The MODPATH program computes particle locations and travel times in three dimensions based on advective flow in a uniformly porous medium. MODPATH can track particles forward in time and space in the direction of ground-water flow, or backward toward recharge areas. Physical, chemical, and biological processes that attenuate chemical constituents in ground water are not considered, and the dissolved contaminant is assumed to not appreciably alter the density of the ground water. MODPATH cannot be used to predict solute concentrations.
The cell-by-cell flow terms from the calibrated steady-state MODFLOW model were used as input to MODPATH. Ground-water travel time in the shallow aquifer system was simulated using a uniform value of effective porosity for each hydrogeologic unit represented in the model. Equation 5 introduced in the preceding discussion quantifies the relation between ground-water-flow velocity, the volume of water moving through the model, and effective porosity of the porous media. The porosity values reported for core samples of the hydrogeologic units at NSA Memphis are total porosity. For coarse-grained, unconsolidated sediments, such as those forming the A1 aquifer, effective porosity will approach the total porosity, but the effective porosity will be somewhat less.
A particle-tracking analysis of ground-water flow in the shallow aquifer system at NSA Memphis was performed for each of three scenarios: (1) effective porosity was assumed to approximate the total porosity values reported for core samples of the hydrogeologic units, (2) minimum effective porosity was used based on reported values in the literature for the type of sediments present at NSA Memphis (Freeze and Cherry, 1979), and (3) intermediate porosity values were used between these endpoints. The residence time of water within the A1 aquifer was simulated by seeding one particle on the lower faces of each active cell of layer 1 and performing a backward-tracking analysis. For effective porosity values ranging from 20 to 33 percent, typical ground-water-flow velocities ranged from about 15 to 25 ft/yr, and average residence times ranged from about 645 to 1,000 years. The variability in the results of particle-tracking analyses (table 11), theoretically, should encompass the range of potential ground-water travel times at NSA Memphis.
Ground-water-flow directions in the A1 aquifer were simulated by seeding the upper faces of layer 1 cells and performing a forward-tracking analysis. Most of the particles traveled for relatively short distances in layer 1 before they were either "captured" by a river node, entered a deeper layer, or exited the model through one of the boundary cells (fig. 23). Most of the ground water (79 percent) moves vertically through the model. The highest rates of vertical movement are within the western half of the study area where a window in the Cook Mountain confining unit was simulated, and under the hypothesized buried river valleys where a window in the Cockfield confining unit was simulated. Vertical flow is also accelerated in the area of the simulated fault, but because the fault is a long narrow feature, a smaller area is affected compared to the windows in the confining units and less water is transmitted than through the windows.
Intermediate porosity values (table 11) were used for particle-tracking analyses of contaminant migration at NSA Memphis. Ground-water-flow paths and times-of-travel within the A1 aquifer were simulated at three sites: (1) the former N-6 hangar area and (2) the grassy area near SWMU 7, both within the Northside AOC; and (3) at SWMU 2 (fig. 24). The contaminants detected within the A1 aquifer are estimated to first have been used about 40 years ago in the mid- to late 1950's. The advective transport of contaminants in the A1 aquifer was simulated by seeding the upper face of the appropriate cell(s) in layer 1 and using forward-tracking analyses. This approach simulates the introduction of contaminants into the A1 aquifer through leakage from the overlying loess or alluvium and subsequent advective transport within the aquifer. Particle locations were plotted after 40 years of travel, simulating a worst case scenario in which the contaminant entered the aquifer as soon as the contaminant began to be used. For simplicity, the assumption was also made that a single release of contaminants occurred.
The location of suspected plumes of contaminants were noted within the Northside AOC (fig. 25). Particle-tracking analyses of the former N-6 hangar area (fig. 26) indicate that ground water moves north-northwest from the suspected source area for about 4,000 feet, and then flows vertically downward along the simulated fault towards layer 3. The average time-of-travel to layer 3, the simulated Memphis aquifer, was about 280 years (table 11). The simulated flow path and travel distance after 40 years compares favorably with the identified extent of migration of the hypothesized plumes (fig. 26).
Particle-tracking analyses of the grassy area near SWMU 7 (fig. 26) indicate that ground water moves north-northwest from the suspected source area until it enters the area-of-influence of the simulated fault and moves downward towards layer 3. The average time-of-travel to layer 3 was about 380 years (table 11). The simulated flow path and travel distance after 40 years also compares favorably with the identified extent of migration of the hypothesized plumes (fig. 26).
The close agreement between the estimated locations of the hypothesized contaminant plumes and the distance traveled in 40 years predicted by the particle-tracking analyses for the two sites within the Northside AOC indicates that the estimates of hydraulic conductivity and effective porosity of the A1 aquifer are reasonably accurate for these areas. Based on the results of particle-tracking analyses, the potential for contaminants to reach the Memphis aquifer in the next 100 years is negligible.
Particle-tracking analyses of the SWMU 2 area (fig. 27) indicate that ground water moves rapidly towards Big Creek Drainage Canal. Out of 40 particles tracked, 39 were removed from the model by river nodes, which simulates ground-water discharge from the A1 aquifer to Big Creek Drainage Canal. The average time-of-travel was about 26 years; however, at present, there is no map of the extent of contaminant migration at SWMU 2 to compare to particle-tracking simulations.
The calibrated flow model and the MODPATH program were not used to evaluate remedial designs at NSA Memphis. The results of the calibrated flow model and MODPATH analyses may simulate the expected direction and extent of contaminant migration if no remedial actions are undertaken.