CGRG Bibliography of Canadian Geomorphology
Search Results
Author : Baillifard, F.; Couture, R.; Jaboyedoff, M.; Locat, J.; Locat, P.; Kirkwood, D.; Robichaud, G.; Hamel, G.; and Rouiller, J.-D.
Date : 2004.
Title : Towards a GIS-based rockslide hazard assessment along the Quebec City Promontory, Quebec, Canada.
Publication : 2nd Swiss Geoscience Meeting, Lausanne, 2004.
Issue :
Page(s) :
Abstract
An original risk assessment method is presented and based on the fact that risk assessment along linear objects at risk (e.g. roads or railways) is often simpler to implement than for planar objects at risk (e.g. inhabited areas), because of the unavailability of a landslide database for planar objects at risk. The method is illustrated by the example of the risk assessment along the Quebec City Promontory cliff, that is oriented south-eastwards along the north shore of the St. Lawrence River. From 1775 to 2003, 53 landslides have affected the historical district of the city and the highway that are situated between the toe of the cliff and the St. Lawrence River. These landslides have caused 88 fatalities and injured 70 people. The purpose of the presented method is to furnish continuous hazard and risk assessment along a profile parallel to the objects at risk. In order to compare the results of the simulation with the observed distribution distribution of landslides density, a 100 m long interval is chosen. Deduced from the study of the landslide database, five instabilities factors are modelled by mean of a GIS: (1) dip direction, (2) slope, (3) erodability, (4) watersheds, and (5) results of rockslide kinematic tests. A simulated landslide density is calculated by summing and single (a single instability factor is higher than its threshold) and multiple criteria (several instability factors are simultaneously higher than their respective thresholds). The value “1” is assigned to each instability factor or combination of instability factors that exceed their threshold. The thresholds are based on either physical or statistical arguments. The simulated density shows a good agreement with the observed density. The landslide database makes it possible to calculate frequencies based on the calculated landslide density: the overall frequency along the profile is distributed according the simulated landslide density, for each 100 m-long section of the profile. Making the approximation that each block that fails reaches the bottom of the cliff, a probability of propagation of 100% can be assumed. As a consequence the rockfall hazard is equal to the frequency (probability of failure). The probability for a house to get hit is equal to the annual frequency, multiplied with the percentage of houses in each 100 m-long section of the profile. The risk is then calculated by multiplying this result with the vulnerability and the value of the object at risk. As shown by either the 2-dimensional or the 3-dimensional examples presented, the use of thresholds for each instability factor, which is often easier than the estimation of a true value, is promising. This facilitates hazard assessment, even when only using field observations. Moreover, assigning an appreciation parameter to the result leads to approach integrating fuzzy logic.
Bibliography of Canadian Geomorphology