Post on 21-Jul-2016




3 download

Embed Size (px)




    Zachary 0. Toledo and J. Boone Kauffman2

    ABSTRACT: Intact riparian zones are the product of an incrediblycomplex multitude of linkages between the geomorphic, hydrologic,and bioti features of the ecosystem. Land-use activities that severor alter these linkages result in ecosystem degradation. We exam-ined the relationship between riparian vegetation and channel mor-phology by sampling species composition and herbaceous rootbiomass in incised (down-cut and widened) versus unincised(intact) sections of unconstrained reaches in three headwaterstreams in northeastern Oregon. Incision resulted in a composition-al shift from wetland-obligate plant species to those adapted todrier environments. Root biomass was approximately two timesgreater in unincised sections than incised sections and decreasedwith depth more rapidly in incised sections than in unincised sec-tions. Thtal root biomass ranged from 2,153 g m2 to 4,759 g m2 inunincised sections and from 1,107 g m-2 to 2,215 g m-2 in incisedsections. In unincised sections less than 50 percent of the total rootbiomass was found in the top 10 cm, with approximately 20 percentin successive 10-cm depth increments. In contrast, incised sectionshad greater than 60 percent of the total root biomass in the top 10cm, approximately 15 percent in the 10 to 20 cm depth, less than 15percent in the 20 to 30 cm depth, and less than 10 percent in the 30to 40 cm depth. This distribution of root biomass suggests a positivefeedback between vegetation and channel incision: as incision pro-gresses, there is a loss of hydrologic connectivity, which causes ashift to a drier vegetation assemblage and decreased root structure,resulting in a reduced erosive resistance capacity in the lower zoneof the streambank, thereby allowing further incision and widening.(KEY TERMS: riparian ecosystems; erosion sedimentation; rootbiomass; root distribution; channel morphology; incision.)


    Riparian areas are zones of direct interactionbetween terrestrial and aquatic ecosystems (Gregoryet al., 1991). While riparian areas comprise only 1 to 2percent of the land area in arid systems (Kauffman

    and Krueger, 1984), they are disproportionately sig-nificant in terms of biological production and diversity(Gregory et al., 1991; Naiman and Descamps, 1997;Patten 1998, Kauffman et al., 2001). Riparian areasare valuable to society through their multitude ofecosystem functions and processes, such as floodabatement, habitat for migratory birds and aquaticspecies (Naiman et al., 1993), maintenance of regionalbiodiversity (Naiman and Descamps, 1997), andwater quality control and nutrient cycling (Green andKauffman, 1989). Riparian areas have received con-siderable attention by scientists and managers (John-son et al., 1985; Abell, 1989; Clary et al., 1992;Tellman et al., 1993; Feller, 1998; Koehler andThomas, 2000; Wigington and Beschta, 2000) becauseof concerns centering on their widespread degradation[National Resource Council (NRC), 1992; Beschta,19971.

    Riparian areas throughout the western UnitedStates have been altered or degraded through land-use activities including hydrologic alterations(Dominick and O'Neill, 1998), beaver removal(Naiman et al., 1988), and livestock grazing (Fleischn-er, 1994; Dwire et al., 1999; Kauffman and Pyke,2001). Like many western landscapes, eastern Oregonriparian areas have been affected by additional land-use activities, including mining, logging, splash dams,and road building (McIntosh et al., 1994).

    Riparian-stream degradation occurs when hydro-logic, geomorphic, or biotic processes are disruptedsuch that interactions or linkages between these fea-tures are disrupted (Figure 1). For example, channelincision can sever linkages between floodplains andstreams, which then alters the biotic communities

    'Paper No. 01046 of the Journal of the American Water Resources Association. Discussions are open until August 1, 2002.-

    'Respectively, Fisheries Biologist, Mason, Bruce and Girard, Inc., 707 SW. Washington Street, Suite 1300, Portland, Oregon 97205; andProfessor, Department of Fisheries and Wildlife, Oregon State University, 104 Nash Hall, Corvallis, Oregon 97331 (E-MaillToledo:


  • Toledo and Kauffman

    (Junk et al., 1989; Robertson et al., 2001). Incisionmay be expected when there is an imbalance betweenerosive and resistive forces acting on the bank andbed material (Schumm, 1999). Shifts in the forces act-ing on the bank can be caused by a change in thehydrology, geomorphology, or vegetation, which influ-ence each other in a positive feedback response (Fig-ure 1; Kauffman et al., 1997). The positive feedbackresponse can be described as follows: channel mor-phology influences the water availability to the ripari-an area (floodplain) (Leopold and Maddock, 1953),water availability affects plant species compositionthat comprises the riparian communities (Hupp andOsterkamp, 1996; Otting, 1998; Chapin et al., 2000),and both aboveground and belowground vegetationcomponents affect the channel morphology by influ-encing both erosive and resistive forces (Hickin, 1984;Thorne, 1990; Hupp, 1999).

    Figure 1. Linkages Between Geomorphology; Hydrology,and the Biota. Altering one of these features will result

    in a positive feedback response to the others(modified from Kauffman et al., 1997).

    Vegetation has been shown to strongly influencechannel morphology (Gregory and Gurnell 1988).Hickin and Nanson (1984) found that unvegetatedchannels could exhibit double the lateral channelmigration rates of vegetated channels. Smith (1976)found that bank sediment with a 5 cm root mat and aroot volume of 16 to 18 percent had 20,000 times theresistance of equivalent bank sediment without vege-tation or roots. Clifton (1989) reported that whenWickiup Creek (in central Oregon) was allowed to re-vegetate following livestock exclusion, the channelgained 60 cm of sediment within ten years and the

    channel cross-sectional area had decreased by 94 per-cent after 50 years.

    We hypothesized that there are four potentialresponses of root biomass to a decrease in water avail-ability associated with channel incision (Figure 2).There can be: (a) a lower total root biomass but with asimilar distribution within the soil horizons; (b) alower total root biomass and an increased rate of losswith depth; (c) no change in total root biomass or dis-tribution; or (d) an increased level of total rootbiomass. The second response, (b), would have thegreatest potential to affect channel morphology anderosion potential because there would be an overalldecrease in root biomass, particularly at greaterdepths. The objective of this study was to examine therelationships between riparian vegetation, rootbiomass, and channel morphology in incised andunincised stream channels.


    The study reaches are located in the Blue Moun-tains of northeastern Oregon. During preliminaryreconnaissance, we examined 21 streams and selectedthree that met our criteria: an unconstrained alluvialchannel; a low level of recent human impact (i.e.,enhancement structures, livestock grazing, etc.); andthe presence of a hydrologic knickpoint (i.e., an areawhere an abrupt change of elevation and slope gradi-ent occurs (Brooks et al., 1997) that separates anupstream, unincised section from a downstream,incised section. Starting points of the sections werechosen with regards to tributary junctions, location ofknickpoints, and changes in valley form. Incised sec-tions were determined using previously collectedwater table data and/or physical parameters such asbank height and active channel width. Incised andunincised sections were similar in floodplain geomor-phology. The causes of channel incision likely includedroads, mining, and grazing.

    Crane Creek (445308"N, 11823'SO'W; elevation1,680 m) is a third-order tributary to the North ForkJohn Day River. The knickpoint was a road culvert;however, it was not determined whether the culvertinitiated or halted the headcut (knickpoint). Theunincised section was approximately 100 m upstreamfrom the incised section.

    Little Fly Creek (4503'45"N, 118'30'15'W; eleva-tion 1,460 m) is a third-order tributary to the GrandeRonde River. The knickpoint was also a culvert, whichhas halted the upstream migration of the headcut(Figure 3). The incised section was approximately200 m downstream of the unincised section.


  • Root Biomass in Relation to Channel Morphology of Headwater Streams

    Root Biomass


    Root Biomass


    bRoot Biomass

    7. /- (


    Root Biomass

    D 7?( d

    Figure 2. Four Hypothesized Responses of Root Biomass to Channel Incision: (a) Less Overall Root Biomass ThanUnincised Stream, Same Rate of Loss With Depth; (b) Less Overall Root Biomass Than Unincised Stream, and anIncreased Rate of Loss With Depth; (c) No Change From an Unincised Stream; and (d) Increased Level of Biomass.


    Little Fly Creek

    1999Figure 3. The Incised Section of Little Fly Creek in 1977 (left) and 1999 (right). The man (circled) was standing

    at the headcut in 1977. The tree (circled) to the left of both photos provides reference and scale.Note the change in channel width and bank height - an obvious disconnect from the floodplain.


  • Thiedo and Kauffman

    Squaw Creek (4507'44N, 1183223"W; elevation1,370 m) is a second-order tributary to the GrandeRonde River and has been the subject of a previousvegetative and hydrologic study, which described thepresence of the hydrologic knickpoint using watertable elevation (Otting, 1998). The end of the unin-cised section is approximately 10 m upstream fromthe beginning of the incised section.


    The length of each sampled section was at least 20-times the average active channel width, with a mini-mum length of 40 m. Average valley bottom width andaverage geomorphic floodplain width were measuredat high flow conditions from ten equally spaced tran-sects located perpendicular to the valley sides. Thegeomorphic floodplain was defined as the near hori-zontal surface formed by fluvial deposition (lateraland vertical accretion) from the channel, which is fre-quently flooded under present conditions (i.e., activelybeing formed; Knighton, 1998). The entrenchmentratio (floodplain width to active channel width) wasdetermined for each section.

    Channel measurements were taken at cross-chan-nel transects spaced at one meter during low flow con-ditions (Robison, 1988; Robison and Beschta, 1989).These measurements included active channel width,wetted width, thalweg depth, bank height, overhang-ing bank presence, and depth of overhang. Activechannel width was determined using visual cluessuch as scour, vertical accretion of fine sediment, andmorphology (Harrelson et al., 1994). Thalweg depth isthe point of greatest scour, or the maximum depthalong the cross-channel transect. Bank height wasdefined as the height from the water surface to thefirst dominant topographic plane of the streambank(floodplain).

    Vegetation composition was collected within a 25-cm x 25-cm plot at one-meter intervals for 40 m alongthe green line of both banks for each section. Thegreen line is defined as the first perennial vegetationabove the stable low water line (Bauer and Burton,1993). All plants found in the plot were identified tospecies (Hitchcock and Cronquist, 1973; Walters andKeil, 1996) and their cover was recorded. Each specieswas assigned into a "wetland indicator category"ranging from wetland-obligate to upland-obligate(Table 1) based upon classifications in Reed (1986).Wetland indicator categories provide insight into thehydric environment occupied by each species, andthus provide some indication of hydrologic connectivi-ty at the immediate streambank.

    TABLE 1. Wetland Indicator Categories (from FederalInteragency Committee for Wetland Delineation, 1989).

    Probability ofOccurrence inin Wetlands

    Wetland Indicator Category (:percent)Obligate Wetland > 99Facultative Wetland 67-99Facultative 34-66Facultative Upland 1-33Obligate Upland

  • Root Biomass in Relation to Channel Morphology of Headwater Streams

    the incised sections, the unincised sections were char-acterized as narrow channels with deep water depthsat low flow, wide floodplains, low banks, and high per-centages of overhanging banks (Figure 4). Averageactive channel width was less than 200 cm in theunincised sections while greater than 350 cm in theincised sections. Average wetted width was less than160 cm in the unincised sections (Table 2) and greaterthan 180 cm in the incised sections. Average thalwegdepth was significantly greater in unincised sectionscompared to incised sections.

    Floodplain width was less than 10 m in all incisedsections while ranging from 15 to 34 m in the unin-cised sections (Table 2). Entrenchment ratios were 6.5to 18 times greater in the unincised sections than inthe incised sections. Even with the broad valley bot-toms, all of the incised sections had entrenchmentratios at or below two, with the floodplain less thantwo times as wide as the active channel. Thus only anarrow corridor of floodplain remained for aquatic-terrestrial interactions during overbank flows.

    Unincised sections had mean bank heights lessthan 42 cm while the incised sections had mean bankheights of 43 to 150 cm. For example, the mean bankheight at Little Fly Creek was 41 cm and 150 cm inthe unincised and incised sections, respectively. Over59 percent of the sampled streambank length in theunincised sections had overhanging banks (59 percentat Squaw Creek to 100 percent in Crane Creek) whilethe frequency of overhanging banks along the sam-pled streambank length in the incised sections wasalways less than 50 percent (i.e., 31 percent at SquawCreek and 43 percent at Crane Creek). The averagedepth of overhanging banks was approximately 1.5 tothree times greater in the unincised sections than theincised sections.

    Species and Community Composition

    Calculation of streambank plant species composi-tion based upon wetland indicator categories providesa good indication of the degree of hydrologic connec-tivity between the channel and terrestrial zones (Fig-ure 5). Obligate-wetland and facultative-wetlandspecies dominated the unincised sections while facul-tative-upland species largely dominated the incisedsections. For example, obligate-wetland and faculta-tive-wetland categories composed 80 percent of thespecies in the unincised section of Crane Creek and86 percent of the species in the unincised section ofSquaw Creek. Whereas obligate-wetland and faculta-tive-wetland categories composed only 41 percent ofthe species in the unincised section of Crane Creekand 53 percent of the species in the unincised sectionof Squaw Cree...


View more >