Effects of model dimensionality (2-D vs 3-D) on simulated groundwater flow conditions in post-glacial marine clay deposits in the context of slope stability analysis

Traditional slope stability analyses rely on 2-D cross-section models that assume rivers act as groundwater divides. However, these simplifications may misrepresent groundwater flow and affect calculated factors of safety. This study evaluates when such assumptions are appropriate by comparing 2-D and 3-D hydrogeological models for slopes of meandering rivers in postglacial marine clays. Results show that for models representing one side of a valley, 2-D modeling is appropriate or conservative for straight slopes without an underlying till layer, as hydraulic head differences remain below 0.5m. If a till layer is present, imposing a no-flow boundary beneath the river produces higher simulated heads (up to 3m) yielding conservative factors of safety at the cut banks for meandering and straight slopes. However, in most cases, when a till layer is present, groundwater flow beneath the river necessitates modelling both valley sides to avoid unrealistic flow paths. In such cases, 2-D models that include both sides of the valley remain suitable for straight slopes with underlying till, as well as at the point bar in meandering slopes, regardless of till presence. In all other configurations, 2-D models produce lower hydraulic heads and factors of safety by up to 10% relative to 3-D simulations.