&ER&
United States
Environmental Protection
Agency
Environmental Research
Laboratory
Corvallis OR 97333
Research and Development
EPA-600/S3-82-003 Jan. 1983
Project Summary
Agricultural Runoff and
Reservoir Drawdown
Effects on a
2760-Hectare Reservoir
Byron H. Shaw
The 2760-hectare Big Eau Pleine
Reservoir in Marathon County, Wiscon-
sin has experienced frequent winter fish
kills and summer algae blooms since its
construction in 1937. From 1974 to
1979 a study of the reservoir and its
945-km2 watershed was conducted in
an attempt to identify and quantify the
sources of water quality problems and
recommend management practices to
reduce these problems.
Land use and nutrient loading studies
in the watershed identified agricultural
runoff, especially animal waste, as the
major source of nutrient loading. Total
phosphorus loss from the watershed
averaged 0.59 kg/ha/yr for the 4-year
period with approximately 60 percent
occurring during the spring snowmelt
and runoff season. Hydrologic and soil
erosion modeling indicated that the
greatest runoff and soil erosion occurred
during spring snowmelt and that much
of the erosion and runoff originates on
the lower slopes and alluvial soils.
Reservoir studies identified Aphanizo-
menon flos-aquae as the major bloom-
producing alga. Chlorophyll a values for
the four summers averaged 105 ju g/l,
ranging from 65 in 1978 to 120 in
1976. Yearly variations in chlorophyll a
did not correlate with differences in
yearly or seasonal total phosphorus
loading from external sources. Internal
phosphorus loading appeared to be
more important in determining summer
algae blooms; much of the internal load-
ing is believed to be from drawdown-
related resuspension of sediments.
Total phosphorus levels in the reservoir
begin to increase at about the same
time as summer drawdown begins.
Winter oxygen problems in the reser-
voir were related closely to reservoir
drawdown. Sediment oxygen demand,
long-term BOD studies, and reservoir
monitoring showed that while consider-
able oxygen was lost over winter due to
biological reactions, the reservoir would
not go anaerobic. Winter drawdown
was found to result in scouring of sedi-
ments high in BOD as the reservoir was
gradually drawn down to the old river
channel. This scouring resulted in rapid
loss of the remaining oxygen as draw-
down moved progressively down the
reservoir.
Recommendations include controlling
animal-waste spreading during winter
and increased use of conservation prac-
tices, especially on lower slope portions
of the watershed, including fencing
streams at least 30 feet from the stream
channel. Recommended reservoir man-
agement changes include delaying sum-
mer drawdown to minimize internal
phosphorus loading, delaying winter
drawdown to at least mid-January, and
increasing minimum pool volume by 25
percent.
This Project Summary was developed
by EPA's Environmental Research Lab-
oratory, 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 order-
ing information at back).
Introduction
The 2700-hectare Big Eau Pleine Re-
servoir in Marathon County, Wisconsin,
has a history of excessive algal growth
and winter fish kills. Several major kills
in the early and mid-1970s resulted in
demands by the public for corrective
action. Other studies have labeled the
reservoir very eutrophic and identified
agricultural runoff as the major source
of nutrients.
The present study, initiated in 1 974,
was an attempt to identify the relation-
ship of nutrient sources and reservoir
management to the trophic conditions
and the winter oxygen problem and to
make recommendations for alleviating
these problems. Computer models were
used to characterize the watershed hy-
drology, soil erosion and nutrient runoff
characteristics. Another computer model
was used to quantify the limnological
processes occurring in the reservoir.
The 945-km2 watershed is character-
ized by rolling topography; its silt loam
surface soils are underlain by fairly im-
permeable dense glacial till or residuum.
This creates ideal conditions for surface
runoff during snowmelt or heavy rains.
The reservoir, built in 1937 to augment
the flow of the Wisconsin River, has an
annual water level fluctuation of about
9 meters, amounting to over 95% of its
total volume.
Results
There were three major components
of the project:
1. A land use survey covering 3 per-
cent of the total watershed area, with
use of the universal soil loss equation to
determine conservation needs on each
field in the sample.
2. A modeling project to relate hydro-
logic characteristics in seven sub-basins
of the Hanann Creek sub-watershed to
nutrient loss and land use.
3. A soil erosion modeling project to
relate sediment and associated nutrient
loss to land use and watershed charac-
teristics.
The soil erosion modeling project
showed that 60 to 70 percent of erosion
occurred during spring snowmelt and
runoff. This finding required the calcula-
tion of a new rainfall factor for the
watershed to use in the soil loss equa-
tion, e.g., a factor of 184 compared to
125-150 based on the Soil Conserva-
tion Service annual erosion index. This
finding is consistent with data on nutri-
ent loading which is also highest in early
spring.
Results from both the soil erosion and
the hydrologic models indicate that the
largest overland flow and soil erosion
occur on the lower slope portions near-
est the stream. Soil transport to the
stream was lowest where there was a
permanent vegetation strip between
cultivated fields and the stream chan-
nel, supporting the concept of vegetated
buffer strips to reduce sediment and
nutrient transport to streams. Winter
manure spreading on these lower slope
areas should be avoided.
The land use survey revealed the
following percentages of total land use:
corn, 13.8; oats, 14.4; hay, 29.5; pas-
ture, 22.3; forest, 14.0; idle land, 2.1;
other, 4.2.
Results of the soil erosion project indi-
cated that the average soil loss for the
watershed was 4.82 metric tons per
hectare compared to the "allowable soil
loss" established by the Soil Conserva-
tion Service of 6.72 metric tons/ha.
Twenty-four percent of the total land
area exceeded the allowable soil loss.
Cost estimates for erosion control are
presented in Table 1. These costs are
based on 13 soil conservation plans for
erosion control in the Big Eau Pleine
Watershed and on 1978 cost estimates.
Major practices recommended were
diversions and grass waterways with
lesser amounts of contour strips and
terraces. The additional expense of con-
trolling nutrient loss from barnyards and
manure storage facilities would increase
this cost greatly.
Reservoir Studies
Nutrient loading
Nutrient and BOD inflow and outflow
to the reservoir were determined using
daily values collected with an autometer
sampler on the Big Eau Pleine River and
bi-weekly samples from other tribu-
taries and at the dam. Table 2 shows
the annual summary of these data for
the four years of the study. Daily, monthly
and yearly data vary considerably be-
cause of annual differences in rainfall.
Total phosphorus loading alone ranged
between 26.4 to 83.3 metric tons per
year; greatest loading occurred in the
early spring, except in 1978, when
heavy July rain created large summer
runoff events. Nitrogen and BOD load-
ing followed the trend observed for
phosphorus.
Reservoir discharge and retention'of
nutrients was almost as variable as
loading because of rainfall and runoff
variability and the effect of reservoir
management on nutrient retention.
Phytoplankton response to nutrient
loading was determined by chlorophyll
analysis, supplemented by 1*C primary
production studies and plankton counts
during 1975 and 1976. Primary pro-
duction for 1975 and 1976 was 340
and 685 gC/m2, respectively; mean
summer chlorophyll values were 52.2
and 103 mg/l. These data support pre-
vious studies indicating that the reser-
voir is hypereutrophic with summer
phytoplankton dominated by Aphanizo-
menon flos-aquae.
Total phosphorus loading and chloro-
phyll a data presented in Figure 1 and
Table 3 indicate that the annual phos-
phorus loading does not correlate closely
enough to chlorophyll a levels to use
simple phosphorus loading models for
the Big Eau Pleine Reservoir. Seasonal
phosphorus loading as represented in
Figure 1 also shows poor correlation to
chlorophyll a.
In this reservoir, the major processes
affecting annual primary production are
seasonal phosphorus loading from agri-
cultural runoff, internal phosphorus
loading, solar radiation, and retention
Table 1.
Area to be
Treated
Estimates of Cost to Implement Various Soil Conservation Practices for
Erosion Control in the Big Eau Pleine Watershed
Cost per Total Cost
Ha Estimate (Rounded)
Basis for
Treatment
14,928 ha $155.14
23,587 ha $155.14
55,315 ha $155.14
$2,300,000 Areas exceeding 6.7 tons/ha
soil loss using rainfall factor
125-150
$3,7OO.OOO Areas exceeding 6. 7 tons/ha
soil loss using rainfall factor 184
$8,500,000 All cropland
-------
Table 2. Yearly Inflow and Outflow of Nitrogen, Phosphorus, and BOD, and Water Volumes for the Big Eau Pleine Reservoir
1975-1978 (Metric Tons)
1975
In Out
Total
Phosphorus
Reactive
Phosphorus
Kjeldahl
Nitrogen
N03 & N02
Nitrogen
BOD5
Water
(Hectometers)
63.4
41.7
312.2
168.9
650.1
204.5
16.2
8.4
241.4
34.2
504.0
179.7
1976
In Out
50.5
27.0
451.6
195.8
402.6
222.2
56.4
23.5
523.7
114.1
896.6
328.2
1977
In Out
26.4
19.0
168.4
77.4
148.4
132.6
4.8
2.3
66.3
15.9
107.6
63.6
1978
In Out
83.3
53.7
530.1
224.3
565.7
369.2
50.9
20.9
647.1
106.1
1697.8
454.4
Total
In Out
223.6
141.4
1462.3
666.4
1803.2
928.5
128.3
55.1
1478.5
270.2
3206.0
1025.9
Table 3.
Yearly and Seasonal Total Phosphorus Loading, Summer Drawdown, and Average Chlorophyll a Values for the Big Eau
Pleine Reservoir
Total P Inflow (Metric Tons)
Summer Drawdown
Meters
A verage
Chlorophyll a
(mg/l)
Yearly
Jan. 1-July 1
June 1-Sept. 1
June 1-Sept. 1
July-August
1975
1976
1977
1978
63.4
50.4
26.4
83.3
44.2
48.9
9.0
36.6
5.3
1.4
1.5
33.1
1.31
2.53
0.61
0.0
104
126
, 105
65
Chlorophyll a
«---o-~<>--o Total phosphorus -
- 0.02
J
May J O J A J O J A J 0 J A J O
1975 1976 1977 1978
Figure 1 . Chlorophyll a and total phosphorus for the Big Eau Pleine Reservoir
1975-1978.
time. While all are important, the data
generated during this project indicate
that internal loading from sediment dis-
turbance during reservoir drawdown is
a major factor affecting phosphorus
levels and primary production. Figure 2
shows the relationship between reser-
voir volume and total phosphorus con-
centrations. Phosphorus levels are high
following spring runoff, decline rapidly,
increase again when summer drawdown
is initiated and then remain high through-
out the summer. The exception was
during 1978, when heavy rains kept
water levels near capacity for the entire
summer, resulting in a more rapid water
exchange, little sediment disturbance,
more turbid water and consequently
lower chlorophyll levels (Figure 2).
Reservoir Oxygen Relationship
Studies of oxygen demand by runoff
water, phytoplankton and sediments
showed that large increases of both
BOD5 and sediment oxygen demand re-
sult from summer phytoplankton popu-
lations. It was, however, found that
-------
Volume
—* ° Total phosphorus
empty
- 0.02
J A J 0 J AJOJAJOJAJOJAJO
1974 1975 1976 1977 1978
Figure 2. Reservoir volume and total phosphorus concentrations for the Big Eau
Pleine Reservoir 1974-1978.
these oxygen demands alone are unlikely
to result in the complete oxygen deple-
tion that occurs in the reservoir during
the winter. Winter oxygen depletion
was found to occur from the bottom up-
ward and gradually throughout the
water column, as is typical of eutrophic
lakes. However, as winter drawdown
progresses, often there is a complete
loss of oxygen beginning at the upper
end of the reservoir and moving gradually
downstream. This process is believed
to be caused by scouring and resuspen-
sion of sediments high in settled organic
matter, and therefore oxygen demand,
as the reservoir is drawn down to its old
river channel. This often occurs when
the reservoir volume is reduced to about
40 percent of its capacity (Figure 3).
Conclusions
The problems in the Big Eau Pleine
Reservoir are due both to land-use ac-
tivities in the watershed and to reservoir
management.
The main cause of high nutrient
loading to the reservoir is agriculture
and related activities in a watershed
where surface runoff is seasonally high.
Family-owned dairy farms account for
over 90 percent of the land-use in the
watershed; winter spreading of manure
and cattle grazing up to and often
through the stream channel are very
common. These sources of animal
wastes, coupled with fertile soils that
have slow infiltration rates, result in
high nutrient losses to the drainage.
The heaviest nutrient loss from the
watershed occurs during snowmelt and
spring rains which coincide with the
period when the reservoir is being refilled
after winter drawdown. Approximately
60 to 70 percent of the annual total
phosphorus load enters the reservoir
during March and April.
The spring runoff results in initially
high total phosphorus concentrations in
the reservoir (0.2 mg/l), which stimu-
late an early bloom of diatoms and green
algae. However, the phosphorus con-
centration decreases rapidly due to sed-
imentation and algal uptake and only
begins to increase again when water
levels are lowered. Phosphorus is re-
leased by wind agitation of sediments
exposed during drawdown and, to a
lesser extent, from anaerobic sediment
release. This process increases total
phosphorus concentrations throughout
the drawdown period which, in turn,
support exceptionally large blooms of
Aphanizomenon flos-aquae.
Winter oxygen depletion in the reser-
voir and subsequent fish kills are, how-
ever, the major water quality concern.
Oxygen depletion proceeds from the
bottom upward as is common in eutro-
phic lakes. But this process alone would
not result in anaerobic conditions
throughout the entire reservoir. As
drawdown proceeds, the upstream por-
tions of the reservoir are converted to a
river condition, resulting in resuspension
of sediments that have accumulated in
the deeper water during spring runoff
events and summer algae production.
Because these sediments are anaerobic
and high in organic matter, reduced iron
and sulfur, they extract any remaining
oxygen in the water they contact. Thus,
the condition moves downstream, grad-
ually removing remaining oxygen in the
lower reaches of the reservoir as draw-
down and sediment resuspension con-
tinue. During winters with high tributary
flows and delayed drawdowns, less
severe oxygen depletion occurs.
The computer models used in this
project were generally helpful in quanti-
fying environmental processes and in
identifying major factors to be con-
sidered in solving the watershed and
reservoir problems. However, all of the
models used need further refinement;
ideally they should be linked to make
modeling a more useful management
and research tool.
Recommendations
Solutions to the water quality prob-
lems in the Big Eau Pleine Reservoir are
expensive and complex. Recommended
actions fall into two categories—land
use and reservoir management.
A. Land Use
1. Overwinter manure storage should
be provided on each farm; manure
applied to the land should be in-
corporated into the soil without
delay.
2. Stream channels should be fenced
to exclude livestock, except at
-------
t
I
o
T3
I
O
JAJOJAJOJAJOJA JOJAJOJAJOJ
1974 1975 1976 1977 1978 1979
Figure 3. Reservoir volume and dissolved oxygen for surface samples 1974-1979,
well-chosen watering sites. Buffer
strips should be a minimum of nine
meters on each side of the stream.
3. Soil conservation plans should be
prepared and implemented on
each farm to reduce soil erosion
and associated nutrient loss.
4. Small runoff-retention ponds
should be constructed to act as
sediment and nutrient traps and
thus result in higher base flow
levels in the river by increasing
ground-water recharge.
5. Land disposal of cheese factory
wastes should follow the same
procedures as for animal wastes.
However, caution must be used
to avoid overloading soils, since
these wastes are very high in
nitrogen and phosphorus. Lagoon
storage and summer spreading
would prevent much of the pre-
sent nutrient loss.
B. Reservoir Management
Even if all of the proposed land-use
recommendations were followed, there
probably would still be serious water
quality problems in the Big Eau Pleine
Reservoir. The following guidelines would
be helpful in managing the reservoir to
overcome existing problems and those
that wfll persist even after land-use
practices have improved.
1. Summer nutrient levels and result-
ing algae blooms could be reduced:
a. If early spring runoff, which
has the highest phosphorus
levels, could be allowed to
flow through the reservoir be-
fore refilling is started. This
would reduce the amount of
phosphorus retained in the res-
ervoir for use by algae later in
the year. However, this proce-
dure could only be followed
during years in which a large
snow pack exists.
b. Since resuspension of sediment
phosphorus has been shown to
be a significant source for
summer algae blooms, it would
be desirable to hold the reser-
voir at as constant a level as
possible, using short periods of
rapid drawdown instead of
constant slow drawdown.
2. Winter oxygen problems could be
minimized by:
a. Allowing the reservoir to fill as
much as possible during the
fall prior to freeze-up.
b. Delaying any winter draw-
down until at least mid-
January.
c. Stopping drawdown when the
reservoir volume is reduced to
25 percent of the full volume.
If these recommendations are not
feasible for economic or political rea-
sons, it may be possible to minimize
winter fish kills using mechanical aera-
tion; however, even with aeration, a
modified drawdown may be necessary.
Bryon H. Shaw is with the College of Natural Resources, Stevens Point, Wl
54481.
Charles F. Powers is the EPA Project Officer (see below).
The complete report, entitled "Agricultural Runoff and Reservoir Drawdown
Effects on a 2760-Hectare Reservoir, "(Order No. PB82-186 529; Cost: $9.OO,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, OR 97333
U. S. GOVERNMENT PRINTING OFFICE: 1983/659 -095/575
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
PS 0000329
230 S DEARBORN STREET
CHICAGO 1L 60604
------- |