UNIVERSITY OF ENGINEERING AND TECHNOLOGY,

PESHAWAR

GEOSTATISTICAL ORE RESERVE ESTIMATION

(LAB 6)
(COURSE CODE MinE_303L)

NAME: MUGHEERA AYAZ

REG.NO: 20PWMIN0871

SUBMITTED TO:
DR. KHAN MUHAMMAD

DEPARTMENT OF MINING ENGINEERING,

PESHAWAR
 DETERMINATION OF EXPERIMENTAL
VARIOGRAM AND VARIOGRAM MODELLING

6.1.OBJECTIVES:
 To derive the experimental variogram in MS excel
 To model the variogram in the MS excel

6.2.INSTRUMENTS/SOFTWARES:
 MS excel

6.3.THEORY:
The construction of the variogram is not one step
process for in the estimation in the kriging estimation process. While
we are using the kiriging estimator for the estimation then it is
dependent on the variogram. For the construction of variogram, first
of all the samples pairing is done by selecting a suitable lag
distances for pairing and also the lag tolerance for that. The pairing
is done on the basis of the lag distance specified then a scatter plot is
obtained which is the complete description of the data points which
are under the observation. It just excludes the outliers and contains
the description of only those samples which are at a distance as

Figure 1: complete description by a scatter plot after pairing

15/01/2023 Page 2
specified lag distance. This is what we say is scatter plot.

The second one is the scatter plot and z(x) is a sample point at a
position and z(x+h) is the other paired sample separated at a
distance h, then using that scatter plot which is the complete
description, the summary statistics can then be drawn. By using
the scatter plot we can draw any kind of summary statistics as it
provides the complete description.
Any type of variogram can be drawn by using the complete statistics
as provided by the scatter plot. Either we can draw a correllogram or
a medogram out of it.
The variogram is then plotted using the complete description as
provided by the scatter plot. The variogram model is plotted the
way that on Y axis we have semi variance and along x axis. The lag
distance is already specified at which the sample pairing has taken
place. Where the semi variance is the half the mean squared
distance of the sample points.
Semi variance values are found by taking the difference of both
sample points making a pair and then do this for all sample points
that are paired with each other and this difference is squared then.
Then their mean found by dividing by the total number of the
sample points.

TERMINOLOGIES IN THE VARIOGRAM MODELING:

VARIOGRAM: a variogram is a kind of model that is modeled to

show the way how two variables are correlated to each other in
space and time. A variogram model is pre requisite to the kriging.

15/01/2023 Page 3
LAG DISTANCE:
It is the distance or the spacing between
two sample points which are separated by a distance h.
SEMI VARIANCE:
semi variance is the measure of dissimilarity
of the sample points as the function of separating distance h. it is the
half of the mean squared difference of the sample points.

RANGE:
Range is the maximum lag distance at which the sill is
reached.
STABILIZED VARIOGRAM:

The point at which the

dissimilarity is maximum and it specifies the point beyond which the
sample are no more correlate to each other. It corresponds the
maximum distance at which the samples are stuck to be correlated
to each other and beyond that extent there is no more correlation
found.
SILL:
Sill is the variogram value that corrospondse to the range.
It is the measure of maximum semi variance or maximum
dissimilarity.
Sill= nugget + continuity

CONTINUITY:
It is the part of the sill which describes the
structure of the spatial process. It represents the state of being
connected or unbroken for the spatial correlation of the sample
15/01/2023 Page 4
points.
NUGGET:
The part of the sill where there is not any structure of
the spatial process. It shows the region where no correlation exist at
all.
VARIOGRAM RELATION WITH THE SEMI VARIANCE:

Semi variance in the variogram represents the measure of

dissimilarity as a function of separating distance h. In a proper
spatial structure for which the geostatistics could be applied, the
semi variance of the data points at the smaller distance is smaller as
the closer samples are more similar to each other while similarity
decreases with the increase in the distance between them and hence
the semi variance will increase with an increase in the lag distance
between them. The relation between the variogram and the semi
variance can be described by the following relation.
Gamma(h)=covariance(o)-covariance(h)

Where gamma(h) is the semi variance and the c(h) is the

covariance as the measure of similarity of the sample data points.
On the other hand, c(o) is the covariance at point zero which
represents the sill. The semi variance and the covariance are the
mirror images of each other.

6.4.PROCEDURE:
The avriogram is constructed by using the
data as input in the MS excel. The data that is provided at the star
of the practical is as shown in the figure the X coordinates, y

15/01/2023 Page 5
coordinates as distance and the grades.

 Then various lag distances are specified at a spacing of 5.

Figure 2: Data as input for excel

Figure 3: lag distance specified

Then difference between points separated at a distance of 5 is

found. Then condition is set first of all for the first lag distance
that is 5, such that the difference between first and second
distance should be equal to 5.The formula set for this in the
formula bar is as follow.

=IF(ABS(B2-B3)=5,(C2-C3)^2,IF(ABS(B2-B4)=5,(C2-C4)^2,””))

Then answer to this as we get is then extended down to the last one as in

15/01/2023 Page 6
the figure.

Figure 3: value for h=5 for each point as extended

down
Then we copy the above formula from the formula bar and paste it into
very next column for the distance h=10. The changes then made in the
formula such that difference between two columns in the distance should
be equal to 10. The formula then gets the following shape.

=IF(ABS(B2-B4)=10,(C2-C4)^2,IF(ABS(B2-B3)=10,(C2-
C3)^2,""))

Figure 4: value for h=10 for each as extended

down
15/01/2023 Page 7
The result we get by the above formula then extended down to till the
last one.

 We put conditions because some of the consecutive points are

not separated by the distance 5 while their separating distance can
be different between the two executive points.
 Similarly, for h=15 the previous formula is copied and pasted for
it by making the changes in it such that the difference should be
equal to the 15. The formula gets the following shape.
=IF(ABS(B2-B5)=15,(C2-C5)^2,IF(ABS(B2-B4)=15,(C2-
C4)^2,IF(ABS(B2-B3)=15,(C2- C3)^2,"")))

Figure 5: value for h=15 for each as extended down

The result we get by the above formula then extended down to till the
last one.
 Similarly, for h=20, the previous formula is copied and pasted for
it by making the changes in it such that the difference should be
equal to the 20 and also an extra condition applied at the star that
the difference between the first and the sixth one is equal to 20.
The formula gets the following shape.

15/01/2023 Page 8
 =IF(ABS(B2-B6)=20,(C2-C6)^2,IF(ABS(B2-B5)=20,(C2-
C5)^2,IF(ABS(B2-B4)=20,(C2- C4)^2,IF(ABS(B2-
B3)=20,(C2-C3)^2,""))))

Figure 6: value for h=20 for each as extended down

 Similarly, for h=25, the previous formula is copied and pasted for
it by making the changes in it such that the difference should be
equal to the 25 and also an extra condition applied at the star that
the difference between the first and the seventh one is equal to
25. The formula gets the following shape.

=IF(ABS(B2-B7)=25,(C2-C7)^2,IF(ABS(B2-B6)=25,(C2-
C6)^2,IF(ABS(B2-B5)=25,(C2-C5)^2,IF(ABS(B2-B4)=25,(C2-
C4)^2,IF(ABS(B2-B3)=25,(C2-C3)^2,"")))))

15/01/2023 Page 9
 Figure 7: value for h=25 for each as extended down

 Same procedure is repeated for the h=30

 Now also the average of the all these values for each value of h is
found below each column. Similarly the semi variance is also
found for each column for each value of h. which is as follows.
Semi variance is found as the half of the average.

Figure 8: average and semi variance are found

15/01/2023 Page 10
 We are not further going to take the 30m apart samples because
the semi variance at 30 is almost got the maximum value and now
becoming the constant.
 After this a table we make as in the following figure spatially
pasting the semi variance values and semi variance values are not
defined by the experimental variogram at h=0 as will be
described by the model.
Then a graph is made between the lag distance as shown following.

 So, one part is covered that the experimental variogram is found.

Then from the structure of the experimental variogram we decide
which of the authorized model should be used in this case. In the
present case, the structure is much similar to the spherical model
so we continue with the spherical model. As the shape of the
experimental variogram is similar to the spherical model by
15/01/2023 Page 11
 shape.

 So the formula for the spherical model is kept in the excel as

reference so that the condition for the model in the excel can be
taken easily. Then the model is described in the excel
 Where the range, sill and nugget are taken from the experimental
variogram.
 The conditions for the model are put regarding the spherical
model. The formula for the model condition at h=0 is gets the
following shape.
=IF(M3=0,($S$4),IF(M3>=$S$2,($S$3),IF(M3>0,($S$4)+($S$3-
$S$4)*(3/2*(M3/$S$2)-1/2*(M3/$S$2)^3),IF(M3>=($S$2),($S$3)))))
Similarly, at every h, the formula for the model is same except
the change in the column.
 Then modeled variogram is introduced into the experimental
variogram for which the x values taken are lag distances and y
values are taken the model values.

15/01/2023 Page 12
Then a line which is solid is added to the points for modeled variogram.
As the points here for experimental variogram and the model are not
perfectly fitting each other . so a little change in the range will make the
model meet with the experimental variogram perfectly

6.5.RESULT:
This lab we studied about how to construct an experimental
variogram and then how to fit a model to it. We drew the variogram
as shoen in the figure.

15/01/2023 Page 13
 THANK YOU

15/01/2023 Page 14