K.H. Tiedemann (Canada)
Landslides, debris flows, rockslides, rock falls, runout distance, mobility.
We argue that the statistical analysis of landslides using data from comparable geological areas can provide information which is useful for both the prediction of future landslide events and for the mitigation of hazards related to these future landslide events. This paper attempts to help a gap for rockslides and rock falls and makes two main contributions. First, this analysis models mobility for rockslides and rock falls with a view to determining whether they are significantly different geological phenomena. The present study examined 33 rockslides and found a slope coefficient of -0.16 and 14 rock falls and found a slope coefficient of -0.041, with both coefficients statistically significant with small standard errors. Using a pooled regression model, the two samples appear to be significantly different suggesting that disaggregation is useful. Second, this analysis uses a statistically appropriate method to correct for heteroscedastic model errors. The study finds that the corrected and uncorrected standard errors are significantly different, suggesting that the use of heteroescedasticity adjusted errors should be considered in future modelling.
Important Links:
Go Back
