sign in sign up
<<
Home , Check dam

NativeAmerican methods for conservation and restoration of semiarid ephemeral

J.B. Norton, F. Bowannie Jr., P. Peynetsa, W. Quandl!lacy, and S.F.Siebert

ABSTRACT:Combined effects of brush encroachmentand channelentrenchment in the SouthwesternUnited States threaten ecological, hydrological, and agricultural functions of ephemeral streams and associated alluvial surfaces. Restoration is difficult because entrenchmentis widespread,and highly variable flow magnitudesdefy standardizedstructural approaches.Methods developed by the Zuni Indiansduring morethan 2,000years of farmingon dynamicalluvial fans combinebrush removalwith ephemeralchannel- control and show promisefor effectivewatershed-scale conservation and restoration.Zuni utilize severaltypes of simplebrush structures that rely on hydrauliccharacteristics of woody materialto modifyerosive and depositional effects of both small seasonal and irregular storm-flow events. Zuni brush structuresare inexpensiveand quick to build, require no externalmaterial inputs, and avoidthe extensivedisturbance associated with conventional rigid check-damconstruction. Successive surveysat three incised-headwaterephemeral channels on the Zuni Indian Reservationshowed that Zuni techniquesincreased the incidenceof overbankflow during small streamflowevents (acting as permeablecheck ) and large (movingdownstream and forming debris jams) and thereby helped reconnectchannels to alluvial fans.Zuni brush structuresrepresent a potential alternative to capital- and labor-intensive approaches to semiarid watershed conservationand restoration.

Keywords: cutting, conservationfarming, discontinuousephemeral streams, ecological restoration, local knowledge, permeable check dams, runoff agriculture, woody debris, woody speciesencroachment

Arroyo cutting and encroachment by edgein dangerof being lost becauseof drastic woody species have been degrading declinesin traditional land use could be valu- desert and shrub-steppe grasslands of able in addressinga complex, contemporary the Western United States for more than environmentalproblem. 100 years (Archer 1994; Bryan 1925). Large-scale erosion, which severs Although the two processesare closely relat- connections between ephemeral streamflow ed because woody species encroachment and ecologicallyimportant alluvial landforms, increasesrunoff and erosion (Abrahamset al. has defied control for decades(Gellis et al. 1995; Bull 1997), they are typically treated 1995). Conventional approaches,including separately in rangeland conservation and rigid impervious check damsmade of earth, restorationefforts.A comprehensiveapproach concrete, , or logs (Heede 1976; that restoresboth plant communities and sur- face hydrology to more desirableconditions could inlprove the successand cost-effective- ness of ecological restoration efforts. This paper explorestraditional Zuni methods that use brush and trees cleared from degraded rangelandsto control arroyo cutting. It also presentsan example in which local knowl-

Reprinted from dIe Journal of Soil and Woter Conservation Volume 57, Number 5 250 jOURNALOFSOILANDWATERCONSERVATION 5102002 Copyright CI 2002 Soil and Water Conservation Society Seehorn 1992; Zeedyk et al.1995), as well as managedby mdigenous tarnlers who have a the majority of which often falls during soil bioengineering approaches, which use very rich body of traditional knowledge. d1understormsin July,August, and September. live plant materials where sufficient moisture Controlling entrenchment of semiarid The combination of steeperodible sandstone will allow rooting (Gray and Sotir 1996), ephemeral streams is difficult because the and shaleslopes covered by patchy pinyon- address only the erosion part of a persistent extremely variable and dynamic conditions juniper (Pinusedulis-juniperns spp.) and scrub- feedback between gIillying and invasion of create behavior different from that shrub plant conununities wid1 relatively rangelands by woody species. Other restora- where most hydrological principles and ero- intense sununer storms contributes to tion efforts focus on rangeland brush removal sion control methods were developed (Bull dynamic,sand-bed fluvial systemsthat move by cutting, burning, chaining, or other 1997). Small streamflow events that occur large amounts of (Lagasseet al. methods (Connelly et al. 2000) but ignore the once every two to 10 yearscarry most of the 1990). concentration of runoff into that sedimentthat fills and destroysstructures built Zuni subsistenceagriculture dependedon accompanies the shift &om grass to domi- to storewater and control large floods.Large, corn (Zia mays)grown in nonirrigated fields nance by woody species. powerful floods can breach,flank, or undercut in d1esedynamic fluvial settings for more Soils formed in alluvium of headwater check damsdesigned to control erosiondur- than 1,500 years(Kintigh 1985). Socio-eco- stream systems retain runoff, , and ing smallerflows (Gellis et al. 1995). nomic change and assimilation policies nutrients to support biological productivity Though conventional rigid check dams beginning with U.S. occupation of Zuni on-site, protect downstream aquatic resources canmodjfy particular streamreaches (Gellis et lands in d1emid-19dt century causeddrastic (peterson et al. 2001), and support some of al. 1995), the immense extent of ephemeral declinesin agriculture (Clevelandet al. 1995). the oldest agricultures, including that of the drainagenetworks and relativelyhigh cost of Major arroyocutting coincided with this shift Zuni and other Southwestern Native such structures limit their watershed-scale away from traditional farming and led to Americans (Kintigh 1985; Nabhan 1984; effectiveness.Soil bioengineering approaches multiple, U.S.-sponsored erosion-control Sandor 1995), the fam1ers of the Negev (Grayand Sotir 1996) are alsorelatively mate- effortsthroughout d1e20dt century, including (Evenari et al. 1982), fam1ers of the Nile rial- and labor-intensive for such extensive large earthen check darns built by govern- Qenny 1962), and others (e.g., Hillel application,and soil conditions of semiarid- ment agenciesand smaller-scaleefforts by the 1991; Hirst and Ibrahim 1996). Overbank arid ephemeralstreams where arroyo cutting Civilian Conservation Corps (CCC) and flows are crucial to all these functions because is most prevalent may be too dry to support od1erprograms (Gellis et al. 1995). soil productivity is renewed with inputs of fast-growing vegetation. Headwater streams Three actively eroding, low-gradient, moisture, sediment, and organic matter &om often make up as much as 85% of the total entrenched,meandering headwater ephemeral upland hillslopes (Norton et al. 2001). channellength in watersheds(peterson et al. streams-Rosgen "F-type" channels(Rosgen entrenchment eliminates these allu- 2001) and, in semiarid areas,flow only in 1994)-were selected for treatment with vial processes (Bull 1997). The Zuni and responseto rainfall. Large watershedrestora- traditional Zuni brush structures(Laate, Lalio, other agricultural Southwestern Native tion efforts must be simple and quick to and Quandelacysites). These streamsare rep- Americans have used and protected the natu- implement,require minimal capitaland labor resentativeof headwateralluvial systemsand ral productivity of soils derived &om alluvium inputs, and effectively modjfy both small provide important ecological, hydrological, in annual food crop-based farming systems flows and powerful floods. and agricultural functions. for millennia (Hack 1942; Kintigh 1985; Zuni farnlers who still practice traditional The Laatesite is a short ephemeralchannel Nabhan 1984; Sandor 1995). agriculture use simple, low-investment struc- that crossesan orchard high on d1e leeward In general, the diverse ways in which tradi- tures that appear to mitigate destructive side of a sandstonemesa. The channel gradi- tional farmers approach soil and water arroyo cutting during both small and large ent was historically maintained to facilitate conservation reflect deep-seated practical strearnflows.This type of is flooding of d1e sandy eolian soils. In 1996, knowledge about what works and what does often part of brush- and tree-clearingactivi- channelincision threatenedto exclude both not work under particular environnlental ties intended to prepare fields for cultivation mature and recendy establishedfruit trees conditions (Bocco 1991; Siebert and Belsky or to improve livestock grazing lands. Our from streamflows.The Quandelacy site is a 1990). In semiarid environnlents, local objective was to describe and evaluate the continuously incised channel running nom approaches are often adapted to highly performance of indigenous erosion control d1emoud1 of a small through range- variable and unpredictable conditions, a methodsand to evaluatetheir potential appli- lands dominated by big sagebrush(Artemisia characteristic that is increasingly sought in cation in watershed- and ecosystem-scale tridentata).Atthe beginning of our study,this modern management systems (Berkes et al. conservationand restoration efforts. channel ranged from a slot-like gully about 2000). Traditional agricultural systems in 1 m deep by 1 m wide to a 6 m deep,V- dry environnlents depend on incremental Methods and Materials Shapedarroyo and had severalsteep headcuts. methods that emphasize flexibility to take We conducted this study in collaboration The Lalio site is a deeply incised arroyo that advantage of pulses of productivity or to with traditional Zuni farmers on the Zuni empties onto a traditional, hand-cultivated survive drought (Ellis et al. 1993; Niemeijer Indian Reservation in west-central New cornfield. By 1998,d1e arroyo had deepened 1998; Tabor 1995). In the Southwestern Mexico. The Zuni reservation lies on the for a third of the way into the' cornfield, United States, ephemeral stream entrench- southeasternpart of the Colorado Plateauat excluding about three-quarters of the field ment (or arroyo cutting) is a dynamic, 1,800 to 2,400 m elevation and receivesan from flooding. A continuous had formed unpredictable landscape process that has been~ averageof 300 mm of precipitation annually, through d1efield.

I 5102002 VOWME 57 NUMBER 51 251 I A three-part approach for treating and evaluatingstructure effectivenesswas imple- mented at eachstudy site: Treatments.Zuni crewsled by people with lifdong experiencein traditional agricultural practices installed erosion-control structures at the three sitesusing axes,shovels, and post- hole diggers.The Laatesite wasfirst treatedin June 1996,the Lalio site inJune 1998,andthe Quandelacy site in June 1999. Members of the crews recorded labor input, wrote descriptionsof structures,sketched maps and structure designs, and photographed con- struction.The crew leadersare co-authors of this report. Channel morphology.Successive longitudi- nal surveyswere used to document treatment effects on channel morphology (Harrelsonet at. 1994).Surveys were conduct- ed during construction,after flow events,and at the end of the study.(The relativelyremote Laatesite was surveyedonly at the beginning and end of the study period.) Photo points were establishedalong each channel.Surveys included devations of thalweg, structures, tan IX),and K is a constant developed through banks, and water levels identified by high~ Results and Discussion regression analyses for different channel water marks on the arroyo banks. Traditional Zuni brush treatments effec- gradients (K = 0.3 for G S 0.20). Channell connectivity.Pre~ and tively reduced ilie erosive power of small The exact placement of the brush struc- post-study channel dimensionswere used to streamflowevents and large floodS,iliough by tures varied depending upon channel charac- estimatethe probability of overbankflow at very different modes of action. They also teristics and the types of woody material most the Lalio and Quandelacy sites.Manning's effectivelymodified channelmorphology and readily available. Simple brush piles were equation was used to estimatethe peak dis- restored channel/alluvial fan connectivity. the most commonly employed technique charge that would be required to overtop During small flows, ilie structures acted as throughout the channels, with stronger struc- pre- and post-study arroyo banks and then permeable barriers that retained runoff in tures placed in key locations toward distal was applied to two- through 100-yearrecur- temporary pools and accumulatedsediments. ends of the channels and in constricted areas. rence interval equations(Thomas et al. 1997) During larger,more powerful floodS,many of Minjrnizing disturbance to channel beds and to estimate the averagemquency that such ilie structureswere destroyed,but ilie woody banks, using material close at hand, and using overbankflows would occur. debris was redeposited in debris jams that stones only sparingly to compress brush were To place the events and treatment effects altered channel morphology and drastically concepts common to all four kinds of struc- observedin 1999 within the context of long- increasedilie likelihood of overbankflow. tures, which are described below: term Zuni climate, recurrence intervals of TreatmentDesign and LAborRequirements. Brush piles. Limbs of pinyon and juniper both rainfall and streamflowevents were esti- The workers installed four general types of trees and whole, big sagebrush plants were mated. Rainfall volume and intensity were brush structures iliroughout ilie treated piled in channels with larger sterns pointing monitored at a tipping-bucket rain gauge reaches.Construction followed generalprin- downstream (Figure 2a). Material was 2 km from the Lalio site (datacourtesy of the cipalsfrom knowledge about flow eventsbut stomped down repeatedly and layered spar- Zuni Conservation Project, Zuni Pueblo, variedby worker and by streamflowsituation. ingly with stones or larger logs where those New Mexico) and rainfall-event recurrence Spacing between ilie structures (Figure 1) were readily available. Large amounts of sage- intervals were determined with intensity- averagedless than 20 m for all ilie sites,much brush were packed into slot-like channels duration-mquency curves for Albuquerque, closerilian guidelinescalculated wiili Heede's over distances of as much as 30 m by stomp- New Mexico (USWB 1955).The magnitude (1976) spacingrule, which is commonly used ing down layer after layer. Shallow of 1999 streamflowevents was determined by for designing erosion-control projects.This through the Laate Orchard and the Lalio applying Manning's equation to channd spacingrule is basedon ilie equation: cornfield were filled with fine woody materi- al from rubber rabbitbrush (Chyrsothamnus dimensions surveyed after flow events by S = --H using high-water marks to indicate peak KGcosa nauseosus)and broom snakeweed (Gutiemzia stage.These valueswere compared to peak sarothrae).This was the most rapid method and dischargesderived from equationsfor two- where S is spacing(m), H is height (m), had the objective of placing as much debris as through 100-year recurrence interval events G is the channelgradient ratio, a is the angle possible into the channel with a minimum of in the two watersheds(Thomas et at. 1997). correspondingto the channel gradient (G = additional strengthening.

252 JOURNALOF SOIL ANO WATER CONSERVATIONSIO 2002 Figure 2 Types of structures installed at study sites: (a) brush pile with fine branches oriented upstream and stones to compress brush (drawing by Fred Bowannie Jr.); (b) brush check dam with fine branches woven into larger limbs piled across channel. (Tape follows thalweg.); (c) pungie posts immediately upstream from a post-and-wire structure. Posts are about 8 cm in diameter at the base.

(a)

Arroyo Flow . Flow .

Arroyo bed Side View

Top View

Check dams. Simple check damsof inter- Pungie posts. Pungie posts consist of Treatment Effects. Although 1999 was a woven woody material were slightly more clusters of pinyon and juniper limbs and relatively dry year (252 mm, or 80010of nor- intensive than the brush piles. In this sternsset into the arroyo bed to form comb- mal precipitation), 170 mm of rain fell inJuly approach,the largest,longest limbs were laid like structuresthat capturewoody debris and and August, which is 159% of normal precip- acrossthe channel in a pile about 0.5 to 1 m create reinforced debris jams during flow itation (WRCC 2001). Much of this rain high with brushy tops in alternating direc- events(Figure 2c).The postswere set at least carne as unusually intense thunderstorms tions. Smaller limbs were jabbed and woven 60 cm into the ground, angled upstream. (Figure 3), including a greater-than-25-year into the pile so that branchesand leavespre- Placementwas downstreamfrom brush-pile recurrence interval storm on July 5, at more senteda densemat of fine woody material to structuresand upstreamfrom other types to than 18 mm in 15 minutes. The thunder- the upstreamdirection (Figure 2b). Some of protect and reinforce them. Some of the storms generated streamflow events with these dams were reinforced with posts cut pungies were reinforced with post-and-wire widely varying magnitude, including several ttom pinyon or juniper stemsand set imme- braces.Part way through the season,the five- to 25-year recurrence interval flows and diately downstreamttom the brush structure. workers decided to alter this method by four that equaled or exceeded 100-year Post-anJ-wire methods.In this method, leaving fine brancheson the postsso that, if events (one at the Quandelacy site on Aug. '2 workersset two rows of polesabout 1 m apart posts washed out, they would function as and three at the Lalio site on July 5, Aug. 2, acrossthe arroyo channeland packedbrushy woody debris. and Aug. 27), according to measurements material between the polesto createa barrier The crew treated a total of 1.25 km of conducted by hydrologists of the Zuni about 1.5 m high. The tops of the poles were ephemeralstream channel during the study Conservation Project. Because of the remote then wired together.This method requmd period but had to treat segmentsrepeatedly nature of the Laate site, channel din1ensions more effort in sorting and trimming woody after large floods removed structures three were surveyed only during treatment installa- material, setting posts, packing brush, and times at the Lalio site and once at the tion and at the end of the study, so flow wiring braces together. This method was Quandelacysite. Including the repeatedtreat- magnitudes were not estimated for that site. deliberatelyemployed in relativelywide, low- ments, 3.4 km were treated, averaging 93 The structures at the Quandelacy and Lalio energy reaches,at points immediately below person-hourskm-1 of channel.The first treat- sites all withstood the less-than-25-year sharpbends, or in seriesof two to three struc- mentsat the Lalio and Quandelacysites were recurrence interval flows, temporarily retain- tures. Each of these structureswas built by relatively rapid, but the second treatments ing water and causing some sedin1ent deposi- one worker and took two to three hours for required more time to build the reinforced tion (Figures 4 and 5). During larger floods, structuresup to 7 m long. pungie post and post-and-wire structures. however, many of the structures washed out.

15102002 VOWME57 NUMBER5 I 253 I 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 of about 1.5 m at die Lalio site (Figure 6) and 1 m at die Quandelacy site.This caused over- bank flow for nearly 100 m upstream of die debris jam at the Lalio site, which deposited nearly 20 cm of sand on parts of die alluvial fan. Figure 7 shows die pre- and post-study channel profiles at die three study sites.The structures at the Laate site retained sediment and were mosdy buried during die post- treatment survey, except for die most distal structure, which had been installed on bedrock and washed away. The overall effect of die treatments was to restore flooding to larger portions of die allu- vial fans at all three sites.The schematic map of die Lalio site (Figure 8) shows that, at die beginning of the study, flows emerged from the channd near die center of die agricultur- al field. The apex of die fan, or the point where flow leaves die channel and spreads over the alluvial fan, moved upstream after each 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 die agricultural field, but larger, potentially destructive floods would overtop die banks far upstream. Although die response of vegetation in the areas cleared for brush was not measured, increased grass production, particularly blue g= (Boutelouagradlis),was evident compared with adjacent areas. Repeated surveys at channel cross-sections at the Lalio and Quandelacy sites confirm that connectivity between die ephemeral stream channels and alluvial fans was restored (Table 1). At the Lalio site, sediment deposi- tion 100 m upstream from die debris jam reduced the hydraulic radius so that calculat- ed recurrence interval (Thomas 1997) of overbank flow changed from nearly 50,000 years at the beginning of die study to four years at the end. Similarly, at die Quandelacy site, 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 die material for erosion control simultaneously addresses two major ecological degradation processes.Clearing brush and trees that com- pete widi grasses is a common approach to ecological restoration and has been shown to~ Figure 5 The lower part of Lalio arroyo shortly after peak on July 18, 1999. Note the stepwise waterlines behind the structures and the effect of brush clearing along banks.

increase forage production (Archer 1994). consideredfailure of Increased herbaceous ground cover also the structures but increases the hydraulic roughness of soil sur- part of the ongoing faces, causing more and less runoff processof conserving (Bull 1997). The Zuni workers often pruned alluvial fan functions. lower branches on pinyon and juniper trees In thjs traditional rather than cutting whole trees, explaining Zuni approach, as that removing lower branches increased for- much woody materi- age production around trees and increased al aspossible is placed site distances for livestock herding but main- in channels with tained the shade and wind protection provid- minjmal strengthen- ed by the trees. Pruning and/or clearing trees ing. No keys into and brush is often the principal purpose; the bedsor banksand no materials are disposed of in arroyos to control reinforcements with erosion. stonesor gabionsare Our results are supported by a previous used.Instead of rely- study on the Zuni Indian Reservation in ing on rigid check dams,which can cause explained how, although the approach was which Gellis et al. (1995) found that 28 of 47 scouring of streambedsor banks (Gellis et al. learned from older Zuni people, methods large, heavy-equipment-built earth and rock 1995), erosion control is achieved through were constantly modified to fit specific erosion-control structures they evaluated had hydraulic characteristicsof woody material. conditions. The brush structures reflect the breached, and 31 were more than half-filled Although there arefew publishedaccounts of constant experimentation and innovation with sediment in the 20 to 30 years since permeable brush structures for controlling characteristicof traditional ecologicalknowl- construction. In contrast, only five of 23 arroyo cutting, woody debris is commonly edge (Berkeset al. 2000). brush structures had washed out or were used to improve or restoreaquatic habitat in Studies of woody debris suggestthat the breached, while the remaining structures perennial streams. Zuni methods could be effective for increas- continued to function. The brush structures There are similarities between the struc- ing .Woody debris is responsi- were built in the 1970s by a Youth tures built during our study and methods ble for as much as 87% of channel sediment Conservation Corps team led by an elderly usedduring the 1930sby CCC crewsat Zuni storagein small woodland streams,with 39% Zuni farmer. Gellis et al. (1995) attributed and elsewhere(USEPA 1999). However,the or more of the individual pieces forming failures of both the large rock and earthen CCC methods all include stone and depositional sites(Fetherston et al. 1995).As structures and the Zuni brush structures to keying structures into beds or banks,while such,woody debris is a primary determinant lack of maintenance. Because they require no the Zuni methods purposefully avoid these of channel morphology, forming pools and heavy earth-moving equipment or offiite techniques.Nevertheless, the similaritiessug- regulating transport of sediment, organic materials, the brush structures would be easier gest that some of the techniquescould have matter, and nutrients.While lack of woody and less expensive for local people to build been handed down from Zuni people who debrisis often recognizedas a characteristicof and maintain. Unlike capital- and labor- servedon CCC crews.The CCC techniques dryland streams,Minckley and Rinne (1985) intensive rigid darns, modification of Zuni may havebeen adaptedby Zuni people to fit suggestthat streamsidevegetation was abun- brush structures by streamflow events is not local conditions. The workers in our study dant before destruction of riparian forests

15/02002 VOLUME57 NUMBER51 255 I

~ vating the water-surfaceprofile and signifi- cantly increasingflood travel times (Gippel 1995). In one study,woody debris covered lessthan 2% of the streambedbut provided 500/0of total flow resistance(Mmga and Kirchner 2000).After initial treatmentsin our study,woody material coverednearly 50% of the channelbed of the treatedreaches. Implications for Watershed Restoration. The labor and materialsrequired to construct brush piles suggestthat this method could be implemented rapidly and inexpensivelyover entire watersheds.At an averageof93 person- hours kIn-! of entrenchedchannel treated,a 20-personcrew could treat more than 1.7 kIn per eight-hour day. Our work during an unusuallywet and stormy year required that we treat the studyreaches repeatedly after 50- to 100-yearrecurrence events. These unusu- ally frequent and high magnitudestorm flows came almost inunediately after structures were installed,before smaller eventsbegan to bury the woody material with sedimentsand hold them in place.This partly explains the need to repeatedlytreat the reaches. Over the long term and across whole watersheds,treated headwaterchannels may need additional woody material added at about 20-year intervals, after large floods remove woody debris or as brush darns fill with sediment.This observationis supported by Gellis et al. (1995),who found that five of 28 brush structuresneeded maintenance20 yearsafter installation.More frequent debris additionsafter flows that causesedimentation would raise channel-bed elevations and thereby increasethe likelihood of overbank flows and depositionin agricultural fields. These Zuni erosion control practices are closely associatedwith traditional rainfed corn (Zea mays)farming, which provided the basisof Zuni subsistenceuntil the late 19th century (Kintigh 1985).Since then, changing socio-economic conditions, devastatingepi- demics, and u.S. agricultural development policies caused drastic declines in rainfed farming (Oeveland et al. 1995;Ferguson and Hart 1985).The cessationof traditional farm- ing in ephemeralstream coursescoincided with the onset of stream-channelincision at Zuni (Balling and Wells 1990; Bryan 1925). This suggeststhat the absenceof previously widespreadtraditional erosion-controlactivi- ties may have contributed to the severity of channel entrenchment that is now wide- spreadin the Zuni area. The Zuni approachto controlling arroyo

256 JOURNALOF SOil AND WATERCONSERVATION 5102002 Figure 8 Sketch map showing chronology of flood events and reaction of treatments at the Lalio site. cutting may be relevantto large-scalerange- Drawing by Fred Bowannie Jr. See discussion in text. land watershed restoration in the Western United States.Implementing this approach could be as simple as using woody material displacedduring chaining or cutting to build the simple permeablebrush check dams in nearby incisedwatercourses. While this addi- tional work would be relativelyinexpensive, it could effectivelyreduce erosion and increase sedimentationover wide areas. The Zuni methods could be useful in other semiarid regions where incised ephemeralstream channels have sand or finer beds,flow eventsmove large amounts of and suspendedsediments, and there is abundant fine woody material near eroding Debris channels.An important characteristicof the jams Zuni approachis that farmersmodify erosion control methods to fit particular situations (i.e.,channel geometry or the typesof woody material available)or objectives(i.e., control, prevention, or restoration of connectivity between channels and surrounding alluvial + surfaces).As such, the Zuni methods would probably be most applicable if built upon local environmental knowledge and tradi- I , tions. Traditional, low-input erosion-control practices are widely used throughout the flows and die area of alluvial fans flooded by time in strengtheningstructures, Zuni farm- world (e.g., Mexico, Bocco 1991; and recurrenceinterval flows of lessthan or equal ers and livestock herdersfocus on placing as Indonesia, Siebert and Belsky 1990) and to five years.Frequent, low-intensity floods much woody material as possibleover long typically displayadaptations to local environ- are important to renewal of desirable soil channelreaches in an attempt to reduceflood mental and cultural conditions,as well asbasic properties.These methodsare rapid and inex- power, move fan apexes upstream, and principles of potential use elsewhere. pensiveto implement and thus may be useful increasedle ttequency and area of overbank in watershed-scalerestoration efforts.While flows. The Zuni medlods combine clearing Summary and Conclusion most conventional erosion-control methods brush and trees widl erosion control in an Traditional Zuni erosion-control mediods rely on rigid check damsof concrete,stones, efficient, comprehensiveapproach to water- effectively modify channel morphology of gabions,and/or logs keyed into streambeds shedrestoration. entrenched headwater ephemeral streams and banks,the Zuni brush structuresrely on The approachdescribed herein is basedon after small flows and powerful floods in ways hydraulic characteristicsof woody debris to long-term understanding of dle range of that increase die occurrence of overbank modify erosive flows. Rather than investing variability 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 with channel dimensions at the beginning and end of the study period. See discussion in text. Parameter Units Lalio Lalio Quandelacy Quandelacy overbank, overbank, overbank, overbank, pretreatment post-treatment pretreatment post-treatment

Watershed area ha 36 36 17 17 Station' m 320 320 130 130 Slope 0.02 0.006 0.035 0.01 Cross-sectional area m2 4.5 1.08 1.94 0.58 Hydraulic radius m 0.58 0.20 0.44 0.15 Manning's n" 0.026 0.026 0.026 0.026 Peak discharge m3s2 17.02 1.10 8.07 0.63 Recurrence interval'" years never 4 never 3 * 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).

ISIO2002, VOWME57 NUMBER5 I 257 I ment of the SouthwesternUnited States.It Connelly,J.W, M.A. Schroeder,A.R. Sands, and C.E. Braun. international symposium. American Society of Civil incorporatessome methodsbrought to Zuni 2000. Guiddines ro manage sage grouse populations and Engineers. Fort Collins, CO. rheir habirats.Wildlife Sociery Bulletin 28:%7-985. Manga, M. and J.W Ki11:hner. 2000. Stress partitioning in during nearly 100 yearsof U.S. government- Ellis,J.E., M.B. Coughenour, and D.M. Swift. 1993. Climate streams by large woody debris. Water Resources sponsorederosion control but adaptsthem to variability, ecosystem srability, and the implications for Research 36:2373-2379. fit local conservationobjectives and tradition- range and livestock development.p 248. In R.HJ. Minckley, WL. and J.N. Rinne. 1985. Large woody debris in al environmentalknowledge. Traditional Zuni Behnke et al. (editors). Range ecology at disequilibrium: hotcdesert streams: An historical review. Desert Plants soil conservation measures allow local New model.. of narural variability and pastoral adapra- 7:142-153. tion in African savannas. London: Overseas Nabhan, G. 1984. Soil fertility renewal and water harvesting residentsto effectively conserveand restore Development Instirute. in Sonoran Desert agriculture: The Papago example. natural resourceswithout expensiveequip- Evenari, M.L., L. Shanan, and N. Tadmor. 1982. The Negev: Arid Lands Newsletter 20:21-28. ment. These methods may also represent a The challenge of the desert. Cambridge, MA: Harvard Niemeijer, D. 1998. Soil nutrient harvesting in indigenous comprehensive,cost-effective approach to University Press. teras water harvesting in Eastern Sudan. Land Ferguson, TJ., and E.R. Hart. 1985. A Zuni adas. Norman, Degradation & Development 9:323-330. large-scalewatershed restoration applicable OK: University of Oklahoma Press. Norton,J.B., R.R. Pawluk, andJ.A. Sandor. 2001. Farmer- to other semiarid environments around Fetherston, K.L., RJ. Naiman, and R.E. Bilby. 1995. Large scientist collaboration for research and agricul=al the world, particularly when combined woody debris, physical process, and riparian forest development on the Zuni Indian Reservation, New with local knowledge of runoff and erosion development in monrane networks of the Pacific Mexico, USA. Pp 107-120. In W A. Payne and D. characteristics. Northwest. Geomorphology 13:133-144. Keeney (editon). Sustainability in agricul=al systemsin Gellis, A.C., A. Cheama, v: Laahty, and S. Lalio. 1995. transition. American Society of Agronomy. Madison, Assessment of gully-control strucrures in the Rio WI. Acknowledgement Nutria watershed, Zuni Reservation, New Mexico. Peterson, BJ., WM. Wollheim, PJ. MulhoI1and,J.R. Webster, This work was funded by National Science Water Resources Bulletin 31:633-646. J.L. Meyer, J.L. Tank, E. Marti, WB. Bowden, H.M. Foundation grant DEB-9S284S8 and is dedicated Gippel, CJ. 1995. Environmenral hydeaulics oflatKe woody Valett, A.E. Hershey, WHo McDowell, WK. Dodds, S.K. debris in streams and . Journal of Environmenral Hamilton, S. Gregory, and D.D. Morrall. 2001. Control to the memory of Andrew H. Laahty,who lead this Engineering 121:388-395. of nitrogen export fium watersheds by headwater research effort as director of the Zuni Sustainable Gray, D.H., and R.B. Sotir. 1996. Biotechnical and soil bio- streams. Science 292:86-90. Agriculture Project. We thank the Zuni Tribe and engineering slope srabi1ization: A practical guide for Rosgen, D.L. 1994. A c1assification of na=al riven. Catena Zuni Conservation Program for supporting this erosion control. New York, NY:John Wiley &: Sons. 22:169-199. Hack,J. T. 1942. The changing physical environment of the Sandor,J.A. 1995. Searching soil for clues about Southwest research. Craig Goodwin's review contributed to Hopi indian. of Arizona. Papers of the Peabody prehistoric agriculture. In H:WToll (editor). ProceedinS' the accuracy of our report. Museum of American Archaeology and Ethnology, of the New Mexico Archaeological Council Symposium Harvard 35:1-85. on Southwest Agriculture. References Cited Harrelson, C.C., C.L. Rawlins, and J.P. Potyondy. 1994. Seehorn, M.E. 1992. Stream habitat improvement hand- Abrahams,A.D.,AJ. Panons, and J. Wainwright. 1995. Effects Stream channel reference sites: An illustrated guide to book. Technical Publication R8-TP 16. Atlanta, GA: of vegetation change on interrill runoff and erosion, field technique. General Technical Report GTR RM- U.S. Forest Service. Walnut , southern Arizona. Geomorphology 245. U.S> Department of Agriculture (USDA) Forest Siebert, S.E, and J.M. Belsky. 1990. Bench terracing in the 13:37-48. Service. Kerinci uplands of Sumatra, Indonesia. Journal of Soil Archer, S. 1994. Woody plant encroachment into Heede, B.H. 1976. Gully development and control. Research and Water Conservation 45:559-562. Southwestern grasslands and savannas: Rates, patterns Paper RM-169. USDA. Tabor,J.A.1995.improving crop yidds in the Sahel by means and proxintate causes.pp. 13-68. In M. Vavra et aI. (edi- Hillel, D. 1991. Out of the Earth: Civilization and the life of of water-harvesting. Journal of Arid Environments tors). Ecological implications of livestock herbimry in the soil. Berkeley, CA: University of California Press. 30:83-106. the West. Denver, CO: Society for Range Management. Hirst, S.M. and A.M. Ibrahinl. 1996. Effects of flood protec- Thomas, B.E., H. W Hjalmarson, and S.D.Waltemeyer. 1997. Balling, R.CJ. and S.G. Wells. 1990. Historical rainfa1l pat- tion on soil fertility in a riverine area in Methods for estimating magnitude and &equency of terns and arroyo activity within the Zuni River drainage Bangladesh. Communications in Soil Science and Plant floods in the Southwestern United States.Water Supply basin, New Mexico. Annals of the Association of Analysis 27:119-156. Paper No. 2433. Denver, CO: U.S. Geological Sur"")' American Geographers 80:603-617. Jarrett, R.D. 1985. Determination of roughness coefficients U.S. Environmental Protection Agency. 1999. Unified water- Berkes,E.]. Colding, and C. Folke.2000. Rediscovery oftra- for streamsin Colorado. Water-Resources Investigations shed assessment.Albuquerque, NM: USEPA. ditional ecological knowledge as adaptive management. Report 85-4004. Lakewood, CO: U.S. Geological U.S. Weather Bureau. 1955. Rainfal1 intensitycduration-&e- Ecological Applications 10:1251-1262. Survey. quency curves. Technical Paper No. 25. Hydrologic Bocco, G. 1991. Traditional knowledge for soil conservation Jenny, H. 1962. Model of a rising nitrogen profile in Nile Services Division, USWB. Washington, D.C. in Central Mexico. Journal of Soil and Water Valley alluvium, and its agronomic and pedogenic Western Regional Climate Center. 2001. Zuni, New Conservation 46:346-348. implications. Soil Science Sociery of America Mexico, monthly total precipitation (Station 299897). Bryan, K. 1925. Date of channel trenching (arroyo cutting) Proceedins- 26:588-591. Available from WRCC at www.wrcc.dri.edu/cgi- in the arid southwest. Science 62:338-344. Kintigh, K. W 1985. Setdement, subsistence, and sociery in bin/c1iMAIN.pl?nmzuni (verified November). Bul1, W.B. 1997. Discontinuous ephemeral streams. late Zuni prehistory. Tucson, AZ: The University of Zeedyk, B., B. Romero, and S.K. Albert. 1995. Using simple Geomorphology 19:227-276. Arizona Press. structures for flow dispersion in wet meadow restoration. CleveIand,D.A., FJ. Bowannie,D.E Eriacho,A. Laahty, and E. Lagasse, P.E, J.D. Schall, and S.G. Wells. 1990. Runoff and Desired future conditions for Southwestern riparian Perramond.1995. Zuni farming and United States gov- erosional processesin an arid watershed. Pp 614-632. In ecosystems: Bringing interests and concerns together ernment policy: The politics of biological diversity in American Sociery of Civil Engineers (editor). RM GTR-272. Fort Collins, CO: Forest Service. agriculture. Agriculture and Human Values 12:2-18. Hydraulics/hydrology of arid lands: Proc~ of an

258 jOURNALOF sOILAND WATER CONSERVATION 5102002