A Tool for Modelling the Impact of Triggered Landslide Events on Road Networks

Thursday, 18 December 2014: 5:15 PM
Faith E Taylor1, Michele Santangelo2, Ivan Marchesini2, Bruce D Malamud1 and Fausto Guzzetti2, (1)King's College London, Earth and Environmental Dynamics Research Group, Department of Geography, London, WC2R, United Kingdom, (2)CNR Institute of Research on Hydrogeological Hazards in Southern Italy and Islands, Perugia, Italy
In the minutes to weeks after a landslide trigger such as an earthquake or heavy rain, tens to thousands of landslides may occur across a region, resulting in simultaneous blockages across the road network, which can impact recovery efforts. In this paper, we show the development, application and confrontation with observed data, of a model to semi-stochastically simulate triggered landslide events and their impact on road network topologies. In this model, “synthetic” triggered landslide event inventories are created by randomly selecting landslide sizes and shapes from already established statistical distributions. The landslides are then semi-randomly distributed over the region’s road network, where they are more or less likely to land based on a landslide susceptibility map. The number, size and network impact of the road blockages is then calculated. This process is repeated in a Monte Carlo type simulation to assess a range of scenarios. Due to the generally applicable statistical distributions used to create the synthetic triggered landslide event inventories and the relatively minimal data requirements to run the model, the model is theoretically applicable to many regions of the world where triggered landslide events occur. Current work focuses on applying the model to two regions: (i) the Collazzone basin (79 km2) in Central Italy where 422 landslides were triggered by rapid snowmelt in January 1997, (ii) the Oat Mountain quadrangle (155 km2) in California, USA, where 1,350 landslides were triggered by the Northridge Earthquake (M = 6.7) in January 1994. When appropriate adjustments are made to susceptibility in the immediate vicinity of the roads, model results match reasonably well observations. In Collazzone (length of road = 153 km, landslide density = 5.2 landslides km-2), the median number of road blockages over 100 model runs was 5 (±2.5 s.d.), compared to the observed number of 5. In Northridge (length of road = 780 km, landslide density = 8.7 landslides km-2), the median number of road blockages over 100 model runs was 108 (±17.2 s.d.) compared to the observed number of 48. As model development progresses, we hope that this open source tool can be applied to other locations to aid civil protection agencies in exploring the potential impact to the road network of triggered landslide events.