v>EPA
                           United States
                           Environmental Protection
                           Agency
                         Environmental Research
                         Laboratory
                         Corvallis OR 97333
Research and Development    EPA-600/D-83-094  September 1983
ENVIRONMENTAL
RESEARCH    BRIEF
                              Habitat Structure and Fish
                       Communities of Warmwater Streams

                      James R. Karr, Paul L. Angermeier, and Isaac J. Schlosser*
Introduction

The basic goal of clean water legislation passed in the last
two decades is restoration and maintenance of the chemi-
cal, physical, and biological integrity of the nation's waters
The mechanisms required to reach this goal are not entirely
obvious Early efforts concentrated on physical and/or
chemical pollution;  however,  a broader perspective is
required.1'2  Based on the authors' work over the past
decade, five major  sets of variables that  impact the
structure of stream  communities (Figure  1) have been
identified.

Most research on the role of physical habitat characteristics
in the regulation of fish community structure has concen-
trated on cold water systems, with emphasis on salmonids.
The significance of physical habitat in warmwater streams
has been largely ignored by water resource planners. Even
studies that describe physical  habitat of streams rarely
examine the cause and effect interactions of habitat
structure, availability of food resources, and other factors
that shape fish communities.

With this background in mind, a research program was
initiated to evaluate the role of physical habit at in regulating
the structure of fish communities in warmwater streams in
'Co-authors Karr and Angermeier are with the Department of Ecology,
 Ethology, and Evolution, University of Illinois. Champaign, Illinois; Co-
 author Schlosser is with the Department of Biology, University of North
 Dakola, Grand Forks, North Dakota.
                     Figure 1.    Primary variables that affect the structural and
                               functional integrity of an aquatic biota.
                     east central Illinois. The study combined an empirical
                     approach involving observations of fish in relatively natural
                     conditions with  studies of stream areas subjected to
                     extensive modifications by human society. Finally, several
                     experimental field and laboratory studies were designed to
                     clarify aspects of fish community dynamics. This research
                     brief is a summary of several journal articles published and
                     in review. Complete citations for those articles can be found
                     at the end of this brief.

                     Fish Communities Along Physical Habitat
                     Gradients—Natural Gradients

                     Two habitat gradients (upstream to downstream and riffle
                     to pool) were investigated in Jordan Creek in Vermilion

-------
County, Illinois.3 Several major patterns were identified—
from the diversity of the fish community to the apparent
dominant processes regulating community  organization
along each of these gradients (Table 1}.

The major attributes of physical habitat measured in this
study were habrtal  diversity and habitat volume. Habitat
diversity is a complex integration of depth, current velocity,
and substrate attributes. Habitat volume  is measured  as
stream area times mean depth for  a  study region, Both
habitat diversity and habitat volume increased  from up-
stream to downstream and riffle to pool habitats. Temporal
variation in habitat diversity wasgreater in upstream areas
and  habitat volume tended to vary more over time  in
upstream  and  riffle areas.  Seasonal and  year-to-year
variation in rainfall also  caused variation  in  habitats,
especially volume.

Benthic insect  density in  Jordan Creek  was high from
autumn (Oct-Nov) through spring (May-June). Following
emergence of adults in late spring, invertebrate  densities
were low in summer in areas with riparian vegetation.
Insect availability, as indicated by drift samples, increased
along a gradient from silt-sand to gravel-rock substrates.
Potential food availability for top carnivore fish peaked in
late summer and early fall with increased numbers and
biomass of young-of-the-yearfish.
                                            Although  habitat  diversity  was significantly  (p  <.05)
                                            correlated with  fish species diversity,  the relationship
                                            between the two variables varied as a result of seasonal
                                            migration by fishes. These seasonal migrations were tied to
                                            changing flow conditions, food availability, and the search
                                            for suitable spawning and nursery areas. As a result, the
                                            utility of the habitat/ fish  diversity relationship as a
                                            predictive model varied seasonally. In addition, the preci-
                                            sion  of  the relationship was lowest in  more variable
                                            upstream and riffle habitats.

                                            Younger fish (age  classes O-ll) were found primarily in
                                            shallow, temporally variable areas upstream and in riffles.
                                            Relative growth rates were highest during summer but they
                                            did not increase (relative to spring) as much as expected
                                            from seasonal increases  in water temperature. Centrar-
                                            chids had substantially higher growth rates than cyprinids
                                            during early life stages. Net production for age O-ll fish was
                                            highest in upstream and riffle areas because those areas
                                            supported high densities of young, generalized insecti-
                                            vores. Net production of insectivore-piscivores was highest
                                            in downstream and pool habitats. Stream  reaches with
                                            large, stable pools and raceways produced fewer fish due to
                                            shifts in age structure toward fewer, large individuals (age
                                            III+) with slower relative growth rates. Temporal variation in
                                            reproductive success and survival of younger age groups
                                            (O-l) was associated with variation in peak flows. Finally,
Table  1.
Summary of Relative Characteristics of Habitat Structure and Fish Organization Along Two Physical
Gradients in a Headwater Stream
                                                                Relative Position on Gradient
          Characteristics
                                                Downstream or Pool
                                                   Environment
                                                                         Upstream or Riffle
                                                                           Environment
 1.  Habitat Structure

 2.  Fish Community Structure
    a.   Species richness
    b.   Age structure
    c.   Size composition
    d.   Dominant trophic group(s)
3.  Fish Community Function
    a.  Net production
    b.  Absolute and relative
        growth rates of age O-l
        of the dominant trophic
        group(s)

4.  Seasonal and annual stability
    of community attributes; i.e.,
    species richness, trophic
    structure, age structure,
    and production

5.  Hypothesized dominant
    processes regulating
    community organization
                                  Deep, temporally stable
                                           High
                                         Old fish
                                        Large fish
                                   Insectivore-piscivores

                                    Benthic insectivores

                                           Low
                                           High
                                           High
Shallow, temporally variable
           Low
        Young fish
        Small fish
  Generalized insectivores
           High
           Low
           Low
                                   Competitive exclusion
                                   and predation
Recolonization dynamics,
effects of gradual
changes in the physical
environment on competitive
interactions, and temporal
variation in reproductive
success

-------
variation in peak flow was  a  major factor determining
spatial and temporal variation in production.

Therefore, spatial shifts in physical conditions from shallow,
temporally variable areas (upstream and riffles) to deeper,
more stable areas (downstream and pools)  result in
consistent spatial changes in community structure (Table
1). The ultimate mechanisms responsible for these changes
were not precisely documented, although they clearly vary
from  headwater  to downstream or riffle to pool.  We
hypothesize that in shallow, unstable habitats, recoloniza-
tion dynamics, the effect of  gradual changes in physical
conditions  on  competitive  interactions,  and  temporal
variation in reproductive success are more important than
competitive exclusion and predation  as determinants of
community organization (Table  1).

Fish Communities Along  Physical Habitat
Gradients—Human Disturbance

We assessed  the impact  of channel straightening  and
removal of riparian vegetation on trophic structure, repro-
ductive success, and growth rates of fishes in relatively
natural (Jordan Creek - JC) and modified (Big Ditch -  BD)
headwater streams.4 Shallow  habitats and organic sub-
strates (diatoms and/or filamentous algae)  were more
common in BD (not shaded by riparian vegetation) than JC
(shaded) during  low  flow periods  in summer.  Insect
densities in JC were highest in late spring, declining to low
levels by late summer. Insect densities in BD were high
throughout summer.

Fish species in JC were predominantly benthic insectivores
and insectivore-piscivores,  and trophic  structure,  age
structure, and biomass of the fish community were stable
between years and seasons. Recruits made up a small  and
stable portion of community  biomass and were primarily
insectivore-piscivores and generalized insectivores. Young-
er age classes occupied shallow riffle habitats.

In contrast, trophic structure  and recruits in  BD were
predominantly generalized insectivores, omnivores,  and
herbivore-detritivores. Omnivores and herbivore-detriti-
vores were primarily mid-river  species  (quillback  and
gizzard shad). Considerable seasonal and annual variation
in trophic  structure,  total biomass,  and age  structure
                                            occurred in BD associated with annual fluctuations in flow
                                            regime, abundance of organic substrates, and reproductive
                                            success of mid-river species.  Younger age classes had
                                            higher summer growth rates in BD than JC.

                                            The effects of alteration of headwater streams are evident
                                            when placed in  the context of the stream continuum
                                            concept which suggests that interactions between physical
                                            environment and the organic energy  base result in a
                                            relatively predictable pattern of lotic community structure
                                            and function  from headwaters to downstream areas. The
                                            critical effect of stream alterations in the context of this
                                            concept is that alterations create a shallow, temporally
                                            variable physical  environment typical of headwater areas
                                            where most recruitment occurs. Yet, at the same time, the
                                            alterations shift  the energy  base toward  autotrophic
                                            processes which are more typical of mid-river habitats. As a
                                            result, mid-river omnivores and herbivore-detritivores domi-
                                            nate recruitment in modified headwaters. Reduced avail-
                                            ability of benthic invertebrates and altered habitat condi-
                                            tions result in declining  abundance of  insectivores and
                                            carnivores  due to lowered reproductive success. The
                                            authors conclude that land use and channelization activities
                                            in headwater streams have played a major role in the shift
                                            in recent decades of many large river communities in the
                                            midwestern United States from dominance by insectivore
                                            and insectivore-piscivore  fishes to omnivores and herbi-
                                            vore-detritivores.8

                                            Physical Habitat and Fish
                                            Assemblages in Divided Streams
                                            Two 35-m  sections of Jordan Creek were divided in half
                                            longitudinally with 6mm mesh hardware cloth supported by
                                            steel posts.6 On one side of each section, all cover features
                                            (logs, limbs) were removed from in or near the water. On the
                                            other side, a continuous series  of similar  objects was
                                            secured along the  stream. In July and September, samples
                                            of the biomass of fish were 4.8 to 9.4 times as high in the
                                            areas with structurally complex habitats (Table 2). Further,
                                            larger fish, and especially top predators, tended to select
                                            the structured habitat. In this case we  know that water
                                            quality was the same in the structured and unstructured
                                            sides of the stream, yet the numbers of fish are markedly
                                            different. These  improved  habitat conditions  seem  to
                                            provide  two  things:  habitat for  small  fish including a
                                            diversity of substrates for food organisms and hiding places
Table 2.
Fish and Invertebrate Densities and Fish Biomass in Adjacent Sides of a Stream Split with 6mm Mesh
Hardware Cloth
                                               July 1979
                                      No Cover              Cover
                                                                         September 1979
                                                                   No Cover              Cover
Fish

  Number of individuals

  Number >120mm TL

  Total biomass (gms)

Benthos

  Number/0.1 m2
                             34

                              2

                            170



                             92
  46

  28

1606



 383
  4

  2

284



 39
  73

  17

1366



 219

-------
  (cover) from which large fish can prey on smaller species
  This emphasizes the importance of habitat structure as a
  determinant of biotic conditions in a stream.

  Woody Debris in Warmwater Streams

  The importance  of woody  debris to the  structure and
  function of a warmwater stream ecosystem was examined
  by removing woody debris from a series of stream reaches
  Experimental  reaches  were  compared with unaltered
  reaches over a 2-year period.7

  At initiation of the  experiments (June 1980), altered and
  unaltered sites were similar with respect to depth profiles,
  current regimes, and standing stock of organic litter (Figure
  2). Water depth declined in all sites through the summer
  and autumn due to lack of precipitation,  but shifts toward
  shallow depths were especially pronounced in altered sites
  due to the filling of pools by unstable substrates (Figure 2)
  In addition, the abundance of organic litter on the stream
  bottom declined markedly in altered sites  but remained
  relatively constant  in unaltered  sites.6  Litter abundance
    declined in altered sites due to burial by shifting substrate
    and the absence of  retention structures.  In June 1981,
    depth profiles, current regimes, and organic  litter abun-
    dances for unaltered sites were similar  to those in June
    1980. In altered sites,  however, deep areas were not re-
    established, currents were faster, and organic debris  was
    only one-third as abundant as the previous year (Figure 2).
    Seasonal shifts in depth, current and litter abundance in
    1 981 were similar to those observed in 1 980, though  less
    pronounced due  to  the  uncommonly stable flows  that
    occurred through the summer of 1981. In reaches of Jordan
    Creek with stable (rocky) substrates, removal of woody
    debris had  less impact on stream structure and function
    than in reaches with unstable (silt-sand) substrates.

    Fish were  also monitored  throughout the experimental
    period to evaluate effects of habitat changes  on the  fish
    community.7 In June 1 980, mean fish biomass was similar
    (p >05) between groups of altered  and unaltered sites
    (Figure 3) but by October 1 981 mean biomass in unaltered
    sites (1 833gm/35m) was significantly (p <.05) greater than
    that found  in altered sites. Declines  in fish  abundance
                June 1980
                                                 October 1980
                                                                                   June 1981
       1
                                                                                                        Altered
                                                                                                          Sites
                                                                                                      Unaltered
                                                                                                          S/tes
10-  20- 30 +
Depth cm
                             Litter
                                             10-  20-

                                             Depth cm
         Litter
                        jo-  20-  30
                        Depth cm
L/tier
Figure 2.    Depth frequency distribution and litter abundance in ,
            Debris was removed in July 1980
naltered and altered (woody debris removed) reaches of Jordan Creek.

-------
                         Unaltered
                           Sites
                                                                        Altered Sites
                                                    Artificial Cover
                          n = 3
                                                                               No Cover
                                                                                 n= 6
         3.5-1
    -H    3.0 H
    E
    .0
    3S g), their abundance being lower in altered
than unaltered sites. Distributions of small fish (<40 mm
total length) among sites were variable. Small cyprinids
tended to be most numerous in altered sites, while small
centrachids tended to be most numerous in unaltered sites.

These experiments illustrate the  interactions of physical
and biological processes that occur in stream ecosystems.
Woody debris has several  important functions in Jordan
Creek. It affects channel hydraulics and so maintains depth,
current, and substrate diversity, ft acts as a stable substrate
for retaining organic material and supporting macroinver-
tebrates. These functions, in addition to providing cover, are
essential to maintaining habitat quality for fish. In streams
with unstable substrates (silt, sand), woody debris may be
the most important attribute of physical habitat in determin-
ing ecosystem structure and function.
                                              Effects of Fish Consumption
                                              On Food Availability
                                              Paired screen exclosures were erected to assess the impact
                                              offish consumption on benthic invertebrate abundance and
                                              size distribution.8 One exclosure (closed) excluded fish from
                                              a section (0.54 m2) of stream bottom, while the other (open)
                                              permitted fish to enter and exit freely. Exclosures were set
                                              up for 4-week periods in  upstream (silt-sand substrates)
                                              and  downstream (gravel-pebble  substrates) reaches  of
                                              Jordan  Creek.  Closed  exclosures  from  upstream sites
                                              supported  greater  densities  of invertebrates than open
                                              enclosures (Table  3).  These density  differences were
                                              largely due to the greater abundance of chironomid larvae
                                              and copepods, both of which are important food items for
                                              Jordan Creek fish. In addition, large invertebrates (at least
                                              4.7mm long), which are geperally preferred by fish, were
                                              more abundant in closed exclosures than open ones. Data
                                              from downstream exclosures indicated that fish consump-
                                              tion  had no  effect on invertebrate abundance or size
                                              frequency. These results suggest that the  potential for
                                              competition for food among  stream fishes is greatest in
                                              upstream  areas.  If so, stream  modifications  such as
                                              removal of riparian vegetation or woody debris, which may
                                              have dramatic effects on invertebrate availability, can be
                                              expected to have greatest impact on upstream fish popula-
                                              tions. Competition for food among fish in  upstream areas
                                              may be a common phenomenon, particularly during sum-
                                              mer. Thus,  disturbances  that further  destabilize food

-------
resources may ultimately contribute to instability in fish
community structure.
Table 3.
Pair
Invertebrate Densities (No./0.0116m2) from
Paired Exclosures in Jordan Creek. Open
Exclosures were Accessible to Fish, While
Closed Ones were Not
   Month
Open
1
2
3
4
5
6
June
June
June
September
September
September
72
11
11
77
51
140
108
93
34
84
191
210
Effects of Cover and
Current on Predation Rate

A series of experiments were conducted in a laboratory
stream to examine the influence of habitat variables on
predator-prey interactions among fishes.9 Several hypoth-
eses were tested: 1) fish seek cover to avoid predators; 2}
presence of cover decreases predation rates; 3) fish seek
cover to avoid  current.  Experiments  included  various
combinations (presence or absence) of prey, predators,
cover, and current.

Small fish (4 species) were attracted to cover (plastic plants)
in the absence of current, but not in the presence of current
(4-15 cm/s).  In  addition, small fish were  more strongly
attracted to cover when predators (large fish) were absent
than when  predators  were present. The presence  of
predators also inhibited activity (swimming) of small fish.
Small centrarchids exhibited stronger associations with
cover than did small cyprinids.

Presence of cover  did not  affect average predation rates
suffered by small fish in these experiments, though rates
were  more variable in  the presence of cover. However,
centrachids suffered higher mortalities in experiments
without current than  in those with current.

In conclusion, these experiments indicate that small fish
may effectively  alter their  habitat  use so as  to avoid
predators. Furthermore, predation on fish by other fish may
be  more important to  community organization in non-
turbulent (i.e. lentic)  environments than in turbulent (i.e.
lotic) ones.

Conclusion

Overall, the study results clearly show that factors limiting
biotic integrity in warmwater stream ecosystems are not
restricted to water quality. Indeed,  physical habitat  is a
major determinant of biotic integrity. The role  of physical
habitat conditions  includes  direct effects on  fish abun-
dances as well as indirect  effects resulting from complex
                        interactions with channel hydraulics, availability of food
                        (both primary and secondary production), and susceptibility
                        to predators. The interactions of these and other variables
                        are exceedingly complex and require that water resource
                        planners consider factors  in  addition  to  water quality
                        (physical and chemical attributes) in efforts to restore and
                        maintain biotic integrity.
Closed    Literature Cited
                                             1.  Karr, J. R. and P. R. Dudley. 1981. Ecological perspec-
                                                tive on water quality goals. Environmental Manage-
                                                ment. 5:55-68.
                                             2,  Karr, J. R. 1 981. Assessment of biotic integrity using
                                                fish communities. Fisheries. 6:21-27.
                                             3.  Schlosser,  I. J.  1982.  Fish community structure and
                                                function along two habitat gradients in a headwater
                                                stream. Ecological Monographs. 52:395-414.
                                             4.  Schlosser,  I. J. 1982. Trophic structure, reproductive
                                                success,  and growth rate of fishes  in a natural and
                                                modified  headwater stream.  Canadian Journal  of
                                                Fisheries and Aquatic Sciences. 39:968-978.
                                             5.  Karr, J. R., L. A. Toth, andG. D. Carman. 1983, Habitat
                                                preservation for midwest stream fishes: principles and
                                                guidelines. U.S.  Environmental  Protection Agency,
                                                Corvallis, OR. EPA-600/3-83-006. 120 pp.
                                             6.  Angermeier, P.  L. and J.  R. Karr. 1983. Functions of
                                                woody debris in a small warmwater stream. Trans. Am.
                                                Fish. Soc. (in review).
                                             7.  Angermeier, P. L. 1983. Effects of depth and cover on
                                                fish distributions in selected  Illinois  streams. Ecoi.
                                                Monogr. (in review).
                                             8.  Angermeier, P. L. 1983. Effectsof fish consumption on
                                                the abundance of benthic invertebrates in  an  Illinois
                                                stream. Chapter in Ph.D. Dissertation, University  of
                                                Illinois at Urbana-Champaign.
                                             9.  Angermeier, P. L. 1983. Effects of current and preda-
                                                tors on cover use by fish in a laboratory stream. Chapter
                                                in Ph.D. Dissertation, University of Illinois at Urbana-
                                                Champaign.

-------