Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 Missouri River and Floodplain Ecology I observe a great alteration in the Current course and appearance of this pt. of the Missouri. in places where there was Sand bars in the fall 1804 at this time the main current passes, and where the current then passed it is now a Sand bar. Sand bars which were then naked are now covered with willow several feet high. the enterance of some of the Rivers & creeks changed owing to the mud thrown into them, and a layor of mud over some of the bottoms of 8 inches thick. Captain William Clark, August 20, 1806 The Missouri River ecosystem experienced a marked ecological trans- formation during the twentieth century. At the beginning of the century, the Missouri River was notorious for large floods, for a sinuous and mean- dering river channel that moved freely across its floodplain, and for massive sediment transport. But by the end of the twentieth century, the Missouri River bore little resemblance to the previously wild, free-flowing river. This chapter describes these environmental changes in the Missouri River eco- system, contrasting the ecological state of the river and its floodplain before and after the construction of the series of large mainstem dams and reser- voirs. The physical and ecological units of todayâs Missouri River ecosys- tem are then described, followed with a review of the scientific research that forms the basis for our understanding of the ecosystemâs dynamics and the consequences of human actions. The chapter concludes with an overview of the systemâs ecological issues and a commentary from the committee. Seven huge dams were constructed along the Missouri River during the twentieth century, six of them pursuant to the 1944 PickâSloan legislation. The riverâs annual flooding regime was nearly eliminated in those areas under the damâs controlling influences. In another key twentieth century environmental change, the river was confined to a single, uniform channel. It was fixed in place by dikes and revetments downstream from just below 54
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 55 Gavins Point Dam, the most downstream dam, to the riverâs mouth at St. Louis. Ecological changes that accompanied changes in hydrology proceeded more slowly but were of a similar magnitude. Large floodplain areas along the upper Missouri were inundated by the reservoirs. Large areas of native vegetation communities in downstream floodplains were converted into farmland. Many native fish and avian species experienced substantial re- ductions, while nonnative speciesâespecially fishesâthrived in some areas. Scientists and citizens today understand more fully the consequences of similar changes that have occurred on many of the worldâs large rivers and the possibilities for reversing them. In the second half of the twentieth century, the field of large river ecology emerged to provide a scientific basis for river restorations, strategies, and initiatives. This research and the scientific understanding of river ecosystems builds upon a longer history of research on hydrology, geomorphologic processes (the shaping of river channels by water and sediment), vegetation dynamics, and river mechan- ics. The importance of extreme climatic events in the ecological structure and functioning of large river-floodplain ecosystems features prominently in contemporary river science theories (Bioscience vol. 45, # 3, 1995). Recent studies of large river systems emphasize the ecological values of hydrologic connections between a riverâs main channel, backwaters, and floodplain (Gore and Shields, 1995; USGS, 1999; Ward and Stanford, 1995). Aquatic ecosystem restoration practices build upon and comple- ment these theories. For example, current science-based river management paradigms and practices seek to take advantage of the tendency of these large river systems, through natural processes like floods and the transport of sediment, to make and sustain these connections (Bayley, 1995; USGS, 1999). Restoring some ecological functions of a large river system also provides the benefits of maintaining species biodiversity in the river and on its floodplain, as well as restoring habitat for threatened and endangered species. This chapter introduces and explains concepts that underpin the theories and practices of contemporary large river science and restoration. This knowledge helps explain the ecological consequences of human- induced changes in the Missouri River ecosystem and provides the scientific basis for considering public policy decisions to improve ecological condi- tions. THE PRE-REGULATION MISSOURI RIVER Physical Processes: Hydrology and Geomorphology During the period of westward expansion in the United States, the Missouri River was a large floodplain river. It periodically overflowed its
56 THE MISSOURI RIVER ECOSYSTEM banks and water spread across its floodplain, hydrologically connecting the channel, to its floodplain and backwaters. The Missouri River created new channels as its main channel moved laterally across its floodplain. These types of geomorphic changes in the channel were a notable feature of the pre-regulation Missouri, especially below present-day Yankton, South Dakota downstream to the confluence with the Platte River. In this stretch of the river, the distance between the Missouri River bluffs ranges from five to eighteen miles (Schneiders, 1999). Variations in the river channelâs location, form, and volume of sedi- ment transported were driven by changes in river flows. In turn, the riverâs hydrology was greatly influenced by weather extremes. The annual pattern of flows on the pre-regulation Missouri included spring and summer rises that generally occurred in April and June, respectively. The April rise was caused by local snowmelt on the plains and by local rainfall; the June rise was caused by melting snowpack in the Rocky Mountains and rainfall at lower elevations. The spring rise tended to be brief, lasting about one to two weeks, and was relatively localized. The summer rise lasted longer and inundated larger portions of the floodplain. Prior to regulation, the Missouri River was known as the âBig Muddy,â as it carried large amounts of sediment. Erosion tended to be most severe as flood waters were rising, with substantial deposition of sediment occurring as flood waters receded. The Missouri River existed in a dynamic equilib- rium with its floodplain, frequently redistributing sediment between its channel and floodplain, as described below. As Missouri River flows increased in the spring and summer (referred to by hydrologists as the ârising limbâ of a hydrograph), the river would erode sediment from its bed and its banks. The river bed would undergo rapid physical changes during this period; it would be degraded (lowered) because of erosion, the channel would migrate laterally, backwaters and the main river channel would be connected by overbank flows, and shoreline and riparian vegetation, including trees, would be scoured and washed into the channel. The rising water also would replenish the groundwater table, an important process for maintaining floodplain vegetation. As flows re- ceded (the hydrographâs receding limb) and water volume and velocity decreased, the degraded channel would refill with deposited sediment, braided channels and meanders would become isolated from the main chan- nel, and fresh substrates would be deposited for occupation by plants and animals. Prior to channelization and flow regulation, the lower Missouri River was braided to highly sinuous, a form naturally found in rivers with broad floodplains and heavy sediment loads. Hiram Chittenden described the
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 57 riverâs pre-regulation sinuosity in his account of the adventures of nine- teenth century Missouri River steamboat captain Joseph La Barge: The river is like a great spiral stairway leading from the ocean to the mountains. A steamboat at Fort Benton is 2565 feetâtwo and one-times the height of the Eiffel Tower in Parisâabove the level of the sea; yet so gentle is the slope nearly all the way that, in placid weather, the water surface resembles that of a lake. This wonderful evening-up of the slope of the river by the extreme sinuosity of its course is a fact not only inter- esting as a natural phenomenon, but of the utmost importance in the behavior and use of the stream (Chittenden, 1962). The pre-regulation river was characterized by log jams, snags, whirl- pools, chutes, bars, cut-off channels, and secondary channels around bars. The main channel typically had a deep thalweg (the deepest part of the river) that contained the faster-moving flow and a shallower section(s) on one or both sides of the channel (Figure 3.1). The cross-sectional shape of the main channel often exhibited a highly nonuniform velocity distribution (Hesse, 1993). The main river channelâs width was variable, ranging from roughly 1,000 to 10,000 feet wide during normal flow periods to 25,000 to 35,000 feet wide during floods (Schneiders, 1999). Based on surveys com- pleted between 1878 and 1892, the maps in figures 3.2 through 3.4 portray the sand bars, meanders, and extensive floodplain vegetation that charac- terized the pre-regulation riverine ecosystem at Bismarck, North Dakota, Yankton, South Dakota, and St. Joseph, Missouri, respectively. River depth was greatest in spring and early summer and shallowest in December and January. However, there were scour holes downstream of FIGURE 3.1 Typical cross-section of the pre-regulation Missouri River. SOURCE: Rasmussen, 1999.
58 THE MISSOURI RIVER ECOSYSTEM FIGURE 3.2 Nineteenth Century Missouri River Floodplain topography at Bis- marck, North Dakota. SOURCE: Hesse, 1987. log jams and other obstructions, at bends, or where tributaries entered. Depths in these holes varied from a few inches to more than thirty feet (Schneiders, 1999). Mid-channel and point bars were found along the entire length of the Missouri River. The bars shifted frequently in response to changing flows with the larger bars scoured at higher flows. Ecological Processes The rich biodiversity (see Box 3.1) of the pre-regulation Missouri River ecosystem was sustained through a regime of natural disturbances that included periodic floods and attendant sediment erosion and deposition. These disturbances, in turn, supported a variety of ecological benefits, in- cluding commercial and recreational fishing, timber, biomass fuels, wild game, trapping and fur production, clean water, medicines, soil replenish- ment processes, and natural recharge of groundwater. The ecological pro-
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 59 FIGURE 3.3 Nineteenth century Missouri River floodplain topography at Yankton, South Dakota. SOURCE: Hesse, 1987. cesses sustained by periodic flooding in river-floodplain ecosystems are encompassed by the contemporary concept known as âthe flood pulse.â The Flood Pulse Concept The concept of the âflood pulseâ summarizes the effects on biota of the connections between the river channel and floodplain (Junk et al., 1989; Bayley, 1995). The flood pulse describes the predictable rising and falling of water in a natural river-floodplain ecosystem as the principal agent controlling the adaptations of most of the biota. Central to the flood pulse concept is the notion that floods, rather than representing a disturbance to the ecosystem, are part of the natural hydrologic regime, and that the prevention of floods actually represents an ecological disturbance (Bayley, 1995). The flood pulse is essential to the health of river-floodplain ecosys- tems for the following reasons:
60 THE MISSOURI RIVER ECOSYSTEM FIGURE 3.4 Nineteenth century Missouri River floodplain topography at St. Jo- seph, Missouri. SOURCE: Hesse, 1987. 1. Floods add dissolved and particulate organic matter and mineral nutrients to aquatic and terrestrial ecosystems. The river channel and its floodplain both depend on erosion and deposition associated with the channelâs lateral migration. Inundation deposits silts and nutrients that replenish floodplain pools and backwaters. The flooding of terrestrial mineral and organic matter releases nutrients to the water. 2. Many plants rely upon inundation for rapid growth and reproduc- tion. Species such cottonwood and willow are highly dependent upon periodic floods. 3. Many animals (invertebrates, fish, birds, mammals) are adapted to the flood cycle and depend upon the high plant and microbial activity associated with it. Floods provide reproductive cues for many fish species in river-floodplain ecosystems. Furthermore, floods make inundated flood- plain vegetation available as a food source for fish and invertebrates.
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 61 Box 3.1 Biodiversity Biodiversity, the shortened term for biological diversity, is commonly used in the lexicon of twenty-first century environmental sciences. The Oxford Dictionary of Biology describes biodiversity as âthe existence of a wide variety of species (species diversity) or other taxa of plants, animals, and microorganisms in a natu- ral community or habitat, or of communities within a particular environment (eco- logical diversity), or of genetic variation within a species (genetic diversity). The maintenance of a high level of biodiversity is important for the stability of ecosys- tems.â Biodiversity is often associated with the number and kinds of species at a locality. Ecologists describe the number of different species as species richness, while species diversity accounts for both the number of species as well as their relative abundance in a community. It is important to remember that biodiversity includes not only these components, but that it also refers to genetic diversity and may include all manner of organisms ranging from paddlefish to bacteria. Ecosys- tems that support high species diversity often also demonstrate high biodiversity, such that management for species diversity and species richness in natural eco- systems tends to foster the other, sometimes more obscure, elements of biodiver- sity. A key change that occurs in regulated river systems with respect to biodiver- sity is the proliferation of nonnative species. Described by Stanford et al. (1996) as the most pervasive result of habitat alteration in large regulated rivers, this shift usually occurs in communities ranging from fish and invertebrates to riparian and floodplain vegetation. A committee of the National Research Council (2000a) rec- ommended that an indicator that measures native species diversity is an important indicator of human impacts on the environment. Appreciation of the flood pulseâs ecological significance has grown with recognition of the importance of flooding at various scales and magnitudes (Anderson et al., 1996; Wetzel, 2001). Key aspects of floods are frequency, duration, magnitude, and timing. The combination of these variables con- trols plant and animal life associations along rivers and influences all as- pects of the riverine food web (Poff and Ward, 1989; Richter et al., 1996; Walker et al., 1995). The timing of overflow is important to native biota adapted to life in rivers and adjoining riparian ecosystems. Periodic high flows and low flows help maintain the health of large river-floodplain ecosystems by act- ing as âresetâ mechanisms that reinitiate early successional vegetation and serve to limit certain faunal associations that can outcompete species nor- mally restricted to life within a channel (Cummins et al., 1984). Fish spawning, insect emergence, and seed dispersal are commonly triggered by
62 THE MISSOURI RIVER ECOSYSTEM rising waters that, once receding from the floodplain, provide food sources or seed beds for many riverine species. Because these processes sustained the riverâs biological production and diversity, the pre-regulation Missouri River exhibited a rich heterogeneity of habitat. A typical cross-section of the pre-regulation Missouri River contained a deep channel, multiple side channels, oxbow lakes, islands, sand bars and dunes, and backwater habitats interspersed by areas of higher land. These channels and backwater areas provided slower-moving waters critical for the reproduction, shelter, and feeding of fish species. Higher lands contained rich forests, prairie grasses, and thick underbrush that contained a myriad of plant species. Lewis and Clark described this rich pre-regulation biodiversity, noting that the Missouri ânourishes the willow- islands, the scattered cottonwood, elm, sycamore, linden, and ash, and the groves are interspersed with hickory, walnut, coffee-nut, and oakâ (Lewis and Clark expedition, Coues, editor; cited in Schneiders, 1999, p. 35). THE POST-REGULATION MISSOURI RIVER Physical Processes: Hydrology and Geomorphology Flow regulation and channelization substantially changed the Missouri Riverâs historic hydrologic and geomorphic regimes. The primary change was that the extreme high and extreme low flows were lost from the hydrograph downstream of each mainstem dam. This dampening effect below Gavins Point Dam extends downstream to near Nebraska City, Ne- braska (Hesse, 1994), where tributary influences partially restore pre-regu- lation flows to the river. Below Fort Peck Dam, for example, the median high flow was cut in half following the damâs closure (Shields et al., 2000). Not only have high flows been markedly reduced in many areas, low flows have increased considerably. The result of these changes is an annual hydrograph that exhibits far less variability. Figures 3.5 and 3.6 show hydrologic changes in the post-regulation river at Sioux City, Iowa (just downstream of Gavins Point Dam) and at Hermann, Missouri. These figures illustrate hydrologic changes associated with the construction of the mainstem dams. The figures also show that these changes that occurred are not uniform along the Missouri River and that some downstream sections in the state of Missouri have experienced less hydrologic change with regu- lation. But the variability that characterized pre-regulation Missouri River hydrology has greatly diminished along most of the riverâespecially in those reaches directly below the damsâand the spring and summer rises no longer occur in many stretches. The Missouri River also has experienced large changes to its channel structure and dynamics. The channelized portion of the Missouri River
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 63 FIGURE 3.5 Missouri River discharge at Sioux City, Iowa. SOURCE: USACE, undated. FIGURE 3.6 Missouri River discharge at Hermann, Missouri. SOURCE: USACE, undated.
64 THE MISSOURI RIVER ECOSYSTEM begins near Nebraskaâs Ponca State Park (just upstream of Sioux City, Iowa). From Ponca State Park downstream to the Big Sioux River, the river channel is stabilized. The navigation channel then begins at the mouth of the Big Sioux River at Sioux City. The navigation channel, which extends to St. Louis, ranges in width from 600 feet at the Big Sioux confluence to roughly 1,100 feet at St. Louis. Nearly all channelization activities have been conducted as part of the federal Missouri River Bank Stabilization and Navigation Project (passed as part of the 1945 Rivers and Harbors Act), which is executed by the Corps of Engineers. Channelization has been accomplished through a combination of engineering structures, including hardpoints, a variety of revetments, dikes, and sills (Slizeski et al., 1982; see Figure 3.7). FIGURE 3.7 Commonly-used dikes and revetments along the Missouri River nav- igation channel. SOURCE: Slizeski et al., 1982.
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 65 The cross-sectional shape of the Missouriâs channelized portion (735 miles or about one-third of the riverâs length) is approximately trapezoidal. Prior to channelization, the riverâs flow had been swift only in its thalweg (a line connecting the deepest points of the river channel), as the river con- tained sloughs, sandbars, and side channels. But today the river runs swiftly throughout the entire channelized, uniform cross-section. The re- duction in width, along with a decrease in flow resistance because of the uniform cross-section and the clearing of snags and sand bars, has caused an increase in flow velocity, which today measures roughly three miles per hour at usual levels of river discharge (Schneiders, 1999). Regulation of the Missouri Riverâs flows also changed sediment trans- port dynamics. Prior to regulation, the amount of sediment transported past Omaha, Nebraska ranged from 228,570,000 metric tons in 1944 to 39,909,297 metric tons in 1931. From 1940-1952 (the period from the closure of Fort Peck Dam until the closure of Gavins Point Dam), the average annual sediment load transported past Omaha was 148,930,000 metric tons. After 1954, the average sediment load was reduced to 29,487,600 metric tons (Slizeski et al., 1982). Several inter-related processes have caused post-regulation changes in the riverâs rate of lateral migration. The channel downstream of the dams has degraded (deepened). Degradation occurs as a result of the water released from the dams and increased currents caused by structures that have been installed to force water into a single channel. These features downstream of dams mimic the natural tendency of flow to mobilize and transport sediment. However, once mobilized and transported downstream, there is no longer an upstream source of sediment to replace sediment removed by these flows. Replacement sediment that would have main- tained the dynamic equilibrium of the channel is deposited in upstream reservoirs. But there is an abundance of sediment in floodplain areas lateral to channels below the dams that could be used to meet sediment needs in the channel. With the exception of Oahe and Big Bend dams, where downstream reservoirs extend upstream nearly to the tailwaters of these dams, channel degradation has occurred below the dams. In the channelized section from Ponca State Park downstream to St. Louis, channel degradation occurs downstream from Sioux City, Iowa to just above the Missouriâs confluence with the sediment-laden Platte River. Farther downstream, especially near the confluence of the Missouri and Mississippi rivers, the channel bed is gradually aggrading. From the confluence upstream to approximately Mis- souri River mile 12, Missouri River navigation channel depths are fre- quently impacted by Mississippi River flows, causing a backwater effect up the Missouri River that results in reduced velocities and temporary deposi- tion (Mellema, 1986).
66 THE MISSOURI RIVER ECOSYSTEM Channel degradation below the Missouri Riverâs mainstem dams is well documented (Holly and Karim, 1986; Mellema and Wei, 1986; Osterkamp and Hedman, 1982; Sayre and Kennedy, 1978). For example, Figure 3.8 shows that degradation on the order of ten feet or more extends many miles downstream of Garrison Dam. Similarly, below Gavins Point Dam, the river in 1980 had degraded by 8.5 feet. Degradation of the river channel disconnects the river channel from its floodplain. Channel degradation not only makes it more difficult for the river to overflow its banks, but it also affects the floodplain water table. Most importantly, the lack of flooding removes a source of periodic re- charge water for infiltration to the groundwater table. In addition, because the water table (an alluvial aquifer) is hydrologically connected to the river channel itself, there is a consequent lowering of this aquifer in association with the lowering (incision) of the river channel. This lowering of the water table effectively drains water from oxbow lakes and wetlands. Moreover, 1680 1670 1946 1660 Elevation (feet) 1650 1999 1640 1630 1390 1380 1370 1360 1350 1340 1330 River Mile FIGURE 3.8 Average bed elevations below Garrison Dam (located at River Mile 1390) from 1946-1999 (river miles measure distance upstream from the Missouri Riverâs mouth near St. Louis). SOURCE: USACE 1993; 1999.
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 67 in highly-regulated stretches of the river, reduced fluctuations in river stage have resulted in reduced fluctuations in the floodplain water table. These fluctuations are important to maintaining animal and plant species richness in the floodplain, as some species will benefit from a raised water table, while other species will benefit when the water table is lower. Channelization of the lower Missouri River and subsequent degrada- tion of the river channel also have affected tributary streams. In upstream areas, downward incision is occurring along many of the Missouriâs tribu- taries. This process occurs because the slope of the tributary channel bed increases in order to meet the (relatively) newly-lowered elevation of the Missouri River channel bed. Many tributaries continue to adjust to the new river bed elevation. As channel degradation continues to entrench the stream, there are fewer overbank flows than there were prior to degradation, thus reducing interaction between the flow in the channel and floodplain. Rates of chan- nel migration also have decreased. Lateral migration of river channels can occur in areas below dams; however, meandering rates have been markedly reduced downstream of the Missouri mainstem dams because of sharp reductions in peak flows and the armoring of streambanks. Johnson (1992) found that channel erosion and deposition rates (both indicators of river meandering rate) are only 25 percent and 1 percent of pre-dam values, respectively, downstream of Garrison Dam. Similarly, Shields et al. (2000) found that the mean rate of channel migration just downstream of Fort Peck Dam declined from 20 feet per year to 6 feet per year. Ecological Processes Many processes essential to maintaining ecological integrity have been altered in the post-regulation Missouri River. The spring and summer floods have been eliminated in many stretches of the river (although floods still occur in much of the riverâs channelized section, especially in down- stream sections in the State of Missouri). The isolation of the Missouri River from its floodplain caused by river regulation structures has in many stretches largely eliminated the flood pulse and its ecological functions and services. In these areas, the absence of overbank flooding removes a source of water for the growth of vegetation, as well as a medium for fishes to move into floodplain areas to spawn and feed. As a result of these changes, the production and the diversity of the ecosystem have both markedly declined. One of these impacts is a reduced ability for trees to regenerate. On the Missouri River and many of its tributaries, this has especially been the case for the cottonwood, largely as a result of the current low rate of river meandering (Johnson et al., 1976; Johnson, 1992). The habitat through a typical cross-section of the post-
68 THE MISSOURI RIVER ECOSYSTEM FIGURE 3.9 Typical cross-section of the post-regulation, channelized Missouri River. SOURCE: Rasmussen, 1999. regulation Missouri (in the non-submerged portions) has been greatly sim- plified (Figure 3.9). Side channels and backwater areas have been greatly reduced, thereby eliminating important habitat for many species of fishes, birds, and game. The water, sediment, and nutrients previously spread across the floodplain by overbank flows and the meandering river are now primarily restricted to the main channel or contained in the systemâs reser- voirs. These changes, combined with other human activities in floodplain areas, have produced an ecologically impoverished ecosystem. MISSOURI RIVER ECOSYSTEM PHYSICAL AND ECOLOGICAL UNITS Scientific investigations today are conducted in an ecosystem that changed greatly during the twentieth century and that today is fragmented into distinct physical and ecological units. The mainstem dams, along with other flood damage reduction and navigation enhancement projects, parti- tion the river into four sub-units that differ greatly in hydrology, sediment balance, and biota (Figure 3.10). The river can be classified into four sub- units: â¢ Free-flowing (upstream of the reservoir system); â¢ Remnant floodplains (between the reservoirs); â¢ Channelized reach (downstream of the dams and reservoirs; the lower one-third of the river); â¢ Reservoirs These categories vary markedly in the degree to which they mimic the pre-regulation Missouri River. The mainstem Missouri River is further
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 69 FIGURE 3.10 Missouri River basin and river reaches. SOURCE: USGS, undated. influenced by alterations on its tributaries. Dozens of tributaries enter the Missouri River along its course and most have experienced flow and sedi- ment alterations by dams, water diversions, channel modifications, and land use changes to their watershed (e.g., farming and wetland drainage). The transformation of the Missouri River from a free-flowing to regulated river makes the upper basin tributaries, which are generally less regulated in comparison to lower basin tributaries, important components of a compre- hensive, basinwide strategy for ecosystem restoration. Some examples are instructive. Missouri River Tributaries Yellowstone River The Yellowstone River flows 675 miles through Montana to its confluence with the Missouri River at the North Dakota border (Figure 3.10). The Yellowstone River is the longest free-flowing river remaining in the contiguous United States. There are no significant impoundments on the Yellowstoneâs mainstem, but nearly one-third of its drainage area has been dammed and six low-head dams on the main channel divert water for
70 THE MISSOURI RIVER ECOSYSTEM irrigation (Helfrich et al., 1999). While irrigation withdrawals and tribu- tary dams affect the riverâs hydrology and low-head dams restrict upstream movement of some native fishes (Helfrich et al., 1999), the river retains much of the ecological character it exhibited prior to European settlement (Jackson, 1994). At their confluence, the flow of the Yellowstone River is greater than that of the Missouri. Where the Yellowstone River, with its abundant silt load and naturally varying hydrology, meets the Missouri River near the Montana-North Da- kota border, near pre-regulation conditions exist. In fact, the Yellowstone River serves as a refuge for many of the Missouri Riverâs native, warm- water fish (Ryckman, 2000). For example, there are high levels of paddle- fish reproduction in the lower Yellowstone in years with above average streamflow. Additionally, native suckers and chubsâin decline throughout much of the river systemâare fairly abundant and reproduce in the confluence area (Ryckman, 2000). Moreover, the Missouri River down- stream of the confluence is a healthy riparian zone that includes ample cottonwood and willow generation maintained by floods and sediment contributed by the Yellowstone. Bad River The unregulated Bad River empties into the Missouri River at Fort Pierre in central South Dakota just upstream of Lake Sharpe (Figure 3.10). The Bad River is small and intermittent and therefore provides only limited ecological benefits to the Missouri River mainstem when compared to the flows of the Yellowstone River. Moreover, there is only a short distance between the confluence of the Bad and Missouri rivers and Lake Sharpe, leaving little of the sediment-deprived Missouri River downstream of Oahe Dam to benefit from the Bad Riverâs input of sediment-laden water. The proximity of the Bad-Missouri confluence to Lake Sharpe has caused much of the Bad Riverâs delivery of 3.25 million tons of sediment per year to remain near its mouth, reducing channel capacity and increasing flooding in and near Fort Pierre (Thelen and Noeske, 1996). Flow releases from Oahe Reservoir (a few miles upstream of the Bad-Missouri confluence) intended to transport sediment farther into Lake Sharpe cost $12.5 million annually from foregone power revenues (Thelen and Noeske, 1993). The Bad Riverâs high sediment transport rate results from a combination of highly erodable soils and failure to use best management practices on crop- land and rangeland in the watershed (Stukel and Madsen, 2000). Better farming and ranching practices could lessen, but not stop, the sedimenta- tion problem at the confluence and in Lake Sharpe. Thus, the condition of the Missouri River mainstem in the Pierre-Fort Pierre area depends largely on human activities in this tributary watershed.
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 71 Platte River The Platte River enters the Missouri River near Plattsmouth, Nebraska (Figure 3.10). While the Platteâs upper tributaries (South and North Platte rivers) are highly regulated and used for irrigation water, the relative lack of storage reservoirs on the Platte River itself allows considerable amounts of sediment to enter the Missouri River at the confluence. Grain sizes of this sediment range from coarse to fine sand. The sizeable increase in the Missouri Riverâs bedload increases the potential for in-channel bar forma- tion and alluviation on the floodplain during floods. The sediment and flow variability added to the Missouri River by the Platte River offer the potential for improving river ecology; but this potential is limited because of the Missouri Riverâs channelized and highly regulated state both above and below its confluence with the Platte. Kansas River In contrast to the Bad and Yellowstone rivers, the Kansas River is heavily regulated (Figure 3.10). Eighteen reservoirs, with a total flood- control capacity of 7.4 million acre feet, have been constructed on the Kansas River (Perry, 1993). These reservoirs are intended to reduce flood damages and to enhance navigation flows on the Missouri River. Although the Kansas River is large enough to affect conditions in the Missouri River below the Kansas Riverâs mouth, flow regulation and sediment trapping by its reservoirs reduce the potential of the Kansas River to improve ecological conditions in the Missouri River. Physical Units of the Regulated Missouri River Free-Flowing Reaches The only free-flowing reach of the Missouri River lies in Montana, upstream of the mainstem dams. This reach without storage reservoirs extends from the Missouri River source near Three Forks, Montana, down- stream to Canyon Ferry Reservoir, a distance of about 30 miles. However, the much longer reach from Canyon Ferry Dam to Fort Peck Lake is only mildly regulated because of the comparatively small storage capacity of Canyon Ferry Reservoir relative to total river flow and the long distance between Canyon Ferry Dam and the next downstream reservoir (Fort Peck). Contributions from small mountain streams and springs help retain some of the natural flow and temperature patterns in this reach as well. These moderately regulated reaches have retained their essential pre-regula- tion characteristics, including overbank flooding, adequate sediment supply
72 THE MISSOURI RIVER ECOSYSTEM to prevent channel degradation, scattered populations of cottonwood for- ests similar to those observed by Lewis and Clark, and productive native fisheries (Scott et al., 1997). Floodplain vegetation in these reaches is often impacted less by river regulation than by local land use practices, such as grazing (Auble and Scott, 1998). Remnant Floodplains Remnant floodplain sub-units occur between reservoirs (Figure 3.10). The length of these reaches varies considerably. In some cases, the headwa- ters of the mainstem reservoirs extend nearly to the tailwaters of the next upstream dam; there are few remnant floodplains from Lake Oahe down- stream to Fort Randall Dam. In other cases, reservoirs are separated by large stretches of river (e.g., section 3, from Fort Peck Dam downstream to Williston, North Dakota; Figure 3.10). These latter subunits have retained a natural appearance, with a sinuous channel and a wide floodplain often with oxbow lakes, sand dunes, and interspersed patches of natural forest vegetation and agricultural fields. The natural appearance, however, masks fundamentally altered hydrologic and sediment regimes. Nonetheless, many of these subunits are not physically static, and undergo natural degradation and sedimentation processes as altered by flows and releases from upstream dams and tributary inflows. Many of these segments are now incised, which has caused the loss of adjacent wetlands and secondary channels. Streamflow through these remnant floodplain reaches depends pri- marily on releases from upstream dams and secondarily on local tributary inputs. Pre- and post-regulation comparisons of streamflow through these reaches can be striking. For instance, at Bismarck, North Dakota, part of which lies on the floodplain between Garrison (upstream) and Oahe (down- stream) dams, the hydrographâs peak flow has been greatly reduced since the closure of Garrison Dam in 1953. Between 1928 and 1953, about two- thirds of the annual peak flows at Bismarck exceeded 2,500 cubic meters per second; since 1953, no peak has exceeded 2,500 cubic meters per second (Johnson, 1998). Reduced peak flows in the post-regulation period fail to inundate the floodplain, in sharp contrast to the Missouri Riverâs notorious floods before river regulation. Reily and Johnson (1982) showed that Garrison Dam has also changed seasonal flow patterns; peak flows now occur in winter instead of in spring, and minimum flows now occur mainly in spring and fall instead of in winter. The lack of overbank flooding in remnant reaches, except on the lowest terraces during extreme wet periods, has serious ecological consequences. Reiley and Johnson (1982) and Johnson (1992) reported decreased rates of both tree growth and tree population recruitment due to the absence of annual recharge of water and nutrients. Moreover, the reduced post-regu-
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 73 lation peaks in Missouri River discharge have been insufficient to cause lateral meandering of the channel that is needed if recruitment sites for pioneer forest communities dominated by cottonwood and willow are to be created. This diverse community type is in serious decline in much of the Great Plains due to river regulation and land management (grazing) prac- tices (Knopf et al., 1988). Downstream effects on remnant floodplains are less severe below smaller, upstream dams. For example, Ramey et al. (1993) found that Canyon Ferry Dam and smaller dams on tributaries have decreased the magnitude of higher flows (those greater than 1,400 cubic meters per sec- ond) by 14-23 percent at Fort Benton, Montana. Downstream of Fort Benton, regeneration of cottonwood forests in constrained reaches (nar- rowed by high valley walls) depends entirely on such high flows (Auble and Scott, 1998; Scott et al., 1997). In this case, dams have reduced flooding at the expense of cottonwood forest regeneration and growth. Water quality effects (most significantly cold water releases from middle levels or the bottom of reservoirs) are also most pronounced immediately below dams and diminish as one moves downstream. At the downstream end of remnant floodplains, streambed aggradation occurs where sediment carried by the river is dropped in the still water of the reservoir. Where aggradation is substantial and new vegetation curtails sediment redistribu- tion, streamflow may be obstructed (Johnson, 2002). This causes flooding from above (overbank flow) and from below (rising water tables). This phenomenon has caused property damage near Running Water, South Da- kota (upstream of Lewis and Clark Lake); Pierre, South Dakota (upstream of Lake Sharpe); and Bismarck, North Dakota (upstream of Oahe Reser- voir). Tributary streams within remnant floodplain reaches may ameliorate the effects of mainstem dams at specific locations (Johnson, 2002). Their influence depends on many factors, including the degree to which their flow has been regulated and their sediment trapped, their entry point on the mainstem (i.e., distance from mouth to nearest downstream reservoir), and their size (flow volume). Relatively natural tributaries contribute sediment and streamflow. The additional sediment and water contribute to higher peaks, turbidity, and greater flow variability, all of which are important to most native riverine organisms. Moreover, undammed tributaries often provide the shallow water and sandbar habitat for fish spawning and rear- ing destroyed by mainstem dams. Additionally, sediment input from tribu- taries can attenuate channel degradation below dams. The most beneficial ecological effects of the Missouri Riverâs tributary streams occur when relatively large, unregulated tributaries empty into the mainstem some distance upstream of the next reservoir. The Yellowstone River is the best example of this. There are fifteen to fifty miles of the
74 THE MISSOURI RIVER ECOSYSTEM mainstem Missouri River below the confluence, depending on the levels of Lake Sakakawea (Ryckman, 2000). The Yellowstone River adds flow and sufficient sediment to the relatively clear water released from Fort Peck Dam to cause natural cut-and-fill alluviation, riparian vegetation establish- ment, and successful reproduction of native fishes such as the pallid stur- geon and the paddlefish (Helfrich et al. 1999, Ryckman, 2000). In sum, effects of reservoirs on downstream remnant floodplains are slow and progressive. These changes downstream of dams may take de- cades to centuries to achieve their full impacts on remnant floodplain eco- systems. However, the sediment and flow variability that tributaries con- tribute to some remnant reaches of the Missouri River can ameliorate some of a damâs negative impacts. In turn, this also shortens the service period of downstream reservoirs. Channelized Reach Downstream of Gavins Point Dam, the Missouri River has been channelized (narrowed and deepened in a relatively fixed position) from Sioux City, Iowa to its mouth to permit navigation by boats and barges, and its banks were stabilized to enhance utilization of the bankline adjacent to the channel (sections 14-19 in Figure 3.10; Schmulbach et al., 1992). In addition, chutes and side channels have been blocked and diverted, convert- ing the once structurally-complex channels and instream islands into a single thread of deep, fast moving water. Levees have been constructed on both banks along much of the lower river to protect crops and settlements behind them; these levees constrain overbank flows to a narrow zone of the floodplain. This channelized stretch of the river was once highly dynamic: River surveys . . . showed that before 1930, which marked the beginning of major channel control works by the Corps of Engineers, the channel was two to three times its present width, and bars and islands existed in abundance. Floods usually brought great changes as new bars and islands were created, old ones disappeared, and the channel migrated rapidly to bring about extensive floodplain erosion and redeposition. Since 1930 and the construction of the channel control works, insular bars have ap- peared occasionally during low water, but their existence is only tempo- rary; and as a consequence of channel stabilization, significant shifts of the channel even during floods have all but been eliminated (Schmudde, 1963). The riverâs upper portion in South Dakota, Nebraska and Iowa has degraded because of erosive water releases from upstream dams, the trap- ping of sediments in mainstem reservoirs, and insufficient flows to accom- plish lateral channel readjustment. By contrast, the Missouriâs lower reaches (especially downstream of the Platte) have aggraded. In addition to degra-
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 75 dation or aggradation of the channel bed, interaction (material and energy exchange) between the river and the floodplain has been significantly re- duced or totally eliminated. Engineering works on the riverâs main channel have resulted in signifi- cant ecological changes in the channelized reaches. Construction of revet- ments has greatly narrowed and deepened the channel and has fixed its location. This has virtually eliminated shallow water habitat and greatly increased water depth and current velocity. Ecological impacts of these changes on native fish and on streamside vegetation have been strongly negative (Schmulbach et al., 1992). Levees restrict the river to only a small portion of its total floodplain, except during rare floods such as in the 1990s, when some levees were breached and water and sediment moved behind the levees into floodplains (Galat et al., 1998). Overall, the levee system has reduced interaction between the river channel and its floodplain, resulting in the inability of the river to sustain its historic levels of biodiversity. MISSOURI RIVER ECOSYSTEM SCIENCE The Missouri River ecosystem has been the subject of scientific investi- gations that date back to Meriwether Lewis and William Clark and their cataloguing of Missouri River plant and animal species during their epic journey of 1804-1806. The pace of scientific investigations of the ecosys- tem increased notably during the second half of the twentieth century. A recent bibliography of technical reports and scientific investigations lists well over two thousand entries prior to 1997 (Burke et al., 1997). At least several hundred scientific publications have been added between 1997 and the publication date of this report. A comprehensive review and analysis of that entire body of science was beyond this committeeâs means and scope. Furthermore, that research is unevenly distributed across topics; for example, there has been more scien- tific inquiry into select species, such as those on the federal endangered species list, than into other ecological topics such as carbon cycling or plant and animal interactions. In addressing its charge to identify the general state of that information, and to identify and prioritize key scientific ques- tions and information, the committee thus focused its reviews on the two ecosystem components that have received the bulk of scientific attention, fisheries and floodplain vegetation. Research on Fisheries Ecological impacts of large mainstem dams and other human activities in the Missouri River basin were slow to be discovered. The riverâs biota
76 THE MISSOURI RIVER ECOSYSTEM had been at least superficially inventoried between the mid-nineteenth and twentieth centuries (e.g., Aikman, 1929; Allen, 1875; Bailey and Allum, 1962; Bennett, 1931; Fisher, 1945; Gilmore, 1911; Jordan and Meek, 1885; Linsdale, 1928; Perisho and Visher, 1912; Reid and Gannon, 1927; Stevens, 1945). However, connections between key physical processes and key ecological processes remained virtually unstudied for the Missouri River, and most large rivers, until late in the twentieth century (Hesse et al., 1989; Stanford et al., 1996). Detecting change itself during this period was diffi- cult because the Missouri Riverâs baseline conditions were only partially known. Moreover, when changes were documented, the causes were un- clear because of the increasingly complex mix of human and natural factors affecting the river ecosystem. Among the better-documented ecological changes on the Missouri are the development of sport and recreational fisheries in the large mainstem reservoirs, especially in the three largest reservoirsâLake Sakakawea (Gar- rison Dam), Lake Oahe (Oahe Dam), and Fort Peck Lake (Fort Peck Dam). The clear water in the reservoirs provided an advantage to âsight feedingâ native species, such as the walleye, which was a species in relatively low abundance whose numbers increased dramatically with habitat changes caused by the reservoirs. Just as these environmental changes made condi- tions better for some species, other species that were better adapted to pre- regulation conditions, such as the sauger, experienced declines with the replacement of a free-flowing river by the system of reservoirs. The key introduced sport species on the Missouri River is the chinook salmon. Rainbow smelt and spottail shiner have also become established and are a major food source for the salmon and the walleye. Northern pike numbers increased dramatically with construction of the reservoirs. Once the domi- nant sport species immediately following construction of the reservoir sys- tem, the northern pike has declined in numbers and today represents a relatively insignificant portion of the sport catch. White and black crappies responded well to the filling reservoirs and became major panfish species for a few years. The shovelnose and the pallid sturgeon are among the native species that have nearly disappeared in the reservoirs. The paddle- fish has also been extirpated from much of the reservoir system, with remnant populations above Fort Peck Lake, at the confluence of the Yellowstone and Missouri rivers, and near the mouth of the Niobrara River. Symptomatic of the changes that have occurred in the Missouri River and floodplain ecosystem are the appearance of three federally listed threat- ened or endangered species. These are the least tern (Sterna antillarum), piping plover (Charadrius melodus), and a unique fish species, the pallid sturgeon (Scaphirhynchus albus). These species have generated much at-
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 77 tention with respect to prospective changes in Missouri River dam and reservoir operations, as the Corps of Engineers must respond to jeopardy biological opinions issued by the U.S. Fish and Wildlife Service regarding dam operations and the continued existence of these species (Appendix A, Table 4 lists the fish species found along the mainstem of the Missouri River today. The appendix also includes fishes that may exist on the floodplain in small creeks, or in overflow pools and oxbow lakes). The Pallid Sturgeon The pallid sturgeon was listed as endangered throughout its entire range on September 6, 1990, and the species is currently considered close to extinction (Dryer and Sandvol, 1993). Pallid sturgeons were thought to live primarily in large, turbid rivers such as the Missouri, and the Mississippi River downstream from its confluence with the Missouri. It utilized over- flow areas on the floodplain, backwaters, chutes, sloughs, islands, sand- bars, and main channel banklines, pools, and snags (Dryer and Sandvol, 1993). Because it feeds on aquatic invertebrates and fish that prey upon aquatic invertebrates, the lower velocity margins of the main and extra channels were essential habitats for the pallid sturgeon (Carlson et al., 1985). Some information suggested that pallid sturgeon readily utilized off-channel habitats for feeding and nursery and main channels for spawn- ing (Dryer and Sandvol, 1993; Keenlyne, 1989; Zuerlein, 1992). Some scientists have expressed concern that pallid sturgeon cannot reproduce in the Missouri Riverâs channelized and reservoir habitats (Henry and Ruelle, 1992; Ruelle and Henry, 1994). In 1993 it was concluded that sturgeon populations will continue to decline and that riverine habitat alteration and destruction are negatively impacting sturgeon recovery (National Paddle- fish and Sturgeon Steering Committee, 1993). Plains Minnows and Sauger The plains minnow and the sauger are two examples of common Mis- souri River native fish species that declined rapidly in the aftermath of dam construction and channelization. Plains minnows were once considered the most abundant minnow in the portion of the Missouri River in upper Missouri (Cross, 1967; Fisher, 1962; Jones, 1963; Morris et al., 1972; Pflieger, 1975). This small minnow was well adapted to the riverâs turbid environment. It lived among the numerous sandbars, feeding on living and dead plant material, and was an important component of the trophic web of the pre-regulation Missouri River (Hesse, 1994). The plains minnow has experienced a dramatic decline in abundance and is a much smaller compo- nent of the species composition today (Figure 3.11). It rebounded for a few
78 THE MISSOURI RIVER ECOSYSTEM 300 Percent Composition Catch Per Unit Effort 250 200 150 100 50 0 1945 1970-75 1986-93 1993-1997 1998 1999 2000 FIGURE 3.11 Percent composition and catch per unit effort of plains minnows (Hybognathus placitus) in seine hauls of equalized from sites adjacent to Nebraska, Iowa, and Missouri for periods between 1945 and 2000. SOURCE: Hesse, 1994. years during a wet period between 1993 and 1998, but the increase in abundance was quickly reversed, as floodplain connectivity was severed during 1998 and 1999. Saugers were common prior to channelization and impoundment of the Missouri River (Cross, 1967; Jones, 1963; Jordan and Evermann, 1969). The species comprised between 10 and 65 percent of the main channel large-river fish assemblage. They have since declined by as much as 98 percent in some locations in the river (Figures 3.11 and 3.12; from Hesse, 1994). Sauger were important sport fish of exceptional food quality, and recreational anglers fished for sauger before the mainstem dams were built. They are closely related to the walleye except they were widely adapted to the turbid environment of the Missouri River, and they were much more numerous than walleye before river regulation. The list of threatened or endangered Missouri River species continues to grow. Whitmore and Keenlyne (1990) noted that 82 species found along the Missouri River were listed as rare, threatened, or endangered by the seven states bordering the river. Included were 24 fish, 22 birds, 14 plants,
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 79 20 Catch Per Unit Effort 18 16 14 12 10 8 6 4 2 0 1963 1965 1971 1975 1986 1987 1988 1989 1990 1996 1997 1998 1999 2000 FIGURE 3.12 Catch per unit effort of sauger (Stizostedion canadense) by stan- dardized electrofishing from channelized sites an the Missouri River adjacent to Nebraska and Iowa during the period of 1963 through 2000. SOURCE: Hesse, 1994. 8 reptiles, 6 mammals, 6 insects, and 2 freshwater mussels (Appendix B lists threatened and endangered species along the Missouri River). Research on Floodplain Vegetation Prior to twentieth century human-induced environmental changes, the Missouri Riverâs floodplain vegetation was a storehouse of biodiversity. One of the few comprehensive surveys of the floodplain forest flora found 220 species of vascular plants growing in the remnant river section between Garrison Dam and Oahe Reservoir in North Dakota (Keammerer et al., 1975). This inventory was conducted long after extensive forest clearing had occurred and did not include a comparably rich flora of wetland plants found in non-forest communities on the floodplain. Studies of this 75-mile floodplain remnant by Keammerer et al. (1975) and Johnson et al. (1976) revealed a mosaic of aquatic, riparian, and terrestrial communities, includ- ing oxbow lakes, ponds, marshes, sand dunes, shorelines, in-channel is- lands, sand bars, forests, and agricultural fields.
80 THE MISSOURI RIVER ECOSYSTEM Natural vegetation communities along the Missouri featured forests with a wide variety of species. The dominant floodplain trees were cotton- wood (Populus deltoides), green ash (Fraxinus pennsylvanica var. lanceolata), box elder (Acer negundo), and American elm (Ulmus americana). Subdominant trees included peach-leaved willow (Salix amygdaloides) and bur oak (Quercus macrocarpa). Common shrubs and woody vines included dogwood (Cornus stolonifera) wolfberry (Symphoricarpos occidentalis), poison ivy (Rhus radicans), chokecherry (Prunus virginiana), juneberry (Amelanchier alnifolia), woodbine (Parthenocissus inserta), and fox grape (Vitis vulpina). Johnson et al. (1976) determined that these forests formed a succes- sional series of ecological communities, from the youngestâdominated by cottonwood-willow formed on fresh alluvium on low benchesâto the old- estâdominated by ash-box elder-elm on high benches. The river initiated the succession by meandering across its floodplain during floods and erod- ing older forests on the outside of river curves while creating point bars on the inside of curves for pioneer tree establishment. Approximately two- thirds of the floodplain forest flora occurred in the successional cotton- wood forests that depend on river meandering. No equally comprehensive inventories of floodplain vegetation have been published in downstream sections of the river; however, general de- scriptions are available and general ecological relationships are known. Downstream from North Dakota, the overall floodplain flora becomes considerably richer, particularly the woody component. For example, twenty-one species of trees were found on the Missouri River floodplain between Sioux City, Iowa and Rulo, Nebraska (Vaubel, 1973), approxi- mately three times more than along the Missouri River in central North Dakota (Johnson et al., 1976). On the lower portions of the Missouri River, cottonwood and willow were the dominant species on recently deposited and exposed sandbars, as they were throughout the length of the Missouri (Galat et al., 1996; Hesse et al., 1988). Later successional species were more diverse than in northern reaches of the river. For example, box elder, silver maple (Acer saccharinum), red mulberry (Morus rubra), and several elms replaced cot- tonwood and willow and formed an intermediate successional stage. The mature forest included several species of oaks (Quercus spp.), hickories (Carya spp.), black walnut (Juglans nigra), basswood (Tilia americana), hackberry (Celtis spp.), and sycamore (Platanus occidentalis). Weaver (1960) reported that along the Nebraska section of the Missouri River, forest, shrubs, and coarse grasses occupied most of the active floodplain, while higher terraces were nearly covered with prairie. Much of this diverse and extensive floodplain forest was cleared before significant regulation of the river. A large portion of these woodlands was
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 81 removed to provide fuel for steamboats during the nineteenth century, and more recently for agriculture. In the section of the Missouri River in North Dakota between Garrison Dam and Oahe Reservoir, 38 percent of the floodplain forest was cleared for agriculture between 1881-1938, an aver- age rate of approximately 0.7 percent per year (Johnson et al., 1982). An additional 18 percent of the forest was cleared for cultivation between 1938-1979, an average rate of approximately 0.5 percent per year. The majority of the woodlands along the river were removed to provide fuel for steam-powered vessels during the nineteenth century, and more recently for agricultural purposes. Thus, clearing activities claimed approximately 56 percent of the original floodplain forest by 1979. Considerably more forest has been cleared since then, but the amount has not been quantified. Clearing of the Missouriâs floodplain forest occurred much earlier along the riverâs lower portions. Bragg and Tatschl (1977) estimated that 76 percent of the floodplain in Missouri was forested in 1826. By 1937, the percentage had dwindled to 17 percent, and by 1972 had dropped to 13 percent. The percentage of the floodplain cultivated increased accordingly, from 18 percent in 1826 to 83 percent in 1972. Hesse et al. (1988) estimated cover changes on the floodplain for a larger area, from St. Louis upstream to Rulo, Nebraska, including the area studied by Bragg and Tatschl in 1977. Cultivated land increased 43-fold between 1892 and 1982 (from 2,339 hectares to 100,091 hectares), while during the same period forests decreased by 41 percent, wetlands by 39 percent, sandbars by 97 percent, and grasslands by 12 percent (Hesse et al., 1988). The authors suggested that higher floodplain ground was cleared of forest prior to the dams. Dam construction and the cessation of flooding then stimulated another round of forest conversion to agriculture down- stream of and between the reservoirs, this time on lower ground. Significant changes in Missouri River fauna also occurred long before river regulation. For example, some of the nearly 160 species and their habitats first described by Lewis and Clark (Burroughs, 1961) went extinct or were extirpated from the region because of hunting and loss of habitat. These species included bison, elk, grizzly bear, wild sheep, swans, and the Carolina parakeet. These animals used the Missouri River valley as protec- tion from harsh winters and summer heat. The beaver was essentially trapped out of much of the northern Great Plains. Wetland drainage, stream channelization, and trapping kept beaver numbers low during much of the first half of the 1900s (Bennitt and Nagel, 1937). More recently, due partly to the declining value of wild fur, beaver are more numerous even along portions of the channelized reaches (Larry Hesse, River Ecosystems, Inc., personal communication, 2001). Botanical research from 1960 to 2000 on the upper Missouri River centered on the effect of dams on floodplain forest succession, particularly
82 THE MISSOURI RIVER ECOSYSTEM the effect of altered hydrology and sediment regime on river meandering and on cottonwood regeneration. Natural succession patterns were worked out in remnant reaches shortly after the dams were built. No comprehen- sive studies of forest succession were conducted prior to construction of the mainstem dams. Botanical research was concentrated in two Missouri River reachesâthe unchannelized floodplain downstream from Gavins Point dam (Wilson, 1970), and the remnant floodplain between Garrison Dam and the headwaters of Oahe reservoir, near Bismarck, North Dakota (Johnson et al., 1976). Later investigations of cottonwood regeneration in the reach of the Missouri Riverâs section designated as a national monu- ment in Montana (formerly designated Wild and Scenic) were conducted by Scott et al. (1997) and others. The studies in North Dakota reconstructed (post hoc) the patterns and processes of forest succession on the Missouri River floodplain under natu- ral, pre-regulation conditions, concluding that the key geomorphic process directing vegetation succession was the riverâs meandering channel. The Missouri River moved across its floodplain during high flow periods; its rate of movement was especially rapid during floods. In places, the river had moved several miles in less than a century (Johnson, 1992). During episodes of meandering, the outer bank of the river eroded while the inner bank accreted. On the riverâs outer curve, forests in various stages of successional development, along with other floodplain habitats such as sand dunes and marshes, were undermined. On the riverâs inner curve, new land formed that was ideal for forest regeneration. During floods, the Missouri River channel was filled with fallen trees, which eventually settled to the bottom of the river and became snags. Cottonwood and willow (both tree and shrub growth forms) were the first woody species to colonize the newly accreted areas (point bars). They dominated the pioneer vegeta- tion throughout the length of the Missouri River, but were especially im- portant in the upper, more westerly, reaches where few other tree species could grow. The success of cottonwood and willow was attributed to their specific adaptations to riparian zones, which includes: rapid seed germination to grow immediately after spring floods, rapid root and height growth en- abling tolerance to flooding, drought, and sedimentation, tolerance to the often low soil fertility on sandbars, and the ability to reproduce vegeta- tively, particularly after physical damage from floods or from beaver. Evi- dence for the success of cottonwood was its historic dominance of the floodplain vegetation; over eighty percent of the extensive forests on the floodplain of the pre-regulation Missouri River in North Dakota had cot- tonwood as their most important tree (Johnson et al., 1992). Research also uncovered a key fact about cottonwood: It cannot repro- duce successfully in its own forests. As such, it depended on the creation of
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 83 new land from an actively meandering river for its persistence and promi- nence in the vegetation. It behaved as a disturbance-dependent, fugitive species that declined in abundance in stable portions of the floodplain while increasing in other portions recently reworked by the Missouri River. Later successional tree species replaced it over time in stable areas. The Missouri Riverâs pre-regulation floodplain was a mosaic of forests with a wide range of ages, from young cottonwood-willow forests a decade or two old, to forests of ash, box elder, elm, and oak that were old enough to have lost all traces of the cottonwood pioneer element. Of the forests studied by Johnson et al. (1976), approximately two-thirds were early to mid-successional (dominated by cottonwood), while one-third were domi- nated by later successional tree species. High biodiversity both within forest communities and across the flood- plain could not be maintained without the rejuvenating forces of floods and channel meandering. Johnson et al. (1976) found the highest tree diversity mid-way through succession, a period when all tree species grew together, with cottonwood and willow in the overstory, and with ash, box elder, elm, and oak in the understory. Hibbard (1972) found similar patterns for forest-dwelling animals; the number of species of both small mammals and birds peaked in forests of mid-successional age. This ephemeral, species- rich stage was historically sustained by new forests created by a meandering river. Hibbard (1972) found that the floodplain forest community provided nesting habitat for a wide range of bird species, from open country birds in the youngest, post-flood communities, to shrub-loving bird species in middle-aged cottonwood communities, to forest-dwelling birds in the most mature forests. Peaks in both species number and population of birds were reached in older successional forests because of their high vertical stratifica- tion (Hibbard, 1972). The large size and hollow trunks and branches of older cottonwood trees provided nesting cavities for woodpeckers and wrens (Knopf, 1986). More than 50 species of songbirds were found by Liknes et al. (1994) along the upper Missouri River, and approximately 50 percent were neotropical migrants. Dean (1998) found 39 species of neotropical migrants utilizing Missouri River floodplain forests as stopover habitat. Model calculations suggest that without changes to the current man- agement regime, cottonwood forests will essentially be lost as a significant community on remnant floodplains in less than a century (Johnson, 1992). The cottonwood forests that remain on the floodplain between and imme- diately below dams cannot be sustained by the current low river meander- ing rates. Both erosion and deposition rates (an index of river meandering rate) have decreased substantially since the closure of the mainstream dams. Deposition rates have fallen to one percent of their pre-regulation levels and erosion rates have fallen to twenty-five percent of their pre-regulation lev-
84 THE MISSOURI RIVER ECOSYSTEM els. Cottonwood forests are forecast to be replaced by those dominated by green ash. These future forests, assuming that they escape clearing for agricultural expansion, are likely to be considerably lower in tree and bird diversity primarily because of the loss of pioneer plant species, loss of vertical structural complexity, and the loss of nesting cavities found mostly in old cottonwood trees. Reduced channel meandering was not the only impact of flow regula- tion by dams on floodplain vegetation. Reily and Johnson (1982) found that seasonal shifts in flow, the near-elimination of overbank flooding, and a lowering of the water table below the floodplain in late spring, reduced the growth of trees occupying remnant floodplains between reservoirs. Peak river flow no longer occurs early in the growing season. It is consequently out of phase with the vernal growth pattern typical of floodplain trees. The absence of flooding except on the lowest benches represents a loss of soil moisture to the floodplain compared to pre-regulation conditions. Climatic conditions in the upper Missouri River region are relatively dry. As such, overbank flooding was important for the growth and persistence of trees. Several scientists have recommended measures to regenerate cotton- wood forests in sections of the Missouri River affected by mainstem dams. Johnson (1992), for example, suggested planting native pioneer trees, such as cottonwoods and willows, on marginal agricultural land unless a more natural flow regime could be restored. While planting can, with time, maintain certain important ecosystem properties, such as cavity-nesting habitat for birds, it cannot restore certain other properties, such as the high species diversity of early successional forests. Pre-regulation forests were established on relatively low benches and were repeatedly aggraded by siltation from floods. As a result, these communities supported a signifi- cant proportion of wetland species. This species diversity cannot be re- stored by tree planting on relatively high benches. COMMITTEE COMMENTARY The Missouri River ecosystem experienced a variety of human-caused environmental changes during the twentieth century. Before these changes, the Missouri River experienced annual floods, with occasional massive flooding. It meandered freely across its floodplain, it carried large amounts of debris and snags, especially during floods, and it eroded, transported, and deposited voluminous amounts of sediment. These dynamic geomor- phic processes promoted erosion on the riverâs undercutting banks and deposition on its inner banks. This pre-regulation, physical variability was important to sustaining the river systemâs biological diversity and produc- tion.
MISSOURI RIVER AND FLOODPLAIN ECOLOGY 85 Engineering works constructed during the twentieth century aimed to enhance conditions for navigation and to provide protection against floods by reducing this variability. Mainstem dams and channelization greatly changed the riverâs physical systems. Much of the riverâs huge sediment load was deposited in the massive reservoirs, resulting in sediment imbal- ances and marked channel incision below the dams. The massive mainstem reservoirs submerged stretches of free-flowing river and floodplain forest. Changes in habitat allowed some native species that were challenged by pre-regulation conditions to thrive. Some species that thrived under pre- regulation conditions, such as the pallid sturgeon and sauger, experienced sharp reductions. The riverâs periodic flooding also was greatly reduced, and even eliminated, in stretches under the damsâ stabilizing influences. This resulted in the reduction or loss of ecologically-beneficial flood pulses and low flows. The hydrologic connections between the river channel, floodplain, and backwater areas were greatly disrupted. In the channelized portions, river meandering was eliminated and the ecological diversity of the river and floodplain was greatly simplified. A large body of scientific research has identified these ecological changes, including declines in many native species and a general decline in the overall integrity of the ecosystem that have accompanied the mainstem dams and other human influences on the ecosystem. As in all large ecosystems, uncertaintiesâsome of which are essentially irreducibleâremain in the scientific understanding of the Missouri River and floodplain ecosystem. Nonetheless, the scientific research provides a good picture of the funda- mentals of the ecosystemâs pre- and post-regulation ecological structure and function. Although much contemporary research retains a species-specific focus, some scientists and organizations are integrating research across disciplines and across the entire river system. This broader perspective and inquiry will complement the existing science in promoting a more systematic under- standing of the Missouri River ecosystem. Mainstem dams and reservoirs, channelization, and a management re- gime that promotes hydrologic stability have all contributed to reductions in the ecosystemâs dynamic properties. Current scientific inquiry is thus hindered in its ability to investigate these dynamic processes and how they affect the ecosystem and its cadre of organisms. The greatest unknowns in the science of this ecosystem are in understanding its responses to changes in the current management regime.