Cellular Automata are examples of Parallel Computing models (von Neumann, 1966; Burks, 1970; Lindenmayer, 1971; Wolfram, 1984). They are a valid alternative to differential equations in simulating complex natural phenomena (Toffoli, 1984)--and, particularly, debris flows (Johnson, 1970). In fact, the evolution of this type of phenomenon can be described in "a-centric" terms--i.e. exclusively, in terms of local interactions among the elementary portions into which the phenomenon can be subdivided (Petitot, 1977). In the present paper, the latest release (S <sub>2</sub> ) of the model SCIDDICA, specifically developed for simulating debris flows, is briefly discussed. On 5-6 May 1998, numerous rapid/extremely rapid debris flows were triggered by exceptional rainfalls in Campania, mostly on the slopes of the Pizzo d'Alvano massif (Brancaccio et alii, 1998; Del Prete et alii, 1998; Guadagno et alii, 1998; Calcaterra et alii, 1999). Hundreds of small debris slides originated in the volcaniclastic mantle, on the uppermost portions of the slopes, and turned to fast-moving debris flows. These phenomena generally involved the entire depth of volcanic soil cover, eroding it down to the bedrock, and increasing their initial volume by scraping off soil and vegetation along the path. Landslides hit the urbanised areas located at the base of the slopes, killing 161 people and leaving more than 1,000 others homeless. Carbonate rocks of Mesozoic age constitute the skeleton of the Campanian Apennines (D'Argenio et alii, 1973; Ippolito et alii, 1975; Bonardi et alii, 1988). Pyroclastic fall deposits of the Somma-Vesuvius volcanic complex (Arno et alii, 1987), locally altered and/or reworked, mantle the Pizzo d'Alvano massif. The thickness of the soil mantle ranges from a few centimetres to several metres, depending on slope, morphology and superficial erosion processes. The characteristics of the soil mantle (e.g. grain size, petrography, and thickness) vary, according to parent eruption and subsequent geomorphic history (De Gennaro et alii, 1998; Rolandi et alii, 1998). The Curti debris flow has been selected as "representative", among the whole population of hundreds of debris flows triggered in Campania in May 1998. Since its S <sub>1</sub> release, SCIDDICA has proved to satisfactorily simulate the path of the Curti debris flow (Avolio et alii, 1999b; D'Ambrosio et alii, in press). Moreover, new simulations performed through the S <sub>2</sub> release better predict the area affected by the landslide, once the geological and morphological settings are properly given to the model (as input matrixes). After a phase of thorough testing and parameter evaluation, to be carried out by considering a significant sample of real cases in the same study area, SCIDDICA could conveniently be applied to evaluate the most probable spatial evolution of future debris flows of the same type. Nevertheless, more "elementary" processes still need to be considered, and included into the transition function of the model (here simplified, in order to allow first simulations on a PC platform). On the other hand, simulations indicated that some of the processes considered could also be disregarded when modelling very rapid mudflows like the Sarno ones. Finally, oncoming testing of the model-to be carried out in "parallel environment" by applying genetic algorithms--will also allow a better definition of confidence intervals related to experimental parameters.

You do not currently have access to this article.