Specifications

University of Pretoria etd – Combrinck, M (2006)
100 (Figure 5-3) contains four data points at the end that are filtered out. No data are
discarded by the SDTC filter and therefore the first two options yield identical results.
The advantages of filtering out the last four points (17-20) are:
a more realistic maximum depth of investigation (706 m instead of 3943 m)
data points 15 and 16 still indicate high conductance values in the third case while in the
first two cases they have been “contaminated” during smoothing by the consecutive
noisy points to give lower than expected conductance values.
Similar results are observed for stations 400, 450 and -1550.
From these examples it is seen that the SDTC filter on its own also serves to discard
some of the last erratic data points although it was not originally designed with this in
mind. However, it is not as effective as removing the noisy points from the raw input
data.
A final comparison of the second and third filter options are done by means of
contoured imaged conductivity sections of line 4950.
(Note: This line has been chosen
specifically because it contains stations with clean as well as noisy data.)
These sections are shown in
Figure 5-8. Only imaged conductivities at depths of less than 800 m below the surface
were contoured; deeper than that the data plot as individual, incoherent points and is not
suitable for contouring. There is not much to choose between the two approaches when
presented in this way. The only significant difference occurs between stations 400 to 600.
Here the application of only the SDTC filter results in fairly high negative values (dark
blue). The same feature is visible on the bottom section (SDTC plus noise filter) but it is
less pronounced. Taking into account the fact that there are only three points on station
500, it is probably due to noisy data (see station 450, Figure 5-6) and the smoother,
double filtered bottom section is preferable to the top one.
109