A number of geodynamic models have been proposed for the southeastern margin of the Tibetan plateau and include a range of deformation processes. One unresolved issue is whether crustal and mantle flow occurs, and if it does, how flow contributes to the mass balance of the India‐Asia collision. To address this question, new magnetotelluric data were collected and used to derive a three‐dimensional electrical resistivity model of the crust and upper mantle beneath the Red River Fault (RRF) zone and adjacent areas. The most prominent features of the model are (1) a resistor in the upper‐middle crust directly beneath the trace of the RRF; (2) a major change in upper mantle resistivity across the RRF; and (3) a significant conductor in the upper mantle northeast of the RRF, which extends upward into the crust, and which requires a melt fraction of up to 3%. The model suggests that the lower crustal conductors may be due to melt/fluids derived from the mantle, rather than from outward flow from Tibet. The most likely source of fluids and melts could be upwelling mantle flow related to the Hainan mantle plume. The change in resistivity across the RRF implies a change in lithospheric strength may explain the present‐day localization of deformation and uplift at this location. The resistivity model may also give insights into the distribution of ore deposits in Ailao Shan, since many mineral deposits are derived from magmatic fluids generated in the mantle at the edges of regions of thick lithosphere.

Neural networks (NNs) are efficient tools for rapidly obtaining geoelectric models to solve magnetotelluric (MT) inversion problems. Training an NN with strong predictive power requires numerous training samples to prevent underfitting. To reduce the computational burden of generating a large number of training samples, this work analyzes the influence of the sample features and distribution on the training effect for an NN and proposes an efficient method of sample generation. This innovative method consists of three steps: 1) geoelectrically simplifying the features; 2) removing unnecessary features on the basis of realistic geological characteristics; and 3) mapping the samples to a higher-dimensional space. Numerical examples based on simple stratified models show that the number of samples can be reduced to below one-millionth of the original number while improving the predictive effect of the NN. The performance and effectiveness for processing more complex structures are verified by the inversion results obtained for a public data set, COPROD2. We conclude that this advanced method can generate high-quality training samples at a greatly reduced computational cost. The analysis of the sample features and distribution not only advances the state of research on the use of machine learning in geophysical inversion but also is a forward-looking study on the mechanisms of underfitting, tracing the source of these phenomena back to the training samples used.

The exploration of podiform chromites in the Indus Yarlong Zangbo suture zone of southern Tibet has proved difficult because most known deposits pinch out and then reappear in the same direction. Several ground-based geophysical approaches such as gravity, magnetic, and controlled-source audio-frequency magnetotelluric (CSAMT) methods have been applied to explore for these chromite deposits but have mostly failed to delineate prospective areas. We have evaluated a successful podiform chromite exploration case history that is based on AMT. More than 8000 AMT stations were used in this study within a [ ] area of the ophiolite belt. Line separations were 80 or 40 m, and the station separation was 20 m. We implemented Bostick conversion and nonlinear conjugate gradient inversions for data interpretation, whereas 2D resistivity sections and 3D resistivity imaging were used to elucidate the inner structure and distribution of rock faces within the Luobusa ophiolite. Results from rock physics and drilling further indicate that resistivity-anomaly domains from these AMT results are correlated with rock faces in terms of fresh harzburgite, altered harzburgite and dunite, and they can thus be connected to concealed deposits. Therefore, we have developed three resistivity-anomaly models for chromite exploration, and we delineated several prospective regions containing exploitable deposits within the Luobusa ophiolite. Seven of the nine verified boreholes discussed in this paper intersected with chromite deposits; one comprises the largest and highest grade chromite deposit in China to date. Our AMT results provide the impetus for future chromite exploration in Tibet and enable a refined understanding of the structure and distribution of rock faces within the Luobusa ophiolite.

The pore structure of tight sandstones has significant influence on the elastic properties, and also determines the fluid-flow-related wave dispersion and attenuation mechanisms. However, in those velocity dispersion models, only the compliant pore or crack with fixed aspect ratio and concentration has been taken into account, which is not completely realistic in reservoir rock. A procedure is presented to obtain the pore aspect distribution of compliant pore (crack) from the pressure dependence of velocities. Based on the pore aspect distribution of compliant pore, the existing squirt fluid model of Gurevich is extended to consider complex pore structure of reservoir sandstones, especially when the aspect ratio has a relatively wide distribution. The extended Gurevich's squirt-flow model is consistent with Gassmann's equation at low-frequency limit, and with Mavk-Jizba model at high-frequency. To illustrate the validation of the extended Gurevich's squirt-flow model, we compare predictions of our squirt model with laboratory measurement of two water-saturated tight sandstones at ultrasonic frequency. The new model on the basis of the pore structure is more accurate in predicting the pressure dependence of compression and shear velocities for these two samples. Considering aspect ratio distribution of compliant pore in the tight sandstone samples, the new model for predicting dispersion curve shows a relatively wide squirt-flow relaxation frequency range, which even covers the typical seismic frequency and sonic logging frequency. This observation suggests that squirt flow be still important even in the seismic frequency band for tight sandstone with special pore structure.

Magnetotelluric (MT) inversion and forward modeling are closely linked. The optimization and iteration processes of the inverse algorithm require frequent calls to forward modeling. However, traditional numerical simulations for forward modeling are computationally expensive; here, deep learning (DL) networks can simulate forward modeling and significantly improve forward speed. Applying DL for forward modeling in inversion problems requires a high-precision network capable of responding to fine changes in the model to achieve high accuracy in inversion optimization. Most existing MT studies have used a convolutional neural network, but this method is limited by the receptive field and cannot extract global feature information. In contrast, the Mix Transformer has the ability to globally model and extract features. In this study, we used a Mix Transformer to hierarchically extract feature information, adopted a multiscale approach to restore feature information to the decoder, and eliminated the skip connection between the encoder and decoder. We designed a forward modeling network model (MT-MitNet) oriented toward inversion. A sample dataset required for DL forward was established using the forward data generated from the traditional inverse calculation iteration process. The trained network quickly and accurately calculates the forward response. The experimental results indicate a high agreement between the forward results of MT-MitNet and those obtained with traditional methods. When MT-MitNet replaces the forward computation in traditional inversion, the inversion results obtained with it are also highly in agreement with the traditional inversion results. Importantly, under the premise of ensuring high accuracy, the forward speed of MT-MitNet is hundreds of times faster than that of traditional inversion methods in the same process.