Part 1: Calibrate a MODFLOW model for steady-state conditions using the river package to model the East River. Use the MODFLOW-NWT version of MODFLOW to help with model convergence. There are 6 monitoring wells close by (MW-1 through MW-6) that should be used as calibration targets (see below). Use the hydraulic conductivity of the sand aquifer and the conductance of the river bed sediments as the two adjustable parameters. Use a vertical anisotropy of 10:1. The recharge is constant at 8 inches per year, evapotranspiration is constant at 0.010 ft/d with an extinction depth of 6 ft. The model grid should have enough vertical discretization to account for the location of well screens. You can neglect the effect of the three residential drinking water wells because their extraction rate is only 200 gal/d per well. From Part 1, provide the following:

• Map of the model grid and a summary of all pertinent model information.
• Contour maps of steady-state heads (using linear contours) in plan view for Layer 1 and whatever is your lowermost layer. Use 0.5 ft contour intervals.
• Contour maps of heads in profile. Use the “Front View” sowing heads in the row that contains MUN-1. Use 0.5 ft contour intervals and z-exaggeration/magnification of 20.
• As part of your calibration, report the values for K and riverbed conductivity and the final RMSE and nRMSE. Perform a sensitivity analysis for K and the riverbed conductivity. Prepare tables of nRMSE and plots to summarize the results of the analysis.

Questions to answer: Which parameter is the model most sensitive? Which process is most responsible for water leaving the model? The river or evapotranspiration?

Part 2: Model the spill of brine using MT3D using constant concentrations cells for chloride equal to 10,000 mg/L. The spill location is provided on Figure 1. The approximate width and length of the spill is 200 ft by 200 ft.

When setting up your MT3D model ….

• Use the “3rd order TVD ultimate” package to solve the advection part of the transport equation explicitly.
• Use the GCG solver package to solve the dispersion and source/sink part of the transport equation implicitly. Use the default GCG solver options.

To model the brine, use an effective porosity of 0.15 and longitudinal dispersivity equal to 10 ft. Assume the transverse horizontal and transverse vertical dispersivities are one-tenth and one-hundredth of the longitudinal dispersivity. Note: in GMS “effective porosity” is entered in the Basic Transport Package and is just called “porosity”.

Run the model for 18,250 d (approximately 50 years). Use 1 stress period with 50 time steps, one for each year. Set the Output Control to save concentrations every 365 d. After running MT3D, set contours to Color Fill, use 100 mg/L as the minimum contour, use red for the max concentration, blue for the min concentration and use 8 to 10 contours.

From Part 2, provide the following:

• Contour maps of brine in plan view at various times.

Questions to answer: Where does the plume go? Does the plume move vertically? Are the residential wells impacted? Explain.

Part 3: Re-run MODFLOW and MT3D with the municipal well (MUN-1) extracting water for the following three cases.

Hint: if your cells go dry in the top layer (as indicated by red triangles or missing cells), you may need to lower the elevation of the top-most layer and rerun the model.

From Part 3, provide the following:

• Contour maps of steady-state heads (using linear contours) in plan view for Layer 1 and whatever is your lowermost layer. Use 0.5 ft contour intervals.
• Contour maps of heads in profile. Use the “Front View” sowing heads in the row that contains MUN-1. Use 0.5 ft contour intervals and z-exaggeration/magnification of 20.
• Contour maps of brine in plan view at various times.

Questions to answer: What effect does the pumping well have on heads and flow direction in the aquifer?

What effect does the pumping well have on the plume? On the concentration of the residential wells over time? What is the impact to the water quality of MUN-1? Is the result what you would expect?

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Part 4: Model the effect of removing 90% of the source area after 50 years. We will do this by changing the flow model to transient and adding a second stress period for another 50 years.

Stress Period 1: 0 to 18,250 d (50 time steps), constant concentration of 10,000 mg/L, MUN-1 pumping at 15,000 ft3/d

Stress Period 2: 18,250 to 36,500 d (50 time steps), constant concentration of 1,000 mg/L, MUN-1 not-pumping.

Important:

• Make sure to map the head solution from Part c) as the initial/starting heads.
• Make sure to specify specific yield = 0.1 and specific storage = 1×10-6 /ft in all grid cells
• Make sure the second stress period does NOT use ICBUND to specify the chloride constant concentration cells (use the source-sink mixing package). Set the Output Control to save concentrations every 365 d.

From Part 4, provide the following:

• Contour maps of brine in plan view at various times.

Questions to answer: Do the residential wells “clean up”? If so, how long will it take? What is the response in the residential wells if there is 100% removal of the source area?

Present your findings in a modeling report to the town of Hill Valley.