A glimpse behind the scenes of the process of interpreting inversion results with an illustrative example featuring VOXI Earth Modelling. This case demonstrates the need for an iterative process revolving around discussion and the assessment of assumptions that go into an inversion. [Video]
Geophysical Inversion Modelling
Inversion algorithms were rated as the #1 exploration innovation in the 100 Innovations in the Mining Industry report published by Minalliance. Over the past decade, geophysical inversion has proved its effectiveness in exploring for ore deposits and major oil reserves around the world.
What is Geophysical Inversion Modelling?
Inversion enables resource explorers to extract more insight from geophysical data by converting geophysical measurements into 3D images of the subsurface that can be integrated with other surface and subsurface geologic observations. Insights generated from geophysical inversion have helped to improve prospecting and focus drill targeting, particularly in deeper and more complex subsurface environments.
Geophysical surveys are conducted to seek information about the underlying and hidden geology that gives rise to the values observed on the Earth’s surface. For example, the observed gravity field varies as a function of the underlying rock density, and a magnetic survey varies as a function of rock susceptibility. But the 3D relationship is very complex. What is seen at the surface depends on the depth and shape of various features, and there are many distributions of density or susceptibility or other rock properties that can explain what is observed. The process of 3D inversion aims to produce the most likely distribution of physical rock properties that explain what is observed.
Three key trends have been driving exploration companies to use inversion more routinely:
- The increasingly important role of geophysics as an exploration method for exploring deeper, at depths of hundreds of metres below the surface.
- The requirement to reduce the risk of high stakes resource exploration through efficient assessment of large tracts of ground, predictive multi-parameter modelling and target delineation.
- Advances in technology that have reduced the time and effort required to transform large geophysical data into useful visualizations of the subsurface through 3D inversion.
Over the past decade, inversion has proved its effectiveness in exploring for ore deposits and major oil reserves around the world. Within mineral exploration, inversion modelling has aided in developing exploration potential within the iron ore and nickel belts of Western Australia; the Stuart Shelf and Olympic Dam in South Australia; iron oxide-copper-gold in Africa, South America and Australia; copper in Mongolia; and nickel laterite in Colombia. Within the oil industry, inversion has reduced uncertainty when exploring the dense sedimentary section that surrounds the salt bodies in the Gulf of Mexico, and offshore West Africa and South America where the geology is more complex and less predictable.
What are the basic principles of inversion?
The process of 3D inversion seeks to produce a 3D distribution of physical rock properties (e.g. density) that explains an observation measured in the field (e.g. a gravity response). This process is inherently non-unique, and so the experience of the interpreter and the a priori information used to validate an inverse model is important.
The forward problem: Forward modelling is the process of calculating a response (e.g. a gravity measurement) from a given earth model (e.g. a density distribution). The computation itself may be challenging, but the concept is simple – because when you calculate a forward response, there’s only one answer.
Forward modelling is an important step in determining the value of a particular geophysical survey. Once an exploration team has a good idea of the type of target they seek, they construct a hypothetical, 3D earth model and calculate what the response of various types of surveys would be. Some surveys will make the target easy to see, others won’t. The value of forward modelling is that it helps make wise exploration decisions when planning geophysical surveys.
The inverse problem and non-uniqueness: The opposite of the forward problem is the inverse problem. Instead of finding the single possible response to a given earth model, inverse modelling will help determine what 3D distribution of physical properties yields a measured field response. The catch is that there are many models that can create the same surface response. This is called non-uniqueness. For example, a broad dense body close to the surface creates a gravity response that is similar to a very dense compact body deep in the Earth.
The question becomes: how can you create inversion results that are useful representations of the subsurface? The answer lies in other pieces of the puzzle that are known. Things like overburden thickness, lithology from drill data, and borehole assay results. This known information can help constrain the inverse problem to a limited number of plausible models. The subset of models will have certain commonalities that can then help the explorer make an interpretation that agrees with all the pieces of the puzzle (lithology, geochemistry, and structural information).
The most useful models are the result of exploring the inversion model space by running many scenarios with different constraints and sensitivity to other geological information. So, new algorithms and faster computers have a huge impact on the success of geophysical inversion for exploration. Similarly, the ability to easily integrate and use supplementary information to better constrain the inversion is critical to producing reliable models.
Advances in geophysical 3D inversion technology
The University of British Columbia (UBC) led the way when they established the UBC-Geophysical Inversion Facility in 1989 with funds from the B.C. Science and Technology Fund. With industry support, UBC went on to develop modelling and inversion programs and utility codes that form the basis for most of the geophysical inversion solutions deployed in research institutes and large mineral exploration companies.
More recent advances have resolved technical barriers to using inversion in early stage exploration, removing the complexities in setting up inversion parameters for modelling reliability, and increasing processing speed for rapid and continual iteration of results to support time-critical exploration decision making.
Cloud-powered inversion: In 2012 Geosoft released VOXI Earth Modelling, a cloud-based geophysical inversion software service that generates 3D voxel models from airborne or ground gravity and magnetic data. By harnessing the processing power of the cloud, VOXI has made large, multi-parameter geophysical inversion modelling faster, more responsive and effective as a tool to assist with predictive modelling for project generation and target delineation, as well as more advanced exploration.
Behind VOXI’s speed and agility is cloud technology engineered by Geosoft to conduct the complex geo-computing via the internet, with minimal drain on the explorer’s personal computer systems. The VOXI Earth Modelling cloud service is powered by Microsoft Windows Azure.
Since the VOXI service went live, thousands of models have been generated in the cloud. Explorers are routinely creating earth models in minutes, with models as large as 12.5 million voxi cells are being processed in under an hour using the VOXI service. This type of modelling could consume a full day of people and computing resources using traditional technology.
Advanced inversion techniques: Geosoft has also introduced advanced geophysical modelling techniques such as Magnetization Vector Inversion (MVI) and Iterative Reweighting Inversion Focusing (IRIF) and the Cartesian Cut-Cell method (CCC) to improve modelling accuracy. MVI allows the direction of the magnetization to vary within the model and thus take into account the combined effects of remanence, demagnetization, anisotropy and induced magnetization. The result can be a more realistic representation of rock magnetization, particularly at low magnetic latitudes, or when remanent magnetization is suspected. IRIF takes a smooth earth model and uses it as a reweighting constraint when running a new inversion of the same data. It can improve modelling results in regions where there are few or no subsurface constraints, by increasing the geological resolution of unconstrained inversions. The CCC method allows the topographic surface to be modelled accurately by removing the effect of cell edges.
Inversion of gravity gradiometry: VOXI supports inversion and forward modelling of gravity gradiometry data to yield a detailed 3D model of the rock density. The resulting detailed density model can be used to interpret and to target regions for potential oil, gas and mineral deposits. In addition to generic vertical gravity gradient data two common data types, Falcon AGG data and Bell Geoscience/ ARKeX LM FTG data, can be imported directly into the model.
Drillhole data and wireframe models can also be used to constrain VOXI inversion models. Geosoft’s wireframing tools will create reference models from drillhole data, as well as digitize geological sections, to build wireframe models, and create voxel constraints in VOXI.
Geosoft VOXI Earth Modelling is offered as a software service accessible within the Geosoft Oasis montaj collaborative exploration platform.
What is the value of geophysical inversion for resource exploration?
Inversion is applicable at every stage of the exploration program.
- At the project generation stage, a 3D inversion of existing gravity or magnetic data can provide visual clues to what is happening geologically in the subsurface even if there is no outcrop to be found.
- At the prospect targeting phase, explorers can use inversion results to improve their geological models for more effective drill program planning and follow up.
- As exploration progresses to the advanced stage, the inversion process will provide an increasingly accurate picture of structural geology and their extensions in the subsurface.
How can I maximize my success with geophysical 3D inversion?
Some of the recent advances in inversion, such as VOXI's Iterative Reweighting Inversion Focusing (IRIF), and Magnetization Vector Inversion (MVI), are helping to define new targets for global mineral projects.
At a gold project in the Yukon, for example, IRIF is helping geologists reconcile the geophysics with the geology and identify new drill targets in an area of no outcrop.
At a gold-silver mine in Mexico, inversion modelling is outlining the geometry and depth to skarn mineralization targets while MVI is solving the problem of magnetic remanence that can distort traditional magnetic inversions.
In a broader sense, mining companies are using 3D inversion to process large volumes of geophysical data that would have overwhelmed traditional modelling technology in order to define the controls on known mining camps and predict where the next mining camps might lie.
By incorporating inversion into their exploration programs, explorers can reduce the time and effort required to visualize and understand deeper subsurface environments and they can save on drilling costs by generating more accurate targets.
What are some of the most common, impactful things you can do to improve your 3D geophysical inversion models? Geosoft's Taronish Pithawala, technical lead for geophysical modelling, presents the 'Top 5 Inversion Best Practices' in this web series. [Video]
New inversion methods are now helping to enhance earth models and interpretation of Norway's Karasjok Greenstone Belt. The project serves as a case example for the use of geological and geophysical techniques both for mineral exploration and geology in 3D. [Article]
Deposits may have become harder to find in recent decades, but by eliminating instrument noise, making 3D geophysical inversion an integral part of the exploration process and harnessing the unprecedented computing power of the cloud, Rio Tinto aims to see what is currently hidden. [Article]
Magnetization Vector Inversion (MVI) is a technique introduced by Geosoft to help eliminate erroneous assumptions about magnetization. After a year of application in mineral exploration, project examples confirm that MVI results provide a more reliable representation of subsurface geology. [Article]
From validating geology on grassroots project to finding new ore around existing mines, geophysical inversion is taking some of the risk out of high stakes mineral exploration. Recent technology developments have turned inversion into a fast and responsive tool that can be used at every stage to delineate targets with greater speed and accuracy. [Article]
Explorers can visualize the subsurface in a more instructive way by using 3D inversion of geophysical data. This case study demonstrates the use of integrated geological and magnetic data inversion modelling for defining drilling targets in deep and complex exploration plays. [Case Study]
Eventually, as the number of constraints in inversion models increases, geoscientists will be able to map regions of alteration in proximity to orebodies. The latest innovation has focused on generating inversion models with greater ease and confidence, making their output more reliable and informative as an aid for mineral exploration. [Article]
Magnetic Vector Inversion, a simple approach to the challenge of varying direction of rock magnetization
This paper includes case-study work that demonstrates the application of Magnetic Vector Inversion to particularly challenging and well known magnetic anomalies in both Brazil and Australia. [Conference Paper]
This paper compares the results of Magnetization Vector Inversion (MVI) with Susceptibility Inversion using airborne magnetic data from the Crixas area, Goias State, Brasil. MVI was found to provide better 3D understanding and interpretation in low latitude areas as well as in the presence of remanent magnetization. [Conference Paper]
An introduction to Magnetization Vector Inversion (MVI), a technique which incorporates both remanent and induced magnetization in 3D inversion for correct interpretation of magnetic field data. [Conference Paper]
Interpretation and modelling of the Pedirka Basin (central Australia) using magnetics, gravity, well-log and seismic data
The Pedirka Basin of central Australia remains a frontier for petroleum exploration. In this case study, publicly available high quality magnetics and regional gravity datasets was combined with available seismic sections and well logs to assist in gravity 2D section modelling and magnetic 3D geophysical inversions. [Conference Paper]
Geophysical inversion modelling is being used to characterize deeper prospects and complex fields, providing new insight and reducing risks within petroleum and mineral resource exploration. This Gravity Case Study of the Podolsky Deposit, Sudbury Basin, demonstrates the use of gravity interpretation and modelling to resolve structural ambiguity and generate new opportunities in Canada’s Sudbury Basin. [Conference Paper]