The thesis investigates and proposes a suitable form of collecting, organizing and analysing landslide data in order to improve the knowledge of landslide processes in Central America. The study recommends the organization of existing and new data in a national landslide database for Nicaragua in a digital format. The database is intended to support the scientific community and local and national authorities in landslide hazard assessment, emergency management, land-use planning and the development of early warning systems. Valuable information on landslide events has been obtained from a great variety of sources, such as landslide inventory maps, technical reports, natural hazard catalogues, newspapers, historical monographs and scientific papers. Through analyses of landslide data stored in the database, the thesis presents the status of landslide knowledge in Nicaragua both at national and local scales and characterizes landslides in terms of spatial and temporal distribution, types of slope movements, triggering mechanisms, number of casualties and damage to infrastructure. The research collected spatial and temporal information for about 17000 landslides that occurred in mountainous and volcanic terrains, between 1570 and 2003.Information is mainly recorded for the period 1826 to 2003, with a large number of events (62% of the total number) during the disastrous Hurricane Mitch in October 1998. Data on human losses and damages in historical documents were used to show consequences of landslides. The consequences in Nicaragua have up until now been greatly underestimated. Debris flows have been the most common types of recorded events, both in volcanic and non-volcanic areas, but other types, including rockfalls and slides, have also been identified. Intense and prolonged rainfall, often associated with tropical cyclones, and seismic and volcanic activity, represent the most important landslide triggers. The influence of topographic (elevation, slope angle, slope aspect) and lithologic parameters was analyzed for the northern part of Nicaragua.Data, mainly from 348 fairly well documented events in Nicaragua and 19 from other Central American countries, have been treated statistically to search for possible correlations and empirical relationships to predict run-out distances for different types of landslides, knowing the height of fall or the volume. The mobility of a landslide, expressed as the ratio between height of fall (H) and run-out distance (L) as a function of the volume and height of fall; and the relationship between the height of fall and run-out distance were studied for rock falls, slides, debris flows and debris avalanches. The results showed that debris flows and debris avalanches at volcanoes have the highest mobility and reach longer distances compared to other types of landslides in the region. Among them, the longest (~25 km) and the most catastrophic in Nicaragua occurred at the Casita volcano on October 1998 triggered by the rainfall associated with Hurricane Mitch.The Casita event started as a flank collapse which rapidly evolved into a debris avalanche and transformed into a lahar while moving downstream. In this thesis field data were combined with mechanical models, to provide a better understanding of the initial flank collapse. New insight into the geology, tectonics, pre- and post-failure geometry, and stratigraphy of the scarp area were provided and the failure sequence and pre- and post failure slope stability were modelled combining data from previous published and unpublished works and observations. The flank collapse was constrained by confirming that it included three failure stages that occurred continuously during an interval of seconds to a few minutes, involving both the northern and southern area of the scarp as previously proposed by Vallance et al. (2004). The study established the most likely failure mechanism of the first stage and identified the most likely mechanism of failure of the second stage. Analyses indicate that sliding in the volcanic breccia overlying a unit of clay-rich pyroclastic deposits with low shear strength controlled the failure mechanism and the stability in the southern area of the scarp. The failure then propagated into the overlying loose materials. Slope stability analyses for undrained conditions indicate that the remaining southern slope becomes unstable if it gets saturated during intense and/or prolonged rainfall, i.e. same conditions as those produced during Hurricane Mitch. In dry conditions the slope is stable as long as the berm of colluvium at the toe is not removed.
List of papers
Devoli G., Morales A., Høeg K. (2007). Historical landslides in Nicaragua - Collection and analysis of data. Landslides: 4 (1), 4-18.
Devoli G., Strauch W., Chávez G., Høeg K. (2007a). A landslide database for Nicaragua: A tool for landslide hazard management. Landslides: 4 (2), 163-176.
Devoli G., De Blasio F., Elverhøi A., Høeg K., (accepted, 8th of March 2008). Statistical analysis of landslide events in Central America and their run-out distance. Geotechnical and Geological Engineering Journal:
Devoli G., Cepeda J., Kerle N., (submitted). The 1998 Casita volcano flank failure revisited – new insights into geological setting and failure mechanisms. Submitted to the Engineering Geology Journal (27th of November 2007)