Landslide failure and runout susceptibility in the upper T. Ceno valley (Northern Apennines, Italy)

Abstract

The 'Conditional Analysis' multivariate statistical method was used to evaluate Landslide Susceptibility (LS) in an area of the Italian Northern Apennines. An Inventory Landslide map, containing 518 landslides, and seven landslide-related factor maps (lithology, elevation, slope angle and aspect, profile and tangential curvatures, bedding/ slope relations) were processed using a shell script that automatically carries out the whole procedure producing a final map with five Failure Susceptibility (FS) classes. The procedure was applied separately to the most frequent landslide typologies, namely, rotational slides, flows, and complex landslides, and for each of the 127 different combinations of the seven factors. To define areas potentially affected by the down-slope movement of the depleted material, four runout belts with different probabilities were distinguished around the classes with highest FS. By overlaying the resulting map with the map of the elements at risk, namely, settlements, roads, and streams, the spatial risk in the area was assessed.

Landslide susceptibility assessment in the Upper Orcia Valley (Southern Tuscany, Italy) through conditional analysis: a contribution to the unbiased selection of causal factors

“…A well-structured susceptibility evaluation procedure should consider the terrain conditions preceding the landslide events, since the failure occurrence could cause strong topographic modifications of these areas Fabbri, 1999, 2008;Fernandez et al, 2003;Ayalew and Yamagishi, 2005;Nefeslioglu et al, 2008;Clerici et al, 2010). To test the best feature to be used for each landslide type, our procedure provides for the application of the susceptibility analysis using, for landslide inventory, the depletion zones and the outer buffer areas from the depletion zones, the latter to be dimensioned as a function of the input data resolution.…”

“…Once the factor values have been correctly classified, it is necessary to select the correct number of factors, because, again, the greater the diversity of vUCUs, the lower the extent of the spatial units and, in turn, the lower the significance of the statistical analyses (Clerici et al, 2006(Clerici et al, , 2010. However, filtering techniques to cancel out or merge small and insignificant areas can introduce bias or errors in the procedure .…”

“…Some factor selection procedures, such as those proposed by Chung et al (2002), Remondo et al (2003) Clerici et al (2010), provide for selecting the most significant factors after having computed all the possible susceptility maps from all the possible combinations of the potential causal factors and having tested the results by means of a validation method.…”

“…2). If more factors than the needed number have a G-value above the mean, we propose a method of exclusion based on the correlation between these factors, in order to avoid that the vUCUs provide redundant information about the landslide distribution (Clerici et al, 2010). Thus we computed a unique correlation matrix for all the possible pairs of potential factors using Pearson correlation when comparing two numerical variables (such as slope and drainage density), the Cramer Index when comparing two nominal variables (such as lithology and land use), and the η 2 index when comparing a numerical variable to a nominal variable (such as lithology and amplitude of relief) (Pearson, 1896;Cramer, 1999).…”

Forward logistic regression for earth-flow landslide susceptibility assessment in the Platani river basin (southern Sicily, Italy)

“…Examples of previous studies Conditional analysis (CA) Clerici et al (2010, Conoscenti et al (2008, Costanzo et al (2012a, 2012b, Irigaray et al (2007), Jiménez-Peralvárez et al (2009), Rotigliano et al (2011, 2012, Vergari et al (2011) Discriminant analysis (DA) Baeza and Corominas (2001), Carrara (1983), Carrara et al (2008), Guzzetti et al (2006), Rossi et al (2010. ) Binary logistic regression (BLR) Atkinson and Massari (1998), Ayalew and Yamagishi (2005), Bai et al (2010), Can et al (2005), Carrara et al (2008), Chauan et al (2010), Conforti et al (2012), Dai and Lee (2002), Davis and Ohlmacher (2002), Erener and Düzgün (2010), Mathew et al (2009), Nandi and Shakoor (2009), Nefeslioglu et al (2008, Ohlmacher and Davis (2003), Van den Eckhaut et al (2006 Classification and regression trees (CART) Felicísimo et al (2012), Vorpahl et al (2012) Artificial neuronal networks (ANN) Aleotti and Chowdhury (1999), Ermini et al (2005), Lee et al (2004), Pradhan and Lee (2010) Original Paper exploited to compare the fitting of the model having only the constant term (all the β p are set to 0) with the fitting of the model that includes all the considered predictors with their estimated non-null coefficients so as to verify if the increase in likelihood is significant; in this case, at least one of the p coefficients is to be expected as different from zero …”