|
|
|
GIS approach to statistical modelling for mineral deposits in the Singhbhum copper belt, Bihar, India, using geological and geophysical parameters
Calculation of Probability Index:
In a region where there are
known mineral occurrences with gap areas in-between, the potential of the whole
region including the gap areas for a particular mineral, say copper, can be
statistically calculated. Instead of calculating one value for the whole region,
the area was divided into a number of smaller areas at regular spacing and the
potential or probability of, say copper, which occur in such smaller areas can
be calculated by giving it a value called Probability Index. The Probability
Index will be 100 over the producing mines and will vary from 0 to 100 over the
gap areas. Where the Probability Index will be high in the gap areas, it will
indicate possibility of finding mineral occurrence. These values were be
contoured to give probability contours.
The geological map of the study
area was considered for transforming the above 36 variables into binary form.
With the help of AutoCad software a grid of 1 km x 1 km was made on a tracing
paper and the tracing paper was overlain on the geological map. For each cell,
over the area, the variables present in that cell were noted. The reserve of
copper in the different deposits and their grade was found out from published
material. Then the total metal accumulation was calculated for each of these
areas with help of the formula Grade/100 x Reserve, 15 locations were considered
where there is a producing mine and where metal accumulation could be
calculated. The cells in which these locations fell were called the control
cells. Normally it was assumed that one location fell in one cell. The adjacent
cells where it was assumed the maximum exploration has taken place were called
the control area. Out of the 36 variables those falling in each of the control
cells were noted and if present was denoted by 1 and if absent was denoted by 0.
So a matrix was prepaired, Table-8, with the 1st. column being the 15 locations,
2nd. column the metal accumulation in these locations and the 1st row being the
36 variables. A further data reduction was done by discarding those variables
which were 0 in all the 15 locations i.e. the variables were absent in these
locations. So the number of variables considered came down to 26.
Once
this matrix was generated stepwise regression was carried out to obtain a
relation between the metal accumulation and the variables.
Multiple
Regression Analysis:
This is done to calculate the metal accumulation in
the unknown cells where the variables are known. Muliple regression analysis
relates the independent variables i.e. the variables that are being considered
to the dependent variable that is the metal accumulation by giving an equation
relating the two.
The equation was applied to each cell over the area
under study and the metal accumulation value in each cell was calculated. This
calculation was done using a spreadsheet software. Program was written to
automate the calculation. The values in each cell were multiplied by a factor of
100 and rounded.
Within the study area the total number of control cells
is 16, and the total number of cells in the control area is 90. In order to
assign values to contours the following procedure was adopted :-
- The sum of all calculated values in the control area was taken;
- This sum was divided by the number of control cells i.e. 16. This gives the
scaling factor of 68;
- The calculated values for all cells in the area was divided by the scaling
factor i.e. 68.
- The values for overlapping blocks of 4 cells were added.
- The results were contoured using a contouring software, SURFER.
The contour so generated was imported to another software, AutoCad
where it was overlain on the 1 km x 1 km grid after bringing both of them to the
same size.
From a large number of variables considered, Characteristic
analysis helped in reducing the variables and selecting the important ones. The
Multiple Regression Analysis helped in further reducing the variables and
ensuring that only those variables were used which were important for the
analysis. This technique helped in giving an equation relating the metal
accumulation to the variables and this aided in predicting the metal
accumulation in unknown areas.
The average amount of copper per control
cell is calculated to be 96,128.13 tons. The area of the unit cell is 2 km x 2
km or 4 sq km. Hence, the probable tonnage of copper per square km (K) amounts
to (96,128.13 x m) / 4 = 24,032 m.
Hence, from the contoured map (Fig.3
and Fig.4) it can be said that where the probability index contour is 0.9, the
probable tonnage of copper in the surrounding 2 km x 2 km cell is 0.9 x 96,
128.13=86,515.317 tons or say 86,500 tons. From the given table we know that the
estimated value 0.9 is reasonably precise, i.e., at least 1,2 or 3 cells will
have 86,500 tons of copper with a probability of 64%.
|
|
|