Earthworks such as cuttings and embankments account form a major part of the entire transport network infrastructure. Large parts of that infrastructure in Europe are susceptible to a range of geohazards, landslides being the most prevalent. These landslides frequently result in direct damage to assets, deaths and injuries, while indirectly also leading to traffic disruptions. There is a need therefore to identify critical assets where remediation efforts should be prioritised in order to prevent such events from occurring. Current state of the art practice involves using qualitative risk matrices, where the hazard and consequence components are determined through subjective visual survey observations. Landslide hazard analysis determines the spatial (susceptibility) and temporal probability of landslides of a certain intensity occurring over an observed area. A number of quantitative methods for landslide hazard and risk assessment have been developed recently, generally these methods are considered more effective due to their reduced subjectivity and their consideration of additional factors. A number of studies outline the application of these methods to natural terrain, but to date these methods have not been developed for transport network earthworks. This study presents and compares the results of two landslide susceptibility analysis approaches for cuttings and embankments on a section of Irish Rail network. The first, “geotechnical” approach uses probabilistic slope stability calculations to rank the assets by their reliability index. The second, “statistical” or “datadriven” approach, uses logistical regression as a statistical tool to obtain the susceptibility ranking of the earthworks, using the database of previous failures on the network as an input. Furthermore, several methods for obtaining the temporal hazard characteristics are presented and applied, these methodologies combine to provide a full hazard assessment map of the network.

landslides; earthworks; susceptibility; hazard; risk