Landslide susceptibility assessment in the coastal area of Essaouira – Morocco

Khouz, Abdelah; Trindade, Jorge; El Bchari, Fatima; Bougadir, Blaid

http://hdl.handle.net/10400.2/20836

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
MEDGU21_AK.pdf		308.69 KB	Adobe PDF	Download

Send Feedback

Authors

Abstract(s)

Landslide susceptibility mapping is an important tool for landslide risk management and urban planning. The main aim of this study is to assess coastal slope landslide susceptibility in Essaouira city (Morocco) using a deterministic approach. The primary data needed for analyzing the risk of landslides include a record of landslides factors that contribute to their occurrence factors that trigger them and elements in the area at risk. Among these the inventory of landslides holds significance. It provides information, about where landslides have happened in the past the types of landslides how they failed, the causes, behind their occurrence and the extent of damage they caused. Historical archives, field data and image interpretation are sources, for creating landslide inventories. However it is unfortunate that these sources are often incomplete making it challenging to conduct a risk assessment. A total of 588 landslides were recognized and mapped in the study area, the terrain units of the study area were divided according to morphometric and operational criteria. A total of 2975 terrain units with 50 m wide were classified in coastal system type, based on information from satellite imagery, aerial photographs, digital elevation model, geological maps, and data and field surveys. Afterward, the landslide susceptibility assessment model was carried out by classifying the terrain units into stable and unstable units. The classification process is based on the quantification of the percentage of the unstable area of each terrain unit. This initial method, for assessing the likelihood of landslides, which involves analyzing land forms using photos, satellite images and fieldwork along, with GIS analyses helped us identify 931 areas of land that are considered unstable. This accounts for 61% of the study area (known as the Rocky subsystem). Additionally, more information was added to a Geodatabase of various independent variables such as morphology (height, slope, curvature, and aspect), geology (structure, lithology, tectonics, soil, and bedrock geotechnical properties), hydrogeology, and climatology. A statistical model that uses information values is employed to analyze instances of landslides and identify the conditions in which they have occurred. This helps determine the combinations of factors that contribute to the likelihood of landslides. To guarantee meaningful validation process, the inventory dataset was subdivided into a test and a validation group. The model building process utilized a test group comprising 40% of the inventory while the validation group was employed to perform cross validation on the results of the model. Finally, the susceptibility map was validated to define the sensitivity analysis for independent variables and combinations of variables.