**Authors:** Filippo Catani, Samuele Segoni, Giacomo Falorni

**Abstract:** The importance of soil thickness, intended here as depth to bedrock (DTB), is well known as a factor conditioning a number of surface and subsurface processes (e.g. landscape evolution, landsliding, soil conservation, soil moisture). This notwithstanding, it is still one of the least understood and difficult to measure physical variables of the entire hillslope system, especially when considered at catchment scale. In areas dominated by active geomorphologic processes such as soil erosion and shallow landsliding, this uncertainty severely constrains the reliability of models concerning sediment production and transfer. Within a river basin soil thickness z can vary as a function of many different and often interplaying parameters among which we can count vegetation cover, lithology, climate, gradient, hillslope curvature, upslope contributing area and land use. Generally, for the interpolation of soil thicknesses between measured values a simple linear relationship with gradient is used, where z is considered inversely proportional to slope gradient at a point. We present here an alternative methodology which links soil thickness to gradient, horizontal and vertical slope curvature and relative position within the hillslope profile. This last parameter is fundamental: points having equal gradient and curvature can have very different soil thicknesses due to their different position on the hillslope. The interpolation model is implemented in a GIS environment and was tested in the Terzona river basin (Italy). Result validation indicates good agreement with field data and average errors are significantly lower than those obtained from other topography-based methods. The use of predicted soil depth to determine derived quantities such as the factor of safety for landsliding potential also provides promising results, confirming that this method can greatly enhance the prediction of soil losses, sediment production, shallow landslide hazard, etc. when utilized in conjunction with distributed hydrological and geomorphological models.

**Authors:** Giulio Barbieri, Stefano Barbieri, Paolo Cambuli

**Abstract:** In this paper a methodology using neural networks (NN) for landslide hazard mapping is presented. The zone studied is located in the central-east part of Sardinia, Italy. It is characterized by several shallow and rotational landslide phenomena. Neural networks represents a innovative methodology for landslide susceptibility zonation. Neural networks are analytic techniques modelled after learning processes similar to the neurological functions of the brain. A NN can be trained to identify certain patterns (e.g. landslides) and thus to predict the result for other data input. Geomorphologic information as slope and aspects are calculated from Digital Elevation Models provided by Sardinian Regional Administration. Other parameters concerning lithology, land use and landslide location were evaluated from field surveys and from the literature. These data were used for neural network learning. The comparison between the results obtained with the proposed methodology and those derived by heuristic techniques of landslide parameter overlay mapping shows a fair agreement.

**Authors:** Federico Preti

**Abstract:** Root reinforcement models are used to analyze published data from laboratory and in-situ shear tests and pullout tests on soils reinforced with synthetic materials and root systems. Analysis of stability was conducted referring to shallow landslides which occurred in Italy. Various modelling approaches are compared and the problems involved in gathering necessary data are taken into account. Understanding the failure mechanism allows one to identify appropriate application of the models to stability analysis. Adopted models can be useful in understanding soil-root behaviour and interpreting test results. This allows one to identify their requirements and limitations. In most cases, the tensile force may be well below the ultimate tension. Root reinforcement is often treated as an additional apparent cohesion at a potential slip surface and included in the resistance term of the calculation of the factor of safety; it is assumed that a slope will fail if the total available resistance has been mobilised to counteract the driving forces along any potential slip plane passing through the soil mass. In that case, the ratio of the maximum available shearing resistance over that required to prevent failure, the basic definition of the factor of safety F for slope stability problems, is at unity (F= 1 and F> 1 if only part of the available resistance needs to be mobilised). The simple models can give approximate results if the tensile force can be evaluated. Overestimation of root cohesion could be corrected by a 0.4 factor. The analyses demonstrated the importance of root geometry, site conditions, and the nature of root displacement, which control the failure mechanism.

**Authors:** Michele Calvello, Giuseppe Sorbino, Leonardo Cascini

**Abstract:** A numerical model to predict landslide movements along pre-existing slip surfaces from rainfall data is presented. The model comprises: a transient seepage finite element analysis to compute the variations of pore water pressures due to rainfall; a limit equilibrium stability analysis to compute the factors of safety along the slip surface associated with varying groundwater pressure conditions; an empirical relationship between the factor of safety and the rate of displacement of the slide along the slip surface; an optimization algorithm for the calibration of analyses and relationships based on available monitoring data. The model is applied to a well-monitored active slide in Central Italy, characterized by very slow movements occurring within a narrow band of weathered bedrock overlaid by a clayey silt colluvial cover.

**Authors:** Alessandro Guerricchio, Valeria Biamonte, Maurizio Ponte, Alessandro Guerricchio, Rosario De Salvo

**Abstract:** A high precision GPS survey has been carried out since 1999 in the area of San Lucido (Calabria Region - Southern ltaly), to evaluate the movements of deep seated gravitational deformations and huge landslides, connected with a left transcurrent fault and reactivated by the 1783 Calabrie earthquake. Even if far about 100 km from the epicentral area of that earthquake, the whole slope of the San Lucido territory was deeply damaged by the reactivation of large landslides. Deep ruptures, produced by this event, involved the Tertiary and Quaternary formations from an elevation of about 800m asl, over a Iength of 5km and an average breadth of 2-3 km. Currently it is believed that slope movement is not exhausted. Indeed it causes strong marine erosion together with stability problems in important infrastructure facilities, as weII as in urbanized areas. Slow subsidence is occurring along the coast owing to the continuation, although sporadic, of activity in the huge ancient landslide. The GPS monitoring data, relative to three measurements campaigns, have allowed to determine the displacement rates in a period of about three years. The surveys are still going on. In particular, it has been possible to measure, within the topographic network, average movements of 3 mm in the baseline between the ancient monastery of S. Maria di Monte Persano and the ruined castle in the San Lucido town, and of 6 mm in the baseline which connects the above mentioned monastery to a point situated in a well in Libertino locality. Deformations with estimeted velocities of 2,105 x 10-8 cm/s have been observed. The monitoring data agree with the displacements of the deformations ascertained by means of extensometer measurements in some recent buildings and in some bridges of the old Paola-Cosenza railway line.

**Authors:** Michael Sheridan, D. Kumar, Bruce Pitman

**Abstract:** Debris flows as opposed to debris avalanches are characterized by the presence of intergranular fluid. Analysis and mitigation of hazards arising from such flows requires high quality modeling tools. Modeling the complex interaction between solid and fluid phases in such flows is difficult and a number of models have been presented in recent years including most recently by Pitman and Le [Phil. Trans. Royal Society A, vol. 363, July 2005 p 1573-1602]. In contrast to mixture theory models, the two-fluid model separately solves for granular and fluid phase mass and momenta with the interaction between phases represented by a drag force-like term. In this paper we present application of the above two fluid model using hyperbolic solution methodology for such debris flows, and apply it to simulating flows over natural terrain at several sites of interest where lahars are a potent hazard using our newly developed TITAN-DF toolkit. The applications include flows Ruapehu volcano, New Zealand and Tungurahua volcano in Ecuador. Adaptive parallel solution methodology and integration with a geographic information system, make it feasible to do realistic simulations on digital representations of natural terrains. The TITAN-DF toolkit contains modules for necessary pre and post-processing. All our developed software is available freely in open source fashion.

**Authors:** Bruno Merz, Heiko Apel, Annegret Thieken

**Abstract:** Flood hazard mapping and flood risk analysis are based on a number of flood scenarios. Each scenario is associated with a certain return period T, e.g. 100, 200, 500 years. Traditionally, a static approach is used to derive these scenarios. Using observed data and flood frequency analysis, flood hazard curves, relating discharge and return periods, are calculated for streamflow gauge locations. By means of a regionalization scheme, T-year water levels and T-year discharges are assigned to ungauged locations within the watershed. These water levels are horizontally extended across the flood plains to obtain the T-year inundation area. Combining the flooded area with flood damage estimation yields the spatial distribution of the flood damage which is attributed to the respective return period. An alternative, dynamic approach is the use of dynamic flood simulation models within a probabilistic framework. Process models that represent the hydrological and hydraulic flood processes in the catchment are coupled in a Monte Carlo simulation. In this way, process understanding can be included into the derivation of the flood hazard scenarios. The contribution compares the potential and limits of the static and the dynamic approaches. Considering as example the river Rhine, it is shown that the static approach may lead to overestimation of the flood risk, since it is not able to represent extreme effects, such as river levee breaches, that have not been observed in the usually short observation period. Furthermore, the static approach may fail to represent the spatial heterogeneity of flood events in the catchment. Therefore, the static approach should be restricted to certain purposes whereas the dynamic approach has a much wider application.