Multi-scale landslide risk assessment in Cuba : discussions, conclusions and recommendations (THESIS VERSION)

Abstract

More research is needed on the optimal ways for generating landslide inventories, in particular to design different inventory systems based either on local reports or on periodical surveys. Existing landslide inventory databases could be enhanced by introducing time and scale dimensions. Time dimensions involve the linking of landslide records of the same movement across time, in order to properly address reactivation history. Scale dimension means the ability to change the representation of the landslide at different scales from a single point up to a set of polygons representing landslide features. Other aspects such as magnitude, time of occurrence, geometric dimensions are very valuable for statistical analysis of landslides and to predict expected magnitude and travel distance. Data about landslide damage is usually missing in landslide inventories and much research could be undertaken in order to link magnitude and damage for vulnerability assessment. The lack of the temporal component in the landslide inventory was a drawback in this research and in many other studies. A key recommendation for this problem is to establish a landslide inventory and to keep it updated.