1720 D. B. OSMUNDSON ET AL. Ecological Applications
<br />Vol. 12, No. 6
<br />Fausch et al. 1988), yet many studies suggest that biotic
<br />factors such as food availability may be more important
<br />(Cada et al. 1987, Filbert and Hawkins 1995, Hughes
<br />1998, Nislow et al. 1999). With the exception of the
<br />flood-pulse concept, wherein main-channel productiv-
<br />ity is enhanced by the entrainment of floodplain nu-
<br />trients and carbon (Junk et al. 1989, Bayley 1995, Pow-
<br />er et al. 1995), few investigations have attempted to
<br />link flow effects to availability of food for fishes.
<br />To restore native fish populations in regulated rivers,
<br />management strategies must be based on an under-
<br />standing, not only of the fishes' life-history attributes,
<br />but also of the ecosystem that sustains them (Power et
<br />al. 1995, Stanford et al. 1996, Wootton et al. 1996).
<br />Petts (1991) identified the need for integration of in-
<br />terdisciplinary research that approaches the study of
<br />regulated rivers on a longitudinal basis and in a hier-
<br />archical manner: hydrology and geomorphology to de-
<br />scribe the influence of flow regime and sediment sourc-
<br />es on river structure, and ecology to link these influ-
<br />ences to riverine fauna. However, large-scale riverine
<br />processes are inherently difficult to study (Schlosser
<br />1991, Johnson et al. 1995, Peterson and Kwak 1999).
<br />With some exceptions, experimental studies are gen-
<br />erally not possible (Power et al. 1995, Poff et al. 1997),
<br />so correlative empirical approaches are required to dis-
<br />cern linkages among flow, sediment, and biota.
<br />The Colorado River, in southwestern United States,
<br />is one of the most regulated rivers in North America
<br />(Fradkin 1981). Nearly all of the pathways by which
<br />regulation can alter or degrade native aquatic fauna are
<br />represented, supplying diverse challenges for the ap-
<br />plied ecologist. Driven by the mandates of the Endan-
<br />gered Species Act, resource managers need solutions
<br />to ecologically complex problems (Stanford 1994,
<br />Holling and Meffe 1996). Herein, we report the results
<br />of our investigations concerning, physical habitat and
<br />trophic relations in a contiguous portion of the upper
<br />Colorado River inhabited by all life stages of Ptych-
<br />ocheilus lucius Girard (Colorado pikeminnow), an en-
<br />dangered piscivorous cyprinid. For this population, low
<br />recruitment and low adult carrying capacity are two
<br />primary constraints that limit population viability (Os-
<br />mundson and Burnham 1998, Osmundson et al. 1998).
<br />In addition to the obvious effects of reduced range from
<br />instream barriers, we hypothesized that the capacity of
<br />the system to support adult pikeminnow may have de-
<br />clined from a reduction in food availability due to in-
<br />creased persistence of fine sediment (particles <2 mm),
<br />a possible effect of river regulation.
<br />We used an interdisciplinary approach to investi-
<br />gating the role of flow regime in riverine trophic struc-
<br />ture, integrating geomorphology with studies of fish
<br />community structure and biomass of lower trophic lev-
<br />els (Petts 1991, Johnson et al. 1995). We first described
<br />the distribution of physical habitat and biota on a lon-
<br />gitudinal basis throughout the present and historical
<br />range of Colorado pikeminnow in the upper Colorado
<br />River mainstem. Then, using empirical relations, we
<br />sought links among sediment distribution, standing
<br />crops of primary producers, consumers, and the prin-
<br />cipal piscivore (Colorado pikeminnow). The relation
<br />between flow and sediment movement was next as-
<br />sessed through standard geomorphological analyses.
<br />Utilizing these empirical relations and historical flow
<br />records, we evaluated the potential for reduced predator
<br />food availability as affected by functional changes in
<br />flow. Our goal was to assess links among flow, sedi-
<br />ment, and biota in a large river system and determine
<br />if such linkages might provide a food-mediated path-
<br />way by which flow regime alteration affects fish as-
<br />semblage structure. We discuss the implications for na-
<br />tive fish restoration efforts in general and provide man-
<br />agement recommendations for conservation of Colo-
<br />rado pikeminnow in the upper Colorado River.
<br />SPECIES OVERVIEW AND STUDY AREA
<br />Colorado pikeminnow
<br />The Colorado pikeminnow, a large, long-lived pis-
<br />civore, along with three other large-river, warm-water
<br />fish species (two chubs, Gila cypha Miller and G. ele-
<br />gans Baird and Girard, and one sucker Xyrauchen tex-
<br />anus Abbott) endemic to the Colorado River Basin, are
<br />classified as endangered under the Endangered Species
<br />Act (U.S. Fish and Wildlife Service 2000). All four
<br />once occurred throughout the basin, ranging from Wy-
<br />oming south to the Gulf of California (Miller 1961).
<br />Due primarily to extensive dam construction during the
<br />1930s to 1960s, populations of Colorado pikeminnow
<br />were extirpated in the lower sub-basin (downstream of
<br />Lee's Ferry) by the 1970s and today occur only up-
<br />stream of Glen Canyon Dam, primarily in the states of
<br />Utah and Colorado (Minckley and Deacon 1968,
<br />Minckley 1973). In the upper sub-basin, the range of
<br />this species was reduced early in the century by di-
<br />versions built for local irrigation projects, and addi-
<br />tionally by the more recent (1960s-1980s) construction
<br />of several large dams. Lotic environments were con-
<br />verted to lentic habitats, downstream water tempera-
<br />tures were reduced by hypolimnetic releases, and mi-
<br />gration routes were blocked (Ono et al. 1983). After a
<br />75% basin-wide loss of range, a viable population oc-
<br />curs today only in the Green River system (Gilpin
<br />1993). A smaller, less viable population occurs in the
<br />mainstem Colorado River and its Gunnison River trib-
<br />utary (Osmundson and Burnham 1998). Populations of
<br />Ptychocheilus lucius appear to require an extensive
<br />length of river with an array of habitat types to meet
<br />the changing needs of different life stages. Larvae
<br />hatched in gravel-cobble substrates of high-gradient
<br />reaches drift 100-200 km downstream to low-gradient
<br />reaches where backwaters formed in silt-sand bars pro-
<br />vide ideal nursery habitat (Haynes et al. 1984, Tyus
<br />and Haines 1991). Insectivory is largely replaced by
<br />piscivory during the first year (Vanicek and Kramer
|