Front Cover
 Letter of transmittal
 Title Page
 Executive summary
 Table of Contents
 List of Figures
 List of Tables
 1. Introduction
 2. Status of riparian ecosystems...
 3. Functions and properties of...
 4. Fish and wildlife resources...
 5. The value of riparian ecosystems:...
 Literature cited

Group Title: Riparian ecosystems : their ecology and status
Title: Riparian ecosystems
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00014876/00001
 Material Information
Title: Riparian ecosystems their ecology and status
Physical Description: xvi, 155 p. : ill., maps ; 27 cm.
Language: English
Creator: Brinson, Mark M
National Water Resources Analysis Group (U.S.)
Eastern Energy and Land Use Team
Publisher: Eastern Energy Land Use Team and National Water Resources Analysis Group, U.S. Fish and Wildlife Service
Place of Publication: Kearneysville W. Va
Publication Date: [1981]
Subject: Aquatic ecology -- United States   ( lcsh )
Ecology -- Research -- United States   ( lcsh )
Wetland ecology -- United States   ( lcsh )
Genre: federal government publication   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: p. 127-154.
Statement of Responsibility: by Mark M. Brinson ... et al..
General Note: "Biological Services Program."--Cover.
General Note: "September 1981."
General Note: "August 1981"--Cover.
General Note: "FWS/OBS-81/17."
 Record Information
Bibliographic ID: UF00014876
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 001245439
oclc - 07902892
notis - AFZ5943
lccn - 81604061
 Related Items
Other version: Alternate version (PALMM)
PALMM Version

Table of Contents
    Front Cover
        Front Cover
    Letter of transmittal
        Page i
    Title Page
        Page ii
        Page iii
    Executive summary
        Page iv
        Page v
        Page vi
        Page vii
    Table of Contents
        Page viii
        Page ix
        Page x
    List of Figures
        Page xi
        Page xii
        Page xiii
    List of Tables
        Page xiv
        Page xv
        Page xvi
    1. Introduction
        Page 1
        Page 2
        Page 3
    2. Status of riparian ecosystems in the United States
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
    3. Functions and properties of riparian ecosystems
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
    4. Fish and wildlife resources in riparian ecosystems
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
    5. The value of riparian ecosystems: Institutional and methodological considerations
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
    Literature cited
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
Full Text

Biological Services Program



Their Ecology and Status



Fish and Wildlife Service

U.S. Department of the Interior

~ -" """""""""""""""""""" ..
r, n
r~--=--_. -,-, -
:* --~L---
aS- -cmFZ=--

Route 3, Box 44
Kearneysville, West Virginia 25430
September 25, 1981

Dear Colleague:

Enclosed is an Eastern Energy and Land Use Team (EELUT) publication entitled
"Riparian Ecosystems: Their Ecology and Status." This report provides a
review and synthesis of available information that can be used to document and
assess the ecological values of riparian ecosystems. Included are chapters
covering: 1) the status of riparian ecosystems; 2) ecological functions and
properties of riparian ecosystems; 3) importance of riparian ecosystems to
fish and wildlife; and 4) considerations in valuation of riparian ecosystems.
This product is intended to serve as a reference document for biologists in
Federal and State water resource and fish and wildlife agencies, and for
private organizations interested in conservation of riparian resources.

Readers should feel free to cc :act us for more information and publications
pertaining to riparian ecosystems, through:

Information Transfer Center
Eastern Energy and Land Use Team
U.S. Fish and Wildlife Service
Route 3, Box 44
Kearneysville, WV 25430
(FTS) 925-5265 or (304) 725-2061, ext. 5265


Edgar A. Pash
Team Leader


September 1981



Mark M. Brinson, Bryan L. Swift,
Reuben C. Plantico, and John S. Barclay

Eastern Energy and Land Use Team
National Water Resources Analysis Group
U.S. Fish and Wildlife Service
Kearneysville, West Virginia 25430


The purpose of this publication is to document and interpret the information that is
available on riparian ecosystems so that the consequences of their alteration and
deterioration can be assessed at a national level. The common functional properties of
these ecosystems and their attractiveness to wildlife make it possible to address
riparian ecosystems as discrete and manageable entities.

The Fish and Wildlife Service has been involved in several efforts that have led to
the development of the document. Much of the earlier concern was for the consequences
of channelization and other stream alterations on fish and wildlife communities. It
was soon recognized that most stream alterations could not be considered separately
from changes in floodplain vegetation and animal communities. The growing body of
literature on riparian ecosystems suggested a strong interdependency between stream and
floodplain processes.

A national symposium held in 1978 on "Strategies for Protection and Management of
Floodplain Wetlands and Other Riparian Ecosystems" was an attempt to focus attention on
the research of individuals that were working largely on ecosystems associated with
streams.' The following year a workshop on riparian ecosystems in Harpers Ferry pro-
duced a number of strategies and alternatives for riparian ecosystem protection and
enhancement in which the Fish and Wildlife Service could potentially participate.2
This more recent effort is a "second generation" state-of-the-art whereby we summarize
and synthesize what is known about riparian ecosystem function, values, and management.

This publication is intended to provide a geographically balanced treatment of
technical information on riparian ecosystems from a nationwide perspective. By fo-
cusing on the common properties of these ecosystems, recommendations and decisions that
affect their management and protection should be simplified. The manuscript is ori-
ented to provide assistance to decisionmakers involved in ecosystem management who must
utilize ecological principles and information.

Any suggestions or questions regarding this report should be directed to:

Eastern Energy and Land Use Team
Route 3, Box 44
Kearneysville, WV 25430

'Johnson, R. R. and J. F. McCormick (tech. coord.). 1978. Strategies for protection
and management of floodplain wetlands and other riparian ecosystems. USDA Forest
Service, Gen. Tech. Rep. WO-12. Washington, D.C. 410 pp.
2Warner, R. E. 1979. Proceedings of a workshop on fish and wildlife resource needs
in riparian ecosystems. Eastern Energy and Land Use Team, U.S. Fish and Wildlife
Service, Kearneysville, WV. 53 pp.


This report describes the func-
tions, values, and management of river-
ine floodplain and streambank ecosys-
tems, henceforth called riparian eco-
systems. The report is composed of
sections on the status of riparian eco-
systems, their ecological function and
properties, wildlife resources, and
valuation considerations. This brief
synopsis of the four sections provides
an overview of the material covered in

In the absence of a comprehensive
inventory of riparian ecosystems in the
U.S.A., existing resource inventories
provide only a rough indication of the
extent and distribution of these eco-
systems. However, when taken together,
the data give a great deal of insight on
the amount of riparian ecosystem in
existence, the quantity of natural area
lost to a variety of other uses, and the
nature of alterations.

One liberal estimate of the amount
of land subjected to flooding (100 year
floodplain) and thus potentially sup-
porting riparian ecosystems is 121 mil-
lion acres, or 6% of the land in the
U.S.A. (excluding Alaska). In reality,
much less exists in a natural or semi-
natural forested condition, and a con-
servative estimate is 23 million acres.
From other sources, we estimate that
approximately 70% of the original flood-
plain forest has been converted to urban
and cultivated agricultural land uses.

Case histories of riparian ecosys-
tem status and condition show large dif-
ferences in loss from place to place,
but as much as 95% loss of natural vege-

station has been reported in some are
Examples for the lower Mississil
Colorado, Sacramento, and Misso
Rivers have been particularly well do<
mented, and, in comparison with es
mates of loss of natural vegetation
uplands, put riparian lands in n
category of the most severely alter
ecosystems in the U.S.A.

Along with data on losses in n
tural floodplain forests, the magnitu
of stream alteration provides an ir
pendent assessment of changing condi
of riparian ecosystems. About 60%"
the major stream segments have be
judged unsuitable for inclusion in
National Wild and Scenic Rivers Sys
because of water resource or other cu
tural developments within riparian co
riders. Numerous examples exist 3
losses in stream length due to chan
realignment and alteration. Losses
surface area of riparian ecosystem
doubtedly occur in larger proport
than loss in stream length because lal
amounts of drainage and forest clearing
usually accompany relatively small I
ductions in stream length. Impoundme
have also inundated significant areas
riparian vegetation, and the downstre;
effects of modified streamflow on ripI
ian ecosystem function have seldom b

Alteration and loss of natural
parian ecosystems, as compared with
land ecosystems, are of particular c
cern because of the greater magnitude (
modification required for conversion
other uses. The potential for res
ration is lower because drainage pr(
cludes most other goods and services t
society that flood-dependent ripar1
ecosystems provide. U




Over geologic time periods, streams
undergo phases of erosive downcutting
and alluvial deposition. At the same
time stream channels migrate back and
forth across floodplains, a process
which results in a continual replacement
and displacement of the plant and animal
communities. In this way a stream is
responsible for "organizing"the flood-
plain into a variety of diverse commun-
ities, many of which are controlled by
the depth, duration, and frequency of

Flooding and flow water are also
responsible for depositing and eroding
sediments. Both the suspended material
and the water that carries it represent
supplies of materials from sources out-
side the floodplain. Upland ecosystems
lack a similar lateral transport system;
consequently this is one of the funda-
mental differences between upland and
riparian ecosystems. Both the abundance
of water and nutrient supply are par-
tially responsible for maintaining the
productivity and vitality of riparian

Primary productivity may be -re-
garded as an indicator of the vitality
of an ecosystem. Not only does primary
productivity initiate organic energy
flow for food webs, but another of its
fundamental functions is to maintain the
structural integrity of the ecosystem.
Studies done on floodplain forests of
the Southeast show that they are among
the most productive ecosystems in the
nation. Riverine wetlands also export a
disproportionate amount of organic mat-
ter as compared with an equivalent area
of upland ecosystem. Thus they augment
the amount of energy and structural car-
bon that downstream aquatic ecosystems,
particularly estuaries, receive from
continental runoff. Instream communi-
ties also are highly dependent on leaf
litter from streamside forests for main-
taining metabolism and ecosystem struc-

Differences in nutrient cycling be-
tween floodplains and upland ecosystems
are related to (1) the influence that
flooding and an "aquatic" phase has on
restricting oxygen availability to soils

and sediments, hence altering the meta-
bolic pathways of microbial communities,
and (2) the aqueous transport system
that provides pathways of exchange
through lateral imports, sedimentation,
and exports of nutrients. Most nutrient
cycling studies conducted in southeast-
ern floodplain forests suggest a high
capacity to absorb and recycle nutri-
ents. In arid riparian ecosystems, the
quantity of water, rather than its qual-
ity, is an overriding factor in ecosys-
tem processes. The potential for flood-
plains to have an influence on the nu-
trient status of floodwaters depends
partly on the length of time and the
quantity of water and nutrients that
come in contact with the floodplain.

It should be possible to predict
the severity of damage that a particular
alteration will have on normal ecosystem
processes based on an understanding of
natural ecosystem function. Alterations
of ecosystems can be categorized as
changes in geomorphic processes and
water delivery patterns, physiological
stress, and biomass removal. Stream
channelization, containment of stream
flow and channel constriction, impound-
ments and diversions, introduction of
toxins, grazing by livestock, timber
harvest, and hunting and fishing corre-
spond with one or more of the three
alteration categories.

From this analysis it is possible
to predict the consequences of the
seemingly diverse sources of intrusions
into riparian ecosystems. If goals of
mitigation are to restore the multiple
services that these ecosystems provide
in their natural condition, some altera-
tions can be mitigated and others clear-
ly cannot. If the principal sources of
energy and material continue to be sup-
plied to the system, there is a high
probability of recovery. If these
sources are blocked or diverted, mitiga-
tion to reverse the damage can occur
only after great investments of time,
energy, and money.


Many of the attributes of riparian
ecosystems that make them attractive to
humans are also responsible for the suc-

cess and maintenance of wildlife popula-
tions. These characteristics include
the presence of flowing water, moist and
nutrient rich soils, relatively high
plant productivity, and corridors for
migration and travel. The structural
complexity of these ecosystems, particu-
larly in comparison with uplands in arid
climates, provides many habitat require-
ments and adds to the landscape diver-
sity of the regional geography.

During the past decade, a large
number of studies have documented that
riparian ecosystems unquestionably pro-
vide essential habitat requirements for
a large diversity of vertebrate species.
More migratory and nesting species of
birds have a higher affinity for ripar-
ian ecosystems than they do for upland
ecosystems. Although catastrophic
flooding may temporarily reduce the
abundance of "terrestrial" vertebrates,
these species are adapted to rapid re-
colonization after flood conditions sub-
side. In fact, certain fish populations
are augmented by enormous increases in
feeding area that floodplain inundation
provides, in addition to the seasonal
supply of leaf fall into the water sur-
face of the stream channel under
non-flooding conditions.

The reasons for dependence on and
affinity for riparian ecosystems by such
a large and disproportionate number of
vertebrates are due to a multiplicity of
factors. The presence of flowing water,
high plant productivity, and nutri-
ent-rich conditions have already been
mentioned as contributing factors.
Perhaps of more fundamental importance,
riparian and floodplain ecosystems re-
present a combination of aquatic and
terrestrial ecosystems that have some-
what separate spatial and temporal di-
mensions. Habitat features change
dramatically with only small topographic
differences, such as the gradient from
an open water stream channel to a dense
gallery forest. The duration and timing
of flooding superimposes a seasonal
dimension on these gradients. For these
spatial and temporal dimensions to be
maintained, it is essential that the
changing geomorphic forces that drive
riparian ecosystems be allowed to orga-
nize and reorganize the plant and animal


Allocating land and water in ripar-
ian ecosystems among various uses and
assessing the relative social values of
these competing uses are issues of im-
mediate and major concern. Riparian
systems are generally considered quite
valuable because of their ecological
values and natural service functions.
However, institutional mechanisms for
allocating resources such as land and
water are designed to serve perceived
human wants and needs. Therefore, the
way in which private and public insti-
tutions allocate natural resources will
determine whether riparian systems are
left relatively undisturbed for
wildlife, timber, specific kinds of re-
creation, natural flood storage, water
quality enhancement, and groundwater re-
charge; or whether they are altered for
agricultural production, navigation
benefits, flood protection, or commer-
cial development. Central to this pro-
cess are the forces and incentives which
drive resource allocation in one direc-
tion or another and the manner in which
preferences and values are weighed in
decisionmaking processes which directly
affect the resources.

The causes of land use patterns in
riparian systems appear to be very com-
plex. In some respects they are. Soy-
bean demand, grazing rights on public
land, tax laws affecting property and
estates, and public flood control pro-
jects are but a few factors which appear
to affect land and water use in flood-
plain ecosystems. However, there are
broader and, in some respects, more
meaningful categories:

1. Market forces affecting private
investment patterns (consumer de-
mand for specific goods and ser-

2. Political forces affecting private
investment (world trade policies,
regional economic development,
public subsidies); and

3. Institutional factors affecting
private and public decisionmaking
which include:

a. f-arket decisionmaking (proper-
ty rights specifications,
failure of markets to capture
costs and benefits of private
transactions, information pro-
blems), and

b. Nonmarket (government) insti-
tutions and activities (taxes,
subsidies, regulations which
affect the incentives of pri-
vate decisionmakers to engage
in particular activities, and
publicly conducted and as-
sisted projects).

Having analyzed these categories of fac-
tors, one can focus on specific poli-
cies, programs, and decisions which de-
termine the fate of riparian systems.

Another distinct aspect of economic
analysis of resource allocation in ri-
parian systems concerns valuation. How
does one value the various competing
uses of riparian systems? This problem
arises most frequently in the context of
public decisionmaking processes whereby

public officials must weigh the value of
one land use versus another (i.e.,
through permitting-licensing activities,
zoning decisions, funding of public
projects, etc.). Typically, public
decisionmakers are confronted with two
very different kinds of information re-
garding values: ecological and eco-
nomic. The decisionmaker is faced with
the dilemma of evaluating noncomparable
values before reaching a decision. How-
ever, ecological values have economic
significance. For example, if riparian
system alteration were to result in lost
natural flood storage, lower water qual-
ity, and fewer wildlife resources, what
is the "cost" of these foregone oppor-
tunities? Since we do not pay land-
owners to maintain land for these pur-
poses, it is difficult to assess soci-
ety's demand for them as expressed
through market prices (reflecting ag-
gregate willingness-to-pay). This ne-
cessitates use of some surrogate value.
We provide a brief review of the ap-
proaches to natural resource valuation
and a critique of each of the method-




PREFACE . . . . . . .. iii I
EXECUTIVE SUMMARY . . . . ... . iv
LIST OF FIGURES . . . . . . . .
LIST OF TABLES . . . . . . .
ACKNOWLEDGEMENTS . . . . . . . .
Scope . . . . . .. . I1
Nationwide Extent of Riparian Ecosystems . . . . 4
Inventories of Floodplain Area . . . ... . 4
Inventories of Streams and Rivers . . . ... 7
Losses of Riparian Ecosystems . . . . . 7
Alterations of Floodplains . . . . . 7
Alterations of Streams and Rivers . . . . ... 12
Assessment of Riparian Ecosystem Status . . . ..... .15
Fluvial Processes. . . . . . . .. 17
Geomorphology . . . . . . 18
Aggradation and Degradation . . . . 18
River Meanders and Topographic Features . . . 20
Hydrology and Hydroperiod . . . ..... ....... 23 |
Surface Water . . . ..... . .... 23
Ground Water ................. . 25
Significance of Fluvial Processes ................. 26
Energy Flow and Biomass Distribution .. . ............. 26 3
Biomass Distribution and Accumulation .. ............. 28
Ecosystem Metabolism. . . . ..... . ..28
Factors Affecting Primary Productivity and Growth . . ... 31
Energy Transfer from Producers to Consumers . . . 32
Export of Organic Matter from Swamp-Draining Streams . ... 32
Energy Flow and Community Structure of
Upland-Draining Streams. . . . . 34
Nutrient Cycling . . . . . 37
Distribution of Nutrients . . . . . 37
Major Flows in the Nutrient Cycle . . . ..... .. 38
Soil-Water Nutrient Exchanges . . . . . 41
The Significance of Hydroperiod and Nutrient Cycling . ... 43


viii i


Diversity Among Riparian Ecosystems. ..... . . . 46
Climate . . . . . . . . 46
Underfit Streams and Downcutting Channels .. ........... 47
Influence of Catastrophic Forces . . . .... 48
Ecological Succession . . . . ... ....... .48
Description of Plant Communities . . . .... 49
Southern Forest Region . . . . ... ..... .49
Central Forest Region . . . . . . 51
Eastern Deciduous Forest Region . . . . .. 53
Northern Forest Region . . . . ... ..... .53
Plains Grassland Region . . . . . .. 54
Mediterranean and Western Arid Forest Regions. . . ... 56
Pacific Northwest and Rocky Mountain Regions . . .... .57
Alaska .. .. . . . . . .. ... 57
Effects of Ecosystem Alteration on the Properties of Ecosystems. .... 60
Stream Channelization . . . . . . 62
Containment of Streamflow and Channel Constriction. . . .. 64
Impoundments and Diversions . . . . .. ... 65
Introduction of Toxic Compounds . . . . ... .. 66
Grazing by Livestock . . . . .. ....... 66
Timber Harvest. . . . .... . . 67
Hunting and Fishing . .. ... . ..... ........ 68

Habitat Values of Riparian Ecosystems for Fish and Wildlife. . .. 69
Predominance of Woody Plant Communities . . . .. 69
Presence of Surface Water and Abundant Soil Moisture. . . .. 70
Diversity and Interspersion of Habitat Features. . . 72
Corridors for Dispersal and Migration . . . . .. 74
Responses of Fish and Wildlife to Habitat Variables. . . ... 75
Vegetation Type . . . . .. . . 75
Size and Shape of Riparian Area . . . . .. .. 76
Stream Type and Hydrologic Patterns . . . .... 78
Adjacent Land Use . . . . ... ........ 79
Elevation . ......... . ... ............. 80
Characteristic Riparian Wildlife Communities . . . ... 80
Birds. .. . .... ...... .. .. ... ... .... 82
Community Characteristics. .. ................ 82
Characteristic Species . . . . ... ...... 84
Mammals . . . . . . . . 86
Amphibians and Reptiles . . .............. 86
Significance of Riparian Ecosystems to Fish and Wildlife .. ..... 87
Comparison of Riparian and Nonriparian Wildlife Communities . .. 87
Dependence of Fish and Wildlife Species on Riparian Ecosystems. .... 95

Valuation: An Institutional Perspective . . . ... 100
Overview. . . . . .. .. ...... 100
Environmental Problems as Economic Problems. ........... 101
The Performance of Institutions. . . . . .. 102



Valuation Problems in the Private Sector: Market Failure . 102
Causes of Resource Misallocation . . ......... 103
Property Rights, Public Goods, and Transaction Costs .. ... 104
Valuation in the Public Sector: Opportunities and Problems ...... .. 105
Problems with "Implicit" Valuation . . . ... 108
Other Variables Affecting Public Evaluation. . . .. 109
A Methodological Perspective on Valuation. . . . .... 109
The Basis for Resource Valuation. . . . . ... 109
Ecological Values and Their Assessment. . . . ... 110
Identifying and Organizing Information
for the Valuation Process. . . . . .... 110
Approaches to Valuation (I): Qualitative and Other Statements
of Value . . . . . . . 113
Wetland Ranking Methods .. . ............. .. 113
Approaches to Valuation (II): Economic Methods . . .... .. 114 I
Assumptions, Methods and Limitations. . . . .. 115
Valuation Techniques . . . . .... 116
Selected Studies of Natural Values . ... ... ...... 118
Approaches to Valuation (III): Life Support or Energy Analysis . 120
Values Based on Life Support . . . . ... 120
Values Based on Energy Analysis, Corrected for Quality . ... 121
LITERATURE CITED ........ .......... ... .... 123









umber Page

1 Distribution of warmwater, coldwater, and intermittent streams
in the United States. . . . . ... . 9

2 Extent of water resource development on streams in the
United States . . . . .. . . . 12

3 Land area covered by riparian vegetation, streams, and
reservoirs in the United States . . . . ... .. 16

4 Two sequences of events leading to the development of the same
surface geometry in terraces and floodplains. . . . .. 19

5 Typical floodplain topographic features, illustrated
diagrammatically, of the Mississippi River near
False River, Louisiana. . . . . ... ...... .22

6 Rates of erosion and diagrammatic side views of stream banks
in permafrost environments . . . ... .... 23

7 General model of floodplain hydrology ... . . ...... .27

8 Effect of increasing annual precipitation on the basal area of
vegetation for upland forests (curve) in comparison to riparian
forests (no pattern).. . . . . . 30

9 Relationship between annual river discharge and cypress tree
growth on the Cache River, Illinois . . . ... 31

10 Changes in structure and function of upland-draining streams
from headwater to mouth . . . . . .. 34

11 Immobilization of phosphorus and nitrogen by decaying leaf
litter in an alluvial swamp . . . .... ..... .42

12 Pathways of nitrogen transformations in an oxidized and
reduced sediment-water system . . . .. .43

13 Phosphorus storage and fluxes in the Creeping Swamp
floodplain ecosystem, North Carolina. . . . 44

14 Seasonal phenology of a tupelo-cypress swamp showing mechanisms
of nitrogen conservation and recycling . . . ... 45



15 Map showing the 2.5 cm isopleth of annual runoff. . . . 46
16 Forest regions in the United States for which riparian plant
communities are described . . . ..... ....... .. 50
17 Idealized profile of species associations in southeastern
bottomland hardwood forests. After Wharton (1978). . . ... 51
18 Changes in height and species composition of floodplain forest
stands along a west to east gradient in Oklahoma. . . ... 54
19 Cross section of the Missouri River in North Dakota showing the
distribution of important tree species. . . . .. 55
20 Profiles of five vegetation types along the Rio Grande from
El Paso to Albuquerque. . . . ... .... 56
21 Profile of vegetation along major rivers in the Sacramento
Valley, California. . . . ... ........ .. 58
22 Cross section of floodplain and terrace communities of the
McKenzie River, Oregon . . . .. . .. 59
23 Zonation of plant communities along an arctic stream. . ... 60
24 Major flows of energy in a floodplain ecosystem ............ 61
25 Changes in channel morphology of the Missouri River (A)
between 1879 and 1954, and in the Gila River (B) from
1914 to 1962 . . . ... . . . 65
26 Daily variation in river stage for the Colorado River at
Lees Ferry during water year 1939 and 1973 . . . ... 66
27 Functions of riparian vegetation as they relate to aquatic
ecosystems. . . . ... . . .. 71
28 Distribution of riparian wildlife species in relation
to streams . . . . . . . .. 71
29 Synchrony of events related to flooding in a floodplain-river
system in the tropics . . . ... . . 73
30 Edges and ecotones in riparian ecosystems . . . ... .74
31 Relationship of wildlife diversity to size of a plant
community type . . . . . . ... 74
32 Riparian zones are frequently used as migration routes by
wildlife, such as mule deer (Odocoileus hemionus) which
travel along streams between high elevation summer range
and low elevation winter range . ..... .. .* ...... 74
33 Fish and wildlife values at small stream impoundments . ... 79



i34 Distance travelled by riparian bird species into agricultural
areas . . . . . . . . 80
35 Number of breeding bird species on 98 riparian census plots . 82
36 Breeding bird densities on 98 riparian census plots . ..... 84
37 Number of bird species in riparian and upland vegetation
types . . . . . . . .. . 95





Number Page
1 Potential and present area of the four predominant riparian
vegetation types in the United States . . . . 5
2 Estimated area of riparian ecosystems in 26 States,
or portions thereof . . . ..... . .. 6
3 Length of streams in the United States. . . . . 8
4 Total length of major stream segments in the U.S.A. and
percentage unsuitable for designation as National Wild and
Scenic or Recreational Rivers . . . ... ... 9
5 Summary of case histories showing losses of riparian ecosystems . 10
6 Area of bottomland hardwoods in the lower Mississippi Valley,
1957 to 1977 by State . . . ..... ... .... .. .. 11 I
7 Summary of Soil Conservation Service channel work through
1972, 1980. . . . ... . . ...... .13
8 Extent of stream alterations in twelve States . . .... .14
9 Deposition rates in forested floodplains. . . . ... 21
10 Dependence of relative stand age on location in a floodplain. . 23
11 Structural characteristics and biomass production of riparian
forests . . . ..... . . ....... 29
12 Concentration and export of organic carbon in drainage waters
for upland- and swamp-draining watersheds . . . ... 33
13 Comparison of litterfall, primary productivity, and respiration
for several sizes of streams. . . . ..... 36
14 Distribution of phosphorus in riverine forests. . . .... 38
15 Litterfall and aqueous flows of phosphorus from the canopy to
the forest floor in riverine swamps . . . . ... 39
16 Summary of decomposition rates of litter in riverine forests. . 40
17 Sedimentation rates of phosphorus in the floodplains of
riverine forests. . . . ... ....... .. 41










1 32





Tree biomass, net annual accumulation, and distribution among
tree species (%) for a floodplain, transition site and upland
along a stream in Illinois ... . . . . 52
Examples of riparian ecosystem alteration and their relationship
to categories of alteration shown in Figure 24 . . ... 6
Importance of aquatic and terrestrial invertebrates in diets
of North American stream fishes . . . . .. 72
Width of riparian buffer strips recommended to protect water
quality and aquatic life in streams . . . . .. 77
References for information on riparian wildlife communities
in the U.S.A . . . . . . . .. 81
Number of breeding bird species on riparian study areas . .. 83
Breeding bird densities in riparian ecosystems. . . . 83
Number of winter bird species on riparian study areas . .. 84
Densities of riparian bird populations in winter. . . 85
Foraging guilds of riparian birds . . . . 85
Most abundant breeding birds on 98 census lots in
riparian vegetation . . . . . . 87
Distribution of common bird species in riparian ecosystems. . 88
Bird species that are locally abundant along streams and rivers .. 91
Number of mammal species in riparian ecosystems . . ... 91
Distribution of common riparian mammals . . . .. 92
Common reptiles in riparian ecosystems . . . ... 93
Common amphibians in riparian ecosystems . . . .. 94
Comparisons of bird densities between riparian and upland
ecosystems . . . . . . . 94
Proportion of wildlife species using riparian ecosystems . .. 95
Number of terrestrial wildlife species dependent on or
preferring riparian ecosystems . . . . 96
Threatened and endangered animal species in riparian ecosystems .. 97
Qualitative list of values of riparian ecosystems . . .. 99
Information requirements for the economic assessment
of riparian ecosystem values . . . . .. 112




The preparation of this report would not have been possible without the cooperation
and assistance of a number of dedicated individuals. Discussions and visits with Roy
Johnson, Richard Warner, and Charles Wharton were of immense value. Review comments by
numerous other persons from Federal agencies have contributed, significantly to this
effort. East Carolina University provided support for the senior author during the
final stages of report preparation.
Much of the impetus for organizing this work on riparian ecosystems was provided by i
Charles Segelquist. His sustained encouragement to the authors and perseverance in
riparian affairs are largely responsible for completion of this report. The various
chapters were prepared by the following individuals: Mark Brinson prepared Chapters 1
and 3 and was responsible for final editing and organization; Bryan Swift wrote Chap-
ters 2 and 4 with the assistance of John Barclay; Reuben Plantico prepared the material
in Chapter 4 with the assistance of Mark Brinson. Karen Byrd typed several drafts of
each of these chapters and the final report.
Credits are due for adaptations of figures used in the manuscript. Figure 1 was
used by permission of the American Fisheries Society. Figure 4 is from Fluvial
Processes in Geomorphology by Luna B. Leopold, M. Gordon Wolman, and John P. Miller,
published by W. H. Freeman and Company, copyright 1964. Figure 6 was reproduced with
modifications with the permission of the National Council of Canada from the Canadian
Journal of Earth Sciences, Volume 9, 1972. Figures 9, 11, and 18 were adapted with I
permissionof the Ecological Society of America. Figure 19 was used with permission of
the Canadian Field-Naturalist. The Northwest Scientific Association granted permission
to produce Figure 22 and Ann Arbor Science allowed a reproduction of Figure 12.








I This document addresses the func-
tions, values, and management of river-
ine floodplain and streambank ecosys-
tems, hereafter called riparian eco-
systems. An abundance of water and rich
alluvial soils are among the more
important attributes that distinguish
these ecosystems from uplands. River
orridors represent lines of convergence
where the energy of flowing water has
delivered and concentrated erodible
Materials from diffuse sources in the
landscape. Because of these special
attributes and life-supporting features,
human society has long perceived their
sefulness as sites for urban settle-
ents, as conduits for transportation,
and as a source for harvestable products
wuch as timber, crops, and wildlife.

In comparison with average stream
flow, catastrophic episodes of stream
Flooding are more important in molding
nd shaping the landscape through
(rosion, sedimentation, alteration of
river courses, and rejuvenation of
vegetation. A major flood may occur
during any given year, and the best we
an do is predict the probability of its
reaching a particular height and car-
Iying a given quantity of materials.
because of this uncertainty humankind
often has found itself poorly adapted to
utilizing the resources and benefits of
these ecosystems.

Depending on the form of the
' iparian ecosystem and the particular
ocality within it, water levels may
ange from prolonged seasonal inundation
of floodplains to periodic rises in the
Iubsurface ground water of streamside
forests. When human intrusions alter
he natural temporal and spatial pattern
of water flow, the essential features

upon which riparian ecosystems depend
are threatened. By the same token,
alteration of these ecosystems may
prevent them from providing valuable
life support services to society such as
maintenance of water quality, flood
water storage, and the production of
quality timber, fish, and wildlife.

This is not to suggest that
riparian ecosystems are immune to man-
agement. On the contrary, judicious
management may be the preferred alter-
native, particularly in the context of
the numerous alterations that have al-
ready occurred in many watersheds.
Distinctions need to be made between the
types of alterations that can be tol-
erated and the degree to which altera-
tions can be made without threatening
the carrying capacity of riparian eco-
systems for providing values and ser-
vices to society. In order to be in a
position to make riparian management
decisions, it is essential that we
understand the function and importance
of the flows of energy and materials
within and through riparian ecosystems.
This is a necessary prelude to estab-
lishing the values of the services that
riparian ecosystems provide society.


The riparian ecosystems discussed
in this report are those associated with
streams and rivers. We include the full
continuum from intermittent headwater
streams with negligible floodplains to
broad meandering rivers, but exclude
flooded coastal features such as salt
marshes and mangrove swamps. The main
focus is on floodplain and streambank
plant and animal communities which are

affected by the stream through addi-
tional water supply, flooding, or
lateral transport of nutrients and
sediments. It is recognized that
riparian ecosystems also may have pro-
found effects on streams. The magnitude
of the interaction will be somewhat site
specific and depends partly on relative
sizes of each. In general, streambank
forests will influence to a greater ex-
tent the ecological processes in small
streams than in large streams. Like-
wise, streams with high discharge
usually will have a greater influence on
riparian forests than small ones, par-
ticularly in areas of the floodplain
that are frequently inundated.

We recognize that "riparian zones"
are not restricted to riverine ecosys-
tems, and that the term is frequently
applied to the more robust vegetation
associated with seeps, springs, meadows,
bogs, margins of ponds and lakes, and a
number of other "wet" features found in
the predominately arid regions of
western U.S.A. Although many of these
wetter areas have important hydrologic
functions and unquestionable wildlife
values (e.g., playa lakes), from a
functional and management standpoint,
they probably have more in common with
non-flowing water systems in more humid
regions, such as certain bog depres-
sions, lakes, prairie pothole marshes,
limestone sinks, and Carolina bays.
Because these predominately stillwater
systems differ from riverine systems,
their management and values should be
approached with fundamental hydrologic
and geologic differences in mind. The
alterations to which riverine and
stillwater systems are subjected also
differ in many instances.

Riverine riparian ecosystems over-
lap a great deal with some of the eco-
system types in the wetland classifica-
tion system of the Fish and Wildlife
Service (Cowardin et al. 1979). How-
ever, we discuss some plant and animal
communities that are not included in the
wetland classification system. This
encompasses areas where streams have the
effect of supplying water, sediments,
and nutrients that would otherwise not
be available under "upland" conditions.
Often these lowland areas are clearly
not areas that are "flooded or saturated
at some time each year" (Cowardin et al.

1979, p. 4) nor do they necessarily have
"predominately hydrophytic cover"
ibidd., p. 3). In addition to the phys-
iological adaptations to flooding,
drought may be an important selective
force for plants in floodplains of arid
climates. However, the physical aspects
of flooding and water flow may be
equally important in determining the
structure and function of riparian com-
munities. This is especially evident
where plant community form and function
are influenced by floods that recharge
groundwater supplies, initiate new com-
munities by removing vegetation, and
provide moist, exposed seedbeds for
germination and growth. Whereas, one of
the main purposes of the wetland classi-
fication system is to "...ensure uni-
formity throughout the United States..."
and one of its principal uses will be
"...the inventory and mapping of wet-
lands..." (Cowardin et al. 1979), the
main purpose of the present document is
to describe the ecological properties
and natural values of riparian systems
and their associated streams.

The Marine and Estuarine Systems of
the wetland classification system are
not included in the riparian category
here because our emphasis is on the eco-
systems associated with the millions of
kilometers of inland streams in the
U.S.A. However, the obvious functional
influence of exports from certain ripar-
ian ecosystems on estuarine and marine
systems is discussed.

The Palustrine and Lacustrine Sys-
tems, where they occur in floodplains
and, in their natural state, become con-
nected to the stream when it floods, are
included in this synthesis as part of
the riparian system. This normally
would include large (>8 ha) and deep (>2
m) oxbow lakes and lakes of levee flank
depressions. Palustrine and Lacustrine
Systems may be either a large or negli-
gible part of a given sector of flood-
plain. The wetland classification sys-
tem does not include wetlands occurring
on the river floodplain as part of the
Riverine System. For the purposes and
uses of the wetland classification sys-
tem (uniformity, inventory, and map-
ping), this may be desirable because the
number of categories is reduced and the
hierarchy simplified by omitting Lacus-
trine and Palustrine Systems in flood-

lains from the Riverine System.
owardin et al. (1979) suggest
that "It is the ground water that con-
Irols to a great extent the level of
ake surfaces, the flow of streams, and
the extent of swamps and marshes" (p.
0). However, under arid climatic con-
itions where evapotranspiration exceeds
ocal precipitation, deprivation of
streamflow would cause the disappearance
f Lacustrine and Palustrine floodplain
features except in anomolous situations
here large rock aquifers provide most
of the water supply.

I Whether the Riverine System of the
etland classification system is in-
cluded as a part of our functional
iparian concept depends on where one
hooses to draw boundaries. Although we
ocus primarily on properties of
streambank and floodplain plant and
Minimal communities, the influence of
hese communities on the stream, and the
stream on these communities, makes it
impossible to discuss one without the
their. The problem with establishing
*oundaries between the two is the tend-
ency to not consider the movement of
later, matter, and organisms that pro-
ides the basis for coupling among eco-
ystems. Thus, the Riverine System is
included to the extent that it plays a
unctional role in maintaining natural
properties and attributes of riparian
Iw Another set of "boundary" problems
s in the headwater portions of streams
where recognizable floodplains cease to
'xist and, at some point, riparian vege-
ation disappears. Usually erosion pre-
bominates and floodplain area is negli-
gible in headwater streams because the
Amount of material available for allu-
ial deposition decreases due to dimin-
shing size of the watershed. There may
be a gradual transition from regions of
lluvial fill to upstream areas where
channels are eroding and the channel is
confined by bedrock.' Leopold et al.

Even sectors of large rivers may be
confined by bedrock and be undergoing
Rapid downcutting. Under these condi-
tions, zones of vegetation that are in-
fluenced by the stream may be quite

(1964) observed that in humid climates
this upper limit of floodplain develop-
ment in stream systems appears to be the
point at which flow in the channel
changes from perennial to ephemeral,
i.e., where groundwater supply is insuf-
ficient to sustain flow through nonstorm

They suggested it is possible that
perennial flow promotes rock weathering
and subsequent sloughing into the chan-
nel, hence initiating lateral deposition
and erosion along a small stream. In
arid climates where intermittent streams
are common because of protracted drought
and high evaporative demand, these cri-
teria would not appear to apply. It is
possible that the vicinity of headward
gully erosion and gully wall collapse
(Leopold and Miller 1956) may represent
the upper limit of floodplains in arid
climates. However, riparian vegetation
often continues upstream from that point
and thus is not restricted to flood-

One of the problems of dealing with
riparian ecosystems from a national per-
spective is the great diversity in vege-
tation, fauna, and geomorphology that
exists. A geographically balanced syn-
thesis of information is difficult to
achieve because of the regional differ-
ences among research approaches. For
example, many nutrient cycling studies
have been done on southeastern flood-
plain forests because of the importance
of these systems for water quality.
Equivalent nutrient cycling studies are
entirely lacking in arid floodplain
forests where water, rather than nutri-
ents, limits ecosystem processes. On
the other hand, the water regimen of
arid riparian floodplains has received
considerable attention, yet equivalent
studies are lacking in the Southeast.
The ecological realities of different
controlling factors in the wide diver-
sity of riparian ecosystems in the
U.S.A. must be recognized and appreci-

narrow relative to broad floodplains
where there are abundant alluvial de-
posits. This is discussed more fully
in the section "Diversity Among Ripar-
ian Ecosystems."



Throughout history, man has alter-
ed, developed, and influenced the extent
and condition of riparian ecosystems,
and today only a portion of the original
floodland area is occupied by natural
vegetation. There has been no single
comprehensive inventory of riparian eco-
systems in the United States to deter-
mine the amount of land area originally
covered by riparian ecosystems and the
proportion of that area presently sup-
porting natural riparian communities.
Data needed to provide this information
with precision are generally unavail-
able, due primarily to the historical
lack of recognition for the distinct and
significant values of riparian ecosys-
tems. However, existing resource inven-
tories provide a rough indication of the
extent and distribution of riparian
plant communities. We have reviewed
documented information from numerous
Federal and State agencies and the
literature on:

1. the past and present extent (area
or length) of major riparian eco-
systems in the United States, and

2. the extent and nature of flood-
plain and stream alterations that
are responsible for losses of ri-
parian ecosystems in the United
States and the environmental qual-
ity of that which remains.

Overall, it appears that more than
70% of riparian ecosystems have been al-
tered, and natural riparian communities
now make up less than 2% of the land
area in the U.S.A. Although a compre-
hensive inventory may be required for
certain management purposes, there are
sufficient data to conclude that these

important ecosystems have not received
adequate protection.


Two approaches were used to provide
insight to the amount and distribution
of riparian ecosystems. An analysis of
inventories on areas that are naturally
prone to periodic flooding provided the
best information on the land area of
riparian ecosystems. We have also ex-
amined inventories on stream length as
an independent estimate of riparian eco-
system extent and status.

Inventories of Floodplain Area
Of an estimated 916 million hec-
tares of land in the entire U.S.A. (769
million without Alaska) (Frey 1979), ap-
proximately 6 to 9% is subject to
flooding. Estimates of the amount of
land subject to flooding vary from 49
million hectares (52 with Alaska) for
100 year floodplains (Maddock 1975), to
54 million hectares (without Alaska)
subject to floodwater and sediment
damage (USDA Conservation Needs Inven-
tory Committee 1971), to 71 million hec-
tares (without Alaska) of non-Federal
rural flood-prone land (USDA Soil Con-
servation Service 1978).

These values probably overestimate
the amount of riparian ecosystem once
present, because the estimated original
area of predominant riparian forest
types totals only 27 million hectares
(Table 1). Moreover, a portion of this
floodplain area can no longer be consid-
ered forested because of extensive al-
teration. For example, only 29% (15


Table 1. Potential and present area of the four predominant riparian vegeta-
tion types in the United States. From Klopatek et al. 1979.

Area (1000 ha) %
Vegetation typea Potential Present decline

Elm-ash forest 2,239 279 88
Northern floodplain forest 7,171 2,227 69
Southern floodplain forest 17,744 6,645 63
Mesquite bosque 71 63 11

S Total 27,225 9,214 66

I aAfter Kuchler (1964).

million hectares) of the Nation's flood-
plains were classified as nonurban and
Nonagricultural land (USDA Conservation
Needs Inventory Committee 1971). Simi-
larly, an estimated 30% (21 million hec-
tares) of non-Federal rural flood prone
Lands are forested (USDA Soil Conserva-
tion Service 1978). Several riparian
forest types have been cleared exten-
sively in the conterminous U.S.A. (Table
1), with losses ranging from as high as
88% for elm-ash forest to as low as 11%
for mesquite bosque (Klopatek et al.
1979). Thus, about 70% of the Nation's
floodplain area has been converted from
natural forest land to urban and culti-
vated agricultural areas.

I Surveys conducted for purposes
other than estimating riparian ecosystem
coverage suggest that these lowlands
constitute less than 30% of the total
floodplain area. Floodplain forest
types now account for about 9.3 million
hectares of the conterminous 48 States
(Table 1). According to a national wet-
land inventory in 1954 (Shaw and Fredine
1956), there were 9 million hectares of
seasonally flooded basins or flats, and
7 million hectares of wooded swamps,
both common forms of riparian wetlands.
However, these areas are not synonymous
I with riparian ecosystems because they
include considerable area of wetland
that is not riparian, and omit less fre-
quently flooded riparian communities.

Riparian ecosystem inventories at
the State level were summed to give a
minimum existing area of 23 million hec-
tares, of which 10.5 million are in the
lower 48 States (Table 2). In the lower
Mississippi Delta, an estimated 2.1 mil-
lion hectares of bottomland hardwoods
were remaining in 1978 (MacDonald et al.
1979a, 1979b). There are about 1.5 mil-
lion hectares of bottomland hardwoods
(12% of State area) in Mississippi
(Mississippi Game and Fish Commission
1978), including a substantial amount
outside of the Delta region. As of
1963, California had nearly 142,000
hectares of riparian vegetation (0.35%
of State area) remaining (California
Department of Fish and Game 1966). The
total riparian area in Arizona is
113,000 hectares (0.4% of State) (Bab-
cock 1968); the area in New Mexico may
be equal or slightly larger (Pase and
Layser 1977). Riparian communities on
Bureau of Land Management lands consti-
tute 287,495 hectares in western states,
5544 in the East, and 12,029,543 in
Alaska (USDI Bureau of Land Management
1980). With the exception of Alaska,
riparian ecosystems are clearly most
abundant in the southeastern states
where an estimated 8.5 million hectares,
or 70% of the total documented area,
were identified. Such data are general-
ly unavailable for northeastern and
northcentral U.S.A.

Table 2. Estimated

area of riparian ecosystems in 26 States, or portions

State Area (hectares) Source

Alaska 12,029,500a BLMb 1980
Arizona 113,153 Babcock 1968
Arkansas 410,765a MacDonald et al. 1979b
California 140,537 Calif. Dept. Fish & Game 1966
Colorado 24,441a BLM 1980
Idaho 22,909a BLM 1980
Kansas 207,406 Spencer 1979
Kentucky 13,760a MacDonald et al. 1979b
Louisiana 1,214,000a MacDonald et al. 1979b
Mississippi 1,457,000 Miss. Game & Fish Comm. 1978
Missouri 38,851a Korte and Fredrickson 1977
Montana 51,216a BLM 1980
Nebraska 115,824 Spencer 1979
Nevada 36,423a BLM 1980
New Mexico 113,314 Pase and Layser 1977
North Dakota 72,481 Spencer 1979
Oregon 71,135a BLM 1980
South Dakota 53,905a Spencer 1979
Southeast (Fla., Ga.
N.C., S.C., Va.) 6,300,000, Langdon et al. 1980
Tennessee 52,610a BLM 1980
Utah 28,934a BLM 1980
Wyoming 18,471a BLM 1980

Total 22,586,635
Total 10,557,135 (without Alaska)

aEstimates were only available for portions of the State and should be consi-
dered an underestimate.
bUSDI Bureau of Land Management. Values cited as BLM (1980) are for "public
land wildlife habitat" only and should be considered underestimates for the
respective States.

Certain specialized riparian com-
munities constitute a significant area
in some regions of the United States.
These areas are of particular interest
to resource managers because of specific
ecological or functional values associ-
ated with them. For example, there were
more than 360,000 hectares of saltcedar
vegetation in the arid western U.S.A. by
1961, and probably well over 400,000
hectares today (Robinson 1965). (This
exotic woody plant has replaced many
native floodplain plant communities

but has very different and limited value
to wildlife.) Beaver ponds occupy about
162,000 hectares of floodplain timber in
the southeastern U.S. (Hill 1976, in
Hair et al. 1978). In the Uinta basin
of Utah alone, 19,733 hectares of vege-
tation are dependent upon irrigation re-
turn flow (Chalk 1979).

Based on these data, it appears
that riparian ecosystems comprise be-
tween 10 and 15 million hectares in the

-d I


I States, or about 1.5% of the U.S.A.
land area. A more precise and compre-
Bnsive inventory may be required for
retain management purposes, but there
are sufficient data to conclude that
these important fish and wildlife habi-
Its are quite limited in extent in most
gions of the country. That riparian
ecosystems cover such a small proportion
Sthe landscape is due to their limited
tent originally (except in the South-
st), and to widespread floodplain al-
terations by man. Brief accounts of
me representative riparian ecosystem
sses are presented later.

Inventories of Streams and Rivers

I Analysis of stream length across
e country provides insight on the dis-
tribution and abundance of riparian eco-
lstems. Stream length generally re-
lects the potential abundance of ripar-
ian systems, and provides a common unit
r measuring the extent of floodplain
There are an estimated 5.1-5.6 mil-
ion kilometers of streams and rivers in
he U.S.A., ranging from the smallest
irst-order tributary to the largest
rivers (Leopold et al. 1964, U.S. Army
orps of Engineers 1978). However, only
bout 1.6 million kilometers were ac-
ounted for in a compilation of State
stream inventories (Table 3). The
Scatter figure may be more useful for
discussion of riparian management poten-
ial, because it represents the extent
of waterways recognized by respective
Itate water resource agencies.

Riparian ecosystems are most exten-
sive in humid and coastal plain regions,
Especially where perennial streams are
Relatively abundant and where warmwater
streams and rivers predominate (Fig-
re 1). Stream length per unit of land
rea (drainage density) is greatest in
ouisiana, high throughout the eastern
half of the U.S.A., but dramatically
power in the West. Similarly, the aver-
ge surface area of streams (USDA Soil
nservation Service 1978) relative to
length is considerably greater east of
the Mississippi River.
There are some 492,000 kilometers
of warmwater fishing streams in the
*.S.A. (Funk 1970). More than 573,000

kilometers of major stream segments
(greater than 40 kilometers long) have
been identified in the Nationwide Rivers
Inventory (U.S. Heritage Conservation
and Recreation Service, pers. comm.
1980) which potentially support exten-
sive riparian communities (Table 4).
Some of the most outstanding riparian
ecosystems in the country are associated
with 141 major rivers (by discharge)
totalling 116,000 kilometers in the
U.S.A. (USDI Geological Survey 1974; see
also Iseri and Langbein 1974). Included
in the above list are the Atchafalaya,
Brazos, Colorado, Columbia, Connecticut,
Gila, Mississippi, Missouri, Rio Grande,
Sacramento, Snake, and St. Lawrence

While coldwater and intermittent
streams are widespread and often support
riparian communities with significant
value to wildlife, their areal extent is
singularly quite limited. Consequently,
most available data on riparian areas
were derived from large river systems,
while the vast extent of small stream-
bank communities has never been quanti-


Historically, riparian ecosystems
have been altered or destroyed to a
largely unknown extent, without protec-
tion from long-term adverse impacts on
their ecological functioning. Causes of
riparian ecosystem degradation are num-
erous, and vary in importance from one
region to the next. Available case his-
tories are presented here to illustrate
the nature of riparian ecosystem losses
across the country.
Alterations of Floodplains

The areal extent of riparian eco-
systems has been reduced by a substan-
tial amount in nearly every region of
the U.S.A. Losses of bottomland vege-
tation have been most dramatic in the
Mississippi Delta, Midwest, and arid
western areas, caused by demand for
water and productive farmlands which
they normally can provide. It is evi-
dent that losses at some locations far
exceed the estimated national average of
70% (Table 5). Some examples are des-
cribed below.

I __ .

Table 3. Length of streams in the United States.a

Total stream Total stream
State length (kilometers)a State (kilometers)

Alabama 11,839 Montana 27,607
Alaska 1.6 million+ Nebraska 19,904
Arizona 1,287 Nevada 11,908
Arkansas 15,315 New Hampshire 20,241
California 46,959 New Jersey 4,184
Colorado 26,554b New Mexico 5,277
Connecticut n/a New York 106,851
Delaware 1,287 North Carolina 6,437
Florida 16,979 North Dakota n/a
Georgia 62,565 Ohio 70,678
Hawaii 2,364 Oklahoma 37,015
Idaho 25,296 Oregon 43,452
Illinois 21,250 Pennsylvania 40,057
Indiana 145,000 Rhode Island n/a
Iowa 30,600 South Carolina n/a
Kansas 16,100 South Dakota 5,544
Kentucky 64,400 Tennessee 30,417
Louisiana 64,887 Texas 128,748
Maine 44,893 Utah 9,864
Maryland 2,736 Vermont 10,461
Massachusetts 17,226 Virginia n/a
Michigan 44,819 Washington 25,608
Minnesota 40,100 West Virginia 36,194
Mississippi 22,700 Wisconsin 43,713
Missouri 91,068 Wyoming 24,853

Total (without Alaska) 1,525,227

aEstimates were obtained by personal communication with
agencies. Although definitions of streams differ from
estimates represent perennial streams that potentially
bDetails available from the authors.
Data not available.

Land use changes on the 9.8 million
hectare Mississippi Alluvial Plain
(mostly riparian) have been documented
(MacDonald et al. 1979a, 1979b). Bot-
tomland hardwood forests covered only
4.8 million hectares in 1937, and were
reduced to 2.1 million hectares by 1977.
Cumulative losses between 1957 and 1977
ranged from 30% to 63% among various
States (Table 6). The rate of clearing
has averaged around 2% per year over the
last 20 years. The majority of bottom-
land forest clearing in the Mississippi

State and Federal
State to State, most
support a fishery.

Delta results from conversion to agri-
culture, particularly for soybeans
(Sternitzke and Christopher 1970, Ster-
nitzke 1976, MacDonald et al. 1979a).

Area of bottomland hardwoods in
southeastern Missouri declined 96% from
an estimated 1.0 million hectares in
1780 to 40,000 hectares in 1975 (Table
5), primarily as a result of lumbering
and drainage for agriculture (Korte and
Fredrickson 1977). Between 1879 and
1972, the total water surface area of



Figure 1. Distribution of warmwater, coldwater, and intermittent streams in the United
States. Map used by permission of the American Fisheries Society.

Table 4. Total length of major stream segments in the
unsuitable for designation as National Wild
tional Rivers.

U.S.A. and percentage
and Scenic or Recrea-

Kilometers of river segments
Unsuitable for %
HCRSa region Total designation unsuitable

Northwest 42,129 7,500 18
Southwest 48,334 5,562 12
Mid-Continent 161,000 129,000 80
South Central 106,911 26,187 24
Lake Central 82,894 71,674 86
Northeast 40,234 31,704 79
Southeast 91,733 73,254 80
Total 573,235 344,881 60

aUSDI Heritage Conservation and Recreation Service.
Stream segments greater than 40 kilometers in length.
cBecause of water resource or other cultural developments in the river

Table 5. Summary of case histories showing losses of riparian ecosystems.

Estimated change %
Description Time period (hectares) loss Source

Bottomland hardwoods in 1700's-1977 9.8 million to 79 MacDonald
lower Mississippi Delta: 2.1 million et al. 1979a
Ark., La., Miss., 1937-1977 4.8 million to 51 MacDonald
Mo., Tenn., and Ky. 2.1 million et al. 1979a
Cottonwood communities along 1600's-1977 2000 to 1133 44 Ohmart
the Colorado River, Arizona et al. 1977

Riparian vegetation along 1935-1978 27,900 to 20,030 28 McNatt 1978
the San Pedro River, Arizona

Riparian forests along the 1850-1977 313,600 to 7,200 98 rlcGill 1975,
Sacramento River, California 1979

Bottomland hardwoods in 1780-1975 1.0 million 96 Korte and
southeastern Missouri to 40,000 Fredrickson
Channel habitats in 1879-1972 49,000 to 25,000 50 Funk and
Missouri River, Mo. Robinson 1974

Two riparian forests in 1871-1969 12,100 to 1,544 87 Barclay 1980
southcentral Oklahoma

the Missouri River (from Rulo, Missouri
to the mouth) was reduced by 50% from
49,000 hectares to 25,000 hectares.
Surface area of unconnected islands in
the Missouri River was 9900 hectares in
1879, and 170 hectares in 1954, a loss
of 98%. Elimination of channel communi-
ties in the Missouri River was the
direct result of stream channel altera-
tions (Funk and Robinson 1974).

In Oklahoma, 12,100 hectares of ri-
parian forest along two streams experi-
enced an 87% reduction in area between
1871 and 1969; about 81% was gone by
1937 (Table 5). These losses were
largely attributable to impacts of chan-
nelization (Barclay 1980).

Riparian vegetation along the
Colorado River is disappearing at a rate
of 1200 hectares per year (Anderson et

al. 1978).

Pure cottonwood communities

have declined from an estimated 2000
hectares to 200 hectares as a result of
altered hydrologic regimes, impound-
ments, and agriculture. There are still
some 1133 hectares of willow-cottonwood
stands along the river (Table 5), but
most are invaded by saltcedar, an exotic
tree species of much lower value to
wildlife (Ohmart et al. 1977).

Between 1935 and 1978, riparian
areas composed of cottonwood, mesquite,
saltcedar, and willow along the San
Pedro River in Arizona increased from
6900 hectares to 14,200 hectares (Cot-
tonwood and willow were actually declin-
ing as a result of eliminating perennial
streamflows.) During that same time,
other marsh, mesquite shrub, river chan-
nel, and streambed thickets of annual or
immature plants decreased from 20,600 to

Table 6. Area of bottomland hardwoods
1957 to 1977.

in the lower Mississippi Valley,

Bottomland hardwood area (1000 hectares) % loss
State 1957a 1967a 1977a 1957-1977

Arkansas 843 537 411 52

Kentucky 21 16 14 36

Louisiana 1,743 1,513 1,214 37

Mississippi 613 478 377 39

Missouri 76 43 28 63

Tennessee 84 66 53 38

Total 3,380 2,653 2,097 38

Net loss during b
previous decade -- 727 556

% loss during
previous decade 21.5 21.0

bFrom Tables A1.1-A1.18
From Tables A3.1-A3.18

in MacDonald et al. (1979b).
and A7.1-A7.6 in MacDonald et

al. (1979b).

5700 hectares. The net loss of riparian
vegetation was 7700 hectares (Table 5).
However, along a 35 kilometer stretch
of that river, riparian communities have
declined from 4300 hectares in 1936 to
2200 hectares in 1972, nearly a 50% re-
duction (Lacey et al. 1975). Stream
channel alteration, irrigation diver-
sion, groundwater pumping, and over-
grazing were all contributing factors to
the alteration or destruction of those
riparian communities (McNatt 1978).

There were nearly 313,600 hectares
of riparian forests along the Sacramento
River in the 1850's (Sands 1978). By
1952, about 11,000 hectares remained,
and in 1972, there were only 7600 hec-
tares (McGill 1975). Native riparian
vegetation was further reduced to 7200
hectares by 1977, or about 2% of the
original area (Table 5). Most recent

losses were the result of converting
high terrace forest land to deciduous
orchard (McGill 1979).

The U.S. Geological Survey mapped
3700 hectares of phreatophytes in a 74
kilometer reach of the upper Gila River
(Gatewood et al. 1950). When examined
in 1958, 16% had been cleared for farm
use (Horton 1972). Clearing continued,
and only 2670 hectares were reported in
1967, a 29% reduction in 23 years (Lacey
et al. 1975). About 45,000 hectares of
floodplain along the lower Gila River
was assumed to have been covered by ri-
parian vegetation in 1860. In 1970 only
6620 hectares (15%) of riparian vegeta-
tion were present, and more than
one-half was saltcedar communities.
When total acreage of this exotic was
subtracted, only 2350 hectares of native
riparian communities remained, or about

5% of the theoretical riparian base
present in 1860 (Haase 1972).

Riparian ecosystems have not been
cleared so extensively in some areas of
the country. For example, the acreage
of bottomland hardwood-cypress forests
in five southeastern states (Florida,
Georgia, North Carolina, South Carolina,
Virginia) remained fairly stable from
1940 to 1980 (Langdon et al. 1980).
Cottonwoods, which were scarce along the
lower South Platte River in the middle
19th century, increased greatly in num-
ber over the next 100 years and may have
peaked in the 1950's, after water re-
source developments reduced the "flashy"
flows to more moderate seasonal fluc-
tuations (Crouch 1979). Mountain ri-
parian areas have not changed as dis-
tinctly as lowland floodplain areas;
there has been some clearing and con-
struction of dams, but in general vege-
tation along mountain streams has been
maintained by near normal ecological
processes (Horton 1972).






200 400 00 o 10o00 1200 1400 10o

Figure 2. Extent of water resource de-
velopment on streams in the United
States. Sources: (1) estimated by
authors; (2) USDA Soil Conserv. Serv.
1980; (3) Little 1973; (4) Little 1973;
(5) pers. comm., USDI Heritage Conserv.
and Recreation Serv. personnel 1980; (6)
estimated by authors.

Alterations of Streams and Rivers

The total nationwide extent of ri-
parian community losses caused by stream
alterations has not been determined.
However, available data indicate water
resource development projects have re-
sulted in substantial disruption of
streamside ecosystems (Figure 2).

During the past century and a half,
mankind has been responsible for the
"development, improvement, or modifica-
tion of at least 320,000 kilometers of
waterways" (Little 1973). This consti-
tutes a direct impact on at least 20% of
the stream length recognized by the
various States, and would equal over
one-half of the total length of warm-
water streams where channel alterations
are most prevalent. 'However, actual
losses in surface area of riparian eco-
systems undoubtedly occur in larger pro-
portion than losses in stream length.
This is because large amounts of drain-
age and forest clearing usually accom-
pany relatively small reductions in
stream length.

Extent of recent channel alteration
activities by Federal agencies has been
documented (USDA Soil Conservation Ser-
vice 1971, 1975, 1980; Little 1973).
Between 1940 and 1971, the Corps of
Engineers assisted 889 stream develop-
ment projects covering a total of 17,827
kilometers of which 9946 kilometers were
completed, 6270 were under construction,
and 1611 kilometers were planned. As of
1972, SCS channel work in the U.S.A.
totalled about 33,800 stream kilometers,
of which 13,911 kilometers were con-
structed or under contract. By 1980, a
total of 17,344 kilometers of SCS chan-
nel alterations were constructed or
under contract, an increase of 483
kilometers per year (Table 7) (USDA Soil
Conservation Service 1980).

Among the 1630 projects administer-
ed by the Corps and SCS by 1971, 45,614
kilometers were channel alterations and
9490 kilometers involved floodplain al-
teration by levee work. About 47% was
to have been carried out in five States
(Louisiana, Mississippi, Arkansas,
California, and North Carolina) and an
additional 25% in five other States
(Texas, Florida, Georgia, Illinois,
Indiana) (Little 1973).

Table 7. Summary of Soil Conservation Service channel work through 1972
and 1980 (from USDA Soil Conservation Service 1972, 1980).

SCS channel work (kilometers)

31 December 1972 12 March 1980b
Region Constructed Constructed Planned

Northeast 2,313 2,686 1,310

Midwest 1,651 2,226 1,799

South 9,429 11,830 13,123

West 520 604 367

Total 13,913 17,346 16,599

a"Constructed" includes
"planned" includes all
bbut not constructed or
Includes 1972 figures.
CTotals were calculated

all channel work completed or under contract;
channel work planned and in an approved project
under contract.

prior to rounding off of regional figures.

Estimates of stream channel altera-
tions by SCS and Corps activities fall
far short of the total carried out by
all agencies and private interests
(Table 8). In Missouri, for example,
3584 (4%) of the total 91,068 stream
kilometers had been channelized, and an
additional 4699 kilometers (5%) were
inundated by impoundments at flood pool
elevation (Missouri Dept. of Conserva-
tion, pers. comm. 1980). In a survey of
351 stream kilometers in Kentucky, 144
kilometers (41%) had been recently al-
tered (Russell 1967). Approximately
one-third of the total length of streams
inventoried (402 of 1236 kilometers in
Montana had been altered from their
natural condition, of which half (222
kilometers) was by relocation, 103 kil-
ometers were rip-rapped, and 66 kil-
ometers were diked (Peters and Alvord
1964). Among 366 perennial streams in
Hawaii, 15% have been channelized,
totalling 151 kilometers and including
57% (31 of 54) on the populous island of
Oahu (Timbol and Maciolek 1978).

Reduction in stream length is a sig-
nificant but often unmeasured aspect of
channelization projects. Loss of stream
mileage from stream alterations may be
very high in some stream corridors. One
stretch of the Missouri River has been
shortened from 875 kilometers in 1870 to
801 kilometers in 1972, a loss of 74
kilometers (Funk and Robinson 1974).
Total length of 13 Montana streams and
rivers was shortened 109 kilometers (9%)
from the original 1236 kilometers by re-
routing of 220 kilometers of stream into
111 kilometers of man-made channel
(Peters and Alvord 1964). Data from
Iowa indicate that stream length across
the State has been reduced 1693 kilome-
ters and possibly as much as 4800 kil-
ometers (Bulkley et al. 1976). Other
examples of stream length reduction are
cited in Table 8.

Impacts of stream alteration clear-
ly extend far beyond the actual develop-
ment site; consequently data from the
Nationwide Rivers Inventory (U.S. Herit-

Table 8. Extent of stream alterations in twelve States.

State Extent of alteration Source

Hawai i










South Dakota

River Basin

Of 366 perennial streams in Hawaii, 15%
have been channelized, totalling 151 kilo-
meters (6% of the State total), and in-
cluding 57% on the populous island of Oahu.

In a survey of 1831 stream kilometers,
698 (38%) had been altered.

An estimated one-third of the State's
natural streams has been channelized.

Total stream length in the State has been
reduced at least 1693 kilometers and pos-
sibly as much as 4800 kilometers.

In a survey of 351 kilometers, 144 (41%)
had been recently altered.

About 3862 kilometers (17%) of the
streams in Mississippi have been altered.

Across the State, 3584 stream kilometers
(4%) have been channelized, and at
flood level an additional 4699 kilometers
(5%) are inundated by impoundments.

One stretch of the Missouri River was
shortened from 875 to 801 kilometers since
1870, a loss of 74 kilometers (8%).

Approximately one-third of stream length
studied (402 of 1236 kilometers) was
altered from the natural condition. Total
length had been reduced 109 kilometers
(9%) by channelization.

Total stream mileage has been reduced
1341 kilometers (6%) by channelization.

An estimated 34,236 kilometers of streams
(48% of the State total) have been

About 20% of the State stream mileage
is altered, including impoundment
of 80% (644 of 805 kilometers) of the
Missouri River.

Over 5600 kilometers (8%) of the total
67,600 stream kilometers are impounded at
normal full pool level. An additional 1770
kilometers (3%) have reservoir-regulated flows.

Timbol and
and Maciolek

Irizarry 1969

D. Rogers, Ill.
Dept. of Conserv.
(pers. comm.)

Bulkley et al.

Russell 1967

B. Freeman, Miss.
Game & Fish Comm.
(pers. comm.)

0. Fajen, Mo.
Dept. of Conserv.
(pers. comm.)

Funk and

Peters and
Alvord 1964

G. Zuerlein, Nebr.
Game & Parks Comm.
(pers. comm.)

A. Spencer, Ohio
Div. of Wildl.
(pers. comm.)

R. Hanten, S.D.
Dept. of Game,
Fish and Parks
(pers. comm.)

Tennessee Valley
Authority 1971

age Conservation and Recreation Service,
pers. comm. 1980) may provide a better
indication of stream condition across
the Nation. Among the 570,000 kilo-
meters of major stream segments, 60%
were judged unsuitable for inclusion in
the National Wild and Scenic Rivers
System because of water resource or
other cultural developments within ri-
parian corridors (Table 4).

Because reservoirs are situated in
floodplains and riparian zones, con-
struction of impoundments has resulted
in significant losses of riparian eco-
systems and their values to wildlife.
The total length of streams inundated by
reservoirs has not been determined, but
probably exceeds 24,000 kilometers. By
January 1, 1980, there were 1608 reser-
voirs with a mean annual pool of 202
hectares or more. This increased the
area of these reservoirs by 409,550
hectares since 1970 to a total of
3,989,000 hectares (Ploskey and Jenkins
1980). If an arbitrary 4:1 ratio of
length to width and triangular shape
were assumed for these reservoirs, they
would extend over an estimated 22,000
kilometers of streams. Among the 1562
reservoirs having a storage capacity of
617 hectare-meters or more, 6,002,000
hectares would be covered at maximum
controllable water level (Martin and
Hanson 1966) and would flood over 27,400
kilometers of stream. In the Tennessee
River Basin, an estimated 5734 kilome-
ters (8%) of the total 67,500 stream
kilometers are impounded at normal full
pool level, and 1814 kilometers (3%)
have reservoir-regulated flows (Tennes-
see Valley Authority 1971). The cre-
ation of Lake Oahe on the upper Missouri
River in South Dakota inundated 90,650
kilometers of land, including all areas
along a 320 kilometer reach of river
(Hirsch and Segelquist 1978).

Prosser et al. (1979) state that
the "loss of terrestrial habitat from
reservoir construction constitutes only
0.6% of all undeveloped lands capable of
supporting wildlife." However, direct
loss of length in major streams is pro-
bably at least 5% nationwide, while ex-
tent of downstream impacts cannot be
estimated. Further, the land area inun-
dated by large reservoirs alone is equal
to 8% of the total 100-year floodplain
area, a value which does not include

the extent of undocumented loss due to
smaller reservoirs.


Existing inventories of floodplain
area and stream length cannot be used
alone or without interpretation to pro-
vide an overview of the status of ripar-
ian ecosystems in the U.S.A. However,
when taken together, these data give a
great deal of insight to the nationwide
amount of riparian ecosystem that was
originally present, the quantity lost to
other uses, and the nature of altera-
tions. From the foregoing data, we
derived some rudimentary estimates of
the status of riparian ecosystems in the
U.S.A. (Figure 3).

One estimate of the amount of land
subjected to riverine flooding (100-year
floodplain) and thus potentially sup-
porting riparian ecosystems is 49 mil-
lion hectares, or 6% of the U.S.A. land
area (excluding Alaska). This figure
may be considered liberal, because the
estimated original area of four pre-
dominant riparian forest types totals
only 27 million hectares (Table 1). Re-
gardless, much less exists in a natural
or seminatural forested condition, and
streamside riparian communities now
constitute only about one-third of the
original area. The extent of bottomland
alterations is known to be much greater
in Arizona, California, and Missouri,
and for certain floodplain forest types.
Because at least 10.5 million hectares
of riparian communities can be accounted
for from State surveys (Table 2), the
nationwide total is probably between 10
and 15 million hectares, or about 1.5%
of the conterminous U.S.A. land area.

The great difference between poten-
tial riparian land area and that now in
a woodland condition reflects the extent
of alteration that has occurred, and
some discrepancies in defining and de-
lineating riparian ecosystem boundaries.
Many of our riparian lands have been
directly destroyed or converted to urban
or agricultural uses that are usually
incompatible with natural ecological
functions (Chapter 3) and wildlife re-
sources (Chapter 4). These alterations
can be considered "acute" because they
severely preclude most other goods and

services to society that riparian eco-
systems provide (Chapter 5). As com-
pared to all other vegetation types in
the U.S.A. (Kuchler 1964), conversion of
floodplain forests to other land uses
puts riparian ecosystems among the most
severely altered landforms in the

In addition to these losses of ri-
parian communities that can be quanti-
fied, stream alterations, pollution,
grazing, and recreation can also reduce
the functional quality of remaining
areas through more subtle "chronic"
impacts. In the northcentral and north-
eastern states, up to 80% of major
stream corridors are interrupted by
water resource or cultural developments.
In the South and West, existing riparian
communities are disturbed by manipula-
tion of streamflows and overflows, and
subjected to problems associated with
consumptive uses of water and grazing.
Numerical estimates of riparian ecosys-
tem area fail to measure these less in-
tensive disturbances.

The significance of riparian eco-
system alterations, whether acute or
chronic, lies in the relative irrever-

sibility of man's impacts. Although
agricultural and water resource develop-
ments can theoretically be reversed, the
economic expense and incentives for
doing so in floodplains are currently
very prohibitive. Most importantly,
reclamation of riparian ecosystems re-
quires restoration of complex natural
hydrologic regimes. However, because
conversion of flood-prone areas to other
uses usually involves permanent drainage
or impoundment, opportunities for miti-
gation and recovery by natural succes-
sion are practically nonexistent.

Despite the outstanding ecological
values of natural riparian ecosystems,
natural plant communities on these lands
have been reduced in extent by 70% over-
all, and as much as 95% in some areas.
The functioning of remaining areas is
threatened by further direct losses and
impacts of man's activities in adjacent
aquatic and upland ecosystems. The
effect of these riparian ecosystem
losses to the well-being of society,
through the degradation of ecological
function, wildlife resources, and pro-
duction of goods and services, will be
apparent in following chapters.

Figure 3. Land area covered by riparian vegetation, streams and reservoirs in the
United States. Sources: (1) Frey 1979; (2) Maddock 1975; (3) Klopatek et al. 1979;
(4) USDA Soil Conserv. Serv. 1978; (5) Ploskey and Jenkins 1980.



All ecosystems have common proper-
ties of energy flow, material cycling,
and community organization; yet no two
ecosystems are organized and function in
exactly the same way. However, riparian
ecosystems have several unifying proper-
ties that set them apart from other eco-
system types.

One of these properties is their
linear form, a consequence of being as-
sociated with streams. As a result, the
abundance of riparian ecosystems depends
on drainage density of streams (kilo-
meters of stream length per square
kilometer of land area) which, in the
northeastern U.S.A. for example, ranges
between 1 and 2.5 km/km (Leopold et al.
1964). Thus, there are few places in
that region that are very distant from a
riparian ecosystem.

Another related property is the
function that riparian ecosystems serve
in providing 'corridors for the transport
of water and erodible material derived
from the landscape. In comparison with
upland ecosystems, riparian areas tend
to be wetter, to have more nutrients
available to them, and to be more fre-
quently subjected to catastrophic water
flow. The convergence of energy and
material from the landscape on riparian
ecosystems is expressed in their nutri-
ent-rich soils and lush growths of vege-

Finally, the property of linearity
and the function as corridors of mate-
rial transport combine to assure that
riparian ecosystems are profoundly con-
nected to other ecosystems upstream and
downstream from them. Few other eco-
system types possess such a large amount
of transition zone relative to the area

that they occupy. These transition
zones are the boundaries at which ter-
restrial and aquatic ecosystems inter-
face and the sites of important ex-
changes of material and energy in the

In spite of these apparent differ-
ences between riparian and upland eco-
systems, it is difficult to find quanti-
tative data on basic ecological charac-
teristics (energy flow, nutrient cy-
cling, community structure) that clearly
distinguish these ecosystem types from
one another. A major problem is that
riparian ecosystems vary greatly among
geographic regions, as do upland ecosys-
tems. One of the purposes of this
chapter is to determine the extent to
which data on riparian ecosystem struc-
ture and function allow us to character-
ize them as unique ecological entities.
Recognition of any unifying characteris-
tics of riparian ecosystems may be use-
ful in assessing the effects of their
alteration and in providing guidelines
to their management.


Plant and animal communities are
sensitive to the edaphic conditions
under which they develop. In riparian
ecosystems, soil moisture is an extreme-
ly important variable because small
topographic variations in a seemingly
level floodplain can mean the difference
between a waterlogged, anaerobic envir-
onment and a well drained, aerated sub-
strate. Many plant species are intol-
erant of even brief periods of inunda-
tion while fewer species are adapted to
survive in constantly waterlogged soil.

As a result, abrupt changes in species
composition may occur in floodplains
with elevational variations of only a
few centimeters.

Natural fluvial processes are re-
sponsible for many of the diverse, often
subtle, topographic features of flood-
plains. An understanding of fluvial pro-
cesses responsible for forming riparian
ecosystems is necessary in order to pre-
dict consequences of alteration or mani-
pulation of the natural system. Altera-
tion of fluvial processes is likely to
create a new set of floodplain features
to which plant and animal communities
must adapt.
Human activities in riparian eco-
systems are frequently oriented toward
stabilizing, rather than maintaining the
dynamic nature of fluvial processes.
The many approaches to stabilizing
stream channels and controlling water
flow are but a few examples of efforts
to counteract dynamic fluvial processes.
However, stabilization processes such as
these have, in many instances, decreased
rather than increased fundamental eco-
system properties such as species diver-
sity and processes such as rates of pri-
mary productivity, nutrient cycling, and
animal production. Fluvial processes
are necessary for the formation and con-
tinued maintenance of riparian ecosys-
tems; therefore, we begin with an over-
view of these processes before examining
the more purely biological properties.


Alluvial portions of valleys where
riverine forests normally occur may be
undergoing aggradation, degradation, or
be in a steady state condition. In the
steady state condition, where the supply
of alluvium from upstream erosion is
balanced by the transport of alluvium
downstream, floodplain features do not
necessarily remain static. In fact,
morphologic features of floodplains con-
tinually change as river channels
meander laterally and in a downstream

Aggradation and Degradation. Under
non-steady state conditions, an alluvial
valley and its stream may aggrade or de-
grade. Over time, these trends of ag-
gradation and degradation may alternate,

resulting in complex stratigraphic se-
quences. Leopold et al. (1964) illus-
trated the hypothetical development of
terraces by means of two sequences of
events that lead to the same surface
geometry (Figure 4). A large-scale ex-
ample of these processes has been des-
cribed for the Mississippi alluvial val-
ley, but the sequence also can occur in
smaller streams.

The Mississippi alluvial valley has
undergone at least five alternating
periods of valley cutting and alluvial
deposition that correspond with glacial
advance and retreat during the Quater-
nary period (Fisk 1944, 1952; Fisk and
McFarland 1955). Glacial advance and
accumulation of water in continental ice
masses resulted in a lowering of the sea
level by several hundred feet. In an
effort to adjust to this lowered base
level, erosion of an extensive valley
system occurred across the Gulf Coastal
Plain. As ice sheets retreated, sea
level rose, and the entrenched valley
system became alluviated during the
interglacial stages. Coarse material was
introduced first from steep tributaries
which built alluvial cones of gravel and
sands. When these materials reached the
Mississippi, they were transported sea-
ward and deposited over wide areas by a
braided river system as aggradation oc-
curred. As the basal portion of the al-
luvium thickened, sediments became finer
because stream gradients were reduced
and did not have the competence to
transport coarse sediments. As sea
level stabilized, the braided channel
was replaced by a single meandering one
through a combination of diminishing
load, smaller particle size, and deeper
scouring action. As a result, the
Mississippi River is now in an overall
balance between aggradation and degra-

Smaller streams have been shown to
undergo similar but less dramatic phases
of downcutting and alluvial filling
(Hadley 1960). Factors which cause
these shifts can be the result of one or
more of the following processes: geo-
logic uplift, change of base level
(usually sea level), or change in cli-
mate. Particularly for smaller flood-
plains, colluvium, or material trans-
ported from valley sides, can be a
source of material for floodplain

Figure 4. Two sequences of events lead
metry in terraces and floodplains. Onl
pold et al. (1964).

deposits. In narrow portions of flood-
plains this material may predominate as
the substrate for floodplain forests.
For example, approximately one-fifth of
the cross sectional area of the alluvium
of Beaverdam Run, Pennsylvania consists
of colluvium (Lattman 1960). The re-
mainder consists of channel fill, lag
deposits (boulders), lateral accretion,
and vertical accretion (including peaty

ing to the development of the same surface geo-
y example D is confined by bedrock. From Leo-

During channel overflow, there is
an opportunity for vertical accretion of
the floodplain through the deposition of
suspended sediment transported from up-
stream. (Flooding from local precipi-
tation does not result in floodplain
accretion.) This deposition is, of
course, a feature which contributes to
the high fertility of floodplain soils.
These deposits represent augmentation of
nutrient capital in those areas of the

floodplain where they occur. The amount
of overbank deposition is proportional
to the hydroperiod (duration and depth
of flooding) and the amount of sus-
pended-sediment load. While sus-
pended-sediment load varies in propor-
tion to the erodibility of the water-
shed, hydroperiod depends on local
floodplain topography combined with
flood frequency of the stream. The
recurrence intervals for bankfull flows
for 19 streams in the United States sum-
marized by Wolman and Leopold (1957)
range from 1.07 to 4.0 years. In the
bottomland forests of the White River
basin in Arkansas, sites where annual
flooding occurs may remain flooded as
much as 40% of the year (Bedinger

Rates of deposition differ greatly
among floodplains and within a given
floodplain. Observations on the rate of
vertical accretion in floodplains range
from a few millimeters per year to over
a meter during a single flood episode
(Table 9). It cannot be determined from
these values whether or not the standing
stock of alluvium is increasing or de-
creasing because few studies report
rates at which floodplain erosion
occurs. The floodplains of Beaverdam
Run, which changed to an aggrading
regime perhaps 200 years ago due to de-
forestation of the area, consists of
vertical accretion in at least the upper
2 m (Lattnan 1960). The floodplain of
the Cimmaron River in southwestern
Kansas has been undergoing vertical ac-
cretion at the rate of 2.1 cm/year since
a major flood destroyed the pre-existing
floodplain features and replaced them
with a valley-wide braided channel
(Schumm and Lichty 1963).

Sudden climatic and man-induced
changes in discharge and sediment load
can reverse trends in aggradation and
degradation of stream channels. These
altered trends, in turn, can be extra-
polated to changes that will occur in
floodplain hydrology and geomorphology.
Lane (1955) proposed the simple and use-
ful relationship

QS a QDso,
in which Q is water discharge, S is the
slope of the channel bed, Q is the
bed-material discharge, and 6s5 is a

measure of the size of the channel bed
material. For example, if a dam is con-
structed on a stream, bed material is
trapped behind the dam and clear water
is discharged downstream. This
decreases Qs on the right-hand side of
the equation which would require a re-
duction in S on the left-hand side,
assuming Q and Dso remain constant.
Consequently, downstream from the dam,
channel-bed slope (S) would decrease, a
phenomenon which is brought about by de-
gradation or net erosion of the stream
channel. This implies an increase in
channel capacity and a lower stage
height for equivalent discharge volume.
Thus, floodplain inundation would occur
with less frequency and involve less
floodplain area, resulting in dryer con-
ditions in the riparian ecosystem. Even
if protective measures were taken to
reduce the rate of channel degradation,
a reduced sediment supply from upstream
and regulated flow below the dam would
result in altered floodplain conditions.
Several other applications of Lane's
equation (Simons et al. 1975) provide
examples of nan-induced stream changes
from which floodplain alterations are

River Meanders and Topographic
Features. Riverine forests grow on a
number of topographic features that are
generally the result of aggradation,
degradation, and meandering of the river
channel itself (Allen 1965). Some typi-
cal floodplain features that are ap-
parent in a section of the Mississippi
River, Louisiana (Figure 5) include:

1. Natural levees adjacent to the
channel which contain coarser
material deposited during
flood overflow.

2. Meander scrolls located on the
inside curve of bends. These
rises and depressions, which
are the result of point bar
deposits, formed as the chan-
nel migrated laterally and

3. Backswamp deposits and sloughs
where finer sediments are de-
posited in meander scroll
depressions or in slack water
along the valley wall.

Table 9. Deposition rates in forested floodplains.

River and locality Deposition Event or period Source

Missouri R., N.D.
Near Bismarck

Lowlands between
Bismarck and Mandan

Cimmarron R.,
SW Kans.

Cache River, 11.

Upper Mississippi R.

Kankakee R., Ill.

Ohio R., Ohio

Connecticut R.

Kansas R.

Rio Grande, N.M.

Alexandra R., Alberta

8 10 cm 1952; largest flood
on record for river

180 cm

5.1 cm/yr Ca. 12 years of
record using tree
age since a des-
tructive flood
0.8 cm/yr Of annual total,
0.06 cm from flood
of 1.13 yr. recurrence

1.7 cm/yr Annual deposition in
backwater lake on

590121 g/m2

0.24 cm

3.47 cm
2.23 cm


Total sedimentation
during spring flood
of which 80% was

Mean deposition during
100 yr. flood, Jan. -
Feb. 1937

March 1936
Sept 1935

July 1951

1.5 cm/yr Mean aggradation
for 16-yr period
between Albuquerque
and Socorro

0.3 cm/yr Fed by glacial melt-
water; average aggra-
dation during past
2,500 yrs.

Johnson et
al. 1976

Schumm and
Lichty 1963

Mitsch et al.

Eckblad et
al. 1977

Mitsch et
al. 1979b

1939, in
Wolman &
Leopold 1957

Jahns 1947,
in Wolman &
Leopold 1957

Carlson &
Runnels 1952,
in Wolman &
Leopold 1957
Thompson 1955

Smith 1976

MacKenzie R., N.W.T.

1.3-1.9 cm

Mean for sand deposition
along point bar for 2
mo. during each of two

Gill 1972a

Figure 5. Typical floodplain topo-
graphic features, illustrated diagram-
matically, of the Mississippi River near
False River, Louisiana. Adapted from
Fisk (1952).

4. Oxbows or oxbow lakes which
are relict meander bends that
have been cut off.

5. Point bars on the inside curve
of river bends where deposi-
tion is rapid.

Streams migrate back and forth
across floodplains and move in a down-
slope direction; consequently all areas
in a floodplain, with the exception of
those formed by colluvial deposits, have
been traversed at one time by the stream
channel. If the rate of meander move-
ment occurs on a time scale similar to
that of ecosystem succession, younger
communities will be encountered on the
inside meander curve (Leopold et al.
1964). Wolman and Leopold (1957) report
rates of channel migration ranging from
10 ft (3 m) to over 2000 ft (610 m) per
year for rivers with drainage basins
greater than 100,000 mi2 (259,000 km2).

It should be possible in some cir-
cumstances to calculate the rate of
lateral channel movement from the gra-
dient of tree age in a transect perpen-
dicular to the inside of a meander curve
(Everitt 1968). On the basis of succes-
sional development in a section of the
Missouri River, it has been demonstrated
that the youngest communities correspond

to the center of the floodplain, while
the oldest ones are located at the edge
(Table 10). Although any area in the
floodplain may be potentially eroded by
river meanders, Johnson et al. (1976)
showed that the center of the Missouri
River floodplain, or the "meander belt",
is eroded more frequently. Rivers in
the southeastern Atlantic States appear
to be migrating southward as indicated
by their proximity to bluffs on the
south side and by the presence of broad
floodplains on the north side. In ex-
tremely broad floodplains, such as the
lower M*ississippi River, large areas of
the floodplain have not been occupied by
the river channel for thousands of years
(Gagliano and van Beek 1975).

Thermo-erosional processes are par-
ticularly significant in bank erosion
and meander rates in regions of perma-
frost. Outhet (1974) has classified
bank types in the Mackenzie River delta
according to their shape and erosional
rates (Figure 6). River channels in
permafrost environments erode the bank
on the outside of meanders as elsewhere;
however, the development of thermo-ero-
sional niches (bank undercutting) and
the presence of structural weaknesses
(ice wedges and other forms of ground
ice) result in large-scale sloughing to
a somewhat greater extent than occurs in
temperate environments. Although near-
shore stream current and thermal ex-
change are usually responsible for niche
development, erosion by wave action may
be significant where a long open-water
fetch is possible on wide rivers. Con-
tinuous removal by high current veloc-
ities all summer is why type 1 banks
have higher rates of erosion than other
types (Figure 6). Type 2 banks are a
result of intermittent removal of
material caused by variations in channel
discharge or variation in wind velocity
or direction. Type 3 banks are a result
of soil flow where ice-rich bank faces
retreat continuously through the summer.
Destruction of cut bank levees is accom-
panied by deposition along their back-
slopes; hence, the levee form is main-
tained without its total destruction
(Gill 1972b). Only where thermo-ero-
sional niches are active (type 1) does
the floodplain become destroyed and
undergo degradation without compensating

Table 10. Dependence of relative stand age on location in a floodplain.
Values are percent of stands measured in each age category and
floodplain location. After Johnson et al. (1976).

Percent of stands measured in each age
category and floodplain location
Relative stand Edge of
age class Meander belt Intermediate floodplain

Young 64 36 0
Medium 18 73 9
Old 8 25 67

Thus fluvial processes have at some
time been responsible for shaping nearly
all floodplain features. These process-
es produce topographically diverse and
spatially heterogeneous conditions that
result in a mosaic of diverse habitats
for plant and animal communities.

Hydrology and Hydroperiod

Riparian ecosystems vary consider-
ably from stream to stream and even in


Sbeas a aStramd -,


-- o,,,) m/yr -5 mS/yr 0 5 re/1y
0 510 1I

Figure 6. Rates of erosion and diagram-
matic side views of stream banks in per-
mafrost environments. Modified from
Outhet (1974).

sectors along a single stream. However,
differences in hydrologic properties are
mainly those of magnitude since all ri-
parian ecosystems are influenced by
flooding, possess topographic features
of fluvial origin, and are dominated to
various degrees by the streams that flow
through them. Surface water hydrology
is the most visible feature of flood-
plain hydrology, but it cannot be fully
understood without considering its
interaction with groundwater.

Surface-Water. The flooding regime
of riparian ecosystems may differ in
depth, frequency, duration, and time of
the year. Some of the factors that may
influence the depth of flooding (defined
as the difference in stage of a stream
at median discharge and a given flood
recurrence interval) include climate,
topography, channel slope, soils, and
geology (Coble 1979). If all these fac-
tors remain constant, then the depth of
flooding depends largely on size of the
drainage basin and storage capacity of
the floodplain surface. Topographic
features of floodplains may also impound
water and cause flooding as a result of
local precipitation independent of
stream discharge. This flooding is par-
ticularly common in oxbows, depressions
between parallel levees, and in back
swamp depressions where drainage pat-
terns to the stream channel are poorly
developed (Figure 5). More commonly,
where floodplains slope gently from the

river channel to uplands, both flooding
frequency and depth from overbank flow
are inversely proportional to floodplain
elevation. Typically, annual floods
occupy a greater area of floodplain than
do less frequent floods. Successively
higher levels of the floodplain occupy
less of the total area.

Duration of flooding is directly
related to the drainage area of the
stream basin upstream from the site in
question. For floodplain areas with
annual flooding on the Ouachita and
White River basins in Arkansas, flood
duration ranges from 10 to 18% of the
year for sites having drainage areas
from 13,000 to 18,000 km and from 5
to 7% of the year for sites having
drainage areas less than 780 km
(Bedinger 1979). This is a consequence
of broader storm hydrograph peaks for
streams with large drainage basins than
those with smaller ones. Sites on
streams having drainage areas of several
10's of thousands of square kilometers
in these river basins typically flood
for as much as 40% of the year. In doing
so they hold flooding waters from the
trunk stream which serves to ameliorate
downstream flooding.

An example of where many factors
that regulate flooding come into play is
the gradient beginning in the eastern
slope of the Appalachian Mountains, con-
tinuing through the Piedmont province,
and terminating along the south Atlantic
seacoast. Mountainous headwater streams
are characterized by small watersheds,
steep slopes, and constricted V-shaped
valleys. The typically shallow soils
have limited storage capacity for water.
Orographic rains result in greater pre-
cipitation than occurs at lower alti-
tudes. Consequently hydrographic peaks
are sharp and frequent, particularly
toward the end of the winter season and
into the spring when evapotranspiration
is low and soil water storage reaches
annual highs. In the rolling topography
of the Piedmont, flood peaks are the
highest among the three physiographic
provinces (Coble 1979) and tend to occur
when frontal weather systems stabilize
over the region and provide abundant
precipitation. Flash floods are less
likely than in the mountains partly be-
cause of larger watershed size and
greater storage capacity of the deeply

weathered soils. Coastal plain rivers
that have their origin in the Piedmont
and mountains tend toward a prolonged
winter hydrographic pulse as a result of
integrating the upstream peaks. Flood-
plains in low elevations of broad al-
luvial valleys may remain flooded for
months at a time.
Generalizations on surface water
hydrology can seldom be made for large
geographic regions. For example, the
physiographic and climatic diversity of
Alaska results in a variety of flooding
regimes. The largest floods occur along
the Pacific Ocean, where the Pacific
Mountains System forms a barrier to
moist air from the ocean, resulting in
high precipitation and rapid runoff in
the fall and winter from the rugged
slopes (Childers 1970). North of this
mountain system precipitation is less,
flood discharge rates are much lower,
and floods are confined to spring and
summer. In interior Alaska and the
north slope drainage, extensive freezing
and rapid warming in the spring may
cause spectacular spring breakup floods
when snowmelt flows into ice-jammed

Where glaciers flow across the
mouths of valleys, water flow may become
blocked and form a lake (Post and Mayo
1971). Catastrophic floods may occur
when glacier dams fail. These events
are especially prevalent in the Pacific
Mountain System of Alaska where outburst
flooding from glacier-dammed lakes may
be annual, once each 2 to 4 years, or
only after several years. Wide flood-
plains may be inundated to unusual
depths, and rapid erosion, deposition,
and stream channel changes may occur.

In the annual cycle of interior and
north Alaskan rivers, five hydrologic
periods can be recognized (MacDonald and
Lewis 1973). The longest period is when
the river is frozen beginning as early
as October and lasting into May. During
this prolonged period the availability
of unfrozen water under ice is criti-
cally important to aquatic invertebrates
and fish and also to several species of
mammals and birds (Wilson et al. 1977).
Rising temperatures in May melt snow and
flow is initially on top of the winter
ice cover during the pre-breakup phase.
The breakup phase may last only several

I ys and may be accompanied by ice jam-
ming, depending on local conditions such
river level when freezing initially
curred and whether the stage rises
efficiently to cause ice to float free-
ly downstream. A post-breakup flood
formally coincides with peak snowmelt.
e summer flow phase may be established
y mid-June when the general trend is of
decreasing discharge except for occa-
ional summer storms that may cause
pid rises in river stage.
Due to the variety of factors that
ntrol flooding regimes, surface water
drology in riparian ecosystems is
highly site specific. To understand the
hydrology of a given area of floodplain,
lth the hydrologic characteristics of
Ie watershed and local groundwater
hydrology must be taken into considera-

iinGround Water. Ground water in the
alluvial aquifer is in intimate connec-
ion with surface water in streams and
oodplain depressions (e.g., oxbow
wakes). The normal gradient and direc-
tion of ground water movement is toward
these surface water features through
ound water discharge. During periods
I high river stages the gradient is
reversed and water moves from the stream
Sthe aquifer. The extent to which the
luvial aquifer is an important area
or discharge and recharge of ground
water depends upon its size. Two ex-
emes were illustrated in Figure 4. In
ample D the floodplain is narrow and
alluvium mostly lacking; under these
nditions the floodplain will have
ttle groundwater storage and a small
fluvial aquifer. In example A, the
groundwater storage of the alluvium is
tentially large and may greatly influ-
ce the surface water hydrology either
serving as a source of water for the
channel at low river stage or as a recip-
nt of water from the channel at high
river stage.

For the lower Missouri River flood-
Elain, Grannemann and Sharp (1979) have
Ehown that the river itself has the most
Important influence on groundwater
levels. During sustained high river
Ages, which normally occur between
spring and early autumn, inflow of lat-
eral seepage keeps groundwater levels
high. The hydraulic gradient is re-

versed as river stage falls from late
autumn through the winter when flood-
plain groundwater supplies base flow to
the river. Grannemann and Sharp (1979)
discuss several other factors that con-
trol groundwater flows and levels in the
floodplain. These include:
1. Distance from the river channel.
Equalization of differences in
water head change more slowly far-
ther from the river than close to
i t.
2. Time elapsed since the river has
risen or fallen. Provided the
river stage does not overtop the
levee system, a sustained flood
peak will contribute more water to
the groundwater system than a
higher flood of shorter duration.

3. Geometry of the river meanders and
valley walls. Where an area of
floodplain is partially encircled
by a sharp river meander or where
floodplain segments are narrow due
to proximity of stream channel and
valley wall, river stage and
groundwater levels will respond to
each other more quickly.

4. Variations in the composition of
alluvium. Thick clay strata and
clay plugs will create a longer
time lag than sand or silt in
groundwater head response to river
stage changes due to the lower
transmissivity of clay sediments.

5. Tributary creeks flowing into the
floodplain. These may cause per-
manent groundwater highs and pro-
mote downvalley flow where they
are oriented parallel to the major
Water table fluctuations in the
floodplain of the upper Sangamon River,
Illinois, are strongly controlled by the
water level in the stream channel (Bell
and Johnson 1974). At middle elevations
between the stream and uplands, ground-
water loss to evapotranspiration during
certain summer periods may exceed the
combined sources of water by infiltra-
tion of groundwater from the river and
drainage from higher elevations. Thus,
even in the absence of overbank flood-
ing, groundwater levels in floodplains

may fluctuate in response to other fac-

Attempts to quantitatively deter-
mine water budgets from inflow/outflow
measurements are restricted to streams
in arid regions where floodplain or bot-
tomland groundwater deposits are sub-
jected to competitive demand by phreato-
phyte vegetation and by withdrawals for
consumptive human use and irrigation.
Figure 7 is a generalized model for a
water budget of the alluvial fill of a
floodplain. Results of a study for the
Gila River floodplain (Gatewood et al.
1950) are superimposed on this figure to
show the magnitude of water movement.
The predominant flows of water for the
various reaches studied were inflows
from upstream and downstream outflows.
Among the total outflow from'the lower-
most reach, only 2.5% was due to evapo-
ration from the river surface and wet
sand bars and 12.3% to evapotranspira-
tion by the bottomland vegetation.
While the value for evapotranspiration
may be an overestimate according to more
recent studies (van Hylckama 1980; R. M.
Turner, pers. comm.), the magnitudes of
flow suggest that groundwater storage
and flow is extremely important to the
maintenance of surface flows. Greatest
groundwater use by evapotranspiration
occurred during the warm months when
flows through the stream sector were
lowest. During the early winter months
groundwater recharge coincided with in-
creasing throughflows.

Significance of Fluvial Processes

The kinetic energy of flowing water
and its capacity to erode, transport,
and deposit materials are responsible
for the origin and necessary for the
maintenance of riparian ecosystems.
Fluvial processes are essential for pro-
ducing and maintaining topographic fea-
tures. If stabilization of water flows
and stream banks interferes with natural
fluvial processes, much environmental
diversity normally present will disap-
pear. Floodplains should be considered
the part of the stream channel that is
utilized to accommodate high flows.
Flooding opens up the riparian ecosystem
to inflows of material from upstream
that would not be available if flooding
were controlled.

The water storage capacity of allu-
vial deposits is particularly critical
for maintaining riparian vegetation dur-
ing the warm season in arid climates
when upstream supplies of water are low.
Where base flow of streams is dependent
on groundwater storage, it may be advan-
tageous to maximize groundwater recharge
through overbank flooding. Plant and
animal communities are adapted to or
even dependent on these pulses of flow
because they evolved under the natural
conditions of flooding.


Energy flow is often regarded as an
indicator of the vitality of an ecosys-
tem. It does not necessarily follow
that ecosystems with high primary pro-
ductivity are inherently more valuable
or in better condition than those with
lower productivity. For example, a
northern bog swamp would undoubtedly
have lower primary productivity than a
southeastern river swamp. However, both
are responsible for contributions to
productivity of the regional landscape
and must be evaluated in the context of
their location. For a given ecosystem,
primary productivity will vary widely
depending on weather conditions, time of
year, water availability, and other en-
vironmental variables. However, indi-
cators of primary productivity such as
litterfall and biomass accumulation pro-
vide insight to the magnitude of energy
flow so that ecosystems can be compared
and factors that control the energy flow
can be evaluated and identified. Envir-
onmental manipulations that either
severely diminish or abruptly augment
energy flow, particularly if the change
is irreversible, may be considered
disruptive to plant and animal communi-
ties as well as other goods and services
derived from ecosystems.

One of the fundamental functions of
primary productivity, in addition to
providing energy flow to food webs, is
that of maintaining the structure and
integrity of ecosystems. During eco-
logical succession in forested ecosys-
tems, large amounts of energy flow
initially are diverted toward the ac-
cumulation of new plant and animal bio-
mass and the formation of more complex



Upstream inflow


Aquifer in valley

- outflows


Recharge to deep
S aquifers

A Recharge of groundwater at high river stage C Pumping from wells for irrigation and consumptive use
B Seepage to river at low river stage Net underground inflow


ol flow

-control flow

main \ \ low


Circles represent sources of energy
and materials supplied
o the ecosystem.

Storage compartment may represent
energy or material such aswater.
sediment, organic matter, and nutrients

Thick arrow indicates that two flows
interact such that one augments the
other in a multiplicative fashion.

Figure 7. General model of floodplain hydrology. Numbers represent fluxes
ntimeters per year for a reach of the Gila River and its floodplain.
Gatewood et al. (195). Symbols after Odum (1971).

of water in
Values from


ecosystem structure. When the quantity
of biomass stabilizes, energy flow con-
tinues to be utilized for the mainten-
ance of existing biomass levels through
replacement of organisms that have died
and are undergoing decay.

Therefore, food production and
maintenance of ecosystem structure are
the two basic ways in which primary pro-
ductivity is important to consumer or-
ganisms. Numerous studies in riparian
ecosystems have documented the capacity
of these ecosystems to maintain high
vertebrate population densities, parti-
cularly in comparison with upland eco-
systems. The extent to which these
higher standing crops of vertebrates
respond to the production of food and
the maintenance of structure will be
discussed in later sections. Here we
examine the nature of biomass distribu-
tion in riparian ecosystems and annual
rates of biomass accumulation.

Biomass Distribution and Accumulation

Aboveground biomass in riparian
ecosystems varies widely ranging from 10
kg/m2 to 119 kg/m2 (Table 11). There is
insufficient information to determine
the basis of this variation, but differ-
ences in stand maturity or age probably
obscure regional trends. However, basal
area, a rough index of the amount of
woody biomass, is available for a larger
number of ecosystems. When the basal
areas of both riparian and upland eco-
systems are compared (Figure 8), it is
apparent that basal area for uplands
follows a more regular pattern and is
under some control by annual precipita-
tion. The curve shown in Figure 8 de-
lineates a hypothetical maximum set of
values for basal areas in upland forests
and shows a decline below about 50 cm
annual precipitation where grasslands
begin to replace forests. In contrast,
the basal areas of riparian forests ap-
pear to be independent of precipitation,
resulting in the presence of floodplain
forests in climates where upland ecosys-
tems support only grassland or desert
vegetation. The more moist conditions
of riparian ecosystems, as compared with
uplands, are a result of the convergence
of runoff along river corridors.

Annual aboveground biomass produc-
tion of riparian forests varies between

339-650 g dry wt/m2 for litterfall
(leaves, fruits, and flowers) and be-
tween 311-1100 g dry wt/m2 for stem wood
production (Table 11). Since litterfall
varies in a predictable fashion with
climatic and edaphic factors (Bray and
Gorham 1964) and leaves are the photo-
synthetic structure, litterfall values
are probably highly correlated with
primary productivity. Lowest and high-
est values for litterfall roughly corre-
spond at respective sites with lowest
and highest values for stem wood produc-
tion, but the correlation between the
two is rather low (r=0.38). The produc-
tion of stem wood biomass accounts for
about 54% of aboveground biomass produc-
tion. The remainder is litterfall
(mostly leaves) which is available to
different groups of consumers depending
on the season.

Production of belowground biomass
and subsequent mortality of roots may be
essential in maintaining levels of or-
ganic matter in soils. No conclusions
can be drawn for belowground biomass
standing crop and production since only
incomplete estimates of the total are
available (Table 11). However, Burns
(1978) reported higher standing stocks
of fine root biomass and greater season-
al differences at undrained, as compared
with drained, cypress strands in Flor-
ida. This suggests that the drier site
(with less aboveground production; Table
11) had slower root turnover rates than
the wetter site with natural flows. All
of the reported belowground root values
exclude stump biomass which may account
for approximately one-half of the total
belowground biomass (Harris et al.
1975). More information is needed to
determine if root biomass distribution
and production respond to other factors
such as hydroperiod, water table depth
or sediment composition and to evaluate
the influence of these variables on
species composition of riparian ecosys-

Ecosystem Metabolism
Biomass distribution and annual
rates of biomass production are rela-
tively static measurements that tend to
obscure seasonal differences in energy
flow. Both temperature and hydroperiod
have a profound influence on the carbon
balance of floodplain soils. Mulholland

m m m Th1 uralg terlg anifass jtiorfipar rest = I I

Stem Basal Biomass(kg/m2) Leaf & fruit Stem wood Total biomass
density area Above- Below- litterfall production production
Forest type (No./ha) (m2/ha) ground grounda (g/m2yr) (g/m .yr) (g/m2.yr) Source

Cypress floodplain,
Fla. 1644 32.5 28.4 -- 521 1086 1607 Brown 1978
Bottomland hardwood,
La. 1710 24.3 16.5 -- 574 800 1374 Conner & Day 1976
and pers. comm.
La. 1235 56.2 37.2 -- 620 500 1120 Conner & Day 1976
and pers. comm.
Ill. 45.2 -- 348 330 648 Mitsch et al. 1977,
Mitsch 1978
Cypress strand,
Fla. 19.2 0.80(to 339 772 1111 Burns 1978
30 cm)
Cypress strand,
drained, Fla. -- -- 10.3 0.31 (to 311 370 681 Burns 1978
30 cm)
Cypress strand,
Fla. 28.6 2.34 (to 650 640 1290 Nessel 1978
40 cm)

Floodplain swamp,
N.C. 705 47.8 27.6 2.70(to 524 585 1384 Mulholland 1979,
40 cm) Brinson et al. 1981h
Fenn, rinn. 3348 25.1 10.0 -- 412 334 746 Reiners 1972

Riverine forest, 3792 59.6- 118.9 1.22 -- .-- Golley et al.
Panama 1975

Floodplain forest,
I1. -- 29.0 -- -- 1250 Johnson & Bell 1976
Transition forest,
Ill. -- 14.2 -- -- 800 Johnson & Bell 1976

Alluvial swamp, 2730 69.0 2.35(to 522 Brinson et al. 1980;
N.C. 40 cm) Brinson et al. 1981b

aRoot biomass, not including stump roots.

Transition from
upland grassland
SUpland stands to fri
ORiparian stands


o 2

3 I

10 20 30 40 50 60

DA og~~s










70 80 90 100 110 120 130 140


150 160 170 180 190 200 210

Figure 8. Effect of increasing annual precipitation on the basal area of vegetation
for upland forests (curve) in comparison to riparian forests (no pattern). Sources for
upland forest: (A) Eggler (1938); (B) Fonda (1974); (C) Gilman (1976); (D) Hough
(1936); (E) McEvoy et al. (1980); (F) Rice (1955); (G) Stearns (1951); and (H) Whit-
taker et al. (1974). Source for riparian forests: (1) Anderson and White (1970); (2)
Barclay (1980); (3) Brinson et al. (1980); (4) Brown (1978); (5) Burns (1978); (6)
Conard et al. (1977); (7) Conner and Day (1976); (8) Crites and Ebinger (1969);
(9) Fonda (1974); (10) Freeman and Dick-Peddie (1970); (11) Fredrickson (1979);
(12) Golley et al. (1975); (13) Hall and Penfound (1939a); (14) Hall and Penfound
(1939b); (15) Hall and Penfound (1943); (16) Hosner and Minckler (1963); (17) Johnson
et al. (1976); (18) Lindauer (1978); (19) Lindsey et al. (1961); (20) Mulholland
(1979); (21) Nessel (1978); (22) Penfound and Hall (1939); (23) Rice (1965); and
(24) Zimmerman (1969).

(1979) obtained detailed measurements of
floodplain forest floor respiration for
2 years under flooded and unflooded con-
ditions in North Carolina. Highest
respiration rates corresponded with
highest temperatures and greater respir-
ation rates were observed for unflooded
conditions. Since unflooded conditions
and high temperatures coincided during
the growing season, a large proportion
of the respiration and carbon dioxide
loss from the forest floor occurred at
this time. Total forest floor respira-
tion averaged 0.95 g C/m2*day taking
into account changes in flooded and un-
flooded portions of the swamp throughout
the year. Of this, 74% was due to the
respiration of unflooded portions, 17%
to flooded portions, 5% to respiration
of the water column, and the remaining
4% to anaerobic respiration. This sug-

gests that alternate wetting and drying
of wetland soils augments losses of
carbon dioxide and prevents organic
matter from accumulating. Where water
level fluctuations are absent, organic
matter frequently accumulates as peat

Primary productivity and respira-
tion measurements for riparian ecosys-
tems are available only for a cypress
forest in Florida studied by Brown
(1978). Gross productivity averaged 26
g C/m2.day, highest of all other cypress
ecosystems that she studied, especially
when compared to the non-riparian cy-
press domes of Florida. Community
respiration was also higher (25 g
C/m2.day) than other cypress ecosystems.
These values are among the highest re-
ported for any ecosystem. The abundance









Sf phosphorus and other nutrient sup-
lies from stream flooding provide
sources necessary to sustain these
high rates of primary productivity.
Sectors Affecting Primary
Productivity and Growth

S The hydrology of riparian ecosys-
ems can have an effect on the meta-
olism and growth of vegetation in three
basic ways. First is water supply,
Uhereby water storage is recharged
through seepage and channel overflow to
loodplains. This is of great import-
nce for plants in arid climates since
t has been shown that riparian forest
communities are maintained in regions
too dry to support upland forests
IFigure 8). Second, nutrient supply in
iparian ecosystems depends partly on
edimentation of particulate matter
transported by overbank flow and partly
n the availability of dissolved nutri-
nts in the water in contact with flood-
lain soils. Finally, in comparison
with stagnant water in non-riverine wet-
ands, flowing water in floodplain
amps ventilates soils and roots so
at gases are exchanged more rapidly.
Oxygen is supplied to roots and soil
Iicrobes; at the same time the release
f gaseous products of metabolism such
as carbon dioxide and methane is en-
hanced. Water flow provides the medium
lor the export of dissolved organic
compounds, some of which are metabolic

I Flood frequency and groundwater
supply are major environmental factors
controlling the growth of floodplain
rees. To determine if reduced flooding
would affect tree growth on the Missouri
iver floodplain, Johnson et al. (1976)
measured radial wood growth representing
I-year periods prior to and following
ood control by reservoirs. Signifi-
ant decreases in growth of older, esta-
blished trees downstream from the reser-
oir occurred after flood control in
three species that germinate under
normal floodplain forest conditions.
Simulation of actual evapotranspiration
rates showed that when water surpluses
rran flooding were absent, low autumn
and winter precipitation in the region
Las insufficient to bring moisture in
he surface soil to field capacity by
he initiation of the growing season.

Even in the more humid climate of Miss-
issippi, bottanland hardwood vegetation
shows accelerated growth when artifi-
cially impounded water is available
until about June (Broadfoot 1967).
Perhaps three-fourths of the annual
radial growth occurs between late April
and late June in southeastern Arkansas
(Phipps 1979). Abundant water supplies
at that time may be critical to support
maximum growth.

Either too much or too little water
can have detrimental effects on growth
of vegetation that is already adapted to
an existing water regime. For cypress
trees in the Cache River floodplain in
Illinois, Mitsch et al. (1979a) reported
an increase in basal area growth rate as
a function of average river discharge
(Figure 9). The slower growth rates
prior to 1937 and following 1966 are
believed to be a result of water levels
raised and maintained by beaver dams.
In the drained portion of a cypress
strand in Florida, Burns (1978) reported
reduced litterfall and root biomass as
compared with a similar site with
natural water flows.

- 40
3 30

20 30 40 50 60

Figure 9. Relationship between annual
river discharge and cypress tree growth
on the Cache River, Illinois. (Mitsch
et al. 1979a).

Reductions in growth attributed to
lower nutrient and sediment supply rates
in the absence of river overflow have
not been documented. The higher fer-
tility of many floodplain soils would
likely have sufficiently large stocks of
nutrients so that effects would be
noticeable only after protracted periods
of nutrient deprivation. Moreover it is
difficult to separate effects of water
and nutrient supply. However, compari-
sons among riverine and stillwater
forested wetlands suggest that sustained
nutrient supply from river overflow is
responsible for higher nutrient cycling
rates (Brinson et al. 1980) and higher
rates of primary productivity (Brown
1978) in riverine forests. These
examples provide indirect evidence that
the nutrient supply to riparian ecosys-
tems can control tree growth and affect
soil fertility. Stimulation of tree
growth due to artificially augmented
nutrient supply has been demonstrated
and will be discussed later.

Energy Transfer from Producers
to Consumers
The preceding discussion on energy
flow illustrates how plant biomass pro-
duction is allocated between the
building of riparian ecosystem structure
and its continued maintenance. These
processes are similar to those that
occur in upland ecosystems, except to
the extent that they are affected by
additional water supply and flooding.
However, riparian ecosystems are unique
in the manner in which some of the
energy as organic matter or organic
carbon is transferred from producer to
consumer organisms. This uniqueness
derives from the fact that litterfall
produced within the riparian ecosystem
may be transported laterally (Bell and
Sipp 1975) and made available to in-
stream animal communities as well as
those downstream from the source of
organic matter production. As compared
with purely aquatic or terrestrial eco-
systems, organic matter produced in
riparian ecosystems has the potential of
supporting a diversity of food webs
within both habitat types.

There appears to be a useful dis-
tinction between swamp-draining and up-
land-draining streams in the manner in
which organic matter is transferred from

the riparian to the aquatic ecosystem.
Upland-draining streams are those that
have negligible or narrow floodplains
that receive organic matter from the
riparian zone principally by litter
falling directly from streamside vege-
tation to the surface of the stream.
Flood events may transport litter from
stream banks into channel and down-
stream. In comparison, swamp-draining
streams are in watersheds that have a
higher proportion of floodplain to up-
land surface area than do upland-drain-
ing streams. Not only do swamp-draining
streams receive litter falling directly
to their channel, but inundation of
broad floodplains provides the oppor-
tunity for additional transport of
organic matter from the floodplain

Export of Organic Matter from
Swamp-Draining Streams. Runoff is the
principal forcing function that influ-
ences export of organic carbon from up-
land watersheds (Brinson 1976). The
extent to which export is augmented by
floodplains and wetlands associated with
a river system probably depends on the
proportion of wetland to upland surface
area. Peculiarities of flow and inun-
dation patterns in floodplains may
directly influence the export of litter
(Bell and Sipp 1975). However, particu-
late forms of organic carbon usually
make up only a small portion of the
total organic carbon supply in rivers,
although the value of the particulate
fraction in providing food for certain
organisms is quite high.

Organic matter export from both up-
land-draining and swamp-draining water-
sheds (Table 12) shows a pattern of both
higher concentration and higher export
rate from watersheds that have extensive
wetland coverage. Mul holland and
Kuenzler (1979) demonstrated that there
was a linear relationship between annual
organic carbon export and runoff for
both watershed types, but that swamp-
draining watersheds export significantly
more organic carbon than upland-draining
watersheds. Rapid leaching of organic
carbon has been demonstrated from newly
fallen leaves of water tupelo (Nyssa
aquatica, a species common in some
southeastern swamps, which would contri-
bute to the organic carbon supply of

Table 12. Concentration and export of organic carbon in drainage waters for upland- and
swamp-draining watersheds. Values are dissolved organic carbon except as in-
dicated for total organic carbon (TOC). Values originally reported as organic
matter were multiplied by 0.5 to estimate organic carbon.

Annual Concentration
Area runoff mean or range Export
Locality (km2) (cm) (mgC/liter) (gC/m2.yr) Source

Swamp-Draining Watersheds
Neuse River, N.C.

Sopchoppy River, Fla.

Oscuro, Guatemala

Amatillo, Guatemala

Mississippi River
Delta, La.

Fahkahatchee River, Fla.

Barron River, Fla.

Lower Satilla River, Ga.

Creeping Swamp, N.C.
CP-10 (1976) 8(
CP-10 (1977) 8(
CP-20 (1976) 3;
CP-20 (1977) 32

Palmetto Swamp, N.C. 5S

Tracey Swamp, N.C. 141

Chicod Swamp, N.C. 13;

Clayroot Swamp, N.C. 11C

Upland-Draining Watersheds

Char Lake, N.W.T. 43

Brazos River, Tex.

Mississippi R. above delta, La.
Missouri River, Neb.

Ohio River. Ill.

Hubbard Brook, N.H.
watershed No. 2 (defor.) 0.1

watershed No. 6 (forest) 0.1

Bear Brook 1.3

Mirror Lake 0.8

Fort River, Mass.

Marion Lake, B.C.

Nanaimo River, B.C.




Malcolm &
Durum 1976




770 89 11.2-12.2


Brinson 1976

10.4(TOC) Day et al.

Carter et al.


12.7-36.2 Beck et al.

Mulholland &
Kuenzler 1979

Mulholland &
Kuenzler 1979
It "

.5 15.8 1.9(TOC) 0.30(TOC) deMarch

Malcolm &
Durum 1976















Hobbie &
Likens 1973,
Bormann et
al. 1974
Fisher &
Likens 1973
Jordan &
Likens 1975

Fisher 1977

Efford 1972

Naiman &
Sibert 1978

Humid tropics
Polochic, Guatemala 5,247

Sauce, Guatemala 300

San Marcos, Guatemala 170

194 1.1-3.7 4.8(TOC) Brinson
1973 & 1976

86 1.7-6.2 3.2(TOC)

85 0.9-1.9 2.2(TOC)

waters flowing through forested wetlands
(Brinson 1977). In contrast, leaching
of soluble organic carbon through well-
drained or upland mineral soil horizons
is relatively slow and inefficient since
residence times for absorbed organic
carbon may be several centuries
(Scharpenseel et al. 1958). Higher
organic matter export from swamp-drain-
ing streams appears to be related to
long retention times of water in contact
with the litter, detritus, and organic
soils of the forest floor.

The significance of particulate
organic detritus to filter feeding crus-
taceans in lacustrine and marine ecosys-
tems is well established. This evidence
suggests that detritus exported to down-
stream ecosystems is an important source
of energy for lakes and estuaries (Seki
et al. 1969, Brinson 1973, Livingston et
al. 1974, Livingston and Duncan 1979).
The correlation between intertidal vege-
tation surface area and commercial
yields of penaeid shrimp (Turner 1977)
as well as the influence of estuaries on
the plankton of the continental shelf
(Turner et al. 1979) extend the concept
of ecosystem coupling to near-shore
waters of the ocean.

The significance of dissolved
organic carbon exports is less apparent,
but concentrations and biological demand
for oxygen are high in surface waters of
many wetlands. This suggests that at
least a portion of the dissolved organic
carbon is readily available for micro-
bial metabolism and thus conversion into
particulate forms for filter feeders
(Correll 1978). Other fractions, parti-
cularly low molecular weight humic and
fulvic acids, have been shown to have
stimulating effects on marine phyto-
plankton (Prakesh et al. 1973) pre-
sumably owing to their capacity to make
certain micronutrients available for
uptake by algae. Flocculation of dis-
solved organic matter induced by the
brackish waters of estuaries may serve
as a mechanism for generating particu-
late forms that would be available for
filter feeders.

It has been demonstrated in several
estuaries that a large proportion of the
organic carbon in estuarine sediments is
derived from terrestrial sources (Rashid
and Reinson 1979, Tan and Strain 1979).

Since it has been shown that forested
wetlands export disproportionately high
amounts of organic carbon in relation to
their surface area, as compared with up-
land regions, sources of organic carbon
from wetlands may be vital in maintain-
ing organic carbon supplies to the
sediments in some estuaries. The
amount of organic matter in estuarine
sediments can in turn affect a number of
other variables including chemical oxi-
dation/reduction gradients, microbial
processes that convert nitrate to gas-
eous nitrogen, sediment/water exchanges
of ammonium and phosphate, and benthic
community species composition. Thus
where wetlands contribute to watershed
exports, stream alteration and wetland
drainage could reduce the concentration
and alter the distribution of organic
carbon in estuarine sediments.

Energy Flow and Community Struc-
ture in Upland-Draining Streams. In the
headwaters of upland-draining streams,
organic matter contributions to flowing
water (beyond those from groundwater
sources) derive principally from leaves
falling directly to the water surface
from streamside vegetation (Figure 10).

Aquatic Primary Producers
Stream |Distribution of
Order Width P/R Ratio Food Habits of
|I I Aquatic Invertebrates

Figure 10. Changes in structure and
function of upland-draining streams from
headwater to mouth. Modified from
Marzolf (1978).




Ce n w


In these situations, water flow is con-
fined to rather discrete channels in
relatively narrow valleys, as compared
with swamp-draining streams in flood-
Iplains where the water flow is usually
sluggish, and small increases in dis-
charge may increase the water surface
area by severalfold. It is obvious that
Streamside or riparian vegetation will
have greater influence on instream
energy flow near the headwaters where
the forest canopy is continuous, and
ave less influence in higher order
treams where the ratio of stream margin
to surface area decreases.
i Inflows and outflows of organic
matter are frequently segregated into
size classes as coarse particulate
>1 mm), fine particulate (<1 mm), and
issolved organic matter (Figure 10).
hole leaf detritus is consumed by
roups of invertebrates that "shred" or
fragment leaves into smaller particles.
these particles are further fragmented
by even smaller organisms while others
ct as collectors or macrogatherers
Cummins 1974). The coarsest fractions
ave the greatest probability of being
processed (either fragmented to smaller
fractions, metabolized by micro-organ-
isms, eaten by invertebrates or leached)
isher (1977) reports that 61% of the
gross input of coarse particulate or-
nic matter to Fort River, Massachu-
tts was metabolized, retained or con-
rted to smaller particles. For the
fine particulate organic matter, only 9%
Us processed and the rest exported.
ssolved organic matter actually showed
net gain which means it was being add-
ed at a greater rate than the stream
system could process it. Consumption
S dissolved and particulate organic
matter is largely through microbial res-
jration (McDowell and Fisher 1976);
wever, immature aquatic insect popula-
ons are largely particle feeders and
are dependent on particulate organic
gtritus and the associated microbial
Immunity for energy and nutrition.
perimental removal of leaf packs in
streams reduces the consumption of dis-
lved and particulate organic matter in
e water (Bilby and Likens 1980). Leaf
cks derived from riparian vegetation
thus provide sites for utilization of
Sanic matter which would otherwise be
orted downstream.



The energy flow or metabolism in
any stream sector can be quantified on a
unit area basis and consists of inputs
from upstream and tributary flows,
direct litterfall, and aquatic primary
productivity (aquatic macrophytes and
benthic and planktonic algae), and of
outputs from respiration and downstream
export (Fisher 1977). It is possible
that natural stream communities adjust
their structure and activities in undis-
turbed watersheds to maintain an ideal-
ized stream metabolism such that accre-
tion of all newly derived organic matter
(from leaf fall or autotrophic produc-
tion) will be consumed within a given
stream segment. For headwater streams
where autotrophic contributions are
negligible and leaf litter inputs
(coarse particulate organic matter) pre-
dominate, invertebrates such as collec-
tors and shredders are in greatest abun-
dance (Figure 10). The rapid leaching
of soluble organic matter from newly
fallen leaves also supplies microbial
communities with an energy source, which
then becomes available to certain col-
lectors as fine particulate organic
matter. Further downstream where the
canopy opens, or in headwater streams
with little shading (Minshall 1978),
grazers assume greater relative abund-
ance. Thus it appears that the presence
of riparian vegetation plays a profound
role in the structure of invertebrate
communities. Since many fish species
are dependent on these invertebrates as
their sole source of food, the riparian
vegetation indirectly plays a role in
fish community structure.

Ecosystem metabolism is an indica-
tor of similar trends in the signifi-
cance of riparian vegetation. The ratio
of gross photosynthesis to ecosystem
respiration (P/R) in the aquatic ecosys-
tem increases from a value of less than
1.0 in shaded headwater streams with a
continuous canopy to greater than 1.0
where autotrophic activity dominates
(Odum 1956). Minshall (1978) points out
that using this ratio alone to charac-
terize stream metabolism may obscure the
significance of the primary pro-
ducer-grazer food chain since a ratio of
less than 1.0 (predominantly heterotro-
phic) does not imply that primary pro-
ductivity is negligible. He warns
against application of ecosystem gen-

eralizations (as in Figure 10) since
there are substantial geographic areas
in arid regions where streamside woody
vegetation is water limited and with-
in-stream autotrophic processes pre-
dominate on an annual basis. Litterfall
inputs, primary productivity, and res-
piration vary widely in streams of dif-
fering size and degrees of shading
(Table 13). Primary productivity varies
inversely with litterfall and thus de-
creases with decreasing stream size in
humid climates. Although Fort River,
Massachusetts has abundant aquatic
macrophyte production, little is grazed,
and it enters the food web as detritus
(Fisher and Carpenter 1976). However, in
arid regions where streamside woody

vegetation is sparse, small streams re-
ceive most of their organic matter from
instream primary production of algae and
aquatic plants. Following a late summer
flash flood in a Sonoran Desert stream,
90% of the preflood algal standing crop
and primary productivity was attained in
2 weeks (Fisher et al. 1980). This
demonstrates the rapid recovery of a
probable food base for consumers after
disruptive floods.

Implications for removal of ripar-
ian vegetation extend beyond those of
disrupting coarse particulate matter in-
puts and shifting energy flows toward a
more autotrophically based food chain.
Removal of vegetation is usually accom-

Table 13. Comparison of litterfall, primary productivity, and respiration for several


sizes of streams.

kq organic matter/n *yr

Litterfall productivity

Ecosystem P/R
respiration ratio

Humid climate

Headwater stream,
canopy continuous
4th order stream, b
canopy discontinuous
Broad stream,
canopy negligible

Arid climate

Small stream,.
trees lacking

aBear Brook, N.H. (Fisher and Likens 1973).
Fort River, lass. (Fisher 1977).
cThames River, England. Only net productivity of plankton
Litterfall from Mathews and Kowalczewski (1969).
Deep Creek, Id. (Minshall 1978).

is available (Iiann et al. 1972).

















Inied by changes in hydrology, sediment
and nutrient loading, and temperature so
that results are neither orderly nor
ledictable. If the consequences were
rely an increase in P/R ratio, inver-
tebrate communities would probably
Ipidly adjust their species composition
d community organizations to new con-
The structural integrity of stream
ds in lower order streams is dependent
part on stabilization by roots and
the presence of snags, logs, and other
structions for creating stable surface
a and a varied and complex substrate
arzolf 1978, Benke et al. 1979, Bilby
and Likens 1980). Moreover, if riparian
Igetation removal is accompanied by
earcutting of the watershed, conse-
uences may include greater pulses in
discharge, higher amounts of annual run-
f, and increased concentrations of
trients and sediments (Likens et al.
177, Bormann and Likens 1979). The
ift to a higher energy, more eutrophic
vironment will produce conditions to
ich only a few of the existing species
of aquatic invertebrates and fishes are
apted. Geomorphic changes in erosion
d sedimentation may accelerate sever-
fold with these disruptions. Where
alternatives to deforestation and land
le changes are not possible, a protec-
ve buffer of riparian vegetation
would remain intact to maintain the
integrity of at least some of the energy
urces and organic matter processing

Nutrient cycling in riparian eco-
stems can influence the water quality
streams and rivers. Riparian ecosys-
*ms along small, low order streams are
buffer zones where excessive nutrients
d sediments from upland disturbances
y be trapped and assimilated. For
rger streams and rivers, overbank flow
of water during flood events provides an
Iportunity for upstream flows to come
contact with the riparian ecosystem.
the absence of a vegetated riparian
zone, water is exported downstream with
little opportunity for nutrient assimi-
tion and transformation, except that
provided in stream channels. In com-
parison with most stream channels,


floodplain forests have greater struc-
tural complexity due to the presence of
more stable sediments, anastomosing
roots, a layer of decomposing leaves and
woody material on the forest floor, and
complex topographic features.

Many of the mechanisms of nutrient
conservation by riparian ecosystems are
universal and differ little from those
found in upland ecosystems. Where fun-
damental differences exist, they are
related to (1) the influence that flood-
ing and an "aquatic" phase has on re-
stricting oxygen availability to soils
and sediments, thus altering metabolic
pathways of microbial communities, and
(2) the aqueous transport system that
provides pathways of exchange between
stream channel and floodplain through
lateral imports, sedimentation, and
exports of elements. These mechanisms
of nutrient assimilation and transfor-
mation are examined in detail.

Most nutrient cycling studies focus
on nitrogen and phosphorus, and have
been conducted in southeastern flood-
plain forests where the presence of
relatively long hydroperiods and broad
floodplains has considerable influence
on water quality of streams and rivers.
In arid riparian ecosystems, water
quantity, rather than its quality, may
be the overriding controlling factor in
ecosystem processes.
Distribution of Nutrients
In forested ecosystems, the distri-
bution of nutrients among ecosystem com-
ponents and annual changes in nutrient
content of these compartments tend to be
proportional to the distribution and
changes in biomass. High or low stand-
ing stocks of nutrients generally cor-
respond with high or low standing stocks
of organic matter in both wetland and
upland forests. For example, data on
phosphorus distribution in riverine
forests show that the rank, from highest
to lowest standing stocks of phosphorus,
is usually (1) soil (total P to approxi-
mately 25 cm depth), (2) aboveground
wood, (3) belowground wood, (4) canopy
leaves, (5) litter layer, and (6) sur-
face water (Table 14). Canopy leaves
and other non-perennial structures such
as flowers and fruits tend to be highly
enriched in phosphorus concentration

Table 14. Distribution of phosphorus in riverine forests.

--- I
g P/m2
Cypress Creeping
Prairie Strand Cache R., Swamp,
Component Cr., Fla.a Fla.b Ill.c N.C.d
Leaves 1.26 0.4e 1.22 1.2
Aboveground wood 3.52 3.6 5.09 5.45
Belowground (lateral --- 6.2 2.82 1.52
Surface water 0.19 0.8 0.176 0.0095
Litter layer --- 2.1 --- 0.45
Soil 46.6h 90.2h 119i 33.65

aBrown (1978); bNessel (1978); cMitsch (1978); dYarbro (1979);
litterfall; f3.2 to 23 cm depth; BBrinson et al. (1981b); hTo
iTo 24 cm depth.

relative to other biomass components, for upland ecosyst
particularly woody ones, but the total wetlands of similar
quantity per unit area is lower. Sedi- et al. 1980). Their
ments represent a large proportion of for nitrogen which t
the phosphorus capital of the ecosystem the importance of f
although only a small proportion of this maintaining the relay
is available for plant uptake at one ity of riverine fore
Annual phospho
Major Flows in the Nutrient Cycle wood also appears to

Major nutrient flows that are most
frequently studied are nutrient return
from the canopy (as litterfall and stem-
flow), decomposition of the litter
layer, increment in wood accumulation,
and sedimentation. Shorter term flows,
such as sediment-water exchanges, are
discussed later. Taken alone, each of
these pathways would give an incomplete
picture of nutrient cycling. However,
when similar pathways are compared for
different ecosystems, patterns may
emerge which provide information on
overall ecosystem fertility. For ex-
ample, phosphorus flows for riverine
forests (Table 15) are higher than those

20 cm depth;

ems and stillwater
latitudes (Brinson
e is a similar trend
ends to substantiate
'luvial processes in
atively high fertil-

rus uptake by stem
correspond to phos-

phorus supply. For a cypress strand in
Florida, phosphorus uptake in stem wood
increased approximately threefold when
nutrient rich sewage effluent was re-
leased into the ecosystem (Nessel 1978).
As compared with other cypress-contain-
ing ecosystems that had lower fluvial
inputs, a floodplain forest in Florida
had greater stem wood production as
measured by annual basal area increment
(Brown 1978). However, because of the
extremely low concentrations of phos-
phorus in stem wood, annual increments
in phosphorus accumulation by this pro-
cess tend to be quite low when compared
to other major flows (Brown 1978, Nessel
1978, Yarbro 1979).

Table 15. Litterfall and aqueous flows of phosphorus from the canopy to the forest floor in
riverine swamps.

precip- kg P/ha-yr
station Litterfall Litter- Total
locality (cm) (kg dry wt/ha) fall Aqueous return Source

Far River 104.7 6428 5.38 1.55 6.93 Brinson
wamp, N.C. et al. 1980
Creeping 124 6010 3.29 1.6 4.9 Yarbro 1979
Swamp, N.C.
Irairie 5970 9.1 --- 9.1 Brown 1978
reek, Fla.
( che River, 105 3480 7.7 1.4 9.1 Mitsch
1. et al. 1979a
press strand, 105.3 8150 6.86 6.86 Nessel 1978

Release of nutrients by decomposi-
ion of leaf litter in riverine forests
s usually sufficiently rapid that there
is little or no accumulation from year
o year. The "half time" of loss is the
ime, in years, that would be required
or one-half of the dry weight to dis-
appear by decomposition. Half times for
eciduous leaves range from less than
.5 year to greater than 1.5 years
Table 16), while woody material and
Pinus spp. leaves decompose more slowly
ind -have longer half times. Stagnant
backwater areas and depressions of
,loodplains tend to accumulate litter
and sometimes peat. In spite of these
exceptions, most of the nutrients of
he litter layer appear to be recycled
n an annual time scale. However, some
studies have shown immobilization of
nitrogen and phosphorus that may con-
inue for several months (Figure 11),
particularly under flooded conditions
during the cool season that follows
autumn leaf fall in temperate zones
*Brinson 1977). This suggests a capac-
ity for short-term accumulation of nu-
Srients from the water, and thus an

influence on water quality, even during
the dormant season when losses of dis-
solved nutrients due to flooding might
be greatest.

Sedimentation of particulate
material on floodplains has been docu-
mented in a number of studies (Table
17). Although these data tend to be
biased by not considering erosion and
scouring as well, considerable quanti-
ties of sediment may accumulate over
large areas, particularly during large
flood events of low recurrence inter-
vals. Estimates of annual phosphorus
deposition by sedimentation range be-
tween 1.72 kg P/ha for a clear stream
floodplain in North Carolina (Yarbro
1979) to 30 kg P/ha for a floodplain
swamp in Florida (Brown 1978). These
sedimentation rates approach or exceed
some of the fluxes first described, al-
though not all of the sediment is im-
mediately available in ionic forms for
plant uptake. Nevertheless, sedimenta-
tion represents a nutrient source that
would otherwise be transported down-

Table 16. Summary of decomposition rates of litter in riverine forests.

Duration of Half times
measurement mm of loss,
Forest type (weeks) Litter type Site mesh years Reference

Cypress strand,
Fla. 52 Site litter, leafy Forest floor 0.8 0.81 Burns 1978
1.6 0.50
Debris pile 0.8 0.92 "
1.6 1.00
52 Site litter, woody Forest floor 0.8 1.54
1.6 1.33
Debris pile 0.8 0.80
1.6 1.78
Cypress strand, 51 Site litter Flooded 0% time 1.6 1.47 Duever et al.
Fla. Flooded 50% time 1.6 3.01 1975
Flooded 61% time 1.6 2.31 "
4 Cypress strand, 52 Taxodium ascendens Ivs Wet site 1.6 1.26 Nessel 1978
o Fla. Dry site 1.6 1.51 "

52 Nyssa sylvatica Ivs Wet site 1.6 0.82
Dry site 1.6 0.91 "
52 Acer rubrum Ivs Wet site 1.6 1.36
Dry site 1.6 0.95
Alluvial swamp, 48 Nyssa aquatic Ivs 1.6 0.37 Brinson 1977
N.C. 48 Nyssa aquatica twigs 1.6 2.48 "

Beaver pond, 75 Salix sp. Ivs 3.5 0.71 Hodkinson 1975
Alberta 75 Juncus tracyi Ivs 3.5 1.69
75 Pinus contorta Ivs 3.5 3.30 "
75 Deschampsia espitosa 1vs 3.5 1.03

Mixed floodplain 50 Fraxinus nigra Ivs 0.05 0.64 Merritt &
forest, Mich. 0.5 0.41 Lawson 1978
8.0 0.14 "

a Half time is the time required for disappearance of one half of the dry weight, according to the exponential decay
formula X/X = e-kt where X is the dry weight initially present and X the dry weight remaining at the end of the
measurement period, t, in years. Half time is calculated as 0.693/k.

m m mm m m m m m m

ITable 17. Sedimentation rates of phosphorus in the floodplains of riverine forests.

ILocality Sedimentation rate Kg/ha Soul


I Cache River,

Prairie Creek,
I Creeping Swamp,

ICreeping Swamp,

Kankakee R.,

3.6 g P/m2 contributed by
flood as sedimentation for
flood of 1.13 yr recurrence
3.25 g P/m2 yr as sedimenta-
tion from river overflow
0.17 g P/m2 yr sedimenta-
tion on floodplain floor
from river overflow
0.315-0.730 g P/m2 yr based
input-output budget of
floodplain (most was fil-
terable reactive phosphorus)
1.357 g P/m2 contributed by
unusually large spring flood
lasting 62-80 days




Mitsch et al. 1979a

Brown 1978

Yarbro 1979

3.15-7.30 Yarbro 1979


Mitsch et al. 1979b

stream if the floodplain did not func-
tion as an area of deposition.
The magnitude and rate of nutrient
uptake by vegetation, return to the
forest floor as litterfall, and nutrient
release by decomposition in southeastern
floodplain forests suggest that they are
capable of retaining nutrients by re-
cycling them as fast or faster than most
other forest types. Possession of a
strong recycling component reduces the
probability that nutrients entering the
system will be lost by leaching from the
soil and by export in throughflowing
water. Sedimentation of phosphorus in
the system provides evidence for sus-
tained supplies of new material for re-
cycling as long as inflow pathways are


(by channel

overflow and

Soil-Water Nutrient Exchanges

When floodwaters come in contact
with the soils of riverine forests or
when runoff from uplands passes through
the riparian zone to headwater streams,
the relatively slow movement of these
water masses provides an opportunity for
mechanisms to function that may alter
the nutrient constituents of the water.
Nitrate (NO ) is often the most abundant
form of nitrogen in stream waters and,
when present in high concentrations,
contributes to water quality problems.
When an anaerobic zone is present near
the surface of poorly drained sediments,



0 1 I 1 I- I- I- I I. I.
N O J F M A M J J A S 0 N D

Figure 11. Immobilization of phosphorus
and nitrogen by decaying leaf litter in
an alluvial swamp. After Brinson

it profoundly affects the pathways of
nitrogen. Denitrification (NO0---N,) in
anaerobic layers depends largely on the
rate of nitrate supply. In the absence
of external inputs of nitrate, it can be
supplied internally by nitrification of
ammonium (NH---NO3) under aerobic con-
ditions. Patrick and Tusneem (1972)
have proposed a scheme whereby ammoni-
fication (organic N--NH,) in an anaer-
obic zone supplies, through diffusion,
the substrate for nitrification in the
aerobic surface layer. Diffusion of
nitrate back to the reduced zone results
in denitrification, so that the nitrogen
gas (N2) produced is not in a form that
can contribute to water quality and
eutrophication problems. These pathways
are illustrated in Figure 12.

Evidence for denitrification is
reported for the Santee River swamp in
South Carolina (Kitchens et al. 1975).
Concentration of nitrate progressively
decreased from the river channel to the
interior of the swamp backwaters, sug-
gesting that increased contact time of
overflow waters with the forest floor
resulted in decreases in nitrate concen-
tration, presumably by denitrification.
More direct evidence is available from a
cypress-tupelo swamp where amended ni-
trate concentrations decreased rather
rapidly from surface water in contact

with organic sediment (Brinson et al.
1981a). The sediments are a permanent
sink for nitrate because it is denitri-
fied when it diffuses to the anaerobic

Although natural rates of denitri-
fication are difficult to determine, the
potential for this process is high and
can be sustained over protracted periods
as long as anaerobic conditions are
maintained and an energy source is
available to drive the process. Conse-
quently, poorly drained areas of ripar-
ian ecosystems can assimilate nitrate at
rates well in excess of natural sup-
plies, whether the source is from ni-
trogen-rich stream water in overbank
flooding or is from nitrogen-rich runoff
from adjacent agricultural land. In
either situation, less nitrate would be
exported to downstream aquatic ecosys-
tems for possible eutrophication if the
riparian zone is protected and natural
hydrologic processes are allowed to

Analysis of exports from watersheds
containing riverine wetlands support
these observations. For small coastal
plain swamp streams in North Carolina,
Kuenzler et al. (1977) showed that con-
centrations and exports of nitrate were
considerably higher for channelized
streams in which the forested wetlands
had been circumvented than for natural
streams in which considerable flooding
occurred during high discharge.

Floodplain forests also show a high
capacity for phosphorus retention and
cycling. Yarbro (1979) developed a
rather complete phosphorus budget for a
swamp floodplain ecosystem in North
Carolina (Figure 13). Inputs to the
ecosystem, mostly from upstream inflows,
exceeded outputs by 3.15 and 7.30 kg
P/ha.yr for each of the 2 years of study
which characterizes the floodplain as a
phosphorus sink. Although most of the
loss appeared to be from soluble reac-
tive phosphorus in the water, there was
a substantial amount of sedimentation
(1.7 kg/ha.yr) of particulate forms.
High forest floor/surface water ex-
changes substantiate the idea that the
sediments are the principal site for
transformation of various forms of phos-
phorus fractions. Transfers between the
forest floor, the deeper soil, and woody

Figure 12.

Pathways of nitrogen transformations in an oxidized and reduced sedi-
system. Modified from Gambrell and Patrick (1978).

biomass were shown to be approximately
half those of surface water/forest floor
exchanges. Estimates of tree wood incre-
ment (0.6-1.2 kg P/ha.yr) suggest that
Ithe vegetation would serve as a sink for
phosphorus only if the forest were ac-
cumulating biomass. This is relatively
small compared to the rate that phos-
Iphorus is recycled by the vegetation,
which suggests rather tight coupling
between litterfall from the canopy, de-
composition of litter, and phosphorus
Uptake by roots. In the absence of a
complex floodplain ecosystem, such as
that which would result from stream
channelization, there would be little
opportunity for phosphorus recycling and
sedimentation. Under channelized condi-
tions, downstream exports would increase
and the phosphorus would likely be made
available to an aquatic ecosystem such
as a lake or estuary.

The Significan'ce of Hydroperiod and
Nutrient Cycling
The importance of seasonal changes
in water level and flow to nitrogen
cycling can be illustrated by consider-
ing the annual cycle of an idealized
stream-floodplain complex. The scenario
begins with a major flood in the winter
of a southeastern swamp forest (Figure
14). Suspended sediments and dissolved
nutrients are transported from the
stream into the floodplain where water
velocity diminishes. Suspended sedi-
ments and the particulate forms of
nitrogen that they contain settle to the
forest floor and the dissolved nitrogen
forms in the water diffuse to the soil
to interact with detritus and sediment
on the forest floor. Deciduous trees of
the floodplain are dormant in the
winter; consequently they are not then



Figure 13. Phosphorus storage and fluxes in the Creeping Swamp floodplain ecosystem,
North Carolina. Modified from Yarbro (1979). See text for explanation.



Storage com.prtment may represent
energy Wo materUl uch as water.
sedimenl. orgaInc mhttr,. and nutrents.

H4el|gon repnuests consumer community
or the processun of energy by animits.

Symbol npreesents Iiary producers
or the proIess of eneMfl by plants.

I _



I Figure 14. Seasonal phenology of a tupelo-cypress swamp showing mechanisms
conservation and recycling.

capable of nutrient accumulation. Mech-
anisms of nutrient removal under these
conditions may include (1) uptake by a
community of filamentous algae that re-
ceives sufficient light for maintenance
only when the forest canopy is leafless
and (2) immobilization by decomposing
microbes that are utilizing the carbon
rich but nutrient poor leaf litter that
fell during the previous autumn.

When the floodwaters warm in the
spring, decomposition of detritus is en-
hanced, thereby releasing nutrients for
plant uptake and growth. Appearance of
leaves in the forest canopy shades the
forest floor, resulting in death of the
filamentous algae. Decomposition of the
algae augments the nutrient supply for
plant uptake. Evapotranspiration by the
forest depresses the water level and
eventually depletes most standing water.
The seasonal events turn full cycle with
leaf fall in autumn and resumption of
flooding in the winter.

of nitrogen

The timing of these seasonal events
and the magnitude and mechanisms of nu-
trient cycling described more fully in
the sections above illustrate two impor-
tant features: (1) the high capacity of
certain riparian forests to recycle nu-
trients such as nitrogen and phosphorus
as compared with the generally lower
rates at which they are imported from
outside the system, and (2) the influ-
ence that contact with the forest floor
has on the nutrients in flood water.
The mechanisms just discussed describe
how floodplain forests can capitalize on
and utilize these inputs.

Of course the potential for these
interactions to occur depends on the
hydroperiod or the length of time and
the quantity of water and nutrients that
come in contact with the floodplain.
Many southeastern river swamps tend to
have geomorphic, hydrologic, and cli-
matic characteristics that are optimal

for strong coupling between streams and

Measures to control flooding or
speed the conveyance of water downstream
tend to deprive riparian ecosystems of
the influx of materials that sustain
their nutrient-rich properties. When
drained and deprived of flooding by
streams, it is likely that disrupted
riparian ecosystems will become sources,
rather than sinks, of nutrients and
sediments for ecosystems downstream due
to the elimination of specialized nu-
trient tranformations that depend on an
"aquatic" phase. Drainage will convert
them from systems characterized by
lateral inputs and outputs to ones of
vertical movement and downward leaching.
Downstream ecosystems must then adapt to
receiving altered rates of organic mat-
ter and inorganic nutrient supply.
Changes in nutrient regimes represent
only one example of a host of other ef-
fects on riparian ecosystems when they
are altered.


A great deal of emphasis has been
placed on the similarity among riverine
ecosystems in the material above. The
underlying theme is that ecosystem
structure and organization is the result
of the energy and pattern of delivery of
flowing water. Hydrologic and geomor-
phic factors, both in the riparian zone
and in the watershed, appear to have a
fundamental influence on differences
observed among riparian ecosystems. It
is the differences, particularly in ve-
getation and factors that influence
vegetation, that will be discussed

Although no attempt is made to pre-
sent a fonnal classification for ripar-
ian ecosystems, broad distinctions exist
among them that fall into useful cate-
gories. Differences in climate, in
spite of the local edaphic properties of
floodplains, have an influence on
species composition of the plant com-
munity. Whether a stream channel is
composed of bedrock or passes over al-
luvial fill will greatly influence the
dimensions of the riparian zone. Within
a given floodplain plant community,
abrupt changes in stream channel adjust-

ment and catastrophic flood events can
be so prevalent and recurrent that the
community is maintained in an early
stage of succession.


The transition from humid to arid
climates does not have nearly the con-
trol on the structure of riparian
forests as it does on that of upland
ecosystems (Figure 8). Presumably this
is due to the fact that floodplains cap-
ture runoff water that is exported from
upland regions and at least part of that
water is available for riparian vegeta-
tion. The line or isopleth separating
areas of less and greater than 2.5 cm
runoff annually in the central U.S.A.
shows good agreement with the separation
between wet and dry climatic zones (Fig-
ure 15). Where runoff is less than 2.5
cm/year there is a greater probability
of encountering intermittent streams in
relatively large drainage basins than is
true in more humid climates. As a re-
sult, riparian vegetation may be sub-
jected to water deficiency as well as
water excess resulting in corresponding
changes in species composition. In
floodplains of arid regions, plants that
can tolerate periods of drought by ex-

Figure 15. Map showing 2.5 cm isopleth
of annual runoff. From Langbein et al.

tending roots to the water table (phre-
Stophytes) and also withstand flooding
are the most likely to survive. Under
humid climates, water is much more
readily available to plant communities
n floodplains and species composition
ill correspond accordingly.
In the western U.S.A. where the
temperature at higher elevations has a
large influence on water balance, the
2.5 cm isopleth circumscribes many of
tIhe mountainous areas. The extent to
which this water supply is available to
floodplain vegetation at lower eleva-
tions depends largely on the amount of
Discharge relative to the volume of al-
luvial fill. The ratio of evapotran-
spiration to precipitation increases
with decreasing altitude under most cir-
Icumstances. It is possible for runoff
from mountainous areas to be lost as
evapotranspiration or in groundwater
Flow by the time it reaches lower alti-
tudes (Thomsen and Schumann 1968). In
areas where there is little alluvial
fill for water storage (e.g., steep
rocky ravines), xeric conditions prevail
and vegetation may differ little from
the surrounding uplands (Zimmermann

I One of the major features that
distinguishes certain arctic drainage
basins from those in warmer climates is
Ithe impermeable layer of frozen ground
(permafrost). As a result, runoff is
from the soil surface so that ground-
water infiltration and storage play an
insignificant role in hydrologic pat-
terns. Permafrost also affects channel
stability and morphology. For example
Idowncutting of the river channel may be
retarded because the streambed remains
frozen during much of the ice-free
period. Although it has been esta-
Iblished that rooted vegetation along
stream banks retards erosion due to the
binding capacity of roots (Smith 1976),
the insulating effects of vegetation in
Spermafrost regions may be more signi-
ficant in maintaining stream banks in a
consolidated, frozen state. On the
Porcupine River removal of river bank
Vegetation increased the depth to summer
thaw from a maximum of 0.3 m with vege-
tation cover to as much as 1 m in
cleared areas (Cooper and Hollingshead

Underfit Streams and Downcutting
Within the context of differences
imposed by climate on riparian ecosys-
tems, a further distinction can be made
between stream systems with bedrock
controlled channels and those with al-
luvial channels. The latter, referred
to as "underfit" streams (Dury 1954a,
1964b, 1965), may have extensive flood-
plains, and are free to adjust their
dimensions, shape, and gradient in re-
sponse to hydraulic changes. Their
channel bed and banks are composed of
material transported by the river under
present flow conditions. By comparison,
bedrock controlled channels are confined
between rock outcrops, and in extreme
cases, have virtually no floodplain so
that only a very narrow margin can be
considered riparian. Of course a given
stream may have alternating sectors of
both conditions which makes generaliza-
tions difficult. However, the distinc-
tion is important when considering the
values and attributes of riparian eco-
systems and their plant and animal com-

Most of the ecosystems described in
the previous sections are those with
clearly distinguishable floodplains and
can be categorized in the underfit
stream type. However, even in the
absence of distinct floodplains,
streamside plant communities are usually
distinguishable from upland communities
in species composition, moisture avail-
ability, and physiognomy. They repre-
sent the riparian zone, although usually
quite narrow compared with floodplains,
that has an abundant water supply, is
characterized by fluctuating water
levels, and is exposed to the abrasive
force of flowing water during floods.
In the situation of lower order streams
that have considerable canopy cover, the
importance of leaf fall has been des-
cribed as essential to maintaining in-
stream energy flow and fish production
(pages 70 71 ). Some of the woody
riparian communities that will be des-
cribed below occupy stream margin envir-
onments that cannot be considered

Influence of Catastrophic Forces
Without doubt, the species composi-
tion of riparian ecosystems is a re-
sponse to multiple factors that are in
some way related to hydroperiod and the
energy of flowing water. However, in
many cases more catastrophic forces
create abrupt episodes of severe and
destructive stress that dominate commun-
ity development. Major floods may elim-
inate large stands of forest by erosion
and bank undercutting, creation of new
channels, and burial under deep deposits
of sediment. Wolman and Leopold (1957)
report that the Kosi River in India mi-
grates across its valley at the rate of
750 m/yr. The disordering effects of
these events serve to maintain an array
of community types in floodplains that
would otherwise mature into more homo-
geneous, even aged stands. Vogl (1980)
cites numerous examples of perturba-
tion-dependent ecosystems where the
maintenance of certain species is
assured by catastrophic events such as
floods, wind, storms, fire, volcanism,
and glaciation.
The abrasive force of ice can be
particularly destructive to vegetation
when ice floes occur in combination with
flooding. Damaged and partially buried
trees in floodplains can be used to re-
construct past flood events (Sigafoos
1964). In Alaska and other areas of
high latitude, the paucity of large
woody vegetation is possibly due to ice
stress and partly due to massive out-
burst floods from glacier dammed lakes
(Post and Mayo 1971). The spectacular
annual floods from 1918-1963 from Lake
George into the Knik River, Alaska were
so regular that the area was designated
as a Natural Landmark by the National
Park Service (Post and Mayo 1971). It
is doubtful if many vascular plants are
able to survive this stress in the ri-
parian zone. However, Brice (1971)
cites an example where balsam poplar
trees survived burial to 2.4 m depth and
later scour that exhumed the trees. The
nearly ubiquitous occurrence of young,
even aged stands of willow and cotton-
wood on point bars and river islands are
evidence of new or renewed environments
created by sediment redistribution
(Lindsey et al. 1961). Thus, the diver-
sity of vegetation both within and among

floodplains is dependent, in part, on
episodes of destructive hydrologic
Ecological Succession
Some reports of ecological succes-
sion in riparian ecosystems have sug-
gested that open water features will
progressively fill in with sediments and
eventually develop into a mixed hardwood
forest community. This is often inter-
preted to mean that the floodplain eco-
system is always approaching some static
and idealized climax condition. This
perception is often in error given the
dynamically changing nature of most ri-
parian ecosystems. Point bars of mi-
grating meanders of streams continually
create new conditions for pioneer com-
munities to become established. If the
stream is in a mode of downcutting
through floodplain alluvium, terraces
will form and become isolated from the
effects of hydroperiod. In the absence
of more frequent flooding, species com-
position will gravitate toward less
flood tolerant species. If the stream
channel is undergoing aggradation, back-
water areas will become less well
drained and be replaced gradually by a
community of species more tolerant to
flooding. On the other hand, increases
in flow and sediment deposition, such as
that experienced by the Atchafalaya
River in the past two decades, may
result in massive amounts of siltation,
a process that leads to better drained
and more elevated conditions (O'Neil et
al. 1975). Catastrophic floods and ice
floes uproot and prune vegetation pro-
viding "open" conditions for species of
plants and animals adapted to rapid
population growth and resource exploi-
tation (Lindsey et al. 1961, Sigafoos
Increasing beaver activity in the
last two decades, particularly in the
bottomland hardwood forests of the
Southeast, have demonstrated the impact
that small changes in hydroperiod can
have on forest communities. It is pro-
bable that elimination of original
beaver populations reduced the hetero-
geneity of floodplain forests and cre-
ated the more uniform forests that are
generally perceived to be the natural










In the absence of hydrologic and
morphic changes in a floodplain,
re is some evidence that secondary
succession will occur more rapidly in
Sodplains than it does in upland
as. For example, Frye and Quinn
79) found that the rate of forest de-
velopment on high floodplain areas of
f Raritan River, New Jersey, occurred
e rapidly than in nearby upland
Sites. The floodplain showed greater
species diversity, equitability, basal
Iea, mean stem diameter, and tree
Thus, changes in riparian ecosys-
Es can be subtle and slow or catas-
ophic and abrupt, but seldom are they
as directional as the classical aq-
Itic-to-terrestrial models of eco-
ical succession would imply. Since a
tiplicity of factors are involved in
community development, the probability
low that these will remain static in
natural floodplain. Some manipula-
ons by humans tend to accelerate
changes while others mute the forces
at are responsible for the maintenance
cyclic phenomena. Since riparian
systems are subjected at different
times to a variety of hydrologic re-
Imes, geomorphic processes, and catas-
ophic forces, generalizations to broad
geographic areas are sometimes difficult
Supply to site specific situations.
imate and biogeography ultimately play
critical role in species composition
of floodplain communities.
scription of Plant Communities
The species composition of some
mon riparian plant communities in the
cited States will be described by geo-
raphical regions (Figure 16). This is
not a classification system for riparian
vegetation but merely an overview of
he dominant species that are most like-
Ty to be encountered in each of the eco-
regions (Teskey and Hinckley 1977a,
1977b, 1978a, 1978b, 1978c; Walters et
1. 1980a, 1980b). That similar species
and genera recur for many regions is not
surprising; it merely confirms that the
Environmental conditions shared by these
ecosystems may be more important than
climatic differences.

1 Southern Forest Region. Bottomland
hardwood forests are located in the

floodplains along major and minor
streams of the Southeast. Vegetation
varies from communities adapted to ex-
tremely long hydroperiods, such as the
water tupelo-baldcypress association, to
oak-hickory communities of "second bot-
tom" forests, some of which may not
flood annually (Figure 17). If the
stream channel has undergone recent re-
orientation, newly formed point bars and
levee deposits may support monospecific
stands of willow (Salix spp.) and mix-
tures of this and cottonwood (Populus
heterophylla), river birch (Betula
nigra) and silver maple (Acer sacchar-
inum). If the river channel remains
stable, species composition may change
to that normally found at higher eleva-
tions because the coarsely textured
sediments drain rapidly after satura-

Areas in deeper depressions that
have long hydroperiods, such as sloughs
and oxbows, will develop water tupelo
(Nyssa aquatica), baldcypress (Taxodium
distichum), and frequently water elm
(Planera aquatica). Communities where
overcup oak (Quercus lyrata) and water
hickory (Carya aquatic) occur are us-
ually among the next most poorly drained
sites. With even shorter hydroperiods,
laurel oak (Q. laurifolia), hackberry
(Celtis laevigata and C. occidentalis),
red maple A. rubrumn, American elm
(Ulmus amerfcana) and green ash
(Fraxinus pennsylvanica) may be common.
Low ridges in the first bottom may be
dominated by sweetgum (Liquidambar
styraciflua) while higher ridges that
have quite short hydroperiods may be
occupied by blackgum (N. sylvatica),
hickories (Carya spp.), and white oak
(Q. alba).
The flats of the second bottom are
likely to have poorer internal drainage
than the high ridges of the first bot-
tom. As a result the species composi-
tion may appear similar to that of the
low ridges of the first bottom. Where
cherrybark oak (Q. falcata var. pago-
daefolia), swamp chestnut oak Q.
michauxii), and water oak (Q. nigra)
occur, hydroperiods are among the
shortest or drainage the best among all
bottomland sites. Live oak (. virgin-
iana) and loblolly pine (Pinus taeda)
are usually confined to the highest
"islands" in floodplain topography.

O Alaska
E Pacific Northwest and Rocky Mountains
SPlains Grasslands
Mediterranean and Western Arid and Semiarid

D Eastern Deciduous Forest
Ei Central Forest
fl Southern Forest
[ Northern Forest


Figure 16. Forest regions of the United States for which riparian plant communities
are described. Terminology after Bailey (1976).







So few virgin bottomland hardwood
stands now exist that cyclic changes in
ancient stands are difficult to recon-
struct. In the Congaree Swamp of South
Carolina, where 11 distinct communities
can be delineated, Gaddy et al. (1975)
suggest that shade tolerant hardwoods
such as laurel oak eventually overtop
sweetgum and other hardwoods for pro-
tracted periods of time. Tree fall is
offered as a mechanism to create canopy
openings so that a mosaic pattern of
communities on the floodplain is main-
Point bar deposition and other new
land forms are initially stocked with

cottonwood and willow. These are suc-
ceeded by silver maple, ash, elm, and
boxelder (A. negundo), a community which
may persist indefinitely in southern
Illinois (Hosner and Minckler 1965).
For more poorly drained sites of the
same region, secondary succession has
been observed to be initiated by button-
bush (Cephalanthus occidentalis), cot-
tonwood, swamp privet (Forestiera acum-
inata), cypress, water tupelo, wiTlow,
green ash, and pumpkin ash (Fraxinus
caroliniana). According to Hosner and
Minckler (1965), further fluvial deposi-
tion or other events that lead to im-
proved drainage will result in replace-
ment of this community by species found

Figure 17. Idealized profile of species

Iod forests. After Wharton (1978).

on successively better drained sites
igure 17).

In narrow bottoms of small streams
where the alluvial soils may be moder-
Iely well drained, cypress and tupelo
nerally are absent. The mixture of
ee species includes those from the
large bottomlands discussed above, from
ist coves, and from mesic uplands
olden 1979). After agricultural aban-
onment there is a distinct trend toward
dominance by light seeded hardwoods
*weetgum, red maple, tulip poplar
iriodendron tulipifera)] that is pro-
vided by mature individuals in uncut
rips left over from incomplete clear-
for agriculture.
The geographic distribution of
ldcypress corresponds approximately
th the distribution of southern flood-
in forests. However, baldcypress is

southeastern bottomland

not an important component of many of
the major floodplains since it tends to
be restricted to the wettest and most
deeply flooded conditions. Some of the
most extensive floodplain areas are
along the lower Mississippi River as
well as large tributaries such as the
Arkansas, Red, Ouachita, Yazoo, and St.
Francis Rivers. Some of the larger
rivers draining in a southerly direction
into the Gulf of Mexico are the Pearl,
Tombigbee, Alabama, Pascagoula, Chat-
tahoochee, Apalachicola, and the Suwanee
Rivers. Those draining from the south
Atlantic coast in a southeasterly direc-
tion include the Altamaha, Ogeechee,
Santee-Cooper, Pee Dee, Cape Fear,
Neuse, and Roanoke Rivers.

Central Forest Region. Bottomland
forests in this region have strong af-
finities with those described for adja-
cent regions (Figure 16). For example,

the studies by Hosner and Minckler
(1965) in southern Illinois have already
been used to characterize the floodplain
vegetation of the Southern Forest
Region. Robertson et al. (1978) show
that the southern floodplain forest type
extends up the Mississippi valley to
southern Illinois and further northward
up the Ohio and Wabash Rivers. To the
east, studies by Lindsey et al. (1961)
conducted on the Wabash River are equal-
ly applicable to the Eastern Deciduous
Forest Region and are discussed below.
The western part of the Central Forest
Region approaches areas where floodplain
forests in the Plains Grassland Region
have been studied intensively (see
below). The admixture of floral com-
ponents from the south, east and west in

the Central Forest Region makes general-
izations about riparian vegetation and
community succession difficult.

In central Illinois, the vegetation
along the Sangamon River illustrates the
rapid transition from floodplains to up-
land forests in species composition,
biomass, and annual biomass accumulation
(Table 18). Silver maple is clearly
dominant in the floodplain, shingle oak
(Quercus imbricaria) and hackberry
(Celtis occidentalis) are codominants in
the transition zone, and white oak dom-
inates the upland community. Total tree
biomass and estimated net biomass accum-
ulation were greatest in the floodplain
followed by the upland and transition
stands. Dutch elm disease and phloem

Table 18. Tree biomass, net annual accumulation, and distribution
among species (%) for a floodplain, transition site and
upland along a stream in Illinois. Biomass percentages
less than 2% are omitted. From Johnson and Bell (1976).

Percent of total biomass
Species Floodplain Transition Upland

Acer saccharinum 73.6 15.7 --
Tleditsia triacanthos 10.9
Fraxinus pennsylvanica 9.4
Platanus occidentaiis 3.6
Euonymus atropur eus -- 9.3 --
Quercus imbricaria 22.3
Carya cordifomis -- 2.2 --
Ce1tis occidentalis -- 27.5
Prunus serotina -- 4.7
Ulmus rubra -- 6.2 --
Ulmus americana -- 5.8 3.3
ercus velutina -- 6.2
uercus alba --- 84.9

Total tree biomass (t/ha) 289 135 227

Estimated net biomass
accumulation of trees 11.5 7.0 10.0
(t/ha -yr)

Frequency of flooding 3-25% 0.5-3% 0.5%


necrosis have contributed to low biomass
of the transition zone by eliminating
all large elm, which probably dominated
the zone prior to 1950 (Johnson and Bell

Eastern Deciduous

Forest Region.

Floodplain forests in this region range
from those located along small to moder-
ate sized streams draining the Appala-
chians to rivers that are relatively
large by the time they pass through the
region. Some of these larger rivers
include the upper Mississippi, Ohio,
Susquehanna, Potomac, and Delaware.
Because of this diversity, generaliza-
tions on riparian vegetation are dif-
ficult to make.

The most intensively studied flood-
plain forests are those on the Wabash
and Tippecanoe Rivers in Indiana (Lind-
sey et al. 1961, Schmelz and Lindsey
1965) which could be included in the
Central Forest Region just discussed
since a few of the study sites are lo-
cated there. First bottoms of the
floodplains tend to be dominated by
black willow (Salix nigra), American
elm, and cottonwod. Second bottoms
that are infrequently flooded are heav-
ily represented by sugar maple (Acer
saccharum), beech (F us grandifolia) ,
American elm, redbud (Cercis canaden-
sis), buckeye (Aesculus gabra) as well
as 16 other species exceeding 10 cm dbh.
In stands on the floodplain of the
Raritan River, New Jersey, Buell and
Wistendahl (1955) mention 14 woody
species. On the inner floodplain where
erosion has produced a series of ridges
and poorly drained sloughs, silver maple
was the dominant tree, followed by
American elm, red maple, and white ash
(Fraxinus americana). In less frequent-
ly inundated and less severely scoured
portions of the floodplain, beech and
tulip poplar were abundant along with
silver maple.

By comparison, the narrow flood-
plains of the Little Tennessee River
system in the Appalachians of western
North Carolina are dominated by river
birch (Wolfe and Pittillo 1977). Other
common species are wild cherry (Prunus
serotina), red maple, black locust
(Robinia pseudo-acacia), and tulip pop-

Successional development on new
sites created by stream migration, as
described for Wissahickon Creek in
southeastern Pennsylvania, may be initi-
ated by silver maple and sycamore
(Platanus occidentalis) following the
herbaceous ragweed cover (Sollers 1973).
This is replaced by a community domina-
ted by white ash, American elm, red
maple, black walnut (Juglans nigra), and
spicebush (Lindera benzoin). With im-
proved drainage, oak-hickory stands
eventually develop. Highest bottoms, or
areas which are inundated only by the
most severe floods, are dominated by
typical upland species (Lindsey et al.
1961). The species composition of
stands at this stage will depend heavily
on the composition of upland forests.
Because of the great diversity in flora
throughout the Eastern Deciduous Forest
Region, there will be a great deal of
geographic variation in the species com-
position of well drained riparian

Northern Forest Region. Riparian
forests in this region have received
little study, possibly because attention
has been diverted to extensive peat bogs
located in the western portion. In com-
parison with the other regions, rivers
tend to be small because many represent
either headwater drainages of the Miss-
issippi River or terminate in the Great
Lakes after flowing a short distance.
The Hudson and Connecticut Rivers in New
England are exceptions.
In blanket peat areas where mineral
rich soil and distinctive water flow
occur, riparian communities develop that
differ from the surrounding low-lying
shrub and sphagnum bog areas. Heinsel-
man (1970) describes these areas with
water flow as rich swamp forest. They
have high densities of northern white
cedar (Thuja occidentalis) which may be
overtopped by black ash (Fraxinus
nigra), larch (Larix laricina) or black
spruce (Picea mariana). Except where
white cedar is dense, speckled alder
(Alnus rugosa) forms a shrub layer.
Speckled alder and black ash usually
disappear in transition from marginal
fen to poorer swamp where water flow is
more sluggish, water is less mineral
rich, and peat is deeper and contains
less inorganic matter. Where more ap-
parent floodplain features exist and

there is little peat accumulation,
American elm may play a larger role,
although black ash is still important
(Janssen 1967).

In the riparian forests along the
Susquehanna, Chemung, and Delaware
Rivers in Applachian Uplands of New
York, there are five characteristic
floodplain features that influence the
species composition of plant communities
(Morris 1977, Morris et al. 1978). They

1. Floodbasins with poorly drained
silts and high organic matter con-
tent that are dominated by wil-
lows, silver maple, cottonwood and
wild cherry.

2. Point bars and stream confluence
areas with well drained silts that
lack willow but have, in addition
to the species listed above,
sycamore and ash.
3. Frequently and destructively
flooded point bars and confluence
areas of sand and silt mixtures
that support black locust, silver
maple, sugar maple, and American
4. Less frequently flooded stable
point bars of coarsely textured
sands that support hickories, in
addition to the two maple species.
5. Seldom flooded Pleistocene ter-
races where pines, oaks, red
maple, and wild cherry dominate.

Plains Grassland Region. As pre-
cipitation decreases from the eastern to
western U.S.A., the isopleth of runoff
reaches 2.5 cm per year in this region,
a value arbitrarily chosen to distin-
guish between humid and arid riparian
ecosystems. Since the natural upland
vegetation is usually savanna, riparian
zones become conspicuous features of the
landscape. Some of the major rivers
that cross this area are the Missouri,
Platte, upper Arkansas, and Canadian

One of the greatest floristic dif-
ferences between arid floodplains and
those of the more eastern regions is the
general absence of oak species, a group

that is particularly abundant in the
bottomland hardwood forests of t
Southern Forest Region and further norl
in the Mississippi valley area of the
Central Forest Region.

Transitions due to moisture al
particularly well illustrated in Okla-
homa where Bruner (1931) distinguished
between the riparian vegetation of thm
eastern, central, and western parts om
the state. Species that occur in more
than one of the parts show decreasing
height in the east to west gradient
(Figure 18). In the east, continuouP
flow of even smaller streams supports
forests rich in species of trees
shrubs, vines and herbs. Baldcypressl
sweetgum, sycamore, river birch, an
black gum are common. Dominants of the
central Oklahoma floodplains, such al
elms, hackberry, walnut, black locust
and honey locust (Gleditsia triacan-



Figure 18. Changes in height and spe-m
cies composition of floodplain forest
stands in a west to east gradient in
Oklahoma. Adapted from Bruner (1931).



those occur also in the east and aug-
ment the species diversity there. In
the arid west, trees are usually rather
widely spaced and neither willows nor
cottonwoods reach the stature that they
attain eastward. Elm and boxelder are
usually found only in valleys or near
streams where the water supply is con-
stant. With only a 2 degree change in
longitude but a 24 cm change in precipi-
tation in central Oklahoma, floodplain
tree species increase from 11 in the
west to 23 in the east (Rice 1965).

In the Missouri River floodplain of
North Dakota where floodplain width
varies from about 1 to 11 km, three
forest types can be distinguished (Fig-
ure 19) (Keammerer et al. 1975, Johnson
et al. 1976). On the lowest and most
frequently flooded area, young cotton-
wood-willow forests have many small
trees 6-12 m tall but have few other
woody species. At higher elevations,
forests consist of older cottonwood
whose tall open canopies overtop bur oak
(Q. macrocarpa) and boxelder. At the

Figure 19. Cross section of the Missouri River in North Dakota showing the distribu-
tion of important tree species. From Keammerer et al. (1975).

highest elevations, floodplain forests
are dominated by green ash, boxelder,
American elm, and bur oak. Canopies are
relatively closed and lack the tall
shrub and sapling layer characteristic
of cottonwood forests.

In the absence of rejuvenation by
flooding due to upstream impoundment in
May 1954, Johnson et al. (1976) state
that cottonwood forests will eventually
disappear since seedbed requirements for
regeneration are lacking. The change
from cottonwood-willow dominance in the
lower floodplain with regulation of
flooding will lead to higher species di-
versity but lower landscape diversity.
Mediterranean and Western Arid For-
est Regions. Some of the major drain-
ages of the arid West are the San Joa-
quin, Sacramento, Salt-Gila, and Rio
Grande-Pecos Rivers. Along these rivers
and their tributary streams, riparian
vegetation provides a striking contrast
to the drought-stressed semidesert and
chaparral of uplands. Species composi-
tion of floodplains includes those that
are confined to more moist areas as well
as those that can survive under drier
upland conditions (Campbell and Green
1968). Differentiation between valley
floor and upland vegetation increases
with increasing drainage area (Zimmer-
mann 1969). Headwaters of intermittent
streams have available little more water
than well drained upland slopes. There
are also dramatic changes in riparian
vegetation with increasing elevation.

Along the Rio Grande between El
Paso and Albuquerque, a distance of
480 km, five vegetation classes can be
described (Campbell and Dick-Peddie
1964). These form a continuum from
south to north with gradual and almost
imperceptible changes between dominant
and subdominant species (Figure 20).

Class 1. In the most xeric class of ri-
parian vegetation, screwbean
(Prosopis pubescens) dominates
and the cover or density is
determined by age of the stand
and moisture availability.

Class 2. Where moisture is greater and
flooding during the growing
season may occur, tamarisk
(also called saltcedar -

Figure 20. Profiles of five vegetation
types along the Rio Grande from El Paso
to Albuquerque. The transition from
class 1 to class 5 is from xeric to more
mesic conditions. Diagrams represent
strips about 25 ft wide and 100 ft long.
From Campbell and Dick-Peddie (1964).

(Tamarix pentandra or T.
chinensis) becomes a competi-
tor with screwbean. In areas
with a high water table and
occasional flooding during the
growing season, tamarisk
thrives at the exclusion of
Class 3. In these dense covers of tam-
arisk, few shrubs and grasses
occur as they do in classes 1
and 2. Class 3 predominates
in the southern sector of the
river and in disturbed areas
to the north.
Class 4. Cottonwood (Populus fremontii) I
stands attain great height
relative to other floodplain
species. Russian olive (Elae-
'ani Goodding wil low (Salix
gooddingii) may become cod"i-T
nants. Mesquite (Proso is
juliflora) occurs occasion y
in the northern localities.


Class 5. These are stands with a dense
overstory of cottonwood and a
separate understory of Russian
olive and Goodding willow.
Tamarisk is found only in dis-
turbed areas.

The introduction of tamarisk and
Russian-olive in the last 50 years has
changed succession and ultimate domin-
ants in some communities. Tamarisk is
in more than 50% of the floodplain plant
communities of the lower Gila River
(Haase 1972).
Elsewhere, Freeman and Dick-Peddie
(1970) noted a trend toward shrub domin-
ance at lower and upper elevations in
southern New Mexico, while trees domin-
ate intermediate elevations. This sup-
ports Zimmermann's (1969) observations
of increasing upland-riparian differen-
tiation with larger drainage area,
though not indefinitely. At the highest
elevations studied (1400 m), species
such as douglas fir (Pseudotsuga
menziesii) and ponderosa pine (Pinus
ponderosa) occur, but are restricted
from distribution at lower elevations
because of high temperatures (Cambell
and Green 1968). The transition of
vegetation across a floodplain in the
Mediterranean Region (Figure 21) illus-
trates the instability of streamside
communities. Forest vegetation develops
only in areas that have not been fre-
quently flooded or that have not under-
gone recent lateral erosion. However,
future generations of cottonwoods are
dependent on the open, moist sand bars
that have resulted from stream insta-

Pacific Northwest and Rocky Moun-
tain Regions. Because of the rugged
local relief of much of these regions,
stream gradients are frequently steep
and channel degradation often predom-
inates. Riparian zones may consist of
narrow interrupted bands along small
streams or as uninterrupted zones in
broad river valleys (Walters et al.
1980b). In mesic sites along streams,
gradients of riparian vegetation are
probably more a result of stand age, as
dictated by time since the last distur-
bance, than the limiting effects of
flooding. Distinct streamside communi-
ties are either a result of new land
being exposed by destructive floods or

the higher local groundwater source
along streams (Fonda 1974). A typical
gradient beginning at streamside for the
western hemlock zone of the Olympic
Mountains is: (1) gravel bars dominated
by Scouler willow (Salix scouleriana);
(2) elevated flats dominated by red
alder (Alnus rubra); with time pioneer
alder gives way to Sitka spruce (Picea
sitchensis), bigleaf maple (Acer macro-
carpum) and black cottonwood (Populus
trichocarpa); and (3) second terraces
occupied typically by Sitka spruce and
western hemlock (Tsuga heterophylla).
This trend is similar to that of the
riparian zone along the McKenzie River
in Oregon (Figure 22). Flooding may
occur annually on the lowest floodplain.

Some species occur only as riparian
species at higher elevations. For ex-
ample, western hemlock and western red-
cedar (Thuja plicata) are restricted
generally to under 550 m but will reach
altitudes of 600 m only along waterways.
In the coastal region of northern Cali-
fornia, redwood (Sequoia sempervirens)
replaces the position of western hem-
lock, Sitka spruce, and Pacific silver
fir (Abies amabilis) found in Oregon and
Washington riparian forests. Not only
is redwood adapted to survive rapid
sedimentation by producing additional
roots, but it is also fire tolerant.

In the Rocky Mountains, species on
wet sites include cottonwood (P. angust-
ifolia), balsam poplar (P. balsamifera),
aspen (P. tremuloides), willows, thin-
leaf alder (Alnus tenuifolia), and berry
bushes (Rubus spp.). At lower eleva-
tions Colorado blue spruce (Picea
pungens) may replace the wet site
species with improved drainage and lack
of disturbance. At even lower elevation
there is a transition to the drier
western arid regions for which the ri-
parian vegetation has already been dis-
Alaska. At least two climatic
zones in Alaska relate to the develop-
ment of riparian vegetation. On the
Arctic slope north of the Brooks Range
where permafrost prevails, willow-alder
communities along streams are in strik-
ing contrast to the shorter tus-
sock-heath tundra and sedge-grass marsh
that surrounds them (Bliss and Cantlon
1957). In contrast, riparian vegetation


River Open Gravel Riparian Valley Oak
Undercutting Plain Bar Forest Forest
blo perceM s r e Jidrt ortory 25M n Uert..,v .-..s5M


sandbar willow cottonwood

herbaceous annuals sierra alder
and perennials box elder

Figure 21.

Profile of vegetation along major rivers in
Fron Conard et al. (1977).

the Sacramento Valley,

in the maritime climate of southeastern
Alaska has similarities in physiognomy
to that of the Pacific Northwest Forest
Region. Changes in floodplain vegeta-
tion fran streamside to upland communi-
ties in Alaska depend largely on whether
the uplands are forested or non-forest-
ed. On the Arctic slope, Sage (1974)
describes three riparian plant communi-
ties. On alluvial deposits that form
gravel and silt bars and islands in
braided streams, usually no vegetation
develops, but in areas not regularly
submerged, Equisetum spp. will develop
as will occasional dwarf willows. Along

small drainage streams, shrub communi-
ties of up to 100 on in height are can-
posed of dwarf birch (Betula nana),
stunted Sitka alder (Alnus _crspa.) nd
willows (Salix pulchra and S. lanata).
A less common community is restricted to
streams and drainage canals in the foot-
hills region which is described as tall
shrub ( 90-100 cm), dominated by felt-
leaved willow (Salix alensis).

In regions where black spruce for-
ests replace the tussock-heath tundra,
more elevated portions of the .loodplain
support stands of balsam poplar which

box elder
sierra alder

valley oak
pipe vult
Wild rost




Black cottonwood


Red alder

Western hemlock

Western redcedar

r i I I I I

o o I C


Figure 22. Cross section of floodplain and
Oregon. From Hawk and Zobel (1974).

are eventually replaced by white spruce
(Picea glauca). Figure 23 illustrates
an ideaTlzed profile for riparian vege-
tation of the Mackenzie River, N.W.T.
(Gill 1972a). In the felt-leaved willow
zone, other species of willow (e.g.,
Salix glauca, S. pulchra, S. arbus-
cuoides)- and Sitka alder may assume im-
portance with increasing stand age.
White spruce appears to assume dominance
only after longer periods without dis-
turbance from flooding. Black spruce
will occur at only the uppermost flood-
plain elevations as described by Drury
(1956) for the upper Kuskokwim River
region just northwest of the Alaska
In southeastern Alaska where a com-
paratively mild marine climate prevails,

terrace communities of the McKenzie River,

Hurd (1971) described the successional
forest stands that followed the reces-
sion of Mendenhall Glacier. Species
composition of the youngest to oldest
communities were quite similar to what
might be expected in a floodplain gra-
dient from streamside to upland. The
youngest stand was dominated by Sitka
alder, with lesser amounts of willow
(Salix sitchensis and S. alaxensis).
Balsam poplar occasionally contributed
to the composition. Later, the poplar
and Sitka spruce dominated. The oldest
stand was a western hemlock--Sitka
spruce mixture which is common through-
out the coastal uplands of southeast
Alaska. It appears that successional
stages after glacial retreat result in
similar gradients in species composition

Incense cedar

Bigleaf maple


Grand fir

r 9





I !

0 50 100 150 200 250 300 350 400 450 500
Horizontal distance (feet)

Figure 23. Zonation of plant communities along an arctic stream. From Gill (1972a).

as the time since disturbance along


Although all ecosystems produce and
respire organic matter, cycle nutrients,
and carry on other processes just des-
cribed, floodplain ecosystems are unique
because these processes are superimposed
on the historical and contemporary work
performed by flowing water. Few other
land forms change as rapidly as flood-
plains where the channel adjusts its
capacity to the natural episodes of
large, infrequent floods and variations
in sediment load. Diverse topographic
features such as oxbow lakes, meander
scrolls, and abandoned channels are
relicts of this work. Although topo-
graphic relief is muted in comparison to
many upland landforms, the presence of
surface water and natural flood events
impose strong control over the microen-
vironments to which plant and animal
communities adapt.

There is
these unique
their related


information on
features and
properties to

predict changes that will occur when ri-
parian ecosystems are altered by manage-
ment of water delivery patterns and by
other human intrusions. These altera-
tions can be perceived as stresses which
change the pattern of energy flow and
the movement of materials to and from
riparian ecosystems.

To better understand the way in
which these alterations interact with
natural ecosystem components, a simpli-
fied model of energy flow is used to
identify major ecosystem processes of
riparian ecosystems (Figure 24). Major
sources of energy and materials are
shown in the circles on the upper left
hand side of the figure -- water, sedi-
ments, nutrients, wind, and sun. Other
symbols represent storage of material
and energy within the ecosystem that are
supplied by the outside sources. Ex-
changes among these storage and inter-
actions with outside sources are indi-
cated by connecting lines of flow.
Where two flows interact, whereby one
flow augments another in a multiplica-
tive fashion, a large arrow is used to
indicate an acceleration of flow. For
example the kinetic energy source of
water flow interacts with sediments and
nutrients to deliver them to riparian



Figure 24. Major flows of
(1978). Refer to text for

energy in a floodplain ecosystem.

Adapted from Brown et al.



Circles represent sources of energy
ind mnterieli supplied
to the ecosystem.
Storage compartment may represent
meergy or material such s water.
sediment. organic nutter. and nutrints.

Thick arrow indicates that two flows
nteract such that one augments the
their in a multiplicitive fashion.

expended energy that is lost without
Sbeng processed fuither such as
disspltion of heat in respiration.

Hu4tagn npaOtseuts onsumr community
or the processing of energy by animals.

Symbol nprmnts primary producers
or the perocsing of nerfly by plants.





ecosystems. (Many subtle, yet important
interactions and feedbacks in this model
have been omitted for simplicity.)
Smaller, downward pointing arrows are
energy sinks that represent necessary
losses of thermal energy, such as
through respiration, for useful work to
occur. When disorder occurs in the
flows of energy among ecosystem compon-
ents, or these components undergo
stresses that prevent proper function-
ing, excessive and wasteful losses of
expended energy to these sinks may
To the right side of the figure
disruptive energy sources are indicated,
again as circles. These represent cate-
gories of alteration or impact that
drain energy away from the stabilizing
flows that maintain ecosystem integrity.
The three groups of alterations--water
delivery and geomorphology, physiolog-
ical stress, and biomass removal--all
interact at different places in the left
to right hand flow of energy.
The closer the alterations interact
with the sources of energy, the greater
the impact on subsequent flows to the
right. Thus, water delivery and geo-
morphic changes will be expressed at all
levels of ecosystem organization. In
contrast, biomass removal will have far
less effect. If the alterations result
in changes of flows close to the primary
energy sources, recovery to the original
unaltered condition will be slow if re-
covery is even possible. Energy drains
more distant from primary energy sources
are less disruptive, and the ecosystem
has a high probability of recovering to
its original condition.
Most real world alterations of ri-
parian ecosystems and associated stream
channels correspond to one or more of
the three energy drains in Figure 24.
If the alteration can be interpreted as
changes in water delivery or geomor-
phology, severe and long lasting changes
can be expected from which there is only
a low probability of recovery to the
original ecosystem. Physiological
stress and biomass removal, depending on
the magnitude and frequency at which
they are imposed, are more likely to be
repaired through natural ecosystem pro-
cesses (succession) or through mitiga-
tion techniques.

Mitigation of damage caused by
water delivery and geomorphic changes is
extremely costly and time consuming.
The costs to restore ecosystems to their
original condition after damage may be
one indicator of the original value of
the work that the ecosystem supplies at
no cost to society if it is allowed to
function naturally.
Exarlples of riparian ecosystem
alteration and their relationship to the
model are outlined in Table 19 and will
be explained in the following discus-
sion. Specific effects may differ de-
pending on individual peculiarities of
the ecosystem undergoing alteration as
well as the nature and severity of the
Stream Channelization
One of the purposes of stream
channelization is to improve the down-
stream conveyance of water. This is
usually achieved by deepening, widening,
and straightening the channel. It re-
presents initially a disruptive geomor-
phic change that would never occur under
natural conditions regardless of the
time frame. In combination with the
effect on water delivery, all essential
sources of energy, with the exception of
sunlight, are either completely elimi-
nated or greatly diverted. Delivery of
water, nutrients, and sediments to the
floodplain ecosystem no longer occurs
through stream channel-floodplain ex-
changes. Absence of the natural hydro-
period and water availability imposes
severe physiological stress on plant and
animal communities.
Increases in channel gradient by
reducing sinuosity will result in
sharper pulses in flow and concentrate
the kinetic energy of flowing water in
time and space. This may initiate ero-
sion and cause gullying, depending on
soil structure and stream gradient, and
result in downstream transport of soil
and nutrients. In small, lower order
stream channels, removal of streamside
vegetation precludes influxes of leaf
litter, the principal energy source for
instream animal communities. Transfor-
mation to a more autotrophically based
food web that might be expected upon
removal of shade will be of little con-
sequence if benthic structure of the




















Table 19. Examples of riparian ecosystem alteration and
shown in Figure 24. Alterations are listed in
impacts on riparian ecosystems, and in inverse
cessation of perturbations.

Riparian ecosystem component
and alterations Structure

Stream channelization Channel depth increased

Containment of streamflow
and channel constriction

Impoundments and diversions:
Upstream in flooded


Introduction of toxic


Heavy metals

Timber harvest followed
by agriculture

Grazing by livestock

Timber harvest followed
by silviculture

Hunting and fishing

Channel gradient increased
and sinuosity decreased

Restricted floodplain

Biomass and water depth

Channel depth increased

Plant biomass

Animal biomass

Plant and animal biomass

Standing stocks of plant
biomass, nutrients, and
streambank deterioration

Plant age structure

Streambank deterioration

Standing stocks of plant
biomass and nutrients

Standing stocks of animal

their relationship to categories of alteration
approximate direct order to the severity of their
order to the time required for recovery following



Decreases in floodplain-channel
exchanges of water, nutrients
and organisms

Sharper pulses in flow, in-
creased effectiveness of
material transport, loss of

Increased channel scour and
greater deposition in narrowed

Primary productivity, nutrient
cycling, upstream-downstream
exchange of organisms

Sediment supply decreased,
scour continues

Primary productivity, trophic
structure, & nutrient cycling

Trophic structure

Primary productivity,
trophic structure, and
nutrient cycling

Decreased primary productiv-
ity, increased nutrient export,
and increased sediment supply
and transport

Primary productivity and
biomass accumulation

Increased sediment supply
and transport

Temporarily decreased trans-
and primary productivity

Grazing and predation

Category of alteration

Water delivery and

Water delivery and

Water delivery and

Water delivery and

Water delivery

Physiological stress

Physiological stress

Physiological stress

Biomass removal and

Biomass removal

Geomorphol ogy

Biomass removal

Biomass removal

stream channel deteriorates and if
greater pulses of water flow and turbid-
ity prevent establishment of primary
producers. Mitigation of these damages
is clearly not possible because the
floodplain has been deprived of the
sources of energy that make it unique
from upland ecosystems.

Snagging, or the removal of woody
obstructions to improve water convey-
ance, has been suggested in the SCS/FWS
Channel Modification Guidelines (44 FR
76299, December 1979) as a preferred
alternative to more severe forms of
channel modification. However, removal
of woody substrates likely causes sig-
nificant declines in overall animal pro-
ductivity, animal diversity, and capac-
ity of the stream to assimilate parti-
culate organic matter (Benke et al.
1979). In a southeastern blackwater
stream, snags were the most productive
habitat available for invertebrates and
many fish species are highly dependent
on this food source.

Containment of Streamflow and
Channel Constriction

Again, geomorphic and water deli-
very are the principal changes in the
natural functioning of the riparian eco-
system when streamflow is contained.
Restricted floodplain storage by levee
containment increases water velocity in
the stream channel and may result in
scour and downcutting. However, the
deposition of sediments, which original-
ly occurred in the floodplain, will be
concentrated between levees and more
rapidly obliterate remaining topographic
features of the floodplain. Large scale
examples of this are occurring along the
Mississippi River (Belt 1975) and its
distributary, the Atchafalaya River (van
Beek 1979). The tendency for these
large rivers to build elevated channels
and levees accelerates when floodplains
are no longer available as areas of
sedimentation. Floodplains outside the
levees will be deprived of materials in
the same way channelization alters ex-
changes between the stream channel and
the floodplain.

Even in the absence of levees,
dikes and jetties contribute to the con-
tainment of streamflow and channel con-
striction. Other, more subtle, human

activities have resulted in a general
tendency toward stabilization of channel
meandering, narrowing of channel width,
and swifter currents. Not only are fun-
damental geomorphic and water delivery
processes affected, but shifts in food
chains can be deduced from activities
that convert broad, sometimes braided
and often intermittent streams into
relatively narrow and swift channels.
If no further alteration occurred, there
would be an increase in riparian vegeta-
tion at the expense of open water; how-
ever other human uses such as agricul-
ture may supplant natural floodplain

For example, the Platte Rivers in
Nebraska and Colorado have undergone a
reduction in width by 80-95% during the
past 100 years (Williams 1978). The
amount of floodplain vegetation has in-
creased considerably at the expense of
aquatic surface area and vegetated
islands. Nadler (1978) attributed this
trend to irrigation practices that pro-
duce more stable flow regimes. Irriga-
tion water, which is withdrawn from the
river and reduces its sediment load,
raises water tables and produces more
uniform streamflow. As a result, ripar-
ian vegetation becomes more dense and
may invade channels during drought
years. The result has been the transi-
tion from relatively straight, wide, and
intermittent streams to narrow and swift
channels with more sinuous configura-

Likewise, in a 830 km reach of the
Missouri River, surface area of the
river was reduced to half (24,618 ha) of
the original area between 1879 and 1972
(Funk and Robinson 1974). Islands,
sandbars, snags, and marshes have been
virtually eliminated (Figure 25a). Con-
struction of dikes and revetments have
been responsible for the surface area
lost, but levees, mainstem dams, and
tributary reservoirs also contributed to
change in channel configuration. Much
of the recently accreted floodplain has
been put into cultivation of crops. The
overall result has been a narrower,
swifter and deeper channel accompanied
by a reduction in habitat diversity,
elimination of some species of fish, and
precipitous declines in commerical
catches of fish.

Figure 25. Changes in channel morphol-
ogy of (A) Missouri River between 1379
and 1954, and (B) Gila River from 1914
to 1962. After Funk and Robinson (1974)
and Turner (1974).

In a similar manner, the Gila
River, Arizona has undergone a gradual
narrowing since 1914 (Fig. 25b). Some of
the obvious reasons for changes in con-
figuration include changes in stream
discharge and periodicity due to water
impoundment. Other less understood
changes involve modification of the ri-
parian plant community by increased fire
frequency and introduction of exotic
species like saltcedar (Turner 1974).
Instream primary and secondary produc-
tivity is probably reduced in greater
proportion than surface area because of
swifter abrasive currents and reduced
penetration of light under the more tur-
bulent and turbid conditions. Instead
of the energy of flowing water being
dissipated over broader areas by shift-
ing sand bars and eroding banks, the
energy is concentrated in the channel
resulting in scour that disorders stream

Impoundments and Diversions

In the inundated reaches upstream
from impoundments, changes from lotic to
lentic conditions are so extreme that
they are too obvious to describe in de-
tail. However, the transformation can
be perceived as a change from a struc-

turally complex riparian ecosystem to a
relatively simple aquatic system. Al-
though the ecological attributes of the
two systems are quite different and dif-
ficult to compare objectively, reser-
voirs usually have construction and
maintenance costs (water weed control,
dam maintenance, etc.) that must be
offset by benefits if society is to gain
from the transformation. In comparison,
floodplain ecosystems require only pro-
tection for them to yield consumables
such as flood water storage, water
quality maintenance, and products from
fish, wildlife, and timber.

Water delivery patterns are altered
downstream from impoundments and the
sediment supply is held mostly in the
reservoir. Other well documented ef-
fects in reservoir regulated streams are
changes in water chemistry (Hannan 1979,
Krenkel et al. 1979, ), effects on chan-
nel morphology (Simons et al. 1975,
Simons 1979), and temperature effects
(Fraley 1979). Although direct effects
on riparian ecosystems may not be as
acute as with other alterations, secon-
dary impacts such as changes in land use
to agricultural crop production are fre-
qently the result. Even if the flood-
plain is not subjected to land use
change, the decrease in sediment supply
below the impoundment will result in
channel scouring and greatly reduce or
eliminate sediment delivery to the
floodplain. For example, the Shasta Dam
on the Sacramento River, California, has
reduced the sediment supply below the
dam and initiated a phase of degradation
(California Department of Water Re-
sources 1979). Erosional-depositional
processes currently in effect have
lowered the channel by 0.3 m at a dis-
tance of 250 km below the dam and are
reducing many high terrace riparian
lands to lower terrace gravel bars.

Changes in the hydrologic regime
also have been dramatic for the Colorado
River in the Grand Canyon (Turner and
Karpiscak 1980). Before Glen Canyon Dam
was built, seasonal variations in dis-
charge were large and daily variations
were low (Figure 26). The variations
were reversed after the dam began oper-
ating in 1963. The result has been
establishment of riparian vegetation
along the river, especially exotic
species such as saltcedar and Russian

I *

2- M




Figure 26. Daily variation in river
stage for the Colorado River at Lees
Ferry during water year 1939 (A) and
water year 1973 (B). From Turner and
Karpiscak (1980).

olive. Stream regulation has had an
enormous detrimental impact on special-
ized fishes that have a narrow tempera-
ture tolerance (Holden 1979).
For decades water diversion and
withdrawal for irrigation in the arid
West has resulted in problems with salt
balance in the Rio Grande (Wilcox 1955),
the lower Colorado, and other major
streams (Skogerboe 1973). Under natural
conditions, floods occasionally rejuve-
nate floodplain soils (Babcock and
Cushing 1942) by leaching salts and re-
ducing salinity levels. With flood con-
trol and the increased evapotranspira-
tion that results from irrigation, there
is an increase in soil salinity, parti-
cularly during periods of low precipi-
tation. Choices of agricultural crops
must necessarily narrow to those toler-
ant of higher salinity until the problem
becomes so acute that agriculture must
be abandoned.

Introduction of Toxic Compounds

Herbicides, insecticides, and toxic
metals, introduced directly to riparian
ecosystems or from the stream by over-
bank flooding can be regarded as a
source of physiological stress to or-
ganisms. If accompanying water delivery
and geomorphic changes are not imposed,
the primary energy sources are main-
tained and recovery is possible if dis-
ruptions are not chronic. In fact, the
capacity of water saturated floodplains
to immobilize heavy metals in organic
rich sediments and to retain pesticides
until they are detoxified (Pionke and
Chesters 1973) can be a useful and im-
portant service for maintaining water
quality (Schlesinger 1979). With up-
lands being managed and utilized at
greater intensity, spills, leaks, and
appearance of man-made products in run-
off are occurring more frequently. The
capacity of floodplains for processing
these residues and the extent to which
they are effective seasonally are not
known. However, alterations that accel-
erate water conveyance will reduce the
capacity of floodplains to perform this
Grazing by Livestock
Grazing effectively removes plant
biomass, alters plant population age
structure, and may change the species
composition of plant communities. These
effects are not restricted to riparian
ecosystems; where rangeland has deteri-
orated under heavy grazing, riparian
vegetation also will be under greater
grazing pressure. Cattle spend more time
in riparian ecosystems than they do in
adjacent uplands in the arid west
(Martin 1979). Reproduction of tree
populations are affected most by heavy
browsing on young plants (Dahlem 1979).
Without population recruitment of young
trees, riparian forests develop unstable
age structure and are biased toward
large, older trees. Along many streams
of arid regions, small stands of relict
cottonwood and sycamore are the only
forest vegetation remaining. Primary
productivity and biomass accumulation of
forests necessarily decline under these
conditions. Owing to the importance of







structural complexity of riparian for-
ests in arid regions (Figure 8), region-
wide abundances of vertebrates and in-
vertebrates are dependent on the main-
tenance of these ecosystems.

Recovery of arid riparian forests
from plant biomass removal in many areas
is prevented by livestock grazing.
Moreover, cottonwood, a major component
of these forests, requires special con-
ditions for regeneration. Barren and
moist sandbars, which are abundant in
shifting, unstable floodplain streams,
provide an ideal seedbed for regenera-
tion of cottonwood. Stream channel con-
striction and flood control considerably
reduce conditions for germination. Cot-
tonwood is particularly well adapted to
colonization following large floods that
may obliterate streamside forests.
Secondary effects of overgrazing
may result in increased runoff from up-
lands and reduction in the stability of
stream channels. Restoration of ripar-
ian vegetation may require not only
reducing or eliminating grazing, but
structural measures to control erosion
as well. Reduction of livestock graz-
ing, construction of check dams, and
other rehabilitation procedures can be
successful in retarding soil erosion and
rapid channel downcutting. A rangeland
restoration study in Colorado demon-
strated that streams were transformed
from intermittent to perennial flow
regimes when restoration procedures
resulted in retention of alluvial fill
and re-establishment of riparian vegeta-
tion (Heede 1971). In this situation an
increase in water storage capacity of
the newly acquired alluvial fill out-
weighed water losses that may have re-
sulted from evapotranspiration by ripar-
ian vegetation. Other management op-
tions are available for improving ri-
parian vegetation and instream condi-
tions (Martin 1979, Platts 1979).
Furthermore, fish populations improve
rapidly when cattle are excluded (Keller
et al. 1979, Van Velson 1979).

Timber Harvest

Forest management practices can
range from the selective removal of
mature trees to the replacement of
natural forest stands by intensive sil-

viculture. Transformation to intensive
agriculture may follow timber harvest.
The capacity of riparian ecosystems to
recover from plant biomass removal will
depend partly on the extent to which
propagules of native species are avail-
able for succession, provided that
drainage patterns and hydroperiod are
not seriously altered. Clearcutting
will cause temporary decreases in evapo-
transpiration, primary productivity, and
probably the capacity to recycle nutri-
ents, whereas selective cutting will
have negligible effects on these proces-
ses. However, in bottomland hardwood
forests of the Southeast, selective
cutting has deteriorated the quality of
wood products ("highgrading") (Maki et
al. 1980) and clearcutting is a pre-
ferred practice by foresters (Putnam et
al. 1960).

In the wettest portions of south-
eastern river swamps, regeneration of
water tupelo by stump sprouting may
result in rapid growth and recovery of
plant biomass. This is possible because
the root stock is maintained alive and
there is less need for the vegetation to
initially divert large amounts of photo-
synthate to belowground parts for
growth. In mixed hardwood floodplain
forests where regeneration may occur by
seeding, the species composition of the
forest will depend on a number of fac-
tors including available seed source,
conditions for germination, competition
among young plants, and light avail-
ability. Ecological succession in
bottomland hardwood forests is poorly
Conversion of forested floodplain
ecosystems to agriculture results in a
severe and more or less permanent reduc-
tion in plant biomass as long as the
affected area is farmed. For example,
aboveground biomass of a cypress-tupelo
stand in Louisiana is 38 kg/m (Conner
and Day 1976) whereas a corn crop ranges
from near zero in the winter to only 0.4
kg/m at peak biomass (Odum 1971). Sec-
ondary practices of flood control and
drainage are more seriously damaging to
ecosystem function than that of biomass
removal. Water delivery changes are in-
volved (Figure 24); consequently there
is little opportunity for ecosystem re-

Pure stands of saltcedar have re-
placed many native cottonwood-willow
communities in arid regions. Saltcedar
is an aggressive competitor and extreme-
ly well adapted to floodplain condi-
tions. It has been successful in domi-
nating large sectors of rivers where
cottonwood-willow communities existed.
Harvest of the original timber, in-
creased frequency of fire, stream chan-
nel constriction, and flood control are
all alterations induced by humans that
have accelerated the dispersal of salt-
cedar in arid riparian ecosystems
(Turner 1974, Everitt 1980).
In efforts to divert water from
maintenance of riparian ecosystems to
use in agriculture, phreatophyte eradi-
cation projects have intentionally re-
moved biomass. There is a great deal of
literature that unequivocally advocates
the benefits of water yield from streams
by means of removing riparian vegetation
(Gatewood et al. 1950, Turner and
Skibitzke 1952, Bowie et al. 1968,
Culler et al. 1970) and most focuses on
an intensively studied reach of the Gila
River in Arizona. Even if the values of
riparian vegetation for organic matter
production, shading and temperature
amelioration of surface water, and
habitat structure (Campbell 1970) are
completely disregarded, extrapolating
the results to unstudied ecosystems is
not warranted because findings vary
greatly under the same climatic circum-
stances (Horton 1972). As early as
1963, it was pointed out that streamflow
augmentation could only be expected
through manipulation of riparian vegeta-
tion under very specific conditions.
These are areas in which (1) the water
supply is adequate to exceed evapotrans-
piration losses after treatment, (2) the
water table or zone of saturation is
within reach of woodland-riparian vege-
tation, and (3) canyon bottom soils
overlaying the water table are of suffi-
cient extent and depth to permit reduc-
tion in evapotranspiration if deep
rooted vegetation is eliminated (Rowe
Even if vegetation is removed, it
can be considered only a temporary con-
dition (Culler 1970) because revegeta-
tion is a predictable consequence of

ecological succession. However, con-
tinual removal of vegetation does not
ensure water salvage. When windspeeds
and temperatures are extremely high,
evapotranspiration from saltcedar dimin-
ishes due to stomatal closure, even
though water is freely available (van
Hylckama 1980). Estimates of salvage-
able water based on the assumption that
riparian vegetation always uses water at
a potential rate may at times be far too
large. The long-term effect of these
disruptive intrusions may be more severe
than just affecting animal biomass and
primary and secondary productivity. Ri-
parian ecosystems of the arid West,
partly because of widespread deteriora-
tion of upland ecosystems, may be ex-
tremely important to the survival of
many species throughout the region.
Hunting and Fishing
Removal of animal biomass is an
alteration that has an excellent oppor-
tunity for recovery as long as the habi-
tat structure and life support system of
the animals are maintained by the prin-
cipal flows of energy. Special consid-
erations must be given to providing suf-
ficient contiguous ecosystem area if
viable populations of predators are to
be maintained. Peculiarities of endan-
gered and threatened species must be
given special attention in addition to
the maintenance of riparian ecosystem
structure and function.
Riparian ecosystems are frequently
managed for game species so that addi-
tional reproductive success of selected
animal populations will support higher
rates of harvesting. When management
techniques cause water delivery or geo-
morphic changes, the primary energy
sources of the ecosystem are being di-
verted. Both short- and long-term
changes of ecosystem function and struc-
ture are predictable under these condi-
tions and they may result in suboptimal
levels of natural function and work.
Since some wildlife management practices
are oriented toward a few game species,
little consideration is given to values
and functions of the ecosystem that sup-
port a high diversity of wildlife





















Biologists, naturalists, and other
'outdoor enthusiasts have long recognized
the high value of streams and riparian
ecosystems to fish and wildlife. How-
ever, quantitative information in sup-
port of these observations has surfaced
only recently. Research conducted in
various areas of the country has con-
firmed that riparian ecosystems are con-
sistently very important to fish and
wildlife on local, regional, and na-
tional scales.

Riparian ecosystems differ from up-
land ecosystems in terms of plant com-
munity type, hydrologic features, soil
type, and topography. These attributes,
along with more subtle environmental
parameters, largely determine the poten-
tial abundance of animal populations at
any particular site. This chapter: (1)
discusses the ecological attributes of
riparian ecosystems that are most impor-
tant to fish and wildlife; (2) presents
a general characterization of riparian
wildlife communities; and (3) examines
the overall significance of riparian
ecosystems to.fish and wildlife.


Undisturbed riparian ecosystems
normally provide abundant food, cover,
and water, and often contain some spe-
cial ecological features or combination
of features that are not found in upland
areas (see Chapter 3). Consequently,
riparian ecosystems are extremely pro-
ductive, and have diverse habitat values
for fish and wildlife.

The importance of riparian ecosys-
tems can be attributed to specific bio-

logical and physical features, includ-

1. Predominance of woody plant com-

2. Presence of surface water and
abundant soil moisture;

3. Close proximity of diverse struc-
tural features (live and dead
vegetation, water bodies, nonve-
getated substrates), resulting in
extensive edge and structurally
heterogeneous wildlife habitats;

4. Distribution in long corridors
that provide protective pathways
for migrations and movements be-
tween habitats.

Most floodplain ecosystems have some
or all of these common attributes that
distinguish them from other ecosystems.
The relationships of these basic fea-
tures to fish and wildlife are described

Predominance of Woody Plant Communities

Riparian areas often support a vari-
ety of plant communities, ranging from
mature hardwood forests to alder swamps
and cattail marshes. However, woody
vegetation predominates in most riparian
environments, while herbaceous riparian
communities are more limited in extent.
Woody riparian communities offer a
variety of wildlife habitat values, and
are very critical to animal populations
where extensive forests are lacking. In
grasslands, rangelands, and intensively
farmed regions of the U.S.A., woody
vegetation along waterways is essential

---I- ---L-LIL~-LY--- --I~L-C'L- --------

for the survival of many fish and wild-
life populations, especially for-
est-dwelling species (Michny et al.
1975, Boerr and Schmidly 1977, Korte and
Fredrickson 1977, Best et al. 1978,
Heller 1978, Thomas et al. 1979b). In
areas where shrub communities and for-
ests have been cleared for agriculture,
woody riparian vegetation may be the
only available cover for.farmland edge
species such as pheasant, dove, and
cottontail (Leite 1972).

Woody vegetation is a primary
structural feature of riparian wildlife
communities. Trees and shrubs are re-
quired for roosting or foraging by most
riparian bird species, ranging from bald
eagle to great blue heron to a variety
of small songbirds (Heller 1978, Swift
1980). Mammals such as white-tailed
deer, beaver, squirrels, and cottontail
are dependent on woody plant materials
for shelter and as part of their diet.
Woody vegetation on the floodplain in-
creases humidity and provides shade that
is attractive to some wildlife species.
The attraction of deer, elk, and other
wild and domestic ungulates to riparian
areas is a result of the thermal cover
and microclimate produced by that vege-
tation (Thomas et al. 1979b).

Dead woody vegetation is an im-
portant component of wildlife habitat in
most forest ecosystems, including ripar-
ian woodlands (Noble and Hamilton 1975,
Conner 1978, Thomas et al. 1979a, Maser
et al. 1980). Standing dead trees or
"snags", which are used extensively by
wildlife, are especially abundant in
beaver ponds (Hair et al. 1978) and
where elms occur (Blem and Blem 1975).
Snags provide nest sites for cav-
ity-dwelling birds, den trees for small
and medium sized mammals, and feeding or
perching sites for many species. Fallen
logs function as cover for wildlife and
as feeding and reproduction sites, but
may hinder movement of larger mammals if
there is too much downed timber. Dead
woody material that is partially sub-
merged in water provides excellent
habitat for aquatic, amphibious and cer-
tain terrestrial species, although too
many logs in a stream channel can act as
a barrier to fish passage (Marzolf 1978,
Maser et al. 1980).

To varying degrees, aquatic inver-
tebrate and fish communities are influ-
enced by streamside vegetation (Figure
27). Roots of woody vegetation along
streams are especially important in bank
stabilization and may provide cover for
fish and other aquatic animals. Leaf
litter from riparian vegetation provides
a substantial proportion of food for
aquatic invertebrates, particularly in
small streams, which in turn constitute
a significant proportion of many fish
species' diets (Table 20). Terrestrial
invertebrates of the riparian zone are
often found in streams and become impor-
tant in the diet of fishes there. The
shading of streams by woody riparian
vegetation has a dramatic effect on
water temperature and the productivity
of the aquatic invertebrate community.
In all but the coldest regions of the
U.S.A., riparian vegetation has a posi-
tive influence on salmonid fishes
(Meehan 1970, Hunt 1979, Chapman and
Knudsen 1980).

Presence of Surface Water and

Abundant Soil Moisture

The mere presence of surface water
is a requirement of many wildlife
species, as an environment for feeding
(e.g., waterfowl, fish-eating birds),
reproduction (e.g., amphibians), travel
(e.g., beaver, muskrats), and escape
(e.g., amphibians, muskrat, and beaver).
Consequently, many species are rarely
found far from water (Figure 28). Water
bodies add a dimension of habitat to
riparian ecosystems (MacArthur 1964,
Hair et al. 1978); increasing the abund-
ance and variety of water bodies contri-
butes to wildlife productivity and di-
versity (Beidleman 1954, Hardin 1975,
Fredrickson 1978).

Seasonal inundation of floodplains
increases potential availability of food
and breeding habitat for some stream
fishes. During annual high water, some
species migrate laterally into flood-
plains to feed among tree roots (e.g.,
catfish, centrarchids), or to spawn on
the inundated forest floor (e.g., blue-
back herring), returning to the channel
when flows slacken and water levels drop
(Figure 29) (Wharton and Brinson 1978,
Welcomme 1979). At the same time,

f tt
2 Li z o
: 2 i Z
(L S
.-o =

m Figure 27.
From Meehan

Functions of
et al. (1977).

riparian vegetation as they relate to aquatic ecosystems.

flooding facilitates transport of or-
ganic detritus to the channel and down-
stream (Welcomme 1979).

Even in the absence of surface
water, soil moisture (during the growing
season at least) may be ultimately re-
sponsible for major differences in
species composition and productivity
between riparian and upland ecosystems.
Abundance and diversity of various song-
bird and small mammal species are re-
lated to soil moisture of plant com-
munities (Johnston and Odum 1956,
Armstrong 1977, Miller and Getz 1977,
Smith 1977, Swift 1980). Several small
mammal species are physiologically
restricted in distribution to areas with
high soil moisture, while others that
use underground runways cannot inhabit
wet sites (Miller and Getz 1977). Moist
soils are required by some bird species
for feeding (e.g., woodcock) and for

GRC AT BLUE HERON' I i km ----, o 4 km
GR[EN ( EtRO --- 4-- .
WOOD DUCK' oo-----------< O m
SALO EAGLE" ---------------- o i
DIPPER .-------.--
ElOc SPECIEs ---,
WOODLAND IOS ----------
BtAVER' --. ------- I
MUSKRAT' '--.-----
RACCOON' lo 400 ..' -1 To I- -
RED BACKED VOLE' ----- ---- ----
-I. .. 1 1 I

0 0 40 4 O 10 0 oo00 140 160 IT0
Figure 28. Distribution of riparian
wildlife species in relation to streams.
Sources: (a) Mathisen and Richards
(1978); (b) Nengel (1965); (c) McGilvrey
(1968); (d) Steenhof (1978); (e) White
(1953), Cornwell (1963); (f) Simpson
(1969); (g) Bradt (1947), Hall (1960);
(h) Liers (1951); (i) Errington (1937);
SSchladweiler and Storm (1959); (k)
Schwartz and Schwartz (1959); (1) Miller
and Getz (1977); (m) Iverson and Turner
(1973); (n) Handley (1948); (o) Organ



above ground canopy & stems 1. Shade-controls temperature &
above channel in stream primary production
2. Source of large & fine plant
3. Source of terrestrial insects

in channel large debris 1. Control routing of water and
derived from sediment
riparian veg 2. Shape habitat-pools, riffles.
3. Substrate for biological

streambanks roots 1. Increase bank stability
2. Create overhanging

floodplain stems & low 1. Retard movement of
lying canopy sediment, water and floated
organic debris in flood flows

Table 20. Importance of aquatic and terrestrial invertebrates in diets of North American stream fishes.
Insect orders represent aquatic life stages unless indicated otherwise.

Stream location
Species and size Stomach contents

Mountain whitefish a Sheep R., Alberta; Ephemeroptera, Trichoptera, Plecoptera
(Prosopium williamsoni) 16.5 m width and Diptera made up 89% of the items
overall. For larger size classes
( 300 mm), contents were up to 40%
of total.
Coho salmon b Whitefish R. estuary; Ephemeroptera most important by weight
(Oncorhynchus kisutch) L. Michigan tributary for yearling fish.
Cutthroat trout Logan R., Utah Ephemeroptera, Trichoptera, and Diptera
(Salmo clarki)' were major volume of food items.
Brook trout (Salvelinus Four streams in Ephemeroptera, Coleoptera, Trichoptera,
fontinalis) and cut- northern Idaho; Diptera and Plecoptera comprised 92%
throat trout (Salmo 5-8 m max. width of items for 2 species. Terrestrial insects
clarki) insignificant.
Brook trout e Unnamed stream, Diptera, Trichoptera, Ephemeroptea, and
(Salvelinus fontinalis) Vermont; 5 m wide Plecoptera major items except during
June and Aug.-Nov. when terrestrial
beetles, grasshoppers, and ants
Black sculpin Upper S. Fork of Ephemeroptera, Diptera, Trichoptera,
(Cottus bailey) Holston R., Virginia; Coleoptera, and Plecoptera comprised
9.4 m wide at low 99% of total food items.
Northern mottled Rock Cr., Oregon; Ephemeroptera and Diptera were major
sculpin (Cottus b. 6 m wide food items of both species.
bairdi) and barred
darter (Etheostoma
f. flabellare)

aThompson & Davies (1976); beck (1974); CFleener
Novak & Estes (1974); Pasch & Lyford (1972).

preferred nesting habitats of others
(e.g., prothonotary warbler). General-
ly, moister sites are more productive of
wildlife, because foods (vegetation,
seeds, insects) are presumably more
abundant there, and vegetation structure
is more favorable to a greater number of
species (Odum 1950, Gaines 1974, Curtis
and Ripley 1975, Hardin 1975, Dickson
1978, Swift 1980).

Diversity and Interspersion of
Habitat Features

Within riparian ecosystems, there
are a great variety of habitat features

(1951); dGriffith (1974); eLord (1933);

that are used by a relatively large num-
ber of fish and wildlife species. Ri-
parian areas are able to support dense
growths of herbaceous, shrub and forest
vegetation, the arrangement of which
determines suitability of a site for
many species. In addition, riparian
environments often provide various
aquatic habitats and nonvegetated sub-
strates that are important to fish and

Riparian ecosystems tend to be very
complex wildlife habitats, due to the
interspersion of the many physical and
biological features present. With

- -

Water level

Level of floodplain

Flow conditions
Slow Increasing High flow Slowing Slow

Condition of plain
Flooded phase
Flooding increasing area Drying Dry

Fish behavior
Dispersed Longitudinal
in migrations Bre
waters Lateral

Feeding and growth

Return to river


Dispersed in
permanent waters
Dry-season migrations

Vegetation: wild condition
Submergence of
Terrestrial grasses terrestrial grasses
1 I

Rapid growth of floating
and rooted aquatics

Growth of
Fires terrestrial grasses

Progressive die-back of aquatics
as increasing areas of floodplain
are exposed

Figure 29. Synchrony of events related to flooding in a floodplain-river system in the
tropics. From Welcomme (1979).

maturity, and as a result of natural
flooding, riparian woodlands often be-
come interspersed with natural drain-
ages, marshes, ponds, and brushland.
This is especially evident at beaver
ponds which are used by a great diver-
sity of wildlife (Kirby 1975, Hair et
al. 1978). Inevitably, wildlife species
that require a combination of riparian
habitat features are more sensitive to
alterations than those requiring only
one component.

Associated with most riparian eco-
systems is substantial development of
edge at the interface between stream
channel and riparian vegetation, and in
the transition from floodplain to upland
plant communities (Figure 30). The

interface between stream and woody plant
communities may be one of the greatest
values to wildlife of riparian ecosys-
tems; many species occur almost entirely
in this zone (Figure 28). Riparian-up-
land edges are very important for many
upland and edge species of wildlife, at
least where woody riparian communities
adjoin relatively open rangeland,
grassland, or farmland (Thomas et al.

Because edges and their ecotones
are usually richer in wildlife than ad-
joining areas (Figure 31), they are an
important component of riparian wildlife
habitats (Hardin 1975, Thomas et al.
1979c). However, excessive manipulation
of floodplain forests to maximize edge

y .--WSHOE (EE)



30. Edges

and ecotones in ri-
Adapted from Thomas

are important as migration and dispersal
routes and as forested connectors be-
tween habitats for wildlife such as
birds, bats, deer, elk, and small mam-
mals (Figure 32) (Blair 1939, Rappole
and Warner 1976, Stevens et al. 1977,
Wauer 1977, Willson and Carothers 1979).
Woody vegetation must be present for
terrestrial species to find needed cover
while travelling across otherwise open
areas. Animals involved in population
dispersal may utilize food and water
from riparian areas during their move-
ments. The value of waterway corridors
for migratory movements may be more
accentuated in arid regions than in
humid, more heavily vegetated areas
(Wauer 1977).

Maintenance of fish populations
often depends on localized dispersal
movements over short distances and
spawning migrations covering hundreds of
kilometers. Fish migrate to satisfy nu-
tritional and reproductive requirements
that may not be met in a particular

20 40 60 80 100 120 140 160 ACRES
8 16 24 32 40 49 57 65 HECTARES

Figure 31.
diversity to
type. From

Relationship of
size of a plant
Thomas et al.


development would adversely affect the
more uncommon species that require con-
tinuous riparian forest cover.

Corridors for Dispersal and Migration

The linear nature of riparian eco-
systems provides distinct corridors that

Figure 32. Riparian zones are frequent-
ly used as migration routes by wildlife,
such as mule deer (Odocoileus hemionus)
which travel along streams between high
elevation summer range and low elevation
winter range. From Thomas et al.

et al.

stream segment, and to maintain popula-
tions throughout a stream (Hall 1972,
Durbin et al. 1979). Reproductive suc-
cess of many species requires unob-
structed access to migration (Davis and
Cheek 1966), which depends on structural
integrity of the stream and its associ-
ated riparian communities.


Despite various environmental at-
tributes common to riparian ecosystems,
there are many ecological variables that
further determine their relative values
as fish and wildlife habitats (Short and
Shamberger 1979). Those variables often
reflect suitability of a site for wild-
life species, and can be used to evalu-
ate and compare riparian habitats with
one another or with nonriparian ecosys-
tems. Most important among riparian
fish and wildlife habitat variables are
vegetation type (composition and struc-
ture), size and shape, hydrologic pat-
terns, adjacent land use, and elevation.

Vegetation Type

Generally, riparian wildlife com-
munities are influenced more by struc-
tural form of vegetation than by species
composition of the plant community. The
type, size, and arrangement of canopy,
shrub, and herbaceous vegetation largely
determine the suitability of a site for
wildlife. Most songbird species have
specific requirements of vegetation
(e.g., dense understory, closed canopy),
as do deer (e.g., twigs within browsing
height), black bear (Landers et al.
1979), bald eagle (Steenhof 1978), a few
small mammals (Miller and Getz 1977),
and many other species. Other species
are able to inhabit several community
types or successional stages. The
variety of wildlife habitats, especially
for birds, is greatest in structurally
diverse woodlands where all three vege-
tation layers are present and where
those layers are distributed in patches
throughout an area (Beidleman 1954,
MacArthur and MacArthur 1961, Austin
1970, Glasgow and Noble 1971, Carothers
et al. 1974, Carothers and Johnson 1975,
Whitmore 1975, Anderson and Ohmart 1977,
Gaines 1977, Stevens et al. 1977, Dick-
son 1978). However, homogeneous ripar-

ian woodlands, such as even-aged planta-
tions, may support a few species not
commonly found in heterogeneous stands
(Dickson 1978).

Riparian wildlife communities are
influenced to some degree by plant
species composition of an area, especi-
ally where there are clear differences
in the food values of the various vege-
tation types. There is probably much
less variation in the riparian community
types of a region than there is in the
structural forms that each type may
take. However, presence of mast (fruits
and nuts) producing trees in a bottom-
land community is especially favorable
to use by wood duck, wild turkey,
squirrels, and other wildlife. Further-
more, various plant species may host
very different invertebrate populations
among the foliage and branches; this
directly affects their value to many
songbird species.

Preferences for certain riparian
vegetation types is most prevalent among
passerine (perching) birds. In Loui-
siana and eastern Texas, oak-gum swamps
had many yellow-billed cuckoos, tufted
titmice, Carolina wrens, and cardinals,
while none of these were among the most
numerous birds in a tupelo swamp (Dick-
son 1978). Cottonwood and willow com-
munities are the most favorable riparian
bird habitats in the West (B. W.
Anderson et al. 1977). Saltcedar, an
exotic plant species, has a low value to
most riparian bird species (Beidleman
1978, Cohan et al. 1978, Conine et al.
1978), but it is valuable as nesting
habitat for white-winged dove (Shaw and
Jett 1959), and a few of the more rare
species, such as Bell's vireo, blue
grosbeak, black-tailed gnatcatcher, and
Gila woodpecker (B. W. Anderson et al.
1977, Cohan et al. 1978). Addition of
native trees to saltcedar stands would
greatly enhance the value of those
sites, as would maintenance of mature
communities rather than early seral
stages (B. W. Anderson et al. 1977). It
is generally believed that hardwoods
support greater breeding bird densities
and number of bird species than soft-
woods (Thomas et al. 1975).

Although little information is
available on herbaceous and non-vegeta-
ted areas of riparian ecosystems it

_ ___ ~L~L II______

seems reasonable that their values to
fish and wildlife differ little from
structurally similar areas in non-ripar-
ian zones. Wildlife communities in ri-
parian marshes are likely dominated by
waterfowl (especially dabbling ducks and
geese), shorebirds (e.g., avocet,
rails), a few songbirds (e.g., black-
birds, wrens, and sparrows), furbearing
mammals, and various amphibians (Hardin
1975, Flake and Vohs 1979). Value of
marshes to wildlife is largely influ-
enced by water regimes, interspersion of
cover and open water, and the composi-
tion and structure of the emergent marsh
plants (Weller 1978).

As a result of continual erosion
and deposition, streams commonly produce
at least two kinds of nonvegetated sub-
strates: barren streambanks; and stream
channel alluvial areas (e.g., outwashes
and sandbars). Prior to invasion by
herbaceous or woody plants, steeply
sloped streambanks provide required
nesting sites for the bird species such
as belted kingfisher, bank swallow, and
rough-winged swallow (Cornwell 1963,
Gaines 1974). Mid-channel sandbars
along the Missouri River provide resting
grounds for migrating waterfowl, basking
areas for softshell turtle, and nesting
sites for the least tern (U.S. Fish and
Wildlife Service 1980). Sandy shoals
are important to turtles for nesting
(Dodd 1978), and for killdeer, spotted
sandpiper and upland sandpiper which
feed near the sand-water interface. The
sandbar-channel combination serves as a
feeding ground and nursery area for many
species of fish. Bald eagle and osprey
feed on fish concentrated in those shal-
low water areas (U.S. Fish and Wildlife
Service 1980a).

Size and Shape of Riparian Area

The size (width and/or area) of a
plant community has a direct relation to
its ecological values. There is no
clear consensus on the minimum size of a
riparian stand that is needed to accomo-
date wildlife populations, protect water
quality, or provide recreation. Various
minimum dimensions have been recommended
for these purposes (Table 21), but addi-
tional research is needed to provide a
more comprehensive data base.

Even very narrow strips of riparian
vegetation are important to instream
aquatic communities and for certain
kinds of wildlife. Species commonly
occurring along streams or shorelines,
such as mink, belted kingfisher, and
riparian edge species, are often able to
establish territories in narrow riparian
woodlands (Curtis and Ripley 1975).
However, narrow riparian woodlands are
unsuitable for species requiring large
areas of forest or considerable isola-
tion from man, such as black bear (Lan-
ders et al. 1979), osprey (Swenson
1979), great blue heron (Scott 1980),
the presumed extinct ivory-billed wood-
pecker (Korte and Fredrickson 1977), and
many forest dwelling songbirds. Reduc-
tion in size of southwestern riparian
woodlands is at least partly responsible
for the regional decline of several
species; Cooper's hawk, red-shouldered
hawk, and yellow-billed cuckoo were
found only where patches were more than
100 m wide (Gaines 1974).

The area of riparian vegetation
most heavily used by terrestrial wild-
life is that within 200 m of a stream
(or open water), although some species
travel as much as 4 km from nesting to
feeding area (Figure 28). A 200m wide
vegetative strip is apparently able to
accommodate breeding territories of most
songbirds (Stauffer and Best 1980).
Many vertebrates, especially riparian
mammals, reptiles, and amphibians, con-
centrate their activities well within
60 m of water (Hairston 1949, Organ
1961, Tilley 1973, Krzysik 1979).

Along with the lateral dimension of
riparian wildlife habitats, the overall
size is also important to many species.
Size of animal territories varies widely
among species, ranging from less than a
hectare for small terrestrial animals to
several square kilometers for birds of
prey and large mammals. Reducing the
size of a community type progressively
eliminates species requiring large areas
of the particular type and favors expan-
sion of species associated with the new
land use and the edges created. For ex-
ample, prothonotary warblers are gener-
ally absent from waterways where the
border of deciduous trees is less than
30 m (100') deep (Simpson 1969). In

Table 21. Width of riparian buffer
aquatic life in streams.

strips recommended to protect water quality and

Function of buffer strip Recommended width Recommended by

Protect water quality from 8 m (25') plus .6 m (2') Trimble 1959
logging per 1% of slope
Protect water quality from 16 m (50') plus 1.2 m (4') Trimble and
logging in municipal watersheds per 1% of slope Sartz 1957

Protect aquatic life from at least 30 m Erman et al. 1977

Protect water quality and fish 25 m (75') plus any addi- USDI Bureau of Land
tional width that supports Management 1979
riparian vegetation.
Protect streams from adverse 30 m (100') U.S. Dept. Agricul-
land management practices ture 1980
Maintain wild or scenic 400 m (.25 mile) Wild and Scenic
values of river corridors Rivers Act
(P.L. 90-542)
Protect aquatic environment at least 15 m Canada Fisheries
and Marine Service

aThese recommendations do not represent conclusions or recommendations of the FWS or
the authors of this report.

contrast, red-shouldered hawks are found
primarily in forested stream valleys
with adjacent clearings (Stewart 1949,
Craighead and Craighead 1956), and are
absent from the center of extensive
forest stands (Brown and Amadon 1968).
While edge species tend to be very ubi-
quitous, species that require large ri-
parian stands are generally less common,
and face declining population levels as
riparian alterations continue. Where
riparian "islands" are created, the size
needed to support potential songbird
diversity near maximum values is at
least 5-6 ha, but is probably as large
as 10 ha for maintaining a diversity of
all wildlife forms (Gaines 1974, Galli
et al. 1976, Emmerich 1978, McElveen
1978, Willson and Carothers 1979).
Larger areas will support additional

species because interspecific competi-
tion and territoriality in a small stand
limit the number of large species that
can coexist.

Width of a riparian woodland also
determines the degree to which impacts
of adjacent land use on water quality
are buffered before reaching the stream.
Optimum width for a riparian buffer zone
varies with stream width, topography,
soil type, type of impact, sensitivity
of the resource, and water quality
standards. Buffer strips reduce erosion
(and pollution), preserve the stream
channel's stability, retard runoff, trap
sediments and nutrients, maintain suit-
able water temperatures for aquatic
life, and provide vegetation and inver-

tebrates as food for birds, and other
wildlife (Curtis and Ripley 1975).

Stream Type and Hydrologic Pattern

Riparian communities are found
along many kinds of streams, varying in
size, shape, velocity, flow patterns,
and water quality. The importance of
stream type to fish and wildlife is
largely a function of the relation be-
tween these variables and habitat com-
ponents already discussed.

As one moves downstream from tribu-
tary to river, flow volume increases,
overbank flooding is more widespread,
and riparian communities are broader and
more distinct than in headwater areas
(especially in mountainous regions). At
the same time, the influence of riparian
vegetation on the adjacent stream de-
creases downstream. Middle-order per-
ennial streams and their riparian com-
munities may be the most heavily used
wildlife areas in a watershed because
they provide very sizable and diverse
habitats (both instream and riparian).

Riparian wildlife are also sensi-
tive to differences in stream type that
are not always reflected by vegetation.
Ephemeral streams often support valuable
woody riparian growth, but lack fish,
the aquatic food base upon which certain
riparian species depend. Similarly,
clear slow-moving water is important to
beaver and muskrat (Flood et al. 1977),
belted kingfisher (Cornwell 1963), and
water snakes (Lagler and Salyer 1947)
because it enhances the production of
aquatic food organisms and the ability
of these species to find food.

Periodic flooding is one of the
most significant phenomena affecting the
use of riparian ecosystems by fish and
wildlife. Although floodplains are very
unpredictable environments, annual
flooding has a generally favorable
effect on productivity of fish (Wharton
and Brinson 1978, Welcomme 1979) and
wildlife (Wharton 1970, Batzli 1977,
Gaines 1977, Fredrickson 1979). The
overflow of streams onto floodplains
directly influences both animal popula-
tions and their habitats. Overbank
flooding is critical for the exchange of
energy, nutrients, and animal popula-

tions between aquatic and terrestrial
portions of riparian ecosystems.

The composition and structure of
riparian plant communities is dependent
upon the prevailing hydrologic regime.
Many bottomland tree species must be
flooded periodically to produce seeds,
and for subsequent development into
seedlings and mature trees (Teskey and
Hinckley 1977a, 1977b, 1978a, 1978b,
1978c; Walters et al. 1980a, 1980b).
However, no woody plants are able to
reproduce on sites that are flooded
throughout the year. Development of
understory vegetation in wetland forests
is reduced by widely fluctuating water
levels during the growing season (Flin-
chum 1977, Brown et al. 1978, Swift
1980). Clearly, the long-term mainten-
ance of existing riparian wildlife habi-
tats depends on the continuation of
natural flooding patterns.

Seasonal and short-term overbank
flooding has profound effects on terres-
trial wildlife. Distributions of ground
dwelling vertebrates are often more
closely related to hydrologic patterns
than to vegetation features. Riparian
mammal populations may be generally im-
poverished (Barclay 1980) or relatively
dense (Arnold 1940), depending in part
on recent hydrologic events (Blair 1939,
Armstrong 1977, Batzli 1977, Miller and
Getz 1977). Short-term floods (several
days) often have little detrimental ef-
fect on wildlife; deer mice, tree squir-
rels, and box turtles apparently take
refuge in unflooded sites or trees
(Stickel 1948, Hoslett 1961, Ruffer
1961). In contrast, severe flooding
(several weeks) temporarily eliminates
and may limit resident small mammal pop-
ulations in a floodplain. Recoloniza-
tion by individuals from nearby unflood-
ed areas occurs slowly (Blair 1939,
Wetzel 1958, McCarley 1959, Turner 1966,
Iverson et al. 1967).

Depth and duration of flooding in a
riparian ecosystem also determines the
availability of foods for waterfowl and
wading birds. In the southeastern
U.S.A., inundation of bottomland hard-
woods during winter creates excellent
feeding areas for hundreds of thousands
of ducks, especially wood ducks and mal-
lards, which feed on the fallen mast

crop (acorns). Wood duck, great blue
heron, and green heron feed primarily in
water less than 0.5 m deep (Martin et
al. 1951, Palmer 1962, Webster and
McGilvrey 1966); a gradual rise or fall
of water levels in the riparian zone
allows maximum use of the area by these
and other species.

Permanent impoundment of streams
has very dramatic consequences on fish
and wildlife habitats in the inundated
floodplain (Figure 33). A rapid in-
crease in fish populations commonly fol-
lows reservoir construction as food re-
sources on the freshly inundated flood-
plain are exploited. A subsequent
decline results as those resources are
depleted, without rejuvenation by alter-
nating wet and dry phases, as occurred
previously. Long-term impoundment of
streams by man or beaver eliminates
habitat of ground nesting, canopy feed-


Habitat for stream-dwellin| fish
Predominantly floodain/terres-
trial wildlife habitat
Streambank habitat for many spae
cializd wildlife species
Natural hydrologic regime provide
exchange pathways for nutrients,
detritus and organisms between
channel and floodplain
Downstrem transport of detritus
and sediments
Corridor for fish and wildlife move-

Habitat for lake-dwelling fish
Prdominantly aquatic fish habi
Streamblnk habitat replaced by
extensive, often unstable shore
line: altered species assemlle
Permanent inundation eliminates
floodplain egetation and vital
pathways of exchange
Retention of detritus and sedi-
ments behind dam
Corridor altered and interrupted

Figure 33. Fish and wildlife values at
small stream impoundments.

ing, and ground foraging birds, includ-
ing prothonotary warbler (Simpson 1969),
Kentucky warbler, and white-throated
sparrow (Dickson 1978), and many ripar-
ian reptiles and amphibians (Dodd 1978).
However, partially impounded riparian
communities can enhance areas for bald
eagle, waterfowl, cavity-nesters, and
flycatching birds (Hair et al. 1978),
and provide protection from predators
for herons, egrets, and red-winged
blackbirds (Dickson 1978).

Adjacent Land Use

Wildlife use of riparian ecosystems
can be influenced by adjacent land use.
Riparian ecosystems surrounded by low
quality wildlife habitats often support
higher density and diversity of birds
during migration than would otherwise be
expected, because populations do not
spread out over the entire area to feed
(Stevens et al. 1977). Nesting birds
can inhabit riparian communities in
higher densities where adjacent agricul-
tural lands produce an abundant food
supply but lack nesting sites (Carothers
et al. 1974). Similarly, carrying ca-
pacity of deer in bottomland hardwood
forests of the lower Mississippi Valley
may double where agricultural crops are
readily available (Glasgow and Noble
1971). Riparian ecosystems surrounded
by forest land do not usually exhibit
such obvious influences of adjacent
wildlife habitat, because resources are
more similar and competing species are
normally present there.

Many bird species find shelter in
riparian vegetation, but feed exten-
sively in surrounding agricultural lands
(Glasgow and Noble 1971, Carothers et
al. 1974, Whitmore 1975, Conine et al.
1978). Of 63 riparian species along the
lower Colorado River, 41 travelled vary-
ing distances into adjacent agricultural
lands (Figure 34). Within those famed
lands, bird densities increased towards
the floodplain, and were positively cor-
related with presence of canals, weedy
margins, and alfalfa. However, total
encroachment of agriculture into the ri-
parian zone would completely eliminate
many species (Conine et al. 1978).

Effects of adjacent land uses are
limited primarily to the vegetative
edges of riparian ecosystems, and are


0.4 0.4 0.8 1.2 1.6 2.0 2.4
Figure 34. Distance travelled by ripar-
ian bird species into agricultural
areas. From Conine et al. (1978).

most important to wildlife in narrow or
small patches of riparian vegetation.
In narrow corridors of streambank vege-
tation, most wildlife species must ex-
tend their territories into adjacent
lands, and are directly affected by the
food resources and wildlife populations
that occur there.

The composition of riparian wild-
life communities is affected by eleva-
tion, especially in the West, where
dramatic changes in climate, topography,
and vegetation are associated with alti-
tude (Noon and Able 1978). In addition,
riparian dependence of many species is
reduced at higher elevations, because
moisture is readily available in nonri-
parian communities as well (Hairston
1949, Johnson et al. 1977).

Abundance and diversity of birds in
lowland riparian ecosystems is signifi-
cantly greater than in high elevation
riparian areas (Finzel 1964, Wooding
1973, Stevens et al. 1977, Burkhard
1978). A similar phenomena may exist
among other vertebrates (Burkhard 1978),
but this has not been confirmed.

Elevation effects on riparian wild-
life communities are often associated
with riparian habitat variables that
have already been discussed (e.g., size,
productivity, and diversity of vegeta-
tion, hydrologic patterns, value as
travel corridors). For example, peren-
nial streams in relatively flat areas
usually support large, distinct riparian
corridors, while riparian vegetation
along mountain streams may be lacking or
barely noticeable. Riparian woodlands
that extend between high mountain and
lowland areas may be important for sea-
sonal movements by elk and deer (Thomas
et al. 1979b), but may not be used by
migrating birds simply because birds fly
between mountain ranges rather than over
them (Stevens et al. 1977).


Surveys of animal communities in
riparian ecosystems reveal that these
areas are inhabited by a great variety
of birds, mammals, amphibians, and rep-
tiles. Certain groups of wildlife tend
to predominate in undisturbed riparian
ecosystems across the U.S.A. However,
the presence or absence of particular
species is often determined by specific
habitat variables, geographic location,
and site specific alterations from human

Partial descriptions of riparian
wildlife communities have been reported
for many areas of the country, but
thorough characterizations are not read-
ily available for most (Table 22). The
value of riparian ecosystems to wildlife
has been most intensively studied in
western arid regions, the Midwest, and
the lower Mississippi Valley where
threats to riparian ecosystems tend to
be greatest.


Table 22. References for information on riparian wildlife communities in the U.S.A.

Region State Referencesa


Pacific Northwest

Rocky Mountain

Arid Southwest


Corn Belt

Lake States

Mississippi Delta









New Mexico

North Dakota

South Dakota







la ine
New Jersey
New Hampshire
New York


West Virginia
North Carolina
South Carolina


AB(12), Gaines (1974, 1977), Goldwasser (1978),
Hehnke and Stone (1978), Ingles (1950), Michny et al.
(1975), Roberts et al. (1977), Sands (1977, 1978).
Hinschberger (1978), Thomas (1979)
Lewke (1975), HcKern (1975)
Anmstrong (1977), Beidleman (1948, 1954), Fitzgerald
(1978), Wooding (1973)
AB(2), Whintore (1975)
AB(1), Brown (1967)

Anderson and Ohmart (1977), Arnold (1940), Carothers
et al. (1974), Johnson et al. (1977), Johnson (1978),
Johnson and Simpson (1971), Stevens et al. (1977),
Szaro (1980)
Austin (1970)
Hubbard (1971), Schmidt (1976)
AB(2), Boerr and Schmidly (1977), Engel-Wilson and
Ohmart (1978), Wauer (1977)

AB(7), Beidleman (1948, 1954), Crouch (1961)
AB(2), Tubbs (1980), Zimmerman and Tatschl (1975)
AB(7), Barclay (1978, 1980), Blair (1939), Heller
Emmerich (1978)

AB(1), Blem and Blem (1975a,b), Yeager (1949),
Yeager and Anderson (1944), Wetzel (1958)
AB(1), Best et al. (1980), Geier (1978), Geier and
Best (1980), Hoslett (1961), Stauffer and Best (1980)
AB(6), New (1972)
AB(6), Leite (1972)
Dawson (1979), Iverson et al. (1967), Kirby (1975)
Dawson (1979), Faanes (1979), Prellwitz (1976)

AB(3), Glasgow and Noble (1971), Kennedy (1977),
Ortego et al. (1976)
Fredrickson (1979)

AB(1), Golet (1976), Miller and Getz (1977)
AB(3), Golet (1976), Swift (1980)
AB(3), Hardin (1975), Malecki and Eckler (1980),
Webb et al. (1972)
Hooper (1967)
AB(1), Dodge et al. (1976), Miller and Getz (1977),
Possardt and Dodge (1978)
AB(2), Ellis (1976), Gill et al. (1975), Hooper (1967)
AB(1), Wharton (1970, 1978)
Dickson (1978)
AB(1), Hair et al. (1978)

Kessel and Cade (1958), Maher (1959), Sage (1974)

a"AS" indicates that breeding bird census data have been published from one or more sites
(number of sites in parenthesis) in American Birds or Audubon Field Notes.

Three groups of wildlife are des-
cribed here: birds, mammals, and herps
(reptiles and amphibians). The purpose
of this section is to identify wildlife
species or groups that are commonly
found in riparian ecosystems. Where
possible, the relative abundance and
diversity of animal communities are des-
cribed. Although some fish communities
are dependent on riparian vegetation,
they are not characterized in this

Birds are probably the most common,
conspicuous, and easily studied form of
wildlife in riparian ecosystems. As a
result, and because of their general
aesthetic popularity, there has been
much research that describes riparian
bird communities.

Community Characteristics. Birds
using riparian ecosystems can be cate-
gorized into at least four groups based
on their seasonal occurrence: (1) sum-
mer (breeding) residents; (2) winter
residents; (3) transients (passing
through during fall and/or spring migra-
tions; and (4) permanent residents (non-
migratory species). As a result of many
factors (migratory and local movements,
reproduction, mortality, and seasonally
changing habitat requirements), bird
populations are distinctly different
from season to season.
Riparian ecosystems are valuable as
breeding habitats for birds everywhere
in the U.S.A. Individual stands of ri-
parian woodland usually have 10 to 50
breeding bird species, with most having
between 20 and 34 (Figure 35, Table 23).
Population densities of birds breeding
in riparian areas generally fall between
40 and 900 pairs per 40 ha (Table 24),
but most often are between 150 and 550
pairs per 40 ha (Figure 36). Presum-
ably, bird density reflects productivity
and is a good measure of the avail-
ability of birds for observation by
birdwatchers, photographers, hikers,
The value of riparian ecosystems to
winter bird populations has received in-
creased attention from biologists re-
cently (Dickson 1978, Szaro 1980). The
species richness of bird communities in




10 20 30 40
Figure 35. Number of .breeding bird
species on 98 riparian census plots
(from Breeding Bird Census data pub-
lished in Audubon Field Notes and
American Birds).

riparian vegetation during winter is
generally comparable to that in summer,
except in the most interior areas of the
U.S.A. (Table 25). The abundance of
winter residents is commonly equal to or
greater than that of summer birds (Table
26), especially where there is a major
influx from northern and inland breeding
grounds (Lewke 1975, Kennedy 1977, Bar-
clay 1980, Szaro 1980).
Riparian ecosystems are also impor-
tant to birds during migration (Rappole
and Warner 1976, Stevens et al. 1977,
Fitzgerald 1978). Many riparian birds
use the same habitats, when available,
during migration as they do on their
nesting grounds (Parnell 1969). Conse-
quently, the number of species found in
a riparian ecosystem during spring and
fall is increased, because it includes
departing and incoming seasonal resi-

Table 23. Number of breeding bird species on riparian study areas.

Community type No. of
and location species Source

Riparian vegetation, Texas 38 Wauer 1977
Cottonwood-willow, Texas 27 Engel-Wilson and Ohmart 1978
Saltcedar, Texas 28 Engel-Wilson and Ohmart 1978
Desert riparian, California 13 Berry 1977
Willow-cottonwood, California 20 Ingles 1950
Cottonwood-willow, California 27 Gaines 1977
Various types, Arizona 18-35 B. Anderson et al. 1977
Bottomland forest islands, Okl. 11-15 Barclay 1978
Mature floodplain forest, Mo. 31 Zimmerman and Tatschl 1975
Young floodplain forest, Mo. 19 Zimmerman and Tatschl 1975
Bottomland hardwoods, Louisiana 16-23 Dickson 1978
Beaver ponds, South Carolina 15 Hair et al. 1978
Riparian forests, New York 33 Malecki and Eckler 1980
Riparian corridor, New York 24 Malecki and Eckler 1980
Alder, New York 26 Hardin 1975
Shrub, Alaska 8 Sage 1974

Table 24. Breeding bird densities in riparian ecosystems.

Plant community type Density
and location (pairs per 40 ha) Source

Cottonwood-willow forest, Ca. 840 Gaines 1977
Willow-cottonwood streambottom, Ca. 197 Ingles 1950
Sacramento Valley riparian, Ca. 240-450 Gaines 1977
Desert riparian, California 863 Berry 1977
Desert bosques, Nevada 44-49 Austin 1970
Floodplain vegetation, Arizona 200-325 Cohan et al. 1978
Cottonwood, Arizona 425-847 Carothers et al. 1974
Mixed riparian vegetation, Arizona 193-322 Carothers et al. 1974
Willow, Colorado 100 Fitzgerald 1978
Cottonwood-willow, Colorado .525-589 Fitzgerald 1978
Cottonwood-willow, Colorado 225-900 Beidleman 1954
Cottonwood, Colorado 319 Bottorff 1974
Saltcedar, Colorado 131-503 B. Anderson et al. 1977
Saltcedar, Texas 486 Engel-Wilson and Ohmart 1978
Cottonwood-willow, Texas 708 Engel-Wilson and Ohmart 1978
Bottomland forests, Oklahoma 400 Barclay 1978
Bottomland hardwoods, Louisiana, Tx. 300-590 Dickson 1978
Riparian vegetation, New York 59-167 Malecki, and Eckler 1980
Riparian communities, Great Plains 137-748 Szaro (1980)

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs