Simulating landslides of different complexity with hexagonal cellular automata.

The paradigms of acentrism and parallelism that are embedded in the definition of Cellular Automata (CA) can be easily and efficiently applied in the simulation of very complex natural processes like landslides. This permits a phenomenological description that is able to overcome resource computational limits, placed in a classical approach and therefore based on the resolution of differential equation systems. We present a general frame and the latest developments of a CA based model for the simulation of debris flow type landslides. Landslides are here viewed as a dynamical system that is subdivided in parts which components evolve exclusively on the basis of "local interactions" in a spatial and temporal discretum, where space is represented by hexagonal cells, which specifications (namely substates) describe the average physical characteristics of the respective area. Such a method permits to start from simple landslides, that can be modeled using few substates and simple local interactions; a more complex landslide can be modeled from the previous model, adding substates and local interactions. SCIDDICA can be considered to exhibit great flexibility in modeling and simulating debris flows. Furthermore it can be applied in some field of intervention such as: The creation of risk maps also with a statistical approach; the simulation of the possible effects of intervention on flows for stream deviation, introducing data which represent alterations of the original conditions (e.g., the construction of a canal or embankment, occlusion of a mud canal etc.). Examples of practical applications on real events involved: the 1992 reactivation of the Tessina landslide in Italy, the 1984 Ontake volcano debris avalanche in Japan and a first partial application to the landslides occurred in the Sarno area of Campania Region (Italy) in the May of 1998.