GSA Annual Meeting in Denver, Colorado, USA - 2016

NEO-PREDICTIVE VARIABLES CREATION FOR MASS MOVEMENT HAZARD ASSESSMENT AT A MEDIUM SCALE USING WEIGHT OF EVIDENCE APPROACH

MASTERE, Mohamed, Ministry of urban and territorial planning, National Institute of Urban and Territorial Planning, Allal El Fassi avenue, Madinat Al Irfane,, Rabat-Instituts Morocco, POBOX:6215, Rabat, 10000, Morocco, mohamed.mastere@gmail.com

Geological hazards represent a research topic that is in constant development. Populated areas are ever expanding and the requirements for the control of the hazards they represent are becoming increasingly significant. Orogenic zones are among the areas affected by different types of hazards. This study focuses on the evaluation of the spatial component of the mass movement (MM) hazard which is the most common hazard in the Northwest part of the Rifain belt in Morocco. It introduces a three-step procedure to assess mass movement susceptibility (MMS) using weights of evidence approach (WOE) in the Bab Taza region (Central Rif, Morocco) known as one of the most areas prone to mass movements (MM). Nine contextual variables that characterize the geological environment (lithology, fracturing, seismicity, slope gradient, elevation, aspect, stream network, precipitation and land-use) were mapped and classified in order to better understand their interrelationships and their respective effects in the onset of MM. Following this, the classification, inventory, description and the analysis of the MM were carried out by interpretation of high remote sensing data associated to the field study. A total of 247 MM of the following types: landslides (45%), rock-falls (48%) and debris flows (7%) were identified, which covered an area of 20 km² and accounting for 3 % of the study area (630 km²).To eliminate the causal dependency a strategy based on the establishment of three neo-predictive variables (NPV) with a geomorphological significance from the parameters responsible of the conditional independency violation was implemented. Several simulations with several combinations of parameters were performed in order to examine the influence of each variable on the predictive power of MMS map. The prediction capability of each model is determined by the area under the curve value and the best combination model was selected to generate the MMS map.