Skip to Main Content U.S. Department of Energy


Water Movement Near the Columbia River

A two-dimensional model that simulates flow pathlines in a vertical cross section oriented perpendicular to the Columbia River has been developed for a location at the Hanford Site. Hydraulic head data from wells and the adjacent river were used to calculate flow direction and velocity in hourly increments for an entire seasonal cycle. River stage cycles extend through a range of several meters, thus exerting a strong influence on water movement in the zone of interaction. By including a fluctuating river stage at the river boundary (center of channel), the model showed that landward of and beneath the shoreline, flow pathlines within the aquifer are deflected downward. The region immediately beneath the shoreline is strongly influenced by river water that infiltrates during high river stage. On the river side of the shoreline, groundwater discharges upward into the river channel, with pathlines converging in the riverbed relatively close to shore. If the model is run assuming a constant, average river stage, these movement features are not represented, thus demonstrating the need to include transient boundary conditions when a fluctuating river stage influences the zone of interaction between groundwater and surface water. The model provides information that supports a variety of applications, including monitoring strategies, contaminant transport models, risk assessments, remedial action design, and compliance requirements for remedial actions.

STOMP was used as the computer code of choice for these simulations. The utility of STOMP code for investigating near-river aquifer conditions had been demonstrated previously at the Hanford Site (Connelly et al., 1999). The code solves transient flow and transport problems in the subsurface environment in one, two, or three dimensions using an integral-volume, finite-difference approach, and has the capability to handle seepage faces along a sloping boundary. Nonlinearities in the governing equations are resolved through a Newton-Raphson iteration. The top of the model extends 5 m into the vadose zone so that the capillary fringe directly above the water table is included in the model calculations.

The principal assumptions made for the simulation include: single hydrologic unit (Hanford formation); no impediment to flow at the river channel interface; an anisotropy ratio of 1 : 10 for vertical-to-horizontal hydraulic conductivity; and a ratio of 100 : 1 for the transmissivity of the uppermost hydrologic unit to the underlying aquitard. Representative values for hydraulic conductivity and effective porosity were selected as part of the calibration process, where model output was compared to actual water levels measured in a nearby monitoring well. The values ultimately adopted were: 107 m (350 feet) per day for the horizontal hydraulic conductivity and 20% for effective porosity. Van Genuchten soil parameters were used for the vadose zone above the unconfined aquifer (Van Genuchten, 1980).

The modeled area extends for a total of approximately 900 m, with 320 m lying beneath the river, which is slightly less than halfway across the channel. The upper boundary condition beneath land areas included a moisture recharge rate of 1 cm/year (Rockhold et al., 1995), while the contact between the uppermost aquifer and underlying aquitard was considered to be essentially no-flow. The left (western) boundary was set at a well for which hourly water level data were available, and the right boundary (eastern) was the center of the river channel. Avariable horizontal grid spacing was used in the flow model. The finest grid spacing is used close to the river, to accommodate the large changes in gradients caused by stage fluctuations. Horizontal grid spacing ranged from 1 to 4 m, while vertical spacing was kept constant at 0.3 m. A 1-m width was adopted for volume flux calculations. Hourly water level data from 100-H Area monitoring wells and the adjacent Columbia River for the year 1998, which is a fairly typical year for river discharge, were used for the model.


Petersen, R.E., and M.P. Connelly. 2004. "Water Movement in the Zone of Interaction Between Groundwater and the Columbia River, Hanford Site, Washington" Journal of Hydraulic Research., Vol. 42, Issue: Extra, pp 53-58.

Connelly, M.P., C.R. Cole, and M.D. Williams. 1999. "Bank Storage Modeling at 100-N Area, Hanford Reservation." The 1999 Pacific Northwest Focus Groundwater Conference, 17-18 February. National Ground Water Association, Portland, Oregon.

Van Genuchen, M.TH. 1980. "A Closed-Form Equation for Predicting the Hydraulic Conductivity of Unsaturated Soils" Soil Sci. Soc. Am. J. 44, pp 892-898

Rockhold, M.L., M.J. Fayer, C.T. Kincaid, and G.W. Gee. 1995. "Estimation of Natural Recharge for the Performance Assessment of Low Level Waste Disposal Facility at the Hanford Site" PNL-10508. Pacific Northwest National Laboratory, Richland, Washington.

Back to Application Table