Hydrogeophysical Inversion of Time‐Lapse ERT Data to Determine Hillslope Subsurface Hydraulic Properties

• Published in: Water resources research 2022-04, Vol.58 (4), p.n/a
• Main Authors: Pleasants, M. S., Neves, F. dos A., Parsekian, A. D., Befus, K. M., Kelleners, T. J.
• Format: Article
• Language: English
• Subjects: brute force calibration; Bulk density; Calibration; Dynamics; Electrical resistivity; electrical resistivity tomography; grid search; Groundwater; Groundwater levels; hillslope hydrology; Hydraulic properties; Hydraulics; hydrogeophysical inversion; Hydrologic models; Hydrology; Inversions; Mathematical models; Methods; Modelling; Moisture content; Mountains; Objective function; Optimization; Parameter sensitivity; Parameters; Physics; Properties; Subsurface water; Tomography; Vertical flow; Vertical mixing; Water content; Water flow
• ISSN: 0043-1397; 1944-7973
• DOI: 10.1029/2021WR031073

Time‐lapse electrical resistivity tomography (ERT) data are increasingly used to inform the hydrologic dynamics of mountainous environments at the hillslope scale. Despite their popularity and recent advancements in hydrogeophysical inversion methods, few studies have shown how time‐lapse ERT data can be used to determine hydraulic parameters of subsurface water flow models. This study uses synthetic and field‐collected, hillslope‐scale, time‐lapse ERT data to determine subsurface hydraulic properties of a two‐layer, physics‐based, 2‐D vertical flow model with predefined layer and boundary locations. Uncoupled and coupled hydrogeophysical inversion methods are combined with a fine‐earth fraction optimization scheme to reduce the number of parameters needing calibration and interpret the influence of the hydraulic parameters on the hydrologic model predictions. Inversions of synthetic ERT data recover the prescribed fine‐earth fraction bulk density to within 0.1 g cm−3. Field‐collected ERT data from a mountain hillslope result in hydrologic model dynamics that are consistent with previous studies and measured water content data but struggle to capture measured groundwater levels. The uncoupled hydrogeophysical inversion method is more sensitive to changes in hydraulic parameter values of the lower hydrologic model layer than the coupled hydrogeophysical inversion method. Time series of minimum objective function value simulations indicate that periodically collected ERT data may recover hydraulic parameters to a similar level of uncertainty as daily ERT data. Using simple hydrologic model domains within hydrogeophysical inversions shows promise for providing reasonable hydrologic predictions while maintaining relatively simple calibration schemes and should be explored further in future studies. Key Points Synthetic resistivity data resolve subsurface hydraulic parameters in uncoupled and coupled hydrogeophysical inversions Field‐collected resistivity data result in hydrologic model dynamics consistent with previous studies and measured data Combining brute force hydrogeophysical inversion and fine‐earth fraction optimization provides a robust framework for parameter estimation