Pollutant Concentration Reconstruction in Porous Media from Localized Measurements

Abstract

Reliable environmental monitoring and groundwater protection heavily rely on the ability to reconstruct full pollutant concentration fields from sparse, partial field measurements. In porous media, dissolved contaminants are transported by groundwater flow while undergoing hydrodynamic dispersion and complex chemical or biological reactions. To address practical contamination challenges, this work investigates the reconstruction of pollutant distributions using water samples collected at discrete locations in unsaturated or saturated soils via sampling tubes equipped with ceramic tips. While these suction samplers provide critical local data, they inherently capture only a fraction of the subsurface environment, leaving wide areas unmonitored. Motivated by this real-world monitoring procedure, we propose steady and time-dependent advection-diffusion-reaction models to simulate contaminant transport. Our main contribution is a generalized unique continuation result demonstrating that, under appropriate physical and mathematical assumptions, two concentration fields coinciding on a non-empty observation region must match throughout the entire space-time domain, even when originating from different source configurations. We investigate this problem from both theoretical and numerical perspectives, backing our mathematical uniqueness results with numerical simulations that validate the efficiency and practical utility of our approach. The theoretical proofs combine regularity properties away from source singularities with unique-continuation principles for parabolic and elliptic partial differential equations, providing a robust framework for environmental risk assessment and site remediation.