SUMMARYApproximately 2,000 m3 (115,000 gallons) of 1 iquid containing radioactive and chemical wastes leaked from the 241-T-106 single-shell tank at the -U.S. Department of Energy's Hanford Site in southcentral Washington in 1973.The leak discharged into the unsaturated, coarse-grained sediments of the Hanford formation which underlie the base of the tank. The extent of the i contaminant plume in the vadose zone was estimated in 1973 and 1978 by gamma spectrometry in dry wells drilled to monitor the leak. The leak provides an excellent opportunity to test the applicability of the PORFLO-3 Version 1.0 computer code for studying flow and transport in the vadose zone.
IThe PORFLO-3 computer code was used to study plume migration for 1 0 6~~ and 137~s. The flow and transport properties of the soils through which the plume has migrated are critical input data for the model but are not available. Consequently, information from a catalogue of Hanford Site soil properties was used. The transient magnitudes and locations of the plume were simulated in three dimensions.Using viscosities and hydraulic conductivities for the tank fluid that are equal to that of water, the plume of 1 0 6~~ simulated for 1973 was larger than that which was measured. By reducing the hydraulic conductivity for saturated conditions in the vertical direction by one-half, the dimensions of the simulated 1 0 6~~ plume approximate those of the measured plume. Such reduction is justified because of the uncertainty inherent in the hydraulic conductivity data and because the tank fluid is more viscous than water. The plume of 1 3 7~s simulated for 1973 approximates that which was measured, assuming a reasonable distribution coefficient to account for the sorption of 1 -cesium by minerals of the Hanford formation.
. -Using the reduced vertical hydraul ic conductivity, the migration of 1 0 6~~ and 1 3 7~s was simulated for the time between 1973 and 1990. (The gamma spectrometry measurement in 1978 suggests that the 1 0 6~~ did not move after 1973.) The plume simulated for 1978 and 1990, however, does not show this behavior; it appears to have expanded significantly since 1973. Because 1 0 6~~ may have a small sorption coefficient, a simulation was made that assumed a distribution coefficient of 0.5 for ruthenium. Use of this coefficient, however, did not result in a stationary 1 0 6~~ plume; the simulated plume continued to expand after 1973. This value is reasonable based on information in the literature. How---ever, the use of the distribution coefficient did not accurately reproduce the lateral spreading of the observed plume, and the simulated plume expanded 4 .significantly after 1973.
I -The simulated plume of cesium for 1978 and 1990 did not move significantly from that measured and simulated for 1973. This static condition likely resulted from the sorption of cesium by minerals of the Hanford formation.
bSimulations of the 241-T-106 tank leak indicate that PORFLO-3 Version 1.0 should be able to represent the transient behavior of a threedimension...