Page 65 - MSc_thesis_R A Kil

6.1.3 Additional data

Despite the non-regional approach of the model building process (see the previous section), trends of
facies distribution in outcrops can be recognised in different areas of the island. These data are therefore
converted into probability maps that can be used as additional input for the model. The tutorial in
appendix E provides further details on the process of making such probability maps in the Petrel software
package.

Facies occurrence and their relative proportions are linked to the outcrop thickness and represented
as pie charts (figure 4.11). The same outcrop thickness is also used to generate probability maps. A
convergent interpolation method converted data points at the pseudo-well locations to probability maps
of a certain facies. The downside of this approach is that no conceptual thinking can be included in the
maps. Therefore the probability maps generated directly from outcrop data are only used as a starting
point. The interpolation method assumes that when no data are present, a certain facies is not present at
that location, resulting in some unusual patterns.

A solution is found by manually creating the probability maps using both the conceptual thinking as
well as the maps generated from outcrop data. This method implies the manual generation of polygons
representing lines of equal probability. In this way, areas without data or parts of the island that are not
quarried to a deep level can be filled in using geological knowledge. The contour lines are subsequently
converted to surfaces and in turn to properties of the static model grid.

Finally, the property volumes presented in figures D.3 and D.4 are included in the MPS facies modelling
algorithm. Probabilities are normalised by Petrel and used as a medium-influence soft data input, which
is the degree of influence recommended by the software. Once the facies model with all the non-scouring
features is created, the scours have to be modelled. This is done with an object modelling based algorithm.
The locations of scours are determined by 1) the well logs; and 2) the probability volumes. In particular
no information is present about the length of the scours. The few examples observed in the field only
showed a cross-section rather than the full length of the scour. Thickness — or depth of the original
scour — can be up to 20 metres, with a width that more or less equalises this depth. Scour-fill facies
only replace the large foresets and undulated beds, because those facies occur in the top layers where the
scours are believed to be occurring.

6.1.4 Facies modelling results

The modelling process of outcrop areas require some trial and error. The algorithm will always honour
the pseudo-well logs with the facies information, but in the inter-well area often more information is
known by the geologist, especially in this case study where parts of the island with high outcrop density
— like Cala Rossa, Bue Marino and Cala Azzurra — show a lot of available information. A quality check

47