The stochastic, or probabilistic, nature of the modeling process provides us with several models. Currently, the geologist interprets these models, selecting those that are most likely. For each of these models, flow analysis can be conducted to determine the overall flow regimes within the aquifer.

Geologically, future work will concentrate on usage of paleocurrent, lithofacies, and geometric data to gain better controls on the 3D geometry and linear stacked behavior of elements. For example, paleocurrent and geometric data will be used for better calculation of sinuosity of channel elements. Also, subsurface work is being conducted to constrain the relative abundance of various elements and to further analyze element geometry, size and behavior.

There is, of course, refining work that needs to be conducted with the mathematical model. The sophistication of the probability generators will increase as more model runs are performed. The initialization step will also improve with the inclusion of subsurface (well log) data. Also, the stacking of consecutive layers is still being developed, and this will provide three-dimensional models that can be input into existing postprocessing programs.

There are also two significant improvements that are currently being developed:

- Channels tend to run from border to border, but are currently treated as an element just like all of the others. We are now treating the channels first, in the initialization phase, separately from the other elements. They will be allowed to grow border to border, then the other elements will grow and be introduced stochastically in the generation phase of the modelling.
- Rather than relying on the geologist to determine whether a model is reasonable, we will let the model determine this. After several models are generated, statistics will be computed on them and compared to the field statistics. The initialization ("p") probabilities will then be adjusted in an attempt to conform the model to the field data. For the non-channel elements, the introduction ("r") probabilities are also adjusted since they are closely related to the "p" probabilities. The result will still be a set of plausible models that can be put into a flow analysis package, but the advantage will be that these models will match the field data without the interpretation of the geologist, giving additional relevance to the modeling procedure.

Previous Page--Field observations and discussion--Terra Cotta Clay Member at Kanapolis Lake ||
Next Page--References

Dakota Home ||
Start of the Stochastic Modeling Report

Kansas Geological Survey, Dakota Aquifer Program

Updated July 5, 1996.

Scientific comments to P. Allen Macfarlane

Web comments to webadmin@kgs.ku.edu

The URL for this page is HTTP://www.kgs.ku.edu/Dakota/vol1/geo/stoch10.htm