RESUMEN
This study presents a numerical model of a large aqueous phase plume of a mixture of chlorinated solvents that has penetrated the fractured dolomitic bedrock near Smithville, Ontario, Canada several decades ago which, since 1989 has been hydraulically controlled by a pump-and-treat remediation system. A multiphase compositional model CompFlow is first applied to simulate the migration of DNAPLs in a discretely fractured porous medium with hydrostratigraphy representing the Smithville site. Results from CompFlow are used to estimate the pure-phase DNAPL distribution in the discrete fractures and rock matrix. Next, CompFlow results are employed to define the source term for a regional-scale transport simulation using HydroGeoSphere (HGS) by treating the layered, fractured dolomitic rocks as an equivalent porous continuum. Transport simulations are conducted both prior to and after the operation of the pump-and-treat system. Results reveal that considerable agreement with the observed mass removal data and TCE plume can be achieved by modifying the composition of the DNAPL source and by reducing the hydraulic conductivity (K) in the source zone region to account for preferential flow around it. Our transport model results support the conceptual model of TCE contamination which posits a mixed source (2 to 4%) of DNAPL with limited contact with actively flowing groundwater that is undergoing equilibrium dissolution. Model results also reveal that the pump-and-treat system has neither been effective in stabilizing the plume nor removing a significant amount of contaminant mass, but that the stability of the plume is instead due to first-order degradation.
Asunto(s)
Simulación por Computador , Restauración y Remediación Ambiental/métodos , Movimientos del Agua , Canadá , Agua Subterránea , Contaminantes Químicos del Agua/químicaRESUMEN
A regional flow and transport model is used to explore the implications of significant variability in Pleistocene and Holocene climates on hydraulic heads and (14)C activity. Simulations involve a 39 km slice of the Death Valley Flow System through Yucca Mountain toward the Amargosa Desert. The long-time scale over which infiltration has changed (tens-of-thousands of years) is matched by the large physical extent of the flow system (many tens-of-kilometers). Estimated paleo-infiltration rates were estimated using a juniper pollen percentage that extends from the last interglacial (LIG) period (approximately 120 kyrbp) to present. Flow and (14)C transport simulations show that groundwater flow changes markedly as a function of paleoclimate. At the last glacial maximum (LGM, 21 kyrbp), the recharge to the flow system was about an order-of-magnitude higher than present, and water table was more than 100 m higher. With large basin time constants, flow is complicated because hydraulic heads at a given location reflect conditions of the past, but at another location the flow may reflect present conditions. This complexity is also manifested by processes that depend on flow, for example (14)C transport. Without a model that accounts for the historical transients in recharge for at least the last 20,000 years, there is no simple way to deconvolve the (14)C dates to explain patterns of flow.