Limitations and considerations for electrical resistivity and induced polarization imaging of riverbed sediments: Observations from laboratory, field, and synthetic experiments

Abstract

Characterization of riverbed sediments is important for understanding groundwater (GW) and surface water (SW) interactions, and their consequent implications for ecological and environmental health. There have been numerous studies using geoelectrical methods for GW-SW interaction studies; however, most applications have not focused on obtaining quantitative information. For instance, although numerous laboratory studies highlight the relationship between geoelectrical properties and relevant parameters (e.g. specific surface area, hydraulic conductivity, and cation exchange capacity), such relationships are not commonly applied to field-scale studies. Furthermore, in addition to the spatial resolution obstacles typically present when applying petrophysical models to field data, geoelectrical data from aquatic environments have complications arising from the presence of a conductive water column overlying a resistive bed. Inadequate consideration of these complications may further preclude the reliable use of such petrophysical models. In this work, laboratory measurements, synthetic modeling, and field measurements were conducted in a third-order river where the riverbed comprises alluvial gravel and underlying red sand. A strong relationship (R2 = 0.72) between imaginary conductivity and specific surface area was observed, and laboratory results were comparable to previous studies. It was demonstrated through synthetic modeling that river stage and channel width, regularization across the river-riverbed interface, and incorrect constraints of both the river conductivity and river stage can have varying influence on inverted geoelectrical images. Reliable geoelectrical images require a priori definition of river stage and conductivity, however inversion constraints using incorrect a priori values result in misleading artifacts. The conductivity image obtained from the field data in this work appeared to reflect the geoelectrical structure anticipated from the laboratory data; however, the phase angle image did not. Although this study focused on riverbed characterization, findings here demonstrate common pitfalls of inversion of aquatic-based geoelectrical data. Primarily, they highlight that synthetic modeling ought to be used to alleviate any uncertainty in the interpretation of geoelectrical models before predictions about GW-SW interactions can be made.