Opportunities for Ecological Improvement Along the Lower

WETLANDS, Vol. 18, No . 4, December 1998, pp. 513-529© 1998, The Society of Wetland Scientists

OPPORTUNITIES FOR ECOLOGICAL IMPROVEMENT ALONG THELOWER COLORADO RIVER AND DELTA

Mark K. Briggs and Steve CorneliusSonoran Institute

7650 E. Broadway Blvd., Suite 203Tucson, Arizona, USA 85710

Abstract : The lower Colorado River mainstem and delta have been severely damaged by a variety ofhuman-related activities, including river impoundment, agriculture, water diversions, introduction of exoticplants and fishes, and ground-water pumping . In some areas, the native wetland habitat that formerly dom-inated this region has disappeared completely . Nevertheless, there are areas where significant wetland habitatpersists as a result of incidental circumstances or purposeful restoration actions . These areas provide impor-tant conservation and restoration opportunities . In this investigation, nine restoration efforts along the lowerColorado River from Parker Dam to the delta region were evaluated to learn how lessons from these ex-periences can benefit future ecological restoration efforts . In addition, we assessed the general ecologicalcondition of this reach to identify critical native wetland plant communities and recommend strategies forprotecting these areas in the future . It is apparent that wetland ecosystems in both the delta and the mainstemwould benefit if effluent waters were allocated to support wetlands rather than allocated to evaporative basins .Other important strategies for improving the ecological condition of the river should include altering reservoirreleases, improving the effectiveness of revegetation efforts, and developing bi-national, collaborative ap-proaches involving local communities and landowners to identify and carry out projects that benefit boththem and the ecological condition of the river.

Key Words : river impoundment, restoration, lower Colorado River, water use, reservoir water releases

INTRODUCTIONThe goal of this investigation was to identify op-

portunities and strategies for improving the ecologicalcondition of the lower Colorado River . This investi-gation focused on the lower portion of the river, whichincludes the mainstem from Parker Dam to the delta,as well as the delta itself. This effort has three prin-cipal objectives . First, the effectiveness of past riparianrestoration efforts along the lower Colorado River(from Parker Dam to the river delta) were evaluatedso that lessons gained from these experiences can beapplied to future ecological restoration activities. Sec-ond, the ecological condition of this reach of the lowerColorado River was assessed, with particular focus onidentifying areas that contain significant native wet-land habitat or show promise for future restoration ac-tivities . Third, courses of action were recommendedfor enhancing damaged areas and maintaining areas ofnatural significance .

CHANGES ALONG THE LOWERCOLORADO RIVER

The Colorado River watershed is a vast system.From its headwaters in the Rocky Mountains to the

513

Sea of Cortez, the river travels over 2,250 km anddrains an area of 632,000 square km, which includes5,180 square km of northern Mexico . Even within thenarrow focus of this investigation-Parker Dam to theSea of Cortez-there are wide variations in land-usepatterns and physical and biological conditions (Figure1). The current hydrologic, physical, and biologicalcharacteristics of this reach and the changes that havetaken place in these parameters since the constructionof Hoover Dam are fairly well documented (Ohmartet al . 1977, Glenn et al . 1992, Abarca et al. 1993,Glenn et al. 1996) and are reviewed only briefly here .

Historic accounts seem to indicate that the lowerColorado River's riparian ecosystems changed littlefrom the time of early Spanish exploration in the 17thcentury to the 1930s when construction of HooverDam was completed (Ohmart et al . 1977). The com-pletion of Hoover Dam in 1935 sparked a wave ofmajor construction and agricultural projects along theriver that have significantly affected the river's eco-logical condition . Today, as the Colorado River flowsfrom Hoover Dam to the delta, it passes through 28dams, irrigates over 1 million hectares of agriculturalland, and serves or supplements water supplies for

5 14

Site #4Dredge Spoil SiteRevegetationHabitat Creation

Site #5Farmers Market BridgeRevegetationHabitat CreationRiverside County

Site #6Three-Finger Lake SitesLand Forming/RevegetationHabitat Creation

Imperial DamLaguna Dam

International BoundaryAm

Morelos Dam - Yuma20 30

40 km

QunaatlonalScale

.~

over 20 million people in the U .S . and Mexico (Col-orado River Basin Salinity Control Forum 1990) .

The Mainstem

The combined environmental impacts of river im-poundment, river diversions, ground-water pumping,spread of non-native species, agricultural activities,and other human activities has had a devastating effecton the river's ecology. In particular, the buffering ofannual overflows and altering of natural channel dy-namics by river impoundment has compromised hab-itat for native fishes and limited the creation of sand-bars and channel islands, which are critical featuresfor the propagation of many native riparian plants(Ohmart et al . 1977)

The completion of Hoover Dam, and then GlenCanyon Dam in 1963, has significantly affectedstreamflow. Prior to the dams, the Colorado River wasa warm, muddy flow with tremendous seasonal fluc-tuation. After dam construction, the river became amuch clearer flow of cold water that fluctuated rela-tively little . Such hydrologic changes have adverselyeffected the river's native population of warm-waterfish, such as the razorback sucker (Xyrauchen texanusAbbott). bluehead sucker (Catostonms discobolusCope), flannelmouth (Catostomus latipinnis Baird and

Bill Williams River

Site # 2No Name LakeRevegetationMitigation

She # 3Cibola NWR Saline SitesRevegetationHabitat Creation/Experiment

She #7Imperial WR CottonwoodSheRevegetationHabitat Creation

She #8Mittry LakeRevegetation/Mitigation

WETLANDS, Volume 18, No. 4, 1998

Site 1Ahakhav Tribal PreserveRevegetationHabitat Enhancement

Site 9Fortuna LakeRevegetationMitigation

Site #10Colorado-Gila RiverConfluenceNatural Regeneration

Figure 1 . Idealized map of Colorado River mainstem from Parker Dam to the international boundary . Sites included in thisinvestigation are labeled and their approximate location is indicated .

Girard), Colorado squawfish (Ptychocheilus lucius Gi-rard), humpback chub (Gila cypha Miller), and bony-tail chub (Gila elegans Baird and Girard), while ben-efiting such non-native fish as the rainbow trout (Sal-mo gairdneri Richardson) and the largemouth bass(Micropterus salmoides Lacepede), which are betteradapted to the river's artificially created clear, cold wa-ters (Minckley 1991). With the exception of the blue-head sucker and speckled dace (Rhinichthys osculusGirard), all of the river's native fish are either endan-gered or under consideration for federal listing.

The distribution and extent of native riparian forests,such as the cottonwood/willow (Populus fremontiiWats./Salix gooddingii Ball (S. Nigra Marsh var. val-licola Dudley)) forests, and wetland ecosystems, suchas the cattail/rush (Typha spp.lJuncus spp.) marsh-lands, have changed significantly along parts of thelower Colorado River system. Ohmart et al. (1977)observed, for example, that cottonwood communitiesalong the mainstem have decreased from over 2,000hectares in the 1600s to less than 200 hectares . Salt-cedar (Tamarix ra,nosissuna Ledeb.), which was intro-duced to the western U .S. during the mid-1800s as asoil stabilizer and ornamental plant, now forms ho-mogeneous stands along significant reaches of the low-er Colorado River (Ohmart et al . 1977) .

Lower Colorado River water is projected to become

Briggs & Cornelius, ECOLOGICAL IMPROVEMENT IN THE LOWER COLORADO RIVER

progressively more saline due to a variety of human-related activities . The U.S. Environmental ProtectionAgency (EPA) estimated that increased salinity con-centrations in the river have been principally causedby out-of-basin exports, irrigation, and reservoir evap-oration-accounting for three percent, 37 percent, and12 percent, respectively, of the increased salinity con-centrations that occurred between 1944 and 1988 (Col-orado River Basin Salinity Control Forum 1990) .

The Delta

Prior to the construction of Glenn Canyon and Hoo-ver Dams, Colorado River water continually reachedthe delta and the Sea of Cortez, providing nutrientsand estuarine habitat for a plethora of marine life . Dur-ing this time, the silt and water that the river broughtto the delta were critical in sustaining dense wetlandplant communities that are estimated to have contained200 to 400 different species (Ezcurra et al . 1988 citedin Glenn et al . 1996). The area occupied by the deltaprior to dam construction is estimated at over 780,000ha and included two below-- sea-level depressions, theSalton Sea and the Laguna Salada (Sykes 1937) . Tidalmarsh and brackish and riparian ecosystems supportedjaguar, beaver, and thousands of migratory and residentwaterfowl . In addition, the delta was the place of set-tlement for the Cocopah-speakers of a Yuman familylanguage who occupied parts of the lower ColoradoRiver and delta for over 2,000 years (Alvarez de Wil-liams 1978). At their height during the early 1600s,the population of the Cocopah communities in the del-ta probably exceeded 6,000 people, who supportedthemselves by fishing, hunting, and gathering in thelush delta environment . As a result of the ecologicaldecline of the delta and outside population pressures,the population of the Cocopah in Mexico has declineddramatically and was estimated in 1980 at 571 (Al-varez de Williams 1978) .

Today, over 1 million ha of land in and surroundingthe delta has been converted to farmland . As a resultof river impoundment and water diversions, river wa-ter rarely flows all the way to the Sea of Cortez, al-tering the natural salinity balance and decreasing theflow of nutrients that supports upper Sea of Cortezfisheries (Glenn et al . 1996). In addition, reduced siltloads due to river impoundment have actually sparkeda period of erosion in the delta, rather than accretion(Thompson 1968) . Therefore, the size of the delta willprobably decrease over time .

The 1944 Mexican Water Treaty allocates roughly1,850 kmj of the lower Colorado River's base flow toMexico per year, but provides a pro rata reduction intimes of shortages (Pontius 1997) . However: even dur-ing times of sufficient flow, much of this water is di-

515

verted to the Canal Central for agricultural irrigationin the Mexicali and San Luis districts of Mexico(Glenn et al . 1996). As a result, mainstem water reach-es delta wetlands only during times of high flow .

The combination of river impoundment and diver-sions has had a devastating effect on delta wetlands .In areas formerly dominated by cattails and riparianforests of cottonwoods or willows, a significantamount of the delta region south of the farmland nowconsists of dry sand, mud, and salt flats dominated bysaltcedar, arrowweed (Pluchea sericea (Nutt.) Coville),and iodine-bush (Allenrolfea occidentalis (Wats .)Kuntze). Freshwater and brackish habitat still remain,but these areas are confined for the most part to agri-cultural wastewater discharge points, artesian springs,and areas influenced by tidal fluctuations .

Despite the tremendous ecological changes thathave occurred in the Colorado River delta, it is im-portant to emphasize that the delta still contains sig-nificant wetland and riparian plant communities. Par-ticularly during the last decade, the amount of Colo-rado River, as well as agricultural return flows, thathas reached the delta seems to have increased (prob-ably due in part to the filling of Lake Powell), helpingto maintain several key intertidal, brackish wetlandsand riparian forests south of the agricultural fields(Glenn et al . 1992, Payne et al . 1992) .

The principal wetlands in the delta are (1) the RioHardy wetlands, which are supported by the Rio HardyRiver and high flow events in the Colorado River ; (2)the Cienega de Santa Clara, sustained by agriculturalrunoff emanating from the Wellton-Mohawk canal andthe Riito drain ; and (3) the El Doctor wetlands, whichare supported by artesian springs (Glenn et al . 1996)[Figure 2]. In addition, riparian forests dominated bycottonwood/Goodding willow/saltcedar have estab-lished in several locations along the main channel ofthe Colorado River just north of the areas influencedby tidal fluctuations of the Sea of Cortez .

These wetland and riparian ecosystems are criticalto a variety of wildlife. The Cienega de Santa Clara,for example, provides habitat for the endangered desertpupfish (Cyprinodon macularis Baird and Girard) andthe Yuma clapper rail (Ralus longirostris yurnanensisVieillot) (Abarca et al. 1993). The delta's ecologicaldecline also appears to be intricately related to the de-cline of two other endangered species: the totoaba fish(Cynoscion macdonaldi Gilbert), which was oncecommon throughout much of the delta (Cisneros-Mataet al. 1995); and the vaquita porpoise (Phoceona sinusNorris and McFarland), a small harbor porpoise thatis heavily dependent on the delta's protected watersand nutrient supply (Moralis and Abril 1994 . Turk-Boyer pers. comm.) .

5 16

METHODS

Background information was collected describingthe current ecological condition of the lower ColoradoRiver, the degree to which that condition has changed,and some of the principal reasons behind the ecolog-ical changes that have occurred . Maps and aerial pho-tographs were collected, and past studies dealing withthe river's ecological condition were reviewed. Ecol-ogists, hydrologists, natural resource managers, resto-ration ecologists, and others with lower Colorado Riv-er experience were interviewed to better understandthe challenges that will have to be overcome beforesignificant progress can be made in improving the riv-er's ecological condition .

Sites were selected in consultation with local expertswith regard to their natural significance or the infor-mation they may convey to future restoration efforts .Sites were then visited and field work conducted withpersonnel who have either studied the area or wereinvolved in the restoration effort .

For each site visited, the following information wasgathered :(1) general background information, including site lo-cation, the size of the restoration project, date of pro-ject implementation and completion, and project ob-jective ;(2) general ecological condition, including the currentcomposition and structure of the site's plant commu-

California

WETLANDS, Volume 18, No . 4, 1998

Arizona

Site #11Rio Hardy WetlandsBackwater area of the ColoradoRiver

Site #12Cienega de Santa ClaraArtificially-supported System

Figure 2 . Idealized map of the Colorado River delta and location of sites considered in this investigation .

nity, obvious signs of disease or perturbation (e.g .,mistletoe (Phoradendron spp.) infestation, significantpopulation of non-native species, leaves showingsymptoms of high salt concentrations), significantchanges in ecological condition (e.g ., dramatic changesin depth to saturated soils or streamfiow characteris-tics), signs of soil salinity problems, and obvious signsof erosion ;(3) restoration strategies employed (where appropri-ate), including information describing the restorationmethods, how the strategies were developed, post-pro-ject maintenance and monitoring, and evaluation of theeffort (how effective the restoration effort was inachieving project objectives), and a general descriptionof the present condition of the site ; and(4) lessons learned, describing how the experiencesgained from this review can be used to improve theeffectiveness of future restoration and conservation ef-forts along the lower Colorado River.

SITE DESCRIPTIONS

In total, 12 sites were evaluated as part of this in-vestigation . Of these, nine are restoration efforts thatwere completed along the river's mainstem, and all butone of these (Three Finger Lake, project #6) used re-vegetation as the principal strategy . The restoration ef-forts ranged significantly in size and scope . Some en-


compassed less than a hectare and involved fewer thana hundred plantings ; others encompassed a much larg-er area and involved thousands of plantings . Three ar-eas of natural significance were also included in thiseffort. These areas were identified by citizens, scien-tists, and natural resource practitioners as areas thatcontain significant amounts of native riparian or wet-land plant communities that are not the result of res-toration efforts (Valdes, Barrera, and Swett pers .comm.) . Of the three sites, one is along the mainstemof the river and the other two are located in the delta .A fourth significant natural area-the El Doctor wet-lands of the delta-is only mentioned briefly in thediscussion . Sites are labeled numerically from up-stream (just south of Parker Dam) to downstream .

Site #1-'Ahakhav Tribal Preserve

Location and Size. This 61-ha revegetation site lieswithin the 405 ha 'Ahakhav Tribal Preserve, which ismanaged by the Colorado River Indian Tribes (CRIT)and lies approximately 3 krn downstream of HeadgateDam.

Objective . To enhance habitat for wildlife and fish,as well as to provide environmental education, outdoorrecreation, and cultural opportunities for tribal andcommunity members .

Completion Date . Planting for this part of the 'Aha-khav Tribal Preserve began in 1996 and will be com-pleted during 1998 . Additional riparian restoration ac-tivities (mostly revegetation efforts) are planned forthe immediate future .

Pre-Project Site Conditions . Pre-project vegetationconsisted primarily of saltcedar, arrowweed, honeymesquite (Prosopis glandulosa Torrey) and screwbeanmesquite (Prosopis pubescens Bentham) . Soils aremostly sandy .

Project Strategy . Prior to project implementation,site characteristics were evaluated with regard to soilsalinity, soil texture, and depth to saturated soils . Un-desirable plants were removed with a bulldozer . Over10,000 screwbean mesquite, honey mesquite, cotton-wood, and Goodding willow seedlings were plantedon over 40 ha of land. The site evaluation allowedrevegetation practitioners to develop a detailed map ofsite conditions, allowing them to place plant materialsin areas characterized by tolerable soil salinity and wa-ter availability conditions. Mesquites were planted inareas characterized by relatively high soil salinity(electroconductivity levels in excess of 2 dS m - ') .Mesquite were placed in the ground as seedlings (asopposed to cuttings or poles) . Cottonwood and willowpoles were started in a nursery and planted in areas

517

characterized by low soil salinity . Once in the ground,all plants were irrigated with a drip irrigation system .Plants will be irrigated until roots are considered tohave reached saturated soils .

Results . Ninety percent of all plants have survived todate. A significant portion of these have grown to aheight of almost 5 m. According to project managers,the high survival rates of planted vegetation demon-strate the importance of mapping site characteristics,particularly soil salinity and water availability char-acteristics . In addition, taking advantage of low-lyingtopographical features where water availability is high(e.g ., secondary channels) was also considered criticalto the survival of obligate riparian species such as cot-tonwoods and willows (Shaffer pers . comm.) .

Site #2-No Name Lake

Location and Size . The 17 ha site is about one riverkm downstream from Angnes Wilson Bridge and liesadjacent to farmland managed by the Colorado RiverIndian Reservation .

Objective . To re-establish native riparian plants in anarea where the river's riparian habitat has been com-promised by agricultural activities and the constructionof Parker Dam.

Completion Date . 1987 .

Pre-Project Site Conditions . Pre-project, on-site veg-etation consisted predominately of arrowweed, withsome saltcedar, screwbean mesquite, and willow . Thewater-table depth was estimated to vary from 4 m onthe upstream end of the site to about 1 .5 m on thedownstream end (Pinkney 1992) .

Project Strategy. The site was cleared and root-ripped in April 1987 . On-site, desirable vegetation wasnot disturbed . A 38-cm-diameter auger was used todisrupt the soil down to the saturated zone . Five daysprior to planting, salts were leached from soils by floodirrigating. Approximately 1,380 cottonwoods, 370 wil-lows, 3,560 honey mesquites, 45 palo verdes (Cerci-dium microphyllum (Ton.) Rose & I.M. Johnston), 80California fan palms (Washingtonia filifera (L. Lin-den) H. Wendl), and 100 quailbrush (Atriplex lentifor-mnis (Tory.) S. Watson) were planted. Trees and shrubswere in cardboard tubes 10 cm in diameter and 40 cmlong. Nitrogen fertilizer was applied using a liquid in-jector system. Undesirable vegetation was controlledby applying "Arsenal," and a systemic insecticide"Orthene" was applied to control physllids . All plantswere irrigated with a drip system (45 .5 1 each day forthe first 30 days; reduced to 3 days a week throughSeptember 1987) (Pinkney 1992) . Cottonwoods and

5 1 8

willows were also planted in an erosional depressionlocated immediately adjacent to the agricultural fields .The depression is roughly 1 .5 ha in area and lies ap-proximately two meters below the elevation of sur-rounding lands .

Results. As of 1990, approximately 53 percent of thetotal numbers of individuals originally planted had sur-vived (Pinkney 1992). Of this amount, more than 90percent were honey mesquite and palo verde . The onlyother plant species with more than 50 percent survivalwas quailbrush . All of these plants occur outside theerosional depression. Of all the cottonwoods and wil-lows that were planted, only those planted in the ero-sional depression survived, where water availabilitywas much greater due to periodic irrigation runofffrom an adjacent agricultural field . Cottonwood andwillow in this depression are over 18 m high and ap-pear healthy . As long as the adjacent field is irrigated,this area will have sufficient moisture to support theexisting cottonwood and willow. As of 1997, honeymesquite, palo verde, and quailbrush have establishedin scattered locations throughout the south end of thesite .

Site #3-Cibola National Wildlife Refuge Saline Site

Location and Size . The Cibola National WildlifeRefuge is located in La Paz County, Arizona, due westof Cibola, Arizona . The Cibola NWR is located ap-proximately at River Mile 99 on the lower ColoradoRiver in La Paz County, Arizona. The revegetation siteis located in the northeast corner of the refuge andconsists of eight 6 m X 9 m plots .

Objective. To better understand the feasibility of es-tablishing native plants in highly saline soils .

Completion Date . September 1986 .

Pre-Project Site Conditions. The site was essentiallydevoid of woody riparian plants at the time of planting .Soils were classified as dense clays, and depth to sat-urated soils was estimated at 1 .5 m. Soil salinityranged from 6,000 ppm to 60,000 ppm (Pinkney1992) .

Project Strategy. Prior to seeding, all plots wereflood irrigated to leach excess salts from the soil pro-file. The plots were seeded and raked in September1986. Four of the eight plots were seeded with quail-brush, while the remaining four plots were seeded withequal amounts of screwbean mesquite, honey mes-quite, and palo verde . Two of the quailbrush plots andtwo of the mesquite and palo verde plots were fertil-ized. The plots were flood irrigated three times in 1986with 2 to 5 cm of water and once in 1987 with 10 to15 cm of water (Pinkney 1992) .


Results . Germination rates for all plants was initiallyhigh . In 1988, quailbrush seedlings were well-estab-lished (more than 1,000 seedlings germinated), and ap-proximately 1,900 screwbean mesquite, 600 honeymesquite, and 25 palo verde seedlings had also estab-lished. However, mortality was high in the followingyears. All palo verde subsequently died as well asmany of the mesquite seedlings. Rabbit damage alsooccurred and seemed to affect the survival and growthof honey mesquite more than screwbean (Pinkney1992). Only three to ten mesquite seedlings werefound in 1997 in each of the four plots and all wereprotected with chicken wire baskets . Experiences heredemonstrated that although it is difficult, it is never-theless possible to revegetate in areas characterized byhigh soil salinity by using appropriate plant materialsand innovative irrigation and planting strategies (Swettpers. comm.) .

Site #4 Cibola National Wildlife Refuge DredgeSpoil Site

Location and Size. This site covers 28 ha and is lo-cated about 8 km east of Palo Verde, California . It isdivided by a levee that parallels the Colorado River .Revegetation work began on the East Dredge Spoilarea in 1977 and on the West Dredge Spoil area in1978 .

Objective. To better understand the feasibility of us-ing revegetation to improve the condition of ecologi-cally-damaged reaches of the Colorado River . This wasthe Bureau of Reclamation's first revegetation experi-ment .

Completion Date . 1978 .

Pre-Project Site Conditions . Russian thistle domi-nated the site prior to project initiation . Soil surveydata indicated that soils on the east side of the site areprimarily sands with a thin clay layer situated 1 m to1 .5 m below the surface. In, contrast, soils on the westside of the site are primarily loam but were coveredby 0.5 m to 3 m of dredge spoil material. Depth to thesaturated zone of the soil profile varied from 3 m to4.5 m (Anderson and Ohmart 1982) .

Project Strategy . In total, approximately 2,000 cot-tonwood, willow, honey mesquite, and blue palo verdetrees (Cercidium floridum Benth.) and shrubs wereplanted as .5 m tall rooted cuttings . A variety of plant-ing techniques were used . Approximately 125 trees ofall species were planted in 20-cm-diameter holes au-gered to a depth of 1 .5 m. Numerous trees were alsoplanted in holes 30 cm in diameter and 3 m deep, aswell as holes 5 cm in diameter and 3 m deep. A dripirrigation system was installed that delivered water at


519

a rate of 15 1 per hour to each planted tree or shrub .Each tree was watered with 121 1 per day for at least150 days .

Results. Dredge Spoil plantings were counted in June1990. At that time, an additional 150 cottonwoods hadestablished naturally . Natural recruitment of honeymesquite and palo verde was also occurring (Pinkney1992). Willow numbers continued to decrease sincethe first inventory was completed . In 1997, there ap-pears to be significant differences in survival rates andoverall plant health between plantings on the west sideof the levee and those on the east side, with establish-ment rates and plant vigor on the west side of the levee(adjacent to the agricultural fields) appearing muchgreater.

Site #5-Farmer's Bridge

Location and Size. This project consists of two siteslocated roughly 200 m from the Colorado River alonga levee that runs parallel to the river due west of Ci-bola, Arizona . The levee can be accessed via Farmer'sBridge road. The southern site is roughly 0.4 ha andthe northern site is roughly 0 .8 ha .

Objective. To re-establish native riparian trees in anarea that has experienced significant ecological de-cline .

Completion Date. 1986 .

Pre-Project Site Conditions. Both sites were void ofwoody plants prior to project initiation . Soils weregenerally sandy, and depth to saturated soils was es-timated to vary between 3 m to 6 m [beneath the sur-face of the soil] (Swett pers . comm .) .

Project Strategy . A total of 65 willow, cottonwood,and screwbean mesquite were planted at the two lo-cations. Poles and seedlings were used in the revege-tation effort . Plantings were irrigated with a drip sys-tem for the first two summers .

Results . Survival of planted species was greater inthe north site than in the south site, possibly due togreater water availability. Twenty trees survive on thesouthern site. Of these, only one willow and six cot-tonwoods were found and all show obvious signs ofwater stress (stunted growth, canopy die-back, thickand yellow leaves) . In comparison, mesquite appearedmuch healthier, averaging roughly 5 m in height withspreading, relatively full canopies . According to theproject manager, the use of plant materials that areadapted to current hydrologic conditions (specificallyconsidering depth to saturated soils) is key to success(Swett pers . comm.) .

Site #6-Three-Finger Lake

Location and Size . The 50-ha site is located in theCibola National Wildlife Refuge on the western sideof the river due north of Paymaster Landing .

Objective. To restore Three-Finger Lake to its pre-1970 condition .

Completion Date. To be completed during 1998 .

Pre-Project Site Conditions . Depending on ColoradoRiver flow, the size of Three-Finger Lake historicallyranged from 8 to 60 ha. In 1970, channelization andrealignment of the Cibola Division of the river wascompleted, diverting waters away from Three-FingerLake and the old river channel. Saltcedar invaded asthe lake dried and wildfires eliminated most of the na-tive vegetation . Today, the site is dominated by a densemonotypic stand of saltcedar.

Project Strategy . Beginning in 1994, approximately50 ha of the lake site was dredged to an elevation of65 m. In addition, one fish pond was created for rear-ing native fish. Native riparian vegetation will beplanted around the dredged areas .

Results. The project is scheduled to be completedduring 1998 .

Site #7-Imperial Refuge Cottonwood and WillowRevegetation Site

Location and Size. This 15-ha site lies within the Im-perial National Wildlife Refuge, which is located nearMartinez Lake, Arizona .

Objective. To re-establish Fremont cottonwood andGoodding willow in an area heavily modified by ag-riculture pressures and overrun by saitcedar.

Completion Date. 1995 (although other revegetationefforts are planned for the future) .

Pre-Project Site Conditions . Wheat and rye wereonce cultivated on this site . Since abandoned, the fieldswere overrun by saltcedar, arrowweed, and other un-desirable species .

Project Strategy . Over 600 cottonwood and willowtrees were planted in January of 1995 . Prior to plant-ing, exotic plants (mainly saltcedar) were cleared fromthe site with a bulldozer. Soil salinity investigationswere performed to guide the development of the plant-ing design and two piezometers were installed to mon-itor water-table fluctuations . Cleared areas were thendisked and leveled, and the site was flood-irrigated justbefore the onset of revegetation to leach excess saltsfrom the soil profile . An auger was used to drill tosaturated soils, which at the time of planting varied

520

between 1 to 2 .5 m below the soil surface . Holes werespaced 6 m apart . Poles collected from the ImperialWildlife Refuge's nursery were stripped of leaves andbranches and placed in the ground. The average lengthof the poles used in the revegetation effort was 4 m,and the average diameter at breast height (dbh) was3.0 cm. Following planting, the site was floodedmonthly, and plants received two liquid fertilizer ap-plications during the growing season . In 1996 and1997, the frequency with which the planted area wasflood irrigated was increased to bi-monthly .

Results. All of the cottonwoods that were plantedhave survived to date . However, by 1997, about 90%of the willows had died . The high mortality rate isbelieved to be due to localized soil salinity problems .As revegetating with native riparian plants often re-quires significant irrigation inputs (particularly duringthe first three summers), having control of the land andthe water rights that go with the land is critical tosuccess (Hill pers . comm .) .

Site #8-Mittry Lake

Location and Size . The 23-ha Mittry Lake revege-tation site is located on the Arizona side of the Colo-rado River roughly 24 km north of Yuma . The MittryLake Wildlife area borders the revegetation site to thewest and the Gila Gravity Main Canal is immediatelyto the east .

Objective. To enhance habitat for wildlife .

Completion Date . April 1986 .

Pre-Project Conditions. Vegetation found on-siteprior to project initiation consisted mainly of sparsedesertscrub on high elevated areas and mixed riparianvegetation along several of the washes on the site'ssouthern end. Soil types were found to be highly var-iable, ranging from large rocks to silts and clays . Wa-ter-table depths were estimated at 2 to 4 m beneathsoil surface (Pinkney 1992) .

Project Strategy. Six revegetation zones were estab-lished on the lake site between March and April 1986 .Each of the zones was fenced, and tree planting holeswere augered to saturated soil (2 to 4 m deep) usingan 38-cm-diameter auger. Holes for shrubs were au-gered 2 m deep using a 19-cm auger. Planting beganin March 1986 and was completed in April 1986 . Alltrees and wolfberry shrubs (Lvcium spp.) were grownin 4 1 cans ; quailbrush plants were grown in smallbiodegradable containers . All trees and shrubs wereirrigated using a drip system. and wire baskets wereused to protect plants from wildlife damage . Irrigationrates varied, but all plants received water each day for


the first 30 days, at a rate of 68 1 per day for trees and11 1 per day for shrubs (Pinkney 1992) .

Results. At the end of the 1988 growing season, onlya few cottonwoods were taller than 4 m . Of the veg-etation planted, willows appeared to have the slowestgrowth rates, possibly due to water stress and damagefrom deer browsing . Estimates of water-table depthsin 1988 indicated that depths were much greater thanthe 2 to 4 m initially estimated. The' largest of theplanted trees were those taking advantage of waterleaking from a canal just uphill from the revegetationsite. Even though wire baskets were used, rabbits anddeer still managed to damage significant numbers ofplanted vegetation. After the third growing season,honey mesquites not significantly damaged by rabbitswere roughly 3 m tall . By 1988, some honey mesquiteswere larger than the planted cottonwoods . Most screw-bean mesquites also grew well and were about 1 .5 mtall after two years of growth . Palo verde growth rateswere generally slower than those of mesquite . Quail-brush grew rapidly and, by the third growing season,some were over one meter tall and producing seed .However, rabbit damage to the lower branches outsideof the protective baskets was common. Wolfberry ap-peared to be stressed during the early portion of thefirst growing season but seemed to grow better duringthe fall months. In heavy textured saline soils, LyciumTrreyi Gray (L. Torreyi var. filiforme Jones) appearedto be more vigorous than Lycium andersonii Gray(Pinkney 1992) .

Site #9-Fortuna Fish Pond

Location and Size . This 3 .2-ha revegetation site sur-rounds the Fortuna Fish Pond and is located about 16km east of Yuma, Arizona at the confluence of the GilaRiver and Fortuna Wash.

Objective. To mitigate for impacts from the construc-tion of the Yuma desalinization plant .

Completion Date. Spring of 1985 .

Pre-Project Site Conditions . Saltcedar, arrowweed,and creosote bush (Larrea tridentata (Moc. & Ses.)Cav.) were the dominant woody species found on thesite prior to project initiation . Soil surveys indicatedgenerally sandy soils with a 15-cm-thick clay layerfound approximately 1 .2 m beneath the soil surface onthe western portion of the site . Depth to saturated soilswas estimated to be less than 3 m . However, soil mois-ture readings indicated that there was a significant dif-ference in water availability above and below the claylayer (25 percent and 3 percent, respectively) (Pinkney1992) .


521

Project Strategy . Approximately 300 cottonwood,willow, and mesquite seedlings (grown in 3 .8 liter con-tainers) were planted in pre-augered holes around theperiphery of Fortuna Pond during the spring of 1985 .In November 1986, approximately 100 cottonwoodand willow cuttings and poles were also planted . Thecuttings were planted in saturated soils around theedge of the pond; no holes were augered for the cut-tings. The poles were 2.5 to 3 m long and were plantedin 5-cm-diameter holes augered to about one meterdeep. Most of the plants were protected initially with1-m-tall chicken wire baskets . In 1986, 1 .5-m-tallwelded wire baskets were placed around some plantsto protect against beaver damage (Pinkney 1992) .Plants were irrigated each day in 1985 from plantingto September. Afterwards, irrigation was cut off untilMay 1986, when the trees began to appear stressed .Daily watering continued to September 1986 at rough-ly 12 1 per day (Swett, pers . comm .) .

Results. Two years following the completion of therevegetation work, approximately 30 percent of thetrees had died. In 1988, more than 70 percent of thecottonwoods and willows along the edge of the pondwere over 3 m tall ; some of the trees along the outletchannel and the edge of the pond were 6 m tall ormore. By 1997, many of the trees were over 10 m talland appeared to be growing vigorously .

Site #10 Colorado River-Gila River Confluence

Location and Size . This is a naturally vegetated siteat the confluence of the Colorado and Gila Rivers, justdownstream from Prison Hill in Yuma, Arizona. Thesite is situated along 4 km of the West Main ColoradoRiver canal and encompasses over 80 ha .

Ecological Characteristics. Most of the vegetationprobably appeared following the high magnitude flowsof 1993 and 1994, although large cottonwood treesalong the southern portion of the site probably cameup following the floods of 1983 . Further work is re-quired to develop a detailed description of this site'scurrent ecological characteristics and the degree that ishas changed since 1983 . Nevertheless, it is apparentthat significant natural regeneration took place, and thearea is now dominated by an extensive riparian forestthat consists of such native species as Fremont cotton-wood, Goodding Willow, mesquite, seep willow (Bac-charis glutinosa Pers.), and cattail .

Conservation Challenges . Taking advantage of suchdramatic natural regeneration by allocating waters tomaintain wetland plant communities that have devel-oped following large flow events could be an effectiveconservation approach. Additional areas that contain

significant desirable riparian habitat need to be iden-tified for possible protection . Even if such areas cannotbe protected, there may be some strategies (e.g ., re-ducing vehicular traffic, altering recreation use, chang-ing livestock management) that could go a long wayin increasing the value of habitat to wildlife .

Site #11-Rio Hardy Wetlands

Location and Size. The Rio Hardy wetlands are inthe Colorado River delta on one of the western mostbranches of the river (Figure 2) .

Ecological Characteristics . Flow into the Rio Hardycomes principally from agricultural runoff and geo-thermal wells discharged into the channel (Payne et al .1992). A third source is backflow from the mainstemof the Colorado River during times of high flow . TheRfo Hardy wetlands are dominated by halophyticplants such as iodine-bush and quailbrush and less salt-tolerant plants such as arrowweed and desert broom(Baccharis emoryi Gray) .

Conservation Challenges . The vegetation of this areahas changed significantly since the turn of the century .As described by McDougal (1904), this area was oncedominated to the north by a riparian forest of cotton-wood and willow and to the south by an extensivetidal-influenced plain of saltgrass . In between therewere scattered mesquite trees and saltbushes . Today,the riparian forest is no longer present, and the wetlandis dominated for the most part by halophytic plants(Glenn et al . 1996). From at least 1977 to 1983, theRio Hardy wetlands were maintained by ponded wa-ters behind a natural dam (Glenn pers. comm .) . Thedestruction of the dam during high flow events of 1983has caused the wetlands to drain and generally decline(Payne et al . 1992), although large flows during 1992and 1993 caused the wetland area to increase from alow of 1,175 ha in 1988 to 24,000 ha in 1993 (Glennet al. 1996) .

Site #12-Cienega de Santa Clara

Location and Size . The Cienega is located on theeastern side of the delta and covers roughly 20,000 ha,with 4,500 ha thickly vegetated (Figure 3) .

Ecological Characteristics . In total, there are 22 wet-land plants in and along the periphery of the Cienega .The main portion of the Cienega is dominated by T .domengensis Pers. and at least eight other hydrophytes(Glenn et al . 1996) . It is supported principally by ag-ricultural runoff from the Wellton-Mohawk canal, withlesser inputs from the Riito Drain .

522

Conservation Challenges. The principal conserva-tion challenge facing the cienega is the possibility thatthe Wellton-Mohawk canal will be turned off. TheWellton-Mohawk canal was planned as a temporaryanswer for agricultural wastewater deposition until thedesalinization plant went on line . Although the SantaClara system is also supported by waters emanatingfrom the Riito Drain, flow in the Wellton-Mohawk ca-nal is much greater and therefore more crucial . If thissource of water is indeed eliminated, a significant por-tion of the Cienega will be compromised. In addition,the Cienega faces challenges from the invasion of ex-otic plants and increasing salinity levels .

DISCUSSION

The ecological condition of the lower Colorado Riv-er has deteriorated significantly over the last 65 years .The cottonwood-willow forests that once dominatedthe mainstem and the delta have been reduced to iso-lated stands and individual trees. Only remnants cur-rently exist of the once immense delta marshlands,with significant portions being tenuously maintainedvia a system of agricultural runoff canals . From anecological standpoint, the loss of these wetland eco-systems is probably one of the major environmentalissues facing the U .S.-Mexico border region today.

River impoundment, the diversion of river watersfor farms and cities, and agricultural activities are theprincipal causes for the dramatic decline of the river'secological condition. The challenges to bringing the


Figure 3 . The Cienega de Santa Clara covers nearly 20,000 ha and provides critical habitat for a variety of wildlife species,including several that are federally endangered .

river back to an improved level of ecologic health aremonumental, requiring comprehensive solutions thataddress the underlying causes of ecologic decline . Pro-viding water specifically for ecological improvement,altering dam releases to promote natural regeneration,conserving water throughout the river's watershed, andaddressing water allocation issues are just some of theregional challenges that will need to be addressed be-fore significant progress can be made in improving theriver's overall ecological condition .

Since the lower Colorado River is rapidly approach-ing the point where every drop of water is specificallyallocated, it will be necessary to "find" the water tomeet the environmental needs of the basin . Given thecurrent socio-political and economic landscape of theColorado River, however, it is likely that years of ne-gotiation may be required before some of these long-term strategies can be implemented. In the meantime,it is reasonable to look for less politically chargedstrategies that can bring immediate, although more lo-cal, benefits to the river's ecological condition . Suchstrategies include improving the effectiveness of res-toration efforts, maintaining existing critical natural ar-eas, and purchasing marginal agricultural land to re-establish native wetland ecosystems .

The recommendations or strategies that are present-ed below run the gamut from policy recommendations(strategies that would require some kind of policychange before they could be implemented) to restora-tion/protection recommendations (strategies that focuson improving the effectiveness of wetland restoration


and protection efforts) . As some of restoration andprotection recommendations (e .g ., improving the ef-fectiveness of riparian revegetation efforts) do not re-quire major changes in the socio-political landscape,they could potentially be implemented on a muchshorter time frame than recommendations which arepurely policy orientated . However, other restorationand protection recommendations (e.g ., augmenting ag-ricultural return flows to the Rio Hardy wetlands)would require a change in policy before they could becarried out .

Improve the Effectiveness of Colorado RiverRestoration Efforts

Evaluating the results of past restoration effortsalong the Colorado River provides a wealth of infor-mation regarding how to improve the effectiveness ofsimilar efforts. Some of the principal lessons drawnfrom these past restoration experiences are describedbelow. They are general in nature and focus more onplanting strategies that can be applied on a regionalbasis than site specific planting techniques or issues .

Develop Project Objectives that are Clear and Specif-ic. As almost all components of a restoration projecthinge on project objective (the personnel that becomeinvolved in the restoration effort, the strategies that areultimately employed, the time frame for project com-pletion, the project budget), an ill-defined objectivecan significantly hinder the effectiveness of the resto-ration effort (Briggs 1996) . In addition to benefitingindividual projects, clear and concise project objec-tives would make evaluating results less problematic,which in turn could greatly benefit future restorationefforts .

The effectiveness of the restoration effort at the For-tuna Fish Pond (Site #9), for example, was almost im-possible to evaluate. The objective of this effort wasto mitigate disturbances caused by the construction ofthe Yuma desalinization plant . Without a detailed anal-ysis of the effects of the construction disturbances,however, there is no way of knowing when restorationefforts have reached the point of ecological compen-sation. A more acceptable approach would be to de-scribe the objective in terms of the habitat type thatneeds to be restored (e.g ., X hectares of cottonwood/willow habitat for southwestern willow flycatchers(Empidonax traillii extitnus AOU). Such a well-de-fined objective will allow restoration managers to de-scribe the restoration endpoint in sufficient detail andin a manner that can be evaluated . For example, theobjective of an effort to restore willow flycatcher hab-itat can be described by plant density, diversity, and

523

vertical complexity if the habitat requirements of thewillow flycatcher are well-enough known .

Consider Using Restoration Strategies Other than Re-vegetation. The Bureau of Reclamation, Bureau ofLand Management, Fish and Wildlife Service, and oth-ers have used revegetation strategies extensively alongthe lower Colorado River to re-establish native riparianplants. Despite some notable revegetation successes, itis obvious that the results of these artificial-plantingefforts pale in comparison to natural regenerative ca-pabilities. Comparing the natural regeneration that wasexperienced at the confluence of the Colorado and GilaRivers following the 1983 and 1993 floods to the re-sults of artificial revegetation efforts underscores theneed to create opportunities for more cost-effective ap-proaches. Such comparisons may be unfair, but theynevertheless highlight the need to evaluate the resto-ration efforts that are being implemented along thelower Colorado River to determine whether the moneybeing spent is having the desired effects .

Revegetation is often limited in its effectiveness be-cause it often does not address the causes of ecologicaldeterioration . Essentially, the goal of revegetation is toreplace lost plants . If the underlying causes for the lossare not adequately addressed or are not well under-stood, it is very likely that human-planted materialswill meet the same demise as those that they are tryingto replace. Most successful revegetation efforts are im-plemented in concert with strategies that address thereasons behind the ecological damage that has oc-curred (Briggs 1996) .

The question that needs to be answered is whetherthe money, time, and energy that is spent on revege-tation can be better directed on implementing othertypes of restoration or conservation efforts? To answerthis important question, it is imperative that river man-agers assess a range of restoration options and attemptto tie strategies to the unique characteristics of specificreaches of the river.

Base Restoration Strategies on a Thorough Evaluationof Site Ecological Conditions . Generally, successfulrestoration projects are based on a sound understand-ing of the site's current ecological condition . Depth tosaturated soils, soil salinity concentrations, presence ofexotic plants, and intensity of use by recreationists andlivestock are just some of the factors that need to beevaluated to develop a sound site restoration plan(Briggs 1996) . Such information is also critical for de-veloping realistic project objectives . Of these factors,evaluating soil chemistry prior to revegetation hasbeen particularly critical for revegetation efforts alongthe lower Colorado River (Anderson 1989). The ex-periences of some riparian revegetation efforts indicatethat flood-irrigating prior to planting may help to re-

524

Figure 4 . Eleven-year old-cottonwoods planted at theFarmer's Bridge revegetation site show obvious signs of wa-ter stress .

move salts from soils to concentrations acceptable toat least some native riparian plant species. However,the riparian revegetation experiences along the lowerColorado River also indicate the need for further re-search on salt and water tolerances of non-agriculturalplant species .

Use Plant Materials that are Adapted to Current SiteEcological Conditions . Revegetation planners oftenattempt to re-establish the exact plant community thatwas present prior to disturbance even though the com-munity may no longer be adapted to the site's currentecological conditions . At the Farmer's Bridge revege-tation effort (Site #5), for example, cottonwoods andwillows were planted in an attempt to re-establishthem in an area where they once dominated. As thedepth to saturated soils at this site is significant, all ofthe cottonwoods and willows that were planted eitherdied or suffer from water stress, while the majority ofthe mesquite that were planted have survived and ap-pear healthy (Figure 4) . In future similar situations, the


use of alternative species, such as mesquite, quailbush, or saltbush that may not have been members ofthe pre-disturbance plant community, yet are native tothe region and are better able to survive in the changedhydrologic environment (or more saline conditions),can significantly improve overall revegetation effec-tiveness .

Take Advantage of Agricultural Return Flows or Run-off. Results experienced at Dredge spoil Site (Site#4) and No Name Lake (Site #2) demonstrate the va-lidity of planting vegetation in areas that are likely toexperience significant agricultural runoff . Both sitesreceive runoff from adjacent agricultural fields duringthe hot, dry summer months. This seems to have hada remarkable effect on survival and growth rates ofplanted trees. At the Dredge Spoil site, for example,the riparian vegetation planted on the levee side im-mediately adjacent to the agricultural fields appearmuch healthier and seem to have experienced greatergrowth rates than those on the opposite side of thelevee .

Future revegetation efforts should take advantage ofsimilar opportunities . In addition, it has been notedthat establishing native, riparian plants immediatelyadjacent to farmland can provide useful habitat for avariety of avian species. Although more research isneeded, results of some riparian revegetation effortsindicate that establishing riparian habitat for avian in-sectivores adjacent to agricultural land can providebenefits for farmers as well (Anderson et al . 1984) .

Take Advantage of Sites that are Inherently More Me-sic . Old river meanders and other low elevation to-pographic features can be characterized by relativelyhigh water availability . Such areas can offer ideal con-ditions for the re-establishment of native, obligate ri-parian plants and should be considered for future re-vegetation efforts . At the No Name Lake site (site #1),the only obligate riparian trees to establish and growwere those planted in the relatively low elevation ofthe lake's depression .

Develop Restoration Projects with a Long-term Out-look. All too often, restoration efforts fail simply be-cause there was no technical expertise or fundingavailable following project completion . By its very na-ture, restoration has a long-term time frame . Successis therefore measured over a protracted time period,requiring an institutional capacity and a political willthat must persist well beyond the time the last plant isplaced in the ground. At the very least, the design ofthe restoration effort needs to include funding and per-sonnel for maintaining irrigation systems and fences,controlling exotic plants, replacing lost plant materials,and monitoring results for three years following pro-


ject completion (monitoring should continue over alonger time frame) .

Identify and Protect Existing Wetland Ecosystems

In both the delta and the lower reaches of the Col-orado River, there remain significant areas occupied bynative, wetland habitat. The cottonwood-willow ripar-ian forest at the confluence of the Gila and ColoradoRivers, the Cienega de Santa Clara, and the marsh-lands of the Rio Hardy are examples . Whether theseareas were created naturally or artificially, with intentor completely by accident, their protection should bea clear short-term priority for lower Colorado Riverconservation efforts. Additional existing natural areasshould be identified and plans to maintain or enhancetheir ecological condition should be developed .

The Cienega de Santa Clara . The Cienega de SantaClara exemplifies this priority . The Cienega is main-tained for the most part by drainage flow from Ari-zona's Wellton-Mohawk Canal. Since the canal's com-pletion in 1978, the Cienega has grown from 200 ha(estimated in 1973) to about 20,000 ha (Glenn et al .1996). If the Yuma desalinization plant begins opera-tion and the Wellton-Mohawk waters are diverted toirrigation districts in Mexacali, the largest remainingwetland in the region may be significantly compro-mised (Zengel et al . 1995, Glenn et al . 1996) . Con-servation efforts need to focus on ways to maintainflow into the Cienega . If the desalinization plant doesgo on line, other sources of water need to be found tomake up for water lost to plant operations . Such alter-native waters could be found by allocating ColoradoRiver water for ecosystem maintenance or divertingagricultural return flows from Mexico southward .

Rio Hardy Wetlands . The Rio Hardy wetlands haveexperienced growth and decline over the past 50 years .During the period of 1947 to 1983, the wetlands ofthe Rio Hardy were maintained by water impoundedbehind a natural bar that existed roughly 35 km fromthe mouth of the Colorado River (Payne et al . 1992) .However, the bar was destroyed by the high flows of1983, and the resulting drainage reduced the size ofthe wetlands from 63,000 ha in 1983 to 1,175 ha in1988. High flows during 1992, however, have in-creased the wetlands to 24,000 ha, but this increase islikely temporary if some type of impoundment struc-ture is not recreated (Glenn et al . 1996) .

Conservation/restoration efforts in the near futureshould focus on augmenting agricultural return flowsto the Rio Hardy wetlands and preventing waters fromescaping once they reach the wetlands . Ducks Unlim-ited engineers have recommended repairing the naturaldam that was destroyed by the flooding of 1983 (Payne

525

et al . 1992). This strategy would probably reducedrainage from the Rio Hardy wetlands and help to re-store at least a portion of the former wetland area .Conducting a feasibility study of this and similar strat-egies to maximize the environmental use of waters en-tering these wetland' areas should be a priority .

The Colorado-Gila River Confluence . The cotton-wood-willow riparian forest at the confluence of theColorado and Gila rivers is a product of natural ripar-ian regeneration following the high flows of 1983 and1993. The expanse of this area demonstrates the oftendramatic resiliency of riparian ecosystems . Such nat-ural regeneration also represents an important resto-ration/conservation opportunity . As in this case, naturehas done the majority of the ecological improvementwork. Flooding has re-worked alluvial sediments tocreate ideal seedbed conditions for riparian plants, andnearby riparian plants have disseminated seed .

Future priorities should include identifying similarnatural areas throughout the basin and developingplans for their protection . As is the issue throughoutthe lower Colorado River, "finding" the water re-quired to maintain these areas is the key challenge . Inaddition, other issues, such as vehicular traffic, devel-opment pressures, water pollution, etc ., may need tobe addressed if the Colorado-Gila River riparian eco-system and other similar areas are to remain viable inthe future .

Appropriate Water Resources for Restoration

Restoration cannot happen unless water is allocatedspecifically for that purpose . However, the 1922 Col-orado River Compact does not dedicate waters tomaintain healthy wetland/aquatic ecosystems . Thesame is true for international treaties between the U .S .and Mexico that contain no language on environmentalconsiderations . Yet, it is highly likely that all of thelegally allocated water will be for human consumptiveuses in the near future by the seven basin states andMexico. Simply put, water management and planningdoes not consider aquatic ecosystems . The fact thatColorado River water is almost completely exhaustedbefore it reaches the delta underscores this point . It istherefore critical to change traditional water policiesthat currently focus solely on human consumptiveneeds so that they reflect a more sustainable manage-ment of the river's waters .

Morrison et al . (1996) pointed out that one of themajor obstacles in the way of gaining water allocationfor restoration is the lack of scientific information thatquantifies the amount of water needed for some degreeof wetland restoration . Essentially, restoration alloca-tions will probably not be made until there is a better

526

understanding of the ecological benefits that the wateris likely to bring .

Expand Revegetation Efforts in the NationalWildlife Refuges

Some of the most extensive (at least in terms of thearea covered) riparian revegetation efforts along thelower Colorado River have been experienced in theImperial and Cibola National Wildlife Refuges . In gen-eral, such protected areas offer great potential for re-vegetation because the land and the water are underthe direct control of the managers, and they are betterable to manipulate restoration strategies to fit site-spe-cific conditions. At the Imperial Wildlife Reserve, forexample, revegetation efforts have established cotton-wood, mesquite, willow, and other species on over ahundred hectares . Moreover, the refuge has been ableto do this without infringing upon the water rights ofothers. In addition, there is room to accomplish more,as less than half of Imperial's 27 million cubic metersconsumptive use allocation is being used (Ellis pers .comm.). In a sense, riparian revegetation has becomea farming operation where cottonwoods and willowsare the crop of choice instead of alfalfa (Swett pers .comm.). Yet, increasing funds to support such effortscan create significant riparian habitat in areas wheresuch habitat has just about disappeared .

Develop a Concerted Binational Effort forRestoring the Delta

Since Mexico has legal entitlements to less than 10percent of the river's annual flow, it is unfair and un-realistic to assume that Mexico should take sole re-sponsibility for restoring the delta . A binational effortthat provides the framework for a variety of coopera-tive cross-border ventures is critical for the future ofthe delta (Morrison et al . 1996). Given the dispropor-tionate Colorado River allocations between the twocountries, water and restoration assistance should beprovided to Mexico from the U .S. In addition, itshould be noted that U .S . interests will probably ben-efit from the recreational qualities that a restored deltawill bring. Birding, camping, kayaking, and other non-consumptive uses could expand tremendously in a re-stored delta region, bringing significant economic ben-efits to local communities and to tourism interests inboth Mexico and she U .S .

Develop and Implement Community-BasedConservation Approaches

Over 20 million people have a direct stake in thecurrent and future use of lower Colorado River water


(Morrison et al. 1996). The great majority of theseusers live far from the river and are interested in itprincipally as a water source . Through state and fed-eral channels, their voices can be heard regarding avariety of river-related issues (e.g ., allocation, waterquality, etc .). Riverside communities have an evenlarger stake in the river's health by virtue of the realand potential economic, social, and environmental val-ues it provides. However, they may not have the ap-propriate opportunity for input into decisions that af-fect their lives even more directly than those livingafar. Providing a forum that will bring community cit-izens together to discuss common problems and iden-tify concrete ways they can engage each other and thewater and land managers would be mutually beneficialto both the river and the towns and communities . Gath-erings of local stakeholders from both sides of the bor-der to discuss the sustainable use of Colorado Riverwater has been initiated and will continue to expand(Nagel, pers. comm.) .

A regional economic study by the Sonoran Institutein 1996 identified nature-based tourism as a priorityoption for building a firmer sustainable business en-vironment and local economic opportunity in the Son-oran Desert borderlands, including the lower ColoradoRiver and delta region (Nimkin 1996) . Opportunitiesof this kind have been developed most extensively inthe Reserva de la Biosfera Alto Golfo y Delta del RioColorado where Conservation International-Mexicoand Secretaria Medio Ambiente Recursos Naturales yPesca (SEMARNAP) have collaborated in promotingthe Cienega Santa Clara as a destination and carryingout nature guide training for the neighboring EjidoLuis Encinas Johnson. Other opportunities may existin association with the Imperial, Cibola and Bill Wil-liams National Wildlife Refuges, and the Rio Hardy inBaja California .

Create Zones of Protection

Providing increased levels of protection can help tomaintain critical wetland areas in the long-term. TheCienega de Santa Clara is part of the much larger Re-serva de la Biosfera Alto Golfo de California y Deltadel Rio Colorado. This protected area was establishedby the Mexican Government in 1993 to conserve theecosystems of the Sonoran Desert, Upper Gulf and theColorado delta, and to provide and protect fishing andtourism activities . This international reserve is recog-nized by UNESCO and offers significant protection tocore ecosystems . Future Colorado River conservationefforts need to focus on protecting additional nativewetland and riparian areas .


527

Modify Reservoir Operations andWater Use Practices

Altering dam releases to enhance or maintain theColorado River's wetland habitat needs to be a futurepriority. The successful completion of the 1996 testflow or "flood" from Glen Canyon Dam to restoreriver beaches may set the stage for further manipula-tions of dam releases for environmental purposes . Ifsuch artificially-produced floods can remove sand andsilt out of reservoir storage to downstream sites, a va-riety of objectives could be accomplished . Deposition-al bars can be created to enhance natural propagationof native riparian plant species, and backwater breed-ing areas for many native fish can also be created im-mediately downstream of newly-formed depositionalbars .

Performing such releases for environmental purpos-es may be more likely than before, as flood-water man-agement along the lower Colorado River in the lastdecade is changing because Lake Powell behind GlenCanyon Dam has nearly filled to capacity . This meansthat flood flows will probably pass through Glen Can-yon Dam much more frequently, increasing theamount of water brought into the lower reaches of theriver and the delta (Payne et al . 1992, Glenn et al .1996). If this scenario is accurate, a variety of resto-ration opportunities that involve innovative manipu-lations of dam releases to benefit river ecology mayexist that were not possible before .

Resolve Other Water-Related Issues

Along the lower Colorado River and in the delta,there seem to be numerous areas that are ecologicallydamaged due, at least in part, to water-related issuesthat are not directly associated with reservoir manage-ment and water allocation. These "other" water-relat-ed issues include addressing ground-water over-pump-ing, water pollution, and improving water conservationmeasures (Morrison et al . 1996). In comparison to al-tering dam releases or allocating water for environ-mental purposes, tackling some of these issues may bean effective way of realizing short-term ecological im-provements. Implementing water conservation mea-sures to reduce ground-water pumping in areas suchas Mexicali, Blythe, and Yuma could reduce draw-down of nearby wetlands. In some areas, lining irri-gation canals with concrete could significantly reducewater loss, possibly freeing some water for environ-mental purposes .

Use Wastewater More Effectively

The river has approached the point of full utiliza-tion, where essentially every drop of water that passes

through is spoken for. Therefore, obtaining significantamounts of water specifically for wetland enhancementwill become increasingly difficult . On the other hand,the amount of wastewater (agricultural drainage andmunicipal effluent) is growing significantly. Althoughthis water has no human use, some is well-suited formaintaining wetland vegetation (Glenn et al . 1997). Asignificant amount of the brackish wetlands in the deltais almost completely sustained by "unusable" agri-cultural runoff. A priority of future wetland conser-vation efforts should be to develop strategies for usingwastewater more effectively . Indeed, at a meeting be-tween the heads of water-management agencies of theU.S. and Mexico, scientists, agricultural, municipal,and environmental representatives suggested that theimproved management of wastewater could reestablishup to 40,500 ha of wetland habitat in the delta (Glennet al. 1997) .

Although the volume of municipal sewage effluentproduced each year is much smaller than that producedfrom agriculture runoff, it is likely to increase withexpanding urbanization . Depending on the quality ofthe effluent, some may also be suitable for maintainingwetlands. Moreover, certain wetland areas could be de-veloped specifically to help clean municipal waste-water, thus providing the dual benefit of wetland hab-itat creation and contaminated water filtration . Ofcourse, safeguards would have to be developed so thatwetlands do not become inexpensive disposal zonesfor raw sewage .

CONCLUSION

Due to a variety of human-related impacts, much ofthe lower Colorado River native wetland ecosystem isdamaged or has disappeared completely. Only rem-nants remain of the magnificent riparian cottonwood-willow forests that once graced significant reaches ofthe lower Colorado River. In the delta, the majority ofthe remaining marshlands are maintained via tenuousassociations with agricultural runoff canals . As eco-logical conditions along the lower Colorado River de-teriorated, so too has the well-being of the people andwildlife that depend on the river for their survival .During the height of their dominion, the population ofthe Cocopah, the original river people, exceeded6,000. Today, scattered communities consisting of afew hundred individuals remain, located for the mostpart on the backwaters of the Rio Hardy and near SanLuis and Somerton, Arizona . Wildlife species such asthe totoaba (a fish species cherished by the Cocopa),the vaquita porpoise, desert pupfish, and the Yumaclapper rail, all of which once thrived in the delta'secosystem, are listed as endangered under the U .S. En-dangered Species Act .

528

From an ecological and conservation perspective,the situation along the lower Colorado River hasreached a critical juncture . To make significant pro-gress in repairing the ecological condition of the river,it is imperative to address difficult and politically-charged bi-national policy issues such as ground-waterdrawdown, water allocation, and dam releases . Someof these strategies may require years of negotiationsand planning, however, before they can be imple-mented. This underscores the need to look for creativesolutions that can be carried out more rapidly. Strate-gies such as using wastewater for environmental pur-poses, developing plans to protect areas characterizedby significant native plant communities (e.g., cotton-wood/willow forest at the confluence of the Coloradoand Gila Rivers), improving the effectiveness of wet-land revegetation techniques, increasing areas protect-ed by state and/or federal agreements, and using com-munity-based conservation approaches would havedramatic, albeit local, effects on enhancing or main-taining the ecological condition of the lower ColoradoRiver.

Of these approaches, immediate attention should begiven to using agricultural runoff and municipal efflu-ent for environmental purposes . Currently, much of thewastewater drains north and is squandered in the en-closed Salton Sea basin . Working with the communi-ties that line the Colorado River should be another toppriority . Such community-based conservation ap-proaches could have the dual benefit of bringing eco-logical relief to the river and economic and/or healthbenefits to riverside residents . As they directly involveriverside peoples, community-based approaches alsohave long-term staying power-a critical element inthe success of any conservation/restoration effort .

ACKNOWLEDGMENTS

This study was supported by grants from the De-fenders of Wildlife and the Mexico Affairs Office ofthe National Park Service. We thank John Fritschieand Craig Miller (Defenders of Wildlife) and HowardO. Ness (US/Mexico Affairs Office, National Park Ser-vice) for their assistance. We also express our sincereappreciation to Anita Alvarez de Williams, GinaWalther, and Claudia Schroeder (El Museo de BajaCalifornia) ; Oscar Ramirez and Alberto Tapia (LaUniversidad de Baja California) ; and John Swett (Bu-reau of Reclamation) for providing background infor-mation and giving us a river and delta tour that wewill not soon forget . In addition, we thank Maria ElenaBarajas and Enrique Flores (Departamento de Ecolo-gia, Estado de Sonora) ; Juan Carlos Barrera and JoseCampoy (Reserva de la Biosfera Alto Golfo de Cali-fornia y Delta del Rio Colorado) ; Mitch Ellis and Car-


men Kennedy (Imperial National Wildlife Refuge) ;Edward Glenn, Rene Tanner, and William Matter (Uni-versity of Arizona) ; Robert D. Ohmart and DuncanPatten (Arizona State University) ; David Reyna -Ortiz(Instituto Nacional de Ecologia) ; Fred Phillips and Ter-ry Shaffer (Colorado River Indian Tribes) ; Lisa Carder(Sonoran Institute) ; and Carlos Valdes Casillas (Insti-tuto Tecnologico y de Estudios Superiores de Monter-rey) for reviewing, editing, and supporting this effort .

LITERATURE CITED

Abarca, EJ ., M .E Ingraldi, and A . Varela-Romero . 1993 . Observa-tions on the desert pupfish (Cyprinodon ,n acularius), Yuma clap-per rail (Rallur longirestris yumanen.sis) and shorebird commu-nities in the Cienega de Santa Clara, Sonora, Mexico. Nongameand Endangered Wildlife Program Technical Report . ArizonaGame and Fish Department, Phoenix, AZ, USA .

Alvarez de Williams, A . 1978. Southwest Cocopah . p. 99-113 . InA. Ortiz (ed .) Handbook of North American Indians, vol . 10 .Smithsonian Institution, Washington, DC, USA .

Anderson, B .W. 1989 . Research as an integral part of revegetationprojects . p. 419-423 . In D.L. Abell (tech . coor .) Proceedings ofthe California Riparian Systems Conference : Protection . Manage-ment, and Restoration for the 1990s . U .S. Department of Agri-culture/Forest Service General Technical Report/PSW-110 .

Anderson, B.W., R.D. Ohmart, and H.A. Allen, Jr. 1984 . Riparianbirds in the riparian/agricultural interface . p. 190-195 . In R.E.Warner and K.M. Hendrix (eds.) California Riparian Systems.University of California Press, Berkeley, CA . USA .

Briggs, M .K. 1996. Riparian Ecosystem Recovery in Arid Lands .University of Arizona Press, Tucson, AZ, USA .

Cisneros-Mata, M ., G. Montemayor-Lopez, and M . Roman-Rodre-quez . 1995 . Life history and conservation of Totaba macdonaldi.Conservation Biology 9 :806-814 .

Colorado River Basin Salinity Control Forum . 1990 . Water qualitystandards for salinity Colorado River system . 106 West 5(H) South,Suite 101 . Bountiful, UT, USA 84010 .

Eddleman, W.R. 1989 . Biology of the Yuma Clapper Rail in thesouthwestern U .S. and northwestern Mexico . U .S . Bureau of Rec-lamation, Yuma Projects Office, Yuma, AZ, USA .

Glenn, E .P. C. Congdon, and J . Garcia . 1997 . New value for oldwater. World and I April 1997 :204-211 .

Glenn . E .P . C . Lee, R . Felger, and S . Zengel . 1996 . Effects of watermanagement on the wetlands of the Colorado River delta, Mexico .Conservation Biology 10 :1175-1186 .

Glenn, E .P . R .S . Felger, A . Burquez, and D .S . Turner. 1992 . Cienegade Santa Clara : endangered wetland in the Colorado River delta,Sonora . Mexico . Natural Resources Journal 32 :817-824 .

Hendrickson . D.A. and A. Varela-Romero . 1989 . Conservation sta-tus of the desert pupfish, Cyprinodon macularius, in Mexico andArizona. Copeia 1989:478-483 .

Lee . C.T 1994 . Mapping wetland dynamics in the lower ColoradoRiver delta using archival landsat map satellite imagery . ArizonaRemote Sensing Center, Office of Arid Lands Studies, Universityof Arizona . Tucson, AZ, USA .

McDougal. D .T. 1904 . Delta and delta vegetation . Botanical Gazette38 :44-63 .

Minckley . W .L. 1991 . Native fishes of the Grand Canyon Region :An Obituary? p. 124-177 . In Colorado River Ecology and DamManagement . National Academy Press, Washington, DC, USA .

Morales-Abril . G . 1994 . Reserva de la Biosfera Alto Golfo de Cal-ifornia y delta del Rio Colorado . Ecologica 3 :26-27 .

Morrison. J .L . S .L. Postel, and PH. Gleick . 1996 . The sustainableuse of water in the lower Colorado River basin . Pacific Instituteand Global Water Policy Project, 1204 Preservation Park Way,Oakland, CA. USA .

Nimkin, D . 1996 . Sustainable development opportunities in the Ar-


izona-Sonora borderlands . Unpublished report prepared by DavidNimkin for the Sonoran Institute, Tucson, Arizona .

Ohmart, D. R., W.O. Deason, and C . Burke. 1977 . A riparian casehistory : the Colorado River. p . 35-47 . In R.R. Johnson and D .A .Jones (tech . coor.) Preservation and Management of Riparian Hab-itat : A Symposium. U .S. Department of Agriculture/Forest Ser-vice General Technical Report/RM-43 .

Payne, J.M ., FA. Reid, and E .C. Gonzalez . 1992 . Feasibility studyfor the possible enhancement of the Colorado River delta wet-lands, Baja California Norte. Mexico . Report to Ducks Unlimited,Inc., and Ducks Unlimited of Mexico, Sacramento, CA, USA .

Pinkney, F.C. 1992 . Revegetation and enhancement of riparian com-munities along the lower Colorado River. U .S . Department of theInterior. Bureau of Reclamation . Ecological Resources Branch,Ecological Resources Division, Denver, CO, USA .

529

Pontius, D . Colorado River Basin study . 1997 . Report to the WesternWater Policy Review Advisory Commission . National TechnicalInformation Service, Springfield, VA, USA .

Sykes, G . 1937 . The Colorado River delta . Carnegie Institution ofWashington, DC, USA . Publication no . 460 .

Thompson, R .W. 1968 . Tidal flat sedimentation on the ColoradoRiver delta, northwestern Gulf of California . Geological Societyof America, Boulder, CO, USA . Memoir 107 .

Zengel, S.A ., V.J . Meretsky, E.P Glenn, R .S. Felger, and D . Ortiz .1995 . Cienega de Santa Clara, a remnant wetland in the Rio . Col-orado delta (Mexico): vegetation distribution and the effects ofwater flow reduction . Ecological Engineering 4 :19-36 .

Manuscript received 12 June 1997 ; revisiof received 29 January1998 ; accepted 13 April 1998 .

lu jR

+

V

.y +

r aa ; rna9

Documents

Opportunities for Ecological Improvement Along the Lower