Results of Seepage Analyses


Seepage characteristics of the Clackamas River were determined by using a combination of three methods: (1) analysis of previous studies that made synoptic measurements, (2) comparisons of daily means of long-term, continuous streamflow records, (3) a synoptic seepage study in 2006. 


Results from Analysis of Previous Studies


Previous studies, with various primary purposes, made synoptic streamflow measurements of the Clackamas River and tributaries. Although some of these studies provided insight into seepage characteristics, on the basis of streamflow measurements and continuous streamflow records collected during base-flow or near-base-flow conditions, some were not suitable for this analysis owing to fluctuating streamflow 
and/or lack of tributary measurements.


Streamflow was measured at the Clackamas River near Clackamas site (site 25) as part of a statewide inventory during the droughts of 1987 and 2001, and was compared to streamflow records at Estacada (site 5, table 2) (Alexander and others, 1987; Herrett and others, 2001). Intervening tributaries (Eagle, Deep, Clear, and Rock Creeks [sites 11, 14, 20, and 23, respectively] were not measured). For the current analysis, tributary flow of 42 ft³/s measured in late summer 2006 was assumed to be equivalent to late summer flow in 1987. In August 1987, the river gained 104 ft³/s (16 percent), which was greater than the uncertainty. In September 2001, large pulses precluded comparison of flows. 


Streamflow measurements of the Clackamas River were made as part of a time-of-travel study in 1992 (Lee, 1995). Streamflow measurements at Barton Bridge (site 13) and Eagle Creek (site 11) in May 1992, compared to streamflow at Estacada (site 5), indicated a loss of 150 ft³/s (11 percent), greater than the uncertainty (table 2). Streamflow at the Barton site was measured in July 1992, but a pulse from River Mill Dam affected the flow.


Streamflow measurements of the Clackamas River and tributaries were made in May, July, and September 1998 associated with a water-quality study (Carpenter, 2003). 
The measurements covered a broad area, extending from upstream of Estacada to the mouth of the river. Unmeasured tributaries precluded use of the May measurements. Knowledge of the travel time of pulses at similar streamflow magnitude, derived from recent streamflow data from the Estacada site (5) and Oregon City site (32), suggests that the July 1998 measurements may have been made during 
unsteady conditions.


Streamflow measurements made at Barton Bridge (site 13), Carver (site 21), and Oregon City (site 33) in September 1998 were compared to recorded streamflow at Estacada. Tributary streamflows were not measured but were estimated to be equivalent to the late summer streamflows of September 2006. The difference in streamflow from Estacada to Barton Bridge was less than the uncertainty. The gain in streamflow between Barton Bridge and Carver of 200 ft³/s (19 percent) was greater than the uncertainty. Although the Clackamas River was measured at Oregon City (site 33), this measurement was invalidated by fluctuating backwater conditions from the Willamette River. 


In summary, analysis of streamflow measurements made during previous studies indicated some gains or losses that exceeded the uncertainty. The August 1987 measurements indicated a gain of 16 percent from Estacada to Clackamas. The May 1992 measurements indicated a loss of 11 percent from Estacada to Barton Bridge. The September 1998 measurements indicated a 19 percent gain from Barton Bridge to Carver and, on a larger scale, a gain of 25 percent for the reach from Estacada to Carver. 


Results from Analysis of Long-Term 
Streamflow Data


Long-term streamflow data consisting of records of daily streamflow at several sites in the Clackamas River basin enabled the estimation of seepage to the stream. 


Seepage in the lower Clackamas River during the annual base-flow period was determined by comparing streamflow at the Estacada site (site 5) to streamflow at sites downstream. The downstream sites are Clackamas River at Clackamas (site 25), from 1963 through 1983, and Clackamas River near Oregon City (site 32) from 2001 through 2006. After accounting for tributary inflows and municipal withdrawals, seepage was inferred from the difference between annual low streamflow at the Estacada site and the site downstream. Streamflow in the Clackamas River is lowest in late summer, and late summer streamflow is similar to what it was before construction of the dam on the Oak Grove Fork at Timothy Lake in 1958 (fig. 4). The August–September average streamflow (the average of all daily mean values during the comparison periods) is about 1,000 ft³/s. Whereas synoptic measurements describe gains and losses on a single day on a mile-by-mile basis, this analysis provides seepage estimates for each year of concurrent streamflow data between the sites, about 20 mi apart.


Tributary inflow was estimated and subtracted from streamflow at the most downstream site. On the basis of streamflow measurements made in September 2006, the combined streamflow of Eagle, Deep, Clear, and Rock Creeks (sites 11, 14, 20, and 23, respectively), was 42 ft³/s, or about 4 percent of the long-term average streamflow at Estacada during August–September. 


Daily total municipal withdrawals were added to daily mean streamflow at the Oregon City site (site 32). This step was not necessary for the early period because no municipal withdrawals occur upstream of the Clackamas site (site 25). Withdrawals do occur between the Clackamas and Oregon City sites, so streamflow in the recent period (at the Oregon City site) was adjusted on the basis of hourly withdrawals for three drinking-water treatment plants.


Gains and losses in streamflow from Estacada to either Clackamas (1963–83) or Oregon City (2001–2006) were small and, on average, were less than the uncertainty (table 2). On a year-by-year basis, gains greater than the uncertainty occurred in 6 of the 26 years, and no losses were greater than the uncertainty (fig. 5). The largest gain was 150 ft³/s, or about 14 percent.


Gains in streamflow between sites in the upper basin during late summer are large in comparison to the minimal net gains in the lower 20 mi of the river. The Clackamas River basin lies within three hydrogeologic areas (Conlon and others, 2005). The upper basin, the part of the basin upstream of Estacada, is primarily within the High Cascade and Western Cascade areas (fig. 1). The lower basin is mainly within the Willamette Lowland area. For this analysis, streamflow was normalized for the size of the contributing area between sites by dividing the calculated streamflow between sites by the size of the contributing area. The locations of the streamflow monitoring sites and the associated contributing areas do not exactly coincide with the boundaries of the hydrogeologic areas. As a result, the headwaters of some tributary basins are partly within the adjacent upgradient area. This analysis is based on average August–September streamflow from water years 1963–83 and 2001–06. No adjustments were made for tributary inflow due to lack of data.


Streamflow gains decreased from upstream to downstream. The gain in streamflow associated with the High Cascade area was calculated as the difference in late-summer streamflow between Oak Grove Fork at Government Camp (Timothy Lake) (site 1) and Oak Grove Fork above Power Plant Intake (site 2) (table 3). The gain in streamflow was 3.2 (ft³/s)/mi², greater than the uncertainty.


The part of the Clackamas River basin draining the Western Cascade area is represented in two segments, (1) an upper segment from Oak Grove Fork above Power Plant intake (site 2) to Clackamas River above Three Lynx Creek (site 3) and (2) a lower segment from Clackamas River above Three Lynx Creek (site 3) to Clackamas River at Estacada (site 5). Gains were about 1.2 (ft³/s)/mi² in the upper segment and 0.8 (ft³/s)/mi²; both were greater than the uncertainty. 


Streamflow gain in the lower Clackamas River basin, which drains part of the Willamette Lowland area, was calculated as the difference in late-summer streamflow between the sites at Estacada (site 5) and sites near the mouth of the Clackamas River (sites 25 and 32 for the 1963–83 and 2001–06 periods, respectively). The increase in streamflow of 0.3 (ft³/s)/mi² was less than the uncertainty. 


The relatively large gain in streamflow between sites in the upper basin was primarily attributable to the contribution of groundwater, which is a key source of basinwide streamflow during the summer. Groundwater inputs in the lower basin, although smaller than those of the upper basin, may, nonetheless, affect water quality in the lower basin. Carpenter (2003) identified high nutrient contributions to the lower Clackamas River from shallow wells and seeps. Owing to the relatively high main-stem streamflow and associated uncertainty, the volume of these inputs may be difficult to determine using the methods described in this report.


2006 Seepage Study


The 2006 seepage study took place over a 2-day period, September 6-7, 2006. Climate conditions were seasonably warm and dry, with only 0.1 in. of precipitation during the previous month at Portland (National Climatic Data Center, 2011). Clackamas River streamflow in August and September is typically the lowest for any two contiguous months of the year (fig. 4). Streamflow at Estacada on September 5, the day prior to most streamflow measurements, ranged from 740 to 810 ft³/s (fig. 3). Streamflow on September 6 ranged from 755 to 780 ft³/s, or about a 2 percent deviation from the daily mean streamflow of 768 ft³/s. Some measurements were made the following day to confirm the September 6 measurements. Streamflow on September 7 ranged from 766 to 785 ft³/s, except for a rise from a dam release that occurred later in the day after measurements were completed. 


The seepage study focused on the lower Clackamas River, between Estacada and Oregon City (fig. 6). Tributaries Eagle, Deep, Foster, Richardson, Clear, Rock, Sieben, Carli, and Cow Creeks were measured. Of the tributaries, only Eagle, Deep, Clear, and Rock Creeks (sites 11, 14, 20, and 23, respectively) were used in seepage calculations because the sum of streamflow for the remainder of the tributary streams was less than 0.1 percent of the streamflow of the Clackamas River. There is one withdrawal from the study reach, at site 27. 


The lower Clackamas River was measured at 13 sites on September 6, 2006, followed by measurements at 2 sites the next day (table 4). Many of the site-to-site differences were within the range of uncertainty, indicating that if gains or losses occur, they are small relative to total streamflow. Gains and losses, indicated by site-to-site differences in streamflow that were greater than the uncertainty, occurred at about a quarter of the sites. The maximum gain or loss on a site-to-site basis was 12 percent. 


The September 6, 2006, measurements were grouped into reaches on the basis of seepage characteristics and the geomorphic setting. Six reaches were designated: A (beginning near Estacada) through F (terminating near Oregon City) (fig. 2, table 2). Gains or losses greater than the uncertainty occurred in three of the reaches. 


Reach A extended from Estacada to just upstream of Eagle Creek (fig. 2). Streamflow measurements in that reach were made at sites 6, 7, 8, 9, and 10. The change in streamflow between sites 6 and 10 was less than the uncertainty (table 2). 


Reach B extended from just upstream of Eagle Creek to Barton Bridge (fig. 2). Measurements were made at sites 10, 12, and 13, and in Eagle Creek (site 11). The loss from the river at sites 10–13 was 58 ft³/s (8 percent), equal to the uncertainty, and is considered significant. This loss may be attributable to infiltration through permeable streambed sediments. The stream channel in this area is subject to lateral migration and changed abruptly during the 1996 flood, capturing a near-stream gravel mine and cutting off a former meander created by a long dike used to protect the gravel mine (Wampler, 2004). Die-off of riparian vegetation following the 1996 avulsion suggests a linkage of the shallow aquifer through gravel deposits to the stream. In addition to changes in the active stream channel, this reach, compared to reach A (adjacent upstream) and reach C (adjacent downstream) features a broader, more meandering stream channel configuration (Burkholder, 2007).


Reach C is near Barton. Main-stem measurements were made at sites 13, 15, and 17, and in Deep Creek (site 14). The change in streamflow between sites 13 and 17 was less than the uncertainty. 


Reach D extends from near Barton to near Carver, just upstream of Rock Creek. Measurements were made at sites 17, 19, and 22. Clear Creek (site 20), which enters in this reach, was also measured. The gain between sites 17 and 22 was 99 ft³/s (14 percent), greater than the uncertainty. Streamflow at the downstream end of this reach was the greatest for any site measured on September 6. The gain in reach D may be attributable to the basin-scale constriction in the landscape at Carver, where the Clackamas River passes through a notch between hills to the southwest and northeast, and subsurface flow may be directed back to the active stream channel. 


Reach E is from near Carver to near Clackamas. Measurements were made at sites 22 and 25. Rock Creek (site 23) enters in this reach and was measured. The loss of 104 ft³/s (12 percent) was greater than the uncertainty. The loss could be attributable to recharge through the streambed to alluvial and Pleistocene Clackamas River terrace deposits. The terrace deposits may occupy part of an abandoned (pre‑catastrophic flood) stream channel extending westward from the northernmost extent of the Clackamas River at about river mile 6.0 (Madin, 1990).


Reach F, the most downstream reach, extends from near Clackamas to near Oregon City. Main-stem measurements were made at sites 25 and 28. There is one withdrawal in this reach. The change in streamflow between sites 25 and 28 was less than the uncertainty.


Streamflow measurements were made the following day (September 7, 2006) at sites 25 and 28. These measurements, coupled with continuous streamflow data from the Estacada site (5); estimated tributary inflow based on measurements from the previous day; and municipal withdrawal, confirmed the results at these sites from the previous day (fig. 6). The net change in streamflow between sites 5 and 25 (incorporating reaches A through E), and between sites 25 and 28 (reach F) were less than the uncertainty. As a whole, the net change in streamflow of the Clackamas River from Estacada to near Oregon City (reaches A through F) on September 6 and 7 was less than the uncertainty.


First posted November 28, 2011