United States Department of Transportation
i
Artificial Intelligence (AI) for Building a Landslide Inventory & Advanced Landslide Warning System in PA
-
2023-08-03
-
[PDF-11.82 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report 01/03/2022 – 08/03/2023
-
Contracting Officer:
-
Corporate Publisher:
-
Abstract:This report presents the results of a study aiming at developing artificial intelligence (AI) models for advanced warning of rainfall induced landslides for unstable slopes above or below state-maintained roadways in Pennsylvania. Two landslide databases for spatial and spatiotemporal analyses of landslides in Pennsylvania and adjacent areas are compiled. Landslide susceptibility maps (LSMs) are generated for PennDOT Districts 11 and 12 and adjacent areas, including northern West Virginia and eastern Ohio. The results indicate that the spatiotemporal machine learning (ML) model can predict landslides, accounting for both spatial terrain factors and temporal rainfall factors, and the model outperforms pure spatial ML models with the same database size. The LSMs generated from this study highlight areas having very low to very high risk of landslide susceptibility with precipitation, which may be used to establish a hierarchy and mitigate risk for slopes at “very high risk” for landslide susceptibility. The maps may also be used for forecasting purposes. For example, they may be used as an aid for planning and programming purposes to address slopes with “very high” landslide susceptibility first. The maps may be used in the event of incoming storms to target slopes with a very high risk of landslide susceptibility so that mitigation or preventative measures (such as temporary road closure) can be employed to ensure safe travel and minimize damage. In addition, the maps may also help to target post-storm roadway/slope inspections to the most critical and high-risk locations first.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
