Loading...
UNCERTAINTY QUANTIFICATION OF LANDSLIDE SUSCEPTIBILITY MAPPING USING BAYESIAN NETWORK
Khabiri, Sahand
Khabiri, Sahand
Citations
Altmetric:
Genre
Thesis/Dissertation
Date
2023-08
Advisor
Committee member
Group
Department
Civil Engineering
Subject
Permanent link to this record
Collections
Research Projects
Organizational Units
Journal Issue
DOI
http://dx.doi.org/10.34944/dspace/9499
Abstract
Landslides account for an average of 25 fatalities and monetary costs exceeding $2 billion USD annually, according to the U.S. Geological Survey (USGS). Landslide susceptibility mapping (LSM) can provide valuable insights for landslide characterization and helps optimize risk mitigation strategies. To enhance the reliability of LSM, uncertainty in landslides characterization must be identified and quantified. There are longstanding and systematic uncertainties in LSM that remain unaddressed or inadequately resolved in the literature, including (1) the impact of boundary geometry on landslide characterization, (2) challenges in defining negative samples (i.e., non-landslide points) for machine learning-based model training, and (3) interpreting the causal relationships among factors influencing landslides and uncertainty propagation in model predictions. To address these knowledge gaps, this work presents results of 1) sensitivity analysis to assess the impact of varying mapping geometry of the landslides on the LSM outcomes, 2) an uncertainty quantification of different negative sample scenarios on LSM development using a Bayesian network model, and 3) a comparative analysis between static and dynamic model structures incorporating a physical slope stability model within a probabilistic machine learning framework. These analyses explore uncertainty derived from mapping, sampling, and modeling perspectives, hence providing valuable insights into future robust LSMs and improved understanding of the causal relationships between geo-environmental conditions and landslide occurrences.
Description
Citation
Citation to related work
Has part
ADA compliance
For Americans with Disabilities Act (ADA) accommodation, including help with reading this content, please contact scholarshare@temple.edu