Relations Between Streamflow, Geomorphology, and Ecology

Key Elements

• Channel morphology and habitat availability depend on the “effective discharge,” (or channel-forming flow typically equivalent to or slightly greater than bankfull discharge), as well as the magnitude, frequency, and timing of less frequent flood events.
• The ability of the channel to respond to various magnitude flood events by creating or enhancing physical habitat depends upon other natural and anthropogenic factors, including physiography, bank materials, sediment supply, and availability of large wood.
• In the lower and middle McKenzie River basins, the elimination of small and large floods has contributed to a net reduction of large pools.
• High magnitude streamflows that exceed bankfull discharge (small and large floods) could be effective in the lower McKenzie River basin in creating and maintaining existing offchannel side channels, alcoves, and other secondary channel habitats.
• Gravel bars along the upper and middle McKenzie River basin reaches likely are less sensitive to discharge regime than bars in the alluvial reaches of the lower basin because steep upper basin reaches support high transport capacities and have fewer depositional settings where gravel bars historically formed.
• Spring Chinook salmon and bull trout are sensitive to the combination of discharge, stream velocity, and stream temperature. Water temperatures below seasonal norms and reduced streamflow in the spring and summer can potentially reduce the growth of fish rearing and affect the timing of smolt outmigrations and adult upstream returns. Warmer and higher streamflow in the fall affect spawning adults and can cause earlier egg hatching and fry emergence leading to reduced survival.
• Channel and floodplain simplification and the decrease in overbank streamflow events are factors in the decline of Oregon chub and Pacific lamprey, which inhabit low-velocity backwaters and isolated floodplain habitats.
• Oregon chub, western pond turtle, and red-legged frogs require diverse offchannel habitats, including side channels, sloughs, oxbow lakes, ponds, and wetlands, which are geomorphic features created and maintained by the interaction of small-large floods.
• Floodplain tree species, black cottonwood and white alder, require high winter and spring streamflow events for seed dispersion, generation of bare sunlit soil patches for germination, and adequate water tables for growth and survival.

In their seminal paper “The natural flow regime,” Poff and others (1997) describe streamflow as the “master variable” because of its strong influence on many critical physicochemical characteristics of rivers, including water temperature, geomorphology, and inchannel and offchannel habitat diversity. These physicochemical responses to streamflow quantity and timing mediate all aspects of aquatic and riparian ecology. Streamflow also directly affects species ecology through, for example, its effect on bedload movement and scour, stream velocity, and change in volume of aquatic habitats. The key streamflow, geomorphology/habitat, and ecology linkages for the McKenzie River basin (as outlined in previous research and literature reviews) are presented below.

Flow and Physical Habitat

Although it is difficult to prescribe precise relations between streamflow and physical habitat because of wide variations in channel morphology, local hydraulics, sediment characteristics and other factors, the literature provides a general understanding of how different magnitude streamflow events affect channel morphology, which in turn influences the availability and complexity of physical habitat. The discharge needed to create and maintain various aspects of physical habitat has been described in various ways by previous authors. For example, “effective discharge” (the flow, or range of flows, that transport the most sediment over the long term; Wolman and Miller, 1960) is often considered to be the channel-forming discharge. Effective discharge is typically considered equivalent to bankfull streamflow, although subsequent authors have shown that effective discharge may actually be greater than bankfull streamflow for gravel bed rivers (Emmett and Wolman, 2001). Other studies illustrate the importance of floods as catalyzing channel change (Grant and others, 1984). Channel complexity is not only a function of the magnitude of individual floods, but their timing and sequence (Wolman and Gerson, 1978; Tal and others, 2004; Wallick and others, 2007). On regulated rivers for which flood flows have been reduced, Kondolf and Wilcock (1996) describe ways in which to link specific sediment or channel maintenance objectives with reservoir releases designed to mimic natural floods (termed “flushing flows”). Stakeholders developing environmental flow guidelines for the McKenzie River basin can consider the magnitude of individual streamflow events when setting objectives for physical habitat and their sequence and timing.

Pool Habitat

It is difficult to differentiate the individual effects of different magnitude flows on pool characteristics along the McKenzie River because there are few studies to draw from, and the two most comprehensive studies of pool habitat (Minear, 1994; Sedell and others, 1991) each span one relatively long time period that encompassed a wide range of flood events as well as other natural and anthropogenic effects. Although more work is needed to better evaluate the role of different magnitude flow events on pool size, depth, and cover, findings from other river systems can be used to develop general hypotheses for the McKenzie River basin.

Pool scour and maintenance occurs under conditions of velocity reversal, in which the riffle gradient decreases (resulting in lower stream velocity) while velocity increases in the pool, leading to pool deepening (Keller 1971, Emmett and others, 1985; O’Connor and others, 1986; Thompson and others, 1999). In general, studies indicate that such conditions occur when discharge exceeds bankfull discharge; however, the magnitude of discharge required for pool scour along the McKenzie River is unknown and could be high. Wesche (1991) report that for mountain streams in Wyoming, scour did not occur until streamflows were 12 times the magnitude of bankfull discharge. Aggradation of pools by fine sediments might occur either on the receding limb of high streamflow events (Bowman, 1977) or during low streamflows (Jackson and Beshta, 1982). In addition, pool complexity is derived from a variety of other factors such as large wood, which requires higher streamflows (perhaps on the order of bankfull or greater) for recruitment and mobilization.

Along the McKenzie River study area, high streamflow events (such as small and large floods) that exceed bankfull probably have the greatest effect on pool complexity and availability (table 21). In the upper McKenzie River basin, there has been little change in the overall magnitude and frequency of peak streamflows (table 12), but for individual years (such as 1995), streamflow reduction by the hydropower dams could result in the elimination of one or more pool-scouring events (fig. 29). In the middle McKenzie River basin, small floods essentially have been eliminated (table 15), and peak discharge in a typical year is only two-thirds of bankfull (fig. 29). This substantial reduction in flow events ranging in magnitude from bankfull discharge to small-floods potentially has resulted in a net loss of pool-maintenance events, and partly could explain the reduction in large pools observed by Minear (1994). However, Minear’s study spanned the period 1937–1991 and did not include the 1996 flood; hence, the reduction in pool frequency possibly may have been less dramatic had the study included this small flood.

In the lower McKenzie River basin, the frequency of small floods decreased by 76 percent, so that now these pool-scouring events occur approximately once every 9 years, whereas they occurred every 2–3 years prior to 1962 (table 15). Large floods enhance pool complexity through pool scour and the creation of channel-spanning blockages of wood and sediment, which can lead to the development of new pools; however, these infrequent, regional flood events (most recently in January 1943, December 1964, and February 1996) occur on a decadal scale.

Offchannel Habitat

Possible changes in the flow regime likely have the greatest effect on offchannel habitat in the lower McKenzie River basin, which historically supported numerous side channels, alcoves, and other secondary channel features. Although determining the exact role of different magnitude streamflows on the creation and maintenance of offchannel habitats requires additional analysis using field observations, historical analysis, and numerical modeling, generalizations can be made based on previous studies. High streamflow pulses (discharge up to bankfull) can enhance offchannel habitat by inundating low-lying features, scouring fine sediments, and disrupting young vegetation. However, larger magnitude streamflows that exceed bankfull discharge (small and large floods) are also needed to carve new features and maintain existing offchannel habitats (table 21; fig. 30). Modeling results and field observations from the nearby upper Willamette River demonstrate that small floods, such as that in February 1996, are able to inundate the floodplain, trigger bank erosion, and carve new channels within erodible, low-lying areas of the active channel, but do not have sufficient power to trigger large-scale avulsions or carve side channels through high floodplain surfaces (Wallick and others, 2007). Dam building and the resultant reductions of peak flows has caused historically frequent small flood events to become relatively rare, with recurrence intervals of almost 50 years for the middle McKenzie River basin and 9 years for the lower McKenzie River basin (table 15; fig. 31). Bankfull streamflows, which are commonly referred to as the “channel forming discharge,” historically occurred on a biannual basis. They have occurred only five times in the period 1962–2008 (fig. 31). Because small floods, large floods, and bankfull events each exert a unique influence on the creation and maintenance of offchannel habitat, reduced frequency of these events has contributed to a reduction in the complexity and availability of side channels, alcoves, and other secondary channel features.

The ability of environmental streamflows to restore offchannel habitat depends on more than flood magnitude; other factors including bank stabilization, geological controls, and incision may limit the ability of environmental flows to carve and maintain secondary channel features. Many areas along the lower McKenzie River that historically displayed abundant side channels were flanked by erodible Holocene alluvium, but are presently stabilized with revetments, which reduce lateral migration and chute formation, hence contributing to simplification of the channel network and minimizing opportunities for the creation of secondary channel features (fig. 27). A logical follow-up study might entail detailed analysis of aerial photographs taken before and after the February 7, 1996, flood, in order to determine the role of streamflow, revetments, and other influences on secondary channel formation.

Depositional Features

The abundance of gravel bars, and in particular the “loose,” recently deposited, unvegetated bars depends on the supply of gravels from upstream sources, the sediment transport capacity of the channel, and the frequency of flow events that can transport gravel (table 21; fig. 32). Gravel bars along the upper and middle McKenzie River basin reaches are probably less sensitive to streamflow than those in lower-basin alluvial reaches because the steep upper reaches support high transport capacities, and there are few depositional settings where gravel bars have historically formed. As a result, the channel bed along much of the upper and middle McKenzie River basin most likely is armored and unlikely to be mobilized except during extreme events. Additionally, the active channel is relatively confined within a floodplain composed of coarse Pleistocene gravels that are too large to be transported by most modern flow events (Stillwater Sciences, 2006b); hence, bank erosion along these reaches is unlikely to recruit an ample supply of transportable gravels.

The large, mobile, sparsely vegetated gravel bars that are located predominantly in the lower McKenzie River basin most likely respond dynamically to discharge regime through variation in size, frequency, and character and form the basis of streamflow–geomorphology relations shown in table 21 and figure 32. Moreover, the channel along the lower McKenzie basin is situated within a wide floodplain composed of sediments deposited during the Holocene. Thus, bank erosion along unrevetted parts of the lower McKenzie River may provide a source of gravels that could be redeposited in bars and other depositional features.

Sediment supply is a key factor that not only exerts a first order control on bar frequency, but also remains unquantified for much of the McKenzie River basin, despite concerns that trapping by the dams has substantially reduced bed-material availability. In the absence of a formal sediment budget for the McKenzie River basin, the flood control dams on the historically flood and sediment-rich tributaries and sediment-rich tributaries of Blue River and South Fork McKenzie River are assumed to have led to substantial reduction in coarse bed material entering the middle and lower reaches of the mainstem McKenzie River. Additionally, streamflow reduction has contributed to vegetation encroachment and stabilization of bar features, which has reduced the area of bare, easily eroded gravel bars, contributing to the overall reduction in sediment supply (fig. 28). Therefore, the ability of different magnitude streamflow events to recruit bed material stored in bank, bar, and floodplain deposits offers a potential opportunity to partially offset reductions in sediment supply caused by the dams.

High streamflow pulses are capable of reducing vegetation on low bar surfaces, remobilizing unvegetated bars, and transporting spawning size gravels. However, large magnitude flood flows are needed to rework bars that have widespread vegetation encroachment, such as those shown in the 2005 aerial photographs (fig. 28). These formerly active bars appear to be unaffected by high streamflow pulses during the post-dam period (fig. 31). Small floods likely would trigger greater remobilization of existing bars by reducing or eliminating vegetation on many bar surfaces. Although small floods, such as the February 1996 event, have occurred infrequently since dam construction, (table 15; fig. 31) these events are likely to promote bank erosion, particularly along unrevetted areas of the lower McKenzie River basin where gravels and wood released from bank erosion could enhance bar formation. Large floods (such as the one in December 1964) can rework and remobilize even densely vegetated bars and potentially trigger large-scale channel changes such as avulsions and rapid meander migration. This in turn can recruit large wood and sediment, which contributes to the size, frequency, and diversity of bar features.

Flow and Exemplar Species

The nine exemplar aquatic and terrestrial species in the McKenize River basin selected for the report include five native fish species (spring Chinook salmon, bull trout, Pacific and brook lamprey, Oregon chub), red-legged frog, western pond turtle, black cottonwood, and white alder. Gregory and others (2007a) summarized the life history and streamflow relationships of the species (with the exception of bull trout) for the Coast Fork and Middle Fork Willamette Rivers. Excerpts from that document are presented below in italics, in total or with slight modification. These exemplar species predominantly are found in a combination of main- and offchannel habitats, and offchannel or riparian/floodplain habitats singularly.

Aquatic Main Channel

In their Coast and Middle Fork Willamette analysis, Gregory and others (2007a) summarized key life history and environmental flow requirements for main- and offchannel fish species, which are also applicable to the McKenzie River basin:

• Fall flood pulses are important for passage of adult salmon.
• Magnitude, timing, and duration of spring streamflows are important cues for upstream migration of adult salmon and downstream migration of salmon smolts.
• Bankfull and overbank streamflows are important for Oregon chub, which inhabit backwaters and isolated floodplain habitats.
• For salmon and steelhead, adequate “incubation streamflows” must be maintained over redds created during spawning.
• Influences of streamflow modifications and thermal regime due to dam operation differ between spring- and fall-spawning native fish species.
• Spring-spawning species include coastal cutthroat trout, lamprey, and all native cyprinids (minnows), suckers, and sculpins.
• Fall-spawning species include spring Chinook salmon and bull trout.
• Non-native species (such as bass, catfish, mosquitofish, carp) typically are spring and early summer spawners.
• Flow modification in any season potentially affects the growth and survival of juvenile fish.
• Flow modifications in spring influence both adults and fry of spring spawning species.
• Flow modifications in autumn influence both adults and fry of fall spawning species.
• In the mainstem Willamette River and major tributaries, native fish may be migrating during any month of the year.
• Dam operations have been linked directly and indirectly to declines in spring Chinook salmon.
• Channel and floodplain simplification have been identified as major factors leading to the declines in Oregon chub, Pacific lamprey, and possibly coastal cutthroat trout.
• Creation of warmer, more lacustrine habitats in the Willamette River favor invasion by bass, carp, catfish, and other nonnative species.

Two recent reviews have compiled extensive information available related to the upper Willamette River spring Chinook salmon (Oncorhynchus tshawytscha) and the contribution of dam operations (U.S. Army Corps of Engineers, 2000) and other anthropogenic factors to its decline (Northwest Power and Conservation Council, 2004). In addition, the NMFS Technical Recovery Team has recently provided an in depth analysis of historical population structures and distributions for spring Chinook salmon (Myers and others, 2006).

From Gregory and others (2007a):

Spring Chinook enter the Columbia River and lower Willamette River in February through April. They move over Willamette Falls by fish ladder in April through June after river temperatures exceed 10°C. Spawning in the upper Willamette River occurs from September through October [fig. 33, in this report]. Some juvenile salmon migrate downstream as fry or fingerlings, but they are not physiologically capable of tolerating the higher salinities of the estuary until late summer of their first year. ...Most spring Chinook from the upper Willamette River enter the ocean as yearlings and return to the Willamette River as 4- to 6-year-old adult fish.

Flow modifications have several potential effects on upper Willamette River spring Chinook. Like most salmonids, Chinook salmon are particularly sensitive to the intertwined parameters of discharge, stream velocity, and stream temperature. The factors are major cues for salmonid life histories, and alteration of the timing of streamflow or stream temperature can alter the growth and survival of all riverine life history stages. Unnaturally cooler water in the spring/summer (along with lower streamflows in spring) can potentially reduce the growth of fish rearing in the river and affect the timing of smolt outmigrations and adult upstream returns. Likewise, warmer temperatures in the fall can affect spawning adults or cause earlier egg hatching and fry emergence from the gravel nests through accelerated development, thereby exposing them to high winter streamflows [table 18 and fig. 33, in this report].

As discussed earlier, releases from the Cougar and Blue River Dams for many years have been unnaturally cooler and warmer in the spring/summer and fall, respectively (figs. 21 and 22). Although the temperature of streamflow releases from Cougar Dam now mimic a more natural pattern, similar infrastructure modifications to Blue River Dam have not been made. However, the downstream thermal effect of Blue River is limited since the streamflow of Blue River is only approximately 10 percent of the streamflow at the McKenzie River near Vida (14162500) streamflow-gaging station.

From Gregory and others (2007a):

An increasing body of evidence from other large river systems suggests floodplain habitats are important low stream velocity refuges and nursery areas for young Chinook. Loss of these areas combined with bank simplification through revetments resulted in deleterious consequences for other salmonid populations. Recent work in the upper Willamette River mainstem has found winter use of floodplains, alcoves, and ephemeral streams by numerous native species, including spring Chinook. During winter streamflows, these areas apparently provide refuge both from high streamflows and introduced predators, as well as food from terrestrial and aquatic sources.

Dam and reservoir operations have also been implicated in fungal and parasitic infections. Smolts that experience high streamflows during outmigration tend to have lower overall rates of infection. Higher streamflows may decrease abundance of intermediate hosts for infections such as whirling disease (Myxobolus cerebralis) and Ceratomyxa shasta.

Bull trout are a nonanadromous salmonid that are coldwater species primarily found in large mountainous upper river basins and their tributaries (Taylor, 2003). They have largely been extirpated from all Western Cascade river basins except for the McKenzie River. Although they can seasonally migrate long distances and probably regularly move down to the mouth of the McKenzie River, they primarily occupy habitats in the tributaries and mainstem upstream of Belknap Hot Springs (RM 118 of the mainstem). Spawning primarily occurs in two subbasins: upper basin tributaries just below Trail Bridge Dam (Anderson and Olallie Creeks) and the South Fork McKenzie River above Cougar Reservoir (Roaring River) (Taylor, 2003). Bull trout have a similar life history pattern to spring Chinook salmon and other salmonids; spawning occurs from September to October in well-sorted gravels/cobbles located in the tail end of pools and riffles (fig. 33). Snorkeling surveys described by Taylor (2003) showed that juveniles primarily use low-velocity areas in marginal pools (backwater and lateral scour pools) with nearby cover (primarily boulders, wood, and other substrates). Juvenile rearing is generally confined to the reach between Trail Bridge Dam and Belknap Springs (Taylor, 2003).

From Gregory and others (2007a):

In contrast to the plethora of data available for spring Chinook, comparatively little is known about the life cycle and flow needs of another native anadromous species, the Pacific lamprey (Lampetra tridentata). Pacific lamprey are large, parasitic only in the marine adult stage, and historically were probably abundant in the Willamette River. The same changes to discharge, temperature, and sediment parameters that affect salmon likely also have led to the observed declines in Pacific lamprey. Adults return in late spring and spend the summer and autumn in the river before spawning as early as February (at Willamette Falls) or as late as July [fig. 34, in this report]; some individual adults may be repeat spawners (like steelhead). Pacific lamprey require small gravels for their nests, but fine silts and clays for larval rearing, which can last up to 7 years. Young lamprey (which are filter-feeders and not parasitic) are particularly susceptible to rapid streamflow fluctuations, and can be stranded if discharges drop rapidly. Water temperatures greater than 22°C cause mortalities of eggs and larvae; however, additional temperature and streamflow requirements are largely unknown. Larval outmigration appears to be triggered by a combination of discharge and temperature, and usually occurs in the spring. The much smaller western brook lamprey (Lampetra richardsonii) has a markedly different life cycle from the Pacific lamprey, but has similar streamflow, temperature, and sediment requirements. Unlike the Pacific lamprey, brook lamprey are neither anadromous nor parasitic. Brook lamprey spawn in late spring as water temperatures rise to 10°C; the eggs drift at night into silty backwater areas, where they hatch and metamorphose up to 6 months later. The filter-feeding ammocoete larvae spend the next 5 years in these areas before metamorphosing to adults, spawning, and dying. Both species require complex low-velocity areas for rearing, and loss of the low-gradient floodplain habitats (described previously) has been cited as a major cause for the observed declines in abundance of both species, particularly Pacific lamprey.

Spring Chinook, bull trout, and brook and Pacific lamprey require a diversity of instream and offchannel habitats. In addition, they are all cold-water species and spawn in similar flowing water habitats (figs. 33 and 34). As summarized in tables 21 and 22, these four species use or require habitat features that are created and maintained primarily by the interaction of small to large floods with local landscape features. Habitat features of particular importance common to these species include large deep mainstem pools, secondary channels, and well-sorted spawning gravels with interstitial streamflow. All four species also are sensitive to the seasonal timing of streamflows and water temperatures and thus are affected by alterations to streamflow and water temperature associated with dam operations.

Aquatic Offchannel

Excerpted from Gregory and others (2007a):

The Oregon chub, Oregonichthys crameri, is endemic to the lowlands of the Willamette River basin and was once widely distributed, but currently is found only in a few isolated locations along the Willamette River and its larger tributaries. Oregon chub inhabit backwaters and isolated floodplain habitats, and were probably once more common inhabitants of these slackwater areas. The loss of floodplain habitats and connectivity to larger river systems is one of the main contributing factors to the decline of the Oregon chub and correlates with the construction of revetments and dams. Exact streamflow and temperature requirements for the chub are largely unknown, although chub in the Middle Fork Willamette apparently require a water temperature of at least 15°C to spawn. Chub are frequently found in the same locales as red-legged frogs, suggesting these two species may respond to similar temperature and discharge regimes.Western pond turtles (Actinemys marmorata) are not limited solely to ponds, but also are found in backwaters, sloughs, marshes, and low-velocity regions of large rivers. Wooded riparian patches near open areas appear to be a predictor for adult turtles: most hibernate in forested floodplains and uplands, and the downed wood provides important basking sites. In addition to requirements for comparatively low velocity habitats, sunny, open areas with little vegetation for nesting habitat are also critical. Nests are constructed during early summer; the young hatch about 3 months later, and remain in the nest until the following spring [fig. 35, in this report].Red-legged frog (Rana aurora) breeding sites are usually found in relatively heavily vegetated locations with significant areas flooded in winter and spring. Breeding sites in the Willamette Valley can be associated with upland ponds as well as floodplain forest wetlands. These breeding sites expand with the onset of winter rains and overbank flood streamflows, and may be dry by midsummer. Red-legged frogs breed and lay their eggs in these shallow ponds during January and February [fig. 35, in this report]; the eggs hatch within 1 to 2 months. Tadpoles spend approximately 3 months before metamorphosing to adults. Red-legged frogs occasionally breed in side channels and sloughs associated with large rivers, and generally lay eggs in areas of little or no current. As with other amphibian species, red-legged frogs may be indicators of a number of environmental insults due in part to their use of different habitats over their life history. Egg masses may be stranded by fluctuating water levels. Loss and alteration of wetlands associated with agriculture and urban areas is likely one of the most critical challenges for red-legged frogs in the Willamette Valley.

Oregon chub, western pond turtle, and red-legged frogs are native species that require diverse offchannel habitats, including mature side channels, sloughs, oxbow lakes, ponds, and wetlands. These geomorphic features are primarily created and maintained by the interaction of small to large floods with the local landscape features as described above and summarized in table 22.

Floodplain Habitat and Riparian Vegetation

From Gregory and others (2007a):

Black cottonwood, Populus trichocarpa, [fig. 36 and table 22, in this report] and riparian willows are considered pioneer species that require bare, moist mineral soils for germination. These surfaces can range from bare gravel bars generated by annual floods or large overbank streamflows that deposit bare soils on the floodplain. The seeds are viable for only 1 to 2 weeks under optimum conditions; once germinated, the rate of streamflow recession is critical. The roots lengthen and follow the decline of the water table; too swift a recession rate, and the seedlings will not survive. The seedlings are highly resistant to inundation and sediment deposition, but are shade intolerant and will not germinate under existing stands. Both cottonwood and willow can also reproduce from broken branches and root fragments; large floods can therefore transport not only seed propagules, but vegetative ones as well.White alder, Alnus rhombifolia, [fig. 36 and table 22, in this report] is an early successional stage species that is found along perennial streams and rivers in lowland valleys. Seeds drop from the trees in late summer or early fall and are dispersed by both wind and water. White alder requires bare mineral soils for germination and can colonize many of the same habitats as cottonwood. Seedlings require continuously moist sites, and will suffer high mortalities under dry conditions. Like cottonwood, it is shade intolerant, and can regenerate from sprouts as well as seeds.

For these floodplain tree species, high winter and spring streamflows are essential for the dispersal of seed and vegetative propagules, generation and maintenance of bare soils for germination, and maintenance of proper water tables for seedling and mature tree growth and survival. These critical life history stages and their relation to pre- and post-dam mean monthly streamflows are presented in figure 36.

First posted February 8, 2010