Introduction


The Columbia Plateau Regional Aquifer System (CPRAS) covers about 44,000 mi² of southeastern Washington, northeastern Oregon, and western Idaho (fig. 1). The population of the region is more than 1.3 million people (U.S. Census Bureau, 2000), and an important agricultural industry valued in the billions of dollars annually has developed despite the arid to semi-arid climate and limited access to surface-water resources. Groundwater availability in the aquifers of the area is a critical water‑resource management issue because of the high demand of water for agriculture, economic development, and ecological needs. Groundwater levels have declined throughout much of the Columbia Plateau (Whiteman and others, 1994, p. B61– B65; Porcello and others, 2009; Burns and others, 2012). A comparison of water-level measurements from 1984 and 2009 from 470 wells in the CPRAS indicates that water levels declined in 83 percent of the wells and that declines were greater than 25 ft in 29 percent of the wells (Snyder and Haynes, 2010).


The primary aquifers of the CPRAS are basalts of the Columbia River Basalt Group (CRBG) and in places, overlying basin-fill sediments. Water-resources issues that have implications for current and future groundwater availability in the region include (1) widespread water-level declines associated with development of groundwater for irrigation and other uses; (2) reduction in base flow to rivers and associated effects on water temperature, water quality, fish, and other aquatic organisms; and (3) current and anticipated effects of global climate change on recharge, base flow, demand, and ultimately, groundwater availability.


The U.S. Geological Survey (USGS) Groundwater Resources Program began a study of the CPRAS in 2007 with the broad goals of (1) characterizing the hydrologic status of the system, (2) identifying trends in groundwater storage and use, and (3) quantifying groundwater availability. The study approach includes updating the regional geologic and hydrogeologic frameworks, documenting changes in the hydrologic status of the system, quantifying the hydrologic budget, and developing a groundwater-flow model for the system. The model, which will be outlined in a separate report, will be used to evaluate and test the conceptual model of the groundwater flow within the system and to evaluate groundwater availability. This report, which describes the relation between groundwater levels and trends and hydrogeologic controls, along with four recently published reports (Kahle and others, 2009; Snyder and Haynes, 2010; Burns and others, 2011; Kahle and others, 2011), provides comprehensive information about the physical hydrogeologic framework of the CPRAS based on historical and current investigations. This study, in part, relied on data collection and analysis conducted as part of a cooperative agreement between the USGS and the Oregon Water Resources Department to better define the hydrologic conditions in the Umatilla basin of Oregon.


Purpose and Scope


The purpose of this report is to describe the compilation of groundwater-level data for the CPRAS that will be used for comparison with a numerical groundwater-flow model and to evaluate the status and trends in the data and their relation to hydrogeologic controls that influence the hydraulic properties of the aquifer or hydraulic stresses from recharge or pumping. The scope of this report includes a regional assessment of the importance of these controls described through the presentation of maps of groundwater elevations, water-level changes, stresses, and hydrogeologic features in the CPRAS. This information will help to develop a broad understanding of how climatic, anthropogenic, and hydrogeologic factors combine to influence groundwater flow, and how climate change and groundwater development may influence the sustainability and availability of the water supply in the region. The analyses presented in this report generally are restricted to the primary CRBG aquifers within the CPRAS to demonstrate important features in key areas, and do not seek to be an exhaustive analysis of any area.


Description of Study Area


A complete description of the Columbia Plateau is available in Kahle and others (2011), parts of which are presented here. The Columbia Plateau (fig. 1) is a structural and topographic basin within the drainage of the Columbia River basin. It is bounded on the west by the Cascade Range, on the east by the Rocky Mountains, on the north by the Okanogan Highlands, and on the south by the Blue Mountains. The Columbia Plateau is underlain by massive basalt flows with thickness estimated to exceed 16,000 ft near Pasco, Washington (Reidel and others, 2002, p. 211, fig. 2.6). Sedimentary deposits locally exceeding 2,000 ft in thickness overlie the basalt over large areas of the plateau.


The Columbia Plateau was divided into four informal physiographic subprovinces that represent structural 
regions—the Yakima Fold Belt, Blue Mountains, Palouse Slope subprovinces, and the Clearwater Embayment (fig. 1). Groundwater characteristics in each of these regions can be different because of variations in stratigraphy, depositional environment, and post-deposition folding and faulting. The presence and importance of flow barriers in the CPRAS has been recognized and discussed in numerous studies (for example, Newcomb, 1959; Porcello and others, 2009). The Yakima Fold Belt includes most of the western half of the plateau and is characterized by a series of east-west trending anticlinal ridges and synclinal basins. The Palouse Slope occupies the northeast quarter of the plateau, is much less deformed, and has a gently southwestward dipping slope. The other structural regions within the CPRAS are the Blue Mountains, a composite anticlinal structure that forms the southeastern extent of the Columbia River basin, and the Clearwater Embayment, which marks the eastward extent of the CPRAS along the foothills of the Rocky Mountains and includes a series of folds extending into Idaho.


Much of the Columbia Plateau is semiarid, the mean annual precipitation for 1895–2007 (Kahle and others, 2011, p. 4) is about 17 in/yr (about 40 million acre-ft/yr) and ranges from about 7 in. in the center of the study area to more than 60 in. in the north-westernmost extent of the study area. The types and amounts of natural vegetation growing on the Columbia Plateau are largely dependent on precipitation and land-surface elevation. The vegetation ranges from sagebrush and grasslands at lower elevations to grasslands and forest at mid elevations to barren rock and conifer forests at the upper elevations. Dry land agriculture mainly includes winter and spring wheat and lentils. Irrigated agriculture includes apples, hops, and other crops.


Overviews of the geology and hydrology of the CPRAS presented summarize detailed descriptions in reports by (1) Kahle and others (2009), who discuss the geologic framework used in this report; (2) Burns and others (2011), who describe the three-dimensional characteristics of the geology of the CPRAS; and (3) Kahle and others (2011), who discuss the hydrogeologic framework and the hydrologic budget components of the CPRAS.


Hydrogeologic Setting


The Columbia Plateau is an intermontane basin between the Rocky Mountains and the Cascade Range that is filled with mostly Cenozoic basalt and sediment. Most rocks exposed in the region are the CRBG, intercalated sedimentary rocks, and overlying younger sedimentary rocks and deposits that include Pleistocene cataclysmic flood deposits, eolian deposits, and terrace gravels of modern rivers. The CRBG consists of a series of more than 300 flows that erupted in various stages during the Miocene, 17 million to 6 million years ago. Individual flows range in thickness from 10 to more than 300 ft (Tolan and others, 1989; Drost and others, 1990). Soil and sediments that were formed or deposited on top of older lava flows and then covered and preserved by a subsequent lava flow are called sedimentary interbeds.


Generalized geologic stratigraphy discussed in this investigation is broadly based on recognized formations within the Columbia River Basalt Group. The primary geologic formations of interest, listed in order of increasing age, include the Saddle Mountains Basalt, the Wanapum Basalt, and the Grande Ronde Basalt (fig. 2; Swanson and others, 1979, p. G4–G8; Whiteman and others, 1994, p. B32–B33). Younger sedimentary deposits cover parts of the CRBG across the study area and are referred to informally as the Overburden. The Overburden consists of undivided, unconsolidated to
semi-consolidated sedimentary deposits ranging from Miocene to Holocene in age (Drost and others, 1990). These include many types of deposits of local and (or) regional extent including flood gravels and slack water sediments, terrace gravels of modern rivers, and eolian deposits that can range in thickness from 0 to 1,300 ft. The Saddle Mountains Basalt formation consists mostly of basalts and associated sedimentary interbeds and is the least extensive and youngest formation of the CRBG. Most of the formation is in the west-central part of the study area, with less continuous occurrences in the Blue Mountains and eastward into Idaho. Thickness of the Saddle Mountains Basalt ranges from 0 to about 1,000 ft. The Wanapum Basalt formation, composed mostly of basalt and sedimentary interbeds, is in the central part of the study area. Much of the formation lies beneath the Overburden and Saddle Mountains Basalt. Thickness of the Wanapum Basalt ranges from 0 to about 1,200 ft. The Grande Ronde Basalt formation is the oldest and most extensive of the basalt formations discussed and comprises the vast majority of the CRBG. This formation underlies most of the study area, except for an area along the southern boundary of the CPRAS in Oregon and along the eastern edge of the aquifer system in Idaho. The Grande Ronde Basalt formation contains basalt and sedimentary interbeds. Thickness of the formation is largely unknown, but may be greater than 15,000 ft near the central part of the basin (Burns and others, 2011, p. 30, fig. 10D, and digital data).


Folding and faulting of the basalts occurred during the period of deposition and more recently. Younger basalt flows commonly were less-voluminous intra-canyon flows, so the distribution and thickness of these lavas are controlled by the shape of the valleys through which the lava flowed. Pleistocene outburst floods shaped the area by causing incision in the basalt, and erosion and deposition of the 
overlying sediment.


These geologic formations correspond to the hydrogeologic units defined for use in the USGS Regional Aquifer-System Analysis (RASA) Program and Groundwater Availability studies of the CPRAS (Whiteman and others, 1994, p. B32–B33; Kahle and others, 2009), and consist of the Overburden unit, the Saddle Mountains unit, the Wanapum unit, and the Grande Ronde unit. The informal term “unit” is used to differentiate from the formal geologic formations simplified for hydrologic discussions. For example, the geologic formations were simplified by grouping the Mabton and Vantage Interbeds into the Saddle Mountains and Wanapum units, respectively. The older rocks bordering and underling the CPRAS form the basement confining unit, referred to as the Older Bedrock hydrogeologic unit, and is composed of various rock types older than the CRBG, generally with much lower permeabilities than the basalts and considered the base of the regional flow system.


Hydrogeologic Characteristics


The hydraulic characteristics of the geologic materials determine how a groundwater-flow system functions and how it will respond to stresses such as pumpage. These characteristics include horizontal and vertical hydraulic conductivity and the storage coefficient. Because of the heterogeneity of the geologic materials that comprise the CPRAS, the hydraulic characteristics can vary considerably. The Overburden deposits are diverse in lithology and the large variation in grain size, depositional regimes, and age of the deposits account for the large range of their hydraulic characteristics (Kahle and others, 2011, p. 20). Each of the CRBG geologic units consist of tens to hundreds of individual layered basalt flows. The layers are highly variable in thickness and extent, but over much of the Columbia Plateau, the lava flows are comprised of flood basalts that form laterally extensive deposits. Hydraulic characteristics vary greatly within and between the individual basalt flows. Horizontal hydraulic conductivities generally are greatest in the interflow zones formed from the combination of basalt flow tops with the base of an overlying basalt flow, and an intervening sedimentary interbed, if present. Flow tops and bases commonly are brecciated, although they can exhibit a wide range of depositional textures. Because flow tops and bottoms commonly have open and highly-connected pore structures, the basalt interflow zones frequently exhibit high horizontal hydraulic conductivity (Lindolm and Vaccaro, 1988). The interflow zones are separated by the low hydraulic conductivity flow interior in which most fractures are cooling joints that often are vertically oriented (columnar jointing). Despite the fact that joints exist, flow interiors commonly are effective confining units. Porosity and permeability generally are lower in the older bedrock than in the Overburden and 
CRBG units.


Kahle and others (2011, p. 21) estimated the median horizontal hydraulic conductivity values for the Overburden, basalt units, and older bedrock as 161, 70, and 6 ft/d, respectively, based on specific capacity data reported in previous studies. Vertical hydraulic conductivity of the geologic units in the CPRAS generally is known to be low but is poorly quantified. Estimates of vertical hydraulic conductivity range from about 0.009 to 2 ft/d for the Overburden unit, although values for some parts of this unit may be as low as 10^-10 to 10^-7 ft/d; values for the CRBG units range from 4×10^-7 to 4 ft/d (Kahle and others, 2011, p. 57).


The storage characteristics of an aquifer are described by the storage coefficient, a dimensionless property defined as the volume of water that an aquifer takes into or releases from storage per unit of surface area per unit change in head. Previous estimates of the storage coefficient in the CPRAS typically range from about 0.1 to 0.2 for the unconfined parts of the Overburden unit and from about 0.01 to 10^-6 for the CRBG basalt units (Kahle and others, 2011, p. 26–27).


Hydrologic Budget Components


The following discussion and estimates of the 
regional-scale hydrologic budget components for the CPRAS is from the recent work by Kahle and others (2011). Mean annual recharge from infiltration of precipitation was estimated based on annual precipitation data and previous model simulation results of recharge. The estimates of the other budget components were developed using a monthly SOil WATer (SOWAT) balance model to determine irrigation water demand, groundwater flux to the underlying modeled soil zone (recharge or discharge), direct runoff, and soil moisture in irrigated areas (Kahle and others, 2011, p. 36). The SOWAT model was developed to make use of estimates of actual evapotranspiration available from a new Simplified Surface Energy Balance method that uses remotely sensed land-surface temperature data (Senay and others, 2007).


Mean annual recharge from the infiltration of precipitation of 4.6 in/yr (10.8 million acre-ft/yr) for 
1985–2007 was estimated for the CPRAS. The spatial distribution in recharge mirrors that of annual precipitation, with the highest recharge (more than 20 in/yr) occurring along the Cascade Range and the Blue Mountains. Mean annual recharge from infiltration of precipitation is less than 1 in/yr for a large part of the study area adjacent to the Columbia and Yakima Rivers where precipitation is limited.


Mean monthly irrigation throughout the study area peaks in July at 1.6 million acre-ft (MAF) (1985–2007 average), of which 0.45 and 1.15 MAF are from groundwater and 
surface-water sources, respectively. Annual use of irrigation water in the study area averaged 5.3 MAF during 1985–2007, with 1.4 MAF (or 26 percent) supplied from groundwater and 3.9 MAF supplied from surface water. Mean annual groundwater recharge from deep percolation of applied irrigation water in the study area was 4.2 MAF (1985–2007); 2.1 MAF (50 percent) occurred within the predominately surface-water irrigated regions of the Yakima Basin, Umatilla Basin, and Columbia Basin Project. The Columbia Basin Project, located in east central Washington, consists of more than 1,000 mi² of land irrigated with Columbia River water through a series of dams, pumping plants, and canals. Annual recharge rates range from less than 5 in/yr in predominately sprinkler-irrigated areas where groundwater is the source to more than 20 in/yr in surface-water supplied areas where conveyance losses and less-efficient application methods are more common.


Annual groundwater use (1984–2009) for purposes other than irrigation was estimated for the study area using information from several agencies. Public-supply groundwater use increased from about 201,000 acre-ft/ yr in 1984 to about 269,000 acre-ft/yr in 2009. Domestic self-supplied groundwater use increased from about 54,600 acre-ft/ yr in 1984 to about 71,200 acre-ft/yr in 2009. Industrial groundwater use decreased from about 53,400 acre-ft/yr in 1984 to about 43,900 acre-ft/yr in 2009. Other groundwater use, including for mining, thermoelectric needs, livestock, and aquaculture combined, increased from 16,900 acre-ft/yr in 1984 to about 43,600 acre-ft/yr in 2009.


Groundwater Occurrence and Movement


Groundwater moves through the regional aquifer system from the uplands to surface drainage features in the lowlands, mainly to the Columbia River and its major tributaries. Groundwater movement is affected by the geometry of the land surface and stream system, the extent, thickness, and hydraulic properties of the aquifers, the presence and nature of geologic structures, and the rate and location of groundwater recharge and discharge. Groundwater flow within the basalt units moves horizontally and vertically in the basalt interflow zones, flow centers, and sedimentary interbeds (Kahle and others, 2011, p. 27). Horizontal hydraulic conductivities generally are greatest in the interflow zones and consequently the interflow zones support most of the horizontal groundwater movement, whereas movement in the typically thick and lower horizontal hydraulic conductivity flow centers mainly is vertical. Therefore, the interflow zones in the basalt sequence form numerous, thin, semiconfined aquifers whose physical and hydraulic characteristics vary horizontally and vertically. Geologic complexity consisting of changes in lithology and folds and faults affect the geometry of flow paths by forming flow barriers or preferential pathways for groundwater flow (Snyder and Haynes, 2010, p. 7; Kahle and others, 2011, p. 27).


Study Methods


Recent and historical water-level data were compiled from a large number of wells located throughout the CPRAS. These data were used to identify “well groups” each consisting of individual wells with similar hydraulic heads (groundwater levels) and temporal trends. Comparisons were made between adjacent well groups to delineate sets of well groups, which define areas of overall similar groundwater-flow conditions with regard to groundwater-flow directions and horizontal and vertical hydraulic gradients (the change in hydraulic head per unit of distance in a given direction). Discontinuities in groundwater-flow conditions between the sets of well groups were used to help infer the presence of features that contribute to control the regional flow of groundwater. These discontinuities can result from: (1) geologic features that influence the hydraulic properties of the aquifer such as changes in lithology or the occurrence of structural folds and faults, or (2) the presence of hydrologic features such as groundwater mounds or troughs caused by stresses such as recharge from irrigation or discharge to pumping.


Sources of Data


The evaluation of groundwater status and trends for the CPRAS relied on information that was routinely collected by many agencies or that was developed in the current or previous studies of groundwater availability in the CPRAS. In particular, this report relies heavily on the development of the geologic framework and three-dimensional geologic model by Kahle and others (2009) and Burns and others (2011), respectively, and the hydrologic budget components developed by Kahle and others (2011).


Groundwater-Level Information 


The primary source of well information, including location, construction, and water-level measurements, was the USGS National Water Information System (NWIS). Additional well and water-level data were provided by the U.S. Department of Energy, Washington State Department of Ecology, Oregon Water Resources Department, Columbia Basin Ground Water Management Area, GSI Water Solutions, Inc., Confederated Tribes of the Umatilla Indian Reservation, and Walla Walla Basin Watershed Council. Due to constraints on project resources, not all wells or water-level measurements from the CPRAS were included in this study. Emphasis was placed on incorporating datasets with the greatest number of wells, which included information on well construction and many water-level measurements made over relatively long periods. Wells are primarily within the extent of the CPRAS, although some wells as much as 20 mi outside of the CPRAS boundary also were included to help identify groundwater-level conditions across the boundary of the CPRAS (fig. 3). Information was acquired for 60,115 wells, although usable water-level data were available for only 39,610 wells. A total of 447,992 water-level measurements were collected from the 39,610 wells. The period of record for the water-level data extended from 1891 to 2010. The longest period of record for water-level measurements at an individual well is 96 years. Measurements at the longest annually monitored well began in 1940 and continue to the present (2011). From these data a subset of 7,772 wells and 147,563 water levels were used for the analyses presented in this report and are on file at the USGS office in Tacoma, Washington (http://wa.water.usgs.gov/projects/cpgw), and available upon request. Much of the data processing for analyzing groundwater levels was accomplished using Excel^® spreadsheet tools developed by Tillman (2009).


Information on land-surface elevation was needed to calculate groundwater elevation from depth-to-water measurements. Land-surface elevations for many wells were reported by the agencies that provided the well information. Land-surface elevations were independently determined using a 10-m lateral resolution Digital Elevation Model (DEM; U.S. Geological Survey, 1999) as verification for these wells and to obtain land-surface elevations for the remaining wells with no reported values. Values of the DEM derived land-surface elevations were used if the difference between the reported and DEM derived elevations differed by greater than 100 ft.


Geologic Model


A geologic model (Burns and others, 2011) provided improved estimates of CRBG and Overburden unit volumes and refined location of large structural features. This model was used to interpret the presence and significance of hydrogeologic controls on the groundwater system presented in this report. An on-line interactive tool was developed to serve point information and cross sections developed from the geologic model to the public (U.S. Geological Survey, 2012).


Data Limitations


The accuracy and representativeness of the groundwater‑level measurements are dependent on various factors pertaining to measurement accuracy, quality assurance procedures, local conditions in the aquifer at the time of measurement, and well construction.


Measurement Accuracy and Data Compilation


Typical methods used to measure the depth to water in a well have precisions ranging from 0.01 ft to several feet. The precision of most measurements is expected to be in the range of about 0.1 ft. However, the accuracy of the determination of the groundwater elevation (how it compares to the actual value) depends on how the depth to water measurement is transformed to a groundwater elevation at a particular location, how it is associated with a particular aquifer, and the associated quality assurance procedures. A major factor affecting measurement accuracy is related to how the elevation for the well was determined, and the errors associated with assigning an elevation to the well. Errors also can relate to the location of the well. Errors in the latitude and longitude coordinates for the site can create errors of hundreds of feet in the assigned elevation of the water level, especially in areas with steep relief. For this report, well locations and elevations were mostly based on information reported from the respective agencies. A comparison was made between reported well elevations and those determined from a 10-m DEM on the basis of the reported latitude and longitude locations and lateral datums from the respective agencies. The median of the absolute values of the elevation differences is 4.3 ft and 90 percent of the differences are less than 44 ft.


The well and water-level data were compiled from many different agencies. Each agency has its own protocols for the description and documentation of measured wells, for the collection and recording of water-level data, and quality assurance procedures. Errors can occur when compiling large datasets due to either data entry, arithmetic errors, or treatment of exceptions when using rule-based algorithms to write all data into a common database. Common errors include incorrect well identification or location, incorrect or incomplete well construction information, errors in depth to water measurements, and incorrect determination of land‑surface elevations. Additionally, some well information and water-level measurements for the same well may be reported by multiple agencies resulting in duplication of information. An effort was made to identify and correct errors in the data used in this study. However, it was not possible to ascertain the quality of the well information or water-level measurement for each well and undoubtedly, errors are present in the data. The large number of wells and measurements used in the analyses in this study should help to minimize the influence of these errors and provide a robust estimate of the groundwater elevation.


Water-level measurements in reports filed by well drillers at the time of new well installations were included in the compilation of water levels. However, measurements of water levels in newly constructed wells may not have been at equilibrium at the time of measurement and therefore may not represent static water-level conditions (Snyder, 2008, p. 11–14). Water-levels where the status indicated the well was dry, obstructed, or influenced by pumping, were excluded from the analyses as these water levels may not represent static conditions.


Representative Sampling


The groundwater-level measurements used in this study are assumed to be representative of the groundwater positions within the aquifers of the CPRAS. However, some of the data may suffer from a number of possible biases. These biases may arise from a variety of sources such as influences due to localized conditions and stresses, well construction, and commingling of groundwater. In addition, the spatial distribution and selection of wells used for the collection of water-level measurements may be biased as a result of increased scrutiny and monitoring in areas with groundwater declines. This can result in the clustering of wells and an over‑representation of wells showing water-level declines.


Localized Conditions and Stresses


Pumping and recharge stresses can cause geographically localized or transient perturbations of water levels. Water-level measurements affected by such localized perturbations may not be representative of true conditions in the aquifer over a broader area.


Recharge in the CPRAS is dominated by precipitation (Kahle and others, 2011), which generally has gradual geographic variation. However, other recharge sources such as irrigation water, artificial recharge, or leakage from streams, canals, or lakes can influence groundwater levels over short distances. Stresses due to discharge by pumping, or to gaining streams, springs, seepage faces, wetlands, or evapotranspiration similarly can have large variations over short distances and influence groundwater levels.


The aquifer properties that influence the movement and storage of water in the aquifer and, therefore the groundwater level in the aquifer, include the extent, thickness, shape, hydraulic conductivity, storage coefficient, and degree of confinement. These properties can be highly heterogeneous over short distances due to variations in the original lithology or process of emplacement, subsequent modification by physical process such as erosion or fracturing due to structural deformation, or chemical changes such as dissolution, alteration, or mineralization. Variations in these properties over short distances may affect the groundwater levels and (or) the timing and magnitude of groundwater changes in response to changes in stresses relative to groundwater levels elsewhere in the aquifer.


Influence of Well Construction


Well construction may substantially influence the water levels in wells. Ideally, a well used for groundwater-level monitoring should be constructed to ensure good hydraulic connection between the well and the intended aquifer and that the water level and water-level fluctuations in the well broadly represent conditions in the aquifer. The well design must take into consideration the placement of open intervals and include appropriately-sized well screens to permit a good hydraulic connection between water in the aquifer and in the well. The use of sanitary and flow seals and well casings, where needed, help to isolate the well and the contributing aquifer from other units to prevent commingling, the collapse of rock into the well, and movement of water between the rock and the outside of the well casing. Incorrect well design, construction defects at the time of installation, insufficient well development (repeated purging and filling of a well to remove fine materials that may clog the well screen), or degradation due to age including silting, corrosion, or bacterial growth may affect how a well responds to changes in the aquifer (Taylor and Alley, 2002, p. 9).


Complete and accurate documentation of the well construction is important to ensure that the water levels measured in the well are properly interpreted. Documentation of well construction for some wells often is incomplete and does not include sufficient information, such as lithology or position of open intervals, to identify the contributing aquifer(s). This creates uncertainty on how to associate a water-level measurement from a well with the appropriate aquifer. Uncertainty with regard to the aquifer represented in a water-level measurement can contribute to misleading interpretations.


Over time, water-level declines in the CPRAS have resulted in dry wells, some of which have been deepened and reconstructed. Because water-levels from a deepened or reconstructed well represent conditions at a different location in the aquifer system, measurements always should be associated with well depth and construction at the time of measurement. However, in most instances the information on the deepening of wells was not associated with the original well for the data compilation sources used for analysis in this study. As a result, water-level measurements made subsequent to the deepening of some wells may remain associated with the depth of the original well. Hydrographs for these wells may show an initial step change in the water-levels measured in the well following deepening that could be mistakenly attributed to other causes. The water-level response of the well subsequent to deepening also may reflect the conditions in the aquifer at the new well depth. To address these limitations, robust methods of analysis were used to ensure that errors at individual wells did not strongly affect overall study conclusions.


Commingling


Commingling is the term used to describe the condition that occurs when a well is constructed so water can move from one aquifer to another through the well bore. This can occur in wells that are open to multiple aquifers through screens or uncased intervals. If the aquifers have different heads, then water will move through the well bore from the aquifer(s) with higher head to the aquifer(s) with lower head. When such commingling occurs, the static water level in the well is a composite water level, averaging conditions between all the aquifers open to the well.


In figure 4, well 3 is a hypothetical multiple completion well with open intervals in aquifers 1 and 2, which results in commingling of waters between aquifers. Because the hydraulic head in aquifer 1 (as represented by the potentiometric surface of the aquifer at the well) is greater than the hydraulic head in aquifer 2 a downward hydraulic gradient enables flow to enter the well 3 bore from aquifer 1, traversing down the well 3 bore and exiting the well 3 bore into aquifer 2. For this example, the transmissivities of aquifer 1 and 2 are equal; therefore, the resulting hydraulic head in well 3, from a transmissivity-weighted average of the heads in aquifers 1 and 2, is exactly one-half the vertical distance between the head in the two aquifers. Well 2 and well 4 are single completion wells with open intervals within aquifer 1 and aquifer 2, respectively, that are in close proximity to commingling well 3. These wells are within the cone of depression and cone of impression (the inverse of a cone of depression) resulting from the commingling in well 3. The hydraulic head in these wells is intermediate to the head at well 3 and the initial unaffected heads in aquifers 1 and 2 as represented by the heads in wells 1 and 5, respectively.


In figure 5, the transmissivity of aquifer 1 is much greater than the transmissivity of aquifer 2; therefore, the head in well 3 is dominated by the hydraulic head of aquifer 1. Because the cone of depression for aquifer 1 is greatly subdued, water levels in proximal wells open to aquifer 1, such as well 2, are minimally affected. However, water levels in proximal wells open to aquifer 2, such as well 4, are strongly affected due to the exaggerated cone of impression for aquifer 2.


The basalt aquifers of the CPRAS consist of a series of permeable interflow zones separated by less permeable flow interiors (Kahle and others, 2011, p. 20). The transmissivities of the aquifers can vary over several orders of magnitude (Kahle and others, 2011). The resulting water levels in commingled wells open to multiple aquifers depend on the relative transmissivity of the aquifers, which is a function of the thickness and the permeability of the aquifers. Figure 6 illustrates some possible examples for aquifers of varying transmissivity (thickness and (or) permeability).


The ratio of water-level fluctuations in a well to the groundwater-level fluctuations in an aquifer penetrated by that well is equal to the ratio of transmissivity of the aquifer in which the fluctuation occurs to the total transmissivities of all aquifers perforated by the well. As a result, the effect of the change in groundwater level of any aquifer where the well is open imposes a smaller change on the water level in the well (Sokol, 1963, p. 1,080).


First posted February 5, 2013