A carregar...
Publicação
Assessment of landslide and fluvio-mass-movement susceptibility on the island of Porto Santo (Madeira, Portugal)
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 860.59 KB | Adobe PDF |
Orientador(es)
Resumo(s)
Numerous studies on landslide susceptibility have been carried out on the island of Madeira where landslides have caused significant economic, social and human losses, such as those associated with the rainfall events of December 2009 and February 2010. On the island of Porto Santo, with a hot semi-arid climate, studies related to landslides, but also to fluvio-mass-movements (FMM) on gully development, are practically non-existent, despite the intense erosion observed on this northernmost island of the Madeira archipelago. In this context, it is important to understand which terrain conditions favour the development of this strong presence of gullies by studying the FMM and the landslide occurrence on natural slopes. In order to achieve this goal, three main objectives were defined: (i) inventory of different types of landslides and FMM, on slopes and gullies, through satellite images photointerpretation from different dates; (ii) susceptibility assessment of different types of landslides and FMM, using a bivariate statistical model (Information Value); (iii) and validation of the susceptibility maps using the ROC (Receiver Operating Characteristic) curves, the Area Under the Curve (AUC), the random partition of the landslides and FMM inventory into training (70%) and validation (30%) groups, the effectiveness ratio of susceptibility classes and the definition of the spatial agreement between susceptibility maps. A total of 1563 landslides and FMM were identified on the island of Porto Santo, categorized into the following types: shallow slides (Sslides, 221 cases); deep-seated slides (6 cases); debris flows (23 cases) and FMM (1313 cases). The FMM’s were further subdivided according to their position in the gully context: FMM Head (401 cases), FMM in the middle gully sector walls (FMM1, 600 cases), FMM in the downstream gully sector walls (FMM2, 300 cases). In addition, 12 FMM complex systems were also individualized. Eight independent predisposing variables for the occurrence of landslides and FMM were used to assess susceptibility: slope, aspect, plan curvature, profile curvature, lithology, soil type, distance to watercourses and land use. Seven susceptibility models were generated, four according to the type of mass movement - Sslides (Model 1), FMM Head (Model 2), FMM1 (Model 3) and FMM2 (Model 4); three models considering different combinations of FMM types - FMM1 and FMM2 (Model 5); FMM Head, FMM1 and FMM2 (Model 6); and a model integrating all FMM types (Model 7). In this study it was possible to distinguish the most important predisposing factors for the occurrence of Sslides and those that favour the occurrence of FMM. As expected, the non-dependence of Sslides on the proximity of the waterline, which is one of the most relevant factors for FMMs, is clear. By comparing the FMM susceptibility models, it is possible to observe a high degree of similarity. Nevertheless, it has been possible to identify some particular terrain characteristics that control the distribution of the different types of FMM in the study area. The distance to the waterline is slightly less relevant for the FMM Head than for FMM1 and FMM2, as the former often seems to be associated with retrogressive erosion. FMM2 is the only type of fluvio-mass movement that shows a strong relationship with steep slopes (30° - 45°), with the lithological unit "beach, river, slope and paleosol and sandstone deposits" and with the land use class "inert extraction areas, waste deposition areas and construction sites", the last two often located at the foot of slopes. In general, all susceptibility models have a very similar predictive capacity, which improves slightly when susceptibility is assessed using individual mass movement types. The spatial agreement between the susceptibility maps of the different FMMs is high.
Descrição
Palavras-chave
Landslides Fluvio-mass-movements Information Value Susceptibility Porto Santo