Citation Program in Applied Geostatistics

Introduction

Welcome
Overview Example
Some Mathematical Background

Objectives of the Workshop

Increase awareness of geostatistics and the situations where

the application of geostatistical techniques could add value
Learn how to assemble the required data for geostatistical
techniques
Introduce statistical and geostatistical techniques -- show some
examples
Provide some statistical / geostatistical tools that will help with
specific tasks (DDH - RC data, multiple variable prediction, 3-D
model building, and uncertainty assessment)
Step through major components of an ore body modeling study
(coordinates, rock type modeling, grade modeling)
Understand the limitations of the resulting numerical models and
the geostatistical techniques
Look at some projects and, perhaps, suggest appropriate
statistical / geostatistical analyses

Lec 01-1 page 1

 Instructors

Dr. Clayton V. Deutsch, P.Eng.

Professor in the School of Mining
and Petroleum Engineering, at the
University of Alberta
Alberta Chamber of Resources
Industry Chair in Mining
Engineering
Canada Research Chair in Natural
Resources Uncertainty
Assessment

Dr. Oy Leuangthong
Assistant Professor in the School
of Mining and Petroleum
Engineering, at the University of
Alberta

Tentative Schedule
Day Topics Covered
1 Math Background
2 Fundamentals of Statistics and Probability
Week 1

3 Sampling, Compositing and Data Integrity

4 Stationarity, Boundaries and Preliminary Trend Modeling
5 Declustering and Debiasing
1 Bivariate Distributions, Variogram/Correlogram Calculation
2 Variogram maps and volumes, Variogram Modeling
Week 2

3 Volume Variance: dispersion and shape change

4 Theory of Kriging
5 Practice of Kriging
1 Multivariate Spatial Distribution
2 Theory of Simulation
Week 3

3 Practice of Simulation
4 Linear Model of Coregionalization and Cokriging
5 Multiple Variable Modeling Approaches
1 Rock Type Modeling: Indicator Technique
2 Rock Type Modeling II: Locally varying probability, Truncated Gaussian Technique
Week 4

3 Trend Modelling and Variance Inflation

4 Post Processing of Simulated Realisations
5 Model Checking and Accuracy

1 Disjunctive Kriging and Uniform Conditioning

2 Multiple Point Statistics
Week 5

3 Other Simulation Approaches: P-field, Direct and Simulated Annealing

4 Decision Making: grade control, conditional bias
5 Project Review

Lec 01-1 page 2

 Historical Perspective

Theory of probability (in its modern form) was formalized in the 1600's
by Blaise Pascal and Pierre de Fermat (Gauss and Bayes were more
recent players)
The foundation for geostatistical techniques was established by people
like Kolmogorov, Weiner, Matern, and Gandin in the early 1900's
Geostatistics was started in the 1960's by Krige and Sichel in South
Africa and Matheron in France. Two of Matheron's first students
(Journel and David) would leave for the USA and Canada and start
new centers of geostatistical research
The application of geostatistical techniques became popular in the
mining industry and meteorology. Now, these techniques are applied
in many diverse applications from fisheries, forestry, environmental
remediation, and so on
Centers for research are numerous, including Stanford, Fountainbleau,
and others such as the University of Alberta

Geostatistics for Mining

Applications (1)
Business Need: make the best possible decisions in the face of
uncertainty. One of the biggest uncertainties is the numerical
description of the ore body.
Statistics is concerned with scientific methods for collecting,
organizing, summarizing, presenting and analyzing data, as well as
drawing valid conclusions and making reasonable decisions on the
basis of such analysis.
Geostatistics is a branch of applied statistics that places emphasis on
(1) the geological context of the data,
(2) the spatial relationship between the data, and
(3) data measured with different volumetric support and precision.
Geostatistics is sometimes referred to as stochastic modeling,
geostatistical reservoir characterization, conditional simulation

Lec 01-1 page 3

 Geostatistics for Mining
Applications (2)
Basic Principles:
work within all known geological (physical) constraints
provide tools to quantify and exploit spatial correlation
algorithms for numerical geological modeling and uncertainty quantification
Doesn't make ore body modeling any easier; just better (if correctly
applied)
Geostatistics is useful for:
Putting geology into numbers
Estimation
Quantifying uncertainty
Sample design
Simulation
Risk analysis
Geostatistics does not:
Replace good additional data
Replace the need for common sense and good judgement
Work well as a black box
Save time

Geostatistics for Mining

Applications (3)
Geostatistical techniques are an indispensable part of the mineral
resource management because quantitative numerical models of
mineral concentrations are required for planning and economic
optimization of the development strategy.
Following are some objectives of studies that use geostatistical tools:
create 3-D block models of grades for mine planning
refine estimates of global reserves
quantify uncertainty in metal content
assess continuity / selectivity concerns
quantify uncertainty in production predictions
Geostatistics provides a variety of tools that are combined in different
ways for different objectives and for particular types of geological
settings
Some Considerations:
Geological understanding / zonation guides geostatistical modeling
Data quality / sampling must be addressed before geostatistical modeling

Lec 01-1 page 4

 Geostatistics for Mining
Applications (4)
When Should Geostatistics be Applied?
At different times in the lifecycle of a deposit:
Exploration Drilling: a pre-feasibility model for in-situ mineral inventory
calculation
Development Drilling: a feasibility model for recoverable reserve calculation
and long term planning
Blast Holes: a production model for grade control and short term planning
Environmental: characterize spatial and temporal interactions between the
mining/milling operation and the local environment
Numerical modeling as a project evolves:
more samples
more geological complexity
tougher questions
The tools that are appropriate at an early stage may not be appropriate
later
Some numerical modeling tools:
estimation (inverse squared distance, kriging, indicator kriging, cokriging)
simulation (Gaussian, indicator, algorithmic)

Why Geostatistics / 3-D Modeling?

Advantages of Geostatistics
Practicality / consistency with data the best approach to model the
spatial distribution of mineral grades
Repeatability / audit-trail
Easy to merge incremental data (provided the conceptual model does
not change)
3-D models lead to better mineral inventories / reserves than set of 2-D
interpretations
Better modeling of variability:
no need for pseudo data
controllable degree of spatial variability
estimates of recoverable reserves are more reliable
Framework to integrate data:
geological interpretation
hard and soft data
data representing different measurement supports
Assessment of uncertainty in process performance due to uncertainty
in geological model

Lec 01-1 page 5

 Key Geostatistical Concepts (1)

Numerical Modeling: At any instance in geological time, there is a single true

distribution of properties in each deposit. This true distribution is the result of a
complex succession of physical, chemical, and biological processes. Although
some of these depositional and diagenetic processes may be understood quite
well, we do not completely understand all of the processes and have no access
to the initial and boundary conditions in sufficient detail to provide the unique
true distribution.

Rock Type 1
Rock Type 2
Rock Type 3
Rock Type 4
Rock Type 5
Rock Type 6

Key Geostatistical Concepts (2)

Uncertainty: All numerical models would be found in error if we were to

excavate that unsampled volume and take exhaustive measurements: There is
uncertainty. This uncertainty exists because of our ignorance/lack of
knowledge. It is not an inherent feature of the deposit.
Uniqueness and Smoothing: Conventional mapping algorithms were devised
to create smooth maps to reveal large scale geologic trends; for mining
applications, however, the extreme high values often have a large affect on the
response (e.g. mine plan).

Mine
Plan
Distribution of
Multiple stochastic models possible responses

Lec 01-1 page 6

 Constructing 3-D Models

The specific process employed for 3-D model building will depend on the
data available, the time available, the type of deposit, and the skills of
the people available. In general, the following major steps are required:
1. Determine the areal and vertical extent of the model and the geological
modeling cell size
2. Establish a conceptual geological model and define zones for modeling
3. For each zone:
1. Define coordinates transformation (e.g. unfolding)
2. Define the number of rock types, the data, and the spatial correlation
3. Generate 3-D rock type model
4. Establish grade values for the different variables and the spatial
correlation
5. Generate 3-D grade models
6. Merge and translate back to real coordinates
4. Verify the model
5. Combine zones into a single model

Software

Introduction to GSLIB
Programs in GSLIB
GSLIB Preliminaries
CCGLIB

Lec 01-1 page 7

 Introduction to GSLIB
GSLIB is an acronym for Geostatistical
Software LIBrary. This name was
originally used for a collection of
geostatistical programs developed by
students and faculty at Stanford
University over the last 15 years.
The original GSLIB inspired the writing of
GSLIB: Geostatistical Software Library
and User's Guide by Clayton Deutsch
and Andr Journel during 1990 - 1992.
This publication (available for about $50
from Oxford University Press) contains a
CD and a user's guide for code based on
the original GSLIB.
Executables and source code available at
www.gslib.com
Most of the original GSLIB code was
rewritten for uniformity and new programs
were added during development
A second edition was completed in
March, 1997
Executables available at www.gslib.com

Programs in GSLIB (1)

There are 39 FORTRAN programs on the GSLIB CD-ROM

PostScript plotting utilities (histograms, probability plots, Q-Q/P-P
plots, scatterplots, location maps, gray/color scale maps)
Utility Programs:
cell declustering
data transformation
IK manipulation: order relations, change of support, p or q values, E-type
estimate,
other theoretical indicator variograms from multiGaussian distribution, linear
system solvers, data calibration (Markov-Bayes), ...
Measures of spatial continuity (Variograms):
direct and cross variograms of any # of variables
1,2, or 3-D regular or irregular spaced data
many different measures (traditional semivariogram, non-ergodic
covariance / correlogram, relative (general and pairwise) variograms,
madogram/rodogram, indicator)

Lec 01-1 page 8

 Programs in GSLIB (2)

Kriging:
1,2 or 3-D grid kriging, cross validation, jackknifing
SK, OK, UK, kriging with external drift
cokriging
indicator kriging
Stochastic Simulation:
Gaussian methods (LU matrix method, sequential)
Indicator methods (cosimulation, Markov Bayes)
Boolean simulation
Simulated annealing
P-field simulation

GSLIB Preliminaries

Input/Output files are ASCII flat files compatible with Geo-EAS and the
Geostatistical Toolbox:

2d data file example:

data2dg.dat) File name
4 Number of columns

}
Xlocation
Ylocation
Column names
ATTRIBUTE
COVARIATE

}
28.5 23.5 3.852 1.003
30.5 45.5 5.543 1.772
Space/Tab delimited data
22.5 18.5 .102 .327
49.5 35.5 4.548 2.833

Graphical output is PostScript page description language (GhostScript

would be a useful supplement).

Lec 01-1 page 9

 Grid Specification
The GSLIB rectangular grid is denoted as:

Could be associated to any Cartesian or stratigraphic coordinate

system.
Output grids are ordered with x cycling fastest, then y, and then z.
Index location of node ix,iy,iz is located by:
loc = (iz-1)*nx*ny + (iy-1)*nx + ix
2-D grids are specified with nz=1. zmn and zsiz are not used.
Grids or simulated realizations are ordered one after another.
Some sort of compressed output format may be necessary.

Example Parameter File

Parameters for GAMV
*******************
START OF PARAMETERS:
../data/cluster.dat \ file with data
1 2 0 \ columns for X, Y, Z coordinates
2 3 4 \ number of varables,column numbers
-1.0e21 1.0e21 \ trimming limits
gamv.out \ file for variogram output
10 \ number of lags
5.0 \ lag separation distance
3.0 \ lag tolerance
3 \ number of directions
0.0 90.0 50.0 0.0 90.0 50.0 \azm,atol,bandh,dip,dtol,bandv

Lec 01-1 page 10

 Review of Main Points
Numerical model is not reality
Get the big features right, then iterate to refine numerical model and
mining plan
Try to anticipate the variability that may be encountered
Mining blocks have significantly less variability than drillhole data:
volume / variance relations
More information will be available at the time of mining: information
effect
Easy to upscale detailed numerical models (by arithmetic averaging),
but it is impossible to downscale too-large modeling cells.
There is a practical limit to number of cells in our numerical models,
due to data storage and processing requirements.
fit-for-purpose modeling - Always ask: What is the problem being
addressed? What is the model going to be used for?
Additional time, data, professional interpretation effort always lead to
better estimates. Is it worth it?

Recall of Notation and Some Math

Notation
Differentiation
Integration
Summation
Matrix

Lec 01-1 page 11

 Some Basic Notation
Upper case letters (X, Z,) are often for random variables (RVs)
Lower case letters (x,z,) are often for outcomes of random variables
Bold font (u, h,) is often reserved for vectors with some component in
coordinate directions
We subscript a list of numbers with characters i,j,k, or , ,
The letter n or N is often used to denote the number of data
The letter L is often used to denote the number of realizations
The letter K is often used to denote the number of rock types
Some other symbols:

Derivatives

The slope of a function

Consider f(x)=x2

Lec 01-1 page 12

 Integration

The area problem

Fundamental Theorem of Calculus

Let f (x) be continuous on [a, b]. If F(x) is any antiderivative of f

(x), then

The second fundamental theorem of calculus holds for f a

continuous on an open interval I and a any point in I, and states
that if F is defined by

then

Lec 01-1 page 13

 Recall of Algebra and Notation

n
Linear Sum: Y = a1 z1 + a2 z2 + + an zn = a z
=1
Linear property of expected value:
E{Y } = E a z = a E{z }
n n

=1 =1
Quadratic form:
2
n n
Y 2 = a z = a a z z
n

=1 =1 =1
e.g.
n=1 :
(a1 z1 )2 = a1a1 z1 z1
n=2 :
(a1 z1 + a2 z2 )2 = a12 z12
+ a1a2 z1 z 2 + a2 a1 z 2 z1 + a22 z 22

 

=1 =2

More Algebra and Notation

n n n n
and Y 2 = a a z z = a2 z2 + a a z z
=1 =1 =1 =1

Expected value of : Y2

E{Y 2 } = E a a z z = a a E z z

Differentiation of a sum:

Y n
=
a0 a0



a z


= z0
=1

with 0 = one of the n 's

Y 2
a a z z = 2 z0 a z

=
a 0 a 0

Lec 01-1 page 14

 Matrix Algebra
a b
c d 2 rows and 2 columns: 2 x 2 matrix

rows columns

Matrix multiplication: multiple (m x n)(n x r) = m x r matrix

a b e ae + bf
c d f = ce + df
(2 x 2)(2 x 1) = 2 x 1

Order matters in matrix multiplication:

a b
[e f ] = [ ea + fc eb + fd ]
c d (1 x 2)(2 x 2) = 1 x 2

Matrix Algebra
a b e
Short form notation
X= A=
& c d f
Matrix Transpose
a c
XT = AT = [e f]
b d
Matrix multiplication:

a b e a b
c d f = XA [e f ] = AT X
c d

Lec 01-1 page 15