from the Earth’s surface to a depth of h (km), and f is the frequency (Hz). Figure 3c shows the depth-frequency curve. Considering that the window length applied to the geomagnetic data time series was 180 min (i.e., 10,800 s), to extract the information of underlying electrical structures effectively, we chose 0.001 Hz (i.e., the period is roughly one-eighth of the window length) as the lowest study frequency. In this work, six frequency bands 0.001–0.005 Hz, 0.005–0.01 Hz, 0.01–0.05 Hz, 0.05–0.1 Hz, 0.1–0.2 Hz and 0.2–0.5 Hz were utilized to calculated Parkinson vectors, and the corresponding depth range are 44–100 km, 30–44 km, 14–30 km, 10–14 km, 7–10 km, and 4–7 km, respectively (

Table 1. The study frequency bands and corresponding depth.

Frequency (Hz)	0.001–0.005	0.005–0.01	0.01–0.05	0.05–0.1	0.1–0.2	0.2–0.5
Depth (km)	44–100	30–44	14–30	10–14	7–10	4–7

). The typical sounding curves of four stations QZ, LJ, SZ, and JX (see the yellow triangles in Figure 1) in our study area are shown in Figure 4.

Figure 5 shows the results of the calculation process for obtaining PA_major and PM_major in the frequency band of 0.001–0.005 Hz at the QZ station in 2019. Figure 5a,c show the real parts of the coefficients of the magnetic transfer functions (Ar and Br). The thresholds (mean + 3σ, σ the standard deviation, black dashed lines in Figure 5a,c) of Ar and Br are 2.32 and 3.75, respectively. Figure 5b shows the distribution of PA in 0°–360° with a 10° interval, and they are mainly distributed at 120° (i.e., PA_major, purple area). The radius denotes the counts of vectors in every azimuth bin. Figure 5d shows the distribution of all PM of obtained PA_major, with the main distribution value of PM (i.e., 0.405, yellow dashed line in Figure 5d) as the PM_major at the QZ station.

Parkinson vectors composed of PA_major and PM_major at all stations were obtained at depths of 4–7 km, 7–10 km, 10–14 km, 14–30 km, 30–44 km, and 44–100 km, denoted by red arrows in Figure 6. In the western region (longitude 72°–92°E), the vectors roughly point east at a depth of 4–14 km (Figure 6a–c), whereas the vectors from one or two stations point in other directions at a depth of 14–100 km (Figure 6d–f), which may be due to the inhomogeneous conductivity structures there. In the center of mainland China (longitude 92°–108°E), there are mainly two regions where the PA_major mostly points toward (or backward) at all study depths, which are indicated by the green (or blue) lines, respectively, in Figure 6a–f. Along the green lines (Figure 6a–f), the PA_major on both sides points to here, while at the blue lines (Figure 6a–f), the PA_major on both sides deviates from the region. This suggests that relatively high- or low-conductivity materials exist in these particular areas owing to the underlying inhomogeneous electrical materials. Note that the low-conductivity materials (blue lines) at two depth ranges of 30–45 km (Figure 6e) and 45–100 km (Figure 6f) are distributed northwest to those at the lower layers (Figure 6a–d). In the northern China (longitude 108°–120°E, latitude 32°–42°N), the vectors point to the areas of longitude 111°–113°E, latitude 35°–41°N at all study depths (i.e., green lines, Figure 6a–f), which indicates the highly conductive structures here. The vectors from stations in the western Bohai Sea point toward the sea, which may be owing to the geomagnetic coast effect at a depth of 4–14 km (Figure 6a–c), whereas at depths between 14 km and 100 km (Figure 6d–f), the vectors from several stations in the inland (southwest direction) area may be affected by highly conductive materials. In southwest mainland China, due to the geomagnetic coast effect, the vectors from the stations along the coast are directed toward the Yellow Sea, East China Sea, and South China Sea at all study depths (Figure 6a–f), which agrees with the observations by [41,51]. In the northeast of mainland China (longitude 115°–132°E, latitude 42°–52°N), the directions of vectors from three stations are almost the same at different study depths (Figure 6a–f). This indicates that high-conductivity materials near the stations in this area are distributed over a wide depth range. Nevertheless, we may obtain more information when more stations are established in this area in the future.

The Parkinson vectors point to conductors, and the backward of Parkinson vectors will point to the resistors [39]. The magnitude of the vector is inversely proportional to the distance from high-conductivity materials concluded from observed data and numerical modelling [43,44,52], that the magnitude is larger when the vector is close to conductors or resistors. Due to that the three obvious electrical materials (i.e., green, and blue lines in Figure 6) indicated by Parkinson vectors are roughly in the north-south direction and to clarify quantitatively the distribution of these materials, we construct the maps of electrical conductivity structures by PMP, defined by Formula (7),

$$PMP=\{\begin{array}{c}1/P{A}_{major}, 0°<P{A}_{major}<180°\\ -1/P{A}_{major}, 180°<P{A}_{major}<360°\end{array},$$

(7)

The 0 value areas of PMP in the maps indicate the conductors or resistors. The conductors are distributed in the areas where the left side $PMP$ values are positive and the right-side ones are negative. For resistors, the condition is reversed. The distributions of conductors and resistors are shown in Figure 7.

Figure 7a–f show the distribution of conductors and resistors in mainland China at depths of 4–7, 7–10, 10–14, 14–30, 30–44, and 44–100 km. We concentrated on the area of longitude 95°–120°E in this work because of the sparse station distribution in other regions (i.e., six stations in longitude 72°–95°E, three stations in latitude, 42°–52°N), and we divided the study area into main three parts from west to east, as shown in Figure 7. In part 1, a high-conductivity band at all study depth ranges (Figure 7a–f) exists, which is distributed along the east of the QTP and extends to the west of Yunnan. In part 2, overall high-resistivity bodies with some regional low-resistivity structures at all study depths (Figure 7a–f) exist, while at a depth between 44 km and 100 km (Figure 7f), the location of all electrical bodies is relatively west compared with the upper layers (Figure 7a–e). At the area of latitude 35°–40°N, the resistors at the depths of 4–14 km and 44–100 km (Figure 7a–c,f) are distributed at the longitudes 100°–108°E, which are nearly in the northwest-southeast direction, while the resistors at the depth of 14–45 km (Figure 7d,e) are located at longitude 108°E in north-south direction. At depths of 4–10 km (Figure 7a) and 14–44 km (Figure 7d,e), we note regional electrical bodies in the area of longitude 110°E, latitude 27°N. In Part 3, we note discontinuously distributed long conductors at an area of longitude 112–113°E, latitude 35°–45°N from 4 km to 100 km (Figure 7a–f). In the eastern neighboring of the conductors in this area, resistors exist at a depth of 4–44 km (Figure 7a–e). In the area of longitude 115–120°E, latitude 32°–38°N, the distribution of the electrical conductivity structures is almost the same at the depths between 4 km and 14 km (Figure 7a–c). In the 14–30 km (Figure 7d) and 30–44 km (Figure 7e) layers, a conductivity body is distributed at the same area of longitude 115°E, latitude 32°–38°N. At a depth of 44–100 km (Figure 7f), there are two resistivity bodies at longitudes 112–118°E and latitudes 34°–40°N.

3. Discussions

Herein, we compared the distribution of the underlying conductivity structures with results from previous studies utilizing the MT method. We retrieved the study areas of previous MT research in Figure 7 for better comparison. As for the long conductivity band in Part 1, previous MT studies have investigated the electrical conductivity structures in different regional areas in the eastern QTP [8,15] and western Yunnan [10], and the corresponding depth ranges of high-conductivity structures are 10–100 km [8,15] and 10–30 km [10], respectively. The long-conductivity band obtained in this study is distributed at a depth of 4–100 km (Figure 7a–f), which is consistent with the region in the eastern QTP at 10–100 km [8,15] (Figure 7c–f). Nevertheless, some discrepancies in depth exist in western Yunnan compared to Ye et al. [10] (i.e., 10–30 km, Figure 7c,d). Meanwhile, distinct geophysical data [53,54,55,56,57] were used to study the area in Part 1. Low-density anomalies at depths between 0 km and 250 km were revealed by the Earth Gravitational Model 2008 [56] over the entire area of Part 1. Low-velocity (Vs.) anomalies were found at depths between 5 km and 100 km in the same area [53,55]. Combining the results of long high-conductivity bands in this study and previous multiple geophysical inversion results in the area of Part 1, the high-conductivity, low-density, and low-velocity materials may be large-scale (i.e., latitude 22°–40°N) plastic bodies [53]. This indicates that the crust and upper mantel here are easily deformed due to the collision of the Indian and Asian continents. This can provide an explanation for the large-scale surface motion in the eastern Qinghai-Tibet Plateau [58].

For Part 2, the resistors from 35°N to 40°N had two types of distributions at different depths. One was toward the northwest-southeast direction (100°–107°E) at a depth of 4–14 km (Figure 7a–c), which is consistent with [23,59] (Figure 7a–c). The high-resistivity characteristic in this area was interpreted by the basalt distribution [59]. As in Li et al. [60], the other resistors were approximately in the north-south direction (107°E) at a depth of 14–44 km (Figure 7d,e). This high-resistivity layer represents a rigid structure at the west edge of the Ordos Block [60]. The discrepancy in conductivity at different depths in this area is consistent with the results of [46], indicating that conductivity significantly varied across the western edge of the Ordos block. As for the resistors at latitudes 30°–35°N, Zhao et al. [16] studied the electrical conductivity structures via the MT method using three profiles in this area, and found that high-resistivity bodies were distributed at depths between 0 km and 60 km, which almost coincide with our results at depths of 4–44 km (Figure 7a–e). In the deep layer of 44–100 km (Figure 7f), the high-resistivity bodies extended to the west compared to the shallow layers (Figure 7a–e); this phenomenon can be found in the MT profile by [16] (Figure 7a–f). The high-resistivity bodies in the shallow layers (Figure 7a–e) are referred to as the Pengguan Massif [61], a region of Proterozoic crystal-line rocks exposed in this area [16], and the deep resistors in Figure 7f may be associated with basement rocks [16]. In the case of the areas at latitude 20°–30°N in part 2, we noted distributed resistive bulks with regional conductors between them at 4–44 km (Figure 7a–e). In contrast, compared with the upper layers, the distribution of electrical bodies below 44 km (Figure 7f) was concentrated relatively west (Figure 7a–e). The spatial distribution of the electrical conductivity structures in this area is roughly consistent with the results of [12] (Figure 7a–e). Resistive bodies indicate the basic features of a stable Precambrian tectonic setting [12] (Figure 7a–e). In summary, the areas of Part 2 are mainly resistive, which suggests that the structures are overall rigid [11,21,62] in central mainland China. This indicates that stress is easily accumulated here and causes the region of strong seismic activity.

In Section 3, we compared the results of this study with those of previous studies [11,19,20,21,22,63]. The conductors in this area are consistent with the results of the MT sounding [11,21,22,63] (Figure 7a–f). Conversely, the resistors roughly correspond with the results of [19,20] (Figure 7a–f). In general, there are high-low-high-low conductive structures from west to east at a depth of 4–100 km (Figure 7a–f), which is consistent with [64] (Figure 7a–f), who investigated the electrical resistivity structures of the lithosphere beneath the entirety of North China (longitude 104°–125°E, latitude 35°–41°E) based on 1° (latitude) × 1° (longitude) standard MT array. The tectonic evolution of the North China Craton has been fully studied using geology, geochemistry, and geochronology [65,66]. The North China Craton is usually divided into three blocks: Western Blocks, Trans-North China Orogen, and Eastern Blocks [66]. Part 3 of this work was located in the North China Craton, and the distribution of the separated electrical bodies from west to east in this region provides geophysical evidence that the Craton is composed of multiple blocks.

Above all, we are confident of the locations of the electrical conductivity structures in this work by comparing them with those of previous studies using the MT method in mainland China. The results of two methods are consistent. This suggests that utilizing Parkinson vectors derived from permanent geomagnetic stations to map conductors and/or resistors is an effective alternative method for investigating the distribution of underlying conductivity structures.

4. Conclusions

In this study, we obtained the distribution of the underlying electrical conductivity structures in mainland China using Parkinson vectors through a magnetic transfer function utilizing permanent geomagnetic data. Generally, the Parkinson vectors in the west of China point to the eastern QTP. In central mainland China, the vectors deviate in this region. As for the northeast of China, the distribution of vectors is complicated, which reflect the complex tectonic activity here. The geomagnetic coast effect is clear at the southeast coast of China. We found a high-conductivity band distributed along the east of the QTP and extending to the west of Yunnan at depths of 4–100 km. Overall high-resistivity bodies at depths of 4–100 km with some regional high-conductivity structures at 4–44 km exist in central China. In addition, the separated high- and low-conductivity electrical bodies are distributed in the northeast of China.

Our results are highly consistent with the electrical resistivity structures obtained by the MT method, thereby indicating the reliability of our method. We divided the results into three parts and found that large-scale plastic high-conductivity bodies exist east of the QTP and west of Yunnan. This can provide an explanation for the large-scale surface motion in the eastern Qinghai-Tibet Plateau. In central mainland China, the highly resistive indicates overall rigid structures, which indicates that stress is easily accumulated here, and causes the region of strong seismic activity. The separated high-and low-conductivity electrical bodies in the North China Craton provide geophysical evidence that the Craton is composed of multiple blocks. The distribution of the underlying electrical conductivity structures in this work can provide an overall perspective for the study of tectonic evolution and geodynamics in mainland China.