Page 61 - Geochemical Anomaly and Mineral Prospectivity Mapping in GIS

60 Chapter 3

whereas anomalies are represented by large crosses. Because extremely low background
values and anomalies are usually fewer, if not absent, compared to other data values in
an exploration uni-element geochemical data set, the large symbols for the former data
values will not dominate a map. Low background values and high background values are
represented by small circles and small crosses, respectively. Background values are each
marked with the smallest symbol – a dot – because they are expected to dominate the
data and its map.
Fig. 3-6A shows a map of spatial distribution of Fe contents in soils based on boxplot
classes defined from the whole data set (Fig. 3-3). The spatial distributions of the Fe
data, based on the boxplot classes of the whole data set, can be explained readily by
variations in lithology. Fig. 1-1A shows very similar distribution of the Fe data, although
the classes were defined based on a-priori knowledge that variations in the Fe data are
influenced strongly by one of the lithologic units. Thus, geochemical data classification
based on a boxplot and the EDA-mapping symbols has strong ability to portray
physically meaningful spatial distributions of uni-element data without assumption of the
normal distribution model or a-priori information about certain factors that influence
variability in a geochemical data set.
Further exploratory analysis of subsets of a uni-element geochemical data set
according to certain criteria could provide further insight into processes that plausibly
influence variations in the data set. For example, based on subsets of the Fe data
according to rock type at sample sites (Fig. 3-5B) and after standardisation according to
equation (3.10), soils on phyllite immediately around the basalt unit are shown to be high
background in Fe compared to soils on phyllite farther away from the basalt (Fig. 3-6B).
The plausible explanation could be contamination of soils on phyllite by soils derived
from the basalt. In contrast, soils at the outer portions of the basalt unit are shown to be
low background in Fe compared to soils at the inner portions of the basalt unit. The
plausible explanation could be contamination of soil on basalt by soils derived from the
phyllite. Thus, uni-element geochemical maps based on resistant classes defined by a
boxplot of a whole data set or data subsets are potentially useful in interpretation of
processes that control variations in the geochemical landscape.
Instead of using different EDA point-symbols, boxplot classes can be represented
with the same and equal-sized point-symbols but with different shades of grey (Fig. 3-4)
or different colours (e.g., Reimann, 2005). Grey-scale or colour-scale representations are
appropriate for interpolated uni-element geochemical data, although the classes must be
defined from a boxplot of original data at sampling points. The symbols or colours used
to represent classes of an exploration uni-element geochemical data set, as defined
through a boxplot, serve to objectively portray in a map the structure and spatial
distribution of that data set with a balanced aesthetic (visual) impression.