engleski

Deterministic landslide susceptibility analyses using LS Rapid software

One of the key objectives of the Japanese- Croatian Project ‘Risk Identification and Land- Use Planning for Disaster Mitigation of Landslides and Floods in Croatia’ within the Working Group 3 (WG3) is landslide susceptibility zonation and hazard mapping of the investigation areas. Using deterministic approach in landslide hazard and risk analysis includes gathering of fundamental data about geometry, soil strength parameters, cover thickness and groundwater level, as well as the application of numerical models in safety factor calculation. Such an approach is often used in geotechnical engineering in slope stability analysis for common engineering problems. Most common is the one the dimensional deterministic model, and one of the mostly used methods inside its framework is stability analysis of infinite slope. Accuracy and reliability of safety factor calculation in each model is defined by the level of field investigations at the location, accuracy of the input data and understanding of the stress history. Detailed study of the whole area can give relatively accurate input parameters for the model, but is hard and demanding work. In some cases, heterogeneous areas are simplified and the gained input parameters are extrapolated to a wider area than the investigated location. These are basic limitations of the deterministic model. Using Geographic Information System (GIS) can ease deterministic as well as probabilistic geotechnical approach, as part of landslide hazard assessment methodology. However, high accuracy in data preview inside GIS cannot replace high impropriety in the assessment of the failure probability which is result of the wrong selection of the geotechnical model or inappropriate failure mechanism. Two dimensional and three dimensional deterministic analyses currently cannot be performed inside GIS, yet the data are being exported in some external software for 2D and 3D slope stability calculation and then restored in GIS as the results preview tool. That is applicable on a very small investigation area with huge amount of available information, where data conversion between software presents the highest challenge. Landslide susceptibility is function of the present slope stability (expressed through the safety factor), together with the existence and activity of triggering factors which cause increasing of active forces or decreasing of strength and, consequently, landslide activation. LS-Rapid software uses three dimensional models for simulation of progressive failure phenomena, developed to assess the sliding initiation and activation of landslides triggered by earthquake, rainfall or their combination. LS-Rapid aims to combine the process of landslide initiation and process of sliding mass movement (dynamic analysis), including the process of the sliding mass volume enlargement on the sliding path. This model is based on the key parameter- shear resistance in the steady state which can be measured or established by investigation. The software offers the recommended values for the input parameters in case when no detail data of the investigation area are available. The basic concept of simulation is based on the analysis of the forces acting on a vertical column inside the moving mass. Increase in pore pressure during the rainfall affects the development of the sliding surface and initiation of the landslide. Detail distribution of pore pressures or the groundwater level inside the slope is usually not known, so the influence of the pore pressure is taken into account trough the pore pressure ratio ru, which gradually increases until the failure appearance in a certain part of the slope. If this approach is applied on the wider area, in which is possible to define the relative position of sliding surface, it would be possible to obtain the values of the critical pore pressure ratio that causes conditions in which failures occur in a specific parts of the investigation area. Connecting this critical pore pressure ratio with distribution of rainfall through the run out coefficient and the infiltration, it is possible to obtain the landslide susceptibility and landslide hazard. In similar way it is possible to use LS-Rapid in seismic analyses to establish landslide susceptibility and landslide hazard caused by earthquake motions. Results of the performed analyses have shown that LS-Rapid can be used a powerful tool in deterministic analyses in landslide prediction, susceptibility and landslide hazard assessment. Deterministic model performed in LS-Rapid provided the critical area for landslide appearance which corresponds to the real conditions in the slope. The validation of the model was carried out by interpretation of stereopairs and engineering geological mapping. It was concluded that landslides inside the zones that in model were characterized as highly susceptible, occurred in the nearest of farthest past. In other words, it can be concluded that the results of performed analyses confirmed that LS-Rapid, except for the detailed analysis of slope stability at one specific location, can also be used for the landslide hazard assessment on wider area. Another conclusion is that the performed analyses using LS-Rapid are more accurate and give more reliable results than alternative methods used in deterministic landslide hazard analyses.

landslide susceptibility zonation; hazard mapping; deterministic approach; GIS; LS-Rapid

nije evidentirano