Geology has lacked an effective experimental approach to many problems. Although some kinds of experiments can be carried out, such as sediment transport on a stream table, most geologic processes cannot be re-created in the laboratory. A mountain range can be observed in the field, but neither the mountain range nor the mountain-building processes can be duplicated in the laboratory. These are shortcomings of which geologists are fully aware.

There is, however, another shortcoming in the geologist's kit of tools--namely, his difficulty in dealing with many interdependent variables or processes that mutually and simultaneously affect each other. However, computers can be used, in conjunction with appropriate mathematical geologic models, to perform kinds of experiments which otherwise could be performed only with difficulty, or not at all. Furthermore, these computer models are ideally suited for "exploring" the effects of interdependent geologic processes that are linked together to form dynamic systems.

The processes of marine sedimentation are highly interdependent. All too commonly, geologists who study stratigraphic sequences attempt to interpret the depositional environments of the strata in terms of only one or two variables--such as water depth or wave energy. Clearly, methods of analyzing the effects of multiple interdependent geologic variables are needed.

Computerized, dynamic mathematical models which represent geologic processes and products in two- or three-dimensional space provide one method of attack. These models are philosophically identical to dynamic oil-reservoir simulation models, and they also share many of their constructional details. Most dynamic geologic simulation models can be constructed from relatively few mathematical "building blocks," which may be grouped into the following categories: (1) materials-balance accounting systems; (2) sources of random variables; (3) Markov chains; (4) flow and transport mechanisms using finite-difference methods of representing diffusion and fluid-flow processes; (5) feedback control; (6) optimization; and (7) graphic display. Virtually all models employ methods in which both space and time are compartmented into small increments.

Experiments with relatively simple sedimentation models suggest that feed-back control exerts much influence on sedimentary sequences. A lag in the isostatic adjustment of the crust in response to the load imposed by deposition of sediment may be responsible for some of the large-scale rhythms in cyclic sedimentation. Simulation models can be used effectively to explore the results of different assumed lag factors in sedimentation/isostatic-adjustment feedback loops.

End_of_Article - Last_Page 555------------