The large Ancona landslide studies revisited including stochastical spatial variability of mechanical parameters.

The great landslide that occurred in Ancona in 1982 caused damage to human livelihoods, buildings, railways, and roads. Many studies have been already carried out from different perspectives. The most common approaches in this field have used deterministic methods in relation to mechanical parameters. However, it is well known that local heterogeneities may have a non vanishing influence on landslides occurrence. Thus, in this paper, we explored the effects of an assumed stochastical spatial variability of the principal parameters, based on experimental measures, on the spatial hazard distribution of selected areas around Ancona, by means of a Probabilistic Approach. Then several trials have been carried out, applying the three dimensional code FLAC-3D, each time considering a different set of some selected mechanical parameters (i.e. density, friction angle, Young Modulus, cohesion), determined by random number generation routines among the “ensemble” of all the possible “realizations” of the selected statistic (Gaussian) to which the set of the numerical values of each parameters have been supposed to belong. Furthermore a “random variability algorithm”, developed by some of the Authors, has been implemented. In particular kinematic evolutions of weakness, characterizing the territory of Ancona, determined by stochastical approaches, have been compared with numerical results carried out by deterministic modelling.