The groundwater flow model used for this demonstration is described on the web page titled "Surficial Sediment MODFLOW Model." This MODFLOW model simulates the movement of groundwater downslope in surficial sand and gravel over impervious bedrock. The MODFLOW model produced a computer file containing groundwater flow rates across cell faces throughout the model (CBC file). The cells are volumes in the grid that comprises the simulated sand and gravel layer. This grid is shown in Figure 1.
The MODPATH model used for this demonstration is described on the webpage titled "Groundwater Flow Path Model MODPATH." MODPATH uses the CBC file and other files generated by MODFLOW to calculate successive locations of a particle moving with the groundwater in the direction of its flow. The particle track begins at a selected source location, and travels to a discharge location. Output from the model includes the locations of the particle at selected times since its travel began at the source location. Figure 1 shows a particle track as dots at one-year increments. This particle track is Path 4 in Figure 1 of the aforementioned MODPATH web page.
The MT3D solute transport model also uses the CBC file and other files generated by the same MODFLOW model. MT3D calculates concentrations of a solute (contaminant) in the groundwater when the contaminant is added at a selected source location. The calculated concentrations include the effect of dispersion of the contaminant due to lateral movement of the water induced by flow around the grains of the subsurface material. The concentrations contoured in Figure 1 of this web page are concentrations that result from inflow of contaminated water (contaminated recharge) at the same source location used for the particle track. The spread of the contamination illustrates the effect of dispersion, inhomogeneity in the subsurface material, and variation of saturated thickness of the sand and gravel (thickness of sand and gravel below the water table). The concentration input at the source was 1,000 milligrams per liter. The size of the source area was 1000 feet by 1000 feet. The solute was treated as not subject to adsorption onto the sand and gravel. The computed concentrations are for the steady-state condition, meaning that sufficient time was elapsed so that the plume is fully developed and concentrations in the sand and gravel are not changing. The amount of contaminant flowing into the system is the same as the amount flowing out.
The results shown in Figure 1 suggest that in a complex groundwater system, a contaminant plume will not necessarily be distributed symmetrically about the flow path from the source. However, there is some error1 in the calculated concentrations and the resultant contour lines. Evidence for error is the presence of contaminant upstream from the source location. Some error is inevitable in this type of calculation. Nevertheless, the results are indicative of the behavior of the real system.
The error is not due to lack of convergence and is not reflected in the mass balance. The solver used was General Conjugate-Gradient (GCG) with the Modified Incomplete Cholesky (MIC) preconditioner. The Convergence criterion was 1e-5, which was reached at 46 iterations. The mass balance error was 5.37e-5.
Posted April 9, 2021
