The Yakima River Basin in south-central Washington has a long history of irrigated agriculture and a more recent history of large-scale livestock operations, both of which may contribute nutrients to the groundwater system. Nitrate concentrations in water samples from shallow groundwater wells in the lower Yakima River Basin exceeded the U.S. Environmental Protection Agency drinking-water standard, generating concerns that current applications of fertilizer and animal waste may be exceeding the rate at which plants can uptake nutrients, and thus contributing to groundwater contamination.

The U.S. Geological Survey (USGS) recently completed a regional scale transient three-dimensional groundwater-flow model of the Yakima River Basin using MODFLOW-2000. The model was used with the USGS particle-tracking code MODPATH to generate advective flowpaths and associated travel times. Analyses used particle backtracking in time from September 2001 through 504 monthly stress periods to October 1959 or until pathlines terminated at a model boundary. The particle starting locations were assigned to 1,000 foot square computational model cells containing one or more of the 121 sampling locations with measured nitrate concentrations greater than the U.S. Environmental Protection Agency drinking-water standard for nitrate (10 milligrams per liter [mg/L]). Of the 2,403 particles, the simulated pathlines for 2,080 reached the water table within the 42-year simulation period, thus identifying the predicted recharge areas for those particles. The median horizontal straight-line distance was 13,194 feet between starting and ending locations for these particles and the median time-of-travel for particles that intersected the water table was 984 days. Well to water-table travel times for 75.4 percent of the particles were less than the average travel time of 3,749 days. Predicted recharge locations for all particles, including those that did not reach the water table in 42 years, were between 50 feet and 34 miles horizontal distance from their starting locations, with a median distance of less than 3 miles away.

Generalized groundwater-flow directions in unconsolidated basin-fill deposits were towards the Yakima River, which acts as a local sink for shallow groundwater, and roughly parallel to topographic gradients. Particles backtracked from more shallow aquifer locations traveled shorter distances before reaching the water table than particles from deeper locations. Flowpaths for particles starting at wells completed in the basalt units underlying the basin-fill deposits sometimes were different than for wells with similar lateral locations but more shallow depths. In cases where backtracking particles reached geologic structures simulated as flow barriers, abrupt changes in direction in some particle pathlines suggest significant changes in simulated hydraulic gradients that may not accurately reflect actual conditions. Most groundwater wells sampled had associated zones of contribution within the Toppenish/Benton subbasin between the well and the nearest subbasin margin, but interpretation of these results for any specific well is likely to be complicated by the assumptions and simplifications inherent in the model construction process. Delineated zones of contribution for individual wells are sensitive to the depths assigned to the screened interval of the well, resulting in simulated areal extents of the zones of contribution to a discharging well that are elongated in the direction of groundwater flow.