4D ERT Inversion Incorporating the Temperature Effects at Embankment Dams

There are 18,000 reservoir dams in Korea. More than 16 % of them are reported to undergo leakage problems and need to be repaired. For the leakage detection, a resistivity monitoring system was installed the system at a test dam site. Moreover, borehole temperatures were monitored together with resistivity data. The resistivity data set automatically collected every 6 hours is stored to the data base via internet. The interpretation of collected resistivity monitoring data is divided into two steps: preprocessing and 4D inversion. First, time series of resistivity data are filtered to suppress a spike type noise and then median filtering is applied to remove the high frequency noise caused by diurnal variation, assuming that resistivity value does not change abruptly with time. Then 4D inversion is performed for the resampled data sets. In the inversion process, the standard deviation of each time-series is used for data weighting and the regularization parameters between data misfits and model constraints are adjusted automatically by comparing their norms. Furthermore, subsurface temperature monitored at the dam is incorporated to correct the temperature effect caused by seasonal temperature variation.