Based on the previous method, dF values from 2018 to 2020 were obtained, and the results are shown in Table 1. The difference in the lithospheric magnetic field between the two measurement points is about −120 nT, indicating the presence of significant magnetic anomalies between the two measurement points. Considering that the horizontal and vertical gradients of the two measuring points are relatively small, this anomaly cannot be a local anomaly of the measuring points, but more likely reflects the characteristics of the regional lithospheric magnetic field.

The changes in the magnetic field of the lithosphere from 2018 to 2020 were -2.2 nT in 2018-2019 and +3.9 nT in 2019-2020. The crustal compression or tension strain generated by tectonic loading in the Qinghai-Xizang Plateau region is usually around 10⁻⁸, which can only produce magnetic field changes of 10⁻¹ nT [30]. In the absence of severe crustal material migration, such as earthquakes and volcanic eruptions, it is generally believed that the changes in the magnetic field of the lithosphere do not change with time [28]. From 2018 to 2020, there was a change of about 6 nT in the lithospheric magnetic field, indicating that there may have been significant changes in crustal material during this period. At the same time, the reverse changes in the lithospheric magnetic field from 2018 to 2019 and from 2019 to 2020 indicate that the lithospheric magnetic field at the two measurement points experienced different effects in 2018, 2019, and 2020, thus exhibiting significant reversal changes.

Table 1. ThedF values from 2018 to 2020.

The seismic coseismic process is accompanied by significant stress release, which can lead to abnormal changes in the geomagnetic field, known as the piezomagnetic effect. The calculation methods for the piezomagnetic effect are divided into analytical methods and numerical calculation methods [31]. The analytical method has a small computational cost and is not prone to distortion, but it can only use uniform medium parameters. Although numerical calculation methods can consider three-dimensional medium parameters, their computational complexity is enormous and distortion may occur at fault planes. Considering the difficulty in obtaining accurate medium parameters, this section uses analytical methods to calculate the piezomagnetic field that may be generated by the Dingri earthquake. The parameters mainly include magnetization intensity (J), stress sensitivity coefficient (β), medium shear modulus (G), and sliding distribution (S) [31]. The values of β and J vary with factors such as lithology and burial depth [23], and precise determination is challenging [19]. Due to the linear relationship between the calculated piezomagnetic field and the product of magnetization and stress sensitivity coefficient (β J) [19], both β and J can be selected as 1 × 10⁻⁹Pa and 1 A/m to calculate the initial piezomagnetic field. When β J changes, the final result is the initial piezomagnetic field multiplied by the corresponding multiple of β J. The shear modulus of crustal media has an impact on the calculated piezomagnetic field [32], especially the mechanical properties of the shallow crust have a significant influence on the calculated piezomagnetic field [31]. Considering that the Dingri earthquake occurs in the crust with low wave velocity, corresponding to a P-wave velocity of about 5 km/s, the shear modulus of the crust is selected as 22 GPa. The magnetic declination and tilt angle refer to the results of IGRF14 [33], which are +0.3˚ (positive values indicate north northeast) and 45.2˚, respectively.

After the Dingri earthquake, different research institutions provided the slip distribution of coseismic faults based on seismic wave data. The rupture surface is roughly 50 to 80 km, and the rupture is mainly distributed north of the epicenter. However, the slip distribution given by different results varies greatly. Based on the preliminary results of the field survey, the surface dislocation in some areas is more than 3 m, and there is a possibility of a greater surface fault. Therefore, this paper selects a slip distribution model that is more consistent with the surface survey for calculation (from the Institute of Qinghai-Xizang Plateau Research, Chinese Academy of Sciences [34], see Figure 2).

The calculation results of the coseismic piezomagnetic field show (see Figure 3) that a significant piezomagnetic field appears at the upper edge and both ends of the sliding surface, with a maximum piezomagnetic field exceeding 1 nT, reflecting the controlling effect of shallow stress changes on the piezomagnetic field. The majority of the area along the sliding surface has a negative piezomagnetic field, with only a small range of positive values appearing on the north side. This is because the sliding amount is mainly distributed in the area north of the sliding surface. The piezomagnetic field at the epicenter is about +0.2 nT, and the piezomagnetic fields at the two measuring points are +0.04 nT and −0.04 nT, respectively, indicating that the piezomagnetic field decays rapidly with distance. Due to the relatively small values of β and J used in this article, most studies have used β and J that are 1.5 to 3 times higher than in this article [13], that is, β J is 3 to 10 times higher than in this article, and some studies have used larger dielectric magnetic parameters. When choosing β J to be 3 to 10 times that of this article, the piezomagnetic field at the epicenter can reach 0.5 to 2 nT, but the piezomagnetic field at the two measuring points is still relatively small, possibly 0.1 to 0.3 nT, which is difficult to observe by instruments.

Figure 2.Coseismic slip model of Dingri M6.8 earthquake.

Figure 3. Coseismic piezomagnetic fields expected from Dingri M6.8 earthquake.

Another magnetic field effect related to earthquakes is the electrokinetic effect, which can generate a flowing electric potential when liquid flows through a discontinuous underground interface, thereby producing a magnetic field [35][36]. This article is based on the analytical method of the electrokinetic effect generated by dislocation sources given in reference [36] to calculate the changes in the coseismic magnetic field of the Dingri earthquake. The model parameters include the length of the fracture surface (l), the burial depth at the top of the fracture surface (a), the burial depth at the bottom of the fracture surface (b), the conductivity on both sides of the fracture surface (σ₁ and σ₂), and the flow potential (S). According to the research results of electrical structure, there are high conductors and high resistance bodies in the crust of the Qinghai-Xizang Plateau and its surrounding areas, and violent earthquakes often occur at the junction of high conductors and high resistance bodies. Therefore, we choose σ₁ and σ₂ as 0.1 S/m and 0.01 S/m here. The flow potential generated by earthquake action is 0.1 to 10 V [35][37], and 1 V is selected in this paper. Based on the above parameters and the coseismic sliding distribution shown in Figure 2 (l is 54 km; a is 0 km; b is 25.3 km), the possible coseismic electromagnetic effects of the Dingri earthquake were calculated, as shown in Figure 4. Unlike the piezomagnetic effect, the range of magnetic field changes in the electrokinetic effect generated by the Dingri earthquake is relatively large. Magnetic fields greater than 0.1 nT are distributed along the rupture surface for about 300 km and perpendicular to the rupture surface for about 200 km. The electrokinetic effect magnetic field at the epicenter is +0.5 nT, and the magnetic fields at the two measuring points are +0.15 nT and −0.1 nT, respectively. The magnetic field changes at both ends of the rupture surface exceed 4 nT.

Figure 4. Coseismic electrokinetic fields expected from Dingri M6.8 earthquake.

The electrokinetic effect is not related to the sliding amount, but is closely related to the size of the sliding surface. The sliding distribution length used in this section is nearly 52 km, while some other models provide larger sliding surface sizes. The geothermal resources in southern Xizang are abundant, and there are hot water explosions, boiling springs, and other resources around the Dingri earthquake [38][39], which may cause an increase in electrical conductivity [40]. During the post-earthquake investigation, a large number of sand blasting and water emitting phenomena were found, indicating that the source of the electrokinetic effect is not limited to the sliding surface. From this, it can be inferred that the variation range of the electrokinetic effect magnetic field is larger, which is sufficient to generate the magnetic field changes observed by the instrument. In particular, the S2 measuring point is located in the high gradient zone of the magnetic field.

Some researchers have analyzed the relationship between the lithospheric magnetic field and earthquakes and structures based on geomagnetic anomaly grid data (EMAG). Earthquakes usually occur in areas where the lithospheric magnetic field is weakly negative or at the junction of positive and negative anomalies [11][13][41], and important active faults are also located in the transition zone between positive and negative anomalies of the lithospheric magnetic field [12][42]. Based on EMAG, the difference in lithospheric magnetic field between the two measurement points is about +18 nT, indicating that the lithospheric magnetic field at the two measurement points is basically the same. However, there is a significant difference from the measured results (about −120 nT). This difference may be due to that EMAG model incorporates no ground-based data and limited aeromagnetic data in the Qinghai-Xizang Plateau region, relying primarily on satellite magnetic measurements that only reflect long-wavelength (>400 km) components of the lithospheric magnetic field. Furthermore, as satellite magnetic surveys are conducted at altitudes exceeding 300 km above the Earth’s surface, the lithospheric magnetic field derived from the EMAG model in the Qinghai-Xizang Plateau area is essentially an inverted result, which leads to significant deviations from actual ground observations. This article only used data from two measurement points, so it is unclear whether the two measurement points are located in the negative anomaly zone and the positive anomaly zone, respectively. Based on the EMAG, it is speculated that the range of differences in the lithospheric magnetic field in this area is relatively small and should belong to local magnetic anomalies.

The lithospheric magnetic field contains rich information on tectonic evolution, crustal structure, and seismogenic background [43]. The composition of the lithospheric material in the Dingri area is relatively complex, including the High Himalayan crystalline complex, the North Ao Formation greenschist to amphibolite metamorphic complex, and the Southern Xizang Tethys sedimentary rocks. The boundaries of different lithologies are a series of normal faults [44][45]. The S1 measurement point is located in the Tethys sedimentary rock area, while the epicenter and S2 measurement points are located in the high-grade metamorphic rock area of the High Himalaya, exposing ferromagnesian granulite [45]. Therefore, the magnetic susceptibility of the medium on the west side of the epicenter should be lower than that on the east side. Based on this, a simple crustal magnetic susceptibility model was established to calculate the possible lithospheric magnetic field (Figure 5). The side with high magnetic susceptibility has a larger magnetic field, while the side with low magnetic susceptibility has a smaller magnetic field, which is consistent with the measured results and confirms the difference in rock properties on both sides of the epicenter.

Figure 5.Forward simulation of the magnetic field resulted from magnetic substances. cm is the susceptibility of the medium. Red, green, blue, and black lines are the magnetic anomaly 2 km, 1 km, 0.5 km, and 0 km above ground, respectively.

The Dengmocuo fault, which gave birth to the Dingri 6.8 magnitude earthquake, is a relatively small normal fault in the Southern Xizang Rift system. It is located on the southern edge of the Xainza-Dingye fault zone and is one of the more active faults in the region [6][46]. Many studies have shown that the normal faults within the Southern Xizang Rift system are regulated by the weak middle crust and serve as important channels for the exposure of lower crustal materials to the surface [2][44]. The high Himalayan metamorphic rocks that appear locally in the Dengmocuo fault indicate that the fault displaced Tethys sedimentary rocks and became a channel for high Himalayan metamorphic rocks to fold back into the upper crust [45]. The low-velocity layer of the middle crust is widely distributed below the Dengmocuo fault, which may correspond to high-temperature partial melting or water-bearing fluids. The low-velocity body on the west side of the fault is more widely distributed [47], resulting in negative magnetic anomalies on the west side of the fault, causing the magnetic field of the lithosphere on the west side to be smaller than that on the east side.

Research based on repeated observation data of the geomagnetic field shows that earthquakes often occur in areas with weak annual variations in the lithospheric magnetic field [22][41] or in areas where the annual variations of the lithospheric magnetic field have been continuously changing for many years [48][49]. This article is based on the repeated observation data of the geomagnetic field from 2018 to 2020. It was found that there was a significant turning point in the lithospheric magnetic field before the 6.8 magnitude earthquake, with an amplitude of about 6 nT. Is this an anomaly before the earthquake? A 5.9 magnitude earthquake occurred on March 20, 2020, with the epicenter located about 15 km northwest of the Dingri earthquake. Is the abnormal change in the lithospheric magnetic field discovered in this study related to the 2020 Dingri 5.9 magnitude earthquake?

From the quantitative calculation of the coseismic piezomagnetic effect and electrokinetic effect in the previous text, it can be seen that the piezomagnetic effect is mainly related to stress changes and the magnetic properties of the medium. Even with larger magnetic parameters, medium strong earthquakes such as the Dingri earthquake are difficult to generate geomagnetic field changes exceeding 1 nT at a distance of 20 km from the force source [50]. Therefore, pre-earthquake structural loading or micro-rupture stress disturbances are more difficult to produce significant geomagnetic field changes [48], which is different from some studies that suggest that pre-earthquake geomagnetic field anomalies may be related to stress changes [49].

The electrokinetic effect is mainly related to the distribution range of fluids, and fluid anomalies related to earthquakes usually have a wide distribution range. Fluid anomalies may exist within a range of several times the length of the fault rupture from the epicenter before and during the same earthquake [51], such as sustained water level anomalies and abnormal geomagnetic fields around the epicenter [52][53]. Based on previous analysis, the electrokinetic effect may generate geomagnetic anomalies exceeding 1 nT, and the range is large. Therefore, the geomagnetic field anomaly before the earthquake may be related to the electrokinetic effect generated by underground fluid anomalies. The Dingri earthquake is more than 5 years away from the anomaly of the lithospheric magnetic field from 2018 to 2020. Statistical results on fluid and magnetic anomalies indicate that the pre-earthquake anomaly for a 6.8 magnitude earthquake can occur from several hours to several months, even 1 year [9][51][54], and the possibility of anomalies occurring 5 years or even longer before the earthquake is unlikely. Therefore, it is inferred that the anomaly of the lithospheric magnetic field from 2018 to 2020 should not be related to the Dingri earthquake, but is more likely to be related to the 5.9 magnitude Dingri earthquake in 2020. The distance between two geomagnetic measurement points for the 5.9 magnitude earthquake in Dingri is about 35 km, while statistical analysis results indicate that the possible pre-earthquake anomaly range for the 5.9 magnitude earthquake in Dingri is about 200 km [55].

Based on the quantitative calculations in the previous text, it can be seen that significant magnetic field anomalies occur near faults and decay rapidly with distance. Although fluid and geomagnetic field anomalies before earthquakes may also occur in locations farther away from faults, the vast majority of anomalies occur in locations closer to faults. Therefore, setting up geomagnetic monitoring points near important faults makes it easier to observe magnetic field anomalies. At present, the geomagnetic repeated observation network is basically spatially uniformly distributed [56], which is advantageous for conducting research on the spatial distribution of the geomagnetic field and can observe some far-field anomalies, but it is not easy to observe more obvious near-fault anomalies.

Besides, temperature variations associated with seismic activity can also alter the geomagnetic field through thermomagnetic effects, where rock magnetism decreases with rising temperature and increases with cooling. According to Okubo et al. (2006) [57], temperature changes of several tens of degrees are required to generate significant magnetic variations (>1 nT), yet the magnitude of stress-induced thermal response is merely 1 mK/MPa [58], rendering the magnetic field changes caused by stress-derived temperature variations negligible. Whether other heat sources (e.g., heat transfer from subsurface fluid migration) could contribute to geomagnetic alterations remains unclear; future studies should incorporate regional thermal infrared anomalies, ground temperature data, and hydrological records to enable quantitative analysis.