Flood Pulsing in Wetlands Restoring the Natural Hydrological Balance

Flood Pulsing in Wetlands:Restoring the NaturalHydrological Balance

edited byBeth A. MiddletonNational Wetlands Research Center,USGS, Lafayette, Louisiana

John Wiley & Sons, Inc.

Flood Pulsing in Wetlands

Flood Pulsing in Wetlands:Restoring the NaturalHydrological Balance

edited byBeth A. MiddletonNational Wetlands Research Center,USGS, Lafayette, Louisiana

John Wiley & Sons, Inc.

This book is printed on acid-free paper.

Copyright 2002 by John Wiley and Sons, New York. All rights reserved.

Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise,except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, withouteither the prior written permission of the Publisher, or authorization through payment of theappropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA01923, (978) 750-8400, fax (978) 750-4744. Requests to the Publisher for permission should beaddressed to the Permissions Department, John Wiley & Sons, Inc., 605 Third Avenue, New York,NY 10158-0012, (212) 850-6011, fax (212) 850-6008, E-Mail: [email protected].

This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with the understanding that the publisher is not engaged inrendering professional services. If professional advice or other expert assistance is required, theservices of a competent professional person should be sought.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in printmay not be available in electronic books. For more information about Wiley products, visit our website at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:Middleton, Beth.

Flood pulsing in wetlands: restoring the natural hydrological balance / by Beth A. Middleton.p. cm.

Includes bibliographical references.ISBN 0-471-41807-2

1. Floodplain ecologyNorth America. 2. Wetland restorationNorth America.

QH541.5.V3 M54 2002333.91'8153'097dc21

2001045615

Printed in the United States of America.

10 9 8 7 6 5 4 3 2 1

Contents

v

Contributors ix

Preface xi

Chapter 1 The Flood Pulse Concept in Wetland Restoration 1

Beth A. Middleton

Chapter 2 Flood Pulses and Restoration of Riparian Vegetation in the American Southwest 11

Julie C. Stromberg and M. K. Chew

Flood Patterns and Riparian Vegetation in the Desert Southwest / 12Flood Pulses and Riparian Restoration / 20Conclusion / 41

Chapter 3 The Role of the Flood Pulse in Ecosystem-Level Processes in Southwestern Riparian Forests: A Case Study from the Middle Rio Grande 51

Lisa M. Ellis, Clifford S. Crawford, and Manuel C. Molles Jr.An Altered River: The Case of the Middle Rio Grande / 53Consequences of the Altered River: Some Obvious Problems / 57

vi CONTENTS

Research at Bosque del Apache National Wildlife Refuge: Floods, Fire,and the Litter Connection / 66Fire: Its Relationship to Flooding and Litter Buildup / 79The Future: Restoration of the Flood Pulse / 88

Chapter 4 The Role of the Flood Pulse in Maintaining Boltoniadecurrens, a Fugitive Plant Species of the Illinois River Floodplain: A Case History of a Threatened Species 109

M. Smith and P. Mettler

The Flood Pulse and Boltonia decurrens / 112Adaptations to Cyclical Flooding / 122Alteration of the Flood Pulse / 125Restoration of the Flood Pulse to the Illinois River Valley / 132Protection for B. decurrens Under the Endangered Species Act / 133Policies and Prospects for the Future / 136

Chapter 5 Conservation and Restoration of Semiarid Riparian Forests: A Case Study from the Upper Missouri River,Montana 145

Michael L. Scott and Gregor T. Auble

Introduction / 146Riparian Forests in Dry Regions / 148The Upper Missouri River, Montana: A Case Study / 151Conclusions / 181

Chapter 6 Implications of Reestablishing Prolonged FloodPulse Characteristics of the Kissimmee River andFloodplain Ecosystem 191

Louis A. Toth, Joseph W. Koebel Jr., Andrew G. Warne, and Joanne Chamberlain

Hydrogeomorphology of the Kissimmee River Basin / 193Flood Pulse Ecology / 203Restoration of the Flood Pulse / 205Restoration Expectations / 208Conclusions / 216

Chapter 7 Flood Pulsing in the Regeneration and Maintenanceof Species in Riverine Forested Wetlands of theSoutheastern United States 223

Beth A. Middleton

Hydrologic Reengineering of Forested Wetlands / 224Regeneration Problems for Plant Species on Floodplains

with Altered Hydrology / 229Restoration Approaches / 262

Index 295

CONTENTS vii

Contributors

ix

Gregor T. Auble, U.S. Geological Survey, Midcontinent Ecological ScienceCenter, Fort Collins, Colorado

Joanne Chamberlain, Kissimmee Division, Watershed Management Depart-ment, South Florida Water Management District, West Palm Beach, Florida

M. K. Chew, Arizona State University, Tempe, ArizonaLisa M. Ellis, University of New Mexico, Albuquerque, New MexicoClifford S. Crawford, University of New Mexico, Albuquerque, New Mex-

icoJoseph W. Koebel Jr., Kissimmee Division, Watershed Management Depart-

ment, South Florida Water Management District, West Palm Beach, FloridaP. Mettler, Southern Illinois University, Carbondale, IllinoisBeth A. Middleton, National Wetlands Research Center, USGS, Lafayette,

LouisianaManuel C. Molles Jr., University of New Mexico, Albuquerque, New

MexicoMichael L. Scott, U.S. Geological Survey, Midcontinent Ecological Science

Center, Fort Collins, ColoradoM. Smith, Southern Illinois University, Edwardsville, IllinoisJulie C. Stromberg, Arizona State University, Tempe, ArizonaLouis A. Toth, Kissimmee Division, Watershed Management Department,

South Florida Water Management District, West Palm Beach, FloridaAndrew G. Warne, U.S. Geological Survey, Water Resources Division,

Caribbean District, GSA Center, Guaynabo, Puerto Rico

PrefaceBeth A. Middleton, National Wetlands Research Center,

USGS, Lafayette, Louisiana

This book is a rst-of-its-kind compilation of the research of leaders in theeld of restoration ecology whose work involves the use of ood pulsingin the restoration of wetlands. The contributed chapters give regional ex-amples of wetland restoration projects in which ood pulsing was a criti-cal part of restoring the hydrodynamic setting for the plants and animals ofoodplains. They thus provide an argument for the widespread incorpora-tion of this approach in restoration projects. Restoration practitioners, aca-demics, and students will nd this book invaluable for the information itdraws together from cutting-edge ideas in the technology of restoration.

Each contributed chapter makes its own case for the importance of oodpulsing in restoration within its own regional setting, as based on researchand monitoring of the projects reported here.

Chapter 1, The Flood Pulse Concept in Wetland Restoration, pro-vides the basic argument for the use of ood pulsing in restorationprojects.

Chapter 2, Flood Pulses and Restoration of Riparian Vegetation inthe American Southwest, describes the impacts of projects that re-duce ood pulsing in the southwestern United States. It also describesresearch that demonstrates the importance of ood pulsing for theregeneration of trees and other vegetation in Sonoran Desert commu-nities.

Chapter 3, The Role of the Flood Pulse in Ecosystem-Level Pro-cesses in Southwestern Riparian Forests: A Case Study from theMiddle Rio Grande, outlines the decline of cottonwoods and shspecies along the Rio Grande. These changes are related to ecosystemprocesses that have been affected by the absence of the ood pulse

xi

xii PREFACE

and changes in insect communities in riparian settings that have led tothe buildup of organic debris.

Chapter 4, The Role of the Flood Pulse in Maintaining Boltonia de-currens, a Fugitive Plant Species of the Illinois River Floodplain: ACase History of a Threatened Species, documents the changes inhydrology of the Illinois River and their relationship to the decline ofBoltonia decurrens following the construction of navigation dams,and agricultural levees that have disrupted the annual ood pulse.

Chapter 5, Conservation and Restoration of Semiarid RiparianForests: A Case Study from the Upper Missouri River, Montana,documents regeneration events in Populus forests along the MissouriRiver as related to the timing of ood pulsing events.

Chapter 6, Implications of Reestablishing Prolonged Flood PulseCharacteristics of the Kissimmee River and Floodplain Ecosystem,describes the most famous case concerning the use of ood pulsing inthe restoration of an entire landscape. The chapter discusses the his-tory of its changes, attempts to restore the original ood pulse, andthe projected reestablishment of communities (sh, insects, birds, andvegetation) along the Kissimmee River.

Chapter 7, Flood Pulsing in the Regeneration and Maintenance ofSpecies in Riverine Forested Wetlands of the Southeastern UnitedStates, describes river regulation projects across the southeasternUnited States. Based on the early life history dynamics of plants, itmakes the case for the incorporation of ood pulsing in riverineforested wetlands and describes the Brushy Lake, Arkansas, project,where a levee was breached to reconnect the channel to the oodplainby ood pulsing.

An extensive reference section is included in each chapter as an aid towetland restorationists and researchers.

The authors extend their special thanks to the many librarians whohelped to locate the volumes of information that contribute to the successof such an endeavor.

1The Flood Pulse Concept

in Wetland RestorationBeth A. Middleton

National Wetlands Research Center, USGS, Lafayette, Louisiana

1

Flood Pulsing in Wetlands: Restoring the Natural Hydrological Balance, edited by Beth A. Middleton.ISBN 0 471-41807-2 2002 John Wiley & Sons, Inc.

The reestablishment of ood pulsing in riverine and tidal systems is be-coming recognized as an essential step in the restoration of wetlandsworldwide. Especially in North America, monitoring of projects that haveincorporated more natural water regimes is now under way. In most in-stances, researchers are still collecting the essential life history data thatwill aid in building a case for the need to recreate ood-pulsed hydrologyin wetland restoration projects. In this book, each chapter examines a casehistory of one these projects, written by a eld researcher close to the heartof this rapidly developing eld.

The ood pulse concept was rst developed to describe seasonalchanges in water levels on Amazonian oodplains and their relationshipsto functional dynamics and the maintenance of species diversity (Junk,1982, 1997; Junk and Howard-Williams, 1984; Junk et al., 1989; NationalResearch Council, 1992; Bayley, 1995) (Figure 1-1). The interconnectionof the river channel and oodplain is critical because functions such asproduction, decomposition, and consumption are driven by the ood pulse(Grubaugh and Anderson, 1988; Sparks et al., 1990) and water uctuationdrives succession (van der Valk, 1981; Finlayson et al., 1989; Niering,

2 THE FLOOD PULSE CONCEPT IN WETLAND RESTORATION

1994; Middleton, 1999a). Although this idea emerged from the study oflarge river ecosystems, there is growing recognition that tidal pulsing isalso important in salt marshes (Niering, 1994; Turner and Lewis, 1997;Zedler and Callaway, 1999) and mangrove swamps (McKee and Faulkner,1999). In addition, isolated restored sedge meadows in the Prairie PotholeRegion have lower species richness than natural wetlands that developedwhile large oods still occasionally interconnected them (van der Valk,1999).

Although the importance of the ood pulse is recognized for a varietyof wetland types worldwide, the idea that it is necessary to reestablish afunctional ood or tidal pulse in damaged systems has been adapted ratherslowly by wetland restorationists. In fact, it is not yet known whetherthe restoration of ood pulsing restores function (Brookes et al., 1996),and at least some evidence shows that in and of itself, ood pulsing isnot enough. For example, merely reopening a tidal channel may not re-store salt marsh function if the soil structure is altered and/or too saline(Haltiner et al., 1997). Regardless of what else may have to be adjusted,reestablishment of the original water dynamics (and sometimes soil condi-tions) is a critical aspect of wetland restoration, even more than reestab-lishing the vegetation.

To restore a wetland, most often what is required is a reversal of the en-gineering that dried the wetland in the rst placethat is, dam removal,dechannelization, remeandering, addition of debris, rediversion of water,cessation of water extraction, levee or polder removal. The reengineeringat a landscape level that is often required for such change is not easy, eitherphysically or politically. However, simpler and widely used approachessuch as damming create static water levels and so are not adequate restora-tion approaches (Middleton, 1999b, 2000).

The alteration of riverine and coastal ecosystems worldwide is so wide-spread as to leave us few examples of systems that still have a natural hy-drologic regime (Sparks et al., 1990; Petts et al., 1992; Junk, 1999). Thisis especially true in temperate areas of the world; in the 139 largest riversystems in Europe, the republics of the former Soviet Union, and regionsnorth of Mexico, 77 percent of their total discharge is affected by damand reservoir operation, interbasin diversion, and irrigation (Dynesius andNilsson, 1994). Water extraction along rivers is also causing salt water in-trusion in fresh and brackish water coastal systems (Muoz and Prat,1989; Prat and Ibaez, 1995). Along rivers in industrialized countries,

3Figure 1-1. Flood pulsing across a forested oodplain in various seasons in North America,related functional dynamics and biotic adaptations. (Adapted from Bayley, 1991, as derivedfrom Junk et al., 1989, in Middleton, 1999b.)


natural ood regimes are almost absent as a result of the reengineering ofwaterways (Bayley, 1995). Nevertheless, a few northern rivers that havebeen reengineered have portions that still ood pulsefor example, theIllinois (Sparks et al., 1998) and the Danube Rivers (Heiler et al., 1995).

After levees were constructed along major rivers such as the Missis-sippi, oodplains were converted to other uses, such as agriculture (Allen,1997). If people move onto a oodplain after the completion of a watercontrol project, it is often politically impossible to initiate the types ofreengineering measures necessary for ood pulsing on the ood or tidalplain. Yet sudden, destructive oods sometimes occur on reengineeredoodplains, so that a certain amount of rethinking is occurring recently. Isit really wise for us to restrict a river to its immediate channel and thus al-low the encroachment of the oodplain, which exposes people to the threatof dangerous oods (Interagency Floodplain Management Review Com-mittee, 1994; Junk, 1999)? In cases where the threat of future ooding islikely, portions of ood or tidal plains may be designated as nature areas toprovide for ood storage (Zinke and Gutzweiler, 1990; Lathbury, 1996).Chronically ooded sites present some opportunities for the use of oodpulsing in restoration, albeit on a small scale. Nevertheless, there are somerecent examples where ood-pulsed conditions have been (or are being)restored on a regional or landscape scale because of public demand, suchas on the Kissimmee River (see Chapter 6). Unfortunately, because of thedanger of ooding private property, restoration projects have usually beenlimited to ineffective measures, such as impounding waterways, that donot provide the biota with the pulsing environment to which they areadapted (Middleton 1999b).

The importance of reestablishing water regimes in sync with seasonalclimate uctuation and water ow in riverine and tidal systems has notbeen fully appreciated in wetland restoration. Organisms have specicadaptations that allow them to tolerate the wet/dry conditions that are apart of a ood-pulsed environment (Junk, 1997; Middleton, 1999a). Notonly does each species have different water requirements and tolerances,these differ for each life stageseed, seedling, and adult (see Chapters 2,4, 5 and 7).

Damming, one of the most common river regulation procedures, isillustrative of the problems created by altered environments for biota(Middleton 1999b). Upstream, the reservoir above the dam becomes per-manently impounded, resulting in a replacement of riparian vegetation

with algal or submerged communities. Downstream from the dam, owsin the stream channel are altered, which changes the nature of the pulsetransmitted to the oodplain (Middleton 1999b). Sediments becometrapped behind the dam, so downcutting and erosion occur in the down-stream channel, further cutting off the channel from the oodplain (Pettsand Lewin, 1979; Hickin, 1983; Petts, 1984).

Permanent ooding lowers the overall species richness along regulatedrivers because the sites never draw down (Nilsson et al., 1997). The dryphase of the ood pulse is critical, because even the most ood-tolerantspecies will eventually die in anaerobic conditions (Crawford, 1983; Arm-strong et al., 1994) even though such species possess many mechanisms tosurvive periods of inundation (Crawford and Braendle, 1996; McKevlin etal., 1998). The long-term effects of impoundment in reservoirs indicatethat when a river margin is permanently ooded, many species are lost, aswas demonstrated in a study of eight Swedish rivers (Jansson et al.,2000b).

Impoundment is often used in restoration as a means of increasing wa-ter levels in a dried wetland, but because of the lack of a ood pulse, re-generationfrom seed dispersal to the seedling recruitment stageisproblematic (Middleton, 1999b, 2000). Dams inhibit the movement of hy-drochorous seeds because of fragmentation and low current velocity, andthis affects seed availability along the corridor; as a result, each impound-ment develops a distinctive ora (Jansson et al., 2000a) (Figure 1-2). In ad-dition, because impoundment reduces dispersal distance, impoundmentsare likely inhabited by individuals that are more closely related to eachother (Jansson et al., 2000a). The impacts of dams on ora were still ap-parent 65 km downstream of dams along six rivers in Virginia (Schneideret al., 1989). River regulation also has severe impacts on fauna; it desyn-chronizes environmental cycles and thus disrupts reproductive cycles ofsh (Welcomme, 1989; Gehrke et al., 1995; Junk, 1997) and migrations ofinvertebrates (Adis et al., 1996).

Successful restoration depends on a better understanding of the life his-tory requirements of plants and animals (Chapters 2 to 6). Seed germina-tion can be critically dependent on ood pulsing, with the high phase ofthe pulse necessary for dispersal and drawdown necessary for germination(Junk and Piedade, 1997; Middleton, 1999b, 2000). Without a ood pulse,the dispersal of some species, such as Taxodium distichum and Populusspp., to suitable elevations for germination during the growing season is

THE FLOOD PULSE CONCEPT IN WETLAND RESTORATION 5


hampered (Chapters 7 and 5, respectively). Certain endangered speciessuch as Boltonia decurrens on the Illinois River cannot germinate and setseed without a ood pulse (Chapter 4). Seed germination can also be sen-sitive to other environmental factors, such as salinity (Galinato and van derValk, 1986; Baldwin et al., 1996), temperature, substrate, pH, and lightquality (Baskin and Baskin, 1998). At the same time, oods remove thedebris that sometimes decreases the germinability of seeds (Chapter 3).Species become increasingly tolerant of ooding as plants mature (Chap-ter 7). By the adult stage, water tolerance is widely variable betweenspecies and forms the basis for the compositional differences of wetlands(Harris et al., 1975; Whitlow and Harris, 1979; Hook, 1984; Theriot, 1993;Middleton, 1999b). Unfortunately, water for restoration purposes in thearid West may be so limited by competing demands by humans thatrestoration may be nearly impossible (Chapter 2).

River regulation impacts the ood pulsing environment experienced byora and fauna on ood and coastal plains, which is critical in the life his-tory dynamics of these species. Without proper attention to the hydrologicsetting created, attempts at wetland restoration will fail. This book reviewsthe case histories of restoration situations where either ood pulsing hasbeen reestablished as part of the project, or extensive studies of the life his-

Figure 1-2. Hypothesized relationships between plant-dispersal vectors, riverbed prole,and composition of the riparian ora in free-owing versus impounded rivers. (A) The ora inthe free-owing river is hypothesized to describe a gradual change downstream, whereas (B)in the regulated river, each impoundment is projected to develop an individual ora (denoted14). (From Jansson et al., 2000a, copyright Ecological Society of America; reprinted bypermission.)

tory requirements of species that are likely to need ood pulsing are beingconducted.

REFERENCES

Adis, J., S. I. Golovatch, and S. Hamann. 1996. Survival strategy of the terricolousmillipede Cutervodesmus adisi Golovatch (Fuhrmannodesmideae, Poly-desmida) in a blackwater inundation forest of central Amazonia (Brazil) in re-sponse to the ood pulse. Memoires du Museum National dHistoire Naturelle169: 523532.

Allen, J. A., and V. Burkett. 1997. Bottomland hardwood forest restoration:overview of techniques, successes and failures, In Wetlands and WatershedManagement: Science Applications and Public Policy, ed. by J. A. Kusler,D. E. Willard, and H. C. Hull, Jr., pp. 328332, April 2326, 1995. Tampa, FL.,Association of State Wetland Managers, Berne, NY.

Armstrong, W., R. Brndle, and M. B. Jackson. 1994. Mechanisms of ood toler-ance in plants. Acta Botanica Neerlandica 43: 307358.

Baldwin, A. H., K. L. McKee, and I. A. Mendelssohn. 1996. The inuence of veg-etation, salinity, and inundation on seed banks of oligohaline coastal marshes.American Journal of Botany 83: 470479.

Baskin, C. C., and J. M. Baskin. 1998. Seeds: Ecology, biogeography, and evolu-tion of dormancy and germination. Academic Press, San Diego, CA.

Bayley, P. B. 1991. The ood pulse advantage and the restoration of river-oodplain systems. Regulated Rivers: Research and Management 6: 7586.

Bayley, P. B. 1995. Understanding large river-oodplain ecosystems. BioScience45: 153158.

Brookes, A., J. Baker, and C. Redmond. 1996. Floodplain restoration and riparianzone management. In River Channel Restoration: Guiding Principles for Sus-tainable Projects, ed. by A. Brookes and F. D. Shields Jr., pp. 201229. JohnWiley & Sons, Chichester, U.K.

Crawford, R. M. M. 1983. Root survival in ooded soils. In Ecosystems of theWorld 4A; Mires: Swamp, Bog, Fen and Moor; General Studies, ed. by A. J. P.Gore, pp. 257283. Elsevier, Amsterdam, The Netherlands.

Crawford, R. M. M., and R. Braendle. 1996. Oxygen deprivation stress in achanging environment. Journal of Experimental Botany 47: 145159.

Dynesius, M., and C. Nilsson. 1994. Fragmentation and ow regulation of riversystems in the northern third of the world. Science 266: 753762.

Finlayson, C. M., B. J. Bailey, and I. D. Cowie. 1989. Macrophyte Vegetation ofthe Magela Creek Floodplain, Alligator Rivers Region, Northern Territory.Supervising Scientist for the Alligator Rivers Region, Australian GovernmentPublishing Service, Canberra.

REFERENCES 7


Galinato, M. I., and A. G. van der Valk. 1986. Seed germination traits of annualsand emergents recruited during drawdowns in the Delta Marsh, Manitoba,Canada. Aquatic Botany 26: 89102.

Gehrke, P. C., P. Brown, C. B. Schiller, D. B. Moffatt, and A. M. Bruce. 1995.River regulation and sh communities in the Murray-Darling River System,Australia. Regulated Rivers: Research and Management 11: 363375.

Grubaugh, J. W., and R. V. Anderson. 1988. Spatial and temporal availability ofoodplain habitat: Long-term changes at Pool 19, Mississippi River. AmericanMidland Naturalist 119: 402411.

Haltiner, J., J. B. Zedler, K. E. Boyer, G. D. Williams, and J. C. Callaway. 1997.Inuence of physical processes on the design, functioning and evolution of re-stored tidal wetlands in California (USA). Wetlands Ecology and Management4: 7391.

Harris, R. W., A. T. Leiser, and R. E. Fissell. 1975. Plant Tolerance to Flooding,RWH-200-7/1/75. Department of Environmental Horticulture, Davis, CA.

Heiler, G., T. Hein, and F. Schiemer. 1995. Hydrological connectivity and oodpulses as the central aspects for the integrity of a river-oodplain system. Reg-ulated Rivers: Research and Management 11: 351361.

Hickin, E. J. 1983. River channel changes: Retrospect and prospect. In Modernand Ancient Fluvial Systems, ed. by J. D. Collinson and J. Lewin, pp. 6183.Blackwell Scientic Publications, Oxford, U.K.

Hook, D. D. 1984. Waterlogging tolerance of lowland tree species of the South.Southern Journal of Applied Forestry 8: 136149.

Interagency Floodplain Management Review Committee. 1994. Sharing theChallenge: Floodplain Management into the 21st Century. AdministrationFloodplain Management Task Force, Washington, DC.

Jansson, R., C. Nilsson, and B. Renflt. 2000a, Fragmentation of riparian orasin rivers with multiple dams. Ecology 81: 899903.

Jansson, R., C. Nilsson, M. Dynesius, and E. Andersson. 2000b, Effects of riverregulation on river-margin vegetation: A comparison of eight boreal rivers.Ecological Applications 10: 203224.

Junk, W. J. 1982. Amazonian oodplains: Their ecology, present and potentialuse. In Proceedings of the International Scientic Workshop on Ecosystem Dy-namics in Freshwater Wetlands and Shallow Water Bodies, pp. 98126. Scien-tic Committee on Problems of the Environment (SCOPE), United NationsEnvironment Program (UNEP), New York.

Junk, W. J. 1997. Structure and function of the large central Amazonian Riveroodplains: Synthesis and discussion. In The Central Amazonian Floodplain,ed. by W. J. Junk, pp. 455472. Springer-Verlag, Berlin.

Junk, W. J. 1999. The ood pulse concept of large rivers: Learning from the trop-ics. Archiv fr Hydrobiologie 115: 261280.

REFERENCES 9

Junk, W. J., and C. Howard-Williams. 1984. Ecology of aquatic macrophytes inAmazonia. In The Amazon: Limnology and Landscape Ecology of a MightyTropical River and Its Basin, ed. by H. Sioli, pp. 269309. Dr. W. Junk Pub-lishers, Dordrecht, The Netherlands.

Junk, W. J., and M. T. F. Piedade. 1997. Plant life in the oodplain with specialreference to herbaceous plants. In The Central Amazon Floodplain: Ecology ofa Pulsing System, ed. by W. J. Junk, pp. 147185. Springer, Berlin.

Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. The ood pulse concept in river-oodplain systems. In Proceedings of the International Large River Sympo-sium (LARS), ed. by D. P. Dodge, pp. 110127. Canadian Special Publicationof Fisheries and Aquatic Sciences, Ottawa, Canada.

Lathbury, M. E. 1996. Toward natural ood control: Floodplain wetlands. WetlandJournal 8: 1013.

McKee, K. L., and P. L. Faulkner. 1999. Biogeochemical Functioning of Restoredand Natural Mangrove Forests in Southwest Florida. Final Report to NOAA/NERRS/SRD. National Wetlands Research Center, Lafayette, LA.

McKevlin, M. R., D. D. Hook, and A. A. Rozelle. 1998. Adaptations of plants toooding and soil waterlogging. In Southern Forested Wetlands Ecology andManagement, ed. by M. G. Messina and W. H. Conner, pp. 173203. LewisPublishers, Boca Raton, FL.

Middleton, B. A. 1999a, Succession and herbivory in monsoonal wetlands. Wet-lands Ecology and Management 6: 189202.

Middleton, B. A. 1999b. Wetland Restoration, Flood Pulsing and DisturbanceDynamics. John Wiley & Sons, New York.

Middleton, B. A. 2000. Hydrochory, seed banks, and regeneration dynamics alongthe landscape boundaries of a forested wetland. Plant Ecology 146: 169184.

Muoz, I., and N. Prat. 1989. Effects of river regulation on the lower Ebro River(NE Spain). Regulated Rivers: Research and Management 3: 345354.

National Research Council. 1992. Restoration of Aquatic Ecosystems. NationalAcademy Press, Washington, DC.

Niering, W. 1994. Wetland vegetation change: A dynamic process, Wetland Jour-nal 6: 615.

Nilsson, C., A. Ekblad, M. Gardfjell, and B. Carlberg. 1997. Long-term effects ofriver regulation on river margin vegetation. Journal of Applied Ecology 28:963987.

Petts, G. E. 1984. Impounded Rivers: Perspectives for Ecological Management.John Wiley & Sons, Chichester, U. K.

Petts, G. E., and J. Lewin. 1979. Physical effects of reservoirs on river systems. InMans Impact on the Hydrological Cycle in the United Kingdom, ed. by G. E.Hollis, pp. 7991. Geo Abstracts Ltd., Norwich, U. K.

Petts, G. E., A. R. G. Large, M. T. Greenwood, and M. A. Bickerton. 1992. Flood-plain assessment for restoration and conservation: Linking hydrogeomorphol-


ogy and ecology. In Lowland Floodplain Rivers: Geomorphological Perspec-tives, ed. by P. A. Carling, and G. E. Petts, pp. 217234. John Wiley & Sons,Chichester, U. K.

Prat, N., and C. Ibaez. 1995. Effects of water transfers projected in the SpanishNational Hydrological Plan on the ecology of the Lower River Ebro (NESpain) and its delta. Water Science and Technology 31: 7986.

Schneider, R. L., N. E. Martin, and R. R. Sharitz. 1989. Impact of dam operationson hydrology and associated oodplain forests of southeastern rivers. In Fresh-water Wetlands and Wildlife, ed. by R. R. Sharitz and J. W. Gibbons, pp.11131122. Ofce of Scientic and Technical Information, Oak Ridge, TN.

Sparks, R. E., J. C. Nelson, and Y. Yin. 1998. Naturalization of the ood regime inregulated rivers. BioScience 48: 706720.

Sparks, R. E., P. B. Bayley, S. L. Kohler, and L. L. Osborne. 1990. Disturbanceand recovery of large oodplain rivers. Environmental Management 14:699709.

Theriot, R. F. 1993. Flood Tolerance of Plant Species in Bottomland Forests of theSoutheastern United States. WRP-DE-6. U.S. Army Corps of Engineers,Vicksburg, MS.

Turner, R. E., and R. R. Lewis III. 1997. Hydrologic restoration of coastal wet-lands. Wetlands Ecology and Management 4: 6572.

van der Valk. A. G. 1981. Succession in wetlands: A Gleasonian approach. Ecol-ogy 62: 688696.

van der Valk, A. G. 1999. Succession theory and wetland restoration. In Proceed-ings of INTECOLs V International Wetlands Conference, ed. by A. J. McComband J. A. Davis, pp. 657667. Gleneagles Press, Adelaide, Australia.

Welcomme, R. L. 1989. Floodplain sheries management. In Alternatives in Reg-ulated River Management, ed. by J. A. Gore and G. E. Petts, pp. 209233.CRC, Boca Raton, FL.

Whitlow, T. H., and R. W. Harris. 1979. Flood Tolerance in Plants: A State-of-the-Art Review. Technical Report E-79-2. U.S. Army Corps of Engineers, Vicks-burg, MS.

Zedler, J. B., and J. C. Callaway. 1999. Tracking wetland restoration: Do mitiga-tion sites follow desired trajectories? Restoration Ecology 7: 6973.

Zinke, A., and K.-A. Gutzweiler. 1990. Possibilities for regeneration of oodplainforests within the framework on the ood-protection measures on the UpperRhine, West Germany. Forest Ecology and Management 33/34: 1320.

11

Flood Pulsing in Wetlands: Restoring the Natural Hydrological Balance, edited by Beth A. MiddletonISBN 0 471-41807-2 John Wiley & Sons, Inc.

2Flood Pulses and

Restoration of RiparianVegetation in the

American SouthwestJulie C. Stromberg and M. K. Chew

Arizona State University, Tempe, Arizona

In mesic regions of North America, to the casual observer, riparian treesare relatively indistinguishable from their nearby upland counterparts. Inarid parts of the southwestern United States, the transition from riparianto upland zones is striking and often abrupt. Forests with multistory treeand shrub canopies along water courses give way to sparse grassland,subshrub, and succulent communities, often within a few meters of theoodplain. Riparian forests along desert rivers provide the water, shelter,and food that sustain obligate riparian animal species and provide a crit-ical buffer for upland species during dry seasons and droughts. Beginningin the late 1800s, however, regional water and land management practicessubstantially reduced the acreage suitable or available for riverine marsh-

12 FLOOD PULSES AND RESTORATION IN THE AMERICAN SOUTHWEST

lands, Populus-Salix forests, and other Sonoran riparian plant communi-ties (Hendrickson and Minckley, 1984; Brown, 1994; Patten, 1998). Inaddition to their direct replacement by irrigated elds and urban areas, ri-parian ecosystems have been degraded by dams that disrupt water andsediment ow, by diversion structures and groundwater wells that dewaterstreams and aquifers, and by the multiple impacts of livestock grazing.

Many riparian restorations have been undertaken to overcome suchdegradations. Flood pulsing has expedited, or has been the major featureof, many ecologically successful projects. Some ood pulsing events fallwithin the category of managed or designer ood releases for the expresspurpose of ecosystem restoration. Others are accidents, whereby oodsreleased from a reservoir for one purpose have collaterally promotedriparian restoration. Natural oods have facilitated the restoration of ri-parian ecosystems along free-owing rivers following the removal ofstressors such as groundwater pumping and livestock grazing.

This chapter discusses examples of ood-assisted restoration efforts,organized by specic restoration goals. It addresses the restoration of re-generative processes for dominant trees (Populus, Salix), the managementcomplications arising from the presence of exotic woody species (Tam-arix), and the broader goal of restoring a large complement of riparianplant species. The role of ood pulses in restoring plant productivity, min-imizing re disturbance, and restoring resilience is discussed. The chapterfocuses on rivers of the Sonoran and Mojave Deserts, occasionally rang-ing beyond these geographic limits. It begins by summarizing regional cli-matic ood patterns, managerial changes to ood regimes, and associatedecological responses of riparian vegetation.

FLOOD PATTERNS AND RIPARIAN VEGETATIONIN THE DESERT SOUTHWEST

Natural Flood Regimes

The rivers of the southwestern United States are prone to rapid but sea-sonally consistent changes in ow. Most oods in the rivers of the Sono-ran and Mojave Deserts result from rainfall, with snowmelt contributing toa lesser degree. Localized but intense convective rainstorms usually de-velop during the July-August monsoon season. Regionwide precipitationand runoff are again high in December and January, following Pacic

FLOOD PATTERNS AND RIPARIAN VEGETATION IN THE DESERT SOUTHWEST 13

0

1

2

3

Mea

n flo

w ra

te (m

3 s-1 )

O N D J F M A M J J A S

San Pedro River- Redington

Figure 2-1. Mean monthly ow rates (m3s-1) for streams in the Sonoran-Mojave Desert tran-sition zone (Santa Maria River-Bagstad) and the Sonoran-Chihuahuan Desert transitionzone (San Pedro River-Redington). Values shown are based on U. S. Geological Survey(USGS) historical daily streams ow data and are averages for a 25-year time period.

0

2

4

6

Mea

n flo

w ra

te (m

3 s-1 )


Santa Maria River- Bagstad

frontal storms, declining through early spring. If tropical storms or hurri-canes penetrate inland, oods occur in late fall, but less frequently. Fol-lowing a southeastward gradient from the Mojave Desert, across theSonoran Desert, toward the Chihuahuan Desert, summer rains becomemore frequent and summer oods more intense (Figure 2-1).

Floods in southwestern desert rivers are extreme in comparison withtheir low ows (Graf, 1988). Channels may carry no surface ows forshort periods in May or June, yet experience instantaneous peak ows ofseveral thousand cubic feet per second during the subsequent monsoonseason. Flood patterns also vary over longer temporal scales (Molles et al.,1992). Both mean and peak annual river ows documented in the regionover the past century have varied by several orders of magnitude between


0

500

1000

1500

2000

2500

3000Pe

ak a

nnua

l flo

w (m

3 s-1 )

19301935

19401945

19501955

19601965

19701975

19801985

1990

Pre-dam Post-dam

Bill Williams River

Figure 2-2. Annual ood magnitude size, pre-dam and post-dam, for the Bill Williams River,in western Arizona. Values shown are based on USGS peak ow data (instantaneous dis-charge, station number 09426000).

dry years and wet years (Figure 2-2). The patterns are nonstationaryand reect drought cycles and the El Nio-Southern Oscillation. In somerivers of the Southwest, large winter oods were relatively more frequentnear the end of the nineteenth century and again toward the end of thetwentieth century. During the 1980s and 1990s, some rivers experiencedback-to-back 100-year return oods in winter. Such oods were less com-mon in the interim. The Santa Cruz River in southern Arizona, for exam-ple, had no large winter oods from the 1930s through the 1960s (Webband Betancourt, 1992). This type of ood pattern results in boom andbust cycles, with episodes of intensive scour and channel widening fol-lowed by rapid and dense regrowth of the vegetation.

The high-energy oods inuence patch dynamics and reinitiate succes-sion. Along Arizonas free-owing (undammed) Hassayampa River, forexample, riparian vegetation changed dramatically after large El Niowinter oods in 1993, which widened the channel from 5 to 50 m anderoded oodplains to an elevation very near the water table (Stromberg etal., 1997) (Figure 2-3). Emergent marshlands characterized by Juncus ar-ticulatus, Typha domingensis, and Scirpus americanus increased vefoldafter the ood. Many mature Populus fremontiiSalix gooddingii forestsand Prosopis velutina woodlands were replaced by young stands of Popu-lus-Salix (Figure 2-4). The long period of ood recession created estab-lishment opportunities for spring-germinating Populus and Salix as well assummer-germinating Baccharis salicifolia, Tessaria sericea, and exoticTamarix ramosissima (Stromberg, 1997) (Figure 2-5). At the same time,


590

591

592

593

594

595

596

Elev

atio

n (m

)

0 25 50 75 100 125 150 Distance (m)

1989 1991 1993

Figure 2-3. Changes in topography of the Hassayampa River oodplain after two majorood events: a ood in 1991 with instantaneous discharge of 368 m3s-1 and a ood in 1993with instantaneous discharge of 745 m3s-1. (Figure is modied from Stromberg et al., 1997.)

parts of the pre-1993 channel were buried by more than 1 m of sediment.Riparian shrublands dominated by the stress-tolerant Hymenoclea mono-gyra established on the coarse ood deposits, too high and dry for marshplants or Populus-Salix seedlings.

Low-energy oods, in contrast, inuence successional processes by ef-fecting other changes, such as depositing layers of ne sediments aroundthe boles of trees and redistributing leaf litter. Rivers in the SonoranDesert, as in other arid regions, carry high sediment loads. In the Has-sayampa River oodplain, annual rates of sedimentation have ranged fromless than 1 to 10 cm, with the sedimentation rate increasing with oodmagnitude (Stromberg et al., 1993b). Deposition declines with distancefrom the active channel for smaller oods (Figure 2-6). Root crowns ofpioneer trees can eventually be buried under several meters of ne sedi-ment. Such sediments may maintain nutrient availability near the soilsurface. Large-seeded, shade-tolerant, competitive trees (Grime, 1979),including Prosopis velutina, P. glandulosa, P. pubsecens, and Juglans ma-jor, establish in the aggraded oodplain soils, aided by summer monsoonrains and the resultant ood spikes that disperse seeds and moisten safe-sites (Stromberg and Patten, 1990; Stromberg et al., 1991).

As the predominant natural disturbances, oods are clearly a majorforce structuring southwestern riparian vegetation (Junk et al., 1989; Poffet al., 1997). Floods create a riparian ecosystem that is dynamic in spaceand time by mobilizing, eroding, and depositing sediment, causing riverchannels to relocate and meander, creating backwater depressions, scour-

16

Figure 2-4. Riparian vegetation along the Hassayampa River during a nonood year (1988,top photo) and one year after the large El Nio oods (1993, bottom photo). (Photographs byJ. Stromberg.)


0

2

4

68

10

12

14

16

Sedi

men

t dep

osite

d (cm

)


aquifers, moistening surface soils, and depositing limiting nutrients on theoodplain.

Human-Altered Flood Patterns

A system of dams and high-capacity reservoirs has been developed on therivers of the southwestern United States. The amount of water capturedand stored per capita (within the service area) is among the highest in thenation (Graf, 1999). The Colorado River, largest in the region, has beentransformed from a powerfully dynamic waterway to a string of storagepools. The completion of Hoover Dam in 1935 effectively disconnectedthe upper and lower Colorado watersheds. Multipurpose dams, such asHoover, store water, control oods, and generate hydropower. Major andminor tributaries of the lower Colorado, including the Gila, Salt, Verde,and Agua Fria, are similarly impounded. Reservoirs such as Lake Pleasant,formed by New Waddell Dam on the Agua Fria, as part of the Central Ari-zona Project, receive and store water that is pumped upgrade and cross-country from the Colorado (Springer et al., 1999). Diversion structures,including New Waddell Dam and Granite Reef Dam on the Salt River,completely reroute ows into canals, leaving a dry riverbed below (Figure2-7). Small rivers, such as the Sonoita and Aravaipa Creeks, have beendammed solely for recreational purposes. Only a few small perennialSonoran Desert rivers, including the Hassayampa, Santa Cruz, and San Pe-dro, remain undammed.

Flood and low-ow patterns are now fundamentally different as a resultof reservoir development and water diversion for off-channel uses. Inmany cases, water that would have arrived in rivers as spring runoff nowappears (if at all) as the sustained high ow of summer water delivery todownstream diversion structures, tailing off and exposing near-channelsediments only as the irrigation season wanes in the autumn. Dammedrivers usually ood less frequently, and the rare oods that do occur tendto be drastic ones that overwhelm water storage or ood control capabili-ties. Dynamic uvial processes, such as channel migration, erosion, andsedimentation, can become static (Shields et al., 2000). Salt levels instream waters and oodplain soils increase as water evaporates from thelarge reservoirs and ushing oods are eliminated.

Nonimpoundment ood control measures, like channelization, expeditewater movement through improved parts of river systems, minimizing


Figure 2-7. A dewatered reach of the Agua Fria River below New Waddell Dam, in centralArizona. (Photograph by J. Stromberg.)

overbank ows, riparian sediment deposition, and bank storage recharge.Regionwide watershed degradation has also modied ood patterns.Widespread soil compaction, reduced plant cover, and loss of A horizonand organic matter in soils have followed overgrazing, poor timber man-agement, and rapid urbanization. As a secondary effect, rapid runoff hasincreased ashiness in the remaining undammed fragments of these al-ready ashy systems, resulting in less inltration and more runoff. The neteffect is larger ood peaks, higher sedimentation rates, and reduced baseows (Trimble and Mendel, 1995).

The range of ecological responses to ood ow alterations has variedwith the nature and extent of hydrologic alteration, the geomorphic set-ting, and the interactive effects of other stressors, including livestock graz-ing (Scott et al., 1996). In the narrow Glen Canyon reach of the ColoradoRiver, ood suppression has allowed narrow strips of marshland and ri-parian forest to persist along the channel, which during pre-dam timeswas scoured by annual oods (Stevens et al., 1995). Many ecological re-sponses have been negative, however. Farther downstream along the Col-orado River, for example, the cumulative effects of river damming and


diversion have nearly obliterated the most extensive riparian wetlands inthe desert Southwest, leaving only a few scattered ecological remnantsamid a vast expanse of unvegetated and degraded land (Glenn et al., 1992;Briggs and Cornelius, 1998). The Lower Colorado River riparian vegeta-tion zone has been narrowed, and the vegetation mosaic has been simpli-ed (Ohmart et al., 1988). As a consequence of the reduced extent andquality of riparian habitat in the Southwest, many animal species that areriparian obligates have undergone population declines. The southwesternwillow ycatcher (Empidonax traillii extimus), a small neotropical mi-grant songbird that nests in dense thickets of Salix spp. and other structuralanalogs along perennial streams, is now on the federal endangered specieslist (Unitt, 1987). The western yellow-billed cuckoo (Coccyzus ameri-canus occidentalis), another neotropical migrant that nests in dense ri-parian vegetation, including Populus-Salix forests, also has declined innumbers. The Yuma clapper rail (Rallus longirostris yumanensis) and theleast Bells vireo (Vireo bellii pusillus) are also on the federal endangeredspecies list (Kus, 1998). Without widespread changes in river manage-ment, the prognosis for these and many other species is grim (Graf et al.,in press).

FLOOD PULSES AND RIPARIAN RESTORATION

Restoration of riparian ecosystems has become a major enterprise in thedesert Southwest. Federal and state agencies, Indian tribes, cities, and pri-vate landowners are spending money and time on efforts to reverse pastdegradations. Goals and approaches vary widely. A decade ago in theSouthwest, goals were somewhat limited, and riparian restoration wassynonymous with cottonwood pole planting (Briggs, 1996). Althoughplantings are still widely used today (Alpert et al., 1999), the practice of ri-parian restoration is maturing beyond single-species plantings to encom-pass an ecosystem approach and a goal of self-sustainability (Goodwin etal., 1997). There is increasing recognition by restoration managers thathydrologic regimes and uvial processes are prime determinants of ripar-ian community structure and that restoration of native biodiversity andecosystem complexity depends on the restoration of uvial dynamism.There is increasing acceptance of the need to restore ows of water, sedi-

FLOOD PULSES AND RIPARIAN RESTORATION 21

ment, and nutrients in sufcient quantities and with appropriate temporalpatterns (Poff et al., 1997; Hill and Platts, 1998; Taylor et al., 1999).

Full restoration of riparian ecosystems requires removing all impedi-ments to natural ow regimes (Schmidt et al., 1998). However, there arefew cases in the desert Southwest in which dams have been removed forthe purpose of habitat restoration. An exception involves Fossil Creekin central Arizona; the impending decommissioning of this small hy-dropower dam is a signicant undertaking. Where dams remain in place,creative ways are being found to rehabilitate, if not fully restore, below-dam ecosystems while still allowing for some degree of municipal or agri-cultural water supply, hydropower production, or ood control (Whittakerand Shelby, 2000). Along several free-owing rivers, ecological stressorshave been eliminated or reduced. At these sites, natural ood pulses are fa-cilitating passive recovery of the riparian vegetation following removal ofthe stressors.

Restoring Dominant Species

Populus fremontii and Salix gooddingii historically were the dominant pi-oneer forest species along Sonoran and Mojave Desert rivers. Precise esti-mates are impossible because rangewide baseline data do not exist, butlosses as high as 90 percent have been postulated for Populus-Salix forestson various rivers such as the lower Colorado River (e.g., Ohmart et al.,1997). These forests are considered a globally imperiled ecosystem typeby The Nature Conservancy (Anderson et al., 1998; Stein et al., 2000).

The population dynamics of Populus fremontii and Salix gooddingii areinuenced by many aspects of a ood regime, including the timing, mag-nitude, duration, and rate of change of any given ood. Both are short-lived pioneer trees that exploit and depend on the ooding cycle throughtemporally specialized reproduction strategies. Plants establish from seedduring occasional ideal years when appropriate ood conditions are pre-sent (Turner, 1974; Everitt, 1995; Scott et al., 1997; Mahoney and Rood,1998; Stromberg, 1998a). The timing of ood ows is critical, as thespecies have evolved to produce seeds that are viable during the brief pe-riod when high spring ows are declining and exposing bare, damp sedi-ments (Horton et al., 1960; Fenner et al., 1984). Large winter oods scourand redeposit oodplain sediments, creating the patchwork of bare min-


eral soils on which plants can establish without competition from over-story trees. S. gooddingii disperse seeds somewhat later in the season thanP. fremontii (although the dispersal periods overlap) and, as the ood wa-ters recede, establish on sites that are lower and closer to the stream. Smalldifferences in the timing of spring ood pulses can inuence the relativesuccess of recruitment of these and other pioneer trees and shrubs (e.g.,Salix exigua) because of phenological differences (Stromberg et al., 1991).Fenner et al. (1985) were among the rst to call attention to the fact thatPopulus fremontii were not regenerating below dams where ood patternshad been altered. They studied the Salt River upstream of Phoenix, Ari-zona, and attributed the lack of Populus regeneration to the lack of springoods, as well to water depletion caused by pumping groundwater. The1980s and 1990s saw a spate of research on the patterns and causes ofPopulus declines below dams throughout the western United States andCanada (Rood and Mahoney, 1990). Mahoney and Rood (1998) observedthat one cause of Populus seedling mortality was rapid recession of ood-waters. Friedman et al. (1998) and Johnson (1998) noted that in somecases, forests of Populus and other woody pioneers expanded immediatelyfollowing dam closure, only to be replaced by later successional specieswithout the return of appropriate regeneration oods.

Research on the autecology of Populus led to the development of re-cruitment models, sometimes called recruitment boxes, that indicatewhen waters should be released from dams and at what drawdown rate, toallow for seedling establishment (Mahoney and Rood, 1998). Several pro-jects have been implemented that have used these models as a basis forrestoring ows to regenerate Populus. However, suitable ows of waterand sediment have yet to be restored in the dammed river reaches that con-stituted the study area of Fenner et al. (1985). Landowners and river advo-cates along the Salt River continue to be concerned about the lack ofadequate Populus regeneration, some because of concern for loss of nest-ing habitat for bald eagles.

The Truckee River in Nevada provides an example of river rehabilita-tion via restoration of a more natural stream ow pattern. Dams, channel-ization, and water diversions had contributed to a loss of age class andstructural diversity within the Populus fremontii forests and a collapse ofnative sh populations, including the endangered cui-ui sh (Chasmistescujus). To stimulate spawning of the sh populations, the U.S. Fish andWildlife Service began managing the Stampede Reservoir in the late


1980s for spring ood release. An ancillary benet was the establishmentof Populus seedlings. However, many seedlings were at possible risk ofmortality, having established on sites that could be ooded in the future.Subsequently, in 1995 a collaboration between The Nature Conservancy,federal agencies, and university researchers resulted in the release of aood tailored from recruitment models, expressly for the purpose of Pop-ulus regeneration (Mahoney and Rood, 1998). Although the prescribedood resulted in added costs to the Bureau of Reclamation, the NevadaNature Conservancy speculates that there will be long-term economic sav-ings due to improved ecosystem functioning: With greater shading of theriver, water temperature will be lower, and less water will be needed tomanage the endangered cui-ui sh, reducing the long-term cost of thespecies recovery program (Chisholm, 1996). Another take-home mes-sage here is that a whole array of ecosystem components may begin to re-cover when restoring a basic ecosystem process, such as the natural owregime of a river (Gourley, 1998). Although it is impossible to manage di-rectly for every species, we increase our odds of providing sustainableecosystem improvement if we take an approach that allows for natural cy-cles of ood disturbance (Bayley, 1991; Stanford et al., 1996).

Flood pulses have accidentally helped to restore small patches ofPopulus-Salix forests to some Southwestern rivers. The Gila River and itsmajor tributaries, the Salt and Verde Rivers, are impounded by seven largewater supply dams as they traverse Arizona before meeting with the Col-orado River near Yuma. The Colorado River itself is one of the most ex-tensively dammed and diverted rivers in the United States, with ows atthe Mexican border now only a small fraction of what they were a centuryago. Most of the water in the Gila and Colorado reservoirs is stored foragricultural use; vast portions of the lower oodplains of these two riversare farmed for lettuce, cotton, alfalfa, and other irrigated crops. During ElNio years in the late 1970s, 1980s, and 1990s, heavy upland precipita-tion, particularly rain on deep snowpack, forced large ows to be releasedthrough some of the dams during winter and spring (Figure 2-8). Releasesmay have become more frequent in recent decades, as the dams have agedand lled with water and sediment (Zamora-Arroyo et al., 2001). Such re-leases were not conceived to provide downstream ecological benets.However, some of the water releases fortuitously corresponded to the re-generation needs of the trees while (presumably) ushing accumulatedagricultural tailwater salts and recharging aquifers. Several young popula-


0

1000

2000

3000

4000

5000

6000Pe

ak a

nnua

l flo

w (m

3s-

1 )

19101920

19301940

19501960

19701980

1990

Gila River

Figure 2-8. Annual ood magnitude in the lower Gila River near Yuma, Arizona. CoolidgeDam, the main water storage dam on the Gila, was completed in 1928. Completion dates fordams on major tributaries range from 1911 (Roosevelt Dam) to 1946 (Horseshoe Dam). Val-ues shown are instantaneous discharges (m3s-1) based on USGS peak ow data (stationnumber 09520500).

tions of Populus and Salix now occur along portions of the lower Gila andColorado Rivers, as well as the lower Salt and Verde Rivers (Rea, 1983;Briggs and Cornelius, 1998). Passive management of oods has also al-lowed for regeneration of Populus fremontii along portions of Utahs Fre-mont River, that are utilized for irrigation storage (Everitt, 1995). Yet treeson some of these rivers may die unless water sources needed for their sur-vival are forthcoming. In addition to working toward securing minimummaintenance ows, there is a need to negotiate dam operating agreementsthat provide for the intentional release of ows during periodic wet yearsin order to mimic the natural ow regime.

Floods have also served as restorative agents for Populus-Salix forestson degraded undammed rivers. The Santa Cruz River, which originates inMexico and ows northward through southern Arizona, has remained freeof dams. Historically, the Santa Cruz was described as a small but lushriver that owed perennially near the U.S./Mexican border (Tellman et al.,1997). During the 1950s and 1960s, however, extensive portions of theSanta Cruz River in the vicinity of Nogales, Arizona, went completely dry.The river lost perennial ow, groundwater tables declined, and most of thePopulus-Salix forests and marsh vegetation died. The 1950s were a periodof extreme drought in the Southwest (Swetnam and Betancourt, 1998).Compounding the drought was extensive pumping of groundwater from


the small alluvial aquifer lining the Santa Cruz River (Stromberg et al.,1993c). During recent decades, however, the Santa Cruz riparian ecosys-tem was revitalized. Abundant rains and runoff, together with the releaseof large amounts of treated municipal wastewater into the channel, con-tributed to aquifer recharge and restoration of the perennial stream ow.Some of this water originated in another watershed, resulting in a net in-crease to the Santa Cruz. Although the restoration of a high water table setthe stage for a high rate of survivorship, it was the occurrence of winterand spring ood ows in the 1970s, 1980s and 1990s that triggered the es-tablishment of the extensive young groves of Populus and Salix that linethe riverbed. The sources of seeds for the natural recovery presumablywere the trees that survived in upstream reaches not subjected to extremegroundwater pumping, or the trees along the edges of irrigated farm elds.

Exotic Species Management

Tamarix ramosissima and related species, large shrubs or trees that are na-tive to Eurasia, have become dominant woody species along rivers of thedesert Southwest (Harris, 1966). Vast portions of the Colorado River, GilaRiver, and Salt River ood plains, for example, are dominated by Tamarixspecies as well as by the native shrub Pluchea sericea. The replacement ofspecies-rich communities by homogenous thickets of single species, bethey native or exotic, can be symptomatic of dam-related reductions in u-vial disturbances and/or imposition of stressors that select for a smallnumber of stress-tolerant species. Many management practices in ripariancorridors, including reduced ood pulsing, have caused oodplain soils tobecome saltier, drier, and nutrient-poor (Stevens, 1989; Busch and Smith,1995). These factors have favored stress-tolerant species such as Tamarix(Glenn et al., 1998; Busch, 1995). Tamarix are capable of great osmotic ad-justment in saline soils (Decker, 1961; Shafroth et al., 1995; Glenn et al.,1998). They are deeper rooted and more able than native Populus and Salixtrees to persist on the available soil moisture at sites where groundwater isbelow rooting depths (Horton, 1972; Busch and Smith, 1995; Stromberg,1998b).

When rivers ood less frequently and at different seasons than their cli-matic legacies dictate, exotic species may gain a recruitment and survivaladvantage. Like the Populus and Salix they have widely replaced, Tamarixannually produce large crops of tiny, wind-dispersed seeds that germinate


on bare, moist soil. Temporally, however, Tamarix are reproductive gener-alists and disperse seed over a much longer time period during the grow-ing season (Shafroth et al., 1998; Roelle and Gladwin, 1999). Theirreproductive strategy allows them to thrive on dammed rivers where highwater ow is delayed by the timing of irrigation water storage and releaseschedules, as well as in the techno-littoral zone of reservoir edges, whereseedbeds are exposed in midsummer during irrigation-driven drawdowns.There are some native riparian pioneer shrubs (e.g., Baccharis salicifolia,Salix exigua) that have a lengthy seed dispersal period. Unlike Tamarix,these seem not to have beneted by such alterations, perhaps partly be-cause of lower fecundity.

Restoration of exotic-dominated sites can be approached by managingfor ecological processes that favor native species and, if necessary, man-aging against the exotics. Restoration efforts at two National WildlifeRefuges provide examples of both approaches. The rst involves the BillWilliams Wildlife Refuge, located on an alluvial tributary of the lowerColorado in west-central Arizona. Alamo Dam was constructed on the BillWilliams River in 1968 to minimize ood pulses into the Colorado River.Over the past 25 years, the size and frequency of winter oods in the BillWilliams River have decreased and summer oods have all but disap-peared (Figure 2-2). Pre-dam, the ten-year recurrence interval ood was1400 m3s-1. Plants would rapidly recolonize after the oods, but periodicscour maintained a more open system than occurs today. Today, the maxi-mum possible controlled releases are 198 m3s-1, the ten-year ood is 148m3s-1, and the mean ow rate is 4 m3s-1. The river is not diverted, and baseows have increased somewhat as the temporal pattern of ow release intothe below-dam system has become less ashy. With fewer scouring owsand less water stress, vegetation has become more abundant than in theabove-dam free-owing Santa Maria River and covers more area thanwould be the case were the dam not present (Shafroth et al., in press) (Fig-ure 2-9). However, most of the inll is T. ramosissima.

One goal of the Bill Williams Refuge managers is to restore Populusand Salix trees to dominance and thereby improve the quality of the ripar-ian habitat. Other goals include maintaining ood control in the ColoradoRiver and the recreational and wildlife benets of Alamo Lake. To accom-plish the rst goal, the U.S. Fish and Wildlife Service, U.S. Army Corps ofEngineers, and university scientists worked together to develop a ow-release plan that calls for high base ows and periodic ood ows. During

27

Figure 2-9. Riparian vegetation along the free-owing Santa Maria River (above) and ow-regulated Bill Williams (below). (Photographs by J. Stromberg.)


1993 and 1995, high spring ows were released into the Bill WilliamsRiver. Recruitment models were retroactively tested, conrming that therecruitment box concepts can be used to enable seedling establishment ofthese target species (Shafroth et al., 1998). However, without the ability torelease large scouring oods from the dam, extensive seedbeds for newgenerations of riparian trees will not be created. Small oods predictablyproduce smaller recruitment bands (Stromberg et al., 1993b). Rates of es-tablishment of Populus and Salix are predicted to decline on the BillWilliams in the future, despite the release of spring ows (Shafroth et al.,in press). Thus, Tamarix will remain a dominant species. Intervention inthe form of mechanical clearing of seedbeds in Tamarix-dominated habi-tat, followed by removal of aggraded sediments, may be necessary to re-store Populus and Salix to dominance. Active restoration measures thatmimic the effects of large oods are needed if natural processes cannot befully restored (Friedman et al., 1995).

Tamarix has also become a dominant species in the Bosque del ApacheWildlife Refuge, as on much of New Mexicos highly regulated RioGrande (Everitt, 1998). Refuge managers here have mimicked the scour-ing effects of large oods by using bulldozers, herbicides, and re to clearextensive stands of Tamarix, which has cost $750 to $1300 per hectare(Taylor and McDaniel, 1998). They released river water onto the bareoodplains in spring, with a seasonal timing and quantity that mimic theood hydrograph of the Rio Grande and thereby favor a diverse assem-blage of pioneer plant species. Long-term monitoring will determinewhether the multilevel canopy, diversity of vegetation structure, and diver-sity of insect life that develop provide superior wildlife habitat to theTamarix thickets that existed previously. This type of wet soil manage-ment is increasingly being used on other regulated rivers, notably inwildlife refuges along the lower Colorado River, where ood pulses can bereleased through water control structures to small, cleared areas of theoodplain or abandoned farm elds. These efforts do not completely elim-inate Tamarix. However, there is growing consensus that total eradicationis neither realistic nor necessary. For example, removal of Tamarix fromPopulus and Salix stands does not cause detectable change in bird abun-dance and diversity, suggesting that it is the presence of the structurally di-verse and insect-rich natives, rather than the complete absence of theexotic, that is of key importance (Weintraub, 1993).


In addition to managing river ows for native species, there are ways tomanage ows so as to decrease the birth rates or increase the death rates ofTamarix. Few such projects have been implemented, and more experimen-tation is warranted. Prolonged late-season inundation has been used as atechnique to increase the mortality of Tamarix seedlings at a Populus del-toides ssp. monolifera restoration site (Roelle and Gladwin, 1999); the in-undation time and duration were selected on the basis of experimentalstudies that compared mortality rates of native and exotic tree seedlings(Gladwin and Roelle, 1998). Restoration projects on dammed rivers couldexperiment with the release of ashy summer ood pulses to kill summergerminants of Tamarix. Although certainly able to survive ood scour andsedimentation as adults, there is evidence that Tamarix are less tolerant ofooding as juveniles than are Populus species. As compared with P. fre-montii, Tamarix ramosissima juveniles exhibited higher mortality result-ing from experimental sedimentation (Levine, 2001) and natural riverooding (Stromberg et al., 1993b). Less ability to tolerate ood scour mayexplain why Tamarix population levels are low relative to those of nativespecies on some free-owing, frequently ooded rivers and may con-tribute to their tendency to proliferate on ood-regulated rivers (DAnto-nio et al., 1999; Shafroth et al., 2002). However, there are conictingopinions on this matter, and certainly there is room for additional study.

A study on the free-owing San Pedro River provides an example of anatural shift from exotic to native dominance over time and evidences a ca-pacity for self-repair in degraded Sonoran riparian ecosystems, once stres-sors are eliminated or adequately reduced. The San Pedro ows northwardfrom Sonora, Mexico, to the Gila River in southern Arizona. Stream owsvary from perennial to ephemeral, depending on local geology and tribu-tary inputs and the extent of groundwater pumping. Historically, ood-plain agriculture and cattle grazing have been the main land uses, but somereaches have recently been set aside by The Nature Conservancy and theBureau of Land Management as conservation areas. Livestock grazing,gravel mining, crop irrigation, and pumping groundwater from the alluvialaquifer have been eliminated from the conservation areas. The compo-sition of the San Pedro riparian plant community has also changed, co-incident with changes in management and weather patterns. Tamarixramosissima became established on the river in the 1950s (a drought pe-riod) and by the 1960s was the dominant woody pioneer in the central


0

5

10

15

20

25

30

35

Mea

n da

ily fl

ow (m

3 s-1 )


'87 '88 '89 '90 '92 '94

Nonestablishment years

Figure 2-10. Stream ow hydrographs during years with and without seedling establishmentof Populus fremontii along the San Pedro River, over a nine-year period. Stream ow valuesare based on USGS historical mean daily ows (station number 09472000).

0

5

10

15

20

25

30

35

Mea

n flo

w ra

te (m

3 s-1 )


1991 1993 1995

P. FREMONTII establishment years

reaches of the river (Stromberg, 1998a). Since then, Populus fremontii andSalix gooddingii have been increasing in relative abundance. In the 1990s,young Populus and Salix outnumbered young Tamarix for the rst time inmore than 50 years in the central reaches.

Flood events were critical to the recovery of Populus and Salix on theSan Pedro River. El Nio winter oods in the 1980s and 1990s were suf-ciently large to scour vegetation and create opportunities for species re-placement and were appropriately timed to favor the native pioneers(Figure 2-10). Under the conditions of livestock exclusion, reduced rates ofupstream groundwater pumping, and frequent winter/spring oods, the na-tive species were able to outcompete the exotic one. Studies have shownPopulus spp. to be more competitive than Tamarix under conditions of ad-


equate water (Stevens, 1989; Sher et al., 2000) and livestock exclusion(with livestock grazing giving a competitive edge to the less palatableTamarix) (Stromberg, 1997). Notably, these patterns of recovery were ap-parent only in the wetter, ungrazed reaches of the San Pedro River;Tamarix continued to dominate in reaches with ephemeral ow, deep wa-ter tables, and livestock grazing (Figure 2-11). There is a need for addi-tional studies that assess the potential for natural recovery of native specieson exotic-dominated sites upon removal of stressors and/or removal of theexotic species. Theoretically, by restoring natural ow regimes and her-bivory patterns, we can tip the ecological balance in favor of native species(Poff et al., 1997).

Restoring Plant Productivity

Riparian forests, supplied with ample water and nutrients, are among themost productive vegetation types in arid regions. Shallow groundwater isthe primary water source for many oodplain trees, including Populusspp., Salix spp., and Platanus wrightii (Busch et al., 1992; Smith et al.,1998; Scott et al., 1999; Stromberg, 2001a). Flood ows are an importantsource of recharge of the alluvial aquifer in many rivers (Workman andSerrano, 1999) and thus can contribute much of the water that sustainsphreatophytic vegetation. The annual radial growth rate of Platanuswrightii along Sycamore Creek in Arizona increased as the number ofwinter oods increased, ranging from 0.1 cm yr-1 in dry years to 0.9 cmyr-1 in wet, ood years (Stromberg, 2001a). Winter ood ows in thissmall river are critical for elevating groundwater to within tree rootingzones (Marti et al., 1999; Stromberg, 2001b).

Flood pulses that occur during the growing season are used as a watersource by some riparian trees in desert regions (Akeroyd et al., 1998).Summer ood pulses are critical in some areas to provide the shallow soilmoisture that sustains seedlings or saplings of Populus species until theirroots grow deep enough to extract water from the aquifer (Cooper et al.,1999). Summer oods provide an important water source for the faculta-tive phreatophytic tree Prosopis velutina by wetting surface soils andrecharging groundwater. This deep-rooted species (with one report of a50 m taproot) attains its largest size when able to access shallow ground-water. Where groundwater is inaccessible, Prosopis trees undergo strong

32

Figure 2-11. Flood pulses have contributed to increased abundance of Populus fremontiiand Salix gooddingii in some parts of Arizonas San Pedro River (above). Tamarix spp. con-tinue to dominate in drier reaches ( below). (Photographs by J. Stromberg.)


0.00

0.20

0.40

0.60

0.80

1.00

Annu

al ra

dial

gro

wth

(cm yr

-1)

-4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0Annual stream flow (ln, m3 s-1)

Figure 2-12. Mean annual radial growth rate of young Platanus wrightii trees in relation tonatural log-transformed annual stream ow rate in an intermittently owing reach ofSycamore Creek, Arizona. Rectangular symbols indicate years with at least one summerood. (Figure is modied from Stromberg, 2001b, with permission of Blackwell Science Ltd.,Osney Mead, Oxford, U.K.)

seasonal increases in productivity in response to the summer monsoonood pulse (Stromberg et al., 1992; Stromberg et al., 1993a).

Flood pulses are often rich in organic matter and nutrients, (Grimmet al., 1981). Storm runoff mobilizes nutrients that have accumulated inand on upland desert soils, and ood waters then deposit them on ood-plain soils (Grimm and Fisher, 1986). Floods also stimulate microbialactivity and increase the rates of decomposition of organic matter, whichcan be low on the dry surface soils of oodplains in semiarid regions(Molles et al., 1998). A high frequency of summer oods has been linkedto increased growth of the tree Platanus wrightii, presumably becausesummer oods replenish the limited supply of nutrients (Stromberg,2001b) (Figure 2-12).

Stevens (1989) noted that waters in the Colorado River have becomedepleted in phosphorous because of sediment trapping in reservoirs. Thisnutrient depletion caused a reduced growth rate of woody riparian pioneerspecies in experimental studies. High salinity levels, such as can developin infrequently ushed soils of regulated rivers, reduce the growth of someriparian tree species (Glenn et al., 1998). However, there have been no sys-tematic studies of the effects of altered or suppressed ood pulses on


woody plant growth and productivity in below-dam reaches of SonoranDesert rivers. Such studies would be useful as a basis for determiningwhether there is a need to restore productivity pulses, in addition to regen-eration pulses.

Minimizing Fire Disturbance

The dynamic uvial processes of ow-regulated rivers are altered in waysthat can variously increase or decrease standing live and dead biomass. Onthe ood-suppressed Bill Williams River and portions of the lower Colo-rado River, post-dam increases in dense stands of re-prone T. ramosis-sima have set the stage for frequent, intense, and large res (Busch andSmith, 1995; Shafroth et al., in press). Along these rivers, more than a thirdof the studied riparian forests burned over a recent 12-year period (Busch,1995). Tamarix ramosissima and the clonal shrub Tesseria sericea re-sprout prolically after res, but Populus fremontii does not, resulting inpostre shifts in vegetation structure.

Fires historically occurred only rarely in regularly ooded Sonoran ri-parian corridors, but they have replaced oods as a primary disturbancefactor on many regulated rivers (Figure 2-13). When oods are sup-pressed, woody debris can accumulate; and vegetation can senescence andincrease in extent, density, and homogeneity. Floods, in contrast, reducere potential by scouring vegetation and creating natural rebreaks, re-moving dead branches and other debris, increasing litter decompositionrates, and increasing the moisture content of the vegetation (Ellis et al.1998; Ellis 2001). In addition, Salix and other pioneer species quicklyrevegetate ood-scoured areas, replacing older, senescent stands withyoung, green wood.

Restoration measures to reduce atypically high re frequencies haveyet to be implemented. On the Bill Williams River, for example, althoughsmall spring oods will be restored, the dense post-dam vegetation re-mains susceptible to re damage because of constraints on ood size.Experiments are needed on this and other rivers to determine whether fre-quent, small summer ood pulses can promote re resistance by increas-ing litter decomposition rates and dislodging dead vegetation. Ultimately,only dam removal may reliably shift the disturbance cycle from re backto oods.

35

Figure 2-13. Fire in Tamarix-dominated habitat along the Rio Grande, New Mexico.This par-ticular re was a controlled burn. (Photograph by John Taylor.)


0

10

20

30

40

Her

bace

ous

cove

r (%)


soils, included Viguiera dentata, Panicum obtusum, Plueraphis mutica,Sporobolus wrightii, and Sporobolus airoides (Stromberg et al., 1996).

There has been little quantitative study of the effects of dams and ood-ow alteration on plant diversity patterns in Sonoran riparian ecosystems.Some studies have found plant biodiversity levels to be low in the densethickets of Tamarix that can develop on ood-suppressed rivers (Brock,1994). Some authors have attributed this to an autogenic effect of theTamarix, but allogenic factors associated with river regulation may be atplay as well. For example, litter accumulation and development of densecanopies (which to some degree are artifacts of ood suppression) cancontribute to diminished understory biodiversity (Nilsson et al., 1999).

Plant species diversity can decline when ood disturbance becomes tooinfrequent or when the spatial and temporal diversity of ood disturbanceand ood-inuenced environmental conditions is reduced (Pollock et al.,1998). Cover and biodiversity of riparian herbs tend to increase on ne-textured soils (Stromberg 1998b; Jansson et al. 2000b) and, conversely,may decline below dams as nes are trapped in reservoirs and down-stream sediments become uniformly coarse (Figure 2-15). Restoring arange of soil conditions on dammed rivers will require restoring sedimentows during ood pulses, possibly through the use of sediment-bypassstructures. Restoring a range of temporal regeneration niches, to enablethe establishment of warm-season and cool-season plants, will requirerestoring the full seasonal range of ood ows.

Dam and reservoir development can also inuence biodiversity levelsby impeding the dispersal of propagules (Jansson et al., 2000a). Studiesare needed to determine whether dams are disrupting such plant dispersalprocesses along Sonoran Desert rivers. A study of dispersal patterns at thefree-owing Hassaympa River showed that most riparian plant specieswere adapted for dispersal by wind or animals (Drezner et al., 2001).Floods may, however, provide a secondary method of dispersing propag-ules laterally within the riparian zone and into downstream reaches. Damsand reservoirs diminish connectivity and may thus impede the travel ofmany animals, forming potential dispersal barriers without particular ref-erence to ood effects.

Along many free-owing rivers, the composition and diversity of theherbaceous vegetation has been altered by land uses, including livestockgrazing and irrigated agriculture. Where such management disturbanceshave been removed, natural oods may be a practical (albeit slow) means


Figure 2-15. A cobble bar along the Verde River, Arizona. Sediments have been trapped inupstream reservoirs, producing coarse soils in this below-dam reach. (Photograph by J.Stromberg.)

of restoring biodiversity, as demonstrated in a study of the HassayampaRiver. The Hassayampa is an alluvial river in central Arizonas SonoranDesert. It has been free-owing for more than a century, following a mas-sive dam failure in 1890. The Nature Conservancys Hassayampa RiverPreserve occupies an 8 km perennial segment of the river. Prior to 1986,the Preserve was a working cattle ranch and parts of the riparian corridorwere used as a trailer park and for commercial orchards. These activitiesprobably catalyzed development of the dense stands of exotic herbs thatoccur in the riparian habitats of the Preserve. Twenty-six percent of the344 plant taxa collected here were exotic, 74 percent of which were annualor biennial grasses and forbs (Wolden et al., 1994). Most of the 88 exoticspecies covered only small areas and were encountered infrequently, butsome became dominant. In 1989, areas adjacent to the channel were vege-tated primarily by the native shrub Baccharis salicifolia and the exoticBermuda grass (Cynodon dactylon). Exotic brome grasses (e.g., Bromusrubens), Bermuda grass, and wild barley (Hordeum murinum ssp. glau-


cum) dominated the Populus-Salix forest understory, and Bromus spp.and H. murinum ssp. glaucum dominated the understory of the Prosopisforests.

Populations of exotic species can persist for a long time after removalof the disturbance factor(s) that facilitated their invasion (Bartolome andGemmill, 1981; Milchunas and Lauenroth, 1995). They may produce self-favoring conditions, may have a long life span, or may dominate the seedsupply. Or abiotic conditions favoring these species may be persistent.Long periods may be required to reverse the physical effects of land use,such as soil compaction. We attempted to accelerate the recovery of nativespecies to the Hassayampa forest understories by imposing several res-toration treatments, including seeding natives and removing exotics byhoeing. We monitored treatment and control plots along stream banks, inmature Populus-Salix forests, and in Prosopis forests, for three years(Wolden and Stromberg, 1997). In separate studies, we monitored herba-ceous cover in plots after a ten-year return ood (Stromberg et al., 1993b).The restoration treatments overall were not very effective. However, fol-lowing the ten-year return ood, we did detect reductions in cover of manyexotics (e.g., Bromus rubens, Melilotus ofcinelis, Cynodon dactylon), in-creases in cover of some native species (e.g., Xanthium strumarium, Dico-ria canescens), and a general increase in diversity.

Other studies also show that oods can naturally restore native speciesto sites oristically altered by grazing (Chaneton et al., 1988). Floods ac-complish many of the same goals of restoration treatments. Flood watersdeposit a diverse seed mixture on fresh substrate while scouring or bury-ing established exotics. Seed banks, distributed like a veneer in the upperlayer of soil throughout the oodplain forests and shrublands, are a keysource for postood reestablishment of many native riparian herbs, in-cluding Bowlesia incana, Calibrachoa parviora, Centaurium calycosum,Mimulus guttatus, and Nicotiana obtusifolia (Boudell and Stromberg1999). Floods can create opportunities for exotic invasion in some circum-stances, but can also facilitate recovery of natives when anthropogenicdisturbances are removed and pressures, such as selection for grazing-tolerant species, are reduced.


Restoring Resilience and Self-Sustainability

Resilience, or resistance to ood disturbance, is considered to be a trait ofhealthy riparian ecosystems. Many riparian plant species have adapta-tions that allow them to resist ood damage or have sufcient regenerativeability to allow for rapid postood redevelopment (Stromberg et al., 1991).However, if subject to stressors such as dewatering or overgrazing, plantcommunities may lose their capacity for self-repair after ood disturbance.Along the Carmel River in California, groundwater pumping caused themortality of streamside vegetation. Without dense vegetation to stabilizebanks during oods or to recolonize them after the fact, the result wasextreme channel widening (Groeneveld and Griepentrog, 1985). Flood-related channel widening has also occurred along portions of the SanPedro River, where the combined inuences of water diversion, beavertrapping, and river incision have constrained the resistance and resilienceof the streamside vegetation (Fonseca, 1998). Livestock grazing, histori-cally ubiquitous in the riparian West, can cause loss or reduction in stream-side plant cover, loss of ne sediments and nutrients from the soil, channeldowncutting, and lowering of the water table (Armour et al., 1991; Trim-ble and Mendel, 1995; Belsky et al., 1999). All of these factors reduce thepotential of the riparian ecosystem to revegetate following ood dis-turbance. Floods, under such circumstances, can have adverse ecologicaleffects.

Riparian restoration projects that do not effectively promote resiliencewill inevitably fail in the long-term. Some restoration projects are beingdesigned to restore resilience and a capacity for self-recovery, by restoringperennial stream ows, regeneration-ood pulses, and connectivity to up-stream riparian ecosystems to ensure a supply of propagules. In othercases, ecosystem processes are not being restored and notions of self-sustainability remain unaddressed. For example, the U.S. Army Corps ofEngineers and municipal authorities are (as of October 2000) engaged inan ambitious and costly project to restore habitat quality to a watershed-decoupled reach of the Salt River through the heart of Phoenix, Arizona.Their congressionally funded technique entails planting some 75,000nursery-grown Populus fremontii, Salix gooddingii, Prosopis velutina, andriparian shrubs along an engineered low-ow channel. Water will be sup-plied to plants over the long term through pumping from a shallow aquifer(derived from urban runoff) into surface pools and a high-pressure spray

CONCLUSION 41

irrigation system. Because of the presence of an estimated 50 poorly doc-umented and economically untreatable landlls along the adjacent historicoodplain, there is neither intention nor capability to restore the hydraulicdynamics that normally drive riparian plant community development. Theproject planners and managers are unable to encorporate major oods as apositive, renewing factor. Rather, they fear expensive damage to park-like plantings and the unlooked-for exposure of some toxic legacy of un-regulated waste disposal (U. S. Army Corps of Engineers, 1998). Theresult must therefore be a sort of semistatic riparian diorama. There likelywill be short-term benets for some animal species. However, restora-tion in this instance resembles nothing so much as traditional urban struc-tural ood control sheathed in a biotic veneer.

CONCLUSION

Free-owing rivers in the arid Southwest can change within minutes froma string of quiet pools or a lazy trickle to wild, dangerous torrents. Despiteengineering efforts to manage and contain stream waters in reservoirs,oods periodically overwhelm these systems and spill onto urban lands,reminding us that control is an illusion. Efforts to address the situation arebeing discussed and experiments initiated to re-wild some of the rivers,albeit in a semicontrolled fashion, to restore some of the ecological, social,economic, and psychological benets of natural waters.

Further studies of ecological processes on free-owing reaches, and ofthe effects of experimental oods on regulated reaches, are criticallyneeded. The many demands and legal constraints on western waters neces-sitate substantial ecological documentation and justication for every dropreleased downstream. Sunbelt cities, including Las Vegas and Phoenix, areamong the most rapidly growing in the United States, putting increasingpressure on diminishing desert water supplies. The problems of riparianmanagement will worsen and the functions of riparian ecosystems will

Flood Pulsing in Wetlands Restoring the Natural Hydrological Balance

Documents

Transcript of Flood Pulsing in Wetlands Restoring the Natural Hydrological Balance