Risk estimation for railways exposed to landslides

Railways have three basic options for managing their exposure to geotechnical hazards: reduce the source of the hazards, spatially avoid the hazards, or reduce the temporal exposure to the hazards. Although railway tracks and the supporting infrastructure are exposed to geotechnical hazards all the time, the most vulnerable components of the railway, the trains and the railway personnel, are only exposed as they pass the hazards. As a result, provided that the occurrence of geotechnical hazards can be reliably predicted, reduction of the temporal exposure of trains and personnel to hazards, during high hazard periods, requires the least capital expense, and is therefore the most economic option. This thesis demonstrates a means of using precipitation indices to identify periods of higher potential landslide hazard for a site in Maple Ridge, BC. Over 100 years of landslide records compiled by the Canadian Pacific railway and others for this site are correlated with daily precipitation data from the Maple Ridge, BC area. Two methods are used to predict the occurrence of landslides using precipitation data. One method consists of using the three parameter Generalized Extreme Value (GEV) frequency distribution analysis of various duration of antecedent precipitation to develop reliable estimates of the return period of each duration antecedent precipitation. Durations of up to a year were considered. The rarest antecedent duration concurrent with the landslide event is assumed to have triggered the landslide. The other method uses the coincidence of up to three elevated antecedent precipitation conditions correlated with the landslide records to identify indices that, when combined, provide reliable prediction of precipitation induced landslide events. An event tree risk analysis for the probability of both train accidents and the probability of death of individual (PDI) railway employees due to a geotechnical hazard is developed. The risk analysis is used to quantify the benefits of using a precipitation induced landslide warning system and to measure the effectiveness of other risk reduction strategies for geotechnical hazards. Risk can be used to evaluate the merits of various mitigative options and their net costs and benefits. It is shown that the PDI of railway personnel as a result of geotechnical hazards on CP is within tolerable limits when compared to published tolerable employee risk. It is shown for the Maple Ridge site that a precipitation induce landslide warning system and a hazard detection system would reduce the probability of a fatality by 39% and 80% respectively. With this information the cost and delays introduced by each system can be compared.