Effects Of Agriculture On Stream Fauna In Central Indiana (june 1983)

United States
Environmental Protection
Agency
Environmental Research
Laboratory
Cor vail is OR 97333
Research and Development
EPA-600/S3-83-020 June 1983
4MER& Project Summary
Effects of Agriculture on
Stream Fauna in Central Indiana
James R. Gammon, Michael D. Johnson, Charles E. Mays, David A. Schiappa,
William L Fisher, and Bradley L Pearman
From 1978 through 1980 the benthic
macroinvertebrate and fish communi-
ties of three stream systems in Central
Indiana were examined. The objective
of this study was to describe the
organization of these communities in
relation to different land use. The
influence of agriculture on the 14
stream segments ranged from virtually
none to intense, and included some
drainage from animal feed lots. The
results of the study suggest the pattern
of change caused by the increasing
development of agriculture in small
watershed streams.
Initially agriculture may lead to an
expanded biomass of fish and macroin-
vertebrates without causing a large
compositional reorganization.
However,chironomids assume a
dominant role for the macroinverte-
brates while other benthic groups
become secondary in importance.
These benthic changes appear to occur
without strongly influencing the fish
community, except for an increase in
standing crop.
Further development of agricultural
stresses causes a sudden, pronounced
shift in the composition of the fish
community. Communities dominated
by insectivores and piscivores
(centrarchids in these streams) are
converted to communities dominated
by omnivores, herbivores, and
detritivores. This alteration may occur
with little or no change in standing crop
biomass. At this stage the density of
non-chironomid insect larvae becomes
reduced.
Additional stress, such as organic
loading from animal feed lots, causes
further reductions in diversity and
density of macroinvertebrates and in
lowered standing crops of fish.
The near-stream, riparian part of the
watershed is vital to the maintenance of
healthy aquatic communities, acting as
a buffer between plowed fields and
farm animals and the aquatic system.
This Project Summary was developed
by EPA's Environmental Research
Laboratory. Corvallis, OR. to announce
key findings of the research project that
is fully documented in a separate report
of the same title (see Project Report
ordering information at back).

Introduction
The negative effects of various diffuse
and sporadic agricultural activities on
water quality are usually assumed, but
rarely measured. In 1977, a Model
Implementation Program (MIP) was
initiated by the U.S. Environmental
Protection Agency (EPA) and the U.S.
Department of Agriculture (USDA) to
demonstrate the benefits of improved
water quality to be gained by
implementing best management
practices. The Indiana Heartland Model
Implementation project was selected to
be part of this program.
The overall Indiana project effort
includes a study of land use (Holcombe
Research Institute, Butler University),
water quality monitoring and soil erosion
modeling (Agricultural Engineering
Dept., Purdue University), efforts to
institute improved agricultural practices
(USDA, Soil Conservation Service), and a
biological investigation (Dept. of Zoology,
DePauw University). The goal of the
biological phase of this project was to
monitor faunal changes that might occur
in the study streams in response to the
institution of better land management
practices in the respective watersheds.
Biological investigations of the streams
included assessments of (1) benthic

-------
macroinvertebrates, (2) fish populations,
(3) riparian vegetational patterns, and (4)
stream morphology and habitat.
  Sampling  stations were established
near the bases of six  subtributaries of
Eagle Creek and near a fixed monitoring
station on Finley Branch. They were also
established  near  the  bases and about
halfway up both north and south forks of
Stotts Creek, as well as on the lower main
stem. In addition  to Eagle Creek (60%-
75% cropland)  and Stotts Creek (53%-
58% cropland) a largely forested water-
shed. Rattlesnake  Creek, was included in
the biological sampling program  in  the
event that suitable  "control" tributaries
were absent from  the study streams.
  Over the three-year period, the aquatic
communities exhibited an unexpected
stability in terms of standing  crop  and
composition. No  discernible  improve-
ment  in  the  fish communities was
observed. The biotic response implied in
this study  is  non-linear  in  that  an
increment  of  improvement   in  water
quality does not necessarily result in an
increment of improvement in the aquatic
community.  Rather, the pattern  dis-
cerned  for these stream communities
suggests that progressive basic improve-
ments in environmental quality could
occur with little or no evidence  of benefit
to the  stream community until a critical
level was reached, whereupon  a sudden
transformation would occur.

Results
  Fish populations were assessed twice
each year using an  electric seine as the
primary collecting  tool.  In addition to
standing  crop  determinations,
community analyses were performed on
the  data  including Shannon  diversity,
evenness, a composite index incorpora-
ting both abundance values and diversity,
cluster analysis, and trophic grouping.
  These community parameters were
examined  in  relation  to  (1)  habitat
structure, (2) Norton stream order, and (3)
mean  depth of  the  stations.  Habitat
structure tended to be fairly uniform from
station to station  (Fisher,  1979)  and
correlated  poorly with the  community
parameters. Habitat evaluation utilized a
numeric ranking  system  with  regard to
mean  depth, bottom substrate, current,
cover,  and  canopy.   There  was   no
correlation between community param-
eters and stream order (III-V in Stotts, ll-lll
in Eagle, and III for Rattlesnake). During
1978  there was  a positive correlation
between mean depth and standing crop.
  The two collecting stations on Rattle-
snake Creek differed widely in standing
crop (Table 1) with 73 kg/ha at the upper
station and 173 kg/ha in the downstream
station.  Shannon  diversities  at  both
stations  were relatively  high,  ranging
from 1.8 to 2.5. The mean size offish was
larger than the fish at most of the agricul-
turalized  stations. The trophic structures
of  the   Rattlesnake  Creek  fish
communities were very comparable, with
more   than  25%  piscivores  (mostly
centrarchids),  40%-50%  insectivores,
and  only   25%-30%  herbivores,
omnivores, and detritivores.
  In the  agricultural  watersheds,  the
standing  crops of fish also varied greatly.
With some exceptions, the mean size of
fish was smaller primarily because the
communities  were  dominated by
minnows, and Shannon diversities were
generally lower.
  Two sites in Stotts Creek (S1  and S5)
supported  fish   communities   whose
trophic structure was fairly comparable to
those  in Rattlesnake  Creek, but other
sites,   particularly those  influenced  by
swine  feed  lots, were  dominated  by
detritivores, omnivores, and herbivores.
  Among the Eagle Creek sites, both E5
and E6 contained populations trophically
similar to those of Rattlesnake Creek, but
other  sites deviated substantially  with
piscivores and insectivores constituting
less than half of the biomass of fish.
  The  cluster analysis produced results
very comparable to the trophic analysis in
that  stations  having  communities
dominated by piscivores and insectivores
(R-stations, E5,  E6, and  S1) tended to
cluster  together,  as  did  stations
                          dominated by omnivores.herbivores, and
                          detritivores.  A  plot  of  the  percent
                          piscivores and insectivores in relation to
                          standing crop illustrates the two basical-
                          ly different  kinds  of fish communities
                          found among the 14 collecting stations
                          (Figure  1).  Subwatersheds  containing
                          more  cropland  were consistently
                          characterized by lower percentages of
                          piscivores and insectivores. Spearman's
                          correlation coefficient was -.714 for the
                          Eagle Creek system  and -.90 for Stotts
                          Creek (Hyde et al. 1982).
                            The analysis of the benthic macroinver-
                          tebrate populations proceeded from two
                          basic assumptions:  (1) that  the water
                          quality in each drainage area would be at
                          least as good as that in the least disturbed
                          stream or reach of that watershed, and (2)
                          that the assessments of water quality are
                          possible using identifications only to the
                          family level and that any further identifi-
                          cation to  genus and  species would not
                          significantly alter those assessments. It
                          was also assumed that the composition of
                          the  benthic community would  change
                          naturally as the year progressed because
                          most of the organisms spend  only part of
                          their lives in that environment.
                            The insects were the dominant group of
                          organisms in the benthic community and
                          the  analyses emphasize those animals.
                          The mean annual diversity based upon
                          density of insect families at each site was
                          used for a preliminary ranking of the sites
                          and also for more  detailed analyses.
                          Biomass  was  less  useful because its
                          magnitude depended upon which instars
                          were  present.  The  growth  rate  and,
Table 1.    Mean Standing Crops and Trophic Composition of Fish in Eagle Creek (E). Stotts
          Creek IS), and Rattlesnake Creek (R) During 1978. 1979. and 198O

                                           % Composition
 Sampling   Standing   	—
  Station  Crop - kg/ha Piscivores  Insectivores  Herbivores   Omnivores   Detritivores
    E1
    E2
    E3
    E4
    E5
    E6
    E8
    S1
    S3
    S5
    S7
    S9
    R1
    R3
 91
152
111
 68
215
142
252
 91
194
177
 69
141
 73
173
11.3
10.4
 9.5
12.0
26.2
30.5
10.5
17.3
 7.9
25.5
 2.4
 3.5
28.2
26.4
Eagle Creek
32.1
35.6
26.1
16.1
47.0
49.9
26.1

8.6
16.1
16.9
24.0
6.4
12.4
7.5
  Stotts Creek
    54.3
    32.6
    37.5
    27.9
    33.5
 6.6
14.4
15.6
32.9
21.8
Rattlesnake Creek
    42.6         2.1
    56.2         5.9
                           46.3
                           32.6
                           40.7
                           42.1
                           19.1
                            7.2
                           54.0
13.9
31.9
18.7
30.1
29.6
           23.8
            9.3
                       0.9
                       5.5
                       6.3
                       5.9
                       1.6
                       3.2
                       2.1
6.3
6.2
3.1
6.7
9.0
            3.3
            2.1

-------

1
400

300

200

100

40
20
10
• 1978
O 1979
D 1980
0 10 20 30 40 50 60 70 80 90 100

% Piscivores and insectivores

Figure 1. Standing crop and proportion ofpiscivores and insectivores of study segments. Each
point is a mean of two determinations.
hence, synchrony of instars varied from
site to site in any month.
The benthic macroinvertebrate popu-
lations of Eagle, Stotts, and Rattlesnake
Creek which ranked highest were used as
standards for evaluating the other sites.
Their respective compositions were
similar with the number of insect families
ranging from 10 to 20.
Chironomidae were major components
of all sites during May and June, but
diminished in importance in July and
August at the "good" sites while
maintaining high densities at the "poor"
sites. During this period benthic diversity
correlated negatively to % cropland in
subwatersheds (r = -.771) of Eagle Creek.
After the decline of Chironomidae at the
"good" sites Trichoptera, especially
Hydropsychidae, became the dominant
insect group with other major
contributions from the Baetidae, Elmidae,
and Simuliidae.
Organic and inorganic pollutants were
major influences on the macroinverte-
brate communities. The nonpoint source
inorganic sediments tended to depress
the density, except for the Chironomidae,
without greatly altering the composition .
Point source pollutants, on the other
hand, altered the taxonomic composition
of the community. Many of the less
common families collected at other sites
never occurred while others occurred
infrequently and, when present, in small
numbers. These may represent
contributions from upstream drift. These
affected communities are dominated by
Chironomidae, Oligochaeta, and Nema-
toda, which on occasion reach higher
than expected densities. The same
pattern was exhibited where pastured
cattle had access to the creek, except for
an initially higher number of families
(represented by few specimens).
Detectable improvements in benthic
communities followed cessation of bridge
building and bulldozing, but not in areas
subjected to persistent, chronic
problems.
Table 2 summarizes the degree to
which stream communities are altered by
differing degrees of agriculturalization.
With regard to fish communities, the term
"moderate" agriculture indicates that all
the normal components are present,
whereas, under "heavy" agricultural
usage a pronounced change occurs as
bass, sunfish, and some insectivores are
lost from the community.

Stream Morphology, Habitat,
and Streambank Vegetation
The major tributaries of the streams
were examined at several sites for (1)
stream habitat, (2) channel morphology,
(3) bottom substrate, and (4) adjacent
land use. All sites were located in the
lower half of the drainage basins.
Stream habitat structure included
seven parameters, including for both
riffles and pools such characteristics as
length, width, depth, flow, aquatic flora,
bottom composition, and cover. Point
determinations were made with a meter
pole at 5 m intervals, a technique similar
to that used by Gormann and Karr (1978).
Flow rates were determined with a
Mikasa flow meter. Seven categories of
bottom composition were identified using
the Wentworth scale.
Two aspects of land use were
examined: (1) the predominant type of
land use adjacent to the stream, and (2)
the presence or absence of a different
kind of interface or buffer strip
immediately adjacent to the stream. The
various types of predominant land use
were (1) cultivated fields of soybeans or
corn, (2) meadows, (3) woods and forests,
and (4) residential areas. Buffer strips or
interfaces, when present, consisted of
either grasses, scattered trees or woods.
Within a geographic region there tends
to be a remarkable consistency in the
dimensions of stream channels in
watersheds of similar size (Dunne and
Leopold, 1978). Most subtributaries were
25 to 75 km2 in drainage basin area and
might be expected to maintain quite
similar channel dimensions if land use
throughout the basins was equitable.
Increased sediment input results in
aggradation with a resultant increase in
channel width and decrease in depth
(Bovee and Milhous, 1978). Reduced
sediment loading would reverse this
tendency. Mean pool depth was
examined in relation to land use and
potential erosion. The deepest pool from
each watershed was assigned a value of
100 and the mean depth of other pools
rated as a percent of this reference depth.
The data were then grouped with respect
to predominant land use and the
presence or absence of different kinds of
buffer vegetational communities. Figure
2 summarizes the findings.

-------
Table 2. Summary of Community Characteristics Under Different Intensities of Agriculture
Degree of Agriculturalization
Characteristic
Minimal
Moderate
Heavy
Macroin vertebrates
Shannon diversity (density of insect families)
Number of insect families
Chironimidae (temporal pattern)

Oligochaeta (density)
Other families (density)
Hydropsychidae
12 to 15
High: May June
Low: July August
Low
High
Dominant throughout summer
1.3 to 1.5
8 to 12
Relatively high throughout
summer
Fluctuating
Declines for uncommon families
Common, but not
always dominant

-------
700

fin
OCX
60
40
20
o

?
j- 80
»5
&
Q
§ 60
Q.
I
Relative M
% S
0

60
40
20
0

A. Rattlesnake Creek
80 80

_
40

0
50
I

[ | w/o buffer
[444 with buffer _^
B. Stotts Creek

-
68
60

34
••••M