A macroscopic collisional model for debris-flows simulation

  • Authors:

  • Donato D'Ambrosio;Giulio Iovine;William Spataro;Hideaki Miyamoto

  • Affiliations:

  • University of Calabria, Department of Mathematics & High Performance Computing Centre, Arcavacata, 87036 Rende (CS), Italy;CNR-IRPI & University of Calabria, Department of Earth Sciences, High Performance Computing Centre, via Cavour, 6-87030 Rende (CS), Italy;University of Calabria, Department of Mathematics & High Performance Computing Centre, Arcavacata, 87036 Rende (CS), Italy;Department of Geosystem Engineering, University of Tokyo, Tokyo 113-8656, Japan

  • Venue:
  • Environmental Modelling & Software
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

SCIDDICA S"4"c is the latest hexagonal release of a family of Cellular Automata models for the simulation of flow-type landslides. It is able to simulate the erosion of the regolith along the flow path, besides branching and re-joining events of the flow masses. Dissipative effects are modelled in terms of not-exclusive velocity-dependent mechanisms, which allow to simulate even complex rheological behaviours. Moreover, it is able to manage the peculiar characteristics of rapid flows, and the effects of mass collisions, by guaranteeing mass conservation. In case of no dissipation, conservation of energy and momentum are also assured. Model calibration has been carried out through parallel Genetic Algorithms, by considering the May 1998 Curti-Sarno (Campania, Southern Italy) debris flow. A preliminary analysis has also been performed, aiming at evaluating the sensitivity of the model with respect to a sub-set of model parameters, the size of the cell, the orientation of the cellular space, and noise in input data. Calibration confirmed the reliability of the model in reproducing the considered case of study. Moreover, sensitivity analyses pointed out its robustness with respect to the considered factors, by highlighting their different weight in affecting the behaviour of the simulations.