Scientific Investigations Report 2008–5059
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5059
Estimates of depth-to-water and water-table elevation are affected by a number of variables, including the types, characteristics, timing, and errors associated with the data, as well as effects due to the interpolation method. Therefore, the maps of depth to water and water-table elevation are approximations; the values represent average conditions and have an associated uncertainty. Further, the actual water-table position will vary temporally as a result of short-term, seasonal, or long-term influences. Data files of depth to water, water-table elevation, and relative uncertainty of the water-table position are available for download as described in appendix B.
Depth to water ranges from 0 ft below the surface along major rivers and streams to a maximum estimated at more than 1,200 ft below land surface on the southern slopes of Larch Mountain (pl. 1). The depth to water for nearly two-thirds of the study area analyzed was less than 100 ft. Areas where the depth to water exceeds 100 ft include the Tualatin Mountains, the Boring Hills, and the foothills of the Cascade Range. Depth to water also exceeds 100 ft in the terrace deposits throughout much of northern and eastern Portland area. Depth to water greater than 300 ft is limited to a few high elevation areas and include parts of the Tualatin Mountains, the slopes of Larch Mountain, the ridge extending southeast from Lake Oswego, and the summits of Rocky Butte, Mount Tabor, Mount Scott, and several other of the Boring Hills (pl. 1). A visual inspection of the depth-to-water map reveals a significant correlation between depth to water and land-surface elevation. Depths to water were deeper in areas with high elevations such as the Tualatin Mountains, Boring Hills, and the foothills of the Cascade Range. Depths to water were shallower in low-lying areas along rivers and streams such as the Columbia, Willamette, Clackamas, and Sandy Rivers and the water table in many places is at or near land surface. The spatial correlation between the interpolated depth to water and the land-surface elevation has a correlation coefficient (R value) of 0.73 indicating a relatively high correspondence between the two; that is, as land-surface elevation increases depth to water increases. These observations are consistent with the concept that depth to water typically is greater beneath hills than valleys (Fetter, 1994, p. 114).
Areas of notably shallow depths to water in addition to the low-lying areas along major rivers and streams include much of the area adjacent to Johnson Creek, the area around Fairview Creek including several small lakes, the area extending from the west end of Lake Oswego southwest to the Tualatin River, and the area consisting of former alluvial channels (Hogenson and Foxworthy, 1965, p. 10, 11, and 28) extending from the confluence of Johnson and Crystal Springs Creeks northward to the Willamette River and southward to the Clackamas River. Another area of relatively shallow depth to water extends southwest from Fairview Creek, through the saddle area between Kelly and Powell Buttes, and intersects Johnson Creek west of Mount Scott (pl. 1). This area has been described by Allison (1978b, p. 193) as an erosional channel formed by the Missoula Floods. The area includes Beggars Tick Marsh and a small depression situated at the west end of Powell Butte, locally known as “Holgate Lake” (Lee, 2002).
The influence on the water table resulting from certain manmade features, where overlying soil and rock have been removed creating shallow depths to water, can be seen on the depth-to-water map (pl. 1). These include sand and gravel pits such as those south of the Portland International Airport, north of Kelly Butte, west of Mount Scott, and along the west side of Gresham as well as road cuts for Interstate Highway 205 in northeastern Portland area and for the railroad through northern Portland, which parallels N. Portland Road. Other influences due to manmade features can be discerned on the depth-to-water map but are actually artifacts resulting from the processing of the digital elevation model data (see section, “Assumptions and Assessment of Errors”). These features include the major elevated roadways and bridges such as those on the Interstate highways, especially near downtown Portland, which produce depths to water that are improbably greater than adjacent areas.
No published depth-to-water maps that cover an extensive part of the study area were located for comparison with the current study. However, the USGS simulated ground-water elevation based on a regional three-dimensional finite-difference ground-water flow model (Morgan and McFarland, 1996). A simulated depth-to-water map can be developed by subtracting the simulated ground-water elevations from the land-surface elevations used in the model. Output data files from the USGS model are available online (U.S. Geological Survey, 2006) and were analyzed to evaluate depth to water. Many of the wells and surface-water features used in the model also were used for the current study. The comparison is extremely good for the central part of the study area, including the areas of northern, northeastern, and southeastern Portland. Agreement also is favorable in the area of the Boring Hills. However, the ground-water model results indicate a substantial depth to water along the Willamette River; whereas, the current study uses a zero depth to water along the Willamette River. This difference may be a function of the relatively coarse horizontal discretization used for the ground-water model, which used a rectangular grid with cells 3,000 ft on a side. Other areas of disagreement between the current study and the ground-water model include the area extending from Troutdale southeast to Sandy and all areas south of Sandy, where the ground-water model appears to substantially overestimate ground-water elevation and, as a result, greatly underestimates the depth to water due to the lack of available control data in this area.
The depth-to-water map (pl. 1) can be used to help identify areas where the occurrence of existing or planned UIC systems may be less suitable. Two factors commonly used to evaluate the suitability of UIC systems at a particular location are the separation distance between the water table and the bottom of a UIC system and the particle size and character of the intervening subsurface materials. Greater separation distances tend to reduce or eliminate certain types of waterborne pollution, and for the same material thickness, finer grained materials, such as clay, tend to yield greater pollution reductions than coarser grained materials, such as gravel. The typical depth of a UIC system in the Portland area is about 30 ft, although UIC systems commonly are much shallower. Assuming a 10-foot separation requirement between the water table and the bottom of a typical UIC system, the minimum depth to water needed would be 40 ft (Oregon Department of Environmental Quality, 2005a, p. 37; 2005b, p. 4, 8). Regions where the depth to water is less than or equal to 40 ft might not have a 10-foot separation distance between the water table and any stormwater injection devices that may be present. UIC systems that do not meet the separation requirements may be considered for one or more of the following actions: additional evaluation with regard to the ability of the intervening subsurface materials to adequately protect the ground water; retrofitting, possibly by adding material such as control density fill (CDF) (a blend of cement, fly ash, sand, and water) to the bottom of the UIC system to increase the separation distance; or decommissioning.
Areas with a depth to water of 40 ft or less include the area adjacent to most rivers and streams, including the Columbia, Willamette, Clackamas, Tualatin, and Sandy Rivers, and Johnson, Fairview, Beaver, Kellogg, and Mount Scott Creeks (pl. 1). Other areas of concern are the area between the Columbia Slough and the bluff to the south, the area around Fairview Creek and several small lakes in the western part of Gresham, the area consisting of a former alluvial channel extending northward from the confluence of Johnson and Crystal Springs Creek (Trimble, 1963, p. 71-72), the area consisting of a former alluvial channel extending southward from the mouth of Kellogg Creek to the Clackamas River, the areas around and between Holgate Lake and Beggars Tick Marsh, and the area immediately south of the western end of Lake Oswego. Shallow depths to water can occur at moderate to high elevations but generally are associated with incised stream channels. Seasonal fluctuation in the water table may result in additional areas where the separation distance may be inadequate for UIC systems based on the assumptions previously discussed. Users requiring information on the seasonal extremes of depth to water can add or subtract one-half the range of seasonal water-table fluctuation (selected based on the appropriate hydrogeologic unit for the area, see section “Estimated Seasonal Water-Table Fluctuations”) from the estimated depth to water.
The elevation of the water table ranged from 11 ft NAVD 88 along most areas of the Columbia and Willamette Rivers to more than 2,000 ft NAVD 88 on the southern slopes of Larch Mountain (pl. 2). A visual inspection of the water-table elevation map shows that the configuration of the water table is similar to that of land surface. Subtle features as well as large, obvious features of the topography are recognizable and represented. The correlation coefficient (R value) for the spatial correlation between the interpolated water-table elevation and the land-surface elevation was 0.96 indicating a high correspondence between the two; that is, as land-surface elevation increases, water-table elevation increases. Water-table elevations are high for topographic features with high elevations, such as the Tualatin Mountains, Boring Hills, and the foothills of the Cascade Range. Water-table elevations are low for areas at low elevations adjacent to the Columbia and Willamette Rivers. These observations are consistent with the concept discussed in the section “Definition of the Water Table” that the water table is often a subdued replica of the land surface.
Comparisons with previous estimates of ground-water elevation in the Portland area generally are favorable. The largest body of recent work that provides extensive maps of ground-water levels is the study of the ground-water hydrology of the Portland Basin by the USGS. Maps consisting of hand-drawn contours of ground-water elevations derived from field measurements are presented by McFarland and Morgan (1996, 20-22, pl. 2-5) for the major hydrogeologic units in the basin. Many of the wells and surface-water features used in their analysis also were used for the current study. The contour map of ground-water elevations for the unconsolidated sedimentary aquifer (McFarland and Morgan, 1996, p. 20-21, pl. 2) shows modest agreement with the estimates from the current study in the area north and west of Powell Butte, although differences can exceed 50 ft in some areas. However, the contours presented by McFarland and Morgan (1996, pl. 2) are indicated as approximate and are relatively unconstrained due to a scarcity of data in this area, making comparisons difficult. Ground-water levels for the Troutdale gravel aquifer (McFarland and Morgan, 1996, p. 22, pl. 3) show a high degree of similarity for this area as well as throughout the area bounded by the Columbia and Clackamas Rivers to the north and south and by the Willamette and Sandy Rivers to the west and east. These contours generally are more highly constrained than those for the unconsolidated sedimentary aquifer, especially in the central part of the area, due to the greater availability of well data for this aquifer. Although the Troutdale gravel aquifer underlies the unconsolidated sedimentary aquifer throughout the terraced areas in the eastern Portland area and Multnomah County, the elevation of the top of the unit generally is close to that of the water-table elevation estimated in this study. The Troutdale gravel aquifer forms the surficial aquifer in many of the areas to the southeast. Many of the wells used in the current study for estimation of the water-table elevation that have open intervals in the Troutdale gravel aquifer also were used in the development of the ground-water level maps by McFarland and Morgan (1996).
Contours of simulated ground-water elevations for the unconsolidated sedimentary aquifer (Morgan and McFarland, 1996, p. 29-30, pl. 3) developed based on a ground-water flow model are in good agreement with the water-table elevations developed in the current study for most of the area. However, in the area immediately north of Powell Butte, the current study indicates a saddle in the water table for the area between Kelly and Powell Buttes that is not represented in the work by Morgan and McFarland (1996, pl. 3), though this may be a result of the resolution of the ground-water model or of the contour interval used. Contours presented for the Troutdale gravel aquifer by Morgan and McFarland (1996, p. 29-32, pl. 4) are in good agreement with the current study for most of the area, with the exception of the area south and east of the city of Sandy. Ground-water elevations for the ground-water model are substantially higher in this area than the values derived in this study. This area has few ground-water elevation measurements, and, as a result, the ground-water flow model and the present analysis are highly unconstrained in this area.
The water-table gradients (slope of the water table) shown on the water-table elevation map are a function of the land-surface relief, permeability of the geologic materials, and variations in recharge and discharge. Steeper (higher) horizontal gradients are indicated where the water-table elevation contours are more closely spaced. Relatively steep water-table gradients occur where the geologic materials are of low to moderate permeability such as the Tualatin Mountains, Boring Hills, and foothills of the Cascade Range, which often are associated with greater topographic relief. Gentler gradients occurring in areas of higher permeability and gentler relief are most common in the low-lying areas along major rivers and streams and in the terrace deposits extending in an area approximately bounded by the Columbia River to the north, the Willamette River to the west, and the Boring Hills to the southeast.
The approximate direction of shallow horizontal ground-water flow can be implied cautiously from gradients depicted on the water-table elevation map (pl. 2). The direction of ground-water flow is indicated as the direction perpendicular to the water-table elevation contours moving from areas of high to low water-table elevations. The water-table elevation map provides an indication of the direction of flow at the surface of the saturated zone but does not provide information regarding lateral or vertical gradients in the saturated zone and how flow directions may change with depth in the flow system. Such information is needed to determine the correct ground-water flow path from a specified location. The map may provide useful indications of the possible direction of shallow ground-water flow over short distances; however, use of a ground-water flow model such as that of Morgan and McFarland (1996) developed for the Portland Basin will provide a more reasonable depiction of the actual direction of ground-water flow in three dimensions over greater distances.
The overall direction of ground-water flow is toward the major ground-water discharge areas consisting of the Columbia, Willamette, and Clackamas Rivers (pl. 2). Local directions of ground-water flow generally are toward adjacent streams and rivers and appear to follow surface drainage patterns in most instances; however, Johnson Creek is a notable exception. Directions of shallow ground-water flow indicate movement toward the Sandy River in the upper reaches of the surface-water drainage, toward the Columbia River in upper midreaches, and toward the Willamette River in some parts of the lower midreaches. Although these patterns are indicative of shallow ground-water flow only, these flow patterns may provide insights on the low unit-area discharge (stream discharge divided by the area of the drainage basin) observed for some parts of Johnson Creek (K.K. Lee, U.S. Geological Survey, written commun., 2007).
Examination of the map of the water-table elevation also reveals the presence of ground-water mounds and depressions. Ground-water mounds are areas where ground water is moving radially away from the center of the mound. The presence of the mounds may be the result of topography, less permeable aquifer materials, and (or) the presence of recharge areas either due to the infiltration of precipitation or some other source, such as losing streams, UIC systems, septic systems, irrigation, or injection wells. Ground-water mounds in the Portland Basin are usually associated with recharge areas located at the top of hills or mountains such as can be seen at the Tualatin Mountains, Mount Tabor, or the Boring Hills (pl. 2). Many of these hills receive high rates of precipitation and also may be composed of less permeable materials.
Closed ground-water depressions are areas where ground water is moving radially toward the center of the depression, indicating ground-water discharge possibly as a result of losses to gaining streams, springs, evapotranspiration, or withdrawal wells. Ground-water depressions alternatively may be the result of discharge of ground water from the surficial aquifer by the downward movement of ground water into the underlying aquifer. Several small closed depressions in the Portland area are associated with gravel pits and quarries. Other small ground-water depressions may include the area on the southwestern side of Rocky Butte and several areas located between the western side of Mount Scott and Milwaukie (pl. 2). The two largest closed ground-water depressions are the area adjacent to the northwestern side of Powell Butte and an area near Sunshine Valley (pl. 2). The depression near Powell Butte may be the result of municipal ground-water pumping for a public-water supply which, until recently, occurred in this area. It is unknown whether the recent cessation of pumping will result in the recovery of ground-water levels in this area or whether the ground-water depression will continue to persist and can, therefore, be attributed to other causes. The depth and extent of the large depression (shown only as a nearly closed depression due to the contour interval) near Sunshine Valley is largely based on the water-level measurement of a single well. Whether this depression is a reasonable representation of the water table in this area or whether it is the result of the use of one or more wells that may be confined or have water levels that are perched is unknown. Many of the closed depressions are not well constrained by the available data and may be a result of inadequacies in the method of interpolation. Further data collection and analysis are needed to better define the depth and extent of closed ground-water depressions and to understand the causes contributing to the occurrence of these depressions.
The 127 wells used to evaluate seasonal water-table fluctuations have trimmed ranges of water-level fluctuations that ranged from 1 to 22 ft, with a mean of 7 ft. Small seasonal water-table fluctuations occur throughout the study area but are concentrated more heavily in the terraced areas between the Willamette and Columbia Rivers north and west of the Boring Hills (fig. 10), where the wells typically obtain water from the unconsolidated sedimentary aquifer. The largest seasonal changes generally occur in the Sandy, Boring, and Damascus areas with few exceptions. These large fluctuations, which ranged as much as 22 ft, occurred in the Troutdale gravel aquifer and older rock hydrogeologic units. The magnitude of the water-table fluctuations is primarily a function of changes in recharge, discharge, and the effective porosity of the aquifer. Mean annual recharge in the Portland area (Snyder and others, 1994, p. 30) is somewhat greater in the areas of larger water-table fluctuations. Discharge due to large volumes of water pumped for agricultural usage is also greater (Collins and Broad, 1993, p. 11). However, the differences in recharge and discharge are greatly magnified in this area of large water-table fluctuations because of the lower effective porosity of the shallow aquifers.
A categorization of the seasonal water-table fluctuations by hydrogeologic unit was evaluated and developed as a result of the qualitative relation observed between seasonal water-table fluctuations and the hydrogeologic unit present. The hydrogeologic unit in the range of fluctuation of the water table (referred to as the “zone of fluctuation” and defined as the part of the aquifer between the minimum and maximum trimmed values of measured depth to water for each well) was determined. The minimum and maximum trimmed depths to water for each well used in the seasonal water-table fluctuation analysis were compared against the depth and thickness of the hydrogeologic units identified for each well (Swanson and others, 1993, p. 8) to determine the hydrogeologic unit in the zone of fluctuation. The hydrogeologic units for some wells located outside the extent of the work by Swanson and others (1993) were identified based on the hydrogeologic classifications of Conlon and others (2005, p. 7-23), and the hydrogeologic units subsequently were correlated to the hydrogeologic units defined in the current study (Conlon and others, 2005, p. 8).
Effective porosities of the hydrogeologic units in the Portland Basin were estimated by Hinkle and Snyder (1997, p. 14 and p. 39-47) based on hydraulic conductivity values calibrated for the USGS Portland Basin ground-water flow model (Morgan and McFarland, 1996, p. 17-19). The mean effective porosity for each hydrogeologic unit in the Portland Basin is shown in table 2. The mean seasonal water-table fluctuation by hydrogeologic unit can be estimated by dividing the mean recharge rate of 22.0 in/yr for the Portland Basin (Snyder and others, 1994, p. 30) by the effective porosities estimated for each hydrogeologic unit. The resulting values of estimated seasonal water-table fluctuation calculated from porosity and recharge are similar to the mean of the measured values of seasonal water-table fluctuations for each hydrogeologic unit (table 2 and fig. 11). This agreement helps to support the conceptualization that the water-table fluctuations are influenced greatly by the hydraulic properties of the hydrogeologic unit in the zone of fluctuation.
U.S. Department of the Interior | U.S. Geological Survey
URL: http://pubs.usgs.gov/sir/2008/5059
Page Contact Information: Publications Team
Page Last Modified: Thursday, 10-Jan-2013 18:51:35 EST
