A comparative review of methods for shallow landslides hazard assessment: critical analysis and proposals

Shallow landslides interest the slopes soil coverage. They always occur after high intensive and long rainfall. Their particular hazard is due to their rapid development, their high velocity and kinetic energy and their elevated spatial distribution making necessary their hazard evaluation. In the last years, several models for the prevision of shallow landslides were proposed. Initially, the more diffused methods of hazard analysis were based on the superimposition of different thematic layers, now re-proposed by Lee & Pradhan (2006). Some authors (Benedetti et al., 2006; Del Maschio et al. 2005; Giannechini, 2005) use the statistical approach, through the concept of pluviometric threshold introduced by Campbell (1975). These authors consider only the pluviometric conditions, the time distribution of the rainfall over a large period or the graphic relationship among the rainfall intensity, the rainfall duration and the shallow landslides. However, the greatest part of models (TRIGRS, Baum et al., 2002; GEOTOP-SF, Simoni et al., 2006; Dapporto et al., 2005; Dietrich & Montgomery, 1995) uses an analytical approach consisting in a coupled analysis of infiltration and stability slope. Some of these models investigate the effects of pore water pressures changes on slope instability. They consider the presence of a water table without a downwards water motion due to the infiltration. Others provide for some simplifications and assumptions like the reduction or annulment of the cohesion value. (Iverson, 2000; Casadei et al., 2004; Gulmares et al., 2003; SHALSTAB, Dietrich & Montgomery, 1995). The simplified model proposed in this paper takes into consideration both the rainfall infiltration in the soil and her influence on the slope stability. For the infiltration analysis the model uses the Green-Ampt method (1911) that considers the downwards advancement of a saturation front from the ground surface in consequence of a project rainfall. The thickness value of saturated soil (h) is used as input Information for the stability analysis according to the infinite slope equation: Sf = (f • γ • h • cosα • cosα + c)/( γ • h • cosα • cosα) where Sf is the Safety Factor, f is the friction coefficient (f = tan ),  is the friction angle, γ is the soil specific weight, h is the thickness value of saturated soil, α is the slope angle and c is the cohesion value. The utilization of this equation allows to connect the project rainfall with the landslides triggering, through the h value. The cohesion annulment wouldn't allow this connection. The object of the proposed approach is to realize a shallow landslides susceptibility map and, at last, a probabilistic map of the project rain for assessing of the hazard due to this phenomenon. In precedent works, the model was been carried out in the Barbera and Barolo viticultural terroirs (respectively, localized In the Monferrato Hills and in the Langhe Cuneesi Hills, in Piedmont). The economic weight of these territories, consequent to prestige of the wines that are produced in them, has led on to the assessment of shallow landslides hazard and risk. The study has allowed to evaluate the differences and the analogies between the various observed models and the proposed one, underlining that the latter, through the infiltration analysis, succeeds in connecting shallow landslides trigger with the rainfall, resulting a valid instrument for modeling and forecasting this type of phenomena. The friction angle value  and the cohesion value c are considered respectively 20° and 0.5 kN/m2 in conditions of total saturation. This values have been used in precedent works (Albanese et al, 2006; Albanese et al, in press 1; Albanese et al, in press 2) and will be calibrated on other areas.