Asadi, Hooshang 
  
4.2.1 Qualitative analysis 
Exploratory analysis, aimed at the delineation of mineralised zones, was carried out using statistical and graphical 
methods. Only the results obtained for gold are discussed in this paper, since the main objective is to try to obtain a 
direct indication of the quality and quantity of gold mineralisation present. The procedures used to analyse and interpret 
the soil geochemistry results for gold are summarised briefly below: 
a) Statistical summaries were used to identify data ranges, presence of any outliers, and number of valid analyses. The 
results obtained highlighted a few data entry errors that were corrected prior to subsequent analysis. Gold values 
showed a high skewness (10.25), and large differences between the arithmetic mean (842 ppb) and the median (33 
ppb), indicating the presence of a log normal distribution. 
b) Plotting of frequency histograms and cumulative probability plots. These confirmed the log normal nature of the 
distribution and indicated the presence of multiple populations. The cumulative frequency curve for gold (Figure 3, 
left hand graph) clearly indicates the presence of at least two distinct log-normal populations. The inflexion point at 
300 ppb was taken as a possible threshold value for separating the two populations. 
  
  
0.1 a 10 100 1000 10000 100000 0.1 1 10 100 1000 10000 100000 1000000 
  
  
  
Figure 3. Log-scaled cumulative probability plots for the gold distribution in soil (left), and in drillcore (right), samples. 
Inferred populations present are indicated by the diagonal lines (blue - low grade population common to both 
sample types; red - high grade population in soils; green - high grade population in drillcore samples) 
c) Classed symbol plots of the gold values. These were used to indicate the spatial distribution of the two populations 
identified on the cumulative frequency curve. A simple two class plot, based on the 300 ppb threshold value, 
indicated that the majority of the high grade population samples fell in a single area corresponding with the outcrop 
area of the Chaldagh and Zarshuran units at the Zarshuran deposit (see Figures 1 and 4). Additional symbol plots 
with a larger number of classes were used to confirm trends identified on the contour plots of gridded data. 
d) Gridding of data. Data was gridded using a quadrant search control, with an anisotropic search ellipse, in order to 
ensure that the unequal spacing, along and between sample lines, did not result in distortions of the results. Several 
different gridding methods were tested, in order to select the optimal approach for interpretation of geological 
related features in the data. Given the strongly skewed distribution of gold values, and the log-normal character of 
the two inferred populations, gridding of the natural logarithms of the gold values was eventually selected as giving 
the most readily interpreted results. Anisotropic inverse distance squared weighting was applied in order to take 
account of a preferred continuity parallel to the strike (NW-SE) of the units present. 
e) Contouring and imaging of results. Various displays were used to create the maps which formed the basis for 
interpretation of the geochemical data. Due to the log normal distribution, it was found preferable to contour data at 
regular intervals according to the log values. The colour shaded example shown in Figure 4 was used to interpret 
"ridges" and "troughs" that are inferred to reflect the influence of geological variations on the soil results. Black and 
white contour plots with more closely spaced contours were also used to highlight sharp discontinuities that may 
  
86 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.