News

Both software have been modified to address changes in the Windows 11 OS.

Details of QCTool upgrades may be found on the QCTool website qctool.ca.

More than 2 years of development have gone into the upgrades in EMIGMA V10. These notes will cover some of the changes.

Magnetics
Our 3D gridded magnetic inversion has been augmented in several aspects. The application already was a multi-core, array processing application but has been upgraded to allow a larger set of data as well as a much larger number of cells in the inversion grids. The inversion now allows the import of the topography over and in the vicinity of the survey. This is critical not only for accurate depth determination but also to address features within the topography. We have also upgraded the ability to introduce known magnetic information either for surface sampling or borehole cores. We have also upgraded to allow the use of CAD models as starting models for the inversions. Joint inversions are now allowed. These can be multiple data components such as the TMI and one or more of its derivatives. Also, joint inversion of airborne and ground data is now provided.

In some instances, magnetic anomalies may conform to Poisson's relation and in which case, it may be interesting to joint invert your magnetic and gravity data. To address this situation, we have added another option to import as a starting model, your gravity model or inversion and vice versa in a gravity interpretation project. In this case, when using a gravity anomaly or grid as a starting model, the software will use Poisson's relation to accurately determine the relation between the density values and produce susceptibility values for the starting model. The reverse is done in the case of gravity interpretation.

EMIGMA has long offered the simulation of models for magnetic vector data. We have now released our 3D gridded inversion to provide 3D inversion of vector data. Vector data must normally be processed prior to inversion and these processing capabilities are now released in our latest version of QCTool(V5.0). First, the vector components must be organized so that they are in a consistent orientation. Now that most instruments have IMU capability, we can now rotate the vector data to a consistent grid orientation. Two choices for grid definition may be made by the user, either geographic north or a user defined grid azimuth. Additionally, QCTool now allows the calculation of the derivatives both of the TMI and of vector data if multiple sensors are present. These gradients must also be de-rotated prior to inversion.

Additionally, in our magnetic tools we have upgraded the magnetic compensation to allow several new capabilities. First, compensation may be done on either TMI or vector data and the compensation coefficients may be calculated using either standard vector data or IMU orientation data. Finally, we have made what we believe is a significant break-through in the sense that high altitude box data is no longer required. Relatively low altitude data such as at 200m can now be used for the compensation coefficients. These new features make the compensation processing of magnetic data for either drones or UAV very accurate and efficient.

TDEM
Major upgrades have been made to handle multiple data sets primarily to deal with the issue of strong non-uniqueness of inloop inversion models and the dangers thus presented. To address the non-uniqueness and inaccuracies in depth determination from inloop data, we have provided multiple component inversion for some time. This includes horizontal as well as vertical data and the use of data at any position from the center of the loop such as out of loop data and data in the loop but not at the center. The new capabilities allow such configurations as ground and airborne data as well as data sets with different loop sizes and base frequencies. We have demonstrated that with the use of two loop sizes, two base frequencies and inloop and out-of-loop data that one can determine a unique 1D model with accurate depth and resistivity information if in the appropriate environment.

Visualizer
EMIGMA has had 3D visualization tools for many years in order to build forward models, examine CAD models and provide some forms of data examination. In recent years, we have provided in our visualization application, the means to examine, in detail, the user's inversion models. These tools allow the user to slice the inversion at various depths and across sections as well as examine inversion results in terms of inversion cell parameters. However, users have long found the application unstable when working with complex models and large inversion grids. Major advances have been accomplished to address these problems and provide significantly faster and more stable viewing.

Resistivity
EMIGMA's 3D resistivity forward modeling algorithms are arguably the most comprehensive, rapid and accurate available and these capabilities give us strong advantages when developing inversion algorithms. Resistivity forward and inverse algorithms are now available for both moving systems (dipole-dipole, pole-dipole, pole-pole, Xhole) and fixed source surveys. For fixed source, EMIGMA offers algorithms for ground as well as surface TX to borehole RX and borehole TX and surface RX.

Both the forward and inverse algorithms are integral equation approaches. This provides significant advantages for both accuracy and speed. With an IE technique, there is no need to extend the grids to "infinity" to meet boundary conditions. The boundary conditions are automatically met by the inclusion of the background layered model that hosts the forward and inverse grids. But this approach offers other helpful techniques when performing inversion. For example, once an anomalous area is identified, the inversion grid may be focused onto that area to delineate the structure.

EMIGMA's inversion algorithms in V10 have had a number of upgrades. One of the most important upgrades is the addition of a non-linear forward solver (LN). Electric, such as observed in resistivity surveys, is one of the strongest scattering phenomena observed in nature. Resistivity contrasts across boundaries can be orders of magnitude and this is reflected in electric field current scattering contrasts across the boundary but current bending and gathering is also an important and strong feature in resistivity data. It has long been shown that standard solvers for electric field scattering are not able to both represent the amplitude of the electric field contrasts nor force the current bending caused by current being divergent free and subsequently the interactions between cells. One initial objective of the EMIGMA project was to address these issues and we have solved many of these issues in our forward algorithms but we have now introduced some aspects of these solutions to the inversions. Additionally, the resistivity inversion has been upgraded accordingly: upgraded to handle larger problems in terms of the size of the inversion grids and the amount of data, upgraded to multi-processing and our new techniques to handle memory issues more efficiently, upgraded in the case of fixed transmitter surveys to handle multiple transmitters as well as multiple orientations of voltage measurements, upgraded to allow the inclusion of geological constraints and to allow both polyhedral and prism starting models.

CSEM
3D CSEM/CSAMT inversion: upgrades include allowing more multiple transmitters, more data components (electric and magnetic data) and a larger frequency suite. Also, included is a non-linear forward component allowing resolution of more high contrast structures. Improvements have also been made to improve speed and stability as well as new starting model capabilities. In QCTool, we have upgraded to allow processing of raw data from full waveform time series to generate the impulse responses for the field data.

Gravity
Important upgrades to gravity processing have been made in QCTool for the isostatic effect. This is a critical correction if topography corrections are required. Topographic corrections must be adjusted for the isostatic effect as over high elevation areas, topographic corrections remove too much of the field and over low areas or on the sea, topographic corrections add too much into the field. The isostatic correction offsets these excessive corrections. This is one reason why performing Bouguer corrections may not be the correct approach for you.

One way to deal with the uncertainty concerning whether to perform Bouguer(terrain) corrections is to introduce the terrain into the inversion problem. In this way, variations in density in the terrain may be considered and shallow anomalies in the terrain may be resolved. In EMIGMA's gravity inversion, the user may import a digital elevation model into the inversion process. In this way, targets in the topography may be studied but also depth to targets will be more accurate than if everything is leveled to the geoid.

Additionally, the gravity inversion has been upgraded to allow polyhedral starting models as well as prisms. This can be used to introduce contact surfaces if available from either drilling or seismic work. Geological information may be introduced into the inversion.

Fourier Processing
Deriving a frequency spectrum for full wave data via FFT techniques is limiting due to the requirement for the time series to be of length 2^N. This is particularly true when trying to recover the phase of a spectral (frequency) element. We have discovered with some manufacturers that the phase estimations can be terribly inaccurate for some frequencies. We have developed processing tools in QCTool to allow the user to process full waveform data whether EM, CSEM or IP via DFT (Discrete Fourier Transform) techniques. Historically, although this approach is known to be more accurate, geophysicists have relied on FFT due to computation times. To address the computation time, we have made the DFT spectral analyses and filtering algorithms multi-core, multi-processing which now makes this approach practical on modern computers running Windows.

FFT techniques are also used for processing potential field data for such targets as derivatives, wavelength filtering, upward continuation, and reduction to the pole. We have implemented our DFT algorithms for such processing. This improves the feature resolution as the cell size for calculating the Fourier processes is now not limited. In particular, this removes crossline artifacts and their effects on the vertical derivatives and thus all types of Fourier processing.