NH030-0019
Application of Scenario Earthquakes for Assessment of Coseismic Landslide Hazard: A Case Study in Costa Rica
Application of Scenario Earthquakes for Assessment of Coseismic Landslide Hazard: A Case Study in Costa Rica
Monday, 14 December 2020
Virtual
Abstract:
We present an analysis of scenario earthquakes to assess landslide probability within tectonically active areas of the Cocos-Caribbean plate boundary in Costa Rica. Costa Rica offers both a rich variety of seismotectonic source areas and a long-recorded history of earthquake-induced landslides (EQIL) dating back to 1772. In order to test the applicability of our global model for EQIL events, we compare modeled landslide distributions with spatially mapped landslide inventories from the M7.6 1991 Limon and M6.1 2009 Cinchona earthquakes. We use the USGS ShakeMap software to compute scenario earthquakes for six tectonic environments within and around Costa Rica, including: (a) subduction interplate events located near the coast, (b) intraslab events within the subducting Cocos plate, as well as crustal events within (c) the Panama Fracture Zone, (d) the North Panama Deformed Belt, (e) the Central Volcanic Range, and (f) the Costa Rica Deformed Belt. We examine a wide variety of fault geometries, depths, magnitudes, and directivity to investigate the impact of each source factor on the spatial distribution of ground shaking parameters. Each scenario is used in conjunction with our global EQIL model to estimate the areas of highest landslide probability relative to the triggering earthquake. We find that source magnitude and depth are the most critical source controls on landslide distribution, followed by fault geometry and directivity. The specific areas affected by an individual source depend on the convolution of ground shaking with earth surface parameters, including surface slope, wetness, lithology, and landcover. By comparing predicted landslide locations with infrastructure and population density, we can also assess the likelihood of landslide-related damage and fatalities for each scenario event using a global landslide fatality model. We find that moderate-sized shallow crustal events within the Central Volcanic Range and the Costa Rica Deformed Belt, together with large, megathrust earthquakes near the coast, have the potential to produce significantly damaging landslide impacts.