Inversion Of MT Data Research Articles

Lithospheric electrical structure and geodynamic model of the red river fault zone and its adjacent areas in southwest China: Constraints from 3-D magnetotelluric imaging

The Red River Fault (RRF) zone is situated on the southeastern boundary of the Tibetan Plateau. To further explore the geodynamic model of the region and the collision mechanism of the Eurasian continent, the electrical structure model of the lithosphere in the RRF was obtained through 3-D inversion of MT data. The preferred model indicated that fault zones were primarily distinguished by low resistivity anomalies, with some exhibiting high resistivity characteristics. It was hypothesized that the low resistivity anomalies may be attributed to the presence of saline fluids or metal sulfide within the fault zones, on the other hand, the high resistivity characteristics suggested inactivation due to fault cooling. Rapid changes in the electrical structure occurred on both sides of the fault zones, causing the upper and middle crust to be divided into distinct secondary tectonic units. Based on the effective viscosity calculation, we analyzed the crust-mantle rheology of the RRF area. We assumed that the RRF area was composed of three geodynamic models: the “crustal flow” model was distributed from the northern part of the Lancang River Fault (LCRF) zones to the southern part of the RRF zones. The “ductile continuum rheology” model was distributed from the north of the RRF to the western part of the Xiaojiang Fault (XJF) zones. The “mantle upwelling” model was distributed in the southeast part of the RRF.

Investigation of the main reservoir of the Tulehu geothermal system (Indonesia) using 3-D inversion of MT data

Tulehu geothermal prospect area is located in Maluku Province, Indonesia. The existence of the geothermal system in this area is recognized by the appearance of hot springs on the western side of Mt Eriwakang and the northern side of Mt. Huwe. Most of these hot springs occur along inferred geological structures. The geological conditions of this area are dominated by Quaternary volcanic rocks, limestone, alluvium, and metamorphic rocks. Several authors have conducted several studies to propose various geothermal conceptual models, from structurally controlled to volcanic-hosted geothermal systems. However, the geoscientific data supporting the hypotheses are not comprehensive yet. There are still gaps in the data coverage as well as in the interpretation of geological, geochemical, and geophysical data, so a comprehensive study should be done before deciding on further stages. Therefore, this paper proposes a new conceptual model of the Tulehu geothermal area by reducing the existing data gap. Interpretation of remote sensing data, geological field observations, age dating analysis, reanalysis of geochemical data, and additional MT data covering Mt. Eriwakang are conducted to reconstruct the conceptual model. Graben-like structures are identified based on the geological data where the product of Mt. Eriwakang is in the middle. Based on the age dating data, the product of Mt. Eriwakang erupted in the Quaternary time (413±6 kA), so it is still quite potential as a heat source. This indication is also supported by the trend of Cl, HCO3, and SiO2, which describe the possible flow from Mt. Eriwakang to the surrounding hot springs. Based on the Na-K-Mg and Silica geothermometers, the reservoir temperature is estimated at 210-240 °C. A three-dimensional (3-D) MT inversion was applied to all the existing and new MT data to obtain a 3-D subsurface resistivity model. The oceanic bathymetry data was also incorporated to mitigate the possible coastal effect from the surrounding seawater to the MT data. The new 3-D MT inversion results show a good agreement with the geological and geochemical indications. The updome-shaped pattern of the subsurface resistivity distribution beneath Mt. Eriwakang supports the geological and geochemical data analysis results concerning the indications of the prospects area around Mt. Eriwakang. These interpretations can strengthen the recommendation to focus on the area around Mt. Eriwakang for further development.

Two-dimensional deep learning inversion of magnetotelluric sounding data

Abstract Conventional linear iterative methods for magnetotelluric sounding (MT) suffer from considerable limitations related to difficulties in selecting the initial model and local optima. On the other hand, conventional intelligent nonlinear methods exhibit slow convergence and low accuracy. In this study, we propose an inversion strategy based on the deep learning (DL) deep belief network (DBN) to realise the instantaneous inversion of MT data. A scaled momentum learning rate is introduced to improve the convergence performance of the restricted Boltzmann machine during the DBN pre-training stage, and a novel activation function (DSoft) is introduced to enhance the global optimisation capability during the DBN fine-tuning stage. To address the difficulty in designing the sample data when prior information is lacking, we employ the k-means++ algorithm to cluster the MT field data and use the clustering results as the prior information to guide the construction of the sample dataset. Then, based on the proposed DBN, we ensure end-to-end mapping directly from the apparent resistivity to the resistivity model. We implement two groups of experiments to demonstrate the validity of both improvements on the DBN. We consider six types of geoelectric model from the test set to demonstrate the feasibility and effectiveness of the proposed DBN method for MT 2D inversion, which we further compare with the well-known least-square regularisation method for several extended geoelectric models and field data. The qualitative and quantitative analyses show that the DL inversion method is promising as it can accurately delineate the subsurface structures and perform rapid inversion.

Efficient Calculation of the Jacobian Matrix for 3D inversion of MT Data

ABSTRACT. The Jacobian or sensitivity matrix is a fundamental component in the implementation of many numerical algorithms for the solution of nonlinear inverse problems. It is also used as a tool for sensitivity analysis and for the assessment of the solutions obtained. However, the computation of the Jacobian by a direct method using finite perturbations of model parameters can be very time consuming and for the 3D case with realistic problem dimensions, it is simply unfeasible. In this work, we present all the steps of a mathematical algorithm based on the adjoint-state method for the assembly of the Jacobian matrix for the 3D magnetotelluric parameter to data mapping obtaining significant reduction in computer execution times. Central to the proposed method is the use of a factorized form of the finite elements matrix for the forward problem. Besides, we solved Maxwell’s equations using a finite elements method using tetrahedral shaped elements formulated in terms of the vector magnetic and scalar electric potentials. We assessed the efficiency of the proposed methodology using a numerical model of a simple sedimentary basin. Our approach outperformed a direct, brute force method in both total execution time and memory usage while maintaining the same level of numerical accuracy. The efficiency of the proposed method makes possible the use of the Jacobian matrix for the inversion of large real 3D MT data sets. Although developed here for the particular case of magnetotellurics the methodology can be extended with reasonable efforts for other similar electromagnetic methods like for example controlled source electromagnetics.Keywords: mathematical models, factored matrix of finite elements, electromagnetism, Efficient calculation of the Jacobian for magnetotelluric data Cálculo Eficiente da Matriz Jacobiana para Dados MT 3DRESUMO. A matriz jacobiana ou matriz sensibilidade é uma componente fundamental na implementação de inúmeros algorítmos numéricos para a solução de problemas inversos não-lineares. Ela também é usada como uma ferramenta na análise de sensibilidade e avaliação dos resultados obtidos. No entanto, o cálculo da matriz jacobiana por um método de diferenças finitas, usando perturbações nos parâmetros de inversão do modelo pode consumir muito tempo e, para o caso 3D com dimensões realistas de problemas geofísicos, é simplesmente inviável. Neste trabalho, apresentamos todas as etapas de um algoritmo matemático baseado no método de estados adjuntos para a montagem da matriz jacobiana para o método magnetotelúrico 3D, obtendo uma redução significativa nos tempos de execução do computador. Central para o método proposto é o uso de uma forma fatorada da matriz de elementos finitos para o problema direto. Para isto, resolvemos as equações de Maxwell usando o método de elementos finitos em malhas 3d tetraedrais, formulados em termos dos potenciais vetorial magnético e escalar elétrico. Nós avaliamos a eficiência da metodologia proposta usando um modelo numérico de uma bacia sedimentar simples. Nossa abordagem se mostrou mais eficiente do que um método direto, de força bruta, tanto em relação ao tempo total de execução quanto ao uso de memória, mantendo o mesmo nível de precisão numérica. A eficiência do método proposto possibilita o uso da matriz jacobiana para a inversão de grandes conjuntos de dados 3D reais de MT. Embora desenvolvida aqui para o caso particular do método magnetotelúrico, a metodologia pode ser estendida para outros métodos eletromagnéticos semelhantes, como por exemplo o método eletromagnético de fonte controlada.Palavras-chave: modelos matemáticos, matriz de elementos finitos fartorada, eletromagnetismo, Cálculo eficiente da jacobiana para dados Magnetotelúricos

Input and output files of ModEM 3-D inversion of MT data in Nanling

Input and output files of ModEM 3-D inversion of MT data in Nanling

Quality over quantity: on workflow and model space exploration of 3D inversion of MT data

3D inversions of magnetotelluric data are now almost standard, with computational power now allowing an inversion to be performed in a matter of days (or hours) rather than weeks. However, when compared to 2D inversions, these are still very computationally demanding. As a result, 3D inversions are generally not subjected to as rigorous testing as a 1D or 2D inversion would be, which has implications when these models are used for geological interpretation. In this study, we explore the parameter space for inversion of continent-scale datasets. The generalisations made regarding the effects of each parameter should also be scalable to smaller surveys and will enable MT practitioners to optimise their results. We have performed testing on a subset of the South Australian component of the eventual Australia-wide AusLAMP (Australian Lithospheric Architecture Magnetotelluric Project). The subset was inverted with different parameters, model setup and data subsets. Specifically, results from testing of the model covariance, the resistivity of the prior model, the inclusion of 'known' information into the prior model, the model cell size, the data components inverted for and the damping parameter lambda were all investigated. In our testing of the 3D inversion software, ModEM3DMT, we found that the resistivity of the starting/prior model had significant effect on the final model. Careful selection of initial lambda value can aid in reducing computational time whilst having a negligible effect on the resultant model, whilst large covariance values and model cell sizes enhanced conductive features at depth.

Experience of using three-dimensional inversion for the large volume of magnetotelluric data

When switching to 3D inversion of MT data, the requirement for computer technology is significantly increased. In this paper we will discuss a few examples of 3D inversion of electromagnetic geophysical field data with the usage of “Lomonosov” supercomputer and show its effectiveness on several geological objects. Each object is associated with a variety of problems: from search for shallow ore to regional hydrocarbon exploration. But all these objects contain a large volume of measurements obtaining qualitative results for which requires a huge amount of time. So that the use of 3D inversion with a high-performance computational complex makes it possible to obtain a qualitative result of solving a wide range of problems.

Three-dimensional Inversion of Electromagnetic Geophysical Data with Parallel Computational Code on Supercomputer Complex “Lomonosov”

Usage of 2D inversion of magnetotelluric data for real geological objects can cause distortion, but it is more often used in commercial projects, because of its effectiveness and great experience. Whereas in the case of 3D inversion is not such a great experience and there are a number of global problems. When switching to 3D inversion of MT data, the requirement for computer technology is significantly increased. In this paper we will discuss a few examples of 3D inversion of electromagnetic geophysical field data with the usage of Lomonosov supercomputer and show its effectiveness on several geological objects. Each object is associated with a variety of problems: from search for shallow ore to regional hydrocarbon exploration. But all these objects contain a large volume of measurements obtaining qualitative results for which requires a huge amount of time. So that the use of 3D inversion with a high-performance computational complex makes it possible to obtain a qualitative result of solving a wide range of problems.

Magnetotelluric studies in the epicenter zone of 2001, Bhuj earthquake

Some earthquakes to Mw5 level are occurring along two faults viz. Kachchh Mainland Fault and South Wagad Fault near the epicenter of Mw7.7 Bhuj earthquake of 2001. These faults are not the causative fault of the 2001 earthquake and suspect for future large earthquake. In absence of detail geophysical surveys, the sub-surface orientation of these faults was not known and MT profiles have proved useful in delineating the subsurface trends. Two MT profiles having lengths of 24km and 15km (with an inter station spacing of 1–3km) are taken about 10km east and west of the epicenter of Mw7.7, 2001 Bhuj earthquake, respectively. The profiles are in nearly N–S direction to detect the faults which trend E–W in the area. From 2-D inversion of MT data of 23 stations, it is found that there are two distinctive conductive features matching with the South Wagad Fault and Kachchh Mainland Fault south of it both being about 10km apart. The South Wagad Fault is seen extending to Banni low lying area 30km west of the present trace.

Magnetotelluric Exploration at Tendaho high temperature geothermal field in North East Ethiopia

SUMMARY Tendaho is one of the high temperature geothermal areas in Afar depression in north east Ethiopia. A total of 129 MT sites were acquired from Tendaho high temperature field. The 2D inversion of MT data from Tendaho high temperature field revealed three main resistivity structures down to a depth of 10 km: low resistivity surface layer underlain by a resistive layer followed by good conducting structure. The low resistivity surface layer show areas with either sediments, lateral flow of geothermal fluids or zeolite-clay alteration zone. Below the conductive layer, is a high resistivity zone that can be correlated to Afar stratoid basalts or epidote alteration zone .The high resistivity structure has been associated with the deep reservoir of the geothermal system. The deep good conductive body is probable heat source of the geothermal system.

A novel anisotropic inversion approach for magnetotelluric data from subsurfaces with orthogonal geoelectric strike directions

The key result of this study is the development of a novel inversion approach for cases of orthogonal, or close to orthogonal, geoelectric strike directions at different depth ranges, for example, crustal and mantle depths. Oblique geoelectric strike directions are a well-known issue in commonly employed isotropic 2-D inversion of MT data. Whereas recovery of upper (crustal) structures can, in most cases, be achieved in a straightforward manner, deriving lower (mantle) structures is more challenging with isotropic 2-D inversion in the case of an overlying region (crust) with different geoelectric strike direction. Thus, investigators may resort to computationally expensive and more limited 3-D inversion in order to derive the electric resistivity distribution at mantle depths. In the novel approaches presented in this paper, electric anisotropy is used to image 2-D structures in one depth range, whereas the other region is modelled with an isotropic 1-D or 2-D approach, as a result significantly reducing computational costs of the inversion in comparison with 3-D inversion. The 1- and 2-D versions of the novel approach were tested using a synthetic 3-D subsurface model with orthogonal strike directions at crust and mantle depths and their performance was compared to results of isotropic 2-D inversion. Structures at crustal depths were reasonably well recovered by all inversion approaches, whereas recovery of mantle structures varied significantly between the different approaches. Isotropic 2-D inversion models, despite decomposition of the electric impedance tensor and using a wide range of inversion parameters, exhibited severe artefacts thereby confirming the requirement of either an enhanced or a higher dimensionality inversion approach. With the anisotropic 1-D inversion approach, mantle structures of the synthetic model were recovered reasonably well with anisotropy values parallel to the mantle strike direction (in this study anisotropy was assigned to the mantle region), indicating applicability of the novel approach for basic subsurface cases. For the more complex subsurface cases, however, the anisotropic 1-D inversion approach is likely to yield implausible models of the electric resistivity distribution due to inapplicability of the 1-D approximation. Owing to the higher number of degrees of freedom, the anisotropic 2-D inversion approach can cope with more complex subsurface cases and is the recommended tool for real data sets recorded in regions with orthogonal geoelectric strike directions.

Saturation with fluids of earthquake focal zones of Eurasian folded belt (according to 2D-3D inversion of MT data)

The information received at realization 2D-3D of inversion magnetotelluric the data, has allowed to estimate together with seismic constructions a mineralization and fluid-filled focal zones of earthquakes within the limits of Northern Caucasus and Тien-Shans, Pamirs, Altai-Sajns, Кorak- Kamchatkas regions. It is shown, that fluid-filled depends on character of seismicity and changes from the tenth shares up to the first interests.

Joint 1D inversion of TEM and MT data and 3D inversion of MT data in the Hengill area, SW Iceland

An extensive study of the resistivity structure of the Hengill area in SW Iceland was carried out by the combined use of TEM and MT soundings. Joint inversion of the collected data can correct for static shifts in the MT data, which can be severe due to large near-surface resistivity contrasts. Joint 1D inversion of 148 TEM/MT sounding pairs and a 3D inversion of a 60 sounding subset of the MT data were performed. The 3D inversion was based on full MT impedance tensors previously corrected for static shift. Both inversion approaches gave qualitatively similar results, and revealed a shallow resistivity layer reflecting conductive alteration minerals at temperatures of 100–240 °C. They also delineated a deep conductor at 3–10 km depth. The reason for this deep-seated high conductivity is not fully understood. The distribution of the deep conductors correlates with a positive residual Bouguer gravity anomaly, and with transform tectonics inferred from seismicity. One model of the Hengill that is consistent with the well temperature data and the deep conductor that does not attenuate S-waves, is a group of hot, solidified, but still ductile magmatic intrusions that are closely associated with the heat source for the geothermal system.

Determination of the regularization level of truncated singular‐value decomposition inversion: The case of 1D inversion of MT data

ABSTRACTThe total mean‐square error (MSE) of the estimated model, defined as the sum of the standard model variance and the bias variance, is used to define the truncation level of the singular‐value decomposition to give a reasonable balance between model resolution and model variance. This balance is determined largely by the data and no further assumptions are necessary except that the bias terms are estimated sufficiently well. This principle has been tested on the 1D magnetotelluric inverse problem with special emphasis on high‐frequency radio magnetotelluric (RMT) data. Simulations clearly demonstrate that the method provides a good balance between resolution and variance. Starting from a homogeneous half‐space, the best solution is sought for a fixed set of singular values. The model variance is estimated from the sum of the inverse eigenvalues squared, up to a certain threshold, and the bias variance is estimated from the model projections on the remaining eigenvectors. By varying the threshold, the minimum of the MSE is found for an increasing number of fixed singular values until the number of active singular values becomes greater than or equal to the estimated number. As a side‐effect, the depth of penetration of a given set of measurements can be estimated very efficiently by simply noting at which depth the final model deviates little from the starting homogeneous half‐space model. A suite of synthetic data is inverted and an example of inversion of one site is shown to illustrate how the truncation is carried out as the non‐linear inversion process proceeds. A field example with a profile across a plume of contaminated groundwater in the Netherlands shows good agreement with the electrical resistivity obtained in a nearby borehole.

2-D inversion of MT data with a variable model geometry

SUMMARY An inversion algorithm which takes into account the varying geometry in magnetotelluric models is considered. The particular problem of estimating the shape of the interface between two different conductivities is studied. This boundary is represented by an implicit function and its parameters are calculated in the inversion process. The assumption of using such a function allows us to simplify the computation of the sensitivity matrix of the model, and gives a stronger numerical stability to the inversion procedure. The finite element method is used to solve the forward problem. The inversion algorithm is based on a Bayesian approach that allows the introduction of a priori information. Several examples with simple geometries show the possibilities offered by the algorithm.