Integrated Geochemical and Geophysical Approach to Mineral Prospecting – A Case Study on the Basement Complex of Ilesa Area, Nigeria - Advances in Data, Methods, Models and Their Applications in Geoscience - page 75

As with all geochemical surveys, the first step in approaching an operational problem is to
conduct an orientation survey. Such a survey normally consists of a series of preliminary
experiments aimed at determining the existence and characteristics of anomalies associated
with mineralization. This information may then be used in selecting adequate prospecting
techniques and in determining the factors and criteria that have a bearing on interpretation
of the geochemical data.
Although the orientation study will provide the necessary technical information upon which
to base operational procedures, the final choice of methods to be used must also take into
account other factors, such as cost of operation, availability of personnel, and the market value
of the expected ore discoveries. The nature of the overburden, whether it is residual or is of
glacial, alluvial, or wind-borne origin, is the first question that must be answered by the
orientation survey. Sometimes it is surprisingly difficult to discriminate between residual and
transported soil. The safest method therefore, is to make critical and careful examination of
complete sections of the overburden at the start of every new field survey. If road-cut
exposures are not available, the soil should be examined by pitting or auguring. Previous
orientation studies carried out by Olorunfemi (1977) and Adewunmi (1984) in parts of
southern Ilesa established that the C horizon is the preferredhorizon for sampling. Details of
laboratory procedures and analysis of the samples were reported by Ariyibi et al. (2010)
5.1.1.1 Statistical results
The multivariate technique, has proven to be viable and credible when applied on
geochemical data as reported by Grunfeld (2003). The Principal Component Analysis (PCA)
which isa multivariate technique, describep observable random variables x 1 , x 2 , ….,x p in
terms of the joint variation of a fewer number, k (<p) variables. The purpose of PCA is to
determine factors (i. e. principal components) in order to explain as much of the total
variation in the data as possible with as few of these factors as possible. This will uncover
their qualitative and quantitative distinctions.
Table1 shows the descriptive statistics of the data. The data are not widely dispersed from
the average when values of standard deviation are compared with the raw data. The
measured values of Fe in the samples are quite large and so account for the large value of
standard deviation (5. 0903) as seen in the table. The covariance matrix is shown in Table 2
and the corresponding correlation coefficients are shown in Table 3 and this was used to
obtain the coefficients of the principal component using MATLAB as shown in Table 4 from
standardized variable.
Observation elements are on the rows of Table 4. For example, Pb is denoted by X 1 and Fe
by X 2 and so on. U 1 , ………,U 8 are the principal components. The loadings or coefficients of
the principal components are on the vertical columns. The magnitude of loadings greater
than or equal to 0. 5 is to be considered for interpretation (Dillon and Goldstein,1984) as this
will give the element with higher association ratio. On the last two rows of Table 4 is the
eigenvalues of covariance matrix of the data and the Hotelling's T 2 statistic which gives a
measure of the multivariate distance of each observation from the centre of the data set. A
plot of the variability (in %) and the Principal component is as shown in Figure 2.
The three principal components with elements having loadings of 0. 5 and above (for which
also the variability is greater than 10%) are: U 1 , U 2 and U 3 and these combined, account for
85. 34 % variability in the data. In U 1 are the elements : Fe and Mn in association (i. e. Fe-Mn)