Electroosmotic flow (EOF) can transport porewater and dissolved neutral solutes through fine-grained porous media, but its magnitude may evolve as electrode reactions alter pH and electrolyte conditions at mineral–water interfaces. Here, we use 24 two-dimensional pore-scale simulations to quantify how interfacial-charge parameterization, background electrolyte concentration, and pore architecture regulate transient electroosmotic displacement. The model couples incompressible flow, multicomponent Nernst–Planck–Poisson transport, time-dependent electrolysis boundary forcing, and chemistry-dependent electroosmotic slip. Zeta potential is calculated from the local aqueous chemistry using either a one-pK electric-double-layer (EDL) parameterization or a two-pK extended triple-layer (ETL) parameterization. The simulations were conducted in four uniformly oriented packings and two composite pore geometries. Under a matched background electrolyte concentration of 10 mM, the geometry-averaged fraction of tracer depleted from the contaminated medium zone after 30 min was 0.998 for the EDL cases and 0.918 for the ETL cases, while the corresponding late-time tailing indices were 0.002 and 0.094. Within the ETL cases, increasing the background electrolyte concentration from 1 to 100 mM reduced the mean 30 min depletion fraction from 0.963 to 0.613 and increased the tailing index from 0.037 to 0.465. None of the 100 mM cases reached 90 % depletion within 30 min. Pore architecture did not alter this group-level ordering but systematically reorganized directional transport and plume geometry within each electrochemical regime. In the uniformly oriented packings, rotation redistributed flow between the gradient-aligned and transverse directions and produced systematic plume reorientation, although fixed-time depletion varied non-monotonically with angle. In the composite geometries, the heterogeneous directional geometry maintained stronger domain-scale directional organization and reached its case-specific early and intermediate progress levels earlier than the non-directional geometry, whereas their fixed-window depletion ranking depended on the electrochemical setting. Comparisons at case-specific relative progress levels separated plume reorganization from differences in transport timescale. These results reveal a two-level control on transient electroosmotic transport: interfacial-charge representation and background electrolyte concentration determine transport magnitude and timescale, whereas pore structure redistributes flow pathways and plume geometry within the resulting regime.