Specifications

University of Pretoria etd – Combrinck, M (2006)

Step 2: Find the first following data point not conforming to S-layer late time behaviour.

Apply the same conditions as in step 1, but look for the first point NOT satisfying these.

The last point passed through the filter is then

(d[i], S[i]).

The value of “1” chosen in conditions (b) and (c) is chosen based on empirical analysis of

the case history 1 data set consisting of 1500 soundings over both resistive and conductive

terrain. It is chosen to rather pass values that may be too high (resulting in excessively high

conductivities) because these can always be filtered out at a later stage in the conductivity

grids. Following this approach means that conductors will not be filtered out, and at worst

will be assigned exaggerated conductivities. It has already been shown in the previous

section that conductivities are normally underestimated by the algorithm and that the actual

numeric values of conductivity should not be taken as the absolute truth especially for

confined conductors. The value of this algorithm is in the identification and location of

subsurface conductors and is not promoted as a complete and final interpretation tool.

4.5.5 Imaged conductivity depth sections generated from synthetic data

The final product of the algorithm as described in this chapter is illustrated in Figures 4-

61and 4-62 for six synthetic models. In Figure 4-61 it is important to note the migration

effects similar to seismic data for vertical plates and low contrast conductors. It is

produced for the same reason of applying one-dimensional assumptions to three-

dimensional data; i.e. plotting all data below the point where it was measured. Migration

routines similar to those found in seismic processing can be developed to correct for this

but falls outside the scope of this study. High conductivity contrasts of the horizontal plate

and prism result in data not conforming to the S-layer differential transform assumptions

and are filtered out – resulting in white areas underneath these conductors. It is clear from

the low contrast data as well that the lower boundaries of confined conductors are not

resolved by this algorithm and that the apparent loss of data underneath conductors is in

fact a useful tool indicating high conductivity contrasts.