The proposed study is aimed at the evaluation and mapping of the landslide susceptibility along a sector of highway «A3 (Salerno-Reggio Calabria)» between Cosenza Sud and Altilia, northern Calabria, using a GIS-based Conditional Analysis method. Landslide inventory map of the area was obtained by detailed field survey and air-photo interpretation. Altogether, 844 landslides were mapped. The types of movement are mainly slides, complex landslides and, subordinately, flows. In order to estimate and validate landslide susceptibility, the landslides were divided in two group. One group (training set) was used to prepare susceptibility map and the second group (testing set) to validate the map. Lithology, distance from tectonic elements, land use, slope, aspect, plan curvature and Stream Power Index (SPI), were assumed as landslide predisposing factors. In order to evaluate landslide susceptibility Conditional Analysis was applied to Unique Conditions Units (UCUs) that are a unique combination of the predisposing factors. Subsequently, the landslide area is determined within each UCU and the landslide density is computed. The outcome of the study was a classification of the study area into four susceptibility classes, ranked from low to very high. The results showed that the 33% of the study area is characterized by a high to very high degree of susceptibility. The overlay of the landslides of testing set with the susceptibility map, showed that over 75% of the landslides is correctly classified, falling in high and very high susceptibility classes. Prediction performances of this map is checked by using both success rate curve and prediction rate curve. The validation results showed that, area under the curve (AUC) for success rate curve was 0.82 whereas for the prediction rate curve was 0.78. Therefore, prediction rate curve revealed that model has good prediction performance. Finally, the landslide susceptibility map provides the baseline information for further evaluations of landslide hazards and related risks.
Suscettibilità di Frana sulle Grandi Vie di Comunicazione: Caso di Studio Autostrada A3, Tratto Cosenza Sud – Altilia (Calabria Settentrionale)
MUTO, Francesco;Critelli S;
2016-01-01
Abstract
The proposed study is aimed at the evaluation and mapping of the landslide susceptibility along a sector of highway «A3 (Salerno-Reggio Calabria)» between Cosenza Sud and Altilia, northern Calabria, using a GIS-based Conditional Analysis method. Landslide inventory map of the area was obtained by detailed field survey and air-photo interpretation. Altogether, 844 landslides were mapped. The types of movement are mainly slides, complex landslides and, subordinately, flows. In order to estimate and validate landslide susceptibility, the landslides were divided in two group. One group (training set) was used to prepare susceptibility map and the second group (testing set) to validate the map. Lithology, distance from tectonic elements, land use, slope, aspect, plan curvature and Stream Power Index (SPI), were assumed as landslide predisposing factors. In order to evaluate landslide susceptibility Conditional Analysis was applied to Unique Conditions Units (UCUs) that are a unique combination of the predisposing factors. Subsequently, the landslide area is determined within each UCU and the landslide density is computed. The outcome of the study was a classification of the study area into four susceptibility classes, ranked from low to very high. The results showed that the 33% of the study area is characterized by a high to very high degree of susceptibility. The overlay of the landslides of testing set with the susceptibility map, showed that over 75% of the landslides is correctly classified, falling in high and very high susceptibility classes. Prediction performances of this map is checked by using both success rate curve and prediction rate curve. The validation results showed that, area under the curve (AUC) for success rate curve was 0.82 whereas for the prediction rate curve was 0.78. Therefore, prediction rate curve revealed that model has good prediction performance. Finally, the landslide susceptibility map provides the baseline information for further evaluations of landslide hazards and related risks.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
