Improving groundwater flow model conceptualisation and calibration with electrical resistivity tomography and self-potential methods

Developing a conceptual model for groundwater flow requires knowledge on the distribution of geological materials, which generally comes from geological observations on outcrops and boreholes, from the interpretation of hydraulic tests or from geophysical surveys. The identification of spatial structures in the subsurface, such as preferential flow paths created by fractured zones, is also critical in developing a reliable conceptual model but it is difficult to achieve. Geophysical methods have been widely used to map the subsurface distribution of geological materials. Recent developments in geophysics, such as the increased use of joint inversion of geophysical and hydrogeological data, have further allowed to quantify the hydraulic conductivity of geological materials. The objective of our work is to demonstrate that the electrical resistivity tomography (ERT) and the self-potential (SP) methods can improve both the conceptual model developed for groundwater flow systems and the calibration of the corresponding groundwater flow model. The use of the two geophysical methods, combined with a groundwater flow model, is presented for a fractured limestone aquifer. The self-potential method relies on passive measurements of the ambient electrical potential at ground surface or in boreholes. One of the mechanisms responsible for the measured signal measured is the transport of dissolved ions with groundwater flow. When this electrokinetic effect is the dominant contribution, the resulting signal is called the streaming potential and it contains information about groundwater fluxes that can be useful to calibrate groundwater flow models. The solution to the SP forward problem was added to the HydroGeoSphere model, which simulates 3D groundwater flow and solute transport in porous media, including fractured geological formations. With this addition, the model can calculate the self-potential signal associated with groundwater flow, given the distribution of Darcy fluxes resulting from the forward flow solution and the electrical resistivity that is, for example, outputted by ERT data inversion. Darcy fluxes are transformed into sources of electrical current by using the streaming potential coupling coefficient. This parameter can be measured either in the laboratory or in-situ from the self-potential signal between two locations where the depth of the water table is known, such as observation wells. We used here both ERT and SP to develop a conceptual model for groundwater flow in a typical carboniferous limestone syncline in South Belgium. The rolling topography in the investigated area results from a succession of calcareous valleys (synclines) and sandstone crests (anticlines). The calcareous synclines form aquifers that are very complex since they are highly fractured and even karstified. A typical calcareous syncline has a width of about 800 m and, using ERT, we could subdivide the syncline into zones of different hydraulic conductivity, based on the degree of fracturation. The zones are oriented along the axis of the syncline and their width ranges between 10 and 40 m. The ERT profiles showed that there is a highly conductive zone, in terms of electrical conductivity, near the syncline fold axis. That zone is interpreted as being highly fractured. Other conductive zones are located symmetrically along both flanks of the calcareous syncline, with respect to the syncline fold axis. The main flow direction is along the axis of the syncline, towards a nearby river. The SP raw signals also showed that, locally, there is a second flow component perpendicular to the axis of the syncline, with groundwater flowing from the flanks of the syncline towards the axis. The conceptual groundwater flow model developed here includes the zones identified with ERT, which were then incorporated into the numerical model. The SP signals were inverted with PEST to calibrate the hydraulic conductivity value of the different zones. HydroGeoSphere was therefore used to simulate first groundwater flow and then the associated self-potential signals in an iterative process. At the start of an iteration, HydroGeoSphere solves the groundwater flow equation given one particular set of hydraulic conductivities and calculates the resulting Darcy fluxes. These fluxes are transformed into sources of electrical current assuming that the electrokinetic effect is the dominant contribution of the SP signals. HydroGeoSphere then calculates the distribution of self-potential given the sources of electrical current and the distribution of electrical resistivity. The hydraulic conductivity values of the zones are then modified and the iteration continues until the model reproduces the measured self-potential signal.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.