DNAPL mapping by ground penetrating radar examined via numerical simulation

Abstract

Abstract Successful remediation of sites contaminated with dense non-aqueous phase liquids (DNAPLs) requires adequate characterisation of the volume and extent of the DNAPL source zone. Ground penetrating radar (GPR) has been proposed to characterise the subsurface distribution of DNAPLs; however, its effectiveness for real applications remains unproven. The objective of this study was to evaluate the potential of GPR to map realistic DNAPL-spill scenarios within heterogeneous subsurface environments, and to monitor the progress of subsequent remedial efforts. This was investigated by creating a novel link between two published numerical models: DNAPL3D-MT to generate a realistic DNAPL scenario, and GPRMAX to simulate a sequence of GPR surveys applied at the surface. A published volumetric mixing model permitted conversion of an evolving hydrogeological domain into a bulk permittivity domain. A field-scale, two-dimensional, surface release of a chlorinated solvent DNAPL into heterogeneous silty sand was employed as a demonstration case, including complete DNAPL remediation by dissolution and its mapping by time-lapsed 100 MHz surface GPR scans. Qualitative and quantitative interpretation of the results reveal that realistic GPR scans are simulated, generating a complex GPR response that is sensitive to both variations in DNAPL saturation and intrinsic subsurface heterogeneity. As a result, deconvoluting the response in the absence of a pristine site scan remains challenging. However, with the aid of newly developed GPR analysis tools presented here, including the combination of ‘sequential scan subtraction’ with ‘normalized radar trace and radar section sum of squares’, changes in DNAPL distribution (mass reduction and remobilization) are demonstrated to be quantifiable. Thus, it is concluded that periodically monitoring the time-lapsed remediation of the source zone with GPR is particularly promising.