Changes in Groundwater Conditions and Groundwater Levels

The accuracy of the groundwater-elevation maps depends on various factors pertaining to the quality, quantity, and spatial distribution of the data, the method of interpolation, and the hydrogeologic properties and stresses within the aquifers. Therefore, the maps of groundwater elevation are approximations whose accuracy is limited by the availability of data. Data density is insufficient to support their use for local or site-specific purposes. These maps are intended to provide a regional overview of groundwater conditions in the CPRAS during spring 2009.

Groundwater Configuration

The configurations of the groundwater-elevation surface in the Saddle Mountains, Wanapum, and Grande Ronde units during spring 2009 are shown in plates 3, 4, and 5. The generalized groundwater-level maps provide a means to estimate regional-scale groundwater flow direction in that groundwater moves from areas of high to low water‑level elevations. However, the groundwater flow system is three-dimensional and consists of vertical and horizontal components of flow that may change with depth in the flow system forming complex flow paths. The overall direction of groundwater flow for each of the basalt units is toward the major groundwater discharge areas, consisting primarily of parts of the Columbia River but also including parts of the Snake and Yakima Rivers. Data density was insufficient to quantify the local directions of groundwater flow, such as in structural basins embedded within the Yakima Fold Belt, although where units occur at or near land surface, groundwater flow generally is expected to be toward streams and rivers that occupy the lowest points in the CPRAS and to follow surface drainage patterns. Additionally, groundwater withdrawals from major pumping centers or widespread recharge from irrigation or canals may produce local to regional variations that may not be reflected in the maps.

Configuration of the groundwater-elevation surfaces for the three basalt units generally compare well with groundwater-elevation surface maps constructed for the three basalt units by Whiteman (1986) on the basis of water levels from spring 1984. The maps by Whiteman (1986) exhibit greater detail because they incorporate a more comprehensive set of data consisting of a larger set of wells and a variety of ancillary data including water levels from years other than 1984, surface-water features, and land-surface topography. The areas where groundwater elevations from 1984 (Whiteman, 1986) and those from 2009 differ (where sufficient data were available for both dates) are most often associated with groundwater-level declines due to pumping since 1984 (see section titled “Water-Level Changes”).

The groundwater-surface configuration for the Saddle Mountains unit is similar to that of land surface (pl. 3). The highest groundwater elevations exceed 2,000 ft are in eastern Klickitat County, Washington, where the highest land-surface elevations occur within the extent of the unit. The lowest groundwater elevations below 500 ft are in two low-lying areas along the Columbia River, along the border between Franklin and Benton Counties, Washington, and near the junction of Benton County and Morrow and Umatilla Counties, Oregon. However, due to the distribution of available data it is difficult to ascertain the overall directions of groundwater flow in the Saddle Mountains unit with much certainty, with the exception of flow that occurs southeastward along the axes of some valleys in the Yakima Fold Belt. A detailed groundwater elevation and flow direction map constructed for the Saddle Mountains unit (as well as the Wanapum, Grande Ronde, and Overburden units) in the Yakima River basin using water-level data from 2001 is presented in Vaccaro and others (2009).

The groundwater-surface configuration for the Wanapum unit is roughly a circular depression centered southwest of Pasco, Washington, with a slight northeast to southwest elongation and a probable outlet that follows the Columbia River to the west (pl. 4). With few exceptions, the highest groundwater elevations exceeding 2,000 ft occur along the eastern and northeastern extents of the unit. The groundwater‑elevation surface generally slopes concentrically inward toward the Columbia and Snake Rivers, reaching minimum elevations of less than 400 ft from the northern end of Benton County extending southeastward to Pasco, Washington, and then extending southwestward to the northeast corner of Gilliam County, Oregon. The general directions of groundwater flow are southwestward from the margin of the Palouse Slope radially towards the Columbia River in the vicinity of Pasco, northward from the Oregon side to the Columbia River, and southeastward along the axes of valleys in the Yakima Fold Belt toward the Yakima or Columbia Rivers.

The configuration of the groundwater surface for the Grande Ronde unit is generally trough shaped, elongated from northeast to southwest in the Pasco, Washington, area and then along a more east-west trend along the Columbia River downstream of Pasco (pl. 5). The highest groundwater elevations occur along the southern, northwestern, and eastern extents of the unit with elevations exceeding 2,000 ft. With few exceptions, the groundwater-elevation surface slopes concentrically inward toward the Columbia and Snake Rivers, reaching minimum elevations of less than 200 ft along the Columbia River between The Dalles, Oregon, and Pasco, Washington. However, data from the central and deepest part of the Grande Ronde unit are sparse or nonexistent leading to high uncertainty in this area. The general directions of groundwater flow are similar to those for the Wanapum unit, with overall directions of flow moving radially from the margins of the unit toward the center of the CPRAS or the Columbia River.

The variography analysis indicated that the maximum range or distance that the groundwater level at a location in the Saddle Mountains unit (that is primarily within the Yakima Fold Belt) could be correlated with the level in another location is about 40 mi along a northwest-southeast direction, but only about 13 mi along a northeast-southwest direction. This difference in range by direction is because the correlation is determined not only by the distance between observations, but also by the direction from one observation to another. For example, water levels in wells along valleys are similar for much greater distances than across valleys, over ridges, and into adjacent valleys. This phenomenon is known as geometric anisotropy and may be explained by the presence of hydraulic barriers (such as those imposed by the faulted ridge tops in the Yakima Fold Belt) or by differences in stresses with direction (such as recharge from precipitation, irrigation, and canals, or discharge due to pumping).

The area of the Wanapum unit primarily extends across the Yakima Fold Belt and Palouse Slope structural regions. The interpolation of groundwater levels was performed separately for each region and then combined to create a map of groundwater levels. Analyses of the Yakima Fold Belt observations yielded a low geometric anisotropy, with a range of about 13 mi along a northwest-southeast direction and 11 mi along a northeast-southwest direction. The Palouse Slope data showed no anisotropy and a range of about 36 mi.

The extent of the Grande Ronde unit includes large parts of the Yakima Fold Belt and Palouse Slope structural regions. Variography indicated that groundwater-level correlations within these two regions are substantially different. Within the Yakima Fold Belt, correlations extend about 38 mi along a northeast-southwest direction, with a slight anisotropy, resulting in correlations extending about 32 mi in a northwest-southeast direction. Within the Palouse Slope, the range is about 61 mi in a north-south direction, with a slight anisotropy indicating a range of about 51 mi in an east-west direction.

Overall, the variography analysis was consistent with previously known groundwater flow features in the study area. The generally short ranges in the Yakima Fold Belt structural region indicate that the geologic complexity due to the large number of folds and faults contributes to compartmentalization of the groundwater flow system. The long ranges in the Palouse Slope result from the fewer number of faults and folds and indicate that the groundwater flow system is more continuous in this much less deformed structural region. The orientation of the greatest correlation of groundwater levels for the Saddle Mountains and Wanapum units within the Yakima Fold Belt is consistent with the dip of the valleys for which sufficient data are available. However, the sparse clusters of data relative to geological structures affect the results presented. As a result, the data are supportive and consistent with the geologic model, but the unqualified use of the parameters described above is not recommended when making predictions of groundwater elevations.

Vertical head differences, calculated as the difference in groundwater elevations between an upper and lower hydrogeologic unit, give an indication of the magnitude and direction of vertical gradients that may influence the vertical component of flow. Accurate head differences depend on the availability of wells that are open to individual hydrogeologic units and are in close proximity (ideally within the same drill hole if properly constructed). Because of the limited availability of this type of data for the three basalt units, only generalized vertical head differences that were calculated by subtracting the estimated water-level surface from one unit from the estimated water-level surface from the underlying unit are discussed. Vertical head differences between the Saddle Mountains and Wanapum units indicate that downward gradients may exist in the Saddle Mountains unit in many of the upper elevation areas. Upward gradients may be in several low-lying areas, including along much of the length of the Columbia River and in some valleys of eastern Yakima County, Washington. Vertical head differences between the Wanapum and Grande Ronde units show no clear spatial patterns possibly as a result of limited data, pumping within either unit, or vertical compartmentalization that may be limiting vertical flow between the units (Porcello and others, 2009).

The depth to water in the basalt units was estimated from the groundwater-elevation maps by subtracting the groundwater elevation from the land-surface elevation (maps of these calculated estimates are not shown). All three basalt units have a modest to strong correlation of depth to water with land-surface elevation, especially where the units are present at the surface. The correlation is strongest for the Saddle Mountains unit. Depth to water was generally shallowest in low-lying or deeply incised valleys and deepest in the higher land-surface elevation and mountainous areas.

Water-Level Changes

Small to moderate net groundwater-level declines between 1984 and 2009 were measured in most wells, although large declines greater than 100 ft and as great as 300 ft and essentially unchanged groundwater levels were not uncommon (pls. 6, 7, 8, and 9; table 1). Of the wells measured in 1984 and 2009, water levels declined in 83 percent of the wells and declines greater than 25 ft were measured in 29 percent of all wells. The magnitude of changes in water level depends on the change in storage within the aquifer, which is a function of the quantity of recharge or discharge added or removed from the aquifer and the storage properties of the aquifer.

The smallest change in groundwater levels in terms of maximum decline, maximum rise, and mean change was measured in the Overburden unit (pl. 6; table 1). The frequency distribution of water-level changes indicates that little change in water levels occurred in most areas, although some areas showed a tendency toward small groundwater declines and a few areas exhibited small groundwater rises. Of the wells measured in spring 1984 and spring 2009, water levels declined in 87 percent of the wells, although declines greater than 25 ft were measured in only 8 percent of the wells. Most of the wells measured in the Overburden unit as part of this analysis are located within the Hanford Site operated by the Department of Energy in northern Benton County, Washington. Water levels in the area of the Hanford Site showed little change except for the central or northwestern areas, which show small to moderate water level declines. These declines can be largely attributed to the substantial decrease in liquid effluent disposal at the Hanford Site during the 1990s (Pacific Northwest National Laboratory, 2009).

Most groundwater levels measured in the Saddle Mountains unit showed small declines, although large declines or small to moderate rises were indicated in some areas (pl. 7; table 1). Water levels declined in 68 percent of the wells and declines were greater than 25 percent in 16 percent of the wells. The magnitudes of the water-level changes were low relative to the other basalt hydrogeologic units. Spatial trends in water-level changes in the Saddle Mountains unit were difficult to discern due to the distribution of the data. An area of small groundwater-level declines was identified in the central area of the Hanford Site. Groundwater declines also were identified in the Yakima River basin. Small water-level rises were measured in the area along the northeast side of the Columbia River within the Palouse Slope.

Groundwater levels in the Wanapum unit mostly showed small or moderate declines, with large declines (greater than 100 ft) in a few areas (pl. 8; table 1). Between 1984 and 2009, water levels declined in 78 percent of the wells measured and declines greater than 25 ft were measured in 38 percent of the wells. Wells within the Yakima Fold Belt structural region generally had larger groundwater declines relative to the Palouse Slope. This may be the combined result of pumping and the increased compartmentalization of the groundwater system by geologic structures present in the Yakima Fold Belt. The largest concentration of declines was in the Yakima River basin, which corresponds to areas of heavy pumping within the Wanapum unit (Whiteman and others, 1994).

Groundwater levels changed in the Grande Ronde unit more than water levels in the other units in terms of maximum decline and mean change (pl. 9; table 1). Moderate declines in water levels were measured in most areas, with declines greater than 100 ft and some greater than 200 ft. Water levels declined in 87 percent of the wells, and declines greater than 25 ft were measured in 60 percent of the wells. Substantial rises in groundwater levels were measured in few wells. Groundwater declines were present in the Grande Ronde unit throughout much of the area where wells were measured. The largest concentrations of declines were in the central northern part of the study area that extends from western Lincoln County, Washington, southward into eastern Grant and southwestern Adams Counties, Washington. Large concentrations of declines also were measured in the part of the Umatilla River basin in northeast Morrow and western Umatilla Counties, Oregon, and in the Pullman, Washington, and Moscow, Idaho, area. These concentrations of declines in water levels correspond with areas of heavy pumping within the Grande Ronde unit (Whiteman and others, 1994).

This report provides a regional assessment of groundwater levels during spring 2009 based on an inventory of 1,752 wells in the Columbia Plateau of Washington, Oregon, and Idaho. Data included in the report are from field investigations and a compilation of published and unpublished well data from many agencies. Water-level change maps are presented for the Overburden, Saddle Mountains, Wanapum, and Grande Ronde units and indicate trends toward water-level declines in many areas since 1984. The distribution of wells monitored in spring of 2009 provide information on where additional monitoring would better define the response of the groundwater system to stresses such as pumping, irrigation, and climate.