Geostatistics
• Classical statistics, random distributions, normal and lognormal
theory. Concept of geo-statistics and its application to mining
• Spatial statistics, Covariogram, definitions, estimation, fitting
• Variogram, semi-variogram, defination, estimation,
experimental variogram, fitting; application, uses of variogram.
• Various model of variograms: random model, spherical model,
expontial model, gaussian model,
• Linear model, logarithmic or de Wijsian model, parabolic model
• Nugget effect, its implication on model, anisotropies
• Numerical calculation of variogram, and graphs for one, two
and three dimensions
• Krigging method for grade and reserve estimation.
• Krigging estimator, krigging error, point krigging, block krigging,
optimal valuation
• Use for geostatistical software for various application
• Case studies of grade estimation
Kriging
5
3 2 6
Estimation of ore grade at A
Here some basic questions may arise that
• Should every sample be considered?
•Should every sample be given equal weight age ?
•If sample 6 is too far then also it should be considered?
Rainfall data in Switzerland using Kriging
Distribution of a mineral in an area
Interpolation of monthly precipitation of
1998 using Kriging
Kriging result example (2)
Jul. 1975 Oct. 2000
-The most spatially heterogeneous - The most spatially homogeneous
- Mean: 204.76 - Mean: 0.30
- SD: 82.71 - SD: 0.22
Interpolation: Estimating the attribute values of
locations that are within the range of available data
using known data values
Estimating a point
here: interpolation
Sample
data
Extrapolation: Estimating the attribute values
of locations outside the range of available
data using known data values
Sample
data
Estimating a point
here: extrapolation
Geostatistical estimation
• Geostatistical estimation is a two stage process:
• i. Studying the gathered data to establish the predictability
of values from place to place in the study area; this study
results in a graph known as a semi-variogram which models
the difference between a value at one location and the value
at another location according to the distance and direction
between them;
• ii. Estimating values at those locations which have not
been sampled. This process is known as 'kriging'. The basic
technique "ordinary kriging" uses a weighted average of
neighbouring samples to estimate the 'unknown' value at a
given location. Weights are optimized using the semi-
variogram model, the location of the samples and all the
relevant inter-relationships between known and unknown
values. The technique also provides a "standard error" which
may be used to quantify confidence levels.
Kriging
• In mining, Kriging is extensively used in the field of
mineral resource and reserve valuation - the
estimation of grades and other parameters from a
relatively small set of borehole or other samples.
• Geostatistics is now widely used in many other
fields. Obviously there are geological and
geographical applications. However, the
techniques are also used in such diverse fields as
hydrology, ground water and air pollution, soil
science and agriculture, forestry, epidemiology,
management of wildlife and weather prediction.
Kriging…….
• Kriging is a geostatistical interpolation technique used in
mining industry for interpolation of input point data and
estimation of a block model (mineral resource model).
• The name “kriging” was given by Prof. Georges Matheron
in honor of the South African mining engineer Danie
Krige.
• The empirical semi variogram provides information on the
spatial auto correlation of data sets.
• However, it does not provide information for all possible
directions and distances.
• For this reason, the kriging is used, and its main
application is the prediction of attribute values at
unsampled locations.
• Kriging is geostatistical techniques used for interpolation
(mapping and contouring) purposes.
Kriging…….
• It is generalized forms of univariate and
multivariate linear regression models, for estimation at a
point, over an area, or within a volume.
• Kriging assumes that the distance or direction between
sample points reflects a spatial correlation that can be
used to explain variation in the surface.
• It is linear-weighted averaging method, similar to other
interpolation methods; however, their weights depend
not only on distance, but also on the direction and
orientation of the neighboring data t the unsampled
location.
• The input data (used to estimate any particular block) are
weighted based on the variogram model, search
parameters and the number of samples used to krige or
estimate a block.
History of Kriging
• The Kriging method was developed by the French mathematician
Georges Matheron based on the Master's thesis of Danie G. Krige.
• Danie G Krige (26 Aug. 1919 – 3 March 2013) was a South
African Statistician and Mining Engineer who pioneered the field
of geostatistics and was professor at the University of the
Witwatersrand, Republic of South Africa.
• The technique of kriging is named after him. Krige's empirical work
to evaluate mineral resources was formalised in the 1960s
by French engineer Georges Matheron.
• Krige sought to estimate the most likely distribution of gold based
on samples from a few boreholes.
• The basic idea of kriging is to predict the value of a function at a
given point by computing a weighted average of the known values
of the function in the neighborhood of the point.
• The method is mathematically closely related to regression
analysis.
Types of kriging
• Linear kriging
• Indicator kriging (nonlinear )
• Disjunctive kriging, (nonlinear )
In this semester we will discuss study only
linear kriging (henceforth“kriging”)
techniques.
Theory of kriging
• Kriging is known by the acronym BLUE: or, the
Best Linear Unbiased Estimator.
• In the kriging process the sum of weights
assigned to the input data is adjusted to one,
and the error of estimation is kept to a
minimum.
• Kriging utilizes the variogram, which does not
depend on the actual value of the variable
(data), rather its spatial distribution and internal
spatial structure.
Theory of kriging……….
• A variogram provides penetrating insight
concerning the data used to construct a
kriging interpolation system.
• When a good variogram model is available,
kriging provides the estimation best
representing the spatial distribution of the
input data.
Common problems associated with the
use of the kriging technique
1. Variograms do not accurately represent the mineralized zone
because of inadequate data.
2. Mathematical models do not accurately fit the variogram
data, or variograms have been improperly interpreted.
3. Kriging is insensitive to variogram coefficients.
4. Computational problems and expense are associated with
repeatedly inverting large matrices.
5. In the kriging weights are derived based on a data-driven
weighting function making the sum of the weights equal to
one, thereby reducing the effect of bias towards input
sample values.
Factors affecting use of kriging
• Type of spatial data distribution and structure,
• It should be applied only if the underlying assumptions of
second-order stationarity are met, i.e., at a minimum the
mean and variance of the sample data remain invariant in
space.
• If the mean and variance of the samples do not meet these
stationarity assumptions, it is advisable to apply an
interpolation technique other than kriging.
• If input samples are clustered, then a declustering application
should be used before variogram analyses.
• The kriging technique takes care of the cluster effect, whereas
other interpolation techniques fail to do so.
Factors affecting use of kriging………
• post kriging analyses, etc. If the input data are located in a near-
perfect grid pattern, it is likely that kriging will produce similar
estimates as other interpolation techniques, but may reduce the
bulls-eye effect (as often seen in inverse distance square estimates).
• Since geology is an integral part of geostatistics, much attention is
required to understand the geology and identify individual domains.
• The other very important factor is the homogeneity, i.e., the data
represents one single homogenous domain. Such a domain can be
characterized by lithology, geochemical characteristics and
geological structures.
• The other basic requirement is sufficient number of data points.
Even though there is no defined criterion for the number of samples
to be used in the kriging process, however approximately 40
samples are needed to reasonably define and model a variogram,
which can be used to produce an acceptable kriged estimate.
Pre Kriging Data Processing
It is essential to analyze the data and establish the
statistical parameters so that a near-robust
kriging profile can be designed for a reasonable
kriging estimate. The pre-kriging data analyses
are summarized below.
• The data set needs to be divided into subsets
each representing a homogenous domain.
• Domains should to be defined based on geology,
geochemistry and geological structures.
• A geological domain (lithology) divided by a fault
should be considered as two domains.
Advantages of kriging
• Helps to compensate for the effects of data clustering,
assigning individual points within a cluster less weight
than isolated data points (or, treating clusters more like
single points)
• kriging provides flexibility to choose the technique that
is most appropriate for a particular mineral deposit. The
popular non-linear techniques include indicator kriging,
for the estimation of discrete variables; lognormal
kriging, for highly skewed distributions; disjunctive
kriging, for non-linear problems; and factorial kriging
analysis, to extract components of a model.
• Gives estimate of estimation error (kriging variance).
• Availability of estimation error provides basis for
conditional simulation of possible realizations of grade.
• A uniquely designed kriging technique can be used to
estimate confidence intervals of block estimates.
Disadvantages of kriging
• Negative weights: It is possible that, due to clustering,
some samples may receive negative weights. Negative
weights may result in negative kriged estimates, which is
highly undesirable.
• Similar results may occur if a highly structured variogram
is used, ignoring the inherent randomness (nugget effect)
in the structure.
• A negative block can be assigned a zero value to avoid
problems in post kriging analyses, such as pit-
optimization. Such results can be avoided if a more
realistic variogram is used and/or data are declustered.
• A deterministic technique: Kriging, being a deterministic
interpolation technique, does not allow assessment of the
risk of resource-estimation.
Applications of kriging
Although kriging was developed originally for applications in
geostatistics, it is a general method of statistical interpolation
that can be applied within any discipline to sampled data from
random fields that satisfy the appropriate mathematical
assumptions. It can be used where spatially-related data has
been collected (in 2-D or 3-D) and estimates of "fill-in" data are
desired in the locations (spatial gaps) between the actual
measurements. To date kriging has been used in a variety of
disciplines, including the following:
• Environmental science
• Hydrogeology
• Mining
• Natural resources
• Remote sensing
• Real estate appraisal
• Integrated Circuit Analysis and Optimization
• Modelling of Microwave Devices
Kriging equations
• The objective of kriging can be simply
expressed as that of determining coefficients
ai’s which when multiplied by the known grades
gi and the resulting products summed will yield
a best estimate of the grade g0. The equation
developed is called a linear estimator and has
the form
g0 = a1g1 + a2g2+ a3g3 …………..
In the simplified kriging equation, the kriged
estimated grade is given by
g0 = a1g1 + a2g2 + a3g3………..
Where: g0 is the location where grade is to be estimated
g1 , g2 , g3 etc are the known samples at given location
a1 , a2 , a3 etc are the assign weights to each samples
Such that a1 + a2 + a3……….. = 1
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Where
h = lag interval
a = range
C0 = nugget
C1 = Sill - C0
The basic kriging equations are:
a1σ11 + a2σ12 + a3σ13 ………..+ ʎ = σ01
a1σ21 + a2σ22 + a3σ23 ………..+ ʎ = σ02
a1σ31 + a2σ32 + a3σ33 ………..+ ʎ = σ03
…………….
Such that a1 + a2 + a3……….. = 1
Where ʎ = Lagrange multiplier
Problem - 1
Determine the Fe % at point X0, if two boreholes
drilled nearby X1 and X2 are having the grade conc as
40 % and 60 % as shown in fig. It is given that C0 =
0.1, C1 = 0.25 and a=250.
i) A = 100 M and B = 150 M
ii) A = 200 M and B = 300 M
iii) A = 200 M and B = 500 M
iv) A = 300 M and B = 300 M
v) A = 100 M and B = 100 M
vi) A = 300 M and B = 500 M
X1 X0 X2
A B
i) A = 100 M and B = 150 M
First let us determine the required distances:
• X0X1 = 100 M X1 X0 X2
• X0X2 = 150 M 150 M
100 M
• X1 X2 = 250
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 100 0.24 0.108
X0X2 150 0.298 0.05
X1X2 250 0.35 0
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0a2 + λ = 0.108
0a1 + 0.35a2 + λ = 0.05
a1 + a2 = 1
0.35a1 + 0a2 + λ = 0.108
0a1 + 0.35a2 + λ = 0.05
a1 + a2 = 1
Solving
a1 = 0.58 a2 = 0.42 λ = - 0.095
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.58*40 + 0.42*60
g0 = 48.4 Ans
ii) A = 200 M and B = 300 M
X1 X0 X2
200 M 300 M
It is given that C0 = 0.1, C1 = 0.25 and a=250.
First let us determine the required distances:
X1 X0 X2
200 M 300 M
• X0X1 = 200 M
• X0X2 = 300 M
• X1 X2 = 500
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 200 0.336 0.014
X0X2 300 C0 + C1 = 0.35 h>a, 0
X1X2 500 C0 + C1 = 0.35 H>a, 0
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0a2 + λ = 0.014
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
0.35a1 + 0a2 + λ = 0.014
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
Solving
a1 = 0.52 a2 = 0.48 λ = - 0.168
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.52*40 + 0.48*60
g0 = 49.6 Ans
iii) A = 200 M and B = 500 M
X1 X0 X2
200 M 500 M
It is given that C0 = 0.1, C1 = 0.25 and a=250.
First let us determine the required distances:
X1 X0 X2
200 M 500 M
• X0X1 = 200 M
• X0X2 = 500 M
• X1 X2 = 700
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 200 0.336 0.014
X0X2 500 C0 + C1 = 0.35 h>a, 0
X1X2 700 C0 + C1 = 0.35 H>a, 0
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0a2 + λ = 0.014
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
0.35a1 + 0a2 + λ = 0.014
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
Solving
a1 = 0.52 a2 = 0.48 λ = - 0.168
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.52*40 + 0.48*60
g0 = 49.6 Ans
iv) A = 300 M and B = 300 M
X1 X0 X2
300 M 300 M
It is given that C0 = 0.1, C1 = 0.25 and a=250.
First let us determine the required distances:
X1 X0 X2
300 M 300 M
• X0X1 = 300 M
• X0X2 = 300 M
• X1 X2 = 600
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 300 C0 + C1 = 0.35 h>a, 0
X0X2 300 C0 + C1 = 0.35 h>a, 0
X1X2 600 C0 + C1 = 0.35 H>a, 0
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0a2 + λ = 0
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
0.35a1 + 0a2 + λ = 0
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
Solving
a1 = 0.5 a2 = 0.5 λ = - 0.175
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.5*40 + 0.5*60
g0 = 50 Ans
v) A = 100 M and B = 100 M
X1 X0 X2
100 M 100 M
It is given that C0 = 0.1, C1 = 0.25 and a=250.
First let us determine the required distances:
X1 X0 X2
100 M 100 M
• X0X1 = 100 M
• X0X2 = 100 M
• X1 X2 = 200
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 100 0.242 0.108
X0X2 100 0.242 0.108
X1X2 200 0.336 0.014
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0.014a2 + λ = 0.108
0.014a1 + 0.35a2 + λ = 0.108
a1 + a2 = 1
0.35a1 + 0.014a2 + λ = 0.108
0.014a1 + 0.35a2 + λ = 0.108
a1 + a2 = 1
Solving
a1 = 0.5 a2 = 0.5 λ = - 0.074
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.5*40 + 0.5*60
g0 = 50 Ans
vi) A = 300 M and B = 500 M
X1 X0 X2
300 M 500 M
It is given that C0 = 0.1, C1 = 0.25 and a=250.
First let us determine the required distances:
X1 X0 X2
300 M 500 M
• X0X1 = 300 M
• X0X2 = 500 M
• X1 X2 = 800
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.1, C1 = 0.25 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 300 C0 + C1 = 0.35 h>a, 0
X0X2 500 C0 + C1 = 0.35 h>a, 0
X1X2 800 C0 + C1 = 0.35 H>a, 0
X0X0=X1X1=X2X2 0 0 0.35
0.35a1 + 0a2 + λ = 0
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
0.35a1 + 0a2 + λ = 0
0a1 + 0.35a2 + λ = 0
a1 + a2 = 1
Solving
a1 = 0.5 a2 = 0.5 λ = - 0.175
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.5*40 + 0.5*60
g0 = 50 Ans
C0 = 0.1, C1 = 0.25 and a=250
X Y g0
i 100 150 48.4
ii 200 300 49.6
iii 200 500 49.6
iv 300 300 50
v 100 100 50
vi 300 500 50
Problem - 2
Determine the Ni % at point X0, if two boreholes
drilled nearby X1 and X2 are having the grade
conc as 12 and 25 % as shown in fig. It is given
that C0 = 0.18, C1 = 0.32 and a=250.
100M
100M
X1 X0
X2
First let us determine the required distances:
100M
100M
X1 X0
X2
• X0X1 = 100 M
• X0X2 = Sqrt(1002+ 2002) M = 223.6 M
• X1 X2 = Sqrt(2002+ 2002) M = 282.84 M
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.18, C1 = 0.32 and a=250
Pair Lag (h) ϒ (%)2 σ
X0X1 100 0.362 0.138
X0X2 223.6 0.495 0.005
X1X2 282.84 C0 + C1 = 0.5 H>a, 0
X0X0=X1X1=X2X2 0 0 0.5
0.5a1 + 0a2 + λ = 0.138
0a1 + 0.5a2 + λ = 0.005
a1 + a2 = 1
0.5a1 + 0a2 + λ = 0.138
0a1 + 0.5a2 + λ = 0.005
a1 + a2 = 1
Solving
a1 = 0.633 a2 = 0.367 λ = -0.1785
The estimated grade is g0 = a1g1 + a2g2
g0 = 0.633*12 + 0.367*25
g0 = 16.77 Ans
Problem - 3
Determine the Zn % at point X0, if three
boreholes drilled nearby X1, X2 & X3 are having
the grade conc as 0.32, 0.23 and 0.33% as shown
in fig. It is given that C0 = 0.02, C1 = 0.16 and
a=200.
100M
100M
X1
X0
X3
X2
First let us determine the required distances:
100M
• X0X1 =Sqrt(1002 +2002 )=224
100M
X1
• X0X2 = Sqrt(2002 +2002 )= 282.84
• X0X3 = Sqrt(1002 +2002 )= 224 X0
• X1 X2 = Sqrt(1002 +4002 )=412.31 X3
• X1 X3 = Sqrt(3002 +3002 )= 424.26 X2
• X2 X3 = Sqrt(1002 +4002 )= 412.31
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.02, C1 = 0.16 and a=200.
Pair Lag (h) ϒ (%)2 σ
X0X1 224 0.18 0
X0X2 282.84 0.18 0
X0X3 224 0.18 0
X1X2 412.31 0.18 0
X1X3 424.26 0.18 0
X2X3 412.31 0.18 0
X0X0=X1X1=X2X2 =X3X3 0 0 0.18
The basic kriging equations are:
Here I = 1 – 3, and j= 1-3
These equations becomes
0.18a1 + 0a2 + 0a3 + λ = 0 a1 = 0.333
0a1 + .18a2 + 0a3 + λ = 0 a2 = 0.333
0a1 + 0a2 + .18a3 + λ = 0 a3 = 0.333
a1 + a2 + a3 = 1 λ = - 0.06
The estimated grade is g0 = a1g1 + a2g2 + a3g3
g0 = 0.333*.32 + .333*.23 + .333*.33
g0 = 0.293 % Ans
Problem - 4
Determine the Zn % at point X0, if three
boreholes drilled nearby X1, X2 & X3 are having
the grade conc as 0.32, 0.23 and 0.33% as shown
in fig. It is given that C0 = 0.02, C1 = 0.16 and
a=250.
100M
100M
X1
X0
X3
X2
First let us determine the required distances:
100M
• X0X1 =Sqrt(1002 +2002 )=224
100M
X1
• X0X2 = Sqrt(2002 +2002 )= 282.84
• X0X3 = Sqrt(1002 +2002 )= 224 X0
• X1 X2 = Sqrt(1002 +4002 )=412.31 X3
• X1 X3 = Sqrt(3002 +3002 )= 424.26 X2
• X2 X3 = Sqrt(1002 +4002 )= 412.31
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.02, C1 = 0.16 and a=250.
Pair Lag (h) ϒ (%)2 σ
X0X1 224 0.177 0.003
X0X2 282.84 0.18 0
X0X3 224 0.177 0.003
X1X2 412.31 0.18 0
X1X3 424.26 0.18 0
X2X3 412.31 0.18 0
X0X0=X1X1=X2X2 =X3X3 0 0 0.18
The basic kriging equations are:
Here I = 1 – 3, and j= 1-3
These equations becomes
0.18a1 + 0a2 + 0a3 + λ = 0.003 a1 = 0.339
0a1 + .18a2 + 0a3 + λ = 0 a2 = 0.322
0a1 + 0a2 + .18a3 + λ = 0.003 a3 = 0.339
a1 + a2 + a3 = 1 λ = - 0.058
The estimated grade is g0 = a1g1 + a2g2 + a3g3
g0 = 0.339*.32 + .322*.23 + .339*.33
g0 = 0.2944 % Ans
Assignment - 2
Q 1: Discuss the various problems associated with the use of the
kriging technique.
Q 2: Determine the Fe % at point X0, if three boreholes drilled
nearby X1, X2 & X3 are having the grade conc as R. N., R.N. + 2
and R N + 5 % as shown in fig. It is given that C0 = 0.06, C1 = 0.24
and a=350. 100M
100M X1
Last date of
submission X0
13th April X3
X2
2018
Problem - 5
Determine the Zn % at point X0, if three
boreholes drilled nearby X1, X2 & X3 are having
the grade conc as 0.32, 0.23 and 0.833% as
shown in fig. It is given that C0 = 0.02, C1 = 0.16
and a=450.
100M
100M
X1
X0
X3
X2
First let us determine the required distances:
• X0X1 = Sqrt(1002 +1002 )=141 100M
100M
• X0X2 = Sqrt(1002 +1002 )=141 X1
• X0X3 = Sqrt(1002 +2002 )= 224 X0
• X1 X2 = 200 X3
X2
• X1 X3 = Sqrt(2002 +3002 )= 361
• X2 X3 = 300
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.02, C1 = 0.16 and a=450.
Pair Lag (h) ϒ (%)2 σ
X0X1 141 0.0927 0.0873
X0X2 141 0.0927 0.0873
X0X3 224 0.1296 0.0504
X1X2 200 0.1196 0.0604
X1X3 361 0.1712 0.0088
X2X3 300 0.1563 0.0237
X0X0=X1X1=X2X2 =X3X3 0 0 0.18
The basic kriging equations are:
Here I = 1 – 3, and j= 1-3
These equations becomes
0.18a1 + .0604a2 + .0088a3 + λ = 0.0873 a1 = 0.390
0.0604a1 + .18a2 + .0237a3 + λ = 0.0873 a2 = 0.359
0.0088a1 + .0237a2 + .18a3 + λ = 0.0504 a3 = 0.251
a1 + a2 + a3 = 1 λ = 0.00677
The estimated grade is g0 = a1g1 + a2g2 + a3g3
g0 = 0.390*.32 + .359*.23 + .251*.833
g0 = 0.416 % Ans
Problem - 6
Determine the Pb % at point X0, if four boreholes
drilled nearby X1, X2 X3 & X4 are having the grade
conc as 10, 15, 20 and 25% as shown in fig. It is
given that C0 = 0.04, C1 = 0.16 and a=200.
100M
X4
100M
X1
X0
X3
X2
First let us determine the required distances:
• X0X1 = Sqrt(1002 +2002 )=224
• X0X2 = Sqrt(2002 +2002 )= 282.84 100M
• X0X3 = Sqrt(1002 +2002 )= 224
100M
X1 X4
• X0X4 = Sqrt(2002 +2002 )= 282.84
• X1 X2 = Sqrt(1002 +4002 )=412.31 X0
• X1 X3 = Sqrt(3002 +3002 )= 424.26
• X1 X4 = 300 X3
• X2 X3 = Sqrt(1002 +4002 )= 412.31 X2
• X2 X4 = Sqrt(4002 +4002 )= 565.68
• X3 X4 = 300
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0.04, C1 = 0.16 and a=200.
Pair Lag (h) ϒ (%)2 σ
X0X1 224 0.2 0
X0X2 282.84 0.2 0
X0X3 224 0.2 0
X0X4 282.84 0.2 0
X1X2 412.31 0.2 0
X1X3 424.26 0.2 0
X1X4 300 0.2 0
X2X3 412.31 0.2 0
X2X4 565.68 0.2 0
X3X4 300 0.2 0
0 00 0.2
The basic kriging equations are:
Here I = 1 – 4, and j= 1-4
These equations becomes
σ11a1 + σ12a2 + σ13a3 + σ14a4 + λ = 0
σ21a1 + σ22a2 + σ23a3 + σ24a4 + λ = 0
σ31a1 + σ32a2 + σ33a3 + σ34a4 + λ = 0
σ41a1 + σ42a2 + σ43a3 + σ44a4 + λ = 0
a1 + a2 + a3 + a4 = 1
0.2a1 + 0a2 + 0a3 + 0a4 + λ = 0
0a1 + 0.2a2 + 0a3 + 0a4 + λ = 0
0a1 + 0a2 + 0.2a3 + 0a4 + λ = 0
0a1 + 0a2 + 0a3 + 0.2a4 + λ = 0
a1 + a2 + a3 + a4 = 1
a1 = 0.25
a2 = 0.25
a3 = 0.25 λ = - 0.05
a4 = 0.25
The estimated grade is g0 = a1g1 + a2g2 + a3g3+ a4g4
g0 = 0.25*10 + 0.25*15 + 0.25*20 + 0.25*25
g0 = 17.5 % Ans
Problem - 7
Determine the Gold content at point X0 & XA, if
three boreholes drilled nearby X1, X2 & X3 are
having the gold conc as 10, 12 and 18 ppm
respectively as shown in fig. It is given that
C0 = 0, C1 = 0.25 and a=400.
100M
100M
X1
X0 X3
XA
X2
First let us determine the required distances:
• X0X1 = Sqrt(2002 +1002 )=224
100M
• X0X2 = Sqrt(2002 +2002 )= 282.8
100M
• X0X3 = 200 X1
• X1 X2 = 300 X0 X3
• X1 X3 = Sqrt(1002 +4002 )= 412.3
• X2 X3 = Sqrt(2002 +4002 )= 447.2
X2
The corresponding values of the variance ϒ and
covariance σ are found using the general formulas
Calculation for variance (ϒ) and covariance (σ) for different lag (h)
C0 = 0, C1 = 0.25 and a=400..
Pair Lag (h) ϒ (ppm)2 σ
X0X1 224 0.188 0.062
X0X2 283 0.221 0.029
X0X3 200 0.172 0.078
X1X2 300 0.229 0.021
X1X3 412 0.25 0
X2X3 447 0.25 0
X0X0=X1X1=X2X2 =X3X3 0 0 0.25
The basic kriging equations are:
Here I = 1 – 3, and j= 1-3
These equations becomes
0.25a1 + .021a2 + 0a3 + λ = 0.062 a1 = 0.1957
0.021a1 + .25a2 + 0a3 + λ = 0.029 a2 = 0.296
0a1 + .0a2 + .25a3 + λ = 0.078 a3 = 0.508
a1 + a2 + a3 = 1 λ = -0.049
The estimated grade is g0 = a1g1 + a2g2 + a3g3
g0 = 0.1957 *10 + .0.296 *12 + 0.508 *18
g0 = 14.653 ppm Ans
Thank you
Case study
Case study of Groundwater Level determination
using Kriging
Study Area
• The study area is located in the north-western part in the state of
Rajasthan, India.
• The study area forms a part of the vast expanse of the Great Indian
Desert, the Thar, and is part of the command area of Indira Gandhi
Nahar Pariyojana (IGNP).
• The climate of the area is arid with extremes of temperature
(maximum upto 50°C and minimum upto 1°C), low erratic rainfall
(annual rainfall of about 250mm, of which 90% is received during
south-west monsoon in the months of June to September)
• The groundwater level in the area is rising since the commencement
of canal irrigation leading to water logging in the area .
• For this study, groundwater level data pertaining to pre-monsoon
(June) and post-monsoon (September) seasons over the years from
1985 to 1990 covering an area of 2100 sq. km were selected.
Plan of canal
network and location
of observation wells
Experimental and fitted semivariogram
for different data sets
Experimental and fitted semivariogram
for different data sets
Basic statistics of the data set
Summary details of fitted variogram
Groundwater level contours (m) by
Kriging Method
Thank you
Geostatistical Software
• ISATIS
• GS+
• WinGslib
• GEOPACK
• S-GeMS (Stanford Geostatistical Modeling Software)
• RGeostats
