The influence of topographic scale in mass wasting susceptibility modeling applied to a pipeline in southeastern Brazil

The purpose of mass wasting susceptibility assessment using a Geographic Information System is to assign, in a regional scale, places were these events are more probable to take place. Fernandes & Amaral (1996) consider that the main purpose of mass wasting susceptibility maps is to provide information about the probability of mass wasting occurrence, dividing the whole area in certain landscape in zones with equal susceptibility. Although these maps have not the purpose of calculate the stability of hillslopes, they should provide information on spatial and temporal probabilities and the magnitudes of the mass wasting in a landslide prone area. The scale used to assess landslide susceptibility in steep hillslopes is very important to better describe the role played by the conditioning factors. The way which the digital elevation model (DEM) is obtained by interpolation or by mesh has an enormous influence in determining the main topographic parameters of this surface, like the contributing area, the flow direction, the slope and the hillslope curvature. These parameters are important to calculate the safety factor (SF) of these hillslopes, which may be used as inputs to deterministic models or to susceptibility indexes hazard maps. Special attention must be given to hillslopes where human occupation has already taken place, including the construction of houses, roads and pipelines. Here we will present a variety of studies on mass wasting susceptibility modeling applied to a pipeline in steep hillslopes of Serra do Mar, close to the city of Rio de Janeiro, where mainly creeping process are occurring. In these numerical experiments we used both deterministic (e.g., SHALSTAB Shallow Landslide Stability Model and TRIGRS Transient Rainfall Infiltration Grid-Based Regional Slope Stability Analysis) and empirical models (e.g., SMORPH Slope Morphology Model), as well as a modification of this model in order to detect areas affected by creep processes. Besides, we also tested a model based on soil and rock properties mapped in the field (IPT Model). In order to characterize the influence of topographic scale in our ability to predict landslide susceptibility, the simulations were carried out in scales 1:1.000, 1:10.000 and 1:50.000.