Native American methods for conservationand restoration of semiarid ephemeralstreamsJ.B. Norton, F. Bowannie Jr., P. Peynetsa, W. Quandl!lacy, and S.F. Siebert

ABSTRACT: Combined effects of brush encroachment and channel entrenchment in theSouthwestern United States threaten ecological, hydrological, and agricultural functions ofephemeral streams and associated alluvial surfaces. Restoration is difficult becauseentrenchment is widespread, and highly variable flow magnitudes defy standardized structuralapproaches. Methods developed by the Zuni Indians during more than 2,000 years of farming ondynamic alluvial fans combine brush removal with ephemeral channel-erosion control and showpromise for effective watershed-scale conservation and restoration. Zuni utilize several types ofsimple brush structures that rely on hydraulic characteristics of woody material to modify erosiveand depositional effects of both small seasonal and irregular storm-flow events. Zuni brushstructures are inexpensive and quick to build, require no external material inputs, and avoid theextensive disturbance associated with conventional rigid check-dam construction. Successivesurveys at three incised-headwater ephemeral channels on the Zuni Indian Reservation showedthat Zuni techniques increased the incidence of overbank flow during small streamflow events(acting as permeable check dams) and large floods (moving downstream and forming debrisjams) and thereby helped reconnect channels to alluvial fans. Zuni brush structures represent apotential alternative to capital- and labor-intensive approaches to semiarid watershedconservation and restoration.

Keywords: Arroyo cutting, conservation farming, discontinuous ephemeral streams, ecologicalrestoration, local knowledge, permeable check dams, runoff agriculture, woody debris, woody

species encroachment

Arroyo cutting and encroachment bywoody species have been degradingdesert and shrub-steppe grasslands ofthe Western United States for more than100 years (Archer 1994; Bryan 1925).Although the two processes are closely relat-ed because woody species encroachmentincreases runoff and erosion (Abrahams et al.1995; Bull 1997), they are typically treatedseparately in rangeland conservation andrestoration efforts.A comprehensive approachthat restores both plant communities and sur-face hydrology to more desirable conditionscould inlprove the success and cost-effective-ness of ecological restoration efforts. Thispaper explores traditional Zuni methods thatuse brush and trees cleared from degradedrangelands to control arroyo cutting. It alsopresents an example in which local knowl-

edge in danger of being lost because of drasticdeclines in traditional land use could be valu-able in addressing a complex, contemporaryenvironmental problem.

Large-scale gully erosion, which seversconnections between ephemeral streamflowand ecologically important alluvial landforms,has defied control for decades (Gellis et al.1995). Conventional approaches, includingrigid impervious check dams made of earth,concrete, gabions, or logs (Heede 1976;

250 jOURNALOFSOILANDWATERCONSERVATION 5102002

Reprinted from dIe Journal of Soil and Woter ConservationVolume 57, Number 5

Copyright CI 2002 Soil and Water Conservation Society

managed by mdigenous tarnlers who have avery rich body of traditional knowledge.

Controlling entrenchment of semiaridephemeral streams is difficult because theextremely variable and dynamic conditionscreate stream behavior different from thatwhere most hydrological principles and ero-sion control methods were developed (Bull1997). Small streamflow events that occuronce every two to 10 years carry most of thesediment that fills and destroys structures builtto store water and control large floods. Large,powerful floods can breach, flank, or undercutcheck dams designed to control erosion dur-ing smaller flows (Gellis et al. 1995).

Though conventional rigid check damscan modjfy particular stream reaches (Gellis etal. 1995), the immense extent of ephemeraldrainage networks and relatively high cost ofsuch structures limit their watershed-scaleeffectiveness. Soil bioengineering approaches(Gray and Sotir 1996) are also relatively mate-rial- and labor-intensive for such extensiveapplication, and soil conditions of semiarid-arid ephemeral streams where arroyo cuttingis most prevalent may be too dry to supportfast-growing vegetation. Headwater streamsoften make up as much as 85% of the totalchannel length in watersheds (peterson et al.2001) and, in semiarid areas, flow only inresponse to rainfall. Large watershed restora-tion efforts must be simple and quick toimplement, require minimal capital and laborinputs, and effectively modjfy both smallflows and powerful floods.

Zuni farnlers who still practice traditionalagriculture use simple, low-investment struc-tures that appear to mitigate destructivearroyo cutting during both small and largestrearnflows. This type of erosion control isoften part of brush- and tree-clearing activi-ties intended to prepare fields for cultivationor to improve livestock grazing lands. Ourobjective was to describe and evaluate theperformance of indigenous erosion controlmethods and to evaluate their potential appli-cation in watershed- and ecosystem-scaleconservation and restoration efforts.

Methods and MaterialsWe conducted this study in collaboration

with traditional Zuni farmers on the ZuniIndian Reservation in west-central NewMexico. The Zuni reservation lies on thesoutheastern part of the Colorado Plateau at1,800 to 2,400 m elevation and receives anaverage of 300 mm of precipitation annually,

Seehorn 1992; Zeedyk et al.1995), as well assoil bioengineering approaches, which uselive plant materials where sufficient moisturewill allow rooting (Gray and Sotir 1996),address only the erosion part of a persistentfeedback between gIillying and invasion ofrangelands by woody species. Other restora-tion efforts focus on rangeland brush removal

by cutting, burning, chaining, or othermethods (Connelly et al. 2000) but ignore theconcentration of runoff into gullies thataccompanies the shift &om grass to domi-nance by woody species.

Soils formed in alluvium of headwaterstream systems retain runoff, sediments, andnutrients to support biological productivityon-site, protect downstream aquatic resources(peterson et al. 2001), and support some ofthe oldest agricultures, including that of theZuni and other Southwestern NativeAmericans (Kintigh 1985; Nabhan 1984;Sandor 1995), the wadi fam1ers of the Negev(Evenari et al. 1982), fam1ers of the NileValley Qenny 1962), and others (e.g., Hillel1991; Hirst and Ibrahim 1996). Overbankflows are crucial to all these functions becausesoil productivity is renewed with inputs ofmoisture, sediment, and organic matter &omupland hillslopes (Norton et al. 2001).Channel entrenchment eliminates these allu-vial processes (Bull 1997). The Zuni andother agricultural Southwestern NativeAmericans have used and protected the natu-ral productivity of soils derived &om alluviumin annual food crop-based farming systemsfor millennia (Hack 1942; Kintigh 1985;Nabhan 1984; Sandor 1995).

In general, the diverse ways in which tradi-tional farmers approach soil and waterconservation reflect deep-seated practicalknowledge about what works and what doesnot work under particular environnlentalconditions (Bocco 1991; Siebert and Belsky1990). In semiarid environnlents, localapproaches are often adapted to highlyvariable and unpredictable conditions, acharacteristic that is increasingly sought inmodern management systems (Berkes et al.2000). Traditional agricultural systems indry environnlents depend on incrementalmethods that emphasize flexibility to takeadvantage of pulses of productivity or tosurvive drought (Ellis et al. 1993; Niemeijer1998; Tabor 1995). In the SouthwesternUnited States, ephemeral stream entrench-ment (or arroyo cutting) is a dynamic,unpredictable landscape process that has been~

VOWME 57 NUMBER 51 251 II 510 2002

the majority of which often falls duringd1understorms in July, August, and September.The combination of steep erodible sandstoneand shale slopes covered by patchy pinyon-juniper (Pinus edulis-juniperns spp.) and scrub-shrub plant conununities wid1 relativelyintense sununer storms contributes todynamic, sand-bed fluvial systems that movelarge amounts of sediment (Lagasse et al.

1990).Zuni subsistence agriculture depended on

corn (Zia mays) grown in nonirrigated fieldsin d1ese dynamic fluvial settings for morethan 1,500 years (Kintigh 1985). Socio-eco-nomic change and assimilation policiesbeginning with U.S. occupation of Zunilands in d1e mid-19dt century caused drasticdeclines in agriculture (Cleveland et al. 1995).Major arroyo cutting coincided with this shiftaway from traditional farming and led tomultiple, U.S.-sponsored erosion-controlefforts throughout d1e 20dt century, includinglarge earthen check darns built by govern-ment agencies and smaller-scale efforts by theCivilian Conservation Corps (CCC) andod1er programs (Gellis et al. 1995).

Three actively eroding, low-gradient,entrenched, meandering headwater ephemeralstreams-Rosgen "F-type" channels (Rosgen1994)-were selected for treatment withtraditional Zuni brush structures (Laate, Lalio,and Quandelacy sites). These streams are rep-resentative of headwater alluvial systems andprovide important ecological, hydrological,and agricultural functions.

The Laate site is a short ephemeral channelthat crosses an orchard high on d1e leewardside of a sandstone mesa. The channel gradi-ent was historically maintained to facilitateflooding of d1e sandy eolian soils. In 1996,channel incision threatened to exclude bothmature and recendy established fruit treesfrom streamflows. The Quandelacy site is acontinuously incised channel running nomd1e moud1 of a small canyon through range-lands dominated by big sagebrush (Artemisiatridentata).At the beginning of our study, thischannel ranged from a slot-like gully about1 m deep by 1 m wide to a 6 m deep, V-Shaped arroyo and had several steep headcuts.The Lalio site is a deeply incised arroyo thatempties onto a traditional, hand-cultivatedcornfield. By 1998, d1e arroyo had deepenedfor a third of the way into the' cornfield,excluding about three-quarters of the fieldfrom flooding. A continuous rill had formedthrough d1e field.

Results and DiscussionTraditional Zuni brush treatments effec-

tively reduced ilie erosive power of smallstreamflow events and large floodS, iliough byvery different modes of action. They alsoeffectively modified channel morphology andrestored channel/alluvial fan connectivity.During small flows, ilie structures acted aspermeable barriers that retained runoff intemporary pools and accumulated sediments.During larger, more powerful floodS, many ofilie structures were destroyed, but ilie woodydebris was redeposited in debris jams thataltered channel morphology and drasticallyincreased ilie likelihood of overbank flow.

Treatment Design and LAbor Requirements.The workers installed four general types ofbrush structures iliroughout ilie treatedreaches. Construction followed general prin-cipals from knowledge about flow events butvaried by worker and by streamflow situation.Spacing between ilie structures (Figure 1)averaged less than 20 m for all ilie sites, muchcloser ilian guidelines calculated wiili Heede's(1976) spacing rule, which is commonly usedfor designing erosion-control projects. Thisspacing rule is based on ilie equation:

HS =

A three-part approach for treating andevaluating structure effectiveness was imple-mented at each study site:

Treatments. Zuni crews led by people withlifdong experience in traditional agriculturalpractices installed erosion-control structuresat the three sites using axes, shovels, and post-hole diggers. The Laate site was first treated inJune 1996, the Lalio site inJune 1998,and theQuandelacy site in June 1999. Members ofthe crews recorded labor input, wrotedescriptions of structures, sketched maps andstructure designs, and photographed con-struction. The crew leaders are co-authors ofthis report.

Channel morphology. Successive longitudi-nal thalweg surveys were used to documenttreatment effects on channel morphology(Harrelson et at. 1994). Surveys were conduct-ed during construction, after flow events, andat the end of the study. (The relatively remoteLaate site was surveyed only at the beginningand end of the study period.) Photo pointswere established along each channel. Surveysincluded devations of thalweg, structures,banks, and water levels identified by high~water marks on the arroyo banks.

Channell alluvial fan connectivity. Pre~ andpost-study channel dimensions were used toestimate the probability of overbank flow atthe Lalio and Quandelacy sites. Manning'sequation was used to estimate the peak dis-charge that would be required to overtoppre- and post-study arroyo banks and thenwas applied to two- through 100-year recur-rence interval equations (Thomas et al. 1997)to estimate the average mquency that suchoverbank flows would occur.

To place the events and treatment effectsobserved in 1999 within the context of long -term Zuni climate, recurrence intervals ofboth rainfall and streamflow events were esti-mated. Rainfall volume and intensity weremonitored at a tipping-bucket rain gauge2 km from the Lalio site (data courtesy of theZuni Conservation Project, Zuni Pueblo,New Mexico) and rainfall-event recurrenceintervals were determined with intensity-duration-mquency curves for Albuquerque,New Mexico (USWB 1955). The magnitudeof 1999 streamflow events was determined byapplying Manning's equation to channddimensions surveyed after flow events byusing high-water marks to indicate peakstage. These values were compared to peakdischarges derived from equations for two-through 100-year recurrence interval eventsin the two watersheds (Thomas et at. 1997).

tan IX), and K is a constant developed throughregression analyses for different channelgradients (K = 0.3 for G S 0.20).

The exact placement of the brush struc-tures varied depending upon channel charac-teristics and the types of woody material mostreadily available. Simple brush piles werethe most commonly employed techniquethroughout the channels, with stronger struc-tures placed in key locations toward distalends of the channels and in constricted areas.Minjrnizing disturbance to channel beds andbanks, using material close at hand, and usingstones only sparingly to compress brush wereconcepts common to all four kinds of struc-tures, which are described below:

Brush piles. Limbs of pinyon and junipertrees and whole, big sagebrush plants werepiled in channels with larger sterns pointingdownstream (Figure 2a). Material wasstomped down repeatedly and layered spar-ingly with stones or larger logs where thosewere readily available. Large amounts of sage-brush were packed into slot-like channelsover distances of as much as 30 m by stomp-ing down layer after layer. Shallow rillsthrough the Laate Orchard and the Laliocornfield were filled with fine woody materi-al from rubber rabbitbrush (Chyrsothamnusnauseosus) and broom snakeweed (Gutiemziasarothrae). This was the most rapid method andhad the objective of placing as much debris aspossible into the channel with a minimum ofadditional strengthening.

--

KGcosa

where S is spacing (m), H is dam height (m),G is the channel gradient ratio, a is the anglecorresponding to the channel gradient (G =

252 JOURNAL OF SOIL ANO WATER CONSERVATION SIO 2002

Figure 2Types of structures installed at study sites: (a) brush pile with finebranches oriented upstream and stones to compress brush (drawing byFred Bowannie Jr.); (b) brush check dam with fine branches woven intolarger limbs piled across channel. (Tape follows thalweg.); (c) pungieposts immediately upstream from a post-and-wire structure. Posts areabout 8 cm in diameter at the base.

(a)

Arroyo bank

Flow .Flow .

Arroyo bed Side View

Top View

Check dams. Simple check dams of inter-woven woody material were slightly moreintensive than the brush piles. In thisapproach, the largest, longest limbs were laidacross the channel in a pile about 0.5 to 1 mhigh with brushy tops in alternating direc-tions. Smaller limbs were jabbed and woveninto the pile so that branches and leaves pre-sented a dense mat of fine woody material tothe upstream direction (Figure 2b). Some ofthese dams were reinforced with posts cutttom pinyon or juniper stems and set imme-diately downstream ttom the brush structure.

Post-anJ-wire methods. In this method,workers set two rows of poles about 1 m apartacross the arroyo channel and packed brushymaterial between the poles to create a barrierabout 1.5 m high. The tops of the poles werethen wired together. This method requmdmore effort in sorting and trimming woodymaterial, setting posts, packing brush, andwiring braces together. This method wasdeliberately employed in relatively wide, low-energy reaches, at points immediately belowsharp bends, or in series of two to three struc-tures. Each of these structures was built byone worker and took two to three hours forstructures up to 7 m long.

15102002 VOWME 57 NUMBER 5 I 253 I

Pungie posts. Pungie posts consist ofclusters of pinyon and juniper limbs andsterns set into the arroyo bed to form comb-like structures that capture woody debris andcreate reinforced debris jams during flowevents (Figure 2c). The posts were set at least60 cm into the ground, angled upstream.Placement was downstream from brush-pilestructures and upstream from other types toprotect and reinforce them. Some of thepungies were reinforced with post-and-wirebraces. Part way through the season, theworkers decided to alter this method byleaving fine branches on the posts so that, ifposts washed out, they would function aswoody debris.

The crew treated a total of 1.25 km ofephemeral stream channel during the studyperiod but had to treat segments repeatedlyafter large floods removed structures threetimes at the Lalio site and once at theQuandelacy site. Including the repeated treat-ments, 3.4 km were treated, averaging 93person-hours km-1 of channel. The first treat-ments at the Lalio and Quandelacy sites wererelatively rapid, but the second treatmentsrequired more time to build the reinforcedpungie post and post-and-wire structures.

Treatment Effects. Although 1999 was arelatively dry year (252 mm, or 80010 of nor-mal precipitation), 170 mm of rain fell inJulyand August, which is 159% of normal precip-itation (WRCC 2001). Much of this raincarne as unusually intense thunderstorms(Figure 3), including a greater-than-25-yearrecurrence interval storm on July 5, at morethan 18 mm in 15 minutes. The thunder-storms generated streamflow events withwidely varying magnitude, including severalfive- to 25-year recurrence interval flows andfour that equaled or exceeded 100-year floodevents (one at the Quandelacy site on Aug. '2and three at the Lalio site on July 5, Aug. 2,and Aug. 27), according to measurementsconducted by hydrologists of the ZuniConservation Project. Because of the remotenature of the Laate site, channel din1ensionswere surveyed only during treatment installa-tion and at the end of the study, so flowmagnitudes were not estimated for that site.

The structures at the Quandelacy and Laliosites all withstood the less-than-25-yearrecurrence interval flows, temporarily retain-ing water and causing some sedin1ent deposi-tion (Figures 4 and 5). During larger floods,however, many of the structures washed out.

deposition for more d1an 100m upstream.Channels upstream from die debris jams

filled widi sediment to within a few centime-ters of die bank from pretreatment depilis ofabout 1.5 m at die Lalio site (Figure 6) and1 m at die Quandelacy site. This caused over-bank flow for nearly 100 m upstream of diedebris jam at the Lalio site, which depositednearly 20 cm of sand on parts of die alluvialfan. Figure 7 shows die pre- and post-studychannel profiles at die three study sites. Thestructures at the Laate site retained sedimentand were mosdy buried during die post-treatment survey, except for die most distalstructure, which had been installed onbedrock and washed away.

The overall effect of die treatments was torestore flooding to larger portions of die allu-vial fans at all three sites. The schematic mapof die Lalio site (Figure 8) shows that, at diebeginning of the study, flows emerged fromthe channd near die center of die agricultur-al field. The apex of die fan, or the pointwhere flow leaves die channel and spreadsover the alluvial fan, moved upstream aftereach flood as the channel filled widi sedi-ment. At the end of die study, small flows(- two- to five-year recurrence interval)remained in the channel and flowed to dieagricultural field, but larger, potentiallydestructive floods would overtop die banksfar upstream. Although die response ofvegetation in the areas cleared for brush wasnot measured, increased grass production,particularly blue g= (Boutelouagradlis), wasevident compared with adjacent areas.

Repeated surveys at channel cross-sectionsat the Lalio and Quandelacy sites confirmthat connectivity between die ephemeralstream channels and alluvial fans was restored(Table 1). At the Lalio site, sediment deposi-tion 100 m upstream from die debris jamreduced the hydraulic radius so that calculat-ed recurrence interval (Thomas 1997) ofoverbank flow changed from nearly 50,000years at the beginning of die study to fouryears at the end. Similarly, at die Quandelacysite, calculated recurrence interval of over-bank flow changed from 3,000 years to three

years.Literature Review and DisCIIssion. Clearing

brush and trees from rangelands and using diematerial for erosion control simultaneouslyaddresses two major ecological degradationprocesses. Clearing brush and trees that com-pete widi grasses is a common approach toecological restoration and has been shown to~

Figure 5The lower part of Lalio arroyo shortly after peak discharge on July 18,

1999. Note the stepwise waterlines behind the structures and theeffect of brush clearing along banks.

~

increase forage production (Archer 1994).Increased herbaceous ground cover alsoincreases the hydraulic roughness of soil sur-faces, causing more infiltration and less runoff(Bull 1997). The Zuni workers often prunedlower branches on pinyon and juniper treesrather than cutting whole trees, explainingthat removing lower branches increased for-

age production around trees and increasedsite distances for livestock herding but main-tained the shade and wind protection provid-ed by the trees. Pruning and/or clearing treesand brush is often the principal purpose; thematerials are disposed of in arroyos to controlerosion.

Our results are supported by a previousstudy on the Zuni Indian Reservation inwhich Gellis et al. (1995) found that 28 of 47

large, heavy-equipment-built earth and rockerosion-control structures they evaluated hadbreached, and 31 were more than half-filledwith sediment in the 20 to 30 years sinceconstruction. In contrast, only five of 23brush structures had washed out or werebreached, while the remaining structurescontinued to function. The brush structureswere built in the 1970s by a YouthConservation Corps team led by an elderlyZuni farmer. Gellis et al. (1995) attributedfailures of both the large rock and earthenstructures and the Zuni brush structures tolack of maintenance. Because they require no

heavy earth-moving equipment or offiitematerials, the brush structures would be easierand less expensive for local people to buildand maintain. Unlike capital- and labor-intensive rigid darns, modification of Zunibrush structures by streamflow events is not

15/02002 VOLUME 57 NUMBER 51 255 I

considered failure ofthe structures butpart of the ongoingprocess of conservingalluvial fan functions.

In thjs traditionalZuni approach, asmuch woody materi-al as possible is placedin channels withminjmal strengthen-ing. No keys intobeds or banks and noreinforcements withstones or gabions areused. Instead of rely-ing on rigid check dams, which can causescouring of streambeds or banks (Gellis et al.1995), erosion control is achieved throughhydraulic characteristics of woody material.Although there are few published accounts ofpermeable brush structures for controllingarroyo cutting, woody debris is commonlyused to improve or restore aquatic habitat inperennial streams.

There are similarities between the struc-tures built during our study and methodsused during the 1930s by CCC crews at Zuniand elsewhere (USEPA 1999). However, theCCC methods all include stone armor andkeying structures into beds or banks, whilethe Zuni methods purposefully avoid thesetechniques. Nevertheless, the similarities sug-gest that some of the techniques could havebeen handed down from Zuni people whoserved on CCC crews. The CCC techniquesmay have been adapted by Zuni people to fitlocal conditions. The workers in our study

explained how, although the approach waslearned from older Zuni people, methodswere constantly modified to fit specificconditions. The brush structures reflect theconstant experimentation and innovationcharacteristic of traditional ecological knowl-edge (Berkes et al. 2000).

Studies of woody debris suggest that theZuni methods could be effective for increas-ing sedimentation. Woody debris is responsi-ble for as much as 87% of channel sedimentstorage in small woodland streams, with 39%or more of the individual pieces formingdepositional sites (Fetherston et al. 1995). Assuch, woody debris is a primary determinantof channel morphology, forming pools andregulating transport of sediment, organicmatter, and nutrients. While lack of woodydebris is often recognized as a characteristic ofdryland streams, Minckley and Rinne (1985)suggest that streamside vegetation was abun-dant before destruction of riparian forests

vating the water-surface profile and signifi-cantly increasing flood travel times (Gippel1995). In one study, woody debris coveredless than 2% of the streambed but provided500/0 of total flow resistance (Mmga andKirchner 2000).After initial treatments in ourstudy, woody material covered nearly 50% ofthe channel bed of the treated reaches.

Implications for Watershed Restoration.The labor and materials required to constructbrush piles suggest that this method could beimplemented rapidly and inexpensively overentire watersheds. At an average of93 person-hours kIn-! of entrenched channel treated, a20-person crew could treat more than 1.7 kInper eight-hour day. Our work during anunusually wet and stormy year required thatwe treat the study reaches repeatedly after 50-to 100-year recurrence events. These unusu-ally frequent and high magnitude storm flowscame almost inunediately after structureswere installed, before smaller events began tobury the woody material with sediments andhold them in place. This partly explains theneed to repeatedly treat the reaches.

Over the long term and across wholewatersheds, treated headwater channels mayneed additional woody material added atabout 20-year intervals, after large floodsremove woody debris or as brush darns fillwith sediment. This observation is supportedby Gellis et al. (1995), who found that five of28 brush structures needed maintenance 20years after installation. More frequent debrisadditions after flows that cause sedimentationwould raise channel-bed elevations andthereby increase the likelihood of overbankflows and deposition in agricultural fields.

These Zuni erosion control practices areclosely associated with traditional rainfedcorn (Zea mays) farming, which provided thebasis of Zuni subsistence until the late 19thcentury (Kintigh 1985). Since then, changingsocio-economic conditions, devastating epi-demics, and u.S. agricultural developmentpolicies caused drastic declines in rainfedfarming (Oeveland et al. 1995; Ferguson andHart 1985). The cessation of traditional farm-ing in ephemeral stream courses coincidedwith the onset of stream-channel incision atZuni (Balling and Wells 1990; Bryan 1925).This suggests that the absence of previouslywidespread traditional erosion-control activi-ties may have contributed to the severity ofchannel entrenchment that is now wide-spread in the Zuni area.

The Zuni approach to controlling arroyo

256 JOURNAL OF SOil AN D WATER CONSERVATION 5102002

Figure 8Sketch map showing chronology of flood events and reaction of treatments at the Lalio site.Drawing by Fred Bowannie Jr. See discussion in text.cutting may be relevant to large-scale range-

land watershed restoration in the WesternUnited States. Implementing this approachcould be as simple as using woody materialdisplaced during chaining or cutting to buildthe simple permeable brush check dams innearby incised watercourses. While this addi-tional work would be relatively inexpensive, itcould effectively reduce erosion and increasesedimentation over wide areas.

The Zuni methods could be useful inother semiarid regions where incisedephemeral stream channels have sand or finerbeds, flow events move large amounts of bedload and suspended sediments, and there isabundant fine woody material near erodingchannels. An important characteristic of theZuni approach is that farmers modify erosioncontrol methods to fit particular situations(i.e., channel geometry or the types of woodymaterial available) or objectives (i.e., control,prevention, or restoration of connectivitybetween channels and surrounding alluvialsurfaces). As such, the Zuni methods wouldprobably be most applicable if built uponlocal environmental knowledge and tradi-tions. Traditional, low-input erosion-controlpractices are widely used throughout theworld (e.g., Mexico, Bocco 1991; andIndonesia, Siebert and Belsky 1990) andtypically display adaptations to local environ-mental and cultural conditions, as well as basicprinciples of potential use elsewhere.

Debrisjams

+

I

flows and die area of alluvial fans flooded byrecurrence interval flows of less than or equalto five years. Frequent, low-intensity floodsare important to renewal of desirable soilproperties. These methods are rapid and inex-pensive to implement and thus may be usefulin watershed-scale restoration efforts. Whilemost conventional erosion-control methodsrely on rigid check dams of concrete, stones,gabions, and/or logs keyed into streambedsand banks, the Zuni brush structures rely onhydraulic characteristics of woody debris tomodify erosive flows. Rather than investing

Summary and ConclusionTraditional Zuni erosion-control mediods

effectively modify channel morphology ofentrenched headwater ephemeral streamsafter small flows and powerful floods in waysthat increase die occurrence of overbank

,

time in strengthening structures, Zuni farm-ers and livestock herders focus on placing asmuch woody material as possible over longchannel reaches in an attempt to reduce floodpower, move fan apexes upstream, andincrease dle ttequency and area of overbankflows. The Zuni medlods combine clearingbrush and trees widl erosion control in anefficient, comprehensive approach to water-shed restoration.

The approach described herein is based onlong-term understanding of dle range ofvariability in dle dynamic fluvial environ-

Table 1.. Estimated discharge for overbank flow (discharge required to reach top of channel banks calculated using Manning's equation withchannel dimensions at the beginning and end of the study period. See discussion in text.

Parameter Units Laliooverbank,

pretreatment

Laliooverbank,

post-treatment

Quandelacyoverbank,

pretreatment

Quandelacyoverbank,

post-treatment

36

320

0.02

4.5

0.58

0.026

17.02

never

36320

0.0061.080.20

0.0261.10

4

17130

0.0351.940.44

0.0268.07never

171300.010.580.15

0.0260.63

3

ham

m2

m

m3s2

years

Watershed areaStation'SlopeCross-sectional areaHydraulic radiusManning's n"Peak dischargeRecurrence interval'"

* Meters from lower end of longitudinal profile.

** Based on roughness coefficients for sand bed channels (Jarrett .1.985).

* * Estimated using regression equations of Thomas (.1.997).

ISIO 2002, VOWME 57 NUMBER 5 I 257 I

ment of the Southwestern United States. Itincorporates some methods brought to Zuniduring nearly 100 years of U.S. government-sponsored erosion control but adapts them tofit local conservation objectives and tradition-al environmental knowledge. Traditional Zunisoil conservation measures allow localresidents to effectively conserve and restorenatural resources without expensive equip-ment. These methods may also represent acomprehensive, cost-effective approach tolarge-scale watershed restoration applicableto other semiarid environments aroundthe world, particularly when combinedwith local knowledge of runoff and erosioncharacteristics.

AcknowledgementThis work was funded by National Science

Foundation grant DEB-9S284S8 and is dedicated

to the memory of Andrew H. Laahty, who lead thisresearch effort as director of the Zuni Sustainable

Agriculture Project. We thank the Zuni Tribe and

Zuni Conservation Program for supporting this

research. Craig Goodwin's review contributed to

the accuracy of our report.

international symposium. American Society of Civil

Engineers. Fort Collins, CO.

Manga, M. and J. W Ki11:hner. 2000. Stress partitioning in

streams by large woody debris. Water Resources

Research 36:2373-2379.

Minckley, WL. and J.N. Rinne. 1985. Large woody debris in

hotcdesert streams: An historical review. Desert Plants

7:142-153.

Nabhan, G. 1984. Soil fertility renewal and water harvesting

in Sonoran Desert agriculture: The Papago example.

Arid Lands Newsletter 20:21-28.

Niemeijer, D. 1998. Soil nutrient harvesting in indigenous

teras water harvesting in Eastern Sudan. Land

Degradation & Development 9:323-330.

Norton,J.B., R.R. Pawluk, andJ.A. Sandor. 2001. Farmer-

scientist collaboration for research and agricul=al

development on the Zuni Indian Reservation, New

Mexico, USA. Pp 107-120. In W A. Payne and D.

Keeney (editon). Sustainability in agricul=al systems in

transition. American Society of Agronomy. Madison,

WI.

Peterson, BJ., WM. Wollheim, PJ. MulhoI1and,J.R. Webster,

J.L. Meyer, J.L. Tank, E. Marti, WB. Bowden, H.M.

Valett, A.E. Hershey, WHo McDowell, WK. Dodds, S.K.

Hamilton, S. Gregory, and D.D. Morrall. 2001. Control

of nitrogen export fium watersheds by headwater

streams. Science 292:86-90.

Rosgen, D.L. 1994. A c1assification of na=al riven. Catena

22:169-199.

Sandor,J.A. 1995. Searching soil for clues about Southwest

prehistoric agriculture. In H:WToll (editor). ProceedinS'

of the New Mexico Archaeological Council Symposium

on Southwest Agriculture.

Seehorn, M.E. 1992. Stream habitat improvement hand-

book. Technical Publication R8-TP 16. Atlanta, GA:

U.S. Forest Service.

Siebert, S.E, and J.M. Belsky. 1990. Bench terracing in the

Kerinci uplands of Sumatra, Indonesia. Journal of Soil

and Water Conservation 45:559-562.

Tabor,J.A.1995.improving crop yidds in the Sahel by means

of water-harvesting. Journal of Arid Environments

30:83-106.

Thomas, B.E., H. W Hjalmarson, and S.D. Waltemeyer. 1997.

Methods for estimating magnitude and &equency of

floods in the Southwestern United States. Water Supply

Paper No. 2433. Denver, CO: U.S. Geological Sur"")'

U.S. Environmental Protection Agency. 1999. Unified water-

shed assessment. Albuquerque, NM: USEPA.

U.S. Weather Bureau. 1955. Rainfal1 intensitycduration-&e-

quency curves. Technical Paper No. 25. Hydrologic

Services Division, USWB. Washington, D.C.

Western Regional Climate Center. 2001. Zuni, New

Mexico, monthly total precipitation (Station 299897).

Available from WRCC at www.wrcc.dri.edu/cgi-

bin/c1iMAIN.pl?nmzuni (verified November).Zeedyk, B., B. Romero, and S.K. Albert. 1995. Using simple

structures for flow dispersion in wet meadow restoration.

Desired future conditions for Southwestern riparian

ecosystems: Bringing interests and concerns together

RM GTR-272. Fort Collins, CO: Forest Service.

Connelly,J.W, M.A. Schroeder,A.R. Sands, and C.E. Braun.

2000. Guiddines ro manage sage grouse populations and

rheir habirats. Wildlife Sociery Bulletin 28:%7-985.

Ellis,J.E., M.B. Coughenour, and D.M. Swift. 1993. Climate

variability, ecosystem srability, and the implications for

range and livestock development. p 248. In R.HJ.Behnke et al. (editors). Range ecology at disequilibrium:

New model.. of narural variability and pastoral adapra-

tion in African savannas. London: Overseas

Development Instirute.

Evenari, M.L., L. Shanan, and N. Tadmor. 1982. The Negev:

The challenge of the desert. Cambridge, MA: Harvard

University Press.

Ferguson, TJ., and E.R. Hart. 1985. A Zuni adas. Norman,

OK: University of Oklahoma Press.

Fetherston, K.L., RJ. Naiman, and R.E. Bilby. 1995. Large

woody debris, physical process, and riparian forest

development in monrane river networks of the Pacific

Northwest. Geomorphology 13:133-144.

Gellis, A.C., A. Cheama, v: Laahty, and S. Lalio. 1995.

Assessment of gully-control strucrures in the Rio

Nutria watershed, Zuni Reservation, New Mexico.

Water Resources Bulletin 31:633-646.

Gippel, CJ. 1995. Environmenral hydeaulics oflatKe woodydebris in streams and rivers. Journal of Environmenral

Engineering 121:388-395.

Gray, D.H., and R.B. Sotir. 1996. Biotechnical and soil bio-

engineering slope srabi1ization: A practical guide for

erosion control. New York, NY:John Wiley &: Sons.

Hack,J. T. 1942. The changing physical environment of the

Hopi indian. of Arizona. Papers of the Peabody

Museum of American Archaeology and Ethnology,

Harvard 35:1-85.

Harrelson, C.C., C.L. Rawlins, and J.P. Potyondy. 1994.

Stream channel reference sites: An illustrated guide to

field technique. General Technical Report GTR RM-

245. U.S> Department of Agriculture (USDA) Forest

Service.

Heede, B.H. 1976. Gully development and control. Research

Paper RM-169. USDA.

Hillel, D. 1991. Out of the Earth: Civilization and the life of

the soil. Berkeley, CA: University of California Press.

Hirst, S.M. and A.M. Ibrahinl. 1996. Effects of flood protec-

tion on soil fertility in a riverine floodplain area in

Bangladesh. Communications in Soil Science and Plant

Analysis 27:119-156.

Jarrett, R.D. 1985. Determination of roughness coefficients

for streams in Colorado. Water-Resources Investigations

Report 85-4004. Lakewood, CO: U.S. Geological

Survey.Jenny, H. 1962. Model of a rising nitrogen profile in Nile

Valley alluvium, and its agronomic and pedogenic

implications. Soil Science Sociery of America

Proceedins- 26:588-591.

Kintigh, K. W 1985. Setdement, subsistence, and sociery in

late Zuni prehistory. Tucson, AZ: The University of

Arizona Press.

Lagasse, P.E, J.D. Schall, and S.G. Wells. 1990. Runoff and

erosional processes in an arid watershed. Pp 614-632. In

American Sociery of Civil Engineers (editor).

Hydraulics/hydrology of arid lands: Proc~ of an

References CitedAbrahams,A.D.,AJ. Panons, and J. Wainwright. 1995. Effects

of vegetation change on interrill runoff and erosion,

Walnut Gulch, southern Arizona. Geomorphology

13:37-48.

Archer, S. 1994. Woody plant encroachment into

Southwestern grasslands and savannas: Rates, patterns

and proxintate causes. pp. 13-68. In M. Vavra et aI. (edi-

tors). Ecological implications of livestock herbimry in

the West. Denver, CO: Society for Range Management.

Balling, R.CJ. and S.G. Wells. 1990. Historical rainfa1l pat-

terns and arroyo activity within the Zuni River drainage

basin, New Mexico. Annals of the Association of

American Geographers 80:603-617.

Berkes,E.]. Colding, and C. Folke.2000. Rediscovery oftra-

ditional ecological knowledge as adaptive management.

Ecological Applications 10:1251-1262.

Bocco, G. 1991. Traditional knowledge for soil conservation

in Central Mexico. Journal of Soil and Water

Conservation 46:346-348.

Bryan, K. 1925. Date of channel trenching (arroyo cutting)

in the arid southwest. Science 62:338-344.

Bul1, W.B. 1997. Discontinuous ephemeral streams.

Geomorphology 19:227-276.

CleveIand,D.A., FJ. Bowannie,D.E Eriacho,A. Laahty, and E.

Perramond.1995. Zuni farming and United States gov-

ernment policy: The politics of biological diversity in

agriculture. Agriculture and Human Values 12:2-18.

258 jOURNAL OF sOILAND WATER CONSERVATION 5102002