Effective Regularization Parameters in 4D ERT Inversion

In 4D inversion of resistivity monitoring data, the time-domain regularization parameter should be optimized to obtain a smooth model along the time axis because the inversion result is largely dependent on the parameter and there are many researches to find the optimum time-domain constraint. However these inversion algorithms use the model parameters or model perturbations at the previous iteration. In this study, we developed a new 4D inversion algorithm that automatically determines the space and time-domain regularization parameters based on model parameters and perturbations at the current iteration. The performance of the developed 4D inversion is examined with synthetic data sets for a given time-lapse model. Finally, we applied the proposed inversion algorithm to the 4D inversion of resistivity monitoring data collected at an embankment dam for the detection of leakage.