Detection and Attribution of Climate Non-Stationarity in Cold Regions Geotechnical Design Using Artificial Intelligence

As our climate changes (non-stationarity), we face new challenges in assessing hazard frequency and assessing the vulnerability of infrastructure to the effects of global warming, such as changing precipitation patterns and increasing ground surface temperature. The lack of appropriate incorporation of future weather and climate information into the design, operation, and management of infrastructure remains a significant barrier to systematically improving climate-resilient infrastructure. In this paper, a spatiotemporal AI-powered platform for the detection and attribution of climate non-stationarity is used to as boundary conditions in a computational simulator to investigate the resiliency of a typical embankment in Canada. It can be seen that considering the climate boundary condition causes the reduction in the factor of safety of the embankment and the change of the performance of the geosynthetic layer under the embankment.