U.S. Geological Survey

By David S. Morgan and William D. McFarland

USGS Water-Supply Paper 2470-B, 83 pages, 29 figures, 4 tables, 9 plates

Available from U.S. Geological Survey, Branch of Information Services, Box 25286, Denver, CO 80225 (303-202-4700).

Abstract

A numerical model of the ground-water flow system in the Portland Basin, Oregon and Washington, was used to (1) test and refine the conceptual understanding of the flow system, (2) estimate the effects of past and future human- caused changes to ground-water recharge and discharge on ground-water levels and stream-flow, and (3) determine priorities for ground-water monitoring and data collection that would facilitate improvements in the utility and accuracy of the model. The model covered an area of 981 square miles that includes most of Multnomah County, Oregon, and Clark County, Washington, as well as parts of Clackamas, Washington, and Columbia Counties in Oregon, and Skamania County in Washington

Model results were most sensitive to the horizontal hydraulic conductivity of the Troutdale gravel aquifer and the Troutdale sandstone aquifer and to the vertical hydraulic conductivity of the confining units and undifferentiated fine-grained sediments. The model was insensitive to the hydraulic characteristics of the unconsolidated sedimentary aquifer because of the strong control that streams and rivers have on water levels in the unit.

The model was calibrated using time-averaged data for 1987-88. Horizontal hydraulic conductivity, vertical hydraulic conductivity, and streambed and riverbed conductances were adjusted until a reasonable fit was obtained between simulated and observed water levels and seepage to streams. Final horizontal hydraulic conductivities had median values ranging from 11 to 110 ft/d (feet per day) in the four aquifer units and from 1 to 8 ft/d in finer grained units. Horizontal to vertical anisotropy ratios ranged from 100:1 in the aquifer units to 1,000:1 in the finer grained units. Approximately 410 observed water levels were available for comparison with simulated water levels in the four aquifer units. The root-mean-squared error ranged from 54 feet in the unconsolidated sedimentary aquifer to 74 feet in the Troutdale sandstone aquifer. Simulated seepage to streams agreed closely with observed seepage at 67 measurement sites.

Recharge to the basin in 1987-88 consisted primarily of 1,440 ft3/s (cubic feet per second) from direct infiltration of precipitation, but was augmented in unsewered urban areas by 62 ft3/s from runoff to drywells and 27 ft3/s from on-site waste-disposal systems. Forty-nine percent of the recharge in the basin enters the system through the Troutdale gravel aquifer and 22 percent enters through the unconsolidated sedimentary aquifer. Simulated discharge in 1987-88 was chiefly by seepage to small rivers and streams (971 ft3/s), followed by seepage to the Columbia and Willamette Rivers and other large water bodies (457 ft3/s) and by pumpage (166 ft3/s). The unconsolidated sedimentary aquifer and Troutdale gravel aquifer supplied 85 percent of 1987-88 pumpage. Most of the ground-water flux through the aquifer system occurred within these two units with 75 percent of total recharge to the system entering through them and 80 percent of total discharge leaving through them in 1987-88. Recharge under predevelopment conditions was 180 ft3/s (12 percent) more than in 1987-88 due to the absence of impervious surfaces associated with urbanization. Simulation of the effects of the increased recharge and no well discharge indicates that water levels may have declined as much as 50 feet in the Troutdale gravel aquifer in southern Clark County in response to municipal pumping. Declines similar in magnitude were simulated in the gravel aquifer in the Troutdale gravel aquifer and Troutdale sandstone aquifer in the vicinity of Sandy, Oregon, in response to pumping for irrigation. Since predevelopment conditions, the combination of reduced recharge and increased pumpage may have reduced discharge to large rivers by 25 percent and reduced discharge to small rivers and streams by 16 percent.

One hypothetical condition for future ground-water development was simulated to test the effects of additional pumping stress on the ground-water system. Pumpage estimates in this condition were based on projected municipal supply demands in Clark County through the year 2010 and on limited use of the city of Portland Columbia South Shore well field. The resulting pumpage was 55 percent (92 ft3/s) greater than 1987-88 pumpage. Equilibrium water-level declines were as much as 20-40 ft within the Troutdale gravel aquifer in Clark County and in the Troutdale sandstone aquifer underlying the Columbia South Shore well field. At equilibrium, much of the additional pumpage was supplied by capture of discharge to the Columbia River (14 ft3/s) and by additional recharge induced from the Columbia River (26 ft3/s).

Further data collection through ground-water-level monitoring and periodic updating of well discharge and recharge from on-site waste-disposal systems and drywells would facilitate future improvements in the model developed in this study.