A Study Of The Accuracy And Precision Of Some Landslide Risk Analyses

F.J. Baynes, I.K. Lee and I.E. Stewart

The results of about 600 landslide risk analyses carried out by the Roads and Traffic Authority of NSW, Australia were reviewed. The analyses relate to a variety of cuts, fills, retaining structures and natural slopes adjacent to the NSW road network and had been assessed using a new procedure based on quantitative landslide risk methodology, but using qualitative rules to assign ratings and matrices to combine the ratings. Measurements of the precision, accuracy and applicability of the procedure are presented. It was found that whilst the results of the landslide risk analyses are suited to prioritising slopes for maintenance programmes, the review demonstrates that it is difficult for experienced practitioners to carry out reliable landslide risk analyses without adequate training.