Integration of hydrological and geophysical data beyond the local scale: Application of Bayesian sequential simulation to field data from the Saint-Lambert-de-Lauzon site, Québec, Canada

• Hydrological and geophysical data were collected at the St-Lambert site, Canada.
• A two-steps Bayesian Sequential Simulation approach is applied to this database.
• We generated stochastic realizations of hydraulic conductivities.
• The results are compared to borehole observations.
• The results are also compared to independent predictions based on a Relevant Vector Machine approach.

SummaryAdequate characterization of aquifer heterogeneity is critically important for the sustainable use, protection, and remediation of groundwater resources. The combined use of hydrological and geophysical measurements is arguably the most effective means of achieving this objective. In this regard, significant progress has been made on the quantitative integration of geophysical and hydrological data at the local scale. However, the extension of such approaches to larger, more regional scales remains a major research challenge. In this paper, we demonstrate the application of a recently developed regional-scale hydrogeophysical data integration approach, which is based on Bayesian sequential simulation, to a field database from Quebec, Canada consisting of low-resolution, surface-based geoelectrical measurements as well as high-resolution direct-push and borehole-based measurements of the electrical and hydraulic conductivities. The results of our study, which involved the integration of data along an approximately 250-m-long survey line, confirm that this novel methodology, with suitable adaptation, is fully applicable to field data and has the potential of providing realistic estimates of the spatial distribution of hydraulic target parameters at the regional-scale. Equally importantly, through the generation of multiple stochastic realizations, the methodology allows for quantitative assessment of the uncertainty associated with the inferred subsurface models, which in turn is essential for interpreting subsequent predictions of the flow and transport characteristics of the studied region.