chapter 8: restoration design restoration design 8 · chapter 8: restoration design final...

CHAPTER 8: RESTORATION DESIGN

FINAL MANUSCRIPT – 5/11/98 8 – 1

Part III

8Restoration Design

Design can be defined as the intentionalshaping of matter, energy, and process tomeet an expressed need. Planning and de-sign connect natural processes and culturalneeds through exchanges of materials, flowsof energy, and choices of land use and man-agement. One test of a successful streamcorridor design is how well the restored sys-tem sustains itself over time while accommo-dating identified needs.

To achieve success, those carrying out resto-ration design and implementation in variable-land-use settings must understand the streamcorridor, watershed, and landscape as a

complex of working ecosystems that influenceand are influenced by neighboring ecosys-tems (Figure 8.1 ). The probability of achiev-ing long-term, self-sustaining functions acrossthis spatial complex increases with an under-standing of these relationships, a commonlanguage for expressing them, and subse-quent response. Designing to achieve stream-or corridor-specific solutions might not resolveproblems or recognize opportunities in thelandscape.

Stream corridor restoration design is stilllargely in an experimental stage. It is knownhowever, that restoration design must con-sider site-specific or local conditions to besuccessful. That is, the design criteria, stan-dards, and specifications should be for thespecific project in a specific physical, climatic,and geographic location. These initiatives,however, can and should work with, ratherthan against, the larger systems of which theyare an integral part.

This approach produces multiple benefits,including:

• A healthy, sustainable pattern of landuses across the landscape.

• Improved natural resource quality andquantity.

Figure 8.1: Stream running through a wet meadow.Restoration design must consider site-specific conditions asan integral part of larger systems.

8.A Valley Form, Connectivity, and Dimension

8.B Soil Properties

8.C Plant Communities

8.D Habitat Measures

8.E Stream Channel Restoration

8.F Streambank Restoration

8.G Instream Habitat Recovery

8.H Land Use Scenarios

STREAM CORRIDOR RESTORATION: PRINCIPLES, PROCESSES, AND PRACTICES

8 – 2 FINAL MANUSCRIPT – 5/11/98

“Leave It Alone / Let It Heal Itself”

There is a renewed emphasis on recovering damaged rivers (Barinaga 1996). Along with this concern,however, people should be reminded periodically that they serve as stewards of watersheds, not justtinkerers with stream sites. Streams in pristine condition, for example, should not be artificially “improved”by active rehabilitation methods.

At the other end of the spectrum, and particularly where degradation is caused by off-stream activities, thebest solution to a river management problem might be to remove the problem source and “let it heal itself.”Unfortunately, in severely degraded streams this process can take a long, time. Therefore the “leave italone” concept can be the most difficult approach for people to accept (Gordon et al. 1992).

• Restored and protected stream corri-dors and associated ecosystems.

• A diversity of native plants and ani-mals.

• A gene pool that promotes hardiness,disease resistance, and adaptability.

• A sense of stewardship for privatelandowners and the public.

• Improved management measures thatavoid narrowly focused and frag-mented land treatment.

Building on information presented in Parts Iand II, this chapter contains design guidanceand techniques to address changes causedby major disturbances and to restore streamcorridor structure and function to a desiredlevel. It begins with larger-scale influencesthat design may have on stream corridorecosystems, offers design guidance primarilyat the stream corridor and stream scales, andconcludes with land use scenarios.

The chapter is divided into seven sections.

Section 8.A: Valley Form, Connectivity,and DimensionThis section focuses on restoring structuralcharacteristics that prevail at the streamcorridor and landscape scales.

Section 8.B: Soil PropertiesThe restoration of soil properties that arecritical to stream corridor structure and func-tions are addressed in this section.

Section 8.C: Vegetative CommunitiesRestoring vegetative communities is a highlyvisible and integral component of a function-ing stream corridor.

Section 8.D: Habitat MeasuresThis section presents design guidance forsome habitat measures. They are often inte-gral parts of stream corridor structure andfunctions.

Section 8.E: Stream Channel RestorationRestoring stream channel structure andfunctions is often a fundamental step in restor-ing stream corridors.

Section 8.F: Streambank RestorationThis section focuses on design guidelines andrelated techniques for streambank stabiliza-tion. These measures can help reduce sur-face runoff and sediment transport to thestream.

Section 8.G: Instream Habitat RecoveryRestoring instream habitat structure andfunctions is often a key component of streamcorridor restoration.

Section 8.H: Land Use ScenariosThis final section offers broad design con-cepts in the context of major land use sce-narios.



(a)

(b)

Figure 8.2: Stream corridors.(a) Stream valley side slopes and (b) floodplain gradients influence streamcorridor function.

Valley form, connectivity, and dimen-sion are variable structural characteris-tics that determine the interrelation-ship of functions at multiple scales.Valley intersections (nodes) withtributary stream corridors, slope ofvalley sides, and floodplain gradientare characteristics of valley form thatinfluence many functions (Figure 8.2).

The broad concept of connectivity, asopposed to fragmentation, involveslinkages of habitats, species, commu-nities, and ecological processes acrossmultiple scales (Noss 1991). Dimen-sion encompasses width, linearity, andedge effect, which are critical formovement of species, materials, andenergy within the stream corridor andto or from ecosystems in the surround-ing landscape. Design should thereforeaddress these large-scale characteris-tics and their effect on functions.

Valley FormIn some cases, entire stream valleyshave changed to the point of obscuringgeomorphic boundaries, makingstream corridor restoration difficult.Volcanoes, earthquakes, and landslidesare examples of natural disturbancesthat cause changes in valley form.Encroachment and filling of flood-plains are among the human-induceddisturbances that modify valley shape.

Stream Corridor Con-nectivity and DimensionConnectivity and dimensions of thestream corridor present a set of design-related decisions to be made. Howwide should the corridor be? How

8.A Valley Form, Connectivity, and Dimension

long should the corridor be? What ifthere are gaps in the corridor? Thesestructural characteristics have a sig-nificant impact on corridor functions.The width, length, and connectivity ofexisting or potential stream corridorvegetation, for example, are critical to



habitat functions within the corridorand adjacent ecosystems.

Generally, the widest and most con-tiguous stream corridor whichachieves habitat, conduit, filter, andother functions (see Chapter 2) shouldbe an ecologically derived goal ofrestoration. Thresholds for eachfunction are likely found at differentcorridor widths. The appropriate widthvaries according to soil type, withsteep slopes requiring a wider corridorfor filter functions. A conservativeindicator of effective corridor width iswhether a stream corridor can signifi-cantly prevent chemical contaminantscontained in runoff from reaching thestream (Forman 1995).

As discussed in Chapter 1, the corridorshould extend across the stream, itsbanks, the floodplain, and the valleyslopes. It should also include a portionof upland for the entire stream lengthto maintain functional integrity(Forman and Godron 1986).

A contiguous, wide stream corridormight not be achievable, however,particularly where competing landuses prevail. In these cases, a ladderpattern of natural habitat crossing the

floodplain and connecting the uplandsegments might facilitate sedimenttrapping during floods and providehydraulic storage and organic matterfor the stream system (Dramstad et al.1996).

Figure 8.3 presents an example ofthese connections. The open areaswithin the ladder pattern are represen-tative of areas that are unavailable forrestoration because of competing landuses.

Innovative management practices thatserve the functions of the corridorbeyond land ownership boundaries canoften be prescribed where land ownersare supportive of restoration. Alteringland cover, reducing chemical inputs,carefully timed mowing, and othermanagement practices can reducedisturbance in the corridor.

Practical considerations may restrictrestoration to a zone of predefinedwidth adjacent to the stream. Althoughoften unavoidable, such restrictionstend to result in underrepresentation ofolder, off-channel environments thatsupport vegetation different from thatin stream-front communities. Restrict-ing restoration to a narrow part of thestream corridor usually does notrestore the full horizontal diversity ofbroad floodplains, nor does it fullyaccommodate functions that occurduring flood events, such as use of thefloodplain by aquatic species(Wharton et al. 1982).

In floodplains where extensive subsur-face hydrologic connections exist,limiting restoration to streamsidebuffer zones is not recommended sincesignificant amounts of energy, nutrienttransformation, and invertebrateactivities can occur at great distances

transitionalupland fringe

floodplain

open

naturalvegetationhills

lopematrix

streamchannel

Figure 8.3:Connections across astream corridor.A ladder pattern of naturalhabitat can restorestructure and functionswhere competing landuses prevail.Adapted from Ecology ofGreenways: Design andFunction of LinearConservation Areas .Edited by Smith andHellmund. ©University ofMinnesota Press 1993.



Seepage1. Sponge effect for hydrologic flows, mimimizing

downstream flooding2. Control of siddolved-substance inputs from matri

1st Order Stream1. Same as for seepage

2nd to 4th Order Stream with Closed Canopy1. Conduit for upland interior species; both sides of

stream so species readily crossing floodplain havealternate routes

2. Control of dissolved-substance inputs from matrix3. Conduit for streambank and floodplain species,

where beaver activities maintain water across thefloodplain and alter hillslope vegetation

4. Minimize hillslope erosion5. Sponge effect for hydrologic flows, minimizing

downstream flooding6. Friction effect, minimizing downstream sediment7. Protect high habitat diversity and species

richness of floodplain

2nd to 4th Order Stream with Open Canopy1. Same as for 2nd to ca. 4th order stream, closed ca2. Provide interior habitat for species conduit, as

migrating open stream intersects hillslopescausing them to be open habitat

5th to 10th Order River1. Conduit for upland interior species, on both sides

of river so species that rarely can cross thefloodplain have a route on each side

2. Provide interior habitat for species conduit, asmigrating open river intersects hillslopescausing them to be open habitat

3. minimize hillslope erosion4. Shade and logs provide fish habitat where

river is adjacent to hillslope5. Source of soil organic matter, an important base

of the river food chain6. Shade and logs provide fish habitat wherever

river is as it migrates across the floodplain7. Genetic benefit to upland species that can use

habitat continuity to infrequently cross floodplain8. Sponge effect for hydrologic flows, minimizing

downstream flooding9. Friction effect minimizing downstream sedimenta10.Protect high habitat diversity and species

richness of floodplain11.Conduit for semiaquatic and other organisms

dependent on river channel resources

interior portion of corridor in upland

meander band

matrix

interior of patch of natural floodplain vegetation

other ecologically-compatible land use

edge portion of corridor in upland

floodplainhillslope

edge of patch of natural floodplain vegetation

Corridor Width Variables

The minimum width of stream corridors based on ecologicalcriteria (Figure 8.4 ). Five basic situations in a river systemare identified, progressing from seepage to river. The keyvariables determining minimum corridor width are listed undereach.

from the stream channel outside thebuffer areas (Sedell et al. 1990).Similarly, failure to anticipate channelmigration or periodic beaver activitymight result in a corridor that does notaccommodate fundamental dynamicprocesses (Malanson 1993).

As previously discussed, restoration ofan ecologically effective streamcorridor requires consideration ofuplands adjacent to the channel andfloodplain. Hillslopes might be asource area for water maintainingfloodplain wetlands, a sediment sourcefor channels on bedrock, and theprincipal source of organic debris inhigh-gradient streams.

Despite these considerations, streamcorridors are often wrongly viewed asconsisting of only the channel and anadjacent vegetative buffer. The widthof the buffer is determined by specificobjectives such as control of agricul-tural runoff or habitat requirements ofparticular animal species. This narrowdefinition obviously does not fullyaccommodate the extent of the func-tions of a stream corridor; but wherethe corridor is limited by immovableresource uses, it often becomes a partof a restoration strategy.

Cognitive Approach: The ReferenceStream Corridor

Ideal stream corridor widths, aspreviously defined, are not alwaysachievable in the restoration design. Alocal reference stream corridor mightprovide dimensions for designing therestoration.

Examination of landscape patterns isbeneficial in identifying a referencestream corridor. The reference shouldprovide information about gap width,landform, species requirements,

Figure 8.4: Factors for determining minimum corridor widths.Stream corridor functions are directly influenced by corridor width.From Forman 1995. Reprinted with permission of Cambridge UniversityPress.



vegetative structure, and boundarycharacteristics of the stream corridor(Figure 8.5).

Restoration objectives determine thedesired levels of functions specifiedby the restoration design. If a nearbystream corridor in a similar landscapesetting and with similar land usevariables provides these functionsadequately, it can be used to indicatethe connectivity and width attributesthat should be part of the design.

Analytical Approach: FunctionalRequirements of a Target Species

The restoration plan objectives can beused to determine dimensions for thestream corridor restoration. If, forexample, a particular species requiresthat the corridor offer interior habitat,the corridor width is sized to providethe necessary habitat. The require-ments of the most sensitive speciestypically are used for optimum corri-dor dimensions. When these dimen-sions extend beyond the land baseavailable for restoration, managementof adjacent land uses becomes a toolfor making the corridor effectivelywider than the project parameters.

Optimum corridor dimensions can beachieved through collaboration withindividuals and organizations whohave management authority overadjacent lands. Dimensions includewidth of edge effect associated withboundaries of the corridor and patternvariations within the corridor, maxi-mum acceptable width of gaps withinthe corridor, and maximum number ofgaps per unit length of corridor.

Designing for Drainage andTopographyThe stream corridor is dependent oninteractions with the stream to sustainits character and functions (see Chap-ter 2). Therefore, to the extent feasible,the restoration process should includeblockage of artificial drainage systems,removal or setback of artificial levees,and restoration of natural patterns offloodplain topography, unless theseactions conflict with other social orenvironmental objectives (e.g., flood-ing or habitat).

Restoration of microrelief is particu-larly important where natural floodinghas been reduced or curtailed becausea topographically complex floodplainsupports a mosaic of plant communi-ties and ecosystem functions as aresult of differential ponding of rain-fall and interception of ground water.Microrelief restoration can be accom-plished by selective excavation ofhistoric features within the floodplainsuch as natural wetlands, levees,oxbows, and abandoned channels.Aerial photography and remotelysensed data, as well as observations inreference corridors, provide an indica-tion of the distribution and dimensionsof typical floodplain microrelieffeatures.

Figure 8.5: A maple in aNew Mexico floodplain.A rare occurrence of aremnant population mayreflect desired conditionsin a reference streamcorridor.



8.B Soil Properties

Stream corridor functions depend notonly on the connectivity and dimen-sions of the stream corridor, but alsoon its soils and associated vegetation.The variable nature of soils across andalong stream corridors results indiverse plant communities (Figure8.6). When designing stream corridorrestoration measures, it is important tocarefully analyze the soils and theirrelated potentials and limitations tosupport diverse native plant andanimal communities, as well as forrestoration involving channel recon-struction.

Where native floodplain soils remainin place, county soil surveys should beused to determine basic site conditionsand fertility and to verify that theproposed plant species to be restoredare appropriate. Most sites with fine-textured alluvium will not requiresupplemental fertilization, or fertiliz-ers might be required only for initialestablishment. In these cases excessivefertilization could encourage compet-ing weed species or exotics. Soilshould always be tested before makingany fertilizer design recommendations.

County soil surveys can provide basicinformation such as engineeringlimitations or suitabilities. Site-spe-cific soil samples should, however, becollected and tested when the restora-tion involves alternatives that includestream reconstruction.

The connections and feedback loopsbetween runoff and the structure andfunctions of streams are described inChapter 2. The functions of soil andthe connection between soil quality,

runoff, and water quality are alsoestablished in that chapter. Theseconnections need to be identified andconsidered in any stream corridorrestoration plan and design. For allland uses, emphasis needs to be placedon implementing conservation landtreatment that promotes soil qualityand the ability of the soils to carry outfour major functions:

• Regulating and partitioning theflow of water (a conduit andfilter function).

• Storing and cycling nutrientsand other chemicals (a sink andfilter function).

• Filtering, buffering, degrading,immobilizing, and detoxifyingorganic and inorganic materials(a filter, sink, and barrierfunction).

Figure 8.6: Distinctvegetation zonesalong a mountainstream.Variable soils result indiverse plantcommunities.



• Supporting biological activityin the landscape (a source andhabitat function).

References such as Field OfficeTechnical Guide (USDA-NRCS)contain guidance on the planning andselection of conservation practices andare available at most county offices.

CompactionSoils that have been in row crops orhave undergone heavy equipmenttraffic (such as that associated withconstruction) can develop a relativelyimpermeable compacted layer (plowpan or hard pan) that restricts watermovement and root penetration (Fig-ure 8.7). Such soils might requiredeep plowing, ripping, or vegetativepractices to break up the pan, althougheven these are sometimes ineffective.Deep plowing is usually expensiveand, at least in the East, should beused only if the planting of a speciesthat is able to penetrate the pan layer isnot a viable option.

Soil MicrofaunaOn new or disturbed substrates, or onrow-cropped sites, essential soilmicroorganisms (particularly mycor-rhizal fungi) might not exist. These aremost effectively replaced by usingrooted plant material that is inoculatedor naturally infected with appropriatefungi. Stockpiling and reincorporatinglocal topsoils into the substrate prior toplanting is also effective (Allen 1995).Particular care should be taken toavoid disturbing large trees or stumpssince the soils around and under themare likely source areas for reestablish-ment of a wide variety of microorgan-isms. Inoculation can be useful inrestoring some soil mycorrhizal fungifor particular species when naturallyinfected plant stock is unavailable.

Soil SalinitySoil salinity is another importantconsideration in restoration becausesalt accumulation in the soil canrestrict plant growth and the establish-ment of riparian species. High soilsalinity is not common in healthyriparian ecosystems where annualspring floods remove excess salts. Soilsalinity can also be altered by leachingsalts through the soil profile withirrigation (Anderson et al. 1984).Because of agricultural drainage andaltered flows due to dam construction,salt accumulation often contributes toriparian plant community declines.

Soil sampling throughout a restorationsite may be necessary since salinitycan vary across a floodplain, even onsites of less than 20 acres. If salinity isa problem, one must select plantmaterials adapted to a saline soilenvironment.

Figure 8.7: Compactionof streamside soil.Compact soils mayrequire deep plowing,ripping, or vegetativepractices to break up theimpermeable layer.



8.C Plant Communities

Vegetation is a fundamental control-ling factor in stream corridor function.Habitat, conduit, filter/barrier, source,and sink functions are all critically tiedto the vegetative biomass amount,quality, and condition (Figure 8.8).Restoration designs should protectexisting native vegetation and restorevegetative structure to result in acontiguous and connected streamcorridor.

Restoration goals can be general (e.g.,returning an area to a reference condi-tion) or specific (e.g., restoring habi-tats for particular species of interestsuch as the least Bell’s vireo, Vireobellii [Baird and Rieger 1988], oryellow-billed cuckoo, Coccyzusamericana [Anderson and Laymon1988]).

Numerous shrubs and trees have beenevaluated as restoration candidates,including willows (Svejcar et al. 1992,Hoag 1992, Conroy and Svejcar 1991,Anderson et al. 1978); alder, service-berry, oceanspray, and vine maple(Flessner et al. 1992); cottonwood andpoplar (Hoag 1992); Sitka and thinleafalder (Java and Everett 1992); paloverde and honey mesquite (Andersonet al. 1978); and many others. Selec-tion of vegetative species may bebased on the desire to provide habitatfor a particular species of interest. Thecurrent trend in restoration, however,is to apply a multispecies or ecosystemapproach.

Riparian Buffer Strips

Managers of riparian systems have longrecognized the importance of bufferstrips, for the following reasons(USACE 1991):

• Provide shade that reduces watertemperature.

• Cause deposition of (i.e., filter)sediments and other contami-nants.

• Reduce nutrient loads ofstreams.

• Stabilize streambanks withvegetation.

• Reduce erosion caused byuncontrolled runoff.

• Provide riparian wildlife habitat.

• Protect fish habitat.

• Maintain aquatic food webs.

• Provide a visually appealinggreenbelt.

• Provide recreational opportuni-ties.

Figure 8.8: Streamcorridor vegetation.Vegetation is afundamental controllingfactor in the functioning ofstream corridors.



Although the value of buffer strips iswell recognized, criteria for theirsizing are variable. In urban streamcorridors a wide forest buffer is anessential component of any protectionstrategy. Its primary value is to pro-vide physical protection for the streamchannel from future disturbance orencroachment. A network of buffersacts as the right-of-way for a streamand functions as an integral part of thestream ecosystem.

Often economic and legal consider-ations have taken precedence overecological factors. For Vermont,USACE (1991) suggests that narrowstrips (100 ft wide) may be adequate toprovide many of the functions listedabove. For breeding bird populationson Iowa streams, Stauffer and Best(1980) found that minimum stripwidths varied from 40 ft for cardinalsto 700 ft for scarlet tanagers, Americanredstarts, and rufous-sided towhees.In urban settings buffer sizing criteriamay be based on existing site controlsas well as economic, legal, and eco-logical factors. Practical performancecriteria for sizing and managing urbanbuffers are presented in the box De-signing Urban Stream Buffers.Clearly, no single recommendationwould be suitable for all cases.

Because floodplain/riparian habitatsare often small in area when comparedto surrounding uplands, meeting theminimum area needs of a species,guild, or community is especiallyimportant. Minimum area is theamount of habitat required to supportthe expected or appropriate use andcan vary greatly across species andseasons. For example, Skagen(USGS, Biological Resources Divi-

sion, Ft. Collins, Colorado; unpubl.data) found that, contrary to whatmight be considered conventionalwisdom, extensive stream corridors insoutheastern Arizona were not moreimportant to migrating birds thanisolated patches or oases of habitat. Infact, oases that were <2.5 miles longand <30 ft in width had more speciesand higher numbers of nonbreedingmigrants than did corridors. Skagenfound that the use of oases, as well ascorridors, is consistent with the ob-served patterns of long distancemigrants, where migration occursalong broad fronts rather than north-south corridors. Because small and/orisolated patches of habitat can be soimportant to migrants, riparian restora-tion efforts should not overlook theimportant opportunities they afford.

Existing Vegetation

Existing native vegetation should beretained to the extent feasible, asshould woody debris and stumps(Figure 8.9). In addition to providinghabitat and erosion and sedimentcontrol, these features provide seedsources and harbor a variety of micro-organisms, as described above. Oldfencerows, vegetated stumps and rockpiles in fields, and isolated shade treesin pastures should be retained throughrestoration design, as long as thedominant plant species are native orare unlikely to be competitors in amatrix of native vegetation (e.g., fruittrees).

Nonnative vegetation can preventestablishment of desirable nativespecies or become an unwantedpermanent component of streamcorridor vegetation. For example,



Designing Urban Stream Buffers

The ability of an urban stream buffer to realize its many benefits depends to a large degree on how well itis planned, designed, and maintained. Ten practical performance criteria are offered to govern how a bufferis to be sized, managed, and crossed. The key criteria include:

Criteria 1: minimum total buffer width.

Most local buffer criteria require that development be set back a fixed and uniform distance from the streamchannel. Nationally, urban stream buffers range from 20 to 200 ft in width from each side of the streamaccording to a survey of 36 local buffer programs, with a median of 100 ft (Schueler 1995). In general, aminimum base width of at least 100 feet is recommended to provide adequate stream protection.

Criteria 2: three-zone buffer system:

Effective urban stream buffers have three lateral zones— stream side, middle core, and outer zone. Eachzone performs a different function, and has a different width, vegetative target and management scheme.The stream side zone protects the physical and ecological integrity of the stream ecosystem. Thevegetative target is mature riparian forest that can provide shade, leaf litter, woody debris and erosionprotection to the stream. The middle zone extends from the outward boundary of the stream side zone,and varies in width, depending on stream order, the extent of the 100-yr floodplain, adjacent steep slopes,and protected wetland areas. Its key functions are to provide further distance between upland developmentand the stream. The vegetative target for this zone is also mature forest, but some clearing may be allowedfor storm water management, access, and recreational uses.

The outer zone is the buffer’s "buffer," an additional 25-ft setback from the outward edge of the middlezone to the nearest permanent structure. In most instances, it is a residential backyard. The vegetativetarget for the outer zone is usually turf or lawn, although the property owner is encouraged to plant treesand shrubs, and thus increase the total width of the buffer. Very few uses are restricted in this zone. Indeed,gardening, compost piles, yard wastes, and other common residential activities often will occur in the outerzone.

Criteria 3: predevelopment vegetative target.

The ultimate vegetative target for urban stream buffers should be specified as the predevelopment riparianplant community—usually mature forest. Notable exceptions include prairie streams of the midwest, orarroyos of the arid West, that may have a grass or shrub cover in the riparian zone. In general, thevegetative target should be based on the natural vegetative community present in the floodplain, asdetermined from reference riparian zones. Turfgrass is allowed for the outer zone of the buffer.

Criteria 4. buffer expansion and contraction.

Many communities require that the minimum width of the buffer be expanded under certain conditions.Specifically, the average width of the middle zone can be expanded to include:

• the full extent of the 100-yr floodplain;

• all undevelopable steep slopes (greater than 25%);

• steep slopes (5 to 25% slope, at four additional ft of slope per one percent increment of slope above5%); or

• any adjacent delineated wetlands or critical habitats.

Criteria 5: Buffer delineation. Three key decisions must be made when delineating the boundaries of abuffer. At what mapping scale will streams be defined? Where does the stream begin and the buffer end?And from what point should the inner edge of the buffer be measured? Clear and workable delineationcriteria should be developed.

Criteria 6. Buffer crossings. Major objectives for stream buffers are to maintain an unbroken corridor ofriparian forest and to allow for upstream and downstream fish passage in the stream network. From apractical standpoint, however, it is not always possible to try to meet these goals everywhere along thestream buffer network. Some provision must be made for linear forms of development that must cross the



stream or the buffer, such as roads, bridges, fairways, underground utilities, enclosed storm drains or outfallchannels.

Criteria 7. Storm water runoff. Buffers can be an important component of the storm water treatment systemat a development site. They cannot, however, treat all the storm water runoff generated within a watershed(generally, a buffer system can only treat runoff from less than 10% of the contributing watershed to thestream). Therefore, some kind of structural BMP must be installed to treat the quantity and quality of stormwater runoff from the remaining 90% of the watershed.

Criteria 8. Buffers during plan review and construction. The limits and uses of the stream buffer systemsshould be well defined during each stage of the development process—from initial plan review, throughconstruction.

Criteria 9: Buffer education and enforcement. The future integrity of a buffer system requires a strongeducation and enforcement program. Thus, it is important to make the buffer “visible” to the community, andto encourage greater buffer awareness and stewardship among adjacent residents. Several simple stepscan be taken to accomplish this.

• Mark the buffer boundaries with permanent signs that describe allowable uses

• Educate buffer owners about the benefits and uses of the buffer with pamphlets, stream walks andmeetings with homeowners associations

• Ensure that new owners are fully informed about buffer limits/uses when property is sold ortransferred.

• Engage residents in a buffer stewardship program that includes reforestation and backyard“bufferscaping” programs

• Conduct annual buffer walks to check on encroachment

Criteria 10. Buffer flexibility. In most regions of the country, a hundred-foot buffer will take about 5% of thetotal land area in any given watershed out of use or production. While this constitutes a relatively modestland reserve at the watershed scale, it can be a significant hardship for a landowner whose property isadjacent to a stream. Many communities are legitimately concerned that stream buffer requirements couldrepresent an uncompensated "taking" of private property. These concerns can be eliminated if acommunity incorporates several simple measures to ensure fairness and flexibility when administering itsbuffer program. As a general rule, the intent of the buffer program is to modify the location of developmentin relation to the stream but not its overall intensity. Some flexible measures in the buffer ordinance include:

• Maintaining buffers in private ownership

• Buffer averaging

• Density compensation

• Variances

• Conservation easements



kudzu will kill vegetation. Generally,forest species planted on agriculturalland will eventually shade out pasturegrasses and weeds, although someinitial control (disking, mowing,burning) might be required to ensuretree establishment.

Plant CommunityRestorationAn objective of stream corridor resto-ration work might be to restore naturalpatterns of plant community distribu-tion within the stream corridor. Nu-merous publications describe generaldistribution patterns for variousgeomorphic settings and flow condi-tions (e.g., Brinson et al. 1981,Wharton et al. 1982), and county soilsurveys generally describe nativevegetation for particular soils. Moredetailed and site-specific plant com-munity descriptions may be availablefrom state Natural Heritage programs,chapters of The Nature Conservancy,or other natural resources agencies andorganizations.

Examination of the reference streamcorridor, however, is often the bestway to develop information on plantcommunity composition and distribu-tion. Once reference plant communi-ties are defined, design can begin todetail the measures required to restorethose communities (Figure 8.10).Rarely is it feasible or desirable toattempt to plant the full complementof appropriate species on a particularsite. Rather, the more typical ap-proach is to plant the dominant speciesor those species unlikely to colonizethe site readily. For example, in thecomplex bottomland hardwood forestsof the Southeast, the usual focus is on

planting oaks. Oaks are heavy-seeded,are often shade-intolerant, and maynot be able to readily invade largeareas for generations unless they areintroduced in the initial planting plan,particularly if flooding has beenreduced or curtailed. It is assumedthat lighter-seeded and shade-tolerantspecies will invade the site at ratessufficient to ensure that the resultingforest is adequately diverse. Thisprocess can be accelerated by plantingcorridors of fast-growing species (e.g.,cottonwoods) across the restorationarea to promote seed dispersal.

Figure 8.9: Remnantvegetation and woodydebris along a stream.Attempts should be madeto preserve existingvegetation within thestream corridor.

Figure 8.10: A thrivingand diverse plantcommunity within astream corridor.Examination of referenceplant communities is oftenthe best way to developinformation on thecomposition anddistribution of plantcommunities at therestoration site.



In areas typically dominated by cot-tonwoods and willows, the emphasismight be to emulate natural patterns ofcolonization by planting groves ofparticular species rather than mixedstands, and by staggering the plantingprogram over a period of years toensure structural variation. Whereconifers tend to eventually succeedriparian hardwoods, some restorationdesigns may include scattered coniferplantings among blocks of pioneerspecies, to accelerate the transition toa conifer-dominated system.

Large-scale restoration work some-times includes planting of understoryspecies, particularly if they are re-quired to meet specific objectives suchas providing essential components ofendangered species habitat. However,it is often difficult to establish under-story species, which are typically nottolerant to full sun, if the restorationarea is open. Where particular under-story species are unlikely to establishthemselves for many years, they canbe introduced in adjacent forestedsites, or planted after the initial treeplantings have matured sufficiently to

create appropriate understory condi-tions. This may also be an appropriateapproach for introducing certainoverstory species that might notsurvive planting in full sun (Figure8.11)

The concept of focusing restorationactions on a limited group of overstoryspecies to the exclusion of understoryand other overstory species has beencriticized. The rationale for favoringspecies such as oaks has been toensure that restored riparian andfloodplain areas do not become domi-nated by opportunistic species, andthat wildlife functions and timbervalues associated with certain specieswill be present as soon as possible. Ithas been documented that heavy-seeded species such as oaks may beslow to invade a site unless planted(see Tennessee Valley Authority Flood-plain Reforestation Projects-50 YearsLater), but differential colonizationrates probably exclude a variety ofother species as well. Certainly, itwould be desirable to introduce aswide a variety of appropriate speciesas possible; however, costs and thedifficulties of doing supplementalplantings over a period of years mightpreclude this approach in most in-stances.

Plant species should be distributedwithin a restoration site with closeattention to microsite conditions. Inaddition, if stream meandering behav-ior or scouring flows have beencurtailed, special effort is required tomaintain communities that normallydepend on such behavior for naturalestablishment. These may includeoxbow and swale communities (baldcypress, shrub wetlands, emergent

Figure 8.11:Restoration ofunderstory plantspecies.Understory species canbe introduced at therestoration site after theinitial tree plantings havematured sufficiently.



wetlands), as well as communitiescharacteristic of newly deposited soils(cottonwoods, willows, alders, silvermaple, etc.). It is important to recog-nize that planting vegetation on siteswhere regeneration mechanisms nolonger operate is a temporary measure,and long-term management andperiodic replanting is required tomaintain those functions of the eco-system.

In the past, stream corridor plantingprograms often included nonnativespecies selected for their rapid growthrates, soil binding characteristics,ability to produce abundant fruits forwildlife, or other perceived advantagesover native species. These actionssometimes have unintended conse-quences and often prove to be ex-

tremely detrimental (Olson and Knopf1986). As a result, many local, county,state, and federal agencies discourageor prohibit planting of nonnativespecies within wetlands or streamsidebuffers. Stream corridor restorationdesigns should emphasize native plantspecies from local sources. It may befeasible in some cases to focus restora-tion actions on encouraging the suc-cess of local seedfall to ensure thatlocally adapted populations of streamcorridor vegetation are maintained onthe site (Friedmann et al. 1995).

Plant establishment techniques varygreatly depending on site conditionsand species characteristics. In aridregions, the emphasis has been onusing poles or cuttings of species thatsprout readily, and planting them todepths that will ensure contact withmoist soil during the dry season(Figure 8.12). Where water tableshave declined precipitously, deepauguring and temporary irrigation areused to establish cuttings and rooted orcontainer-grown plants. In environ-ments where precipitation or groundwater is adequate to sustain plantedvegetation, prolonged irrigation is lesscommon, and bare-root or container-

Streamcorridorrestorationdesignsshouldemphasizenative plantspeciesfrom localsources.

Low Water Availability

In areas where water levels are low,artificial plantings will not survive if theirroots cannot reach the zone ofsaturation. Low water availability wasassociated with low survival rates inmore than 80 percent of unsuccessfulrevegetation work examined in Arizona(Briggs 1992). Planting long poles (20ft) of Fremont cottonwood (Populusfremontii) and Gooding willow inaugered holes has been successfulwhere the ground water is more than 10ft below the surface (Swenson andMullins 1985). In combination with anirrigation system, many planted treesare able to reach ground water 10 ftbelow the surface when irrigated for twoseasons after planting (Carothers et al.1990). Sites closest to ground water,such as secondary channels,depressions, and low sites where watercollects, are the best candidates forplanting, although low-elevation sitesare more prone to flooding and flooddamage to the plantings. Additionally,the roots of many riparian species maybecome dormant or begin to die ifinundated for extended periods of time(Burrows and Carr 1969).

Figure 8.12:Revegetation with theuse of deeply plantedlive cuttings.In arid regions, poles orcuttings of species thatsprout readily are oftenplanted to depths thatassure contact with moistsoil.



Tennessee Valley Authority Floodplain Reforestation Projects—50 Years Later

The oldest known large-scale restoration of forestedwetlands in the United States was undertaken by theTennessee Valley Authority in conjunction withreservoir construction projects in the south during the1940s. Roads and railways were relocated outsidethe influence of maximum pool elevations, but wherethey were placed on embankments, TVA wasconcerned that they would be subject to wave erosionduring periods of extreme high water. To reduce thatpossibility, agricultural fields between the reservoir andthe embankments were planted with trees (Figure8.13). At Kentucky Reservoir in Kentucky andTennessee, approximately 1,000 acres were planted,mostly on hydric soils adjacent to tributaries of theTennessee River. Detailed records were keptregarding the species planted and survival rates.Some of these stands were recently located and studied to evaluate the effectiveness of the originalreforestation effort, and to determine the extent to which the planted forests have come to resemble naturalstands in the area.

Because the purpose of the plantings was erosion control, little thought was given to recreating naturalpatterns of plant community composition and structure. Trees were evenly spaced in rows, and plantedspecies were apparently chosen for maximum flood tolerance. As a result, the studied stands had aninitial composition dominated by bald cypress, green ash, red maple, and similarly water-tolerant species,but they did not originally contain many of the other common bottomland forest species, such as oaks.

Shear et al. (in press) compared the plant communities of the planted stands with forests on similar sitesthat had established by natural invasion of abandoned fields. They also looked at older stands that hadnever been converted to agriculture. The younger planted and natural stands were similar to the olderstands with regard to understory composition, and measures of stand density and biomass were consistentwith patterns typical for the age of the stands. Overstory composition of the planted stands was verydifferent from that of the others, reflecting the original plantings. However, both the planted sites and thefields that had been naturally invaded had few individuals of heavy-seeded species (oaks and hickories),which made up 37 percent of the basal area of the older stands.

Oaks are an important component of southernbottomlands and are regarded as particularlyimportant to wildlife. In most modern restorationplantings, oaks are favored on the assumption thatthey will not quickly invade agricultural fields. Thestands at Kentucky Reservoir demonstrate thatplanted bottomland forests can develop structural andunderstory conditions that resemble those of naturalstands within 50 years (Figure 8.14 ). Stands thatestablished by natural invasion of agricultural fieldshad similar characteristics. The major compositionaldeficiency in both of the younger stands was the lackof heavy-seeded species. The results of this studyappear to support the practice of favoring heavy-seeded species in bottomland forest restorationinitiatives.

Figure 8.13: Kentucky Reservoir watershed, 1943.Planting abandoned farmland with trees.

Figure 8.14: Kentucky Reservoir watershed in1991.Thriving bottomland hardwood forest.



grown plants are often used, particu-larly for species that do not sproutreliably from cuttings. On largefloodplains of the South and East,direct seeding of acorns and plantingof dormant bare-root material havebeen highly successful. Other options,such as transplanting of salvagedplants, have been tried with varyingdegrees of success. Local experienceshould be sought to determine themost reliable and efficient plantestablishment approaches for particu-lar areas and species, and to determinewhat problems to expect.

It is important to protect plantingsfrom livestock, beaver, deer, smallmammals, and insects during theestablishment period. Mortality ofvegetation from deer browsing iscommon and can be prevented byusing tree shelters to protect seedlings.

Horizontal DiversityStream corridor vegetation, as viewedfrom the air, would appear as a mosaicof diverse plant communities that runsfrom the upland on one side of thestream corridor, down the valley slope,across the floodplain and up theopposite slope to the upland. Withsuch broad dimensional range, there isa large potential for variation invegetation. Some of the variation is aresult of hydrology and stream dynam-ics, which will be discussed later inthis chapter. Three important struc-tural characteristics of horizontaldiversity of vegetation are connectiv-ity, gaps, and boundaries.

Connectivity and Gaps

As discussed earlier, connectivity is animportant evaluation parameter of

stream corridor functions, facilitatingthe processes of habitat, conduit, andfilter/barrier. Stream corridor restora-tion design should maximize connec-tions between ecosystem functions.Habitat and conduit functions can beenhanced by linking critical ecosys-tems to stream corridors throughdesign that emphasizes orientation andproximity. Designers should considerfunctional connections to existing orpotential features such as vacant orabandoned land, rare habitat, wetlandsor meadows, diverse or unique vegeta-tive communities, springs, ecologicallyinnovative residential areas, movementcorridors for flora and fauna, or associ-ated stream systems. This allows formovement of materials and energy,thus increasing conduit functions andeffectively increasing habitat throughgeographic proximity.

Generally, a long, wide stream corridorwith contiguous vegetative cover isfavored, though gaps are common-place. The most fragile ecologicalfunctions determine the acceptablenumber and size of gaps. Wide gapscan be barriers to migration of smallerterrestrial fauna and indigenous plantspecies. Aquatic fauna may also belimited by the frequency or dimensionof gaps. The width and frequency ofgaps should therefore be designed inresponse to planned stream corridorfunctions. Bridges have been designedto allow migration of animals, alongwith physical and chemical connec-tions of river and wetland flow. InFlorida, for example, underpasses areconstructed beneath roadways to serveas conduits for species movement(Smith and Hellmund 1993). TheNetherlands has experimented withextensive species overpasses and



underpasses to benefit particularspecies (Figure 8.15). Although nottypically equal to the magnitude of anundisturbed stream corridor lackinggaps, these measures allow for modestfunctions as habitat and conduit.

The filtering capacity of stream corri-dors is affected by connectivity andgaps. For example, nutrient and waterdischarge flowing overland in sheetflow tends to concentrate and formrills. These rills in turn often formgullies. Gaps in vegetation offer noopportunity to slow overland flow orallow for infiltration. Where referencedimensions are similar and transfer-able, restored plant communitiesshould be designed to exhibit struc-tural diversity and canopy closuresimilar to that of the reference streamcorridor. The reference stream corri-dor can provide information regardingplant species and their frequency anddistribution. Design should aim tomaintain the filtering capacity of thestream corridor by minimizing gaps inthe corridor’s width and length.

Buffer configuration and compositionhave also received attention since theyinfluence wildlife habitat quality,including suitability as migration

corridors for various species andsuitability for nesting habitat. Rees-tablishment of linkages among ele-ments of the landscape can be criti-cally important for many species(Noss 1983, Harris 1984). However,as noted previously, fundamentalconsiderations include whether aparticular vegetation type has everexisted as a contiguous corridor in anarea, and whether the predisturbancecorridor was narrow or part of anexpansive floodplain forest system.Establishment of inappropriate andnarrow corridors can have a net detri-mental influence at local and regionalscales (Knopf et al. 1988). Localwildlife management priorities shouldbe evaluated in developing bufferwidth criteria that address these issues.

Boundaries

The structure of the edge vegetationbetween a stream corridor and theadjacent landscape affects the habitat,conduit, and filter functions. A transi-tion between two ecosystems in anundisturbed environment typicallyoccurs across a broad area.

Boundaries between stream corridorsand adjacent landscapes may bestraight or curvilinear. A straightboundary allows relatively unimpededmovement along the edge, therebydecreasing species interaction betweenthe two ecosystems. Conversely, acurvilinear boundary with lobes of thecorridor and adjoining areas reachinginto one another encourages move-ment across boundaries, resulting inincreased interaction. The shape ofthe boundary can be designed tointegrate or discourage these interac-tions, thus affecting the habitat, con-duit, and filter functions.

bridgeroad

Figure 8.15: Underpassdesign.Underpasses should bedesigned toaccommodate bothvehicular traffic andmovement of small fauna.

Restoredplant com-munitiesshould bedesigned toexhibitstructuraldiversityand canopyclosuresimilar tothat of thereferencestreamcorridor.



Species interaction may or may not bedesirable depending on the projectgoals. The boundary of the restorationinitiative can, for example, be de-signed to capture seeds or to integrateanimals, including those carryingseeds. In some cases, however, thisinteraction is dictated by the functionalrequirements of the adjacent ecosys-tem (equipment tolerances within anagricultural field, for instance).

Vertical DiversityHeterogeneity within the streamcorridor is an important design consid-eration. The plants that make up thestream corridor, their form (herbs,shrubs, small trees, large trees), andtheir diversity affect function, espe-cially at the reach and site scales.Stratification of vegetation affectswind, shading, avian diversity, andplant growth (Forman 1995). Typi-cally, vegetation at the edge of thestream corridor is very different fromthe vegetation that occurs within theinterior of the corridor. The topogra-

phy, aspect, soil, and hydrology of thecorridor provide several naturallydiverse layers and types of vegetation.

The difference between edge andinterior vegetative structure are impor-tant design considerations (Figure8.16). An edge that gradually changesfrom the stream corridor into theadjacent ecosystems will softenenvironmental gradients and minimizeany associated disturbances. Thesetransitional zones encourage speciesdiversity and buffer variable nutrientand energy flows. Although humanintervention has made edges moreabrupt, the conditions of naturallyoccurring edge vegetation can berestored through design. The plantcommunity and landform of a restorededge should reflect the structuralvariations found in the referencestream corridor. To maintain a con-nected and contiguous vegetativecover at the edge of small gaps, tallervegetation should be designed tocontinue through the gap. If the gap iswider than can be breached by the

interior gradual edge

Figure 8.16: Edgevegetative structure.Edge characteristicscan be abrupt orgradual, with thegradual boundarytypically encouragingmore interactionbetween ecosystems.



tallest or widest vegetation, a moregradual edge may be appropriate.

Vertical structure of the corridorinterior tends to be less diverse thanthat of the edge. This is typicallyobserved when entering a woodlot:edge vegetation is shrubby and diffi-cult to traverse, whereas inner shadedconditions produce a more open forestfloor that allows for easier movement.Snags and downed wood may alsoprovide important habitat functions.When designing to restore interiorconditions of stream corridor vegeta-tion, a vegetation structure should beused that is less diverse than thevegetation structure used at the edge.The reference stream corridor willyield valuable information for thisaspect of design.

Influence of Hydrology andStream DynamicsNatural floodplain plant communitiesderive their characteristic horizontaldiversity primarily from the organiz-ing influence of stream migration andflooding (Brinson et al. 1981). Asdiscussed earlier, when designingrestoration of stream corridor vegeta-tion, nearby reference conditions aregenerally used as models to identifythe appropriate plant species andcommunities. However, the originalcover and older existing trees mighthave been established before streamregulation or other changes in thewatershed that affect flow and sedi-ment characteristics.

A good understanding of current andprojected flooding is necessary fordesign of appropriately restored plantcommunities within the floodplain.

Water management and planningagencies are often the best sources ofsuch data. In wildland areas, streamgauge data may be available, or on-site interpretation of landforms andvegetation may be required to deter-mine whether floodplain hydrologyhas been altered through channelincision, beaver activity, or othercauses. Discussions with local resi-dents and examination of aerial pho-tography may also provide informa-tion on water diversions, ground waterdepletion, and similar changes in thelocal hydrology.

A vegetation-hydroperiod model canbe used to forecast riparian vegetationdistribution (Malanson 1993). Themodel identifies the inundating dis-charges of various locations in theriparian zone and the resulting suit-ability of moisture conditions fordesired plants. Grading plans, forexample, can be adjusted to alter thearea inundated by a given dischargeand thus increase the area suitable forvegetation associated with a particularfrequency and duration of flooding. Afocus on the vegetation-hydroperiodrelationship will demonstrate thefollowing:

• The importance of moistureconditions in structuringvegetation of the riparian zone;

• The existence of reasonablywell accepted physical modelsfor calculating inundation fromstreamflow and the geometryof the bottomland.

• The likelihood that streamflowand inundating discharges havebeen altered in degradedstream systems or will be



modified as part of a restora-tion effort.

Generally, planting efforts will beeasier when trying to restore vegeta-tion on sites that have suitable mois-ture conditions for the desired vegeta-tion, such as in replacing historicalvegetation on cleared sites that haveunaltered streamflow and inundatingdischarges. Moisture suitabilitycalculations will support designs.Sometimes the restoration objective isto restore more of the desired vegeta-tion than the new flow conditionswould naturally support. Directmanipulation by planting and control-ling competition can often produce thedesired results within the physiologi-cal tolerances of the desired species.However, the vegetation on these siteswill be out of balance with the sitemoisture conditions and might requirecontinued maintenance. Managementof vegetation can also acceleratesuccession to a more desirable state.

Projects that require long-term supple-mental watering should be avoideddue to high maintenance costs anddecreased potential for success. In-versely, there may be cases where theabsence of vegetation, especiallywoody vegetation, is desired near thestream channel. Alteration of stream-flow or inundating discharges mightmake moisture conditions on thesesites unsuitable for woody vegetation.

The general concept of site suitabilityfor plant species can be extended frommoisture conditions determined byinundation to other variables determin-ing plant distribution. For example,Anderson (1996) suggests that restora-tion of native riparian vegetation inarid southwestern river systems may

be limited by unsuitable soil salinities.In many arid situations, depth toground water might be a more directmeasure of the moisture effects ofstreamflow on riparian sites thanactual inundation. Both inundatingdischarge and depth to ground waterare strongly related to elevation.However, depth to ground water maybe the more appropriate causal vari-able for these rarely inundated sites,and a physical model expressing thedependence of alluvial ground waterlevels on streamflow might thereforebe more important than a hydraulicmodel of surface water elevations.

Some stream corridor plant specieshave different requirements at differentlife stages. For example, plants toler-ating extended inundation as adultsmay require a drawdown for establish-ment, and plants thriving on relativelyhigh and dry sites as adults may beestablished only on moist surfaces nearthe water’s edge. This can complicatewhat constitutes suitable moistureconditions and may require separateconsideration of establishment require-ments, and perhaps consideration ofhow sites might change over time.The application of simulation modelsof plant dynamics based on solvingsets of explicit rules for how plantcomposition will change over timemay become necessary as increasinglycomplex details of different require-ments at different plant life historystages are incorporated into the evalua-tion of site suitability. Examples ofthis type of more sophisticated plantresponse model include Van der Valk(1981) for prairie marsh species andPearlstine et al. (1985) for bottomlandhardwood tree species.



Soil bioengineering forfloodplains and uplandsSoil bioengineering is the use of liveand dead plant materials, in combina-tion with natural and synthetic supportmaterials, for slope stabilization,erosion reduction, and vegetativeestablishment.

There are many soil bioengineeringsystems, and selection of the appropri-ate system or systems is critical tosuccessful restoration. Referencedocuments should be consulted toensure that the principles of soilbioengineering are understood andapplied. The NRCS National Engi-neering Handbook, Part 650 (Chapter16, Streambank and Shoreline Protec-tion and Chapter 18, Soil Bioengineer-ing for Upland Slope Protection andErosion Reduction) offers backgroundand guidelines for application of thistechnology. A more detailed descrip-tion of soil bioengineering systems isoffered in Section 8.F, StreambankStabilization Design, of this chapterand in the Appendix A.

REVERSE REVERSE FAST FORWARD

See Chapter 8,Section F for moreinformation on soilbioengineeringtechniques.



8.D. Habitat Measures

Other measures may be used to pro-vide structure and functions. Theymay be implemented as separateactions or as an integral part of therestoration plan to improve habitat, ingeneral, or for specific species. Suchmeasures can provide short-termhabitat until overall restoration resultsreach the level of maturity needed toprovide the desired habitat. Thesemeasures can also provide habitat thatis in short supply. Greentree reser-voirs, nest structures, and food patchesare three examples. Beaver are alsopresented as a restoration measure.

Greentree ReservoirsShort-term flooding of bottomlandhardwoods during the dormant periodof tree growth enhances conditions forsome species (e.g., waterfowl) to feedon mast and other understory foodplants, like wild millet and smartweed.Acorns are a primary food source instream corridors for a variety of fauna,including ducks, nongame birds andmammals, turkey, squirrel, and deer.Greentree reservoirs are shallow,forested floodplain impoundmentsusually created by building low leveesand installing outlet structures (Figure8.17). They are usually flooded inearly fall and drained during lateMarch to mid-April. Draining pre-vents damage to overstory hardwoods(Rudolph and Hunter 1964). Mostexisting greentree reservoirs are in theSouthwest.

The flooding of greentree reservoirs,by design, differs from the natural

flood regime. Greentree reservoirs aretypically flooded earlier and at depthsgreater than would normally occurunder natural conditions. Over time,modifications of natural flood condi-tions can result in vegetation changes,lack of regeneration, decreased mastproduction, tree mortality, and disease.Proper management of green treereservoirs requires knowledge of thelocal system—especially the naturalflood regime—and the integration ofmanagement goals that are consistentwith system requirements. Propermanagement of greentree reservoirscan provide quality habitat on anannual basis, but the management planmust be well designed from construc-tion through management for water-fowl.

Nest StructuresLoss of riparian or terrestrial habitat instream corridors has resulted in thedecline of many species of birds and

Figure 8.17:Bottomland hardwoodsserving as a greentreereservoir.Proper management ofgreentree reservoirsrequires knowledge of thelocal system.



mammals that use associated trees andtree cavities for nesting or roosting.The most important limiting factor forcavity-nesting birds is usually theavailability of nesting substrate (vonHaartman 1957), generally in the formof snags or dead limbs in live trees(Sedgwick and Knopf 1986). Snagsfor nest structures can be created usingexplosives, girdling, or topping oftrees. Artificial nest structures cancompensate for a lack of natural sitesin otherwise suitable habitat sincemany species of birds will readily usenest boxes or other artificial structures.For example, along the MississippiRiver in Illinois and Wisconsin, wherenest trees have become scarce, artifi-cial nest structures have been erectedand constructed for double-crestedcormorants using utility poles(Yoakum et al. 1980). In many cases,increases in breeding bird density haveresulted from providing such struc-tures (Strange et al. 1971, Brush1983). Artificial nest structures canalso improve nestling survival (Cowan1959).

Nest structures must be properlydesigned and placed, meeting thebiological needs of the target species.They should also be durable, predator-proof, and economical to build. De-sign specifications for nest boxesinclude hole diameter and shape,internal box volume, distance from thefloor of the box to the opening, type ofmaterial used, whether an internal“ladder” is necessary, height of place-ment, and habitat type in which toplace the box. Other types of neststructures include nest platforms forwaterfowl and raptors; nest baskets fordoves, owls, and waterfowl; floating

nest structures for geese; and tire nestsfor squirrels. Specifications for neststructures for riparian and wetlandnesting species (including numerousPicids, passerines, waterfowl, andraptors) can be found in many sourcesincluding Yoakum et al. (1980),Kalmbach et al. (1969), Payne (1992),and various state wildlife agency andconservation publications.

Food PatchesFood patch planting is often expensiveand not always predictable, but it canbe carried out in wetlands or ripariansystems mostly for the benefit ofwaterfowl. Environmental require-ments of the food plants native to thearea, proper time of year of introduc-tion, management of water levels, andsoil types must all be taken intoconsideration. Some of the moreimportant food plants in wetlandsinclude pondweed (Potamogetonspp.), smartweed (Polygonum spp.),duck potato, spike sedges (Carexspp.), duckweeds (Lemna spp.),coontail, alkali bulrush (Scirpuspaludosus), and various grasses. Twocommonly planted native speciesinclude wild rice (Zizania) and wildmillet. Details on suggested tech-niques for planting these species canbe found in Yoakum et al. (1980).



Importance of Beaver to Riparian Ecosystems

Beaver have long been recognized for their potential to influence riparian systems. In rangelands, whereloss of riparian functional value has been most dramatic, the potential role of beaver in restoring degradedstreams is least understood.

Beaver dams on headwater streams can positively influenceriparian function in many ways, as summarized by Olson andHubert (1994) (Figure 8.18 ). They improve water quality bytrapping sediments behind dams and by reducing streamvelocity, thereby reducing bank erosion (Parker 1986).Beaver ponds can alter water chemistry by changingadsorption rates for nitrogen and phosphorus (Maret 1985)and by trapping coliform bacteria (Skinner et al. 1984). Theflow regime within a watershed can also be influenced bybeaver. Beaver ponds create a sponge-like effect byincreasing the area where soil and water meet (Figure 8.19 ).Headwaters retain more water from spring runoff and majorstorm events, which is released more slowly, resulting in ahigher water table and extended summer flows. Thisincrease in water availability, both surface and subsurface,usually increases the width of the riparian zone and,consequently, favors wildlife communities that depend on thatvegetation. There can be negative impacts as well, includingloss of spawning habitat, increase in water temperaturesbeyond optimal levels for some fish species, and loss ofriparian habitat.

Richness, diversity, and abundance of birds, herpetiles, andmammals can be increased by the activities of beaver (Bakeret al. 1992; Medin and Clary 1990). Beaver ponds areimportant waterfowl production areas and can also be usedduring migration (Call 1970, Ringelman 1991). In some high-elevation areas of the Rocky Mountains, beaver are solelyresponsible for the majority of local duck production. Inaddition, species of high interest, such as trumpeter swans,sandhill cranes, moose, mink, and river otters, use beaverponds for nesting or feeding areas (Collins 1976).

Transplanting Beaver to Restore StreamFunctions

Beaver have been successfully transplanted into manywatersheds throughout the United States during the past 50years. This practice was very common during the 1950s afterbiologists realized the loss of ecological function resultingfrom overtrapping of beaver by fur traders before the turn ofthe century. Reintroduction of beaver has restored the U.S.beaver population to 6-12 million, compared to a pre-European level of 60-400 million (Naiman et al. 1986). Muchunoccupied habitat or potential habitat still remains,especially in the shrub-steppe ecosystem.

Figure 8.18: Beaver dam on aheadwater stream.Beavers have many positiveimpacts on headwater streams.

Figure 8.19: A beaver pond.Beaver ponds create a sponge-like effect.

Figure 8.20: Beaver habitat.It is advisable to obtain beaver fromhabitat that is similar to where they willbe introduced.



In forested areas, where good beaver habitat already exists, reintroduction techniques are well established.The first question asked should be “If the habitat is suitable, why are beaver absent?” In the case of newlyrestored habitat or areas far from existing populations, reintroduction without habitat improvement might bewarranted (Figure 8.20 ). Beavers are livetrapped from areas that have excess populations or from areaswhere they are a nuisance. It is advisable to obtain beavers from habitat that is similar to where they willbe introduced to ensure they are familiar with available food and building materials (Smith and Prichard1992). This is particularly important in shrub-steppe habitats.

Reintroduction into degraded riparian areas within the shrub-steppe zone is controversial. Conventionalwisdom holds that a yearlong food supply must be present before introducing beaver. In colder climates,this means plants with edible bark, such as willow, cottonwood, or aspen, must be present to provide awinter food supply for beaver (Figure 8.21 ). But often these species are the goal of restoration. In somecases willows or other species can be successfully planted as described in other sections of this document.In other areas, conditions needed to sustain planted cuttings, such as a high water table and minimalcompetition with other vegetation, might preclude successful establishment. Transplanting beaver beforewillows are established may create the conditions needed to both establish and maintain riparian shrubs ortrees. In these cases it may be helpful to provide beaver with a pickup truck load of aspen or other trees touse as building material at or near the reintroduction site. This may encourage beaver to stay near the siteand strengthen dams built of sagebrush or other shrubs (Apple et al. 1985).

Nuisance Beaver

Unfortunately, beaver are not beneficial in all situations, which is all too obvious to those managing damagecontrol. In many cases where they live in close proximity to humans or features important to humans,beaver need to be removed or their damage controlled. Common problems include cutting or eatingdesirable vegetation, flooding roads or irrigation ditches by plugging culverts, and increasing erosion byburrowing into the banks of streams or reservoirs. In addition, beaver carry Giardia species pathogens,which can infect drinking water supplies and cause human health problems.

Control of nuisance beaver usually involves removing the problem animals directly or modifying theirhabitat. Beaver can be livetrapped (Bailey or Hancock traps) and relocated to a more acceptable locationor killed by dead-traps (e.g., Conibear #330) or shooting (Miller 1983). In cases where the water level in adam must be controlled to prevent flooding, a pipe can be placed through the dam with the upstream sideperforated to allow water flow.

beaver dam

winterfood storage

air vent

tunnel entrance

livingchamber

Figure 8.21: A beaverlodge.The living chamber in abeaver lodge is abovewater and used year-round. Deep entrancesenable beavers toobtain food fromunderwater caches inwinter.



8.E Stream Channel Restoration

Some disturbances to stream channels(e.g., from surface mining activities,extreme weather events, or majorhighway construction) are so severethat restoration within a desired timeframe requires total reconstruction of anew channel. Selecting dimensions(width, depth, cross-sectional shape,pattern, slope, and alignment) for sucha reconstructed channel is perhaps themost difficult component of streamrestoration design. In the case ofstream channel reconstruction, streamcorridor restoration design can proceedalong one of two broad tracks:

1. A single-species restorationthat focuses on habitat require-ments of certain life stages ofspecies (for example, rainbowtrout spawning). The existingsystem is analyzed in light ofwhat is needed to provide agiven quantity of acceptablehabitat for the target speciesand life stage, and designproceeds to remedy any defi-ciencies noted.

2. An “ecosystem restoration” or“ecosystem management”approach that focuses designresources on the chemical,hydrologic, and geomorphicfunctions of the stream corri-dor. This approach assumesthat communities will recoverto a sustainable level if thestream corridor structure andfunctions are adequate. Thestrength of this approach is thatit recognizes the complexinterdependence between

living things and the totality oftheir environments.

Although methods for single-speciesrestoration design pertaining to treat-ments for aquatic habitat are includedelsewhere in this chapter, the secondtrack is emphasized in this section.

Procedures for ChannelReconstructionIf watershed land use changes or otherfactors have caused changes in sedi-ment yield or hydrology, restoration toan historic channel condition is notrecommended. In such cases, a newchannel design is needed. The follow-ing procedures are suggested:

1. Describe physical aspects ofthe watershed and characterizeits hydrologic response.

This step should be based ondata collected during theplanning phase, as described inChapter 4.

2. Considering reach and associ-ated constraints, select apreliminary right-of-way forthe restored stream channelcorridor and compute thevalley length and valley slope.

3. Determine the approximate bedmaterial size distribution forthe new channel.

Many of the channel design proce-dures described below require thedesigner to supply the size of bedsediments. If the project is not likelyto modify bed sediments, the existingchannel bed may be sampled using

Review Chap 4's datacollection planningsection.

REVERSE FAST FORWARDFAST FORWARD



procedures reviewed in Chapter 7. Ifpredisturbance conditions were differ-ent from those of the existing channel,and if those conditions must be re-stored, the associated sediment sizedistribution must be determined. Thiscan be done by collecting representa-tive samples of bed sediments fromnearby, similar streams; by excavatingto locate the predisturbance bed; or byobtaining the information from his-toric resources.

Like velocity and depth, bed sedimentsize in natural streams varies continu-ously in time and space. Particularlytroublesome are streams with sedimentsize distributions that are bimodalmixtures of sand and gravel, forexample. The median (D

50) of the

overall distribution might be virtuallyabsent from the bed. However, if flowconditions allow development of awell-defined armor layer, it might beappropriate to use a higher percentilethan the median (e.g., the D

75 ) to

represent the bed material size distri-bution. In some cases, a new channelexcavated into a heterogeneous mix-ture of noncohesive material willdevelop an armor layer. In such acase, the designer must predict thelikely size of the armor layer material.Methods presented by Helwig (1987)and Griffiths (1981) could provehelpful in such a situation.

4. Conduct a hydrologic andhydraulic analysis to select adesign discharge or range ofdischarges.

Conventional channel design hasrevolved around selecting channeldimensions that convey a certaindischarge at or below a certain eleva-tion. Design discharge is usually

based on flood frequency or durationor, in the case of canals, on down-stream supply needs. Channel restora-tion, on the other hand, implies de-signing a channel similar to one thatwould develop naturally under similarwatershed conditions.

Therefore, the first step in selecting adesign discharge for restoration is notto determine the controlling elevationfor flood protection but to determinewhat discharge controls channel size.Often this will be at or close to the 1-to 3-year recurrence interval flow. SeeChapters 1 and 7 for discussions ofchannel-forming, effective, and designdischarges. Additional guidanceregarding streamflow analysis forgauged and ungauged sites is pre-sented in Chapter 7. The designershould, as appropriate to the streamsystem, compute effective discharge orestimate bankfull discharge.

A sediment rating curve must bedeveloped to integrate with the flowduration curve to determine the effec-tive discharge. The sediment loadthat is responsible for shaping thechannel (bed material load) should beused in the calculation of the effectivedischarge. This sediment load can bedetermined from measured data orcomputed using an appropriate sedi-ment transport equation. If measuredsuspended sediment data are used, thewash load, typically consisting ofparticles less than 0.062 mm, shouldbe deleted and only the suspended bedmaterial portion of the suspended loadused. If the bed load in the stream isconsidered to be only a small percent-age of the total bed material load, itmight be acceptable to simply use themeasured suspended bed material loadin the effective discharge calculations.

Review Chap 1 andChap 7's channel-forming, effective, anddesign dischargessections.




However, if the bed load is a signifi-cant portion of the load, it should becalculated using an appropriate sedi-ment transport function and thenadded to the suspended bed materialload to provide an estimate of the totalbed material load. If bed load mea-surements are available, which seldomis the case, these observed data can beused.

Flow levels and frequencies that causeflooding also need to be identified tohelp plan and design out-of-streamrestoration measures in the rest of thestream corridor. If flood managementis a constraint, additional factors thatare beyond the scope of this documententer the design. Environmentalfeatures for flood control channels aredescribed elsewhere (Hey 1995,Shields and Aziz 1992, USACE1989a, Brookes 1988).

Channel reconstruction and streamcorridor restoration are most difficultfor incised streams, and hydrologicanalyses must consider several addi-tional factors. Incised stream chan-nels are typically much larger thanrequired to convey the channel-forming discharge. Restoration of anincised channel may involve raisingthe bottom of a stream to restoreoverbank flow and ecological func-tions of the floodplain. In this type ofrestoration, compatibility of restoredfloodplain hydrology with existingland uses must be considered.

A second option in reconstructingincised channels is to excavate one orboth sides to create a new bankfullchannel with a floodplain (Hey 1995).Again, adjacent land uses must be ableto accommodate the new, excavatedfloodplain/channel.

A third option is to stabilize the incisedchannel in place, and to enhance thelow-flow channel for environmentalbenefits. The creation of a floodplainmight not be necessary or possible aspart of a stream restoration.

In cases where channel sizing, modifi-cation, or realignment are necessary, orwhere structures are required to en-hance vertical or lateral stability, it iscritical that restoration design alsoinclude consideration of the range offlows expected in the future. Inurbanizing watersheds, future condi-tions may be quite different fromexisting conditions, with higher,sharper, peak flows.

If certain instream flow levels arerequired to meet restoration objectives,it is imperative that those flows bequantified on the basis of a thoroughunderstanding of present and desiredconditions. Good design practice alsorequires checking stream channelhydraulics and stability at dischargeswell above and below the designcondition. Stability checks (describedbelow) may be quite simple or verysophisticated. Additional guidanceon hydrologic analysis and develop-ment of stage-discharge relationshipsare presented in Chapter 7.

5. Predict stable planform type(straight, meandering, orbraided).

Channel planform may be classified asstraight, braided, or meandering, butthresholds between categories arearbitrary since channel form can varycontinuously from straight to single-channel meanders to multiple braids.Naturally straight, stable alluvialchannels are rare, but meandering andbraided channels are common and can

Review Chap 7'shydrologic analysisand stage-dischargerelationships sections.




display a wide range of lateral andvertical stability.

Relationships have been proposed thatallow prediction of channel planformbased on channel slope, discharge, andbed material size (e.g., Chang 1988),but they are sometimes unreliable(Chitale 1973, Richards 1982) andgive widely varying estimates of theslope threshold between meanderingand braiding. As noted by Dunne(1988), “The planform aspects ofrivers are the most difficult to predict,”a sentiment echoed by USACE (1994),“... available analytical techniquescannot determine reliably whether agiven channel modification will beliable to meander development, whichis sensitive to difficult-to-quantifyfactors like bank vegetation andcohesion.”

Stable channel bed slope is influencedby a number of factors, includingsediment load and bank resistance toerosion. For the first iteration, restora-tion designers may assume a channelplanform similar to stable referencechannels in similar watersheds. Bycollecting data for stable channels andtheir valleys in reference reaches,insight can be gained on what thestable configuration would be for therestoration area. The morphology ofthose stream types can also provideguidance or additional converginglines of evidence that the planformselected by the designer is appropriate.

After initial completion of these fivesteps, any one of several differentpaths may be taken to final design.Three approaches are summarized inTable 8.1. The tasks are not alwaysexecuted sequentially because trial anderror and reiteration are often needed.

Alignment and AverageSlopeIn some cases, it might be desirable todivert a straightened stream into ameandering alignment for restorationpurposes. Five approaches for mean-der design are summarized in theadjacent box.

For cases where the design channelwill carry only a small amount of bedmaterial load, bed slope and channeldimensions may be selected to carrythe design discharge at a velocity thatwill be great enough to prevent sus-pended sediment deposition and smallenough to prevent erosion of the bed.This approach is suitable only forchannels with beds that are stationaryor move very infrequently—typicallystable cobble- and gravel-bed streams.

Once mean channel slope is known,channel length can be computed bymultiplying the straight line down-valley distance by the ratio of valleyslope to channel slope (sinuosity).Meanders can then be laid out using apiece of string on a map or an equiva-lent procedure, such that the meanderarc length L (the distance betweeninflection points, measured along thechannel) ranges from 4 to 9 channelwidths and averages 7 channel widths.Meanders should not be uniform.

The incised, straightened channel ofthe River Blackwater (Norfolk, UnitedKingdom) was restored to a meander-ing form by excavating a new low-level floodplain about 50 to 65 feetwide containing a sinuous channelabout 16 feet wide and 3 feet deep(Hey 1995). Preliminary calculationsindicated that the bed of the channelwas only slightly mobile at bankfulldischarge, and sediment loads were



Table 8.1

Approach A Approach B (Hey 1994) Approach C (Fogg 1995)

Task Tools Task Tools Task Tools

Determine meander geometry and channel alignment1.

Empirical formulas for meander wavelength, and adaptation of measurements from predisturbed conditions or nearly undisturbed reaches.

Determine bed material discharge to be carried by design channel at design discharge, compute bed material sediment concentration.

Analyze measured data or use appropriate sediment transport function2 and hydraulic properties of reach upstream from design reach.

Compute mean flow, width, depth, and slope at design discharge.4

Regime or hydraulic geometry formulas with regional coefficients.

Compute sinuosity, channel length, and slope.

Channel length = sinuosity X valley length. Channel slope= valley slope/ sinuosity.

Compute mean flow, width, depth, and slope at design discharge.4

Regime or hydraulic geometry formulas with regional coefficients, or analytical methods (e.g. White, et.al., 1982, or Copeland, 1994)3.

Compute or estimate flow resistance coefficient at design discharge.

Appropriate relationship between depth, bed sediment size, and resistance coefficient, modified based on expected sinuosity and bank/berm vegetation.

Compute mean flow width and depth at design discharge.4

Regime or hydraulic geometry formulas with regional coefficients, and resistance equations or analytical methods (e.g. tractive stress, Ikeda and Izumi, 1990, or Chang, 1988).

Compute sinuosity and channel length.

Sinuosity = valley slope/ channel slope.Channel length= sinuosity X valley length.

Compute mean channel slope and depth required to pass design discharge.

Uniform flow equation (e.g. Manning, Chezy) continuity equation, and design channel cross-sectional shape; numerical water surface profile models may be used instead of uniform flow equation.

Compute riffle spacing (if gravel bed), and add detail to design.

Empirical formulas, observation of similar streams, habitat criteria.

Determine meander geometry and channel alignment.

Lay out a piece of string scaled to channel length on a map (or equivalent procedure) such that meander arc lengths vary from 4 to 9 channel widths.

Compute velocity or boundary sheer stress at design discharge.

Allowable velocity or shear stress criteria based on channel boundary materials

Check channel stability and reiterate as needed.

Check stability. Compute riffle spacing (if gravel bed), and add detail to design.

Empirical formulas, observation of similar streams, habitat criteria.

Compute sinuosity and channel length.

Sinuosity = valley slope/ channel slope.Channel length= sinuosity X valley length.


Check stability. Compute sinuosity and channel length.

Lay out a piece of string scaled to channel length on a map (or equivalent procedure) such that meander arc lengths vary from 4 to 9 channel widths.


Check stability.

1 Assumes meandering planform would be stable. Sinuosity and arc-length are known.2 Computation of sediment transport without calibration against measured data may give highly unreliable results for a specific channel

(USACE, 1994, Kuhnle, et.al., 1989).3 The two methods listed assume a straight channel. Adjustments would be needed to allow for effects of bends.4 Mean flow width and depth at design discharge will give channel dimensions since design discharge is bankfull. In some situation channel may be increased to

allow for freeboard. Regime and hydraulic geometry formulas should be examined to determine if they are mean width or top width.

A systematic design methodology has been developed for use in designing restoration projects that involvechannel reconstruction (USACE, WES). The methodology includes use of hydraulic geometry relationships, analyticaldetermination of stable channel dimensions, and a sediment impact assessment. The preferred geometry is acompound channel with a primary channel designed to carry the effective or “channel forming” discharge and anoverbank area designed to carry the additional flow for a specified flood discharge. Channel width may be determinedby analogy methods, hydraulic geometry predictors, or analytically. Currently under development are hydraulicgeometry predictors for various stream types. Once a width is determined for the effective discharge, depth andchannel slope are determined analytically by balancing sediment inflow from upstream with sediment transport capacitythrough the restored channel. Meander wavelength is determined by analogy or hydraulic geometry relationships.Assumption of a sine-generated curve then allows calculation of channel planform. The stability of the channel designis then evaluated for the full range of expected discharges by conducting a sediment impact assessment. Refinementsto the design include variation of channel widths at crossings and pools, variable lateral depths in pools, coarsening ofthe channel bed in riffles, and bank protection.

Table 8.1: Three approaches to achieving final design.There are variations of the final steps to a restoration design, after the first five steps described in the text are done.



low. A carbon copy design processwas used, recreating meander geom-etry from the mid-19th century (Hey1994). The River Neath (Wales,United Kingdom), an active gravel-bed stream, was diverted at fivelocations into meandering alignmentsto allow highway construction. Exist-ing slopes were maintained througheach diversion, effectively illustratinga “slope-first” design (Hey 1994).

Channel DimensionsSelection of channel dimensionsinvolves determining average valuesfor width and depth. These determina-tions are based on the imposed waterand sediment discharge, bed sedimentsize, bank vegetation, resistance, andaverage bed slope. However, bothwidth and depth may be constrainedby site factors, which the designermust consider once stability criteriaare met. Channel width must be lessthan the available corridor width,while depth is dependent on theupstream and downstream controllingelevations, resistance, and the eleva-tion of the adjacent ground surface. Insome cases, levees or floodwalls mightbe needed to match site constraintsand depth requirements. Averagedimensions determined in this stepshould not be applied uniformly.Instead, in the detailed design stepdescribed below, nonuniform slopesand cross sections should be specifiedto create converging and divergingflow and resulting physical diversity.

The average cross-sectional shape ofnatural channels is dependent ondischarge, sediment inflow, geology,roughness, bed slope, bank vegetation,and bed and bank materials. Although

bank vegetation is considered whenusing some of the empirical toolspresented below, many of the analyti-cal approaches do not consider theinfluence of bank material and vegeta-tion or make unrealistic assumptions(e.g., banks are composed of the samematerial as the bed). These toolsshould be used with care. After initialselection of average channel width anddepth, designers should consider thecompatibility of these dimensions withreference reaches.

Reference Reaches

Perhaps the simplest approach toselecting channel width and depth is touse dimensions from stable reacheselsewhere in the watershed or fromsimilar reaches in the region. Thedifficulty in this approach is finding asuitable reference reach. A referencereach is a reach of stream outside theproject reach that is used to developdesign criteria for the project reach.

A reference reach used for stablechannel design should be evaluated tomake sure that it is stable and has adesirable morphological and ecologi-cal condition. In addition, the refer-ence reach must be similar enough tothe desired project reach so that thecomparison is valid. It must be similarto the desired project reach in hydrol-ogy, sediment load, and bed and bankmaterial.

The term reference reach has severalmeanings. As used above, the refer-ence reach is a reach that will be usedas a template for the geometry of therestored channel. The width, depth,slope, and planform characteristics ofthe reference reach are transferred tothe design reach, either exactly or by



Meander Design

Five approaches to meander design are described below, not in any intended order of priority. The first fourapproaches result in average channel slope being determined by meander geometry. These approachesare based on the assumption that the controlling factors in the stream channel (water and sediment inputs,bed material gradation, and bank erosional resistance) will be similar to those in the reference reach (eitherthe restoration reach before disturbance or undisturbed reaches). The fifth approach requiresdetermination of stream channel slope first. Sinuosity follows as the ratio of channel slope to valley slope,and meander geometry (Figure 8.22 ) is developed to obtain the desired sinuosity.

1. Replacement of meanders exactly as found before disturbance (the carbon copy technique).This method is appropriate if hydrology and bed materials are very similar or identical topredisturbance conditions. Old channels are often filled with cohesive soils and may have cohesiveboundaries. Accordingly, channel stability may be enhanced by following a previous channelalignment.

2. Use of empirical relationships that allow computation of meander wavelength, L, and amplitudebased on channel width or discharge. Chang (1988) presents graphical and algebraic relationshipsbetween meander wavelength, width-depth ratio, and friction factor. In addition to meanderwavelength, specification of channel alignment requires meander radius of curvature (Hey 1976) andmeander amplitude or channel slope. Hey (1976) also suggests that L is not usually uniquelydetermined by channel width or discharge. Rechard and Schaefer (1984) provide an example ofdevelopment of regional formulas for meander restoration design. Chapter 7 includes a number ofmeander geometry relationships developed from regional data sets. Newbury and Gaboury (1993)designed meanders for a straightened stream (North Pine River) by selecting meander amplitude to fitbetween floodplain terraces. Meander wavelength was set at 12.4 times the channel width (on thehigh end of the literature range), and radius of curvature ranged from 1.9 to 2.3 times the channelwidth.

L

L meander wavelengthML meander arc lengthw average width at bankfull dischargeML meander amplituderc radius of curvature

arc angle

w

rc MA

ML

Figure 8.22: Variablesused to describe anddesign meanders.Consistent, clearterminology is used inmeander design.Adapted from Williams1986.



Figure 8.23: The naturalmeander of a stream.Rivers meander toincrease length andreduce gradient. Streamrestorations often attemptto reconstruct the channelto a previous meanderingcondition or one "copied"from a reference reach.

3. Basin-wide analysis to determine fundamental wavelength, mean radius of curvature, andmeander belt width in areas “reasonably free of geologic control.” This approach has been usedfor reconstruction of streams destroyed by surface mining in subhumid watersheds of the westernUnited States. Fourier analysis may be used with data digitized from maps to determine fundamentalmeander wavelength (Hasfurther 1985).

4. Use of undisturbed reaches as design models . If the reach targeted for restoration is closelybounded by undisturbed meanders, dimensions of these undisturbed reaches may be studied for usein the restored reach (Figure 8.23 ). Hunt and Graham (1975) describe successful use of undisturbedreaches as models for design and construction of two meanders as part of river relocation for highwayconstruction in Montana. Brookes (1990) describes restoration of the Elbaek in Denmark usingchannel width, depth, and slope from a “natural” reach downstream, confirmed by dimensions of ariver in a neighboring watershed with similar area, geology, and land use.

5. Slope first. Hey (1994) suggests that meanders should be designed by first selecting a meanchannel slope based on hydraulic geometry formulas. However, correlation coefficients for regimeslope formulas are always much smaller than those for width or depth formulas, indicating that theformer are less accurate. Channel slope may also be determined by computing the value required toconvey the design water and sediment discharges (White et al. 1982, Copeland 1994). The mainweakness of this approach is that bed material sediment discharge is required by analytical techniquesand in some cases (e.g., Hey and Thorne 1986) by hydraulic geometry formulas. Sedimentdischarges computed without measured data for calibration may be unreliable.

Site-specific bed material samples and channel geometries are needed to apply these analyticaltechniques and to achieve confidence in the resulting design.



using analytical or empirical tech-niques to scale them to fit slightlydifferent characteristics of the projectreach (for example, a larger or smallerdrainage area).

It is impossible to find an exact replicaof the watershed in which the restora-tion work is located, and subjectivejudgement may play a role in deter-mining what constitutes similarity.The level of uncertainty involved maybe reduced by considering a largenumber of stable reaches. By classify-ing the reference streams, width anddepth data can be grouped by streamtype to reduce the scatter inherent inregional analyses.

A second common meaning of theterm reference reach is a reach with adesired biological condition, whichwill be used as a target to strive forwhen comparing various restorationoptions. For instance, for a stream inan urbanized area, a stream with asimilar drainage area in a nearbyunimpacted watershed might be usedas a reference reach to show what typeof aquatic and riparian communitymight be possible in the project reach.Although it might not be possible toreturn the urban stream topredevelopment conditions, the char-acteristics of the reference reach canbe used to indicate what direction tomove toward. In this use of the term,a reference reach defines desiredbiological and ecological conditions,rather than stable channel geometry.Modeling tools such as IFIM andRCHARC (see Chapter 7) can be usedto determine what restoration optionscome closest to replicating the habitatconditions of the reference reach(although none of the options mayexactly match it).

Application of Regime andHydraulic Geometry Approaches

Typical regime and hydraulic geom-etry relationships are presented inChapter 7. These formulas are mostreliable for width, less reliable fordepth, and least reliable for slope.

Exponents and coefficients for hydrau-lic geometry formulas are usuallydetermined from data for the samestream, the same watershed, streams ofa similar type, or the same physi-ographic region. Because formulacoefficients vary, application of agiven set of hydraulic geometry orregime relationships should be limitedto channels similar to the calibrationsites. Classifying streams can beuseful in refining regime relationships(See Chapter 7's section on StreamClassification).

Published hydraulic geometry relation-ships are usually based on stable,single-thread alluvial channels. Hy-draulic geometry relationships deter-mined through stream classification ofreference reaches can also be valuablefor designing the stream restoration.Channel geometry-discharge relation-ships are more complex formultithread channels. Individualthreads may fit the relationships iftheir partial bankfull discharges areused in place of the total streamflow.Also, hydraulic geometry relationshipsfor gravel-bed rivers are far morenumerous in the literature than thosefor sand-bed rivers.

A trial set of channel properties (aver-age width, depth, and slope) can beevaluated by using several sets ofregime and hydraulic geometry formu-las and comparing results. Greatestweight should be given to formulas



based on sites similar to the projectreach. A logical second step is to useseveral discharge levels in the best-suited sets of formulas. Becausehydraulic geometry relationships aremost compatible with single-channelsand and gravel streams with low bed-material sediment discharge, unstablechannels (aggrading or degradingprofiles) can depart strongly frompublished relationships.

Literature references to the use ofhydraulic geometry formulas forsizing restored channels are abundant.Initial estimates for width and depthfor the restored channel of SeminaryCreek, which drains an urban water-shed in Oakland, California, weredetermined using regional hydraulicgeometry formulas (Riley andMacDonald 1995). Hey (1994, 1995)discusses use of hydraulic geometryrelationships determined using regres-sion analyses of data from gravel bedrivers in the United Kingdom forrestoration design. Newbury andGaboury (1993) used regional hydrau-lic geometry relations based on drain-age area to check width and depth ofrestored channels in Manitoba.

Hydraulic geometry formulas forsizing stream channels in restorationefforts must be used with cautionsince a number of pitfalls are associ-ated with their use:

• The formulas represent hy-draulic geometry only atbankfull or mean annualdischarge. Designers mustalso select a single statistic todescribe bed sediment sizewhen using hydraulic geom-etry relationships. (However,refinements to the Hey and

Thorne [1986] formulas forslope in Table 7.5 should benoted.)

• Downstream hydraulic geom-etry formulas are usually basedon the bankfull discharge, theelevation of which can beextremely difficult to identifyin vertically unstable channels.

• Exponents and coefficientsselected for design must bebased on streams with slopes,bed sediments, and bankmaterials similar to the onebeing designed.

• The premise is that the channelshape is dependent on only oneor two variables.

• Hydraulic geometry relation-ships are power functions witha fair degree of scatter that mayprove too great for reliableengineering design. Thisscatter is indicative of naturalvariability and the influence ofother variables on channelgeometry.

In summary, hydraulic geometryrelationships are useful for preliminaryor trial selection of design channelproperties. Hydraulic and sedimenttransport analyses are recommendedfor final design for the restoration.

Analytical Approaches for ChannelDimensions

Analytical approaches for designingstream channels are based on the ideathat a channel system may be de-scribed by a finite number of variables.In most practical design problems, afew variables are determined by siteconditions (e.g., valley slope and bedmaterial size), leaving up to nine



variables to be computed. However,designers have only three governingequations available: continuity, flowresistance (such as Manning, Chezy,and Darcy-Weisbach), and sedimenttransport (such as Ackers-White,Einstein, and Brownlie). Since thisleaves more unknowns than there areequations, the system is indeterminate.Indeterminacy of the stable channeldesign problem has been addressed inthe following ways:

• Using empirical relationshipsto compute some of the un-knowns (e.g., meander param-eters).

• Assuming values for one ormore of the unknown vari-ables.

• Using structural controls tohold one or more unknownsconstant (e.g., controllingwidth with bank revetments).

• Ignoring some unknownvariables by simplifying thechannel system. For example,a single sediment size issometimes used to describe allboundaries, and a single depthis used to describe water depthrather than mean and maxi-mum depth as suggested byHey (1988).

• Adopting additional governingequations based on assumedproperties of streams withmovable beds and banks. Thedesign methods based on“extremal hypotheses” fall intothis category. These ap-proaches are discussed belowunder analytical approaches forchannels with moving beds.

Table 8.2 lists six examples of analyti-cal design procedures for sand-bed andgravel channels. These procedures aredata-intensive and would be used inhigh-risk or large-scale channel recon-struction work.

Tractive Stress (No Bed Movement)

Tractive stress or tractive force analy-sis is based on the idea that by assum-ing negligible bed material discharge(Q

s = 0) and a straight, prismatic

channel with a specified cross-sec-tional shape, the inequality in vari-ables and governing equations men-tioned above is eliminated. Details areprovided in many textbooks that dealwith stable channel design (e.g.,Richards 1982, Simons and Senturk1977, French 1985). Because themethod is based on the laws of phys-ics, it is less empirical and region-specific than regime or hydraulic

Review Chap. 7'ssection on hydraulicgeometry relation-ships.


Table 8.2: Selectedanalytical proceduresfor stable channeldesign.

Stable Channel Method

Copeland

Domain

Sand-bed rivers

ResistanceEquation

Brownlie

Sediment Transport Equation

Brownlie

Third Relation

Left to designer’s discretion1994

Chang Sand-bed rivers Various Various Minimum stream power1988

Chang Gravel-bed rivers Bray Chang (similar in form to Parker, Einstein)

Minimum slope1988

Abou-Saidaand Saleh

Sand-bed canals Liu-Hwang Einstein-Brown Left to designer’s discretion1987

White et al. Sand-bed rivers White et al. Ackers-White Maximum sediment transport1981

Griffiths Gravel-bed rivers Griffiths Shields entrainment

Empirical stability index1981



geometry formulas. To specify avalue for the force “required to initiatemotion,” the designer must resort toempirical relationships betweensediment size and critical shear stress.In fact, the only difference betweenthe tractive stress approach for designstability analysis and the allowablestress approach is that the effect ofcross-sectional shape (in particular,the bank angles) is considered in theformer (Figure 8.24). Effects ofturbulence and secondary currents arepoorly represented in this approach.

Tractive stress approaches typicallypresume constant discharge, zero bedmaterial sediment transport, andstraight, prismatic channels and aretherefore poorly suited for channelswith moving beds. Additional limita-tions of the tractive stress designapproach are discussed by Brookes(1988) and USACE (1994). Tractivestress approaches are appropriate fordesigning features made of rock orgravel (artificial riffles, revetments,etc.) that are expected to be immobile.

Channels with Moving Beds and KnownSlope

More general analytical approaches fordesigning channels with bed materialdischarge reduce the number of vari-ables by assuming certain constantvalues (such as a trapezoidal cross-sectional shape or bed sediment sizedistribution) and by adding newequations based on an extremal hy-pothesis (Bettess and White 1987).For example, in a refinement of thetractive stress approach, Parker (1978)assumed that a stable gravel channel ischaracterized by threshold conditionsonly at the junction point between bedand banks. Using this assumption andincluding lateral diffusion of longitudi-nal momentum due to fluid turbulencein the analysis, he showed that pointson the bank experience stresses lessthan threshold while the bed moves.

Following Parker’s work, Ikeda et al.(1988) derived equations for stablewidth and depth (given slope and bedmaterial gradation) of gravel channelswith unvegetated banks composed ofnoncohesive material and flat beds inmotion at bankfull. Channels wereassumed to be nearly straight (sinuos-ity < 1.2) with trapezoidal crosssections free of alternate bars. In asubsequent paper Ikeda and Izumi(1990) extended the derivation toinclude effects of rigid bank vegeta-tion.

Extremal hypotheses state that astable channel will adopt dimensionsthat lead to minimization or maximi-zation of some quantity subject toconstraints imposed by the twogoverning equations (e.g., sedimenttransport and flow resistance). Chang(1988) combined sediment transport

Figure 8.24: Lowenergy system withsmall bank angles.Bank angles need to beconsidered when usingthe tractive stressapproach.



and flow resistance formulas withflow continuity and minimization ofstream power at each cross section andthrough a reach to generate a numeri-cal model of flow and sedimenttransport. Special relationships forflow and transverse sediment transportin bends were also derived. Themodel was used to make repeatedcomputations of channel geometrywith various values for input variables.Results of the analysis were used toconstruct a family of design curvesthat yield d (bankfull depth) and w(bankfull width), given bankfull Q, S,and D

50. Separate sets of curves are

provided for sand and gravel bedrivers. Regime-type formulas havebeen fit to the curves, as shown inTable 8.3. These relationships shouldbe used with tractive stress analyses todevelop converging data that increasethe designer’s confidence that theappropriate channel dimensions havebeen selected.

Subsequent work by Thorne et al.(1988) modified these formulas toaccount for effects of bank vegetationalong gravel-bed rivers. The Thorne etal. (1988) formulas in Table 8.3 arebased on the data presented by Heyand Thorne (1986) in Table 7.6.

Channels with Moving Beds and KnownSediment Concentration

White et al. (1982) present an analyti-cal approach based on the Ackers andWhite sediment transport function, acompanion flow resistance relation-ship, and maximization of sedimenttransport for a specified sedimentconcentration. Tables (White et al.1981) are available to assist users inimplementing this procedure. Thetables contain entries for sediment

sizes from 0.06 to 100 millimeters,discharges up to 35,000 cubic feet persecond, and sediment concentrationsfrom 10 to 4,000 parts per million.However, this procedure is not recom-mended for gravel bed channels(USACE 1994). Sediment concentra-tion at bankfull flow is required as aninput variable, which limits the useful-ness of this procedure. Procedures forcomputing sediment discharge, Q

S, are

outlined in Chapter 7. Copeland(1994) found that the White et al.(1982) method for channel design wasnot robust for cohesive bed materials,artificial grade controls, and disequi-librium sediment transport. Themethod was also found inappropriatefor an unstable, high-energy ephem-eral sand-bed stream (Copeland 1994).However, Hey (1990) found theAckers-White sediment transportfunction performed well when analyz-ing stability of 18 flood control chan-nels in Britain.

The approach described by Copeland(1994) features use of the Brownlie(1981) flow-resistance and sediment-transport relations, in the form of thesoftware package “SAM” (Thomas etal. 1993). Additional features includethe determination of input bed materialconcentration by computing sedimentconcentration from hydraulic param-eters for an upstream “supply reach”represented by a bed slope, a trapezoi-dal cross section, bed-material grada-tion, and a discharge. Bank and bedroughness are composited using theequal velocity method (Chow 1959) toobtain roughness for a cross section.A family of slope-width solutions thatsatisfy the flow resistance and sedi-ment transport relations are thencomputed. The designer then selects



Chang equations for determining river width and depth. Coefficients for equations of the form w = k1QK2; d = K4QK5; where w is mean bankfull width (ft), Q is the bor dominant discharge (ft3/s), d is mean bankfull depth (ft), D50 is median bed-material size (mm), and S is slope (ft/ft).

a wc and dc in these equations are calculated using exponents and coefficients from the row labeled “gravel-bed rivers”..k1* = (S D50

-0.5 - 0.00238Q-0.51)0.02.

k4* = exp[-0.38 (420.17S D50-0.5Q-0.51 -1)0.4].

k1** = (S D50-0.5 )0.84.

k4** = 0.015 - 0.025 In Q - 0.049 In (S D50-0.5).

k1*** = 0.2490[ ln(0.0010647D501.15/SQ0.42 )]2.

k4*** = 0.2418 ln(0.0004419D501.15/SQ0.42 ).

Author

Chang

Year

1988

Thorne et al.

1988

Data Domain

Meandering or braided sand-bed rivers with:

k1 k2 k4

Equiwidth point-bar streams and stable canals

0.00238 < SD50-0.5 Q-0.51 and

SD50-0.5 Q-0.55 < 0.05

3.49k1* 3.51k4*

Straight braided streams 0.05 < SD50-0.5 Q-0.55 and

SD50-0.5 Q-0.51 < 0.047

Unknown and unusual

Braided point-bar and wide-bend point-bar streams; beyond upper limit lie steep, braided streams

0.047 < SD50-0.5 Q-0.51 <

indefinite upper limit33.2k1** 0.93 1.0k4**

Same as for Thorne and Hey 1986

Gravel-bed rivers 1.905 + k1*** 0.47 0.2077 + k4***

Adjustments for bank vegetationa

Grassy banks with no trees or shrubs

w = 1.46 wc – 0.8317

d = 0.8815 dc + 0.2106

1-5% tree and shrub cover w = 1.306 wc – 8.7307

d = 0.5026 dc + 1.7553

5-50% tree and shrub cover w = 1.161 wc – 16.8307

d = 0.5413 dc + 2.7159

Greater than 50% tree and shrub cover, or incised into flood plain

w = 0.9656 wc – 10.6102

d = 0.7648 dc + 1.4554

Table 8.3: Equations forriver width and depth.



any combination of channel propertiesthat are represented by a point on theslope-width curve. Selection may bebased on minimum stream power,maximum possible slope, widthconstraint due to right-of-way, ormaximum allowable depth. Thecurrent (1996) version of the Copelandprocedure assumes a straight channelwith a trapezoidal cross section andomits the portion of the cross sectionabove side slopes when computingsediment discharge. Effects of bankvegetation are considered in theassigned roughness coefficient.

The Copeland procedure was tested byapplication to two existing streamchannels, the Big and Colewa Creeksin Louisiana and Rio Puerco in NewMexico (Copeland 1994). Consider-able professional judgment was usedin selection of input parameters. TheCopeland method was found inappli-cable to the Big and Colewa Creeks(relatively stable perennial streamswith sand-clay beds), but applicable toRio Puerco (high-energy, ephemeralsand-bed stream with stable profileand unstable banks). This result is notsurprising since all stable channeldesign methods developed to datepresume alluvial (not cohesive orbedrock) beds.

Use of Channel Models forDesign VerificationIn general, a model can be envisionedas a system by whose operation thecharacteristics of other similar systemsmay be predicted. This definition isgeneral and applies to both hydraulic(physical) and computational (math-ematical) models. The use and opera-tion of computer models has improved

in recent years as a result of betterknowledge of fluvial hydraulics andthe development of sophisticateddigital control and data acquisitionsystems.

Any stream corridor restoration designneeds careful scrutiny because itslong-term impact on the stream systemis not easy to predict. Sound engineer-ing often dictates the use of computermodels or physical models to checkthe validity of a proposed design.Since most practitioners do not haveeasy access to physical modelingfacilities, computer models are muchmore widely used. Computer modelscan be run in a qualitative mode withvery little data or in a highly precisequantitative mode with a great deal offield data for calibration and verifica-tion.

Computer models can be used to easilyand cheaply test the stability of arestoration design for a range ofconditions, or for a variety of alterna-tive channel configurations. A“model” can vary in cost from severalhundred dollars to several hundredthousand dollars, depending on whatmodel is used, the data input, thedegree of precision required, and thelength and complexity of the reach tobe modeled. The decision as to whatmodels are appropriate should be madeby a hydraulic engineer with a back-ground in sediment transport.

The costs of modeling could be smallcompared to the cost of redesign orreconstruction due to failure. If theconsequences of a project failurewould result in a high risk of cata-strophic damage or death, and the site-specific conditions result in an unac-ceptable level of uncertainty when



applying computer models, a physicalmodel is the appropriate tool to use fordesign.

Physical Models

In some instances, restoration designscan become sufficiently complicatedto exceed the capabilities of availablecomputational models. In othersituations, time might be of the es-sence, thus precluding the develop-ment of new computational modelingcapabilities. In such cases the de-signer must resort to physical model-ing for verification.

Depending on the scaling criteria usedto achieve similitude, physical modelscan be classified as distorted, fixed, ormovable-bed models. The theory andpractice of physical modeling arecovered in detail by French (1985),Jansen et al. (1979), and Yalin (1971)and are beyond the scope of thisdocument. Physical modeling, likecomputational modeling, is a technol-ogy that requires specialized expertiseand considerable experience. The U.S.

Army Waterways Experiment Station,Vicksburg, Mississippi, has exten-sively developed the technique ofdesigning and applying physicalmodels of rivers.

Computer Models

Computer models are structured andoperated in the same way as a physicalmodel (Figure 8.25). One part of thecode defines the channel planform, thebathymetry, and the material proper-ties of transported constituents. Otherparts of the code create conditions atthe boundaries, taking the place of thelimiting walls and flow controls in thephysical model. At the core of thecomputer code are the water andsediment transport solvers. “Turningon” these solvers is equivalent torunning the physical model. At theend of the simulation run the newchannel bathymetry and morphologyare described by the model output.This section summarizes computa-tional channel models that can beuseful for evaluation of stream corri-dor restoration designs. Since it is notpossible to include every existingmodel in the space available, thediscussion here is limited to a fewselected models (Table 8.4). Inaddition, Garcia et al. (1994) reviewmathematical models of meander bendmigration.

These models are characterized ashaving general applicability to aparticular class of problems and aregenerally available for desktop com-puters using DOS operating systems.Their conceptual and numericalschemes are robust, having beenproven in field applications, and thecode can be successfully used by

set up model of prototype

executemodel

select model to evaluate

design

new restoration

design

modelresults

evaluateresults

accept or revisedesign

Figure 8.25: Use ofmodels for designevaluation.Modeling helps evaluateeconomics andeffectiveness ofalternative designs.



Model

Discretization and formulation:

CHARIMA Fluvial-12 HEC-6 TABS-2 Meander USGS D•O•T GSTARS

Unsteady flow | stepped hydrograph Y | Y Y | Y N | Y Y | Y N | Y Y | Y N | Y N | Y

One-dimensional | quasi-two-dimensional Y | N Y | Y Y | N N | N N | N N Y | Y Y | Y

Two-dimensional, depth-average flow N N N Y Y Y | Y N N | Y

Deformable bed | banks Y | N Y | Y Y | N Y | N Y | N Y | N Y | Y Y | Y

Graded sediment load Y Y Y Y Y N Y Y

Nonuniform grid Y Y Y Y Y Y Y Y

Variable time stepping Y N Y N N N N Y

Numerical solution scheme:

Standard step method N Y Y N N N Y Y

Finite difference Y N Y N Y Y Y Y

Finite element N N N Y N N N N

Modeling capabilities

Upstream water and sediment hydrographs Y Y Y Y Y Y Y Y

Downstream stage specification Y Y Y Y Y N Y Y

Floodplain sedimentation N N N Y N N N N

Suspended | total sediment transport Y | N Y | N N | Y Y | N N | N N | Y N | Y N | Y

Bedload transport Y Y Y N Y N N Y

Cohesive sediments N N Y Y N Y N Y

Bed armoring Y Y Y N N N Y Y

Hydraulic sorting of substrate material Y Y Y N N N Y Y

Fluvial erosion of streambanks N Y N N N N Y Y

Bank mass failure under gravity N N N N N N Y N

Straight | irregular nonprismatic reaches Y | N Y | N Y | N Y | Y N | N N | N Y | Y Y | Y

Branched | looped channel network Y | Y Y | N Y | N Y | Y N | N N | N N | N N | N

Channel beds N Y N Y Y N Y N

Meandering belts N N N N N Y N N

Rivers Y Y Y Y Y Y Y Y

Bridge crossings N N N Y N N N N

Reservoirs N Y Y N N N N Y

User support

Model documentation Y Y Y Y Y Y Y Y

User guide | hot-line support N | N Y | N Y | Y Y | N N | N Y | N N | N Y | N

Y = YesN = No

Table 8.4: Examples of computational models.



persons without detailed knowledge ofthe core computational techniques.Examples of these models and theirfeatures are summarized in Table 8.4.The acronyms in the column titlesidentify the following models:CHARIMA (Holly et al. 1990),FLUVIAL-12 (Chang 1990), HEC-6(HEC 1991), TABS-2 (Williams andMcAnally 1985), MEANDER(Johannesson and Parker 1985), USGS(Nelson and Smith 1989), D-O-T(Darby and Thorne 1996, Osman andThorne 1988), GSTARS (Molinas andYang 1996) and GSTARS 2.0 (Yang etal. 1998). GSTARS 2.0 is an en-hanced and improved PC version ofGSTARS. HEC-6, TABS-2, andUSGS are federal, public domainmodels, whereas CHARIMA, FLU-VIAL-12, MEANDER, and D-O-T areacademic, privately owned models.

With the exception of MEANDER, allthe above models calculate at eachcomputational node the fractionalsediment load and rate of bed aggrada-tion or degradation, and update thechannel topography. Some of themcan simulate armoring of the bedsurface and hydraulic sorting (mixing)of the underlying substrate material.CHARIMA, FLUVIAL-12, HEC-6,and D-O-T can simulate transport ofsands and gravels. TABS-2 can beapplied to cohesive sediments (claysand silts) and sand sediments that arewell mixed over the water column.USGS is specially designed for gravelbed-load transport. FLUVIAL-12 andHEC-6 can be used for reservoirsedimentation studies. GSTARS 2.0can simulate bank failure.

Comprehensive reviews on the capa-bilities and performance of these and

other existing channel models areprovided in reports by the NationalResearch Council (1983), Fan (1988),Darby and Thorne (1992), and Fan andYen (1993).

Detailed Design

Channel Shape

Natural stream width varies continu-ously in the longitudinal direction, anddepth, bed slope, and bed material sizevary continuously along the horizontalplane. These variations give rise tonatural heterogeneity and patterns ofvelocity and bed sediment size distri-bution that are important to aquaticecosystems.

Widths, depths, and slopes computedduring design should be adopted asreach mean values, and restoredchannels should be constructed withasymmetric cross sections (Hunt andGraham 1975, Keller 1978, Iversen etal. 1993, MacBroom 1981) (Figure8.26). Similarly, meander planformshould vary from bend to bend aboutaverage values of arc length andradius. A reconstructed floodplainshould not be perfectly flat (Figure8.27).

Channel Longitudinal Profile andRiffle Spacing

In stream channels with significantamounts of gravel (D

50 > 3 mm)

(Higginson and Johnston 1989), rifflesshould be associated with steep zonesnear meander inflection points. Rifflesare not found in channels with beds offiner materials. Studies conducted byKeller and Melhorn (1978) and con-firmed by Hey and Thorne (1986)indicate pool-riffle spacing should



vary between 3 and 10 channel widthsand average about 6 channel widthseven in bedrock channels. Morerecent work by Roy and Abrahams(1980) and Higginson and Johnston(1989) indicates that pool-riffle spac-ing varies widely within a givenchannel.

Average riffle spacing is often (but notalways) half the meander length sinceriffles tend to occur at meander inflec-tion points or crossovers. Rifflessometimes appear in groups or clus-ters. Hey and Thorne (1986) analyzeddata from 62 sites on gravel-bed riversin the United Kingdom and foundriffle spacing varied from 4 to 10channel widths with the least squaresbest fit at 6.31 channel widths. Rifflespacing tends to be nearer 4 channelwidths on steeper gradients and 8 to 9channel widths on more gradual slopes(R.D. Hey, personal communication).Hey and Thorne (1986) also developedregression formulas for riffle width,mean depth, and maximum depth.

Stability AssessmentThe risk of a restored channel beingdamaged or destroyed by erosion ordeposition is an important consider-ation for almost all restoration work.Designers of restored streams areconfronted with rather high levels ofuncertainty. In some cases, it may bewise for designers to compute risk offailure by calculating the joint prob-ability of design assumptions beingfalse, design equation inaccuracy, andoccurrence of extreme hydrologicevents during project life. Gooddesign practice also requires checkingchannel performance at dischargeswell above and below the design

condition. A number of approachesare available for checking both thevertical (bed) and horizontal (bank)stability of a designed stream. Thesestability checks are an important partof the design process.

Vertical (Bed) Stability

Bed stability is generally a prerequisitefor bank stability. Aggrading channelsare liable to braid or exhibit acceler-ated lateral migration in response tomiddle or point bar growth. Degrad-ing channels widen explosively whenbank heights and angles exceed acritical threshold specific to bank soil

Plan

Station

Elev

atio

n

Profile

a

b

c

d

e

f

g

a1

a1a

b1

b1b

c1

c1c

d1

d1d

e1

e1e

f1

f1 f

g1

g1g

Figure 8.26: Exampleplan and profile of anaturally meanderingstream.Channel cross sectionsvary based on width,depth, and slope.

Figure 8.27: A streammeander and raisedfloodplain.Natural floodplains riseslightly between acrossover and an apex ofa meander.



type. Bed aggradation can be ad-dressed by stabilizing eroding chan-nels upstream, controlling erosion onthe watershed, or installing sedimenttraps, ponds (Haan et al. 1994), ordebris basins (USACE 1989b). Ifaggradation is primarily due to deposi-tion of fines, it can be addressed bynarrowing the channel, although anarrower channel might require morebank stabilization.

If bed degradation is occurring orexpected to occur, and if modificationis planned, the restoration initiativeshould include flow modification,grade control measures, or otherapproaches that reduce the energygradient or the energy of flow. Thereare many types of grade controlstructures. The applicability of aparticular type of structure to a spe-cific restoration depends on a numberof factors, such as hydrologic condi-tions, sediment size and loading,channel morphology, floodplain andvalley characteristics, availability ofconstruction materials, ecologicalobjectives, and time and fundingconstraints. For more information on

various structure designs, refer toNeilson et. al. (1991), which providesa comprehensive literature review ongrade control structures with an anno-tated bibliography. Grouted boulderscan be used as a grade control struc-ture. They are a key component in thesuccessful restoration of the SouthPlatte River corridor in Denver, Colo-rado (McLaughlin Water Engineers,Ltd., 1986).

Grade control structure stilling basinscan be valuable habitats in severelydegraded warm water streams (Cooperand Knight 1987, Shields and Hoover1991). Newbury and Gaboury (1993)describe the construction of artificialriffles that serve as bed degradationcontrols. Kern (1992) used “riverbottom ramps” to control bed degrada-tion in a River Danube meanderrestoration initiative. Ferguson (1991)reviews creative designs for gradecontrol structures that improve stream-side habitat and aesthetic resources(Figure 8.28).

Horizontal (Bank) Stability

Bank stabilization may be necessary inrestored channels due to floodplainland uses or because constructed banksare more prone to erosion than “sea-soned” ones, but it is less than ideal ifecosystem restoration is the objective.Floodplain plant communities owetheir diversity to physical processesthat include erosion and depositionassociated with lateral migration(Henderson 1986). Bank erosioncontrol methods must be selected withthe dominant erosion mechanisms inmind (Shields and Aziz 1992).

Bank stabilization can generally begrouped into one of the following

Figure 8.28: Gradecontrol structure.Control measures candouble as habitatrestoration devices andaesthetic features.



three categories: (1) indirect methods,(2) surface armor, and (3) vegetativemethods. Armor is a protective mate-rial in direct contact with the stream-bank. Armor can be categorized asstone, other self-adjusting armor(sacks, blocks, rubble, etc.), rigidarmor (concrete, soil cement, groutedriprap, etc.) and flexible mattress(gabions, concrete blocks, etc.).Indirect methods extend into thestream channel and redirect the flowso that hydraulic forces at the channelboundary are reduced to a nonerosivelevel. Indirect methods can be classi-fied as dikes (permeable and imperme-able) and other flow deflectors such asbendway weirs, stream “barbs,” andIowa vanes. Vegetative methods canfunction as either armor or indirectprotection and in some applicationscan function as both simultaneously.A fourth category is composed oftechniques to correct problems causedby geotechnical instabilities.

Guidance on selection and design ofbank protection measures is providedby Hemphill and Bramley (1989) andHenderson (1986). Coppin andRichards (1990), USDA-NRCS(1995), and Shields et al. (1995d)provide additional detail on the use ofvegetative techniques (see followingsection). Newly constructed channelsare more susceptible to bank erosionthan older existing channels, withsimilar inflows and geometries, due tothe influence of vegetation, armoring,and the seasoning effect of claydeposition on banks (Chow 1959). Inmost cases, outer banks of restored ornewly constructed meanders willrequire protection. Structural tech-niques are needed (e.g., Thorne et al.1995) if immediate stability is re-

quired, but these may incorporateliving components. If time permits,the new channel may be constructed“in the dry” and banks planted withwoody vegetation. After allowing thevegetation several growing seasons todevelop, the stream may be diverted infrom the existing channel (R.D. Hey,personal communication).

Bank Stability Check

Outer banks of meanders erode, buterosion rates vary greatly from streamto stream and bend to bend. Observa-tion of the project stream and similarreaches, combined with professionaljudgment, may be used to determinethe need for bank protection, or ero-sion may be estimated by simple rulesof thumb based largely on studies thatrelate bend migration rates to bendgeometry (e.g., Apmann 1972 andreview by Odgaard 1987) (Figure8.29). More accurate prediction of therate of erosion of a given streambankis at or beyond the current state of theart. No standard methods exist, butseveral recently developed tools are

Figure 8.29: Channelexhibiting acceleratedlateral migration.Erosion of an outer bankon the Missouri River is anatural process; howeverthe rate of erosion shouldbe monitored.



available. None of these have beenused in extremely diverse settings, andusers should view them with caution.

Tools for predicting bank erosion maybe divided into two groups: (1) thosewhich predict erosion primarily due tothe action of water on the streambanksurface and (2) those which focus onsubsurface geotechnical characteris-tics.

Among the former is an index ofstreambank erodibility based on fieldobservations of emergency spillways(Temple and Moore 1995). Erosion ispredicted for sites where a powernumber based on velocity, depth, andbend geometry exceeds an erodibilityindex computed from tabulated valuesof streambank material properties.Also among this group are analyticalmodels such as the one developed byOdgaard (1989), which contain rathersophisticated representations of flowfields, but require input of an empiri-cal constant to quantify soil andvegetation properties. These modelsshould be applied with careful consid-eration of their limitations. For ex-ample, Odgaard’s model should not beapplied to bends with “large curva-ture.”

The second group of predictive toolsfocuses on banks that undergo massfailure due to geotechnical processes.Side slopes of deep channels may behigh and steep enough to begeotechnically unstable and to failunder the influence of gravity. Fluvialprocesses in such a situation serveprimarily to remove blocks of failedmaterial from the bank toe, leading toa resteepened bank profile and a newcycle of failure, as shown in Figure8.30. Study of bank failure processesalong incised channels has led to aprocedure for relating bank geometryto stability for a given set of soilconditions (Osman and Thorne 1988,Thorne and Osman 1988). If banks ofa proposed design channel are to behigher than about 10 feet, stabilityanalysis should be conducted. Theseanalyses are described in detail inChapter 7. Bank height estimatesshould allow for scour along the

Stage I

Stage II

Stage III

Bank Angle (deg)

Ban

k H

eig

ht

(ft)

10˚0˚

10˚

20˚

30˚

45˚

stable

unstableunreliable

90˚

I

Bank Angle (deg)

Ban

k H

eig

ht

(ft)

10˚0˚

10˚

20˚

30˚

30˚

45˚

stableII

unstable

unstable

unreliable

90˚

Bank Angle (deg)

Ban

k H

eig

ht

(ft)

10˚0˚

10˚

20˚

45˚

stable III

unreliable

90˚

Figure 8.30: Bankfailure stages.Stability of a bank willvary from stable tounstable depending onbank height, bank angle,and soil conditions.



outside of bends. High, steep banksare also susceptible to internal erosion,or piping, as well as streambanks ofsoils with high dispersion rates.

Allowable Velocity Check

Fortier and Scobey (1926) publishedtables regarding the maximumnonscouring velocity for given chan-nel boundary materials. Differentversions of these tables have appearedin numerous subsequent documents,notably Simons and Senturk (1977)and USACE (1991). The applicabilityof these tables is limited to relativelystraight silt and sand-bed channelswith depths of flow less than 3 feetand very low bed material loads.Adjustments to velocities have beensuggested for situations departingfrom those specified. Although slightrefinements have been made, thesedata still form the basis of the allow-able velocity approach.

Figure 8.31 contains a series of graphsthat summarize the tables and aid inselecting correction factors for flowdepth, sediment concentration, flowfrequency, channel curvature, bankslope, and channel boundary soilproperties. Use of the allowablevelocity approach is not recommendedfor channels transporting a significantload of material larger than 1 mm.The restoration design, howevershould also consider the effects ofhydraulic roughness and the protectionafforded by vegetation.

Perhaps because of its simplicity, theallowable velocity method has beenused directly or in slightly modifiedform for many restoration applica-tions. Miller et al. (1983) used allow-able velocity criteria to design man-

made gravel riffles located immedi-ately downstream of a dam releasing aconstant discharge of sediment-freewater. Shields (1983) suggested usingallowable velocity criteria to sizeindividual boulders placed in channelsto serve as instream habitat structures.Tarquin and Baeder (1983) present adesign approach based on allowablevelocity for low-order ephemeralstreams in Wyoming landscapesdisturbed by surface mining. Velocityof the design event (10-year recurrenceinterval) was manipulated by adjustingchannel length (and thus slope), width,and roughness. Channel roughnesswas adjusted by adding meanders,planting shrubs, and adding coarse bedmaterial. The channel width-to-depthratio design was based on the pre-mining channel configuration.

Allowable Stress Check

Since boundary shear stress is moreappropriate than velocity as a measureof the forces driving erosion, graphshave also been developed for allow-able shear stress. The average bound-ary shear stress acting on an openchannel conveying a uniform flow ofwater is given by the product of theunit weight of water (γ, lb/ft3) timesthe hydraulic radius (R, ft) times thebed slope S:

τ = γRS

Figure 8.32 is an example of allow-able shear stress criteria presented ingraphical form. The most famousgraphical presentation of allowableshear stress criteria is the Shieldsdiagram, which depicts conditionsnecessary for initial movement ofnoncohesive particles on a flat bedstraight channel in terms of dimension-



2.0 fps

4.5

5.0

5.5

6.0

6.5

7.0

Plasticity Index

Basi

c Ve

loci

ty (f

ps)

sediment laden flow3.5

10 12 14 16 18 20

GC

SC

ML,OL,SM

CL,GM

MH,OH

CH

22 24

4.0

2.03.04.05.06.07.08.09.0

10.011.012.013.0

Basi

c Ve

loci

ty (f

ps)

0.0

Basic Velocity for Discrete Particles of Earth Materials

1.0

3.0

3.5

4.0

4.5

5.0

5.5

Plasticity Index

Basi

c Ve

loci

ty (f

ps)

sediment free flow2.0

10 12 14 16 18 20

GC

ML,OL,SM

GM,CL,SC

MH,OH

CH

22 24

2.5

Water Depth (feet)

Corr

ecti

on F

acto

r D

Corr

ecti

on F

acto

r A

8 614 12 10

F

0.41.5 2.0Cotangent of Slope Angle (z)

2.5

bank slope

density

Corr

ecti

on F

acto

r B

3.0

0.6

0.8

1.0

0.80.2 0.4

Void Ratio (e)

Grain Size (inches)

0.6

Corr

ecti

on F

acto

r C e

0.8 1.0 1.2 1.4

11086421

Sand Gravel

sediment free

sediment laden

Enter chart with D75 particle sizeto determine basic velocity.

CobbleFine S

12

0.9

1.0

1.1

1.2

CH,MHCL,ML

SM,SC,GM

,GC

frequency of design flow

1.01 2

Flood Frequency (percent chance)3

Corr

ecti

on F

acto

r F

4 5 6 7 8 9 10

1.2

1.4

1.6

1.8

2.0

depth of design flow

0.92 4 6 8 10 12 14 16 18 20

1.0

1.1

1.2

1.3

1.4

1.5

alignment

0.716Curve Radius ÷ Water Surface Width

4

0.8

0.9

1.0

14

18

Discrete ParticlesSediment Laden Flow

D75 > 0.4mmD75 < 0.4mm

Sediment Free FlowD75 > 0.2mmD75 < 0.2mm

Coherent Earth MaterialsPI > 10PI < 10

Channel Boundary Materials

basic velocity chart value x D x A x2.0 fps

basic velocity chart value x D x A x2.0 fps

basic velocity chart value x D x A x F

Allowable Velocity

Allowable Velocities for Unprotected Earth Channels

Basic Velocities for Coherent Earth Materials (vb)

Notes:In no case should the allowable velocity be exceeded when the 10chance discharge occuregardless of the desigflow frequency.

Figure 8.31: Allowablevelocities forunprotected earthchannels.Curves reflect practicalexperience in design ofstable earth channels.Source: USDA SoilConservation Service 1977.



less variables (Vanoni 1975). TheShields curve and other allowableshear stress criteria (e.g., Figure10.5,Henderson 1966; Figure 7.7, Simonsand Senturk 1977) are based onlaboratory and field data. In simplestform, the Shields criterion for channelstability is (Henderson 1966):

RS/[(SS-1)DS] < a constant

for DS > ~ 6 mm

where Ss is the specific gravity of the

sediment and DS is a characteristic bed

sediment size, usually taken as themedian size, D

50, for widely graded

material. Note that the hydraulicradius, R, and the characteristic bedsediment size, D

S, must be in the same

units for the Shields constant to bedimensionless. The dimensionlessconstant is based on measurementsand varies from 0.03 to 0.06 depend-ing on the data set used to determine itand the judgment of the user (USACE1994).

These constant values are for straightchannels with flat beds (no dunes orother bedforms). In natural streams,bedforms are usually present, andvalues of this dimensionless constantrequired to cause entrainment of bedmaterial may be greater than 0.06. Itshould be noted that entrainment doesnot imply channel erosion. Erosionwill occur only if the supply of sedi-ment from upstream is less than thattransported away from the bed by theflow. However, based on a study of 24gravel-bed rivers in the Rocky Moun-tain region of Colorado, Andrews(1984) concluded that stable gravel-bed channels cannot be maintained atvalues of the Shields constant greaterthan about 0.080. Smaller Shieldsconstant values are more conservative

with regard to channel scour, but lessconservative with regard to deposition.If S

S = 2.65, and the constant is as-

sumed to be 0.06, the equation abovesimplifies to D

50 = 10.1RS.

Allowable shear stress criteria are notvery useful for design of channels withbeds dominated by sand or finermaterials. Sand beds are generally inmotion at design discharge and havedunes, and their shear stress values aremuch larger than those indicated bythe Shields criterion, which is forincipient motion on a plane bed.Allowable shear stress data for cohe-sive materials show more scatter thanthose for sands and gravels (Grissingeret al. 1981, Raudkivi and Tan 1984),and experience and observation withlocal channels are preferred to pub-lished charts like those shown in Chow(1959). Models of cohesive soilerosion require field or laboratoryevaluation of model parameters orconstants. Extrapolation of laboratoryflume results to field conditions isdifficult, and even field tests aresubject to site-specific influences.Erosivity of cohesive soils is affected

Figure 8.32: Allowablemean shear stress forchannels withboundaries ofnoncohesive materiallarger than 5 mmcarrying negligible bedmaterial load.Shear stress diminisheswith increasedsuspended sedimentconcentrations.From Lane 1955.

Shea

r St

ress

(lb

s/ft

2 )

D50 (mm)

0.02

0.01

0.10

0.20

0.40

0.70

1.00

0.2 0.3 0.5 0.7 1 2

C = fine suspended sedimentconcentration

3 4 50.1

1,000 < C < 2,000 ppmC ≥ 20,000 ppm

C < 1,000 ppm



by the chemical composition of thesoil, the soil water, and the stream,among other factors.

However, regional shear stress criteriamay be developed from observationsof channels with sand and clay beds.For example, USACE (1993) deter-mined that reaches in the ColdwaterRiver Watershed in northwest Missis-sippi should be stable with an averageboundary shear stress at channel-forming (2-year) discharge of 0.4 to0.9 lb/ft2.

The value of the Shields constant alsovaries with bed material size distribu-tion, particularly for paved or armoredbeds. Andrews (1983) derived aregression relationship that can beexpressed as:

RS/[(SS-1)D

i ] < 0.0834 (D

i/ D

50)-0.0872

When the left side of the above ex-pression equals the right, bed-sedi-ment particles of size D

i are at the

threshold of motion. The D50

value inthe above expression is the mediansize of subsurface material. Therefore,if D

50 = 30 mm, particles with a

diameter of 100 mm will be entrainedwhen the left side of the above equa-tion exceeds 0.029. This equation isfor self-formed rivers that have natu-rally sorted gravel and cobble bedmaterial. The equation holds forvalues of D

i/D

50 between 0.3 and 4.2.

It should be noted that R and Di on the

left side of the above equation must beexpressed in the same units.

Practical Guidance: AllowableVelocity and Shear Stress

Practical guidance for application ofallowable velocity and shear stress

approaches is provided by the NaturalResources Conservation Service(NRCS), formerly the U.S. SoilConservation Service (SCS)(1977),and USACE (1994). See Figure 8.31.

Since form roughness due to sanddunes, vegetation, woody debris, andlarge geologic features in streamsdissipates energy, allowable shearstress for bed stability may be higherthan indicated by laboratory flumedata or data from uniform channels.It is important to compute cross-sectional average velocities or shearstresses over a range of discharges andfor seasonal changes in the erosionresistance of bank materials, ratherthan for a single design condition.Frequency and duration of dischargescausing erosion are important factorsin stability determination. In cobble-or boulder-bed streams, bed movementsometimes occurs only for dischargeswith return periods of several years.

Computing velocity or shear stressfrom discharge requires design crosssections, slope, and flow resistancedata. If the design channel is notextremely uniform, typical or averageconditions for rather short channelreaches should be considered. Inchannels with bends, variations inshear stress across the section can leadto scour and deposition even whenaverage shear stress values are withinallowable limits. The NRCS (formerlySCS) (1977) gives adjustment factorsfor channel curvature in graphicalform that are based on very limiteddata (see Figure 8.31). Velocitydistributions and stage-dischargerelations for compound channels arecomplex (Williams and Julien 1989,Myers and Lyness 1994).



Allowable velocity or shear stresscriteria should be applied to in-channelflow for a compound cross sectionwith overbank flow, not cross-sec-tional average conditions (USACE1994). Channel flow resistancepredictors that allow for changingconditions with changing dischargeand stage (e.g., Brownlie 1983 forsand-bed channels) should be usedrather than constant resistance values.

If the existing channel is stable, designchannel slope, cross section, androughness may be adjusted so that thecurrent and proposed systems havematching curves of velocity versusdischarge (USACE 1994). Thisapproach, while based on allowablevelocity concepts, releases the proce-dure from published empirical valuescollected in other rivers that might beintrinsically different from the one inquestion.

Allowable Stream Power or Slope

Brookes (1990) suggested the productof bankfull velocity and shear stress,which is equal to the stream power perunit bed area, as a criterion for stabil-ity in stream restoration initiatives.This is based on experience withseveral restoration initiatives in Den-mark and the United Kingdom withsandy banks, beds of glacial outwashsands, and a rather limited range ofbankfull discharges (~15 to 70 cfs).These data are plotted as squares,triangles, and circles in Figure 8.33.

Brookes suggested that a stream powervalue of 2.4 ftlb/sec/ft2 discriminatedwell between stable and unstablechannels. Projects with stream powersless than about 1.0 ftlb/sec/ft2 failedthrough deposition, whereas those

with stream powers greater than about3.4 ftlb/sec/ft2 failed through erosion.

Since these criteria are based onobservation of a limited number ofsites, application to different streamtypes (e.g., cobble-bed rivers) shouldbe avoided. However, similar criteriamay be developed for basins of inter-est. For example, data points repre-senting stable reaches in the ColdwaterRiver watershed of northwesternMississippi are shown in Figure 8.34as stars. This watershed is character-ized by incised, straight (channelized)sand-bed channels with cohesivebanks. Slopes for stable reaches weremeasured in the field, and 2-yeardischarges were computed using awatershed model (HEC-1) (USACE1993).

Figure 8.33: Brookes’stream power stabilitycriteria.Stream power is theproduct of bankfullvelocity and shear stress.

Ch

an

nel Slo

pe

Bankfull Discharge per Unit Width, ft2 s-1

10.0001

0.001

0.01

0.1

10

failure through erosiongenerally successfulfailure through depositionlines of constant stream powerstable reaches, Coldwater River basin, Mississippi

100

0.685 ft lbs s -1ft -2

2.4 ft lbs s -1ft -2

6.85 ft lbs s -1ft -2



Brookes’ stream power criterion is oneof several region-specific stabilitytests. Others include criteria based onslope and shear stress. Using empiri-cal data and observation, the Corps ofEngineers has developed relationshipsbetween slope and drainage area forvarious watersheds in northwestern

Mississippi (USACE 1989c). Forexample, stable reaches in threewatersheds had slopes that clusteredaround the regression line:

S = 0.0041 A-0.365

where A is the contributing drainagearea in square miles. Reaches withmuch steeper slopes tended to bedegradational, while those with moregradual slopes tended to beaggradational. Downs (1995) devel-oped stability criteria for channelreaches in the Thames Basin of theUnited Kingdom based entirely onslope: channels straightened during the20th century were depositional ifslopes were less than 0.005 and ero-sional if slopes were greater.

Sediment Yield and Delivery

Sediment Transport

If a channel is designed using anempirical or a tractive stress approach,computation of sediment-transportcapacity allows a rough check todetermine whether deposition is likelyto be a problem. Sediment transportrelationships are heavily dependent onthe data used in their development.Inaccuracy may be reduced by select-ing transport functions appropriate tothe stream type and bed sediment sizein question. Additional confidencecan be achieved by obtaining calibra-tion data; however, calibration data arenot available from a channel yet to beconstructed. If the existing channel isreasonably stable, designers cancompute a sediment discharge versusstreamflow relationship for the exist-ing and proposed design channelsusing the same sediment transportfunction and try to match the curves asclosely as possible (USACE 1994).

Allowable Shear Stress

The shape of the bed material size distribution is an importantparameter for determining the threshold of motion of individualsediment sizes in a bed containing a mixture of sand andgravel. Beds composed of unimodal (particle-size distributionshows no secondary maxima) mixtures of sand and gravelwere found to have a narrow range of threshold shearstresses for all sizes present on the bed surface. Forunimodal beds, the threshold of motion of all grain sizes onthe bed was found to be estimated adequately by using theShields curve for the median grain size. Bed sedimentscomposed of bimodal (particle-size distribution shows onesecondary maximum) mixtures of sands and gravels werefound to have threshold shear stresses that are still a functionof grain size, although much less so than predicted by theShields curve. For bed material with bimodal sizedistributions, using the Shields curve on individual grain sizesgreater than the median size overestimates the threshold ofmotion and underestimates the threshold of motion for grainsizes less than the median size. Critical shear stresses forgravel beds may be elevated if gravels are tightly interlockedor imbedded.

Jackson and Van Haveren (1984) present an iterativetechnique for designing a restored channel based onallowable shear stress. Separate calculations wereperformed for channel bed and banks. Channel designincluded provision for gradual channel narrowing as the bankvegetation develops, and bank cohesion and resistance toerosion increase. Newbury and Gaboury (1993) use anallowable tractive force graph from Lane (1955) to checkstability of channel restoration initiatives in Manitoba streamswith cobble and gravel beds. Brookes (1991) gives anexample of the application of this method for designing urbanchannels near London. From a practical standpoint, boundaryshear stresses can be more difficult to measure andconceptualize than velocities (Brookes 1995). Allowableshear stress criteria may be converted to allowable velocitiesby including mean depth as a parameter.

The computed shear stress values are averages for the reachin question. Average values are exceeded at points, forexample, on the outside of a bend.



If information is available regardingsediment inflows into the new chan-nel, a multiyear sediment budget canbe computed to project likely erosionand deposition and possible mainte-nance needs. Sediment load can alsobe computed, using the hydraulicproperties and bed material gradationsof the upstream supply reach and asuitable sediment transport function.The USACE software SAM (Copeland1994) includes routines that computehydraulic properties for uniform flowand sediment discharge for singlecross sections of straight channelsusing any of 13 different sedimenttransport functions. Cross sectionsmay have complex geometry andboundary materials that vary along thesection. Output can be combined witha hydrograph or a flow duration curveto obtain sediment load.

HEC-6 (USACE 1993) is a one-dimensional movable-boundary, open-channel-flow numerical model de-signed to simulate and predict changesin river profiles resulting from scourand deposition over moderate timeperiods, typically years, althoughapplications to single flood events arepossible. A continuous dischargerecord is partitioned into a series ofsteady flows of variable discharge andduration. For each discharge, a watersurface profile is calculated, providingenergy slope, velocity, depth, andother variables at each cross section.Potential sediment transport rates arethen computed at each section. Theserates, combined with the duration ofthe flow, permit a volumetric account-ing of sediment within each reach. Theamount of scour or deposition at each

section is then computed, and the crosssection geometry is adjusted for thechanging sediment volume. Computa-tions then proceed to the next flow inthe sequence, and the cycle is repeatedusing the updated cross section geom-etry. Sediment calculations are per-formed by grain size fractions, allow-ing the simulation of hydraulic sortingand armoring.

HEC-6 allows the designer to estimatelong-term response of the channel to apredicted series of water and sedimentsupply. The primary limitation is thatHEC-6 is one-dimensional, i.e., geom-etry is adjusted only in the verticaldirection. Changes in channel width orplanform cannot be simulated. An-other Federal sediment routing modelis the GSTARS 2.0 (Yang et al. 1998).GSTARS 2.0 can be used for a combi-nation of subcritical and supercriticalflow computations without interrup-tion in a semi-two-dimensional man-ner. The use of stream tube concept insediment routing enables GSTARS 2.0to simulate channel geometry changesin a semi-three-dimensional manner.

The amount and type of sedimentsupplied to a stream channel is animportant consideration in restorationbecause sediment is part of the balance(i.e., between energy and materialload) that determines channel stability.A general lack of sediment relative tothe amount of stream power, shearstress, or energy in the flow (indexesof transport capacity) usually results inerosion of sediment from the channelboundary of an alluvial channel.Conversely, an oversupply of sedimentrelative to the transport capacity of theflow usually results in deposition ofsediment in that reach of stream.



Bed material sediment transportanalyses are necessary whenever arestoration initiative involves recon-structing a length of stream exceedingtwo meander wavelengths. A recon-struction that modifies the size of across section and the sinuosity forsuch a length of channel should beanalyzed to ensure that upstreamsediment loads can be transportedthrough the reconstructed reach withminimal deposition or erosion. Differ-ent storm events and the averageannual transported bed material loadalso should be examined.

Sediment Discharge Functions

The selection of an appropriate dis-charge formula is an important consid-eration when attempting to predictsediment discharge in streams. Nu-merous sediment discharge formulashave been proposed, and extensivesummaries are provided by Alonso(1980), Brownlie (1981), Yang (1996),Bathurst (1985), Gomez and Church(1989), Parker (1990), and Garbrechtet al. (1995).

Sediment discharge rates depend onflow velocity; energy slope; watertemperature; size, gradation, specificgravity, and shape of the bed materialand suspended-sediment particles;channel geometry and pattern; extentof bed surface covered by coarsematerial; rate of supply of fine mate-rial; and bed configuration. Large-scale variables such as hydrologic,geologic, and climatic conditions alsoaffect the rate of sediment transport.Because of the range and number ofvariables, it is not possible to select asediment transport formula that satis-factorily encompasses all the condi-

tions that might be encountered. Aspecific formula might be more accu-rate than others when applied to aparticular river, but it might not beaccurate for other rivers.

Selection of a sediment transportformula should include the followingconsiderations (modified from Yang1996):

• Type of field data available ormeasurable within time, bud-get, and work hour limitations.

• Independent variables that canbe determined from availabledata.

• Limitations of formulas versusfield conditions.

If more than one formula can be used,the rate of sediment discharge shouldbe calculated using each formula. Theformulas that best agree with availablemeasured sediment discharges shouldbe used to estimate the rate of sedi-ment discharge during flow conditionswhen actual measurements are notavailable.

The following formulas may be con-sidered in the absence of any measuredsediment discharges for comparison:

• Meyer-Peter and Muller (1948)formula when the bed materialis coarser than 5 mm.

• Einstein (1950) formula whenbed load is a substantial part ofthe total sediment discharge.

• Toffaleti (1968) formula forlarge sand-bed rivers.

• Colby (1964) formula forrivers with depths less than 10feet and median bed materialvalues less than 0.8 mm.



• Yang (1973) formula for fine tocoarse sand-bed rivers.

• Yang (1984) formula for graveltransport when most of the bedmaterial ranges from 2 to 10mm.

• Ackers and White (1973) orEngelund and Hansen (1967)formula for sand-bed streamshaving subcritical flow.

• Laursen (1958) formula forshallow rivers with fine sand orcoarse silt.

Available sediment data from a gagingstation may be used to develop anempirical sediment discharge curve inthe absence of a satisfactory sedimentdischarge formula, or to verify thesediment discharge trend from aselected formula. Measured sedimentdischarge or concentration should beplotted against streamflow, velocity,slope, depth, shear stress, streampower, or unit stream power. Thecurve with the least scatter and sys-tematic deviation should be selected asthe sediment rating curve for thestation.

Sediment Budgets

A sediment budget is an accounting ofsediment production in a watershed. Itattempts to quantify processes oferosion, deposition, and transport inthe basin. The quantities of erosionfrom all sources in a watershed areestimated using various procedures.Typically, the tons of erosion from thevarious sources are multiplied bysediment delivery ratios to estimatehow much of the eroded soil actuallyenters a stream. The sediment deliv-ered to the streams is then routed

through the watershed.

The sediment routing procedureinvolves estimating how much of thesediment in the stream ends up beingdeposited in lakes, reservoirs, wet-lands, or floodplains or in the streamitself. An analysis of the soil texturesby erosion process is used to convertthe tons of sediment delivered to thestream into tons of silt and clay, sand,and gravel. Sediment transport pro-cesses are applied to help make deci-sions during the sediment routinganalysis. The end result is the sedi-ment yield at the mouth of the water-shed or the beginning of a projectreach.

Table 8.5 is a summary sedimentbudget for a watershed. Note that theinformation in the table may be frommeasured values, from estimates basedon data from similar watersheds, orfrom model outputs (AGNPS,SWRRBWQ, SWAT, WEPP, RUSLE,and others. Contact the NRCS Na-tional Water and Climate Data Centerfor more information on these mod-els). Sediment delivery ratios aredetermined for watershed drainageareas, based on sediment gauge dataand reservoir sedimentation surveys.

The watershed is subdivided intosubwatersheds at points where signifi-cant sediment deposition occurs, suchas at bridge or road fills; where streamcrossings cause channel and floodplainconstrictions; and at reservoirs, lakes,significant flooded areas, etc. Sedi-ment budgets similar to the table areconstructed for each subwatershed sothe sediment yield to the point ofdeposition can be quantified.

A sediment budget has many uses,including identification of sediment



sources for treatment (Figure 8.34). Ifthe goal for a restoration initiative is toreduce sedimentation from a water-shed, it is critical to know what type oferosion is producing the most sedi-ment and where that erosion is occur-ring. In stream corridor restoration,sediment yield (both in terms ofquantity and average grain size diam-eter) to a stream and its floodplainneed to be identified and considered indesigns. In channel stability investiga-tions, the amount of sand and gravel

sediment entering the stream from thewatershed needs to be quantified torefine bed material transport calcula-tions.

Example of a Sediment Budget

A simple application of a sedimenttransport equation in a field situationillustrates the use of a sediment bud-get. Figure 8.35 shows a streamreach being evaluated for stabilityprior to developing a stream corridorrestoration plan. Five representativechannel cross sections (A, B, C, D,and E) are surveyed. Locations of thecross sections are selected to representthe reach above and below the pointswhere tributary streams, D and E,enter the reach. Additional crosssections would need to be surveyed ifthe stream at A, B, C, D or E is nottypical of the reach.

An appropriate sediment transportequation is selected, and the transportcapacity at each cross section for bedmaterial is computed for the same

Figure 8.34: Erodedupland area.Upland sediment sourcesshould be identified in asediment budget.

Table 8.5: Example ofa sediment budget fora watershed.

ErosionSource

ProtectionLevel

AcresorMiles

Average Erosion Rate (tons/acre/year or tons/bank mile/year)

AnnualErosion(tons/year)

SedimentDeliveryRatio (percent)

SedimenttoStreams

SedimentDepositedUplands &Floodplains(tons/year)

Sediment Deliveredto Blue Stem Lake

(tons/year)

(percent)

Sheet, rill, and ephemeral gully

Adequate Cropland 6000 3.0 18,000 30 5400 14,380 3620 33.7

Inadequate Cropland 1500 6.5 9750 30 2930 7790 1960 18.3

Adequate Pasture/hayland 3400 1.0 3400 20 680 2940 460 4.3

Inadequate Pasture/hayland 600 6.0 3600 20 720 3120 480 4.5

Adequate Forestland 1200 0.5 600 20 120 520 80 0.7

Inadequate Forestland 300 5.5 1650 20 330 1430 220 2.1

Adequate Parkland 700 1.0 700 30 210 560 140 1.3

Inadequate Parkland 0 0 0 30 0 0 0 0.0

Adequate Other 420 2.0 840 20 170 730 110 1.0

Inadequate Other 0 0 0 20 0 0 0 0.0

Classic gully N/A N/A 600 40 240 440 160 1.5

Streambank

Slight 14 50 100 700 5400 140 560 5.2

Moderate 10.5

3.5

150 1580 100 1580 320 1260 11.7

Severe 600 2100 100 2100 420 1680 15.7

Total erosion 43,520 Total sediment toBlue Stem Lake

10,730



flow conditions. Figure 8.35 shows thesediment loads in the stream and thetransport capacities at each point.

The transport capacities at each pointare compared to the sediment load ateach point. If the bed material loadexceeds the transport capacity, deposi-tion is indicated. If the bed materialtransport capacity exceeds the coarsesediment load available, erosion of thechannel bed or banks is indicated.

Figure 8.35 compares the loads andtransport capacities within the reach.The stream might not be stable belowB due to deposition. The 50-tons/daydeposition is less than 10 percent ofthe total bed material load in thestream. This small amount of sedi-ment is probably within the area ofuncertainty in such analyses. Thestream below C probably is unstabledue to the excess energy (transportcapacity) causing either the banks orbottom to be eroded.

After this type of analysis is complete,the stream should be inspected forareas where sediment is building up orwhere the stream is eroding. If theseproblem areas do not match the predic-tions from the calculations, the sedi-ment transport equation may be inap-propriate, or the sediment budget, thehydrology, or the channel surveys maybe inaccurate.

Single Storm Versus Average AnnualSediment Discharge

The preceeding example predicts theamount of erosion and deposition thatcan be expected to occur over one dayat one discharge. The bed materialtransport equation probably used onegrain size of sediment. In reality, avariety of flows over varying lengthsof time move a variety of sedimentparticle sizes. Two other approachesshould be used to help predict thequantity of bed material sediment

500

500tons/day

750tons/day 700 tons/dayBed material load transport capacity at C 900 tons/day

Bed material load transport capacity at D 150 tons/day

Bed material load transport capacity at E 250 tons/day

Bed material load transport capacity at B 500 tons/day

Bed material load transport capacity at A 400 tons/day

150 tons erosion below C (750 - 900 = -150 tons)

Transport capacity at C 900 tons

Load to C 500 tons transported below B+ 250 tons from tributary E750 tons to C

50 tons deposition below B (550 - 500 = 50)

Transport capacity at B 500 tons

Load to B 400 tons transported below A+ 150 tons from tributary D550 tons to B

Transport capacity at A 400 tons

tributary D

tributary E

Note:Numbers representtons/day bed materialload in stream.

cross-section B

cross-sectionC

cross-sectionA

cross-sectionD

cross-sectionE

400 tons/day400 tons/day

150to

ns/d

ay

25

0tons/day

tons /day

Bed Material Load Routing Computations

Figure 8.35: Sedimentbudget.Stream reaches shouldbe evaluated for stabilityprior to developing arestoration plan.



transported by a stream during a singlestorm event or over a typical runoffyear.

To calculate the amount of sedimenttransported by a stream during a singlestorm event, the hydrograph for theevent is divided into equal-lengthsegments of time. The peak flow orthe average discharge for each seg-ment is determined. A spreadsheet canbe developed that lists the dischargesfor each segment of a hydrograph in acolumn (Table 8.6). The transportcapacity from the sediment ratingcurve for each discharge is shown inanother column (Figure 8.36). Sincethe transport capacity is in tons/day, athird column should include the lengthof time represented by each segmentof the hydrograph. This column ismultiplied by the transport capacity tocreate a final column that representsthe amount of sediment that could betransported over each segment of thehydrograph. Summing the values in

the last column shows the total bedmaterial transport capacity generatedby that storm.

Average annual sediment transport in astream can be determined using aprocedure very similar to the stormprediction. The sediment rating curvecan be developed from predictiveequations or from physical measure-ments. The annual flow durationcurve is substituted for the segmentedhydrograph. The same type of spread-sheet described above can be used, andthe sum of the values in the last col-umn is the annual sediment-transportcapacity (based on predictive equa-tions) or the actual annual sedimenttransport if the rating curve is based onmeasured data.

Sediment Discharge After Restoration

After the sediment transport analysisresults have been field-checked toensure that field conditions are accu-rately predicted, the same analyses are

F

G

E

B

C

D

A

SegmentofHydrograph

155

80

390

280

483

500

100

SegmentDischarge(ft3/s)

Column 1 Column 2

530

90

4500

1700

6000

6500

150

TransportCapacity(tons/day)

Column 3

.42

.42

.42

.42

.42

.42

.42

SegmentTime(days)

Column 4

221

38

Total tons transported over the storm 8112

1875

708

2500

2708

62

ActualTransport(tons)

Column 5

Hours

Dis

char

ge

(ft3

/s)

0 10

A B C D E F G

20 30 40 50 60 700

100

200

300

400

500

600

Table 8.6: Sedimentdischarges forsegments of ahydrograph.The amount of sedimentdischarged through areach varies with timeduring a stream flowevent.



Figure 8.36: Sedimentrating curve.A "sediment rating curve"rates the quantity ofsediment carried by aspecific stream flow at adefined point or gage.

Fig

. 8.4

Susp

end

ed S

edim

ent

(to

ns/

day

)

1 .1.2

.4.6

.81

24

68

1020

4060

8010

020

040

060

010

0050

0010

,000

24681020406080100

200

400

600

800

1000

Discharge (cfs)



repeated for the new cross sectionsand slope in a reconstructed stream orstream reach. Plans and designs maybe modified if the second analysisindicates significant deposition orerosion could occur in the modifiedreach. If potential changes in runoff orsediment yield are predicted to occurin the watershed above a potentialrestoration site, the sediment transportanalyses should be done again basedon these potential changes.

Stability ControlsThe risk of a restored channel's beingdamaged or destroyed by erosion ordeposition can be reduced if economicconsiderations permit installation ofcontrol measures. Control measuresare also required if “natural” levels ofchannel instability (e.g., meandermigration) are unacceptable in therestored reach.

In many cases, control measuresdouble as habitat restoration devicesor aesthetic features (Nunnally andShields 1985, Newbury and Gaboury1993). Control measures may becategorized as bed stabilization de-vices, bank stabilization devices, andhydrologic measures. Reviews ofcontrol measures are found in Simonsand Vanoni (1975), Senturk (1977),Petersen (1986), Chang (1988), andUS Army Corps of Engineers (1989b,1994), and are treated only brieflyhere. Haan et al. (1994) providedesign guidance for sediment controlon small watersheds. In all cases,sediment control systems should beplanned and designed with the geo-morphic evolution of the watershed inmind.



8.F Streambank Restoration

Even where streams retain relativelynatural patterns of flow and flooding,stream corridor restoration mightrequire that streambanks be tempo-rarily (years to decades) stabilizedwhile floodplain vegetation recovers.The objective in such instances is toarrest the accelerated erosion oftenassociated with unvegetated banks,and to reduce erosion to rates appro-priate for the stream system andsetting. In these situations, the initialbank protection may be providedprimarily with vegetation, wood, androck as necessary (refer to TechniquesAppendix).

In other cases, land development ormodified flows may dictate the use ofhard structures to ensure permanentstream stability, and vegetation is usedprimarily to address specific ecologi-cal deficiencies such as a lack ofchannel shading. In either case (per-manent or temporary bank stabiliza-tion), streamflow projections are used(as described in Chapter 7) to deter-mine the degree to which vegetationmust be supplemented with moreresistant materials (natural fabrics,wood, rock, etc.) to achieve adequatestabilization.

The causes of excessive erosion maybe reversible through changes in landuse, livestock management, floodplainrestoration, or water management. Insome cases, even normal rates of bankerosion and channel movement mightbe considered unacceptable due toadjacent development, and vegetationmight be used primarily to recoversome habitat functions in the vicinity

of "hard" bank stabilization measures.In either case, the considerationsdiscussed above with respect to soils,use of native plant species, etc., areapplicable within the bank zone.However, a set of specialized tech-niques can be employed to help ensureplant establishment and improvehabitat conditions.

As discussed earlier in this chapter,integration of woody vegetativecuttings, independently or in combina-tion with other natural materials, instreambank erosion control projects isgenerally referred to as soil bioengi-neering. Soil-bioengineered bankstabilization systems have not beenstandardized for general applicationunder particular flow conditions, andthe decision as to whether and how touse them requires careful consider-ation of a variety of factors. On largerstreams or where erosion is severe, aneffective approach involves a teameffort that includes expertise in soils,biology, plant sciences, landscapearchitecture, geology, engineering, andhydrology.

Soil bioengineering approaches usu-ally employ plant materials in the formof live woody cuttings or poles ofreadily sprouting species, which areinserted deep into the bank or an-chored in various other ways. Thisserves the dual purposes of resistingwashout of plants during the earlyestablishment period, while providingsome immediate erosion protectiondue to the physical resistance of thestems. Plant materials alone aresufficient on some streams or some



bank zones, but as erosive forcesincrease, they can be combined withother materials such as rocks, logs orbrush, and natural fabrics (Figure8.37). In some cases, woody debris isincorporated specifically to improvehabitat characteristics of the bank andnear-bank channel zones.

Preliminary site investigations (see

Figure 8.38) and engineering analysesmust be completed, as described inChapter 7, to determine the mode ofbank failure and the feasibility ofusing vegetation as a component ofbank stabilization work. In addition tothe technical analyses of flows andsoils, preliminary investigations mustinclude consideration of access,maintenance, urgency, and availabilityof materials.

Generalizations regarding water levelsand flow velocities should be takenonly as indications of the experiencesreported from various bank stabiliza-tion projects. Any particular site mustbe evaluated to determine how vegeta-tion can or cannot be used. Soilcohesiveness, the presence of gravellenses, ice accumulation patterns, theamount of sunlight reaching the bank,and the ability to ensure that grazingwill be precluded are all consider-ations in assessing the suitability ofvegetation to achieve bank stabiliza-tion. In addition, modified flowpatterns may make portions of thebank inhospitable to plants because ofinappropriate timing of inundationrather than flow velocities and dura-tions (Klimas 1987). The need toextend protection well beyond theimmediate focus of erosion and toprotect against flanking is an impor-tant design consideration.

As noted in Section 8.E, streambankstabilization techniques can generallybe classified as armor, indirect meth-ods, or vegetative methods. Theselection of the appropriate stabiliza-tion technique is extremely importantand can be expressed in terms of thefactors discussed below.

Figure 8.37: A stabilized streambank.Plant materials can be combined with other materials such as rocks, logs orbrush, and natural fabrics. [(a) During and (b) after.]

(a)

(b)



Careless Creek, Montana

In the Big Snowy Mountains of central Montana, Careless Creek begins to flow through rangelands andfields until it reaches the Musselshell River. At the beginning of the century, the stream was lined with ariparian cover, primarily of willow. This stream corridor was home to a diversity of wildlife such aspheasant, beaver, and deer.

In the 1930s, a large reservoir was constructed to the west with two outlets, one connected to CarelessCreek. These channels were meant to carry irrigation water to the area fields and on to the MusselshellRiver. Heavy flows during the summer months began to erode the banks (Figure 8.39a ). In the followingyears, ranchers began clearing more and more brush for pasture, sometimes burning it out along a stream.

“My Dad carried farmer’s matches in his pocket. There was a worn spot on his pants where he would strikea match on his thigh.,” said Jessie Zeier, who was raised on a ranch near Careless Creek, recalling how hisfather often cleared brush.

Any remaining willows or other species were eliminated inthe following years as ranchers began spraying riparianareas to control sagebrush. This accelerated thestreambank erosion as barren, sometimes vertical, banksbegan sloughing off chunks of salted gumbo in heavy runoff.

As demand for irrigation water increased downstream, theerosion problem continued. Careless Creek was nowcarrying not only water, but entire banks, corrals, andfences, as well as tons of silt, and salt from shale deposits,to the Musselshell River. Conflicts between landowners andwater users arose over the quantity and quality of the water.

Emotions were running high. Groups began workingtogether to resolve problems. A Technical Advisory SteeringCommittee was developed to help the planning effort. Manyorganizations took part, including the Upper and LowerMusselshell Conservation Districts; Natural ResourcesConservation Service; Montana Department of NaturalResources and Conservation; Montana Department of Fish;Wildlife and Parks; Deadman’s Basin Water UsersAssociation; U.S. Bureau of Reclamation; Central MontanaRC&D; City of Roundup; Roundup Sportsmen; countycommissioners; and local landowners.

As part of the planning effort, a geographic informationsystem resource inventory was begun in 1993. Theinventory revealed about 50 percent of the banks along the18 miles of Careless Creek were eroding. The inventoryhelped to locate the areas causing the most problems.Priority was given to headquarters, corrals, and croplands,where stabilization of approximately 5,000 feet ofstreambank has taken place, funded by EPA monies.

Passive efforts have also begun to stabilize the banks.Irrigation flows in Careless Creek have been decreased for the past 5 years, enabling some areas, such asthe one pictured, to begin to self-heal (Figure 8.39b ). Vegetation has been given a chance to root aserosion has begun to stabilize. Other practices, such as fencing, are being implemented, and futuretreatments are planned to provide a long-term solution.

Figure 8.39: Careless Creek.(a) Eroded streambank (May 1995) and(b) streambank in recovery (December1997).(a)

(b)



Effectiveness of TechniqueThe inherent factors in the propertiesof a given bank stabilization tech-nique, and in the physical characteris-tics of a proposed work site, influencethe suitability of that technique for thatsite. Effectiveness refers to the suit-ability and adequacy of the technique.Many techniques can be designed toadequately solve a specific bankstability problem by resisting erosive

forces and geotechnical failure. Thechallenge is to recognize which tech-nique matches the strength of protec-tion against the strength of attack andtherefore performs most efficientlywhen tested by the strongest process oferosion and most critical mechanismof failure. Environmental and eco-nomic factors are integrated into theselection procedure, generally makingsoil bioengineering methods veryattractive. The chosen solution,however, must first fulfill the require-ment of being effective as bank stabili-zation; otherwise, environmental andeconomic attributes will be irrelevant.Soil bioengineering can be a usefultool in controlling streambank erosion,but it should not be considered apanacea. It must be performed in ajudicious manner by personnel experi-enced in channel processes, biology,and streambank stabilization tech-niques.

Stabilization TechniquesPlants may be established on upperbank and floodplain areas by usingtraditional techniques for seeding orby planting bare-root and container-grown plants. However, these ap-proaches provide little initial resis-tance to flows, and plantings may bedestroyed if subjected to high waterbefore they are fully established.Cuttings, pole plantings, and livestakes taken from species that sproutreadily (e.g., willows) are more resis-tant to erosion and can be used loweron the bank (Figure 8.40). In addi-tion, cuttings and pole plantings canprovide immediate moderation of flowvelocities if planted at high densities.Often, they can be placed deep enoughto maintain contact with adequate soil

Figure 8.38: Erodedbank.Preliminary siteinvestigation andanalyses are critical tosuccessful streambankstabilization design.

livefascinebundle

live stake

dead stout stake driven on 2-foot centers each way, minimum length 2 1/2 feet

dead stout stake

geotextile fabric

baseflow

streambed

live stake

2 ft

wire securedto stakes brush mattress

Note:Rooted/leafed condition of the living plant material is not representative at the time of installation.

live and dead stout stake spacing2 feet on center

branchcuttings

16 gaugewire

Figure 8.40: Cuttingsystems.Details of brushmattresstechnique.Source: Chapter 16Engineering Handbook,NRCS 1997.



moisture levels, thereby eliminatingthe need for irrigation. The reliablesprouting properties, rapid growth, andgeneral availability of cuttings ofwillows and other pioneer speciesmakes them particularly appropriatefor use in bank revegetation projects,and they are used in most of theintegrated bank protection approachesdescribed here (see Figure 8.41).

Anchored Cutting Systems

Several techniques are available thatemploy large numbers of cuttingsarranged in layers or bundles, whichcan be secured to streambanks andpartially buried. Depending on howthese systems are arranged, they canprovide direct protection from erosiveflows, prevent erosion from upslopewater sources, promote trapping ofsediments, and quickly develop denseroots and sprouts. Brush mattressesand woven mats are typically used onthe face of a bank and consist ofcuttings laid side by side and interwo-ven or pinned down with jute cord orwire held in place by stakes. Brushlayers are cuttings laid on terraces duginto the bank, then buried so that thebranch ends extend from the bank.Fascines or wattles are bundles ofcuttings tied together, placed in shal-low trenches arranged horizontally onthe bank face, partially buried, andstaked in place. A similar system,called a reed roll, uses partially buriedand staked burlap rolls filled with soiland root material or rooted shoots toestablish herbaceous species in appro-priate habitats. Anchored bundles oflive cuttings also have been installedperpendicular to the channel on newlyconstructed gravel floodplain areas todissipate floodwater energy and

encourage deposition of sediment(Karle and Densmore 1994).

Geotextile Systems

Geotextiles have been used for erosioncontrol on road embankments andother upland settings, usually incombination with seeding, or withplants placed through slits in the fabric(Figure 8.42). In self-sustainingstreambank applications, only natural,biodegradable materials should beused, such as jute or coconut fiber(Johnson and Stypula 1993). The

Figure 8.42: Installationof geotextile.Geotextile should bemade of biodegradablematerials such as jute orcoconut fibers.

Figure 8.41: Results oflive staking along astreambank.Pioneer species are oftenmost appropriate for usein bank revegetationprojects.



typical streambank use for thesematerials is in the construction ofvegetated geogrids, which are similarto brush layers except that the fill soilsbetween the layers of cuttings areencased in fabric, allowing the bank tobe constructed of successive "lifts" ofsoil, alternating with brush layers.This approach allows reconstruction ofa bank and provides considerableerosion resistance (see Green Rivercase study). Natural fibers are alsoused in "fiber-schines," which are soldspecifically for streambank applica-tions. These are cylindrical fiberbundles that can be staked to a bankwith cuttings or rooted plants insertedthrough or into the material.

Vegetated plastic geogrids and othernondegradable materials can also beused where geotechnical problemsrequire drainage or additional strength.

Integrated SystemsA major concern with the use ofstructural approaches to streambankstabilization is the lack of vegetationin the zone directly adjacent to thewater. Despite a long-standing con-cern that vegetation destabilizes stonerevetments, there has been littlesupporting evidence and even someevidence to the contrary (Shields1991). Assuming that loss of convey-ance is accounted for, the addition ofvegetation to structures should beconsidered. This can involve place-ment of cuttings during construction,or insertion of cuttings and polesbetween stones on existing structures.Timber cribwalls may also be con-structed with cuttings or rooted plantsextending through the timbers fromthe backfill soils.

Trees and Logs

Tree revetments are made from wholetree trunks laid parallel to the bank,and cabled to piles or deadman an-chors. Eastern red cedar (Juniperusvirginiana) and other coniferous treesare used on small streams, where theirspringy branches provide interferenceto flow and trap sediment. The princi-pal objective to these systems is theuse of large amounts of cable and thepotential for trees to be dislodged andcause downstream damage.

Some projects have successfully usedlarge trees in conjunction with stone toprovide bank protection as well asimproved aquatic habitat (see casestudy). Large logs with intact rootwads are placed in trenches cut intothe bank, such that the root wadsextend beyond the bank face at the toe(Figure 8.43). The logs are over-

thalweg channel

baseflow

streambed

rootwad

8- to 12-footlength

existing vegetation, plantings orsoil bioengineering systems

diameter of log =16-in min.

footer log

boulder 1 1/2 timesdiameter of log

Figure 8.43: Revetmentsystem.Details of rootwad andboulder technique.Source: Chapter 16Engineering Handbook,NRCS 1997.



lapped and/or braced with stone toensure stability, and the protrudingrootwads effectively reduce flowvelocities at the toe and over a rangeof flow elevations (Figure 8.44). Amajor advantage of this approach isthat it reestablishes one of the naturalroles of large woody debris in streamsby creating a dynamic near-bankenvironment that traps organic mate-rial and provides colonization sub-strates for invertebrates and refugehabitats for fish. The logs eventuallyrot, resulting in a more natural bank.The revetment stabilizes the bank untilwoody vegetation has matured, atwhich time the channel can return to amore natural pattern.

In most cases, bank stabilizationprojects use combinations of thetechniques described above in anintegrated approach. Toe protectionoften requires the use of stone, butamounts can be greatly reduced iflarge logs can also be used. Likewise,stone blankets on the bank face can bereplaced with geogrids or supple-mented with interstitial plantings.Most upper bank areas can usually bestabilized using vegetation alone,although anchoring systems might berequired. The Green River bankrestoration case study illustrates onesuccessful application of an integratedapproach on a moderate-sized river inWashington State.

Figure 8.44: Installationof logs with intact rootwads.An advantage to usingtree revetments is thecreation of habitat forinvertebrates and fishalong the streambank.



Green River Bank Restoration Initiative, King County, Washington

The King County, Washington, Surface Water Management Division initiated a bank restoration initiative in1994 that illustrates a variety of project objectives and soil bioengineering approaches (Figure 8.45 ). Theproject involved stabilization of the bank of the Green River along a 500-foot section of a meander bend

that was rapidly migrating into the adjacent farm field.The project objectives included improvement of fishand wildlife habitat, particularly for salmonids.

Site investigations included surveys of stream crosssections, velocity measurements at two dischargelevels, soil characterizations, and assessment of fishuse of existing habitat features in the area. Thestreambank was vertical, 5 to 10 feet high, andcomposed of silty-clay-loam alluvium with gravellenses. Flow velocities were 2 to 5 fps for flows of200 and 550 cfs. Fish were primarily observed inareas of low velocities and/or near woody debris, andalong the channel margins.

In August, large woody debris was installed along thetoe of the bank. The logs were cedar and fir, 25 feetlong and 28 to 36 inches in diameter, with root wads6 to 8 feet in diameter. The logs were placed intrenches cut 15 feet back into the bank so that theroot wads extended into the channel, and large (3- to4-foot diameter) boulders were placed among thelogs at the toe. Log and boulder placement wasdesigned to interlock and brace the logs and preventmovement. The project used approximately 10 logsand 20 boulders per 100 lineal feet of bank. InSeptember, vegetated geogrids were installed abovethe toe zone to stabilize the high bank (Figure 8.46) .The project was completed with installation of avariety of plants, including container-grown conifersand understory species, in a minimum 25-foot bufferalong the top of the bank.

Within 2 months of completion, the site wassubjected to three high flows, including an 8,430-cfsevent in December 1994. Measured velocities alongthe bank were less than 2 fps at the surface and lessthan 1 fps 2 feet below the surface, indicating theeffectiveness of the root wads in moderating flowvelocities (Figure 8.47) . Some surface erosion andwashout of plants along the top bank occurred, and asubsequent event caused minor damage to thegeogrid at one location. The maintenance repairsconsisted of replanting and placement of additional

logs to halt undermining of the geogrid. The 1995 growing season produced dramatic growth of the willowcuttings in the geogrid, although many of the planted trees in the overbank zone died (Figure 8.48) . Initialobservations have documented extensive fish use of the slow-water habitats among the root wads at thetoe of the bank, and in scour holes created by flows deflected toward the channel bottom.

Typical Cross-Section of Restored BankSection View

Typical Detail – Log PatternPlan View

Figure 8.45: Construction details.Source: King County Surface Water Management Division.

(a)

(b)



Figure 8.47: Completed system.Note calm water along bankline during highflow.

Figure 8.48: Completed system after one year.Note dramatic willow growth from vegetated geogrid.

The site continues to be carefully monitored, and theeffectiveness of the approach has led to theimplementation of similar designs elsewhere in theregion. The project designers have concluded that futureprojects of this type should use small plants rather thanlarge rooted material in the overbank zone to reducecosts, improve survival, and minimize damage due toequipment access for maintenance or repair. Based ontheir observations of fish response along the restoredbank and in nearby stream reaches, they alsorecommend that future projects incorporate a greatervariety of woody debris, including brushy material andtree tops, along the toe and lower bank.Figure 8.46: Partially installed vegetated geogrid.

Installed above the toe to stabilize high bank.



8.G Instream Habitat Recovery

As described in Chapter 2, habitat isthe place where a population lives andincludes living and nonliving compo-nents. For example, fish habitat is aplace, or set of places, in which asingle fish, a population, or an assem-blage of fish can find the physical,chemical, and biological featuresneeded for life, including suitablewater quality, passage routes, spawn-ing grounds, feeding and resting sites,and shelter from predators and adverseconditions (Figure 8.49). Principalfactors controlling the quality of theavailable aquatic habitat include:

• Streamflow conditions.

• Physical structure of thechannel.

• Water quality (e.g., tempera-ture, pH, dissolved oxygen,turbidity, nutrients, alkalinity).

• The riparian zone.

• Other living components.

The existing status of aquatic habitatswithin the stream corridor should beassessed during the planning stage(Part II). Design of channels, struc-tures, or restoration features can beguided and fine tuned by assessing thequality and quantity of habitats pro-vided by the proposed design. Addi-tional guidance on assessing thequantity and quality of aquatic habitatis provided in Chapter 7.

This section discusses the design ofinstream habitat structures for thepurpose of enhancing physical aquatichabitat quality and quantity. It shouldbe noted, however, that the best ap-proach to habitat recovery is to restorea fully functional, well-vegetatedstream corridor within a well-managed

Figure 8.49: Instreamhabitat.Suitable water quality,passage routes, andspawning grounds aresome of thecharacteristics of fishhabitat.



watershed. Man-made structures areless sustainable and rarely as effectiveas a stable channel. Over the longterm, design should rely on naturalfluvial processes interacting withfloodplain vegetation and associatedwoody debris to provide high-qualityaquatic habitat. Structures have littleeffect on populations that are limitedby factors other than physical habitat.

Instream Habitat FeaturesThe following procedures to restoreinstream habitat are adapted fromNewbury and Gaboury (1993) andGarcia (1995).

• Select stream. Give priority toreaches with the greatestdifference between actual(low) and potential (high) fishcarrying capacity and with ahigh capacity for naturalrecovery processes.

• Evaluate fish populations andtheir habitats. Give priority toreaches with habitats andspecies of special interest. Isthis a biological, chemical, orphysical problem? If a physicalproblem:

• Diagnose physical habitatproblems.

-Drainage basin. Trace water-shed lines on topographicaland geological maps to identifysample and rehabilitationbasins.

-Profiles. Sketch main stemand tributary long profiles toidentify discontinuities thatmight cause abrupt changes instream characteristics (falls,former base levels, etc.).

-Flow. Prepare flow summaryfor rehabilitation reach usingexisting or nearby records ifavailable (flood frequency,minimum flows, historicalmass curve). Correct fordrainage area differences.Compare magnitude andduration of flows duringspawning and incubation toyear class strength data todetermine minimum andmaximum flows required forsuccessful reproduction.

-Channel geometry survey.Select and survey samplereaches to establish the rela-tionship between channelgeometry, drainage area, andbankfull channel-formingdischarge (Figure 8.50).Quantify hydraulic parametersat design discharge.

-Rehabilitation reach survey.Survey rehabilitation reachesin sufficient detail to preparechannel cross section profilesand construction drawings andto establish survey referencemarkers.

Figure 8.50: Surveyinga stream.Channel surveysestablish baselineinformation needed forrestoration design.

Man-madestructuresare lesssustainableand rarelyas effectiveas a stablechannel.



-Preferred habitat. Prepare asummary of habitat factors forbiologically preferred reachesusing regional references andsurveys. Identify multiplelimiting factors for the speciesand life stages of greatestconcern. Where possible,undertake reach surveys inreference streams with provenpopulations to identify localflow conditions, substrate,refugia, etc.

• Design a habitat improvementplan. Quantify the desiredresults in terms of hydraulicchanges, habitat improvement,and population increases.Integrate selection and sizingof rehabilitation works withinstream flow requirements.

-Select potential schemes andstructures that will be rein-forced by the existing streamdynamics and geometry. Thefollowing section providesadditional detail on use ofhabitat structures.

-Test designs for minimum andmaximum flows and set targetflows for critical periodsderived from the historicalmass curve.

• Implement planned measures.

-Arrange for on-site locationand elevation surveys andprovide advice for finishingdetails in the stream.

• Monitor and evaluate results.

-Arrange for periodic surveysof the rehabilitated reach andreference reaches, to improvethe design, as the channel ages.

Instream Habitat StructuresAquatic habitat structures (also calledinstream structures and stream im-provement structures) are widely usedin stream corridor restoration. Com-mon types include weirs, dikes, ran-dom rocks, bank covers, substratereinstatement, fish passage structures,and off-channel ponds and coves.Institutional factors have favored theiruse over more holistic approaches torestoration. For example, it is ofteneasier to obtain authority and fundingto work within a channel than toinfluence riparian or watershed landuse. Habitat structures have been usedmore along cold water streams sup-porting salmonid fisheries than alongwarm water streams, and the volumi-nous literature is heavily weightedtoward cold water streams.

In a 1995 study entitled Stream Habi-tat Improvement Evaluation Project,1,234 structures were evaluated ac-cording to their general effectiveness,the habitat quality associated with thegiven structure type, and actual use ofthe structures by fish (Bio West 1995).The study determined approximately18 percent of the structures needmaintenance. Where inadequate flowsand excessive sediment delivery occur,structures have a brief lifespan andlimited value in terms of habitatimprovement. Furthermore, the studyconcluded that instream habitat struc-tures generally provided increased fishhabitat.

Before structural habitat features areadded to a stream corridor restorationdesign, project managers shouldcarefully determine whether theyaddress the real need and are appropri-ate. Major caveats include the follow-ing:



• Structures should never beviewed as a substitute for goodriparian and upland manage-ment.

• Defining the ecological pur-pose of a structure and siteselection are as important asconstruction technique.

• Scour and deposition arenatural stream processesnecessary to create fish habitat.Overstabilization thereforelimits habitat potential,whereas properly designed andsited structures can speedecological recovery.

• Use of native materials (stoneand wood) is strongly encour-aged.

• Periodic maintenance ofstructures will be necessaryand must be incorporated intoproject planning.

Instream Habitat StructureDesignDesign of aquatic habitat structuresshould proceed following the stepspresented below (Shields 1983).However, the process should beviewed as iterative, and considerablerecycling among steps should beexpected.

• Plan layout.

• Select types of structures.

• Size the structures.

• Investigate hydraulic effects.

• Consider effects on sedimenttransport.

• Select materials and designstructures.

Each step is described below. Con-struction and monitoring follow-upactivities are described in Chapter 9.

Plan Layout

The location of each structure shouldbe selected. Avoid conflicts withbridges, riparian structures, andexisting habitat resources (e.g., standsof woody vegetation). The frequencyof structures should be based on thehabitat requirements previouslydetermined, within the context of thestream morphology and physicalcharacteristics (see Chapter 7). Careshould be taken to place structureswhere they will be in the water duringbaseflow. Structures should be spacedto avoid large areas of uniform condi-tions. Structures that create poolsshould be spaced five to seven channelwidths apart. Weirs placed in seriesshould be spaced and sized carefullyto avoid placing a weir within thebackwater zone of the downstreamstructure, since this would create aseries of pools with no interveningriffles or shallows.

Select Types of Structures

The main types of habitat structuresare weirs, dikes (also called jetties,barbs, deflectors (Figure 8.51), spurs,etc.), random rocks (also called boul-ders), and bank covers (also calledlunkers). Substrate reinstatement(artificial riffles), fish passage struc-tures, and off-channel ponds and coveshave also been widely employed. Factsheets on several of these techniquesare provided in the Techniques Appen-dix, and numerous design web sitesare available (White and Brynildson1967, Seehorn 1985, Wesche 1985,

See Chapter 1,Section C for anintroduction tostormflow andbaseflow.

REVERSE REVERSE FAST FORWARD

Preview Chapter 9,for an introduction toconstruction andmonitoring follow-upactivities.



Orsborn et al. 1992, Orth and White1993, Flosi and Reynolds 1994).

Evidence suggests that traditionaldesign criteria for widespread bankand bed stabilization measures (e.g.,concrete grade control structures,homogeneous riprap) can be modified,with no functional loss, to better meetenvironmental objectives and improvehabitat diversity. Table 8.8 may beused as a general guide to relatestructural type to habitat requirement.Weirs are generally more failure-pronethan deflectors. Deflectors and ran-dom rocks are minimally effective inenvironments where higher flows donot produce sufficient local velocitiesto produce scour holes near structures.Random rocks (boulders) are espe-cially susceptible to undermining andburial when placed in sand-bed chan-nels, although all types of stonestructures experience similar prob-lems. Additional guidance for evaluat-ing the general suitability of variousfish habitat structures for a wide rangeof morphological stream types isprovided by Rosgen (1996). Seehorn(1985) provides guidance for smallstreams in the eastern United States.The use of any of these guides shouldalso consider the relative stability ofthe stream, including aggradation andincision trends, for final design.

Size the Structures

Structures should be sized to producethe desired aquatic habitats at thenormal range of flows from baseflowto bankfull discharge. A hydrologicalanalysis can provide an estimate of thenormal range of flows (e.g., a flowduration curve), as well as an estimate

of extreme high and low flows thatmight be expected at the site (seeChapter 7). In general, structuresshould be low enough that their effectson the water surface profile will beslight at bankfull discharge. Detailedguidance by structural type is pre-sented in the Techniques Appendix.For informal design, empirical equa-tions like those presented by Heiner(1991) can be used to roughly estimatethe depth of scour holes at weirs anddikes.

Figure 8.51: Instreamhabitat structure.Wing deflector habitatstructure.Source: Chapter 16Engineering Handbook,NRCS 1997.

existing bank

8-12 feettop width

ig.

2:12:1

2:1

Cross Section not to scale

1:1

1:1

length of jetty(varies)

rock riprap

Front Elevation not to scale

design flow

baseflowstreambed

key into streambed,approx. D100



Channel Type

Low St. Check Dam

Medium St. Check Dam

Boulder Placement

Bank Boulder Placement

Single Wing Deflector

Double WingDeflector

Channel Constrictor

Bank Cover

Channel Type

Half Log Cover

Floating Log Cover

Meander Straight

Migration Barrier

"V" Shaped Log

Gravel Placement

A1 N/A N/A N/A N/A N/A N/A N/A N/A

A2 N/A N/A N/A N/A N/A N/A N/A N/A

B1-1 Poor Poor Good Excellent Poor Poor Poor Good

B1 Excellent Excellent N/A N/A Excellent Excellent N/A Excellent

B2 Excellent Good Excellent Excellent Excellent Excellent Excellent Excellent

B3 Fair Poor Poor Good Poor Poor Poor Poor



C1-1 Poor Poor Fair Excellent Poor Poor Poor Good

C1 Good Fair Fair Excellent Good Good Fair Good

C2 Excellent Good Good Excellent Good Excellent Excellent Good

C3 Fair Poor Poor Good Fair Fair Fair Good

C4 Fair Poor Poor Good Poor Poor Poor Fair

C5 Fair Poor Poor Good Poor Poor Poor Poor

C6 N/A N/A N/A N/A N/A N/A N/A N/A

D1 Fair Poor Poor Fair Fair Fair Fair Poor

D2 Fair Poor Poor Fair Fair Fair

Submerged Shelter Gravel Traps

Fair Poor

A1 N/A N/A N/A N/A Excellent Good Poor Poor

A2 N/A N/A N/A N/A Excellent Excellent Excellent Poor

B1-1 Good Good Good Excellent Fair Good Good Fair

B1 Good Excellent Excellent Excellent Excellent Excellent Excellent Fair

B2 Excellent Excellent Good Excellent Good Good Good Good

B3 Poor Fair Fair Fair Poor Poor Poor Poor



C1-1 Good Good Good Excellent Poor Fair Fair Fair

C1 Good Good Good Excellent Poor Fair Good Fair

C2 Good Excellent Excellent Excellent Poor Good Excellent Excellent

C3 Fair Good Fair Good Poor N/A N/A N/A

C4 Poor Good Fair Good Poor Poor Poor Poor

C5 Poor Good Fair Good Poor Poor Poor Poor

C6 N/A N/A N/A N/A Poor Poor Fair Fair

D1 Poor Poor Poor Poor Poor Poor N/A Poor

D2 Poor Poor Poor Poor Poor N/A Poor Poor

Key:

Excellent - No limitation to location of structure placement or special modification in design.

Good - Under most conditions, very effective. Minor modification of design or placement required.

Fair - Serious limitation which can be overcome by placement location, design modification or stabilization techniques. Generally not recommended due to difficulty of offsetting potential adverse consequences and high probability of reduced effectiveness.

Poor - Not recommended due to morphological character of stream type and very low probability of success.

Not Applicable- Generally not considered since habitat components are not limiting.

Note : A3, A3-a, A4, A4-a, A5, A5-a channel types are not evaluated due to limited fisheries value.

Table 8.9: Fish habitat improvement structures —suitability for stream types.From Rosgen 1996



Investigate Hydraulic Effects

Hydraulic conditions at the designflow should provide the desiredhabitat; however, performance shouldalso be evaluated at higher and lowerflows. Barriers to movement, such asextremely shallow reaches or verticaldrops not submerged at higher flows,should be avoided. If the conveyanceof the channel is an issue, the effect ofthe proposed structures on stages athigh flow should be investigated.Structures may be included in astandard backwater calculation modelas contractions, low weirs, or in-creased flow resistance (Manning)coefficients, but the amount of in-crease is a matter of judgment orlimited by National Flood InsuranceProgram ordinances. Scour holesshould be included in the channelgeometry downstream of weirs anddike since a major portion of the headloss occurs in the scour hole. Hy-draulic analysis should include estima-tion or computation of velocities orshear stresses to be experienced by thestructure.

Consider Effects on SedimentTransport

If the hydraulic analysis indicates ashift in the stage-discharge relation-ship, the sediment rating curve of therestored reach may change also,leading to deposition or erosion.Although modeling analyses areusually not cost-effective for a habitatstructure design effort, informalanalyses based on assumed relation-ships between velocity and sedimentdischarge at the bankfull dischargemay be helpful in detecting potentialproblems. An effort should be madeto predict the locations and magnitude

of local scour and deposition. Areasprojected to experience significantscour and deposition should be primesites for visual monitoring after con-struction.

Select Materials

Materials used for aquatic habitatstructures include stone, fencing wire,posts, and felled trees. Priority shouldbe given to materials that occur on siteunder natural conditions. In somecases, it may be possible to salvagerock or logs generated from construc-tion of channels or other projectfeatures. Logs give long service ifcontinuously submerged. Even logsnot continuously wet can give severaldecades of service if chosen fromdecay-resistant species. Logs andtimbers must be firmly fastenedtogether with bolts or rebar and mustbe well anchored to banks and bed.Stone size should be selected based ondesign velocities or shear stress.



8.H Land Use Scenarios

As discussed in Chapter 3, moststream corridor degradation is directlyattributable to land use practices and/or hydrologic modifications at thewatershed level that cause fundamen-tal disruption of ecosystem functions(Beschta et al. 1994) (Figure 8.52).Ironically, land use practices, includ-ing hydrologic modifications, canoffer the opportunity for restoringthese same degraded stream corridors.Where feasible, the objective of therestoration design should be to elimi-nate or moderate disruptive influencessufficiently to allow recovery ofdynamic equilibrium over time (NRC1992).

If chronic land use impacts on thestream or riparian system cannot becontrolled or moderated, or if someelements of the stream network (e.g.,headwaters) are not included in therestoration design, it must be recog-nized that the restoration action mayhave limited effectiveness in the long-term.

Restoration measures can be designedto address particular, site-specificdeficiencies (an eroding bank, habitatfeatures), but if they do not restoreself-maintaining processes and thefunctions of a stream corridor, theymust be regarded as a focused “fix”rather than an ecosystem restoration.In cases where land use practices arethe direct cause of stream corridordegradation and there is a continuingdownward trend in landscape condi-tion, there is little point in expendingresources to address symptoms of the

problem rather than the problem itself(DeBano and Schmidt 1989).

Design Approaches forCommon EffectsAgriculture, forestry, grazing, mining,recreation, and urbanization are someof the principal land uses that canresult in disturbance of stream corridorstructure and functions. A watershedanalysis will help prioritize and coor-dinate restoration actions (Platts andRinne 1985, Swanson 1989) and mayindicate critical or chronic land useactivities causing disturbance bothinside and outside the stream corridor.Addressing these in the restorationplan and design, may greatly improvethe effectiveness and success ofrestoration work.

Restoration measures designed inresponse to these effects may besimilar across land uses. Sedimentand nutrient management in urban,agricultural, and forest settings, for

Figure 8.52: Sediment-laden stream.Most stream corridordegradation can beattributed to impactsresulting fromsurrounding land uses.



instance, may require the use of bufferstrips. Although the buffer strips havemany common design characteristics,each setting has site-specific factors.

Dams

Dams alter the flow of water, sedi-ment, organic matter, and nutrients,resulting in both direct physical andindirect biological effects in tailwatersand downstream riparian and flood-plain areas (see Chapter 3). Streamcorridors below dams can be partiallyrestored by modifying operation andmanagement approaches. Impactsfrom the operation of dams on surfacewater quality and aquatic and riparianhabitat should be assessed and thepotential for improvement evaluated.The modification of operation ap-proaches, where possible, in combina-tion with the application of properlydesigned and applied best manage-ment practices, can reduce the impactscaused by dams on downstreamriparian and floodplain habitats.

Best management practices can beapplied individually or in combinationto protect and improve surface waterquality and aquatic habitat in reser-voirs as well as downstream. Severalapproaches have been designed forimproving or maintaining acceptablelevels of dissolved oxygen (DO),temperature, and other constituents inreservoirs and tailwaters. One designapproach uses pumps, air diffusers, orair lifts to induce circulation andmixing of the oxygen-poor but coldhypolimnion with the oxygen-rich butwarm epilimnion, resulting in a morethermally uniform reservoir withincreased DO. Another design ap-proach for improving water quality in

tailwaters for trout fisheries involvesmixing of air or oxygen with waterpassing through the turbines at hydro-power dams to improve concentrationsof DO. Reservoir waters can also beaerated by venting turbines to theatmosphere or by injecting compressedair into the turbine chamber (USEPA1993).

Modification to the intakes, the spill-way, or the tailrace of a dam can alsobe designed to improve temperature orDO levels in tailwaters. Installingvarious types of weirs downstream ofa dam achieves similar results. Thesedesign practices rely on agitation andturbulence to mix reservoir releaseswith atmospheric air to increase levelsof DO (USEPA 1993).

Adequate fish passage around dams,diversions, and other obstructions maybe a critically important component ofrestoring healthy fish populations topreviously degraded rivers andstreams. A fact sheet in Appendix Ashows an example for fish passages.However, designing, installing, andoperating fish passage facilities atdams are beyond the scope of thishandbook. Further, the type of fishpassage facility and the flows neces-sary for operation are generally sitespecific. Further information on fishpassage technology can be found inother references, including Environ-mental Mitigation at HydroelectricProjects - Volume II. Benefits andCosts of Fish Passage and Protection(Francfort et al., 1994); and FishPassage Technologies: Protection atHydropower Facilities (Office ofTechnology Assessment, Congress ofthe United State, Washington DC,OTA-ENV-641).



Adjusting operation procedures atsome dams can also result in improvedquality of reservoir releases anddownstream conditions. Partialrestoration of stream corridors belowdams can be achieved by designingoperation procedures that mimic thenatural hydrograph, or desirableaspects of the hydrograph. Modifica-tions include scheduling releases orthe duration of shutoff periods, insti-tuting procedures for the maintenanceof minimum flows, and making sea-sonal adjustments in pool levels and inthe timing and variation of the rates ofdrawdowns (USEPA 1993).

Modifying operation and managementapproaches, in combination with theapplication of properly designed bestmanagement practices, can be aneffective approach to partially restor-ing stream corridors below dams.However, dam removal is the onlyway to begin to fully restore a streamto its natural condition. It is impor-tant to note, however, that unlessaccomplished very carefully, withsufficient studies and modeling and atsignificant cost, removing a dam cancause more damage downstream (andupstream) than the dam is currentlycausing until a state of dynamicequilibrium is reached. Dam removallowers the base level of upstreamtributaries, which can cause rejuvena-tion, bed and bank instability, andincreased sediment loads. Damremoval can also result in the loss ofwetlands and habitat in the reservoirand tributary deltas.

Three options should be considered—complete removal, partial removal,and staged breaching. The option isselected based on the condition of the

dam and future maintenance requiredif not completely removed, and on thebest way to deal with the sedimentnow stored behind the dam. Thefollowing elements must be consideredin managing sediment:

• Removing features of damsnecessary to restore fish pas-sage and ensure safety.

• Revegetation of the reservoirareas.

• Long-term monitoring ofsediment transport and riverchannel topography, waterquality, and aquatic ecology.

• Long-term protection of mu-nicipal and industrial watersupplies.

• Mitigation of flood impactscaused by long-term riveraggradation.

• Quality of sediment, includingidentification of the lateral andvertical occurrence of toxic orotherwise poor-quality sedi-ment.

Water quality issues are primarilyrelated to suspended sediment concen-tration and turbidity. These areimportant to municipal, industrial, andprivate water users, as well as toaquatic communities. Water qualitywill primarily be affected by any siltand clay released from the reservoirsand by reestablishment of the naturalsediment loads downstream. Duringremoval of the dam and draining of thelake, the unvegetated reservoir bot-toms will be exposed. Lakebeds willbe expected to have large woodydebris and other organic material. Arevegetation program is necessary tocontrol dust, surface runoff, and



erosion and to restore habitat andaesthetic values. A comprehensivesediment management plan is neededto address the following:

• Sediment volume and physicalproperties.

• Sediment quality and associ-ated disposal requirements.

• Hydraulic and biologicalcharacteristics of the reservoirand downstream channel.

• Alternative measures forsediment management.

• Impacts on downstream envi-ronment and channel hydrau-lics.

• Recommended measures tomanage sediment properly andeconomically.

Objectives of sediment managementshould include flood control, waterquality, wetlands, fisheries, habitat,and riparian rights.

For hydropower dams, the simplestdecommissioning program is todismantle the turbine-generator andseal the water passages, leaving thedam and water-retaining structures inplace. No action is taken concerningthe sediments since they will remain inthe reservoir and the hydraulic andphysical characteristics of the riverand reservoir will remain essentiallyunchanged. This approach is viableonly if there are no deficiencies in thewater-retaining structures (such asinadequate spillway capacity or inad-equate factors of safety for stability)and long-term maintenance is ensured.In some cases, decommissioning caninclude partial removal of water-retaining structures. Partial removalinvolves demolition of a portion of the

dam to create a breach so that it nolonger functions as a water-retainingstructure.

For additional information, see Guide-lines for the Retirement of Hydroelec-tric Facilities published by the Ameri-can Society of Civil Engineers(ASCE) in 1997.

Channelization and Diversions

Channelization and flow diversionsrepresent forms of hydrologic modifi-cation commonly associated with mostprincipal land uses, and their effectsshould be considered in all restorationefforts (see Chapter 3). In some cases,restoration design can include theremoval or redesign of channel modi-fications to restore preexisting eco-logical and flow characteristics.

Modifications of existing projects,including operation and maintenanceor management, can improve somenegative effects without changing theexisting benefits or creating additionalproblems. Levees may be set backfrom the stream channel to betterdefine the stream corridor and reestab-lish some or all of the natural flood-plain functions. Setback levees can beconstructed to allow for overbankflooding, which provides surface watercontact with streamside areas such asfloodplains and wetlands.

Instream modifications such as uni-form cross sections or armoringassociated with channelization or flowdiversions may be removed, anddesign and placement of meanders canbe used to reestablish more naturalchannel characteristics. In manycases, however, existing land usesmight limit or prevent the removal ofexisting channel or floodplain modifi-



cations. In such cases, restorationdesign must consider the effects ofexisting channel modifications or flowdiversions, in the corridor and thewatershed.

Exotic Species

Exotic species are another commonproblem of stream corridor restorationand management. Some land useshave actually introduced exotics thathave become uncontrolled, whileothers have merely created an opportu-nity for such exotics to spread. Again,control of exotic species has somecommon aspects across land uses, butdesign approaches are different foreach land use.

Control of exotics in some situationscan be extremely difficult and may beimpractical if large acreages or well-established populations are involved.Use of herbicides may be tightlyregulated or precluded in many wet-land and streamside environments, andfor some exotic species there are noeffective control measures that can beeasily implemented over large areas(Rieger and Kreager 1988). Whereaggressive exotics are present, everyeffort should be made to avoid unnec-essary soil disturbance or disruption ofintact native vegetation, and newlyestablished populations of exoticsshould be eradicated.

Nonnative species such as salt cedar(Tamarix spp.) and Russian olive(Elaeagnus angustifolia) canoutcompete native plantings andnegatively affect their establishmentand growth. The likelihood of suc-cessful reestablishment often increaseswhen artificial flows created byimpoundments are altered to favor

native species and when exotics suchas salt cedar are removed beforerevegetation is attempted (Briggs et al.1994).

Salt cedar is an aggressive, exoticcolonizer in the West due to its longperiod and high rate of seed produc-tion, as well as its ability to withstandlong periods of inundation. Salt cedarcan be controlled either by clearingwith a bulldozer or by direct applica-tion of herbicide (Sudbrock 1993);however, improper treatments mayactually increase the density of saltcedar (Neill 1990).

Controlling exotics and weeds can beimportant because of potential compe-tition with established native vegeta-tion, colonized vegetation, and artifi-cially planted vegetation in restorationwork. Exotics compete for moisture,nutrients, sunlight, and space and canadversely influence establishment ratesof new plantings. To improve theeffectiveness of revegetation work,exotic vegetation should be clearedprior to planting; nonnative growthmust also be controlled after planting.General techniques for control ofexotics and weeds are mechanical(e.g., scalping or tilling), chemical(herbicides), and fire. For a review oftreatment methods and equipment, seeU.S. Forest Service (1965) andYoakum et al. (1980).

AgricultureAmerica’s Private Land—A Geogra-phy of Hope (NRCS 1997) challengesall of us to “regain our sense of placeand renew our commitment to privatelandowners and the public.” It sug-gests that as we learn more about thecomplexity of our environment,



The Multispecies Riparian Buffer System in the Bear Creek, IA Watershed

Introduction

The Bear Creek Watershed in central Iowa is a small (26.8 mi2) drainage basin located within the DesMoines Lobe subregion of the Western Corn Belt Plains ecoregion, one of the youngest and flattestecological subregions in Iowa. In general, the land is level to gently rolling with a poorly developed streamnetwork. Soils of the region are primarily developed in glacial till and alluvial, lacustrine, and windblowndeposits. Prior to European settlement of the region (ca 1847) the watershed consisted of the vasttallgrass prairie ecosystem, interspersed with wet prairie marshes in topographic lows and gallery forestsalong larger order streams and rivers. Native forest was limited to the Skunk River corridor into which BearCreek flows.

Subsequent conversion of the land, including the riparian zone, from native vegetation to row crops,extensive subsurface drainage tile installation, dredge ditching and grazing of fenced riparian zones haveresulted in substantial stream channel modification. Records suggest that artificial drainage of marshesand low prairies in the upper reaches of the Bear Creek watershed was completed about 1902, with ditchdredging completed shortly thereafter. While the main stream pattern appears to have remained about thesame since that time, significant channelization continued into the 1970s. Additional intermittent channelshave developed in association with new drainage tile and grass waterway installation. Present land use inthe Bear Creek watershed is typical of the region with over 87% of the land area devoted to row cropagriculture.

Landscape modifications and present land-use practices have produced nonpoint source pollution in thewatershed which landowners have addressed by implementing soil conservation practices (e.g. reducedtillage, terracing, grass waterways) and better chemical input management (e.g. more accurate and bettertimed applications). It has only been recently that placement or enhancement of riparian vegetation or“streamside filter strips” has been recommended to reduce sediment and chemical loading, modify flowregime by reducing discharge extremes, improve structural habitat, and restore energy relationshipsthrough the addition of organic matter and reduction in temperature and dissolved oxygen extremes.

The Riparian Management System (RiMS)

The Agroecology Issue Team of the Leopold Center for Sustainable Agriculture, Iowa State University,Ames, IA, is conducting research on the design and establishment of an integrated riparian managementsystem (RiMS) to demonstrate the benefits of properly functioning riparian buffers in the heavily row-cropped landscape of the midwestern U.S.. The purpose of the RiMS is to restore the essential ecologicalfunctions that riparian ecosystems once provided. Specific objectives of such buffers are to intercepteroding soil and agricultural chemicals from adjacent crop fields, slow floodwaters, stabilize streambanks,provide wildlife habitat, and improve the biological integrity of aquatic ecosystems. The regionalization ofthis system has been accomplished by designing it with several components, each of which can bemodified to fit local landscape conditions and landowner objectives.

The Agroecology Issue Team is conducting detailed studies of important biological and physical processesat both the field and watershed scale to provide the necessary data to allow resource managers to makecredible recommendations of buffer placement and design in a wide variety of landscapes. In addition,socioeconomic data collected from landowners in the watershed are being used to identify landownercriteria for accepting RiMS. The team also is quantifying the non-market value placed on the improvementin surface and ground water quality.

The actual development and establishment of the RiMS along Bear Creek was initiated in 1990 along a 0.6mile length of Bear Creek on the Ron and Sandy Risdal Farm. The buffer strip system has subsequentlybeen planted along 3.5 miles of Bear Creek upstream from this original site. The RiMS consists of threecomponents: 1) a multi-species riparian buffer (MRB), 2) soil bioengineering technologies for streambankstabilization, and 3) constructed wetlands to intercept and process nonpoint source pollutants in agriculturaldrainage tile water.



Multi-species Riparian Buffer (MRB)

The general MRB consists of three zones. The rapid growth of this buffer community can change a heavilyimpacted riparian zone into a functioning riparian ecosystem in a few short years. The combinations oftrees, shrubs and native grasses can be modified to fit site conditions (e.g. soils, slope), major bufferbiological and physical function(s), owner objectives, and cost-share program requirements.

Soil Bioengineering

It has been estimated that greater than 50% of the stream sediment load in small watersheds in theMidwest is the result of channel erosion. This problem has been worsened by the increased erosive powerof streams resulting from stream channelization and loss of riparian vegetation. Several different soilbioengineering techniques have been employed in the Bear Creek watershed. These include the use ofwillow posts and stakes driven into the bank, live willow fascines, live willow brush mattresses andbiodegradable geotextile anchored with willow stakes on bare slopes. Alternatives used to stabilize thebase of the streambank include rock and anchored dead plant material such as cedar or bundled maple.

Constructed Wetlands

Small, constructed wetlands which are integrated into the riparian buffer have considerable potential toremove nitrate and other chemicals from the extensive network of drain tile in the midwest. To demonstratethis technology, a small (600 yd2) wetland was constructed to process drainage tile water from a 12 acrecropped field. The wetland was constructed by excavating a depressional area near the creek andconstructing a low berm. The subsurface drainage tile was rerouted to enter the wetland at a point thatmaximizes residence time of drainage tile water within the wetland. A simple gated water level controlstructure at the wetland outlet provides control of the water level maintained within the wetland. Cattailrhizomes (Typha glauca Godr.) collected from a local marsh and road ditch were planted within the wetlandand native grasses and forbs planted on the constructed berm. Future plans include the construction ofadditional tile drainage wetlands within the Bear Creek watershed.

System Effectiveness

Long-term monitoring has demonstrated the significant capability of the RiMS to intercept eroding soil fromadjacent cropland, intercept and process agricultural chemicals moving in shallow subsurface water,stabilize stream channel movement, and improve instream environments, while also providing wildlifehabitat, and quality timber products. The buffer traps between 70-80% of the sediment carried in surfacerunoff and has reduced nitrate and atrazine moving in the soil solution to levels well below the maximumcontaminant levels specified by the USEPA. Streambank bioengineering systems have virtually stoppedbank erosion along treated reaches and are now trapping channel sediment. The constructed wetland hasreduced nitrate in the tile drainage water by as much as 80% depending on the season of the year. Wildlifebenefits have also appeared in a very short time with a nearly five fold increase in bird species diversityobserved within the buffer strip versus an adjacent, unprotected stream reach.

While the RiMS function is being assessed through experimental plot work with intensive processmonitoring economic benefits and costs to landowners and society also are being determined.Landowners surveys, focus groups, and one-on-one interviews have identified the concern that waterquality should be improved by reducing chemical and sediment inputs by as much as 50%. Landownersare willing to pay for this improved water quality as well as volunteer their time to help initiate theimprovements.

While the RiMS can effectively intercept and treat nonpoint source pollution from the uplands it should bestressed that a riparian management system cannot replace upland conservation practices. In a properlyfunctioning agricultural landscape, both upland conservation practices and an integrated riparian systemcontribute to achieving environmental goals and improved ecosystem functioning.

Support for this work is from the Leopold Center for Sustainable Agriculture, the Iowa Department ofNatural Resources through a grant from the USEPA under the Federal Nonpoint Source ManagementProgram (Section 319 of the Clean Water Act), and the USDA (Cooperative State Research Education andExtension Service), National Research Initiative Competitive Grants Program, and the Agriculture inConcert with the Environment Program.



harmony with ecological processesthat extend across all landscapesbecomes more of an imperative thanan ideal. Furthermore, conservationprovisions of the 1996 Farm Bill andaccompanying endeavors such as theNational Conservation Buffer Initia-tive (NRCS 1997) offer flexibility tocare for the land as never before. Thefollowing land use scenario attemptsto express this flexibility in the contextof comprehensive, locally led conser-vation work, including stream corridorrestoration.

This scenario offers a brief glimpseinto a hypothetical agricultural settingwhere the potential results of streamcorridor restoration might begin to

take form. Computer-generatedsimulations are used to graphicallyillustrate potential changes broughtabout by restoration work and associ-ated comprehensive, on-farm conser-vation planning. It focuses, conceptu-ally, on vegetative clearing, instreammodifications, soil exposure andcompaction, irrigation and drainage,and sediment or contaminants as themost disruptive activities associatedwith agricultural land use. Althoughan agricultural landscape typical of theMidwest was selected for illustrativepurposes, the concepts shown canapply in different agricultural settings.

Hypothetical Existing Conditions

Reminiscent of the highly disruptiveagricultural activities discussed inChapter 3, Figure 8.53 illustrateshypothetical conditions that focusprimarily on production agriculture.Although functionally isolated contourterraces and a waterway have beeninstalled in the nearby cropland, thescene depicts an ecologically deprivedlandscape. Many of the potentialdisturbance activities and subsequentchanges outlined in Chapter 3 come tomind. Those hypothetically reflectedin the figure are highlighted in Table8.10.

Hypothetical Restoration Response

Previous sections of this chapter andearlier chapters identified connectivityand dimension (width) as importantstructural attributes of stream corri-dors. Nutrient and water flow, sedi-ment trapping during floods, waterstorage, movement of flora and fauna,species diversity, interior habitatconditions, and provision of organic

Figure 8.53:Hypotheticalconditions.Activities causingchange in thisagricultural setting.



Potential Effects

ExistingDisturbance Activities

Vege

tati

ve C

lear

ing

Chan

neliz

atio

n

Stre

ambe

d D

istu

rban

ce

Soil

Expo

sure

or

Com

pact

ion

Cont

amin

ants

Woo

dy D

ebri

s Re

mov

al

Pipe

d D

isch

arge

/Con

t.O

utle

ts

Activity has potential for direct impact. Activity has potential for indirect impact.

Decreased landscape diversity

Point source pollution

Nonpoint source pollution

Dense compacted soil

Increased upland surface runoff

Increased sheetflow with surface erosion rill and gully flow

Increased levels of fine sediment and contaminants in stream corridor

Increased soil salinity

Increased peak flood elevation

Increased flood energy

Decreased infiltration of surface runoff

Decreased interflow and subsurface flow to and within the stream corridor

Reduced ground water recharge and aquifer volumes

Increased depth to ground water

Decreased ground water inflow to stream

Increased flow velocities

Reduced stream meander

Increased or decreased stream stability

Increased stream migration

Channel widening and downcutting

Increased stream gradient and reduced energy dissipation

Increased flow frequency

Reduced flow duration

Decreased capacity of floodplain and upland

Increased sediment and contaminants

Decreased capacity of stream

Reduced stream capacity to assimilate nutrients/pesticides

Confined stream channel with little opportunity for habitat development

Increased streambank erosion and channel scour

Increased bank failure

Loss of instream organic matter and related decomposition

Increased instream sediment, salinity, or turbidity

Increased instream nutrient enrichment, sedimentation, and contaminants leading to eutrophication

Table 8.10: Summaryof prominentagriculturally relateddisturbance activitiesand potential effects.



Potential Effects

ExistingDisturbance Activities

Vege

tativ

e Cl

earin

g

Chan

neliz

atio

n

Stre

ambe

d Di

stur

banc

e

Soil

Expo

sure

or C

ompa

ctio

n

Cont

amin

ants

Woo

dy D

ebris

Rem

oval

Pipe

d Di

scha

rge/

Cont

.Out

lets

Activity has potential for direct impact. Activity has potential for indirect impact.

Highly fragmented stream corridor with reduced linear distribution of habitatand edge effect

Loss of edge and interior habitat

Decreased connectivity and dimension (width) within corridor and to associated ecosystems

Decreased movement of flora and fauna species for seasonal migration, dispersal repopulation

Reduced stream capacity to assimilate nutrients/ pesticides

Increase of opportunistic species, predators

Increase exposure to solar radiation, weather, and temperature

Magnified temperature and moisture extremes in corridor

Loss of riparian vegetation

Decrease source of in stream shade, detritus, food, and cover

Loss of edge diversity

Increased water temperature

Impaired aquatic habitat

Reduced invertebrate population

Loss of wetland function

Reduced instream oxygen

Invasion of exotic species

Reduced gene pool

Reduced species diversity

Table 8.10: Summary ofprominent agriculturallyrelated disturbanceactivities and potentialeffects (continued).



materials to aquatic communities weredescribed as just a few of the func-tional conditions affected by thesestructural attributes. Continuousindigenous vegetative cover across thewidest possible stream corridor wasgenerally identified as the most condu-cive to serving the broadest range offunctions. This discussion went on tosuggest that a long, wide streamcorridor with contiguous vegetativecover is a favored overall characteris-tic. A contiguous, wide stream corri-dor may be unachievable, however,where competing land uses prevail.Furthermore, gaps caused by distur-bances (utility crossings, highwaysand access lanes, floods, wind, fire,etc.) are commonplace.

Restoration design should establishfunctional connections within andexternal to stream corridors. Land-scape elements such as remnantpatches of riparian vegetation, prairie,or forest exhibiting diverse or uniquevegetative communities; productiveland that can support ecologicalfunctions; reserve or abandoned land;associated wetlands or meadows;neighboring springs and stream sys-tems; ecologically innovative residen-tial areas; and movement corridors forflora and fauna (field borders, wind-breaks, waterways, grassed terraces,etc.) offer opportunities to establishthese connections. An edge (transitionzone) that gradually changes from oneland use into another will softenenvironmental gradients and minimizedisturbance.

With these and the broad designguidelines presented in previoussections of this chapter in mind,Figure 8.54 presents a conceptualcomputer-generated illustration ofhypothetical restoration results. Table8.11 identifies some of restorationmeasures hypothetically implementedand their potential effects on restoringconditions within the stream corridorand surrounding landscape.

Figure 8.54:Hypothetical restorationresponse.Possible results of streamcorridor restoration arepresented in thiscomputer-alteredphotograph.



Ta

Potential Resulting Effects

Restoration Measures

Wet

land

s

Ripa

rian

Hab

itat

Upl

and

Corr

idor

s

Win

dbre

aks/

Shel

terb

elts

Nat

ive

Plan

t Co

ver

Stre

am C

hann

el R

esto

rati

on

Upl

and

BMPs

for

Agr

icul

ture

Measure contributes directly to resulting effect. Measure contributes little to resulting effect.

Increased landscape diversity

Increased stream order

Reduced point source pollution

Reduced nonpoint source pollution

Increased soil friability

Decreased upland surface runoff

Decreased sheetflow, width, surface erosion, rill and gully flow

Decreased levels of fine sediment and contaminants in stream corridor

Decreased soil salinity

Decreased peak flood elevation

Decreased flood energy

Increased infiltration of surface runoff

Increased interflow and subsurface flow to and within stream corridor

Increased ground water recharge and aquifer volumes

Decreased depth to ground water

Increased ground water inflow to stream

Decreased flow velocities

Increased stream meander

Increased stream stability

Decreased stream migration

Reduced channel widening and downcutting

Decreased stream gradient and increased energy dissipation

Decreased flow frequency

Increased flow duration

Increased capacity of floodplain and upland

Decreased sediment and contaminants

Increased capacity of stream

Increased stream capacity to assimilate nutrients/pesticides

Enhanced stream channel with more opportunity for habitat development

Decreased streambank erosion and channel scour

Decreased bank failure

Gain of instream organic matter and related decomposition

Decreased instream sediment, salinity, or turbidity

Table 8.11: Summary of prominent restoration measures and potentialresulting effects.



Potential Resulting Effects

Restoration Measures

Wet

land

s

Ripa

rian

Hab

itat

Upl

and

Corr

idor

s

Win

dbre

aks/

Shel

terb

elts

Nat

ive

Plan

t Cov

er

Stre

am C

hann

el R

esto

ratio

n

Upl

and

BMPs

for A

gric

ultu

re

Measure contributes directly to resulting effect. Measure contributes little to resulting effect.

Decreased instream nutrient enrichment, siltation, and contaminants leading to eutrophication

Connected stream corridor with increased linear distribution of habitat and edge effect

Gain of edge and interior habitat

Increased connectivity and dimension (width) within corridor and to associated ecosystems

Increased movement of flora and fauna species for seasonal migration, dispersal repopulation

Decrease of opportunistic species, predators

Decreased exposure to solar radiation, weather, and temperature

Decreased temperature and moisture extremes in corridor

Increased riparian vegetation

Increased source of in stream shade, detritus, food, and cover

Increase of edge diversity

Decreased water temperature

Enhanced aquatic habitat

Increased invertebrate population

Increased wetland function

Increased instream oxygen

Decrease of exotic species

Increased gene pool

Increased species diversity

Table 8.11: Summary of prominent restoration measures and potentialresulting effects (continued).



culverts, and lack of erosion controlmeasures on road prisms, cut-and- fillslopes, and ditches are problemscommon to a poor road design (Stonerand McFall 1991). The most extremeroad system rehabilitation requires fullroad closure. Full road closure in-volves removal of culverts and resto-ration of the streams that werecrossed. It can also involve the rip-ping or tilling of road surfaces toallow plant establishment. If naturalvegetation has not already invadedareas of exposed soils, planting andseeding might be necessary.

Full closure might not be a viablealternative if roads are needed toprovide access for other uses. In thesecircumstances a design to restricttraffic might be appropriate. Volun-tary traffic control usually cannot berelied on, so traffic barriers like gates,fences, or earth berms could be neces-sary. Even with traffic restriction,roads require regular inspection forexisting or potential maintenanceneeds. The best time for inspection isduring or immediately after largestorms or snowmelt episodes so theeffectiveness of the culverts and roaddrainage features can be witnessedfirst-hand. Design should addressregular maintenance activities includ-ing road grading, ditch cleaning,culvert cleaning, erosion controlvegetation establishment, and vegeta-tion management.

Buffer Strips in Forestry

Forested buffer strips are generallymore effective in reducing sedimentand chemical loadings in the streamcorridor than vegetated filter strips(VFS). However, they are susceptibleto similar problems with concentrated

ForestryStream corridors are a source of largevolumes of timber. Timber harvestingand related forest management prac-tices in riparian corridors often neces-sitate stream corridor restoration.Forest management may be an on-going land use and part of the restora-tion effort. Regardless, accessing andharvesting timber affects streams inmany ways including:

• Alteration of soil conditions.

• Removal of the forest canopy.

• Reduction in the potentialsupply of large organic(woody) debris (Belt et al.1992).

Forest Roads

The vast majority of the restorationdesign necessary following timberharvest is usually devoted to the roadsystem, where the greatest alterationof soil conditions has taken place.Inadequate drainage, poor location,improperly sized and maintained

BMP Implementation and Section319 of the Clean Water Act

Section 319 of the Clean Water Act of 1987required the states to identify and submitBMPs for USEPA approval to help controlnonpoint sources of pollution. As of 1993,41 of 50 states had EPA-approved voluntaryor regulatory BMP programs dealing withsilvicultural (forest management) activities.The state BMPs are all similar; the majoritydeal with roads. Montana, for example, hasa total of 55 specifically addressed forestpractices. Of those 55 practices, 35 dealwith road planning and location, roaddesign, road maintenance, road drainage,road construction, and stream crossings.



flows. Buffers constructed as part of aconservation system increase effec-tiveness. A stiff-stemmed grass hedgecould be planted upslope of either aVFS or a woody riparian forest buffer.The stiff-stemmed grass hedge keepssediment out of the buffer and in-creases shallow sheet flow through thebuffer.

Most state BMPs also have specialsections devoted to limitations forforest management activities in ripar-ian “buffer strips” (also referred to asStreamside Management Zones orStreamside Protection Zones).

Budd et al. (1987) developed a proce-dure for determining buffer widths forstreams within a single watershed inthe Pacific Northwest. They focusedtheir attention primarily on mainte-

nance of fish and wildlife habitatquality (stream temperature, foodsupply, stream structure, sedimentcontrol) and found that effective bufferwidths varied with the slope of adja-cent uplands, the distribution of wet-lands, soil and vegetation characteris-tics, and land use. They concludedthat practical determinations of streambuffer width can be made using suchanalyses, but it is clear that a genericbuffer width which would providehabitat maintenance while satisfyinghuman demands does not exist. Thedetermination of buffer widths in-volves a broad perspective that inte-grates ecological functions and landuse. The section on design approachesto common effects at the beginning ofthis chapter also includes some discus-sion on stream buffer width.

Yes, number per 1000 feet, dependent on stream widthb

75% current shadeaFixed minimum (75 feet)

Idaho

State

Class I*

NoneNoneFixed minimum (5 feet)

Class II**

Yes, number per 1000 feet, dependent on stream width and bed material

50%, 75% if temperature > 60ºF

Variable by stream width (5 to 100 feet)

Washington Type 1, 2, and 3*

25 per 1000 feet, 6 inches diameter

NoneNoneType 4**

Yes; number to be determined by canopy density

50% overstory and/or understory; dependent on slope and stream class

Variable by slope and stream class (50 to 200 feet)

California Class I and Class II*

Nonee50% understoryeNonebClass III**

Yes; number per 1000 feet and basal area per 1000 feet by stream width

50% existing canopy, 75% existing shade

Variable, 3 times stream width (25 to 100 feet)

Oregon Class I**

None75% existing shadeNonefClass II specialprotection**

StreamClass

Buffer Strip Requirements

Width Shade or Canopy Leave Trees

* Human water supply or fisheries use.

** Streams capable of sediment transport (CA) or other influences (ID and WA) or significant impact (OR) on downstream waters.a In ID, the shade requirement is designed to maintain stream temperatures.b In ID, the leave tree requirement is designed to provide for recruitment of large woody debris.c May range as high as 300 feet for some types of timber harvest.d To be determined by field inspection.e Residual vegetation must be sufficient to prevent degradation of downstream beneficial uses.f In eastern OR, operators are required to "leave stabilization strips of undergrowth... sufficient to prevent washing of sediment into

Class I streams below."

Table 8.12: Bufferstrip requirements bystate.



Stream corridors have varied dimen-sions, but stream buffer strips havelegal dimensions that vary by state(Table 8.12). The buffer may be onlypart of the corridor or it may be all ofit. Unlike designing stream corridorsfor recreation features or grazing use,designing for timber harvest andrelated forest management activities isquite regimented by law and regula-tion. Specific requirements vary from

Pacific Northwest Floods Of 1996

Floods, Landslides, and Forest Management—‘The Rest of TheStory’

Warm winds, intense rainfall, and rapid snowmelt during thewinter of 1995-96 and again in the winter of 1996-97 causedmajor flooding, landslides, and related damage throughout thePacific Northwest (Figure 8.55 ). Such flooding had not beenseen for more than 30 years in hard-hit areas. Damage to roads,campgrounds, trails, watersheds, and aquatic resources waswidespread on National Forest Service lands. These eventsoffered a unique opportunity to investigate the effects of severeweather, examine the influence and effectiveness of various forestmanagement techniques, and implement a repair strategyconsistent with ecosystem management principles.

The road network in the National Forests was heavily damagedduring the floods. Decisions about the need to replace roadsare based on long-term access and travel requirements.Relocation of roads to areas outside floodplains is a measurebeing taken. Examination of road crossings at streams concludedwith design recommendations to keep the water moving, alignculverts horizontally and longitudinally with the stream channel,and minimize changes in stream channel cross section at inletbasins to prevent debris plugs.

Many river systems were also damaged. In some systems,however, stable, well-vegetated slopes and streambankscombined with fully functioning floodplains buffered the effectsof the floods. Restoration efforts will focus on aiding naturalprocesses in these systems. Streambank stabilization andriparian plantings will be commonly used. Examination ofinstream structure durability concluded that structures aremore likely to remain in place if they are in fourth-order orsmaller streams and are situated in a manner that maintains aconnection between the structure and the streambank. Theywill be most durable in watersheds with low landslide/debristorrent frequency.

state to state; the state Forester’s officeor local Extension Service can provideguidance on regulatory issues. USDANatural Resource Conservation Ser-vice offices and Soil and Water Con-servation District offices also aresources of information. Refer to Beltet al. (1992) and Welsch (1991) forguidance on riparian buffer stripdesign, function, and management.Salo and Cundy (1987) provide infor-mation on forestry effects on fisheries.

Figure 8.55: 1996 Landslides.(a) April landslide: debris took out thetrack into the Greenwater River and(b) July landslide: debris took out theroad and deposited debris into theriver.

(a)

(b)



GrazingThe closer an ecosystem is managed toallow for natural ecological processesto function, the more successful arestoration strategy will be. In streamcorridors that have been severelydegraded by grazing, rehabilitationshould begin with grazing manage-ment to allow for vegetative recovery.

Vegetative recovery is often moreeffective than installing a structure.The vegetation maintains itself inperpetuity, allows streams to functionin ways that artificial structures cannotreplicate, and provides resiliency thatallows riparian systems to withstand avariety of environmental conditions(Elmore and Beschta 1987).

Designs that promote vegetativerecovery after grazing are beneficial ina number of ways. Woody species canprovide resistance to channel erosionand improve channel stability so thatother species can become established.As vegetation becomes established,channel elevation will increase assediment is deposited within and alongthe banks of the channel (aggradation),and water tables will rise and mayreach the root zone of plants on formerterraces or floodplains. This aggrada-tion of the channel and the rising watertable allow more water to be storedduring wet seasons, thereby prolong-ing flow even during periods ofdrought (Elmore and Beschta 1987).

Kauffman et al. (1993) observed thatfencing livestock out of the riparianzone is the only grazing strategy thatconsistently results in the greatest rateof vegetative recovery and the greatestimprovement in riparian function.However, fencing is very expensive,requires considerable maintenance,

and can limit wildlife access—anegative impact on habitat or conduitfunctions.

Some specialized grazing strategieshold promise for rehabilitating lessseverely impacted riparian and wet-land areas without excluding livestockfor long periods of time. The effi-ciency of a number of grazing strate-gies with respect to fishery needs aresummarized in Tables 8.13 and 8.14(from Platts 1989). They summarizethe influence of grazing systems andstream system characteristics onvegetation response, primarily from awestern semiarid perspective. Somegeneral design recommendations forselecting a strategy include the follow-ing (Elmore and Kauffmann 1994):

• Each strategy must be tailoredto a particular stream or streamreach. Management objectivesand components of the ecosys-tem that are of critical valuemust be identified (i.e., woodyspecies recovery, streambankrestoration, increased habitatdiversity, etc.). Other informa-tion that should be identifiedincludes present vegetation,potential of the site for recov-ery, the desired future condi-tion, and the current factorscausing habitat degradation orlimiting its recovery.

• The relationships betweenecological processes that mustfunction for riparian recoveryshould be described. Factorsaffecting present condition(i.e., management stress vs.natural stress) and conditionsrequired for the stream toresume natural functions need



Strategya Level to Which Riparian Vegetation is Commonly Used

Control of Animal Distribution (Allotment)

Streambank Stability

Brushy Species Condition

Seasonal Plant Regrowth

Stream Riparian Rehabilitation Potential

Fishery Needs Ratingb

HeavyContinuous season-long (cattle)

Poor Poor Poor Poor Poor 1

HeavyHolding (sheep or cattle) Excellent Poor Poor Fair Poor 1

HeavyShort duration-high intensity (cattle)

Excellent Poor Poor Poor Poor 1

Heavy to moderate

Three herd-four pasture (cattle)

Good Poor Poor Poor Poor 2

Heavy to lightHolistic (cattle or sheep) Good Poor to good

Poor Good Poor to excellent

2-9

Moderate to heavy

Deferred (cattle) Fair Poor Poor Fair Fair 3

HeavySeasonal suitability (cattle)

Good Poor Poor Fair Fair 3

Heavy to moderate

Deferred-rotation (cattle) Good Fair Fair Fair Fair 4

Heavy to moderate

Stuttered deferred-rotation (cattle)

Good Fair Fair Fair Fair 4

Moderate to heavy

Winter (sheep or cattle) Fair Good Fair Fair to good

Good 5

Heavy to moderate

Rest-rotation (cattle) Good Fair to good

Fair Fair to good

Fair 5

ModerateDouble rest-rotation (cattle)

Good Good Fair good Good 6

Moderate to light

Seasonal riparian preference (cattle or sheep)

Good Good Good Fair Fair 6

As prescribedRiparian pasture (cattle or sheep)

Good Good Good Good Good 8

NoneCorridor fencing(cattle or sheep)

Excellent Good toexcellent

Good toexcellent

Good Excellent 9

LightRest-rotation with seasonal preference (sheep)

Good Good to excellent

Good to excellent

Good Excellent 9

NoneRest or closure (cattle or sheep)

Excellent Excellent Excellent Excellent Excellent 10

a Jacoby (1989) and Platts (1989) define these management strategiesb Rating scale based on 1 (poorly compatible) to 10 (highly compatible with fishery needs)

Table 8.13:Evaluation and ratingof grazing strategies.



Grazing System

SteepLow Sediment Load

ModerateLow Sediment Load

ModerateHigh Sediment Load

FlatLow Sediment Load

FlatHigh Sediment Load

Shrubs +Herbs +Banks 0

No grazing Shrubs +Herbs +Banks 0 to +


Shrubs +Herbs +Banks +




Winter or dormant season

Shrubs +Herbs +Banks 0 to +






Early growing season

Shrubs +Herbs +Banks 0 to +





Shrubs –Herbs +Banks 0 to -

Deferred or late season

Shrubs –Herbs +Banks 0 to -

Shrubs –Herbs +Banks 0 to +

Shrubs –Herbs +Banks +



Shrubs –Herbs +Banks 0 to –

Three-pasture rest rotation







Deferred rotation


Shrubs –Herbs +Banks + to 0




Shrubs +Herbs +Banks 0 to –

Early rotation Shrubs +Herbs +Banks 0 to +

Shrubs +Herbs +Banks + to 0





Rotation Shrubs –Herbs +Banks 0 to –





Shrubs –Herbs –Banks 0 to –

Season-long Shrubs –Herbs –Banks 0 to –

Shrubs –Herbs –Banks –





Spring and fall Shrubs –Herbs –Banks 0 to –



Shrubs –Herbs –Banks – to 0

Shrubs –Herbs –Banks 0 to +


Spring and summer





Shrubs –Herbs –Banks 0 to +

SteepHigh Sediment Load

Note: – = decrease; + = increase; 0 = no change. Stream gradient: 0 to 2% = flat; 2 to 4% = moderate; > 4% = steep. Banks refers to bank stability.

Table 8.14: Generalized relationships between grazing systems, stream system characteristics, and riparianvegetation response.



to be assessed. Anthropogenicfactors causing stream degra-dation must be identified andchanged.

• Design and implementationshould be driven by attainablegoals, objectives, and manage-ment activities that willachieve the desired structureand functions.

• Implementation should includea monitoring plan that willevaluate management, allow-ing for corrections or modifica-tions as necessary, and a strongcompliance and use supervi-sion program.

The main consideration for selecting agrazing system is to have an adequatevegetative growing season betweenthe period of grazing and timing ofhigh-energy runoff. It is impossible toprovide a cookie-cutter grazing strat-egy for every stream corridor; designshave to be determined on the ground,stream by stream, manager by man-ager. Simply decreasing the numberof livestock is not a solution to de-graded riparian conditions; rather,restoring these degraded areas requiresfundamental changes in the ways thatlivestock are grazed (Chaney et al.1990).

Clearly, the continued use of grazingsystems that do not include the func-tional requirements of riparian vegeta-tion communities will only perpetuateriparian problems (Elmore andBeschta 1987). Kinch (1989) andClary and Webster (1989) providegreater detail on riparian grazingmanagement and designing alternativegrazing strategies. Chaney and others(1990 and 1993) present photo histo-

ries of a number of interesting grazingrestoration case studies, and of theshort-term results of some of theavailable grazing strategies.

MiningPost-mining reclamation of streamcorridors must begin with restorationof a properly functioning channel.Because many of the geologic andgeomorphic controls associated withthe pre-disturbance channel may havebeen obliterated by mining operations,design of the post-mining channeloften requires approaches other thanmimicking the pre-disturbance condi-tion. Channel alignment, slope, andsize may be determined on the basis ofempirical relations developed fromother streams in the same hydrologicand physiographic settings (e.g.,Rechard and Schaefer 1984, Rosgen1996). Others (e.g., Hasfurther 1985)have used a combination of empiricaland theoretical approaches for designof reclaimed channels. Total recon-struction of stream channels is treatedat length in Section 8.E. Other sec-tions of the chapter address stabiliza-tion of streambanks, revegetation offloodplains and terraces, and restora-tion of aquatic and terrestrial habitats.Additional guidance is available inInterfluve, Inc. (1991).

Surface mining is usually associatedwith large-scale disturbances in thecontributing watershed, therefore, arigorous hydrological analysis of pre-and post-mining conditions is criticalfor stream corridor restoration ofdisturbed systems. The hydrologicanalysis should include a frequencyanalysis of extreme high- and low-flow events to assess channel perfor-mance in the post-mining landscape.



Oven Run, Pennsylvania

The effects of abandoned mines draining into the surrounding lands cause dramatic changes in the area(Figure 8.58(a) ). Runoff with high levels of minerals and acidity can denude the ground of vegetation,expose the soil, and allow erosion with the sediment further stressing streams and wetland. Any efforts torestore streams in this environment must deal with the problem if any success is to be likely.

The Natural Resources Conservation Service, formerly known as the Soil Conservation Service, has beenworking on the Oven Run project along with the Stonycreek Conemaugh River Improvement (SCRIP) toimprove water quality in a 4-mile reach above the Borough of Hooversville. SCRIP is a group of local andstate government as well as hundreds of individuals interested in improving the water quality in an area onPennsylvania’s Degraded Watersheds list.

The initial goal of improving water quality resulted in improving habitat and aesthetic qualities. The watercoming into Hooversville had higher-than-desired levels of iron, manganese, aluminum, sulfate, and acidity.Six former strip mines, which had a range of problems, were identified. They included deep mine openingsthat have large flows of acid mine drainage, acid mine seepage into streams, eroding spoil areas, areas ofponded water that infiltrate into ground water (adding to the acid mine drainage), and areas downhill ofseepage and deep mine drainage that are denuded and eroding.

Control efforts included grading and vegetating the abandoned mine to reduce infiltration through acid-bearing layers and reduce erosion and sedimentation, surface water controls to carry water around thesites to safer outlets, and treating discharge flow with anoxic limestone drains and chambered passivewetland treatments (Figure 8.58(b)) . Additionally, 1,000 feet of trees were planted along one of the sitestreams to shade the Stoneycreek River. Average annual costs for the six sites were estimated to be$503,000 compared to average annual benefits of $513,000.

The sites are being monitored on a monthly basis, and 4 years after work was begun the treatments havehad a measurable success. The acid influent has been neutralized, and the effluent is now a net alkaline.

Iron, aluminum, and manganese levels have been reduced,with iron now at average levels of 0.5 mg/L from averagelevels of 35 mg/L.

Figure 8.58: Stream corridor (a) before and(b) after restoration.

(a)

(b)



Hydrologic modeling may be requiredto generate runoff hydrographs for thepost-mining channel because water-shed geology, soils, vegetation, andtopography may be completely alteredby mining operations. Thus, channeldesign and stability assessments willbe based on modeled runoff ratesreflecting expected watershed condi-tions. The hydrologic analysis forpost-mining restoration should alsoaddress sediment production from thereclaimed landscape. Sedimentbudgets (see Chapter 7) will be neededfor both the period of vegetationestablishment and the final revegetatedcondition.

The hydrologic analyses will providerestoration practitioners with the flowand sediment characteristics neededfor restoration design. The analysesmay also indicate a need for at leasttemporary runoff detention and sedi-ment retention during the period ofvegetation establishment. However,the post-mining channel should bedesigned for long-term equilibriumwith the fully reclaimed landscape.

Water quality issues (e.g., acid minedrainage) often control the feasibilityof stream restoration in mined areasand should be considered in design.

RecreationBoth concentrated and dispersedrecreational use of stream corridorscan cause damage and ecologicalchange. Ecological damage primarilyresults from the need for access for therecreational user. A trail often willdevelop along the shortest or easiestroute to the point of access on thestream. Additional resource damagemay be a function of the mode ofaccess to the stream: motorcycles andhorses cause far more damage tovegetation and trails than do pedestri-ans. Control of streambank access indeveloped recreation sites must be partof a restoration design. On undevel-oped or unmanaged sites, such controlis more difficult but still very neces-sary (Figure 8.56).

Rehabilitation of severely degradedrecreation areas may require at leasttemporary use restrictions. Evenactively eroding trails, camp andpicnic sites, and stream access pointscan be stabilized through temporarysite closure and combinations of soiland vegetation restoration (Wenger1984, Marion and Merriam 1985,Hammitt and Cole 1987). Closure willnot provide a long-term solution ifaccess is restored without addressingthe cause of the original problem.Rather, new trails and recreation sitesshould be located and constructedbased on an understanding of vegeta-tion capabilities, soil limitations, andother physical site characteristics.Basically, the keys to a successfuldesign are:

• Initially locating or moving useto the most damage-resistantsites.

Figure 8.56: ControlledaccessControl of streambankaccess is an importantpart of the restorationdesign.Source: J. McShane.



• Influencing visitor use.

• Hardening use areas to makethem more resistant.

• Rehabilitating closed sites.

UrbanizationFew land uses have the capacity toalter water and sediment yield from adrainage as much as the conversion ofa watershed from rural to urban condi-tions; thus, few land uses have greaterpotential to affect the natural environ-ment of a stream corridor.

As a first step in hydrologic analyses,designers should characterize thenature of existing hydrologic responseand the likelihood for future shifts inwater and sediment yield. Initially,construction activities create excesssediment that can be deposited indownstream channels and floodplains.As impervious cover increases, peakflows increase. Water becomescleaner as more area is covered withlandscaping or impervious material.The increased flows and cleaner waterenlarge channels, which increasessediment loads downstream.

Determine if the watershed is (a) fullyurbanized, (b) undergoing a new phaseof urbanization, or (c) is in the begin-ning stages of urbanization (Riley,1998).

An increase in the amount of impervi-ous cover in a watershed leads toincreased peak flows and resultingchannel enlargement (Figure 8.57).Research has shown that imperviouscover of as little as 10 to 15 percent ofa watershed can have significantadverse effects on channel conditions(Schueler 1996). Magnitudes ofchannel-forming or bankfull flood

events (typically 1- to 3-year recur-rence intervals) are increased signifi-cantly, and flood events that previouslyoccurred once every year or two mayoccur as often as one or two times amonth.

Enlargement of streams with subse-quent increases in downstream sedi-ment loads in urbanized watershedsshould be expected and accommodatedin the design of restoration treatments.Procedures for estimating peak dis-charges are described in Chapter 7,and effects of urbanization on magni-tude of peak flows must be incorpo-rated into the analysis. Sauer et al.(1983) investigated the effect ofurbanization on peak flows by analyz-ing 199 urban watersheds in 56 citiesand 31 states. The objective of theanalysis was to determine the increasein peak discharges due to urbanizationand to develop regression equationsfor estimating design floods, such asthe 100-year or 1 percent chanceannual flood, for ungauged urbanwatersheds. Sauer et al. (1983)developed regression equations basedon watershed, climatic, and urbancharacteristics that can be used to

Figure 8.57: Stormwater flow on a pavedsurface.Impervious surfacesincrease peak flows andcan result in channelenlargement.Source: M. Corrigan.



estimate the 2, 5, 10, 25, 50, 100, and500-year urban annual peak dischargesfor ungauged urban watersheds. Theequation for the 100-year flood incubic feet per second (UQ100) isprovided as an example:

UQ100 = 2.50 A.29 SL.15 (RI2+3)1.76

(ST+8)-.52 (13-BDF)-.28 IA.06 Q100.63

where the explanatory variables aredrainage area in square miles (A),channel slope in feet per mile (SL), the2-year, 2-hour rainfall in inches (RI2),basin storage in percent (ST), basindevelopment factor (BDF), which is ameasure of the extent of developmentof the drainage system (dimensionless,ranging from 0 to 12), percent imper-vious area (IA), and the equivalentrural peak discharge in cubic feet persecond (RQ100) in the exampleequation above.

Sauer et al. (1983) provide the allow-able range for each variable. The twoindices of urbanization in the equationare BDF and IA. They can be used toadjust the rural peak discharge RQ100(either estimated or observed) to urbanconditions.

Sauer et al. (1983) provide equationslike the one above and graphs thatrelate the ratio of the urban to ruralpeak discharge (UQx/RQx) for recur-rence intervals x = 2, 10, and 100years. The 2-year peak ratio variesfrom 1.3 to 4.3, depending on thevalues of BDF and IA; the 10-yearratio varies from 1.2 to 3.1; and the100-year ratio varies from 1.1 to 2.6.These ratios indicate that urbanizationgenerally has a lesser effect on higher-recurrence-interval floods becausewatershed soils are more saturated andfloodplain storage more fully depleted

in large floods, even in the ruralcondition.

More sophisticated hydrologic analy-ses than the above are often used,including use of computer models,regional regression equations, andstatistical analyses of gauge data.Hydrologic models, such as HEC-1 orTR-20, are often already developed forsome urban watersheds.

Once the flood characteristics of thestream are adjusted for urbanization,new equilibrium channel dimensionscan be estimated from hydraulicgeometry relationships developedusing data from stable, alluvial chan-nels in similar (soils, slope, degree ofurbanization) watersheds, or otheranalytical approaches. Additionalguidance for design of restored chan-nels is provided earlier in this chapterin the section on channel reconstruc-tion.

Changes in flooding caused by urban-ization of a watershed can be mitigatedduring urban planning through prac-tices designed to control storm runoff.These practices emphasize the use ofvegetation and biotechnical methods,as well as structural methods, tomaintain or restore water quality anddampen peak runoff rates. Strategiesfor controlling runoff include thefollowing:

• Increasing infiltration ofrainfall and streamflow toreduce runoff and to removepollutants.

• Increasing surface and subsur-face storage to reduce peakflows and induce sedimentdeposition.



• Filtration and biologicaltreatment of suspended andsoluble pollutants (i.e., con-structed wetlands).

• Establishment and/or enhance-ment of forested riparianbuffers.

• Management of drainage fromthe transportation network.

• Introduction of trees, shrubs,etc., for various restorationpurposes.

In addition to changes in water yield,urbanization of a watershed frequentlygenerates changes in its sedimentyield. In humid climates, vegetativecover prior to urbanization often isadequate to protect soil resources andminimize natural erosion, and thecombination of impervious area andvegetation of a fully urban watershedmight be adequate to minimize sedi-ment yield. During the period ofurbanization, however, sediment yieldsincrease significantly as vegetation iscleared and bare soil is exposed duringthe construction process. In more aridclimates, sediment yield from an urbanwatershed may actually be lower thanthe yield from a rural watershed due tothe increased impervious area andvegetation associated with landscap-ing, but the period of urbanization(i.e., construction) is still the time ofgreatest sediment production.

The effect of urbanization on sedimentdischarge is illustrated in Figure 8.59,which contains data from ninesubbasins in a 32-square-mile area inthe Rock Creek and Anacostia RiverBasins north of Washington, DC(Yorke and Herb 1978). During theperiod of data collection (1963-74),three subbasins remained virtually

rural while the others underwent urbandevelopment. In 1974, urban landrepresented from 0 to 60 percent ofland use in the nine subbasins. Thesedata were used to develop a relationbetween suspended sediment yield andthe percentage of land under construc-tion. This relation indicated thatsuspended sediment yield increasedabout 3.5 times for watersheds with 10percent of the land area under con-struction. However, suspended-sediment yields for watersheds wheresediment controls (primarily sedimentbasins) were employed for 50 percentof the construction area were onlyabout one-third of these for areaswithout controls. The effect of con-trols is seen in the figure. The threecurves present growing season data forthree periods of increasing sedimentcontrol: 1963-67, when no controlswere used on construction sites; 1968-71, when controls were mandatory;

Figure 8.59: Sediment-transport curves forgrowing season storms.The effect of urbanizationon sediment discharge isillustrated from datacollected in a 32-square-mile area.

10

Sto

rm S

edim

ent

Dis

char

ge

(to

ns)

Storm Runoff (cfs-days)

50 100 50010

50

100

500

1000

5000

10,000

1963-671968-711972-74

1000



and 1972-74, when controls weremandatory and subject to inspectionby county officials. It further illus-trates that storm runoff is not the onlyfactor affecting storm sediment dis-charge as evidenced by the significantscatter about each relation.

In addition to sediment basins, man-agement practices for erosion andsediment control focus on the follow-ing objectives:

• Stabilizing critical areas alongand on highways, roads, andstreets.

• Siting and placement of sedi-ment migration barriers.

• Design and location of mea-sures to divert or exclude flowfrom sensitive areas.

• Protection of waterways andoutlets.

• Stream and corridor protectionand enhancement.

All of these objectives emphasize theuse of vegetation for sediment control.Additional information on BMPs forcontrolling runoff and sediment inurban watersheds can be found in theTechniques Appendix.

In theory, a local watershed manage-ment plan might be the best tool toprotect a stream corridor from thecumulative impact of urban develop-ment; however, in practice, few suchplans have realized this goal (Schueler1996). To succeed, such plans mustaddress the amount of bare groundexposed during construction and theamount of impervious area that willexist during and after development ofthe watershed. More importantly,success will depend on using thewatershed plan to guide development

decisions, and not merely archiving itas a one-time study whose recommen-dations were read once but neverimplemented (Schueler 1996).

Key Tools of Urban StreamRestoration Design

Restoration design for streams de-graded by prior urbanization mustconsider pre-existing controls and theireffects on restoration objectives.Seven restoration tools can be appliedto help restore urban streams. Thesetools are intended to compensate forstream functions and processes thathave been diminished or degraded byprior watershed urbanization. The bestresults are usually obtained when thefollowing tools are applied together.

Tool 1. Partially restore thepredevelopment hydrological regime.The primary objective is to reduce thefrequency of bankfull flows in thecontributing watershed. This is oftendone by constructing upstream stormwater retrofit ponds that capture anddetain increased storm water runoff forup to 24 hours before release (i.e.,extended detention). A commondesign storm for extended detention isthe one-year, 24 hour storm event.Storm water retrofit ponds are oftencritical in the restoration of small andmidsized streams, but may be imprac-tical in larger streams and rivers.

Tool 2. Reduce urban pollutant pulses.A second need in urban stream restora-tion is to reduce concentrations ofnutrients, bacteria and toxics in thestream, as well as trapping excesssediment loads. Generally, three toolscan be applied to reduce pollutantinputs to an urban stream: storm waterretrofit ponds or wetlands, watershed



pollution prevention programs, and theelimination of illicit or illegal sanitaryconnections to the storm sewer net-work

Tool 3. Stabilize channel morphology.Over time, urban stream channelsenlarge their dimensions, and aresubject to severe bank and bed ero-sion. Therefore, it is important tostabilize the channel, and if possible,restore equilibrium channel geometry.In addition, it is also useful to provideundercuts or overhead cover to im-prove fish habitat. Depending on thestream order, watershed imperviouscover and the height and angle oferoded banks, a series of differenttools can be applied to stabilize thechannel, and prevent further erosion.Bank stabilization measures includeimbricated rip-rap, brush bundles, soilbioengineering methods such aswillow stakes and bio-logs, lunkerstructures and rootwads. Grade stabili-zation measures are discussed earlierin this chapter and in Appendix A.

Tool 4. Restore Instream habitatstructure. Most urban streams havepoor instream habitat structure, oftentypified by indistinct and shallow lowflow channels within a much largerand unstable storm channel. The goalis to restore instream habitat structurethat has been blown out by erosivefloods. Key restoration elementsinclude the creation of pools andriffles, confinement and deepening ofthe low flow channels, and the provi-sion of greater structural complexityacross the streambed. Typical toolsinclude the installation of logcheckdams, stone wing deflectors andboulder clusters along the streamchannel.

Tool 5. Reestablish Riparian Cover.Riparian cover is an essential compo-nent of the urban stream ecosystem.Riparian cover stabilizes banks,provides large woody debris anddetritus, and shades the stream. There-fore, the fifth tool involves reestablish-ing the riparian cover plant communityalong the stream network. This canentail active reforestation of nativespecies, removal of exotic species, orchanges in mowing operations to allowgradual succession. It is often essen-tial that the riparian corridor be pro-tected by a wide urban stream buffer.

Tool 6. Protect critical stream sub-strates. A stable, well sorted streambedis often a critical requirement for fishspawning and secondary production byaquatic insects. The bed of urbanstreams, however, is often highlyunstable and clogged by fine sedimentdeposits. It is often necessary to applytools to restore the quality of streamsubstrates at points along the streamchannel. Often, the energy of urbanstorm water can be used to createcleaner substrates—through the use oftools such as double wing deflectorsand flow concentrators. If thickdeposits of sediment have accumulatedon the bed, mechanical sedimentremoval may be needed.

Tool 7. Allow for recolonization of thestream community. It may be difficultto reestablish the fish community in anurban stream if downstream fishbarriers prevent natural recolonization.Thus, the last urban stream restorationtool involves the judgment of a fisherybiologist to determine if downstreamfish barriers exist, whether they can beremoved, or whether selective stockingof native fish are needed to recolonizethe stream reach.

chapter 8: restoration design restoration design 8 · chapter 8: restoration design final...

Documents