Post on 15-May-2018
transcript
Advanced Geophysical Interpretation Centre
3D modeling of the Quest Projects
Geophysical Datasets
Undercover Exploration workshopKEG-25 April 2012
Nigel Phillips
Rock Physical Properties: density–sus. cross-plot
after Williams 2007
What are the dominant causes of physical properties?
Mineralogy Texture
Grain size Porosity
Serpentinisation
Mineralisation
Igneousdifferentiation
Metamorphism
Weathering
Rock Physical Properties – Processes: density–sus. cross-plot
after Williams 2007
How do geologic processes change
physical properties?
Inversion Essentials: What is Inversion?
Inversionprocessing
Model Inversion estimates Earth models based upon
data and prior knowledge.
??
Data
Measurements over
the Earth are data.
source: UBC-GIF
Energy from source
Earth’sphysical properties
Measurements = Data
Pre-processing
Prior information
Inversion within it’s proper context …
Inversion.
Physical property
distributions = MODELS
Inversion Theory
Main objective:
Choose a model that emulates geology and fits the data,
but doesn’t fit the noise in the data.
2 Challenges:
There are an infinite number of possible models � non-uniqueness
How do we choose one?
We don’t know how noisy the data are.
Inversion Theory: Non-uniqueness
Why an infinite number of models?
The data (100’s – 100,000’s values) are not sufficient to uniquely determine 10,000,000’s earth model parameters.
� Under-determined problem.
The physical phenomena that we are exploiting (gravity, electromagnetic propagation) is usually a decaying as a function of depth or distance, and is not sufficient to uniquely describe the earth model.
Questions to consider:
Consider the simple problem that involves two unknowns (model parameters), x and y. We have one datum, 2.
x+y=2
What is the value of x and y?
� Infinite number of solutions
Consider some candidate models for the x and y parameters:
a: (0,2)
b: (1,1)
c: (2,0)
d: (-1,3)
Which one do we choose?
Using prior information to choose optimal models
Encode prior knowledge in a form that can be “optimized”.
i.e. build a mathematical rule or norm to test sizes of possible
models, then choose the “smallest”.
• The people-in-the-room analogy:
source: UBC-GIF
Questions to consider:
Consider the simple problem that involves two unknowns (model parameters), x and y. We have one datum, 2.
x+y=2
What is the value of x and y?
� Infinite number of solutions
Consider some candidate models for the x and y parameters:
a: (0,2) positive gradient
b: (1,1) flattest
c: (2,0) negative gradient
d: (-1,3) has smallest value
norms are described mathematically
How to pick one of infinitely many solutions?
Geophysical prior knowledge:
Values are positive, and/or within bounds
Physical Properties: Estimates for host rock properties
Point-location values from drill hole information
Logical prior knowledge:
Find a “simple” result - as featureless as possible.
This sacrifices resolution but prevents over-interpreting the data.
Geologic prior knowledge:
Character of the model (smooth, discontinuous)
Some idea of scale length (or size) of the bodies
Structural Constraints
Challenge: Describe geology mathematically
Narrow down the number of options using prior knowledge.
Questions to consider:
Consider the simple problem that involves two unknowns (model parameters), x and y. We have one datum, 2.
x+y=2
But we really have:
x+y=2 (+/- unknown error)
Sources of noise:
Instrument noise (sensitivity, accuracy, t0)
Location noise (GPS)
Geologic noise – near surface geology not of interest
Modelling errors – discretization limitations
Operator mistakes
Topography resolution
Inversion Methods
Potential Fields
UBC-GIF GRAV3D and MAG3D
Inversion for a smoothly varying heterogeneous
3D physical property distribution
Recovers important information about subsurface:
• Depth to top of feature
• Centroid of feature
• Approximate physical property
Geological and physical property constraints
Can be used to guide the result closer to
the true earth solution.
True Model
Recovered Model
Unconstrained Synthetic Example
Inversion Methods
Potential Fields
UBC-GIF GRAV3D and MAG3D
Inversion for a smoothly varying heterogeneous
3D physical property distribution
Recovers important information about subsurface:
• Depth to top of feature
• Centroid of feature
• Approximate physical property
Geological and physical property constraints
Can be used to guide the result closer to
the true earth solution.
Half of model constrained
���� improves other half
True Model
Recovered Model
Constrained Synthetic Example
Inversion models in contextGeophysical inversions are non-unique and generated from noisy data.
Be aware of this and use responsibly.
Logical (non-geologic) constraints are a good starting point and add value to the data.
Use prior information to further narrow down the range of suitable models.
Common Earth Models
Geologic models + Geochemical models + Geophysical models
Honour all the data provide the most comprehensive, quantitative view of the subsurface.
Modelling Objectives
• Provide useful 3D physical property products
• For direct employment in regional exploration
• Provide guidance to the regional structure
• Help geologic mapping
• Help target prospective geology, alteration, or mineralization.
• Exploration criteria for different styles of mineralization can be applied based
on multiple physical properties.
• Depth of overburden analysis
• Guide detailed follow-up survey design
Products
• 3D inversions of potential field data
• Interpolated 3D conductivity model based on 1D EM inversions
• Integrated 3D Physical Property Classification Models
• Accessible deliverables for visualization and quantitative 3D analysis
• Detailed infill areas
Summary of data
• Sanders airborne gravity
• GSC gravity compilation
• (Geotech magnetic)
• Aeroquest magnetic
• GSC magnetic compilation
• Geotech VTEM data
• Aeroquest AeroTEM data
Gravity Data
Airborne acquisition by Sander Geophysics
East-West lines with 2000m line spacing
Regional GSC data also used for regional signal
mGal
Terrain Corrected
Bouguer Anomaly
(2.67 g/cm3)
Magnetic Data
Airborne acquisition by Geotech and Aeroquest
East-West lines with 4000m line spacing
Regional GSC data also used for regional signal
nT
Total Magnetic Intensity
Summary of ModelsFour large survey areas and 6 small infill areas:
Bell, Endako, Equity, Huckleberry, Granisle, and Morrison.
Potential Fields:
3D Density Contrast Model (UBC-GIF Grav3D)
3D Magnetic Susceptibility model (UBC-GIF Mag3D)
500m x 500m x 250m cells
Tiled inversions (full compilation = ~100 million cells)
Airborne EM
Late time conductivity map
3D (interpolated) conductivity model (UBC EM1DTM)
Depth of system penetration estimate
Conductive Plates (EMIT Maxwell)
GIS compilation in Gocad
Airborne EM Modelling
1D Inversions
Inversion for a smoothly varying heterogeneous 1D conductivity distribution
Laterally Constrained� Inversion parameters are tuned to the changing geology
Background/Late-Time Conductivity
Depth of Investigation based on cumulative conductance
Plate Modelling
Alternative to the 1D interpretation for use when the layered earth assumption is inadequate.
Electromagnetic Data
Airborne acquisition by Geotech (VTEM system)
East-West lines with 4000m line spacing
Channel 19 dBz/dt data
27 time channels used
~78m flight height
nT/s
Inversion Methods
Airborne EM
UBC-GIF EM1DTM
Inversion for a smoothly varying heterogeneous
1D conductivity distribution
Laterally Parameterized/Constrained Inversion
Neighbouring stations used to determine
appropriate inversion parameters
Reduces modelling artefacts
Background/Late-Time Conductivity
Background Conductivity: Plan View
S/m
Inversion Methods
Airborne EM
UBC-GIF EM1DTM
Inversion for a smoothly varying heterogeneous
1D conductivity distribution
Laterally Parameterized/Constrained Inversion
Neighbouring stations used to determine
appropriate inversion parameters
Reduces modelling artefacts
Background/Late-Time Conductivity
Depth of Investigation based on cumulative conductance
Conductivity Flight-line Section
Inversion Results
Airborne EMConductivity Model
Log conductivity [S/m]
Fences shown through full
3D conductivity model
Conformable with topography
Inversion Results
Log conductivity [S/m]
Airborne EM Conductivity ModelZoom
Fences shown through full
3D conductivity model
Conformable with topography
Using the Results
Quest Block C:
• Conductivity model: East-West flight line model sections
• Density Contrast iso-surface at a value of 0.05g/cm3
• North-South magnetic susceptibility section
Inversion Results
Physical Property Classification
•Common 3D Discretization Mesh
•Each Physical Property classified into High, Medium, and Low
•Density Contrast and Magnetic Susceptibility
Two-phase system � 9 classifications
•Density Contrast, Magnetic Susceptibility, and Conductivity Background
Three-phase system � 27 classifications
Inversion Results
3D Physical Property
Classification
Surficial Plan View
of 3D classification model
Inversion Results
3D Physical Property
Classification
Surficial Plan View
of 3D classification model
Using the Results
• Regional Interpretation
• Integrated Interpretation
• Interpretation with Physical Properties
• Constraining Information
• Target Customization
• Survey Design
• Integrated Modelling
• Common Earth Model Development
• 3D GIS Regional Targeting
Qualitative
Quantitative
Using the Results
Evaluation of physical property classification based on known
mineral occurrences:
1. Compute spatial correlation of physical property classifications with known mineral occurrences in 3D.
2. Determine which class has the highest correlation.
3. Perform for Density Contrast and Magnetic Susceptibility two-phase system, and Density Contrast, Magnetic Susceptibility, and Conductivity Background three-phase system.
Using the Results – evaluation of physical property classification
Distance to known
mineral occurrences.
Targeting Workflow
Establish clear exploration objectives and criteria
Target generation from the Common Earth Model
Target ListX Y Z3045 2465 10103456 4576 11453104 6543 12793303 4531 1109…
Mineral Potential Cube
Summary3D density contrast, magnetic susceptibility, and conductivity models have been produced.
The models provide more useful information than the data alone.
While being aware of the limitations, the models can be used to promote detailed follow-up through 3D-GIS targeting analysis.
The infill areas provide examples of what information can be extracted from these
data.
Introduce new information as it is acquired to test the validation of these models, and to help improve upon them as more focussed targets are resolved.