Incorporating electrical sounding survey into geostatistical electrical resistivity tomography for high-resolution characterization of karst aquifer

Abstract

Aquifer characterization is vital for researching hydrogeology and engineering safety in karst regions. However, the significant heterogeneity of karst aquifers poses great challenges to tomography methods. Our study introduces a cost-effective framework to enhance geostatistical electrical resistivity tomography (GERT) under sparse observations by integrating electrical sounding (ES) methods. The effectiveness of this approach is demonstrated by conducting sandbox experiments. It shows that the discrete apparent electrical conductivity points derived by the ES survey help distinguish between matrix and conduits. Meanwhile, the GERT can effectively characterize the heterogeneity of karst aquifers with adequate nonredundant observations. Subsequently, apparent electrical conductivity is employed to derive prior information (initial guess field, variance, mean, and correlation length) to improve estimation under sparse observation patterns. However, the raw apparent electrical conductivity point data or Kriging field does not significantly enhance GERT estimation due to discrepancies between apparent and actual electrical conductivity. Moreover, the Kriging field is converted to the conductivity level acquired by the unconditional equivalent homogeneous approach via the feature scaling method. Afterward, observations required for GERT are reduced by 73.03% at a similar imaging resolution by integrating the prior information of the scaled Kriging field. This study highlights the potential of ES information for improving the performance of GERT in karst terrains.