Rockfall hazard and risk analysis for Monte da Lua, Sintra, Portugal

The prediction of rockfall trajectories below a rock cliff is essential in susceptibility, hazard and risk maps, particularity close to populated areas. The Monte da Lua hill area in Portugal, a tourist destination close to the historic city of Sintra (UNESCO World Heritage), is a typical granite boulder chaos landscape where from time to time rockfalls occur, the last such event having occurred on 29 January 2002. This area is therefore suitable to develop a rockfall study in order to provide hazard and risk maps a basis for mitigation measures. A preliminary investigation of the area leads to the identification of 188 potentially dangerous boulders. Detailed locations and geotechnical characteristics in terms of geometry, strength and context were sampled for each boulder. Digital elevations at 1 × 1 m resolution, known rockfall trajectory and building locations are provided in a GIS project for the study together with the spatial database of boulder characteristics. The modelling approach was conducted in two steps: (1) discrimination of the boulders in terms of static and dynamic mobility behaviour with multivariate analysis; (2) stochastic simulation of rockfall trajectories. The rockfall trajectory algorithm proposed is straightforward and is only dependent on elevation data, initial location of boulders and a friction angle. Due to the slope of the area, it assumes that rockfall is always of the rolling or sliding type. The friction angle was calibrated on the basis of the rockfall travel distance recorded on 29 January 2002 and generates simulated "realistic" trajectories. A smaller friction angle increases all simulated trajectories, leading to more "pessimistic" scenarios. The combined analysis of trajectories and potential damage to buildings and discrimination in terms of static and dynamic behaviour provides a final table in which all 188 sampled boulders are classified in one of the five risk grades.