Suivi spatio-temporel annuel des écoulements d’eau et de contaminants dans une infrastructure verte expérimentale par imagerie géoélectrique

Abstract

The intensification of extreme precipitation, combined with increasing urbanization and surface imperviousness, exerts growing pressure on urban stormwater management systems. In this context, Green Infrastructure (GI) are increasingly deployed to reduce peak hydraulic loads and promote passive infiltration. However, their seasonal hydrogeochemical behavior remains poorly documented, particularly in cold climates where the use of de-icing salts represents a significant source of contaminants. This lack of knowledge limits design optimization and the assessment of their impact on subsurface infrastructures and receiving environments. This thesis presents the monitoring of subsurface geoelectrical dynamics using time-lapse cross-borehole electrical resistivity tomography (TL-ERT) within and around a GI located in Laval (Quebec) aiming to image the hydrogeochemical processes. This geophysical setup is complemented with conventional instrumentation (piezometers, water content and temperature probes) along with geochemical and isotopic monitoring. This approach helps characterize infiltration dynamics across the entire profile, whereas traditional methods provide only point scale information. The project began with the finalization of electrical piezometers for on-site installation. Validation of the installation’s proper functioning was conducted during the first geoelectrical imaging in april 2024, allowing for the start of continuous monitoring in fall of the same year. This monitoring was subsequently automated through the deployment of a custom system developed in-house based on the OhmPi project, which has been providing multiple daily measurements since April 2025. Data processing is based on geophysical inversion incorporating a temperature correction of the conductivity models to ensure long-term data comparability. Finally, the use of Time Series Clustering (TSC) methods enabled the synthesis of the entire dataset to identify hydrogeological units and assist in building a conceptual model of the infrastructure over a full seasonal cycle. The results highlight complex seasonal dynamics, strongly influenced by the spring melt. The observed variations in electrical conductivity are initially driven by the movement of chlorides from de-icing salts, before becoming primarily governed by fluctuations in water content starting from late summer onward. This study demonstrates that borehole TL-ERT, supported by TSC as a synthesis tool, accurately captures the behavior of the GI at scales that are inaccessible to conventional methods. This approach could be combined with multiphysics modeling to derive key engineering parameters.

Date	15 May 2026
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Adrien Dimech (Supervisor) & François Duhaime (Co-supervisor)