Advancing Electrical Resistivity Tomography as an Environmental Monitoring Tool in Seasonally Frozen Ground: Linking Lab and Field Scales

Date
2021-03-15
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
When time-lapse electrical resistivity tomography (ERT) is used as an environmental monitoring tool, the effects of temperature variability should be taken into account to avoid erroneous interpretations of subsurface processes. In regions where the near-surface freezes seasonally, removing the effects of temperature is a non-trivial task. One limiting factor is the poorly understood relationship between electrical resistivity and temperature under frozen conditions. A laboratory experiment was conducted that quantified this relationship and related resistivity to unfrozen water content and fluid resistivity using a modified version of Archie’s equation. A second limiting factor is that the standard inversion has a limited ability to resolve sharp boundaries and large contrasts in resistivity, like those seen between frozen and unfrozen regions. A synthetic time-lapse ERT modelling study showed that because standard ERT inversion techniques were unable to accurately recover the resistivity of the frozen surface layer, temperature corrections applied to these models performed poorly. In this synthetic experiment, modifications to data acquisition (burying electrodes or reducing electrode spacing) and regularization strategy (increasing lateral smoothness, reducing regularization across boundaries, or using an L1 model misfit norm) did not appreciably improve the outcomes of temperature corrections in partially frozen ground. A hybrid inversion strategy was developed to incorporate prior information about the geometry of the frozen layer in the inversion. The hybrid inversion used a parametric approach with only depth and a single resistivity to describe the frozen layer, while the rest of the model space was described with a large number of voxels. A synthetic experiment showed that the hybrid inversion improved resolution of a frozen surface layer and features beneath it compared to a standard smooth inversion. The improved resistivity models resulted in lower errors in the temperature-corrected data. However, resolution of the frozen layer was still limited by the information content of the data, regardless of inversion strategy. Altogether, this study improves our understanding of the physical processes and relationships that govern resistivity at subzero temperatures, highlights limitations of standard data processing strategies, and demonstrates the efficacy of a hybrid inversion approach for ERT data collected in partially frozen ground.
Description
Keywords
Electrical resistivity tomography, Frozen ground, Geophysical inversion
Citation
Herring, T. (2021). Advancing Electrical Resistivity Tomography as an Environmental Monitoring Tool in Seasonally Frozen Ground: Linking Lab and Field Scales (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.