A Review of Dynamic Soil Slope Analysis under Seismic Loading: From Classical to Probabilistic and Data-Driven Approaches

Dynamic deformation of soil slopes under seismic loading remains a critical issue in geotechnical earthquake engineering. This review traces the progression of analytical and numerical approaches for assessing seismic slope performance, highlighting key advances and persisting limitations. Early methods, including pseudo-static analysis and Newmark's rigid block model, provided a foundation but were unable to capture nonlinear soil behavior, topographic amplification, and spatial variability of ground motion. The development of continuum-based and numerical techniques, particularly finite element and finite difference methods, has enabled two and three dimensional modeling that incorporates nonlinear responses, soil heterogeneity, and seismic wave propagation. Probabilistic approaches, such as Monte Carlo simulations and stochastic vibration theory, have further advanced the quantification of uncertainty in soil properties and seismic inputs. Despite these improvements, challenges remain, including high computational demand, limited hydro-mechanical coupling, and insufficient validation through real-world case histories. Current research gaps involve the limited treatment of aftershocks, climate change impacts, and complex three-dimensional topography. Future directions emphasize the integration of machine learning, probabilistic frameworks, and multi-hazard analysis to enhance predictive reliability. By synthesizing findings from over ۵۰ studies, this article provides a roadmap for advancing seismic slope stability assessment toward more robust, coupled, and data-driven methodologies.

کلیدواژه ها:

Seismic slope stability ، Dynamic deformation ، Newmark's method ، Probabilistic analysis ، Uncertainty quantification