Mining and Oil/Gas Exploration Case Studies
3D Geophysics Inversion Case Studies
3D ZTEM Inversion
The ZAxis Tipper Electromagnetic (ZTEM) method is an airborne natural source EM technique developed by Geotech Ltd. that is effective at mapping largescale geologic structures. Because natural source electromagnetic signals have no geometric decay as they penetrate the earth, the ZTEM method has a greater depth of investigation than many other controlled source methods. In this case study we inverted two ZTEM field datasets using the new OcTree based ZTEM inversion code. The first case study is one of the largest EM inversions ever ran with over 70 million cells in the inverted model. The result maps the geology with great detail over a massive area. The second example is from the Mt. Milligan porphyry deposit in British Columbia. The inversion result images the largescale structures in the area as well as the deposit scale features.


Physical Property Integration: Conductivity, Chargeability and Magnetic Susceptibility
The study was done on a South American AuCu porphyry deposit in which multiple geophysical datasets were inverted in 3D by Computational Geosciences Inc. (CGI). Datasets included dipoledipole and poledipole DC resistivity and induced polarization, controlled source audio magnetotellurics (CSAMT), and ground magnetics collected over a 3 km square area. Each survey was performed separately between 1997 and 2008 and had previously been inverted in 2D. The survey was reinverted by CGI in order to gain a better understanding of the physical property distributions on the property. To accomplish this, threedimensional models of conductivity, chargeability, and magnetic susceptibility were recovered on the same mesh. This allowed for the practical interpretation of correlations of different physical properties with each other, as well as with known geological structures.


Enhanced Reservoir Monitoring using Coupled Electromagnetics and Flow Modeling
Remote time lapse monitoring of reservoirs can provide valuable information to meet production goals. Remote monitoring requires technology that can detect movement and changes in the reservoir during production and flooding events. In this case study, we demonstrate how an injection event can be modeled and monitored using flow simulation software and electromagnetic data. First the injection event is simulated to predict the fluid flow. This information is then used as a constraint when inverting the collected geophysical data. The combination of flow simulation and electromagnetic inversion provides an enhanced monitoring technique for reservoir characterization.


3D Inversion of Airborne TimeDomain Electromagnetic Surveys
Airborne timedomain electromagnetic (EM) surveys are eeffective tools for mineral exploration, geologic mapping and environmental applications. These surveys can be an economical way to explore large prospective regions. Traditionally, the surveys have been interpreted using time constant analysis, conductivity to depth imaging (CDI) or possibly 1D inversions. These methods assist in a simple interpretation of the data, however because they do not fully model the physics in 3D, they can fail to accurately represent environments such as real world structures and geological targets. Airborne EM datasets are characterized by large volumes of data, as each EM sounding implies a new transmitter location. As a result, inverting this airborne electromagnetic surveys in 3D is a computationally difficultcult problem that until recently has not been possible for the exploration community. Computational Geosciences Inc. (CGI) utilizes advanced mathematics and computer science to allow us to solve problems which were previously deemed too difficultcult. This is done using multiple meshes, each spanning the full model domain. By using adaptive OcTree mesh rerefinement each mesh is optimally designed for computational eefficiency on the local domain dedefined by a subset of transmitters. This methodology maximizes the value of your airborne electromagnetic surveys by enabling fast and accurate 3D inversions of large domains.


Design and Inversion of 3D TimeDomain EM Survey
The information and value that can be obtained from geophysical surveys will always be limited by the method and survey design. Optimal survey design can be achieved by leveraging a basic understanding of the subsurface structure to meet specificc exploration objectives. Computational Geosciences Inc. (CGI) helps maximize the value of geophysical exploration by assisting in the optimal survey design, inverting the data to recover high resolution 3D models, and consulting on the next steps in the exploration program.


3D Inversion of Time Domain Electromagnetic Data for Ground Water Aquifers
Airborne timedomain electromagnetic (AEM) surveys have been shown to be an effective tool for ground water aquifer imaging. The data are difficult to invert in 3D because of the number of sources and the size and scales of the computational domain. To solve the 3D AEM inverse problem we partition the forward problem into multiple meshes. Each mesh spans the full computational domain but uses fine mesh cells around the selected transmitters and receivers. This mesh refinement methodology results in a forward modelling mesh that has far fewer cells than the full inversion mesh. Since the forward modelling operation is the bottleneck for AEM inversions, this procedure results in a highly parallel algorithm that can handle arbitrarily large datasets and can deal with many scales of detail in the data. The advanced 3D inversion capabilities of Computational Geosciences Inc. (CGI) are demonstrated on a SkyTEM field dataset from the Horn River Basin in British Columbia.


3D Inversion of Time Domain Electromagnetic Data for Thin Dipping Conductors
Airborne timedomain electromagnetic (AEM) surveys have been shown to be an effective tool for imaging conductive metalrich targets. The data are difficult to invert in 3D because of the number of sources and the size and scales of the computational domain. To solve the 3D AEM inverse problem we partition the forward problem into multiple meshes. Each mesh spans the full computational domain but uses fine mesh cells around the selected transmitters and receivers. This mesh refinement methodology results in a forward modelling mesh that has far fewer cells than the full inversion mesh. Since the forward modelling operation is the bottleneck for AEM inversions, this procedure results in a highly parallel algorithm that can handle arbitrarily large datasets and can deal with many scales of detail in the data. The advanced 3D inversion capabilities of Computational Geosciences Inc. (CGI) are demonstrated with a VTEM35 field dataset from Geotech Ltd. over the Caber deposit.


3D Inversion of Tensor Gravity Data
Airborne gravity gradiometry surveys are becoming commonplace in mineral and hydrocarbon exploration. Each survey can consist of up to six gravity tensor components at millions of measurement locations. Threedimensional inversion is desirable to recover a threedimensional subsurface model that simultaneously fits all components of the collected data. Highresolution models can be used as an exploration tool by examining the recovered physical parameters rather than transforms of the observed tensors. Up until now, the number of data and model parameters associated with these data sets made an inversion difficult to carry, out even with substantial computational resources. In this work, we present a finitevolume, differentialequation method for gravity gradiometry data inversion. A principal benefit of the differential equation formulation arises from not explicitly forming the dense matrix that is required through the integral equation approach. Forward modeling involves solving a linear system of sparse, discrete operators by preconditioned conjugate gradients. This can save a significant amount of time over the entire inversion process. The number of data does not directly affect the size of the inversion as the fields are solved on the model and then interpolated to the observation locations. To demonstrate the effectiveness of our method, we present an inversion of the Bathurst Mining Camp region that consists of 1.4 million data measurements and a mesh of 24 million cells.

