What is it about?

3-dimensional inversion of Magnetotelluric (MT) data is shown. Various 3D data are inverted using the newly developed code. All data used in this study are previously used by other authors to be inverted. This algorithm shows that all kinds of MT data can be inverted successfully even when there are noise and distortion on the data. Furthermore, the implemented techniques made the new algorithm considerably fast.

Featured Image

Why is it important?

The newly developed algorithm is made to be very versatile and feature-rich. The hybrid forward modeling engine in the forward modeling part and the mesh decoupling technique in the inversion part of the developed code made this study unique and considerably faster. Also, the inversion capability of various data types is implemented. In addition to the impedance tensor and magnetic transfer function, their phase tensors can also be inverted. In addition to that, inter-site tensors and their respective phase tensors can also be used as input data in the algorithm. To deal with the distortion on the data, along with the inversion of the phase tensor data, the estimation of the distortion tensor is also considered and implemented.


3D inversion of electromagnetic data is still a challenge for many people due to its compute-intensive nature. To overcome that problem various techniques are considered in this study. When many small techniques, that can be used to accelerate the inversion code, is used, the total inversion times can be reduced greatly. These techniques are chosen to be the deployment of the potential formulation of the Maxwell equations, the hybrid finite-difference and finite-elements numerical technique to solve the potentials formulation, the mesh decoupling technique to reduce the forward modeling mesh and the use of the GPUs in the numerical solutions. In addition to those, it is very important to find out which data type can best resolve which kind of anomaly. For that reason, every data type that exists in 3D Magnetotelluric inversion is used to invert a model. It is also investigated that if using the phase tensor data or the estimating the distortion tensor as a parameter could be better to deal with the distortion on data. In addition to those, the effectiveness of the inter-site tensor data is investigated. Furthermore, their distortion-free phase tensor data are also used in the inversion. In the end, the inter-site tensor data with multiple base stations and joint inversion of the inter-site data and single-site data (i.e, impedance, magnetic transfer function) are shown in this study.

Deniz Varilsuha
Istanbul Teknik Universitesi

Read the Original

This page is a summary of: 3D Inversion of the Magnetotelluric data by using a hybrid forward modeling approach and mesh decoupling, Geophysics, June 2020, Society of Exploration Geophysicists, DOI: 10.1190/geo2019-0202.1.
You can read the full text:




The following have contributed to this page