Time–Lapse Electrical Resistivity Tomography for Evolving Water–Bearing Fractures Ahead of Tunnels: An Improved Inversion Framework and Synthetic Verification

Water–bearing fractures and seepage–prone zones ahead of tunnel faces may evolve rapidly under excavation–induced disturbance, making early identification and process tracking essential for risk mitigation. Cross–hole electrical resistivity tomography (ERT) is sensitive to fluid–controlled conductivity contrasts, but time–series interpretation based on independently inverted snapshots is often unreliable due to ill–posedness, noise, and temporal inconsistency. In this study, we propose an improved time–lapse ERT inversion framework for monitoring evolving water–bearing fractures ahead of tunnels. The method is formulated as a baseline–anchored, Occam–consistent difference inversion that directly estimates resistivity changes relative to an initial state, incorporating error–aware weighting of differenced data and anisotropic regularization adapted to cross–hole sensitivity, so that temporal coherence is enforced during inversion rather than through post hoc differencing. Synthetic verification is conducted using three dynamic scenarios representing horizontal, vertical, and diagonal migration of conductive water–bearing pathways between boreholes. Quantitative comparison against independent inversion across all scenarios and time steps demonstrates that the proposed framework substantially reduces the root mean square error and mean relative error of the recovered resistivity, while significantly improving the spatial correlation coefficient between the recovered and true models, with the largest improvements observed in the diagonal–migration scenario. The reconstructed change maps exhibit more compact anomaly geometry and delineate evolution corridors aligned with the prescribed trajectories, whereas independent inversion produces diffuse and epoch–dependent change patterns. These results indicate that the proposed time–lapse inversion framework provides a more reliable basis for interpreting evolving seepage–related conductive structures in tunnel–ahead investigations.