Predicting the resistivity signature of internal erosion in freshwater embankment dams: anomaly polarity, detectability, and a full-scale benchmark

Seokhoon Oh 

Electrical resistivity tomography (ERT) is a standard tool for embankment-dam inspection, and low-resistivity anomalies are conventionally read as leakage or internal erosion. High-resistivity signatures of internal erosion have been reported repeatedly but never derived from petrophysical first principles. We show why: for the freshwater reservoirs and clayey cores of most inland dams, about 97 % of intact-core conduction is clay surface conduction, so erosion that removes fines and replaces them with fresh water raises resistivity, and a closed-form criterion, σw(F*−1) > B·Qv, places typical inland reservoirs a factor of 3–10 inside the resistive-damage regime — explaining, and reconciling, the contradictory anomaly polarities reported in the literature. For a reference crest-survey geometry, defect signals follow Amax ≈ 45(D/z)²; combined with a voltage-dependent, site-calibratable error model and geostatistical background variability, this implies that pipe-class defects (≤3 m) are not reliably detectable in single-epoch data, while conventional anomaly thresholds can generate spurious indications in defect-free ground; degraded reaches (≥5–10 m) are resolved by a single survey or time-lapse — the latter also suppressing false positives — whereas pipe-class defects need active contrast or complementary methods. A reconstruction of the published Älvkarleby full-scale test embankment from published inputs alone reproduces the observed anomaly polarity for all six engineered defects, the detection ranking, and the necessity of multi-year averaging. We provide the phase diagram, detectability charts, a reciprocal-measurement site-calibration procedure and an openly archived workflow that allow the expected polarity and the detection limit of a planned survey to be computed before it is commissioned. The results suggest that, in freshwater clay-core dams, low-resistivity anomalies should generally be read as possible downstream consequences of leakage — wetting or fines redeposition — rather than as direct images of core damage.