WSP04273 - Laserfiche WebLink

Home Browse Search

C"').. lJ'j. C'J N c.:.. c:,;, Cross-Sectional Model of Flow and Solute Transport Cross-sectional models were developed to determine how the proposed slurry wall and how aquifer and solute properties could affect the migration and distribution of salts in the shallow aquifer of the Whitney area. Under the proposed strategy. the slurry wall was intended to inhibit flow near the base of the aquifer. Flow would occur primarily near the top of the aquifer and would also become fresher over time because most of the flow through the detention basin would occur well above the contact between the potentially salt-laden Muddy Creek Formation and the overlying alluvial aquifer, The effectiveness of the slurry wall in decreasing the dissolved-solids concen~~(}ns (and load$) leaving the detention basin was examined in this cross-sectional model by assuming that the incoming ground water initially contained no dissolved solids buI that the Muddy Creek Formation contained dissolvable salts. Under these conditions. albeit somewhat unrealistic, the effect of the proposed slurry wall could be readily observed, because concentration gradients from dissolving salts in the Muddy Creek Formation would be pronounced. Simulations were made to determine the effects of differing the dispersivity, the ratio of horizontal to vertical hydraulic conductivity (anisotropy), the volumetric ground- water inflow, and the aquifer thickness. The results of these tests were used to evaluate the effectiveness of the proposed slurry wall and were used. more specifically, to detennine the aquifer or transport property that would be most detrimental to the success of the proposed s~tegy. The conceptual model extends along segment X- Y of section W-Z in figures 4 and IIA. The top of the Muddy Creek Formation is just above the bottom of the model and contains a source of dissolvable salts; the water table represents the model top. Ground-water inflow enters the model area from the left upgradient boundary, whereas the slurry wall, with an outflow area at its top. represents the right boundary. Surface-water inflow or evapo~nspiration were not included in this simulation, so that they would not complicate the interpretation of the simulation results of the individual aquifer and solute properties. The entire aquifer was assumed to have a constant hydraulic conductivity of 30 ftld in all the simulations, Constructing the MathematU:al Model A finite-element rectangular grid having 44 rows and 20 columns connected by 945 nodes was used to simulate ground-water flow and solute transport (fig. 15). Grid spacing was constant with elements having a length (in the direction of flow) of 164 ft and a thickness (vertical dimension) of 0.82 ft The total length of the modeled region extends 3,280 ft upgradient from the slurry wall (which is the downgradient model boundary). The total thickness of the modeled region is 36 ft, with the upper 30 ft representing the alluvial aquifer and the bottom 6 ft representing the top part of the Muddy Creek Formation. Element thicknesses were selected to be small, to accurately simulate concentration gradients even when vertical mixing was on the order of molecular diffusion; nonetheless, some numerical dispersion may occur. Numerical dispersion is a numerical error that tends to artificially increase mixing due to the type of grid network used for the study. The finer the network, the more accurate the results and the less the numerical dispersion, Because the degree of vertical mixing was important in detennining the effectiveness of the slurry wall, a small element thickness was chosen. The boundary conditions included a constant solute source at the 21 nodes representing the base of the alluvial aquifer (topmost part of the Muddy Creek Formation; fig. 15). Each node was arbitrarily specified to have a concen~tion of 37.500 mgIL. which is similar to that of seawater. A no-flow boundary was used to represent the water table, which was virtually at land surface. This was considered to be a good approximation because heads remained fairly constant at the water table except immediately upgradient from the slurry wall. A constant aunospheric pressure was applied to the upper right node (node 945) because at least one specified pressure-boundary node needs to be incorporated to correctly calibrate the pressures at the remaining nodes in the model. The boundary alon5 the left (upgradient) side of the grid was specified as having a constant fluid-source flux of 4.63xlO-5 ft Is, This flux is equivalent to taking a volume of unit width along segment X- Y of section W-Z from the same area used to estimate the total -34-