The greater Umpqua River Basin is located in Douglas County in southwestern Oregon and stretches from the Cascade Mountain crest to the Pacific Ocean at Reedsport, Oregon (Fig. 2). The spawning sites of the petitioned cutthroat trout populations are in the North and South Umpqua Rivers and their tributaries, which combine to form the mainstem Umpqua River about 11 km northwest of Roseburg, Oregon. The drainages of the North and South Umpqua Rivers together make up about 2/3 of the greater Basin drainage, and each river is about 170 km long. The mainstem Umpqua River flows in a northwesterly direction another 180 km to the ocean. Together, the three rivers form one of the longest coastal basins in Oregon, approximately 340 km in length, with a drainage area of over 12,200 sq. km. Major tributaries of the mainstem Umpqua River include Calapooya (River Kilometer [RKm] 164), Elk (RKm 78), and Scholfield Creeks (RKm 18) and the Smith River (RKm 18). The estuary of the Umpqua River is one of largest on the Oregon coast and has a large seawater wedge that extends as far inland as Scottsburg, Oregon at RKm 45.
The North Umpqua River has long been known for its rugged terrain, world class trout fishing, and poetic inclinations:
The Umpqua is a quality of spirit. Its shining length is scarred with lava ledges and outcroppings, its folded bedrock and igneous serrations polished by centuries of snowmelt and spates. Its gorge is still cloaked in forest, their vaulted choirs of spruce and sugar pine and fir softly carpeted with moss and fiddlebacks and pine needles. The forests are mirrored in the emerald pools, all stillness and shadows, with the hush of Mont Saint-Michel or Chartes (Schwiebert 1979).
What makes these "emerald pools" possible is that the North Umpqua River, almost alone among the coastal rivers of Oregon, begins and remains for a portion of its length high enough in the Cascade Mountains that its snowpack usually lasts until summer. This snowpack melts into porous soils and sustains a strong summer water flow. The headwaters of the North Umpqua River are located at an elevation of over 1,830 m on the slopes of the High Cascade Mountain Range near Maidu Lake. The river is separated from the headwaters of the Willamette River by the Calapooya Mountains. The South Umpqua River also begins at an elevation of around 1,830 m on the slopes of the Rogue River Mountain Range, but quickly descends (Hayes and Herring 1960; USDA 1990, USDA and BLM 1992). Only 3% of the South Umpqua River watershed lies above 1,700 m, but over 20% of the North Umpqua River watershed exceeds this elevation (Hayes and Herring 1960).
There is little rain June to October in southern-central Oregon, and most regional streams barely flow during those months; but because of the snowmelt stored in deep pumice and other volcanic soils, water flow in the upper North Umpqua River remains relatively constant, with deep, swift, and cold water. Downstream in the North Umpqua River, and even more so in the South Umpqua River, the flow is erratic, with shallower, slower moving, and warmer water. The importance of the snowmelt stored in deep pumice and other volcanic soils of the North Umpqua River can be seen in that the drainage area feeding the North Umpqua River at the stream gauge above Copeland Creek is the same size as the South Umpqua River drainage above Tiller (1,230 km2), but summer flow on the North Umpqua River is 20 times that of the South Umpqua River (Hayes and Herring 1960).
In recent years, another factor has slowed the reduction of flow during summer in the North Umpqua River. Presently, the entire summer flow of the upper North Umpqua River goes through Pacific Power and Light's (PP&L) Soda Springs powerhouse below Soda Springs Dam; PP&L policy limits river fluctuation to approximately 50 cfs change per hour when the river is flowing above 1,000 cfs in summer (USDA and BLM 1992).
Long-term water temperature data are available from the upper North Umpqua River in Steamboat Creek Basin (from 1969) and from the lower river at Winchester Dam (from 1946). The data from Steamboat Creek have been extensively analyzed and modeled by Hostetler (1991) and Holaday (1992), primarily to evaluate the effects of forestry management practices on stream temperatures. Dambacher (1991) evaluated the effect habitat changes in the Steamboat Creek Basin had on steelhead abundance and distribution.
Although data are lacking on stream temperatures prior to extensive logging in the Basin, post-logging increases in water temperature in the North Umpqua River Basin were identified in 1969 by Brown et al. (1971). They found substantial increases in stream temperatures (over 8° C in a 1280 m distance of Cedar Creek) by measuring areas above and below clearcuts. A variety of studies (reviewed in Meehan 1991) have reported that stream temperatures increase in a forest after clear-cut logging. Maximum temperatures are observed several years after the logging has occurred, followed by gradual recovery and decreasing stream temperatures as forest canopies and riparian vegetation regrow.
In 1969, after most clear-cut logging in riparian areas of the Basin had ended (Hostetler 1991), the USFS placed temperature recording devices at several locations in Steamboat Creek Basin. These devices provide a 20-year record of stream temperatures during recovery from riparian logging in the Basin. Hostetler used a time-series to model components of this stream temperature decline over the period from 1969 to 1989. He found a significantly decreasing temperature trend in all streams and reaches affected by logging. Streams unaffected by logging did not have a significant temperature trend. He also found that since the mid-1980s, some stream temperatures had increased in the upper and lower reaches of Steamboat Creek, although air temperature had not. Overall, Hostetler concluded that streams affected by pre-1969 logging were still recovering from elevated stream temperatures caused by loss of riparian vegetation and shading canopies. Hostetler cautioned that although he found stream temperatures had generally decreased, most streams were still near the upper limit of tolerance for juvenile steelhead.
Holaday (1992) also evaluated stream temperatures from 1969 to 1990 and found that maximum daily stream temperatures in lower Steamboat Creek did not change significantly during the time period. He further compared these results with stream temperature readings taken in 1960 from an area not previously clear-cut. The results from this analysis suggested to Holaday that historical maximum stream temperatures in lower Steamboat Creek may have been high.
The direct effects of stream habitat changes on distribution, abundance, and movement of juvenile steelhead in Steamboat Creek Basin were evaluated by Dambacher (1991). He reported that in 1987 and 1988, 5-day mean maximum summer temperatures in tributaries and mainstem segments of Steamboat Creek. He found tributary temperatures from 10°C to over 20° C, while Steamboat Creek mainstem temperatures ranged from 13° to over 27°C. All mainstem temperatures exceeded the preferred temperatures (13°C) for steelhead and even exceeded the lethal limit (24°C) for days at a time.
Dambacher (1991) suggested that steelhead and cutthroat trout were able to survive in these warm waters by moving to lower-temperature refugia. He found that when tributaries containing cooler water (such as Big Bend Creek or Canton Creek) entered the river mainstem, water temperatures in the mainstem were reduced by as much as 3°C . He also found distinct diurnal water temperature regimes in the Basin , where daily minimum temperatures were as much as 5°C lower than maximum temperatures. Fish would remain at the confluence of cooler temperature tributaries during the day and move upstream at night during minimum river temperatures.
Dambacher (1991) also analyzed the water temperature data collected from 1969 to 1989 at USFS temperature monitoring sites in Steamboat Creek. He found that although there was a decreasing trend in the mean 10-day maximum summer water temperatures in Steamboat Creek Basin, the water temperature still consistently exceeded 14°C in the tributaries and almost 20° in the mainstem. Although data was not available for every year from every monitoring site, Dambacher found that in 12 of the 19 years reported, mainstem temperatures above Canton Creek exceeded the lethal limit for steelhead. Dambacher concluded that the decreasing trend he observed represented a cumulative response from recovery of multiple clear-cut sites upstream of the temperature monitoring stations. Further, he believed that the "single most important factor limiting juvenile steelhead production in the Steamboat Creek Basin is high summer water temperature" (Dambacher 1991, p. 98). River-migrating or sea-run cutthroat trout would have to pass through the Steamboat Creek drainage or other basins with similar forestry management regimes (Dambacher 1991, Holaday 1992) to reach spawning areas in upper tributaries of the North Umpqua River. Further, cutthroat trout apparently are less tolerant of high water-temperatures and experience a lower lethal limit than steelhead (Golden 1975, Bell 1986).
Comprehensive analysis or modeling of these data has not been published, but simple regression analysis of the average yearly and maximum yearly temperatures from 1946 to 1993 reveals a positive trend for both parameters (R2 = 0.19, P = 0.002 for average temperature and R2 = 0.21, P = 0.001 for maximum temperatures) (Fig. 4). When the data are divided at 1969 (the year when Hostetler and Holaday began their analyses), neither the average temperatures for the years 1946 to 1968 nor for the years 1969 to 1993 show a trend (R2 = 0.002, P = 0.83 and R2 = 0.07, P = 0.2 respectively). However, analysis of maximum temperatures reveal a warming trend from 1946 to 1968 (R2 = 0.44, P = 0.0005), but not from 1969 to 1993 (R2 = 0.004, P = 0.76) (Fig. 4).
Figure 4. Average yearly water temperature for 1946 to 1992, and maximum yearly water temperature for 1946 to 1993 taken at Winchester Dam (Umpqua RKm 116) on the North Umpqua River (ODFW 1993a). The regression analysis suggests that overall average and maximum water temperatures have significantly increased since 1946 (R2 = 0.19, P = 0.002 for average temperature, and R2 = 0.21, P = 0.001 for maximum temperatures).
Observations of maximum monthly temperatures during September, October, and November from 1946 to 1993 do not show any distinct trend in water temperatures (R2 = 0.054, 0.01, and 0.04 respectively and P = 0.11, 0.43, and 0.16 respectively), nor is a trend evident when the years are divided by pre- and post-1969. (Fig. 5)
Correlation of these trends with fish passage is difficult to interpret (Fig. 4; see also subsection Temperature Tolerance in Cutthroat Trout and the Discussion and Conclusion sections), but sustained increases in river temperatures occurred at the same time as the collapse of cutthroat trout numbers crossing the dam. Prior to 1954, the highest maximum July temperature was 21.7°C; by July 1958, the maximum temperature was 25°C.
Figure 5. Monthly maximum water temperatures for July, August, September, and October. Temperatures were taken at Winchester Dam on the North Umpqua River from 1946 to 1993.
Regardless of the trend, water temperature readings during many years since the mid-1950s, reveal that maximum water temperatures approached experimental lethal limits for cutthroat trout (Golden 1975, Bell 1986; see also subsection Temperature Tolerance in Cutthroat Trout) at Winchester Dam during the times when sea-run trout were passing. This would suggest that during certain times of the year, a major portion of the North, South, and mainstem Umpqua River exceeded the preferred temperature range of cutthroat trout. However, temperature increases during September at Winchester Dam, do not seem a reasonable explanation for the break between summer and fall peaks in cutthroat trout migration across Winchester Dam. Water temperatures in September at Winchester Dam are consistently lower than July and August (when fish pass the dam, see Fig. 4) and no trend of increasing water temperature was found at the dam in September from 1946 to 1993.
On the North Umpqua River, anadromous fish passage extends to around RKm 113, where historically impassable natural barriers (and presently the 35-m high Soda Spring Dam) terminate salmon migrations (Lauman et al. 1972).
The dam at Winchester (RKm 11) on the North Umpqua River was built in 1890 and considered a "definite barrier" to fish passage at low water; however, fish could apparently surmount the dam through a modified spillway at the north end (FCO and OSGC 1946). Although the fish ladder at the dam was modified in the early 1980s to improve fish passage, issues regarding the use and future of the dam are still highly controversial (Blumm and Kloos 1986). Present fish passage facilities are reported to be satisfactory at all flow levels (D. Loomis, see footnote 2).
Large natural barriers to fish passage in the Umpqua River watershed, such as Smith River Falls, South Umpqua Falls, and Steamboat Falls, have been laddered to facilitate fish passage (Lauman et al. 1972; D. Loomis, see footnote 2).