ABSTRACT
This chapter proposes a practical methodology for modelling geological complexity of natural formations using soft data. The approach presented is an extension of the Markov chain theory used by W. C. Krumbein to synthesize a stratigraphie sequence. The sedimentary nature of geological deposits in a fluvial environment exhibits more variability in the vertical direction than in the lateral direction. Many geological processes display a Markovian property. A Markov chain is completely described when the state space, transition probabilities and initial probabilities are given. The core of stochastic modelling in geology is to get a plausible description of the subsurface with limited amount of data. The horizontal transition probability matrix can be estimated from geological maps describing the formation extensions in horizontal plane. These maps may be obtained from geological surveying. Information about the transition probabilities may be provided directly from geological experience, geologically similar formations but better known or analogous outcrops.
