Three-dimensional modeling was conducted with layered models to further develop the conceptual model of CT distribution in the vertical and lateral direction beneath the 216-Z-1A tile field and 216-Z-18 cribs and to investigate the effects of soil vapor extraction (SVE). Base case and sensitivity analysis simulations considered migration of dense, nonaqueous phase liquid (DNAPL) consisting of CT and co-disposed organics in the subsurface beneath the two disposal sites as a function of the properties and distribution of subsurface sediments and of the properties and disposal history of the waste. Simulations of CT migration were conducted using the Subsurface Transport Over Multiple Phases (STOMP) simulator.Simulation results support a conceptual model for CT distribution where CT in the DNAPL phase is expected to have migrated primarily in a vertical direction below the disposal trench. None of the simulations predicted that CT in the DNAPL phase would move across the water table below the 216-Z-18 site. Movement of CT in the DNAPL phase across the water table below the 216-Z-1A site was only predicted in simulations with smaller disposal areas and larger volumes, compared to the base case simulation, and in isotropic porous media. Because uncertainties in disposal area and volume exist, movement of CT in the DNAPL phase across the water table in the subsurface below the 216-Z-1A site should be considered as a possibility. However, even if DNAPL moved across the water table in the past, there may not currently be a DNAPL phase in the groundwater beneath the 216-Z-1A site because of dissolution. Results also show that the Hanford 1a geologic unit, located just beneath the 216-Z-1A and 216-Z18 disposal areas, retains more CT DNAPL within the vadose zone during infiltration and redistribution than other hydrologic units. During simulated SVE operations, CT in this unit remained in the subsurface while DNAPL in other layers was effectively removed. Additional characterization of the Hanford 1a unit below the two disposal sites would provide valuable information about the quantity of DNAPL phase CT remaining in the vadose zone. A significant amount of the disposed CT DNAPL may have partitioned to the vapor phase and subsequently into water and sorbed phases. As for the 216-Z-9 site, it is predicted that any continued migration of CT from the vadose zone to the groundwater is likely to occur through interaction of vapor phase CT with the groundwater and not through continued DNAPL migration. The results indicated that SVE appears to be an effective technology for vadose zone remediation, but additional effort is needed to improve simulation of the SVE process through an enhanced understanding of rate-limited volatilization. A total of 34 three-dimensional simulations have been conducted based on a layered EarthVision™ geologic model, which is an interpretation of available geologic data. These simulations consist of one base case simulation and 33 sensitivity analysis simulations. These simulations examined the infilt...