Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin Based on EIGEN-6C4 Satellite Gravity and Aeromagnetic Data

The Gonghe-Guide Basin is situated in the northeastern edge of the Tibetan Plateau, one of China’s main targets for geothermal research and exploitation. We present a new crustal structure of the entire basin by using EIGEN-6C4 satellite gravity and aeromagnetic data aimed to further recognize the geothermal mechanisms. The data are processed using the wavelet multi-scale decomposition method and the iterative compact depth from extreme points imaging method. The gravity inversion results reveal distinct low-density regions in the middle-upper crust at 15–35 km depth, which is most likely caused by partial melting. The magnetic inversion results show negative or no magnetism within a similar depth range. But at shallow depths of the same horizontal positions, there is high positive magnetism, which can be interpreted as granite. The results correspond well with the magnetotelluric and seismic results and can be an effective supplement. To confirm and validate this conclusion, a 2D geological model of a profile from a typical area is created to show the detailed tectonic. Based on the new crustal structure results, the suggested geothermal target is the low/negative density corresponding to low/negative magnetism located at deep depths and high/positive magnetism located at shallow depths. Then we established a geological conceptual model to illustrate this process. This indicates that tectonic movement is taking place in the deep part of the earth in the Gonghe-Guide Basin. The research on the geological structure and geothermal heat source mechanism in the Gonghe-Guide Basin can provide a primary reference for research on geothermal resources in other areas with similar geological structures.


Introduction
The Gonghe-Guide Basin is a structurally complex basin located at the northeastern margin of the Tibetan Plateau. It is currently one of the main targets for geothermal research and exploitation in China (Xu et al., 2018a;Zhao et al., 2009). It is also considered one of the most crucial regions for studying the uplift mechanisms of the Tibetan Plateau (Clark & Royden, 2000;Gao et al., 2020;Zhang et al., 2011). In particular, with the increasing demand for clean energy, detecting geothermal energy requires new exploration techniques for widespread implementation (Li et al., 2015;Wang et al., 2018aWang et al., , 2018bWang et al., 2018aWang et al., , 2018b. A better understanding of the tectonic activity, geological structures, and distribution characteristics of deep materials will also allow for increased geothermal detection and exploitation. Furthermore, as an essential foundation for geothermal energy exploration in China , research on the geological structure and geothermal heat source mechanism in the Gonghe-Guide Basin can provide a primary reference for research on geothermal resources in other areas that have similar geological structure. The Gonghe-Guide Basin is a faulted basin surrounded by faults and folded uplifted mountains. The foothills of the Qinghainanshan Fault border it to the north, the Animaqing suture zone (an extension from Kunlun Fault [KLF]) to the south, the Wahongshan Fault to the west, and the Duohemao Fault to the east (Zheng et al., 2010) (Figs. 1 and 2a). The main faults are extensive NW trending sinistral strike-slip thrust faults (Fang et al., 2005;Zhang et al., 2020a). Hydrothermal and hot dry rock (HDR) have been found throughout the Gonghe-Guide area (Gao et al., 2018;Zhang et al., 2018a). HDR exceeding 200°C have been recently detected in the GR1 borehole (red dot in Fig. 1) with temperatures up to 236°C at a bottom depth of 3705 m, a first in China (Zhang et al., 2018b). The Gonghe-Guide Basin is now considered to have great potential for geothermal energy exploration and development.
Previous studies have obtained numerous geophysical surveys and extensive subsurface data to delineate crustal structures across the Gonghe-Guide Basin, using magnetotelluric (MT), seismic, gravity, and magnetic methods (Gao et al., 2018(Gao et al., , 2020Li et al., 2018;Zhao et al., 2020a). Three-dimensional MT imaging indicates a conductive layer in the middle-upper crust in the southeastern Gonghe-Guide Basin and a possible magma chamber near Gonghe town (Gao et al., 2018). Zhang et al. (2020b) and Tang et al. (2021) also found partial melting at depths of 15-35 km in the Gonghe-Guide Basin using MT data. In addition, seismic results show low-velocity zones at depths of 1-10 km, 25-40 km, and approximately 60 km in the basin . With Figure 1 Geological map of the Gonghe-Guide Basin after Fang et al. (2005) and Zhang et al. (2006). Red solid lines represent faults: LJSF Lajishan fault, QHNF Qinghainan fault, QHNSF Qinghainanshan fault, DHMF Duohemao fault, XJF Xinjie fault, WGBF Wayuxiangka-Guinan Buried fault, WHSF Wahongshan fault, GNSF Guinan south fault, KLF Kunlun fault. The dashed black lines are the basin boundaries multiple geophysical data sets, Zhao et al. (2020b) proposed a conceptual geothermal model for the Gonghe Basin with MT, seismic, Moho depth, and Curie point depth evidence. Zhao et al. (2020a) used 2D manual inversion and 3D cross-gradient joint inversion of the gravity and magnetic data collected from Qiabuqia, and indicated that this area is overlain by sedimentary layers approximately 1000-1500 m in thickness. Wang et al. (2021) inversed Moho depth and Curie point depth in and around the basin using gravity and magnetic data using an improved Parker-Oldenburg algorithm. Then, they established different thermal models and concluded that the high heat flow in the Gonghe-Guide Basin coacted with radiogenic heat.
In summary, studies have primarily focused on seismic and MT profiles and obtained the characteristics of local areas (Gao et al., 2018;Zhao et al., 2020b). Inversion interpretation of gravity data and magnetic data have also been performed, emphasizing the inversion of the interface, which reflects the fluctuation characteristics of the Moho surface and Curie point depth (Zhao et al., 2019(Zhao et al., , 2020bWang et al., 2021). However, subject to the accuracy and resolution of previous gravity and magnetic data, no complete three-dimensional crust structure has been established despite the limitations of the data Figure 2 a Topographic data of the Gonghe-Guide Basin, b satellite Bouguer gravity anomaly. Graphical indications as described in Fig. 1 Vol . 180, (2023) Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin 2737 processing methods. Therefore, it is difficult to research the three-dimensional distribution characteristics of the entire basin geological body. Furthermore, there is a lack of comprehensive research on a large-scale and overall lithological distribution of the entire basin, from shallow to deep. Therefore, there are many discrepancies in the interpretation of the geothermal mechanisms across the entire Gonghe-Guide Basin.
To obtain the crust structure and further describe the geothermal mechanism of the Gonghe-Guide Basin, we present a detailed study of the Gonghe-Guide Basin based on EIGEN-6C4 satellite gravity and aeromagnetic data. The lithological distribution and crustal structure of the whole basin are inversed. In addition, to verify and reflect the accuracy of inversion, a 2D geological model from typical area is created to show the detailed tectonics. Finally, target areas of geothermal resources are proposed, including their geophysical characteristics for detection, and the geological conceptual model is then used to determine the relationships between the deep crustal structure and geothermal source.

Geological Setting
The Gonghe-Guide Basin is located north of the KLF and between the Qaidam Basin and Qinling Orogen (Zhang et al., 2006). Tectonically, the Gonghe-Guide Basin is a Cenozoic intermontane basin controlled by the KLF and Altyn Tagh Fault (ATF) (Fig. 1;Fang et al., 2005). To the north, it is bordered by the foothills of the Qinghainanshan Fault (QHNSF), to the south by the Animaqing Suture Zone (extension from KLF), to the west by the Wahongshan Fault, and to the east by the Duohemao fault (Zhang et al., 2006). Between these first-order strike-slip faults, numerous secondary thrust faults, including the faults along Qaidam Basin, Qinghainanshan (QHNS), and Gonghenanshan (GHNS), contributed to the shortening and uplift of the mountains (Zhang et al., 2020a). Within the basin, the QHNS and GHNS are two prominent narrow ranges and thus comprise the range and basin landscapes in the interior of the northeastern Tibetan Plateau ( Fig. 2a; Craddock et al., 2014;Zhang et al., 2012).
The geological map ( Fig. 1) shows that surface is covered by the Quaternary system interior of the basin and its peripheral sporadic Neogene sediments (Sun et al., 2011). The Neoarchean-Early Proterozoic Jinshuikou Group is exposed in the southwest orogenic belt around the basin. The Early Proterozoic Daken-Daban Group is exposed in the northwest. The early Middle Triassic Longwuhe Group is exposed in the south, north, and east of the basin margin orogenic belt. Indosinian granites are primarily exposed in the west, north, and east of the orogenic belt around the basin. Multiple intrusions of different types of magma in the Indosinian period formed a compound granite batholith, which constitutes the main body of the basement of the Gonghe-Guide basin . In the post-collision period, lithosphere delamination occurred due to crustal thickening, and Late Triassic hot spring rocks developed on the east and west sides of the southern basin (Shi et al., 2018).
Many hydrothermal springs have been discovered along the NNW extending faults (Xue et al., 2013), and the hot water comes from the deeper part of the basin (* 1000 m) (Fang et al., 2009). Thermal survey results also reveal excellent geothermal conditions in this basin due to the geothermal gradient and terrestrial heat flow. Based on geophysical detection results, there is a conductive layer in the middle-upper crustal structures in the southeastern Gonghe Basin and a possible magma chamber near Gonghe town (Gao et al., 2018). There is a partial melting body at depths of 15-35 km in the western Gonghe Basin (Zhang et al., 2020b). The results have also been verified by seismic results . However, seismic and MT methods mainly indicated profile results (2D cross-section). It is difficult to display the geological distribution characteristics of the entire basin. Therefore, gravity and magnetic data are used in this paper to further explore the relationship between the geothermal heat source mechanism and crustal structure.

EIGEN-6C4 Satellite Gravity Data
Satellite gravity observation technology has high coverage, precision and resolution, which significantly compensates for the deficiency of ground gravity measurements. To effectively analyze the density structure of the Gonghe-Guide Basin, the EIGEN-6C4 satellite gravity data were used in this study. EIGEN-6C4 is a static global combined gravity field model up to degree and order 2190 (Foerste et al., 2014). The gravity model has high resolution (9 km) and precision (2.73 mGal). Figure 2a is the topographic data, and Fig. 2b shows the Bouguer gravity anomaly with a grid spacing of 5 km of the Gonghe-Guide area. The Bouguer gravity anomaly ( Fig. 2b) ranges from -520 to -320 mGal in the study area. The gravity anomaly characteristic agrees well with the topographic data (Fig. 2a). The low values in the southwest are the Kunlun Mountains, with high elevations and, therefore, a thick crust. The gravity value increases from the southwest part to the northeast. In the Gonghe-Guide Basin, the gravity anomaly has relatively few dramatic changes. The gravity anomalies and fault structure zones also correspond well.

Aeromagnetic Data
Magnetic anomaly maps provide insight into the subsurface structure and composition of the Earth's crust. The aeromagnetic anomaly grid used in our study area was derived from China Geological Survey (CGS). The aeromagnetic grid has a 2 km resolution grid of the magnetic intensity anomaly. The total field magnetic anomalies in the study area range between -250 nT and 360 nT (Fig. 3a). The magnetic anomalies in the Gonghe-Guide Basin are generally weak and evenly distributed. The prominent high magnetic anomalies are distributed in the structural belt around the basin. The dipole nature of magnetic anomalies makes them more challenging to interpret in terms of geological structure. Thus, the initial process of removing/minimizing the inclination effect is by transforming the TMI map into the Differential Reduction to the Pole (DRTP) map (Arkani-Hamed, 2007). This DRTP transform reduces the magnetic anomalies to the pole and corrects for variation in inclination and declination over the study area, assuming that all magnetization is induced. The mean inclination and declination values for the center of the study area ( Fig. 3a) are 55.37°and -1.29°, respectively.

Methodology
This paper carries out multi-scale data processing and analysis of satellite gravity and aeromagnetic data to obtain more precise geological structure interpretation results.

Wavelet Multi-Scale Decomposition Method for Potential Field Data Separation
The observed potential field data are the sum of gravity and magnetic effects at various depths in the subsurface half-space. The target anomalies must first be separated from the observed gravity and magnetic anomalies to study a specific geological problem using gravity and magnetic data. A variety of methods are proposed for separating gravity anomalies, such as upward continuation (Jacobsen, 1987), matched filtering (Spector & Grant, 1970), polynomial fitting (Telford et al., 1990), Wiener filtering (Pawlowski & Hansen, 1990), preferential continuation (Pawlowski, 1995), wavelet transformation multiple-scale decomposition (Fedi & Quarta, 1998) and nonlinear filtering (Keating & Pinet, 2011).
The wavelet multi-scale decomposition (WMSD) method is helpful for decomposing the gravity and magnetic data to obtain residual and regional anomalies at different depths (Fedi & Quarta, 1998;Xu et al., 2015Xu et al., , 2017. Yang et al. (2015) used this method to obtain the gravity effects at various depths of the Tibetan Plateau and analyze the geological structure at different depths. Xu et al. (2018aXu et al. ( , 2018b used an improved WMSD method to analyze the gravity anomaly of the Tibetan Plateau. The quality of results depends on the wavelet basis function and decomposition terms. It has been shown that the Coif3 wavelet basis function can be used to obtain a The residual gravity anomaly of the Gonghe-Guide Basin obtained by using the WMSD method. Graphical indications as described in Fig. 1 Vol . 180, (2023) Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin 2741 reasonable decomposition result under suitable conditions of data characteristics (Xu et al., 2017). The decomposition terms need to be specific for different data and depend on the depth of the equivalent layer. Therefore, the Coif3 wavelet basis function was used to implement the potential field separation in this paper. First, we decomposed the gravity and magnetic data, and multi-scale data were obtained. The average depths of the equivalent layers were estimated using a radially averaged logarithm power spectrum. Then, the residual anomaly (generated by crustal structures) and regional anomaly (generated by Moho) were obtained. Finally, the crustal structure was inverted using the residual anomaly.

Iterative Compact Depth from Extreme Points (ICDEXP) Imaging Algorithm
Inversion of potential field data is a powerful interpretation tool that provides a meaningful description of the source distribution (magnetization or density). Inversion methods are computationally expensive; moreover, the source models depend on the a priori information and constraints (Pilkington, 1996). In studying the crustal structure, it is often difficult to carry out inversion calculations directly because of the large amount of data and few constraints. In addition to inversion, imaging methods may provide a fast preliminary picture of the source distribution. Various imaging methods have been proposed and used to solve different problems (Fedi and Pilkington, 2012). They can provide an initial source model to be improved with more refined inversion algorithms.
Another popular method is the depth from extreme points (DEXP) imaging method, which has been successfully applied in target locations (Paletti et al., 2020), geological source imaging (Milano et al., 2016;Paoletti et al, 2016), and other studies. Based on DEXP imaging, Baniamerian et al. (2016) proposed a new algorithm called ''compact depth from extreme points,'' which iteratively produces different source distribution models with an increasing degree of compactness and, correspondingly, increasing source-density values. Liu et al. (2020) compared the ICDEXP method with the inversion method and found that the ICDEXP method does not require any matrix inversions, so extensive RAM is not needed, and the solution can converge faster than the inverse algorithm. Furthermore, the results of the ICDEXP and inversion methods are very similar. Therefore, this paper used the ICDEXP method to invert residual anomalies and obtain crustal structure. Figure 4 shows the results of WMSD, and the depth was calculated using the radially averaged logarithm power spectrum. In the wavelet decompositions, the regional component of the fourth term of wavelet decomposition corresponds to a depth of 55 km, the average depth to the Moho (Gao et al., 2020;Shen et al., 2016). Therefore, its regional component is thought to be caused by Moho. The distribution characteristics correspond to the Moho undulation. Therefore, we can use the result to calculate the depth of Moho, and the sum of the first, second, third, and fourth detail features of the wavelet decomposition can be used as the residual anomaly. The depths of the first ? second, third, and fourth wavelet decomposition are 5 km, 20 km, and 45 km. Compared with the original gravity anomaly, the characteristics of positive and negative local field alternation are more prominent. To obtain the density of the crustal structure, the fluctuation characteristics of the deep Moho were removed, resulting in the elimination of the overall low characteristics in the southwest and high characteristics in the northeast. In the interior of the basin, there are large areas of negative gravity anomalies. There are also apparent negative anomalies near the Wahongshan and Xinjie faults. Overall, there are numerous low-density anomaly field sources in the crust.

Results of Potential Data Separation
The gradient zone with sharp changes in the NNW direction between 101°E and 102°E in the east of the working area appears to be caused by the Xinjie fault. The fault is * 20 km wide and has a prominent semi-flower structure (Zhang et al., 2020a). It is the western boundary of the Guide Basin and the boundary between the Guide and Gonghe Basins. It mainly exposes banded Indosinian granites formed by sliding melting carried by the fault (Zhang et al., 2007). The exposed abundance of hot springs near the Xinjie fault is thought to correspond to this low-density gravity anomaly zone. The NW trending gradient zone with abnormally sharp changes distributed in the west is caused by hidden faults such as the Guinan Nanshan fault, Wahongshan-Wenquan fault, and Wayuxiangka-Guinan fault. Intermittent strip or elliptical low-density gravity anomalies correspond to negative landforms such as intermountain basins and valleys. Generally, it is inferred that NW trending structures control the distribution of low-density bodies near the Gonghe Basin in the west, and the distribution of low-density bodies near the Guide basin in the east is controlled by NNW trending structures.
Similarly, we use the WMSD to separate the residual and regional anomalies of magnetic data. Firstly, residual results of the first, second, third, and fourth wavelet decomposition (Fig. 5) were obtained, and the equivalent depths were calculated using a radially averaged logarithm power spectrum. To compare with geological information, the depths of first ? second, third, and fourth wavelet decomposition are 3 km, 9 km, and 15 km. Then the sum of the first, second, third, and fourth detail features of the wavelet decomposition can be used as the residual anomaly because the magnetism disappears under the Curie point depth. The residual magnetic anomaly obtained after removing the regional field can be regarded as the distribution of magnetic anomaly generated by the substances above the Curie point depth. The main distribution directions of magnetic anomaly strips are NNW and NW, which are consistent with the main distribution directions of gravity anomalies.
The main distribution direction in the western Gonghe Basin is NW and NNW in the eastern Guide Basin. The weak magnetism is distributed in the Gonghe-Guide Basin. The monotonic and unvaried magnetic fields are generated by the overlying sedimentary rocks. Around the basin, positive and negative magnetic anomalies are distributed alternately in strips, which are close to the strike of main faults. The maximum value is 350 nT, which may be caused by granite intruding along the fault and its alteration zone. The magnetic anomaly is more prominent in the boundary characteristics of the basin, and the basin is controlled by NNW and NW deep faults. This may be related to the Gonghe-Guide Basin being a faulted basin controlled by late-stage faults and magmatic intrusions. Thick sedimentary layers have accumulated in the basin, showing low magnetic anomalies. The late tectonic movement allowed the granite magma to intrude along the NNW and NW faults.
These granite intrusions and the surrounding hydrothermal alteration often show high magnetic anomalies. In addition, the positive anomalies scattered in the basin may be caused by concealed granite.

Inversion Results
According to the results of WMSD, the field value variation characteristics at different depths were obtained, and the depth information was estimated. The average depth of the Moho surface is 55 km. Therefore, D4 is a regional field, and we have obtained the residual field of the crust by using the sum of the first, second, third, and fourth wavelet decomposition, as shown in Fig. 6. Based on this information, the inversion of crustal structure density was calculated by using the ICDEXP algorithm. During the inversion calculation, for the satellite gravity data, the maximum depth was set as 50 km, the structure index was 3, the number of iterations was 20, and the upper and lower density limits were 1 and -1. The three-dimensional structure result and vertical slice diagram are shown in Fig. 7. The vertical slice AA' is a profile located outside the basin. The densities of this slice are low, and the negative data correspond to the Wahongshan fault.
The vertical slices BB' and CC' are located near Tanggemu and Gonghe, located in the west and east basin, respectively. The negative density data are prominent. In profile BB', the two negative densities correspond to the west Gonghe Basin and the Wahongshan fault. The vertical slice CC' goes through the Gonghe Basin and near the town of Guinan. Two negative densities are located near the towns of Gonghe and Guinan, which may be caused by deep partial melting. In the three-dimensional diagram and vertical section, the depth range of the density anomaly is clearly displayed, concentrated at a depth ranging between 15 and 35 km. Cross-section DD' passes through the town of Guide. The negative densities are also very obvious in this profile. The burial depth is 15 km * 35 km, corresponding to the Xinjie fault and deep partial melting near the town of Guide.
For display convenience, we intercepted the results of different depths (5,10,15,20,25,30,35 and 40 km), as shown in Fig. 8. The inversion results of different depths varied between methods. The sedimentary layers at shallow depths (up to 5 km) have little density fluctuation, consistent with the previous conclusion that the Cenozoic sedimentary thickness is greater than 5 km (Sun et al., 2011). The low-velocity bodies from near the surface to a depth of about 5-10 km are Quaternary loose or weakly diagenetic strata, and there is no strong density anomaly in the shallow part. At 10 km and beyond, the high-and low-density anomalies exhibit a clear staged pattern. In addition to the shallow sedimentary layer with little density fluctuation, the resolution of satellite gravity cannot obtain shallow details. The distribution of density anomalies changes significantly from 15 to 35 km. At depths greater than 35 km, the difference between high-and low-density anomalies begins to decrease. This shows that the density anomalies are mainly concentrated in the range of 15 * 35 km.
At depths of 15 km * 35 km, the low-density anomaly is mainly distributed in the corresponding main fault zones, and the difference between highand low-density anomalies is small and is concentrated in the Wahongshan fault on the west side of the basin, the Xinjie Fault, and the Duohemao fault on the east side. Moreover, partial melting of rocks at a certain depth causes a change in volume and density, which is expressed as low gravity in the gravity anomaly. The partial melt is the heat source of geothermal heat, and the heat flow rises through the upwelling channel provided by the fracture zone, forming a hydrothermal system. The Xinjie fault corresponds well with known hot spots, so it can be inferred that the geothermal energy in this area is mainly concentrated in these low-density bodies and their vicinity.
At depths [ 35 km, the density distribution characteristic changes significantly. At a depth of 40 km near the Xinjie fault zone, there are apparent lowdensity anomalies, and the low-density anomalies at other locations tend to disappear. This phenomenon can explain why the temperature increases successively from south to north. Generally speaking, the faults on the west side of the Gonghe-Guide Basin extend deep, mainly consisting of large faults extending into the lower crust.
Similarly, based on the wavelet multi-scale decomposition method, Fig. 9 shows the residual magnetic anomaly, which is the magnetic material feature above the Curie point depth. During the inversion, for the aeromagnetic data, the maximum depth was set as 30 km, the structure index was 3, the number of iterations was 20, and the upper and lower magnetization limits were 1 and -1. The threedimensional inversion results were obtained, as shown in Fig. 10. To compare with gravity data, slices were taken at the same locations as Fig. 7.
The magnetic variation characteristic from shallow to deep is prominent. In the AA' profile, magnetism is strong because of the intense tectonic Vol. 180, (2023) Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin 2749 activity. However, it is negative at the Wahongshan fault. Similarly, in the BB' section, there is a significant negative magnetic anomaly on the west side of the Gonghe Basin with a depth range of 6 * 24 km. The magnetism slowly dissipates with depth. In the CC' section, a strong positive magnetic body in the basin's center extends to 24 km and below. It is speculated that this is a hidden granite body. No partial melting has occurred. Near the towns of Chaka, Tanggemu, Gonghe, and Xinjie, small-scale high magnetic anomalies at shallow depths and low magnetic negative anomalies at deeper depths are found, indicating shallow granite bodies and deep partial melting areas. This is consistent with the gravity results and the depth range. The same characteristics occur in section DD'; the characteristics of the Guide area are very similar to Gonghe, both of which are shallow granite bodies with deep weak or negative magnetic anomaly characteristics, reflecting deep partial melting.
To facilitate analysis and display, slices at different depths were also made, as shown in Fig. 11. At 3 km and shallower depths, there are no significant magnetic anomalies, indicating no apparent magnetic body in the shallow part dominated by sedimentation. At depths less than 3 km, there is weak magnetism and no apparent magnetic anomaly. It is speculated that sedimentary rocks are the main cause of these characteristics. Furthermore, the resolution of the shallow data also can be a factor.
In all the results at depths [ 6 km, there are scattered magnetic anomalies, mainly concentrated in the edge of the basin, faults, fold zones, and granite outcropping areas in the western part of the basin. It is speculated that they are related to granite intrusions. The magnetic field inside the basin is consistent, and there is no obvious abnormal body. The magnetic bodies are mainly concentrated in the center of the basin.
At a depth of 6 km, an east-west elliptical positive magnetic anomaly appears in east Gonghe County. This location also corresponds with a negative gravity anomaly, speculated as the Qiabuqia dry hot rock mass. It has been confirmed that the Qiabuqia thermal rock mass is about 21 km long in the east-west direction and 14 km wide in the northsouth direction. The plane is nearly elliptical and develops stably within the depth of 21 km (Zhang et al., 2020a).
At a depth of 9 km, an NW bead-like positive magnetic anomaly is highlighted near the Wahongshan in the southwestern Wayuxiangka. There are dense EW and NW-W faults, numerous Permian and Triassic granites, and hot springs. There is also an apparent positive magnetic anomaly at a depth of 24 km, indicating that this area is an excellent geothermal exploration area.
In general, the gradient zone of magnetic anomaly from 15 to 24 km is gradually flattened, indicating that the existence of high underground temperatures leads to the disappearance or weakening of magnetism. It is speculated that high temperatures have demagnetized the material.
To further confirm and verify the correctness of the above interpretation, based on the inversion results, a 2D geological model of the CC' profile is created to show the detailed tectonic framework (Fig. 12). In the basin, the experimental density and magnetic susceptibility of granites are 2.532-2.555 g/ cm 3 and 0.0456 9 10 -3 * 13.847 9 10 -3 SI separately (Zhao et al., 2020a(Zhao et al., , 2020b. The partial melting is widespread in the upper-middle crust, where the density is set as 2.4-2.5 g/cm 3 . Correspondingly, middle and lower crust densities are set as 2.8 g/cm 3 and 3 g/cm 3 , respectively. To fit the magnetic data, we set the magnetic susceptibility of granite as 0.1 9 10 -3 SI for Ordovician granite and 0.05 9 10 -3 SI for Indosinian granite. The geological model of the CC' profile corresponds well with our inversion results, in which the partial melting is located deeper and the granites are shallow.

Crustal Partial Melting
Previous studies have shown that partial melting and fluids usually cause significant S-wave low-speed anomalies, and S-wave low-speed anomalies may also superimpose the effects of temperature increases and partial melting simultaneously (McKenzie et al., 2005). Furthermore, the MT results show that highconductivity regions can be interpreted as partial 2750 W. Zhou et al. Pure Appl. Geophys. melting in the crust (Tang et al., 2021;Zhang et al., 2021). However, the characteristics of density and magnetism have not been elucidated.
In this study, we use the wavelet multi-scale field separation method and three-dimensional fast iteration imaging to implement the inversion calculation of large-scale data. We make full use of the advantages of gravity and magnetic field data in the study of regional geological structures. Effective residual anomalies were obtained based on the wavelet multi-scale field separation method, and subsurface material distribution characteristics were inverted.
In the residual anomalies of the Gonghe-Guide Basin, a wide range of negative gravity and magnetic anomalies are displayed, which reflect the existence of low-density and magnetism geological bodies underlying the basin to a certain extent. The negative gravity anomalies correspond to the low resistivity and low velocity of previous results (Zhang et al., 2017). In addition, a temperature change is mainly manifested in the demagnetization of magnetic minerals and negative or weak magnetic anomalies. Specifically, when the rock temperature rises to the Curie temperature, rock magnetism disappears (Zeng et al., 2012). Based on this, we preliminarily believe that the negative or large consistent low-magnetic field area in the Gonghe-Guide Basin may be related to the demagnetization of magnetic minerals caused by deep thermal factors. Based on low density and weak magnetism characteristics, it is reasonable to speculate partial melting regions in the deep crust. However, it is not sufficient to extrapolate solely from residual anomalies.
In the study of large-scale crustal structural characteristics, the fast iteration imaging method (ICDEXP) reflects the advantages of calculation. The inversion results of gravity and magnetic data are consistent with the results of MT and seismic analysis, and the existence of partial melting regions in the Gonghe-Guide Basin is further shown. From the inversion results, the density distribution is uniform from the near-surface to a depth of * 5 km, indicating Quaternary loose or weakly diagenetic stratum. Low-density geological bodies dominate the middle and lower crust from * 15 to 35 km, mainly distributed in the towns of Gonghe, Chaka, Tanggemu, Xinjie, Chaka, Xinjei, Wayuxiangka and Guide of the basin. In these areas, there is also certain low/weak magnetism. This characteristic can be inferred as a part of the partial melting regions. We determined the characteristics by using gravity and magnetic data inversion. Thus, we concluded that partial melting regions in the crust are generally characterized by low resistivity, low velocity, low density, and weak or no magnetism.

Hot Dry Rock Target Areas and Their Characteristics
The idea of a partial melting area as a geothermal target needs further discussion. We found that density and magnetism correspond well at deep depths, with low density and low magnetism characteristics. However, as the main area of dry hot rock in the Gonghe-Guide Basin, the shallow parts must have granite distributed throughout (Gao et al., 2020;Tang et al., 2021). Therefore, the shallow depths should have specific characteristics consisting of high magnetic anomalies. Consequently, we further analyze the target distribution of the dry hot rocks.
The gravity and magnetic data show the partial melting regions of the entire basin and its surroundings, making it easier to delineate the existence of target areas. The results of the 3D inversion imaging prove this point. In the inversion results, the Gonghe-Guide basin shows a wide range of low-density and weak or low magnetic geological bodies, which suggests that there are high-temperature geological bodies or partial melting layers deep in the crust.
We can conclude that the partial melting regions are the key geothermal prospecting target but not completely. Another key factor is the shallow signature. Because the granite also has low density, magnetic data must be used to distinguish geothermal targets. Based upon this study, the targets are the low/ negative density corresponding to low/negative magnetism in the deep and high/positive magnetism in the shallow region. According to the geological map of the study area, the strong magnetic anomaly is mainly due to the existence of intrusive granite rocks and granodiorite in the study area. Therefore, in the Gonghe-Guide Basin, the key geothermal prospecting target may be located near the towns of Guide, Vol. 180, (2023) Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin 2751 Gonghe, Xinjie, Chaka, Wayuxiangka, Tanggemu, Xinhai. In addition, the Wahongshan and Xinjie faults may also be ideal geothermal prospecting targets.

The Geothermal Source Mechanism
Determining the geothermal source mechanism is the main problem in geothermal exploration in the Gonghe-Guide Basin, and it is also a common problem in geothermal resource exploration in other similar basins. Different authors have proposed different mechanisms. There are three primary hypotheses: (1) granite is the primary heat source (Zhang et al., 2018a;Zhang et al., 2019); (2) the geothermal source is located in deeper sections and transmits to the shallow part through the flowershaped faults (Zhang et al., 2018b;Gao et al., 2018); (3) the geothermal source is from the deep mantle .
These three views have certain rationality but also have shortcomings. The geothermal mechanisms are uncertain and have yet to be proven. The mechanisms of conduction and heat accumulation are also unclear. Overall, these disputes affect the understanding of dry hot rocks themselves and restrict the evaluation and efficient development and utilization of geothermal resources.
The hot springs in the region are primarily concentrated along faults. Hot springs with higher temperatures are not directly formed in the basin due to the influence of sedimentary layers. This shows that the heat source in the study area mainly comes from the crust and at depth. According to the inversion results in this paper, several large faults, such as Duomaohe, Xinjie, and Qinghainanshan, extend deep into the crust. These faults also have large high-temperature hot springs, suggesting that large-scale faults play a vital role in heat conduction and supply. The Gonghe-Guide Basin is a faulted basin with a thick sedimentary cover and fault zones controlled by the stress field of the Tibet Plateau. Due to the existence of large-scale high-temperature bodies and the negative topography of the basin, the interior of the basin exhibits negative density and weak magnetism.
Many fault zones around the basin allow hightemperature magma to intrude along the fault. The high-temperature alteration of the intrusive bodies and surrounding rocks makes prominent positive and negative alternating magnetic anomaly bands near the fault zone. The geothermal springs in the Gonghe-Guide Basin are mainly exposed near the fault zones on the east and west sides of the basin. Because of the numerous granite intrusions near the fault zone on the southwest side of the basin, magnetic anomalies controlled by structures are distributed in strips. It is speculated that the geothermal source here may be dominated by a large amount of deep molten materials.
In addition, there has been no volcanic or magmatic activity in the Gonghe Basin and its surrounding orogenic belt since the Cenozoic. The average heat generation rate of radioactive elements in drill core is 3.20 ± 1.07 lW/m 3 (Zhang et al., 2020c), which is slightly higher than the global average value (2.1-2.5 lW/m 3 ) of the heat generation rate of radioactive elements in Mesozoic Cenozoic granite (Artemieva et al., 2017). However, it is much lower than the basement granite in the area where the dry hot rock is produced in the Cooper Basin in Australia, with a heat generation rate of up to 7 * 10 lW/m 3 from radioactive elements within the rock (Beardsmore, 2004;McLaren et al., 2003). The value of granite heat generation is closely related to the geological age, and the global average value of radiogenic heat generation of Middle-Cenozoic granite (3.09 ± 1.62 lW/m 3 ) is close to the average heat generation rate of radioactive elements in the granite of the Gonghe-Guide Basin (Tang et al., 2020), which means that there are no obvious radioactive anomalies in the granite of the Gonghe-Guide Basin. Therefore, the residual heat from the granite magma and the heat generated by the decay of radioactive elements are not likely to be the main heat source.
In summary, as shown in Fig. 13, we suggest that the essential geothermal source comes from the mantle caused by substantial tectonic activity since the early Cenozoic (Craddock et al., 2014;Gao et al., 2020), causing partial melting in the crustal structure, and the heat flows from deep to shallow along the faults. The Quaternary sediments covering the surface of the Gonghe Basin serve as the geothermal system's caprock. The heat source structure of the Gonghe-Guide Basin is closely related to the geological structure in the crust.

Conclusions
The crustal structure of the entire Gonghe-Guide Basin was obtained based on the satellite gravity and aeromagnetic data, using the wavelet multi-scale decomposition field separation method and the iterative fast imaging inversion method, providing evidence for the existence of underground partial melting regions. The results show that the location of deep low density and low magnetic anomalies corresponds well to the partial melting regions and corresponds to the results of previous MT and seismic methods. In addition, the inversion results in this paper show that the regions near the towns of Guide, Gonghe, Xinjie, Chaka, Wayuxiangka, Tanggemu, Xinhai, and Xinjie have abnormal characteristics of low density, shallow high magnetism, and deep low magnetism. These areas have good geothermal target potential, consistent with the existing geological and borehole data. Therefore, in the inversion of gravity and magnetic anomalies, regions with similar characteristics can be used as targets of geothermal resources. Combined with other geological data, we suggest that geothermal genesis is occurs as heating from deep to shallow due to mantle material upwelling and early Cenozoic Indo-Chinese tectonic movement. The most important heat source comes from the mantle and tectonic activity, causing partial melting in the crust. The heat continues to flow along the faults into shallow strata, heating the surrounding rocks to adequate temperatures for geothermal exploration.

Funding
This research was partly supported by the National Natural Science Foundation (42004068)

Data Availability
The gravity data can be download from ICGEM (http://icgem.gfz-potsdam.de/home). The aeromagnetic data associated with this research are confidential and cannot be released.

Declarations
Conflict of Interest The authors have no relevant financial or non-financial interests to disclose.
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. Vol. 180, (2023) Crustal Structure and Geothermal Mechanism of the Gonghe-Guide Basin 2753