Statewide Seismic-Induced Landslide Risk Assessment in Virginia: A probabilistic approach
College
College of Engineering & Technology (Batten)
Department
Civil & Environmental Engineering
Graduate Level
Doctoral
Presentation Type
No Preference
Abstract
Landslide hazards, driven by triggers such as earthquakes and rainfall, pose significant risks to infrastructure and community resilience. However, the assessment of these hazards has been constrained by the limited resolution and qualitative-based evaluations of the existing map in Virginia. This paper addresses these limitations by adapting a novel probabilistic approach (Saygili and Rathje, 2009) to develop a seismic-triggered, high-resolution, quantitative statewide landslide hazard map for Virginia. The method used the rigid sliding block approach and the Newmark (1965) displacement model to evaluate seismic-triggered landslide susceptibility, which is defined by a yield acceleration ky, the minimum horizontal acceleration of earthquake shaking that initiates the block sliding, resulting in a landslide. The yield acceleration is calculated by topographic slopes, soil shear strength parameters (including cohesion, internal friction angle, and unit weight of the soil), and groundwater table. To create a high-resolution map, we derived a topographic slope map (with a 10 m spacing) for Virginia from high-resolution digital elevation models. The soil shear strength parameters were inferred from a statewide surface geological map. The groundwater table was interpolated from water wells data. Next, we used probabilistic seismic hazard analysis (Cornell, 1968) to evaluate the probabilistic occurrence of seismic triggers such that the resultant ground shaking exceeds ky. A series of probabilistic assessments of seismic triggers for different return periods (the duration that, on average, one exceedance event would occur), including 100, 475, 975, and 2475 years (the most commonly considered in seismic design), were produced. Within a shorter return period, fewer extreme seismic events would occur, which would trigger fewer landslides. Therefore, we expect landslide hazards to be weaker during a shorter return period than a longer return period. After convolving the map by probabilistic assessments of seismic triggers at different return periods, the corresponding landslide hazard maps at those return periods were generated. We observed that higher landslide hazards appeared in the western regions of Virginia, primarily mountainous areas with steeper slopes. The hazard maps are in 10-m spacing that can provide a very accurate risk assessment for statewide infrastructure (e.g., transportation roads). In our future research, we will extend the analyses to conduct multi-triggered landslide hazards assessment by including rainfall-triggered landslides.
Keywords
Landslide Risk Assessment, Seismic-Triggered Landslides, Geohazards, Virginia, Probabilistic Analysis
Statewide Seismic-Induced Landslide Risk Assessment in Virginia: A probabilistic approach
Landslide hazards, driven by triggers such as earthquakes and rainfall, pose significant risks to infrastructure and community resilience. However, the assessment of these hazards has been constrained by the limited resolution and qualitative-based evaluations of the existing map in Virginia. This paper addresses these limitations by adapting a novel probabilistic approach (Saygili and Rathje, 2009) to develop a seismic-triggered, high-resolution, quantitative statewide landslide hazard map for Virginia. The method used the rigid sliding block approach and the Newmark (1965) displacement model to evaluate seismic-triggered landslide susceptibility, which is defined by a yield acceleration ky, the minimum horizontal acceleration of earthquake shaking that initiates the block sliding, resulting in a landslide. The yield acceleration is calculated by topographic slopes, soil shear strength parameters (including cohesion, internal friction angle, and unit weight of the soil), and groundwater table. To create a high-resolution map, we derived a topographic slope map (with a 10 m spacing) for Virginia from high-resolution digital elevation models. The soil shear strength parameters were inferred from a statewide surface geological map. The groundwater table was interpolated from water wells data. Next, we used probabilistic seismic hazard analysis (Cornell, 1968) to evaluate the probabilistic occurrence of seismic triggers such that the resultant ground shaking exceeds ky. A series of probabilistic assessments of seismic triggers for different return periods (the duration that, on average, one exceedance event would occur), including 100, 475, 975, and 2475 years (the most commonly considered in seismic design), were produced. Within a shorter return period, fewer extreme seismic events would occur, which would trigger fewer landslides. Therefore, we expect landslide hazards to be weaker during a shorter return period than a longer return period. After convolving the map by probabilistic assessments of seismic triggers at different return periods, the corresponding landslide hazard maps at those return periods were generated. We observed that higher landslide hazards appeared in the western regions of Virginia, primarily mountainous areas with steeper slopes. The hazard maps are in 10-m spacing that can provide a very accurate risk assessment for statewide infrastructure (e.g., transportation roads). In our future research, we will extend the analyses to conduct multi-triggered landslide hazards assessment by including rainfall-triggered landslides.