United States
Environmental Protection
Agency
Robert S. Kerr Environmental Research
Laboratory
Ada OK 74820
EPA-600/2-79-143
August 1979
Research and Development
&EPA
Livestock Feedlot
Runoff Control by
Vegetative Filters
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. US Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. 'Special" Reports
9, Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-143
August 1979
LIVESTOCK FEEDLOT RUNOFF CONTROL BY VEGETATIVE FILTERS
by
Dale H. Vanderholm, Elbert C. Dickey, Joseph A. Jackobs,
Roger W. Elmore, and Sidney L. Spahr
University of Illinois
Urbana, Illinois 61801
Grant No. R804341-01-1
Project Officer
R. Douglas Kreis
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
The Environmental Protection Agency was established to coordinate
administration of the major Federal programs designed to protect the
quality of our environment.
An important part of the agency's effort involves the search for
information about environmental problems, management techniques and new
technologies through which optimum use of the nation's land and water
resources can be assured and the threat pollution poses to the welfare
of the American people can be minimized.
EPA's Office of Research and Development conducts this search
through a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental
Research Laboratory is responsible for the management of programs to:
(a) investigate the nature, transport, fate and management of pollutants
in groundwater; (b) develop and demonstrate methods for treating waste-
waters with soil and other natural systems; (c) develop and demonstrate
pollution control technologies for irrigation return flows, (d) develop
and demonstrate pollution control technologies for animal production
wastes; (e) develop and demonstrate technologies to prevent, control
or abate pollution from the petroleum refining and petrochemical in-
dustries, and (f) develop and demonstrate technologies to manage pol-
lution resulting from combinations of industrial wastewaters or indus-
trial/municipal wastewaters.
This report contributes to the knowledge essential if the EPA
is to meet the requirements of environmental laws that it establish
and enforce pollution control standards which are reasonable, cost
effective and provide adequate protection for the American people.
a
William C. Galegar, Director
Robert S. Kerr Environmental
Research Laboratory
iii
-------
ABSTRACT
This research program was initiated with the overall objective of de-
termining if contaminated runoff from small livestock feedlots could be
successfully controlled by infiltration and treatment on vegetated field
areas (vegetative filters). A secondary objective was to develop standard
design criteria for vegetative filters.
The vegetative filter can be described simply as a system in which a
vegetative area such as pasture, grassed waterway or even cropland is used
for treating runoff by infiltration, settling, dilution, filtration, and
absorption of pollutants. Four full-scale vegetative filters were designed
and installed on feedlots in central and northern Illinois. Two configur-
ations were used-channelized flow and overland flow. After settling for
partial solids removal, runoff was applied directly to the filters and
allowed to flow from the inlet to the outlet section. Monitoring included
measurement, sampling and analysis of influent runoff, effluent runoff,
runoff at intermediate points, ground water, soil, and forage produced on
the filter area.
Runoff from most smaller rainfall events infiltrated completely, re-
sulting in no discharge. Runoff from larger events partially infiltrated
and partially discharged. Discharge samples analysis indicated a removal
of over 95 percent of nutrients and oxygen demanding materials in the
applied runoff on a mass balance basis and 80 percent removal on a concen-
tration basis. Removal was found to be directly related to flow distance
or contact time with the filter. Greater flow depths with channelized flow
required greater contact time or flow distance than did shallow overland
flow to achieve the same level of treatment.
Design criteria were developed for overland flow and channelized flow
vegetative filters. These include the basic philosophy that small runoff
events will be infiltrated and runoff from larger events will be allowed
to discharge after being treated to an acceptable degree.
This report was submitted in fulfillment of Contract No. R804341-01-1
by the departments of Agricultural Engineering, Agronomy, and Dairy Science,
University of Illinois at Urbana-Champaign under the partial sponsorship
of the U.S. Environmental Protection Agency. This report covers a period
from February 9, 1976 to May 8, 1978 and work was completed as of May 8,
1978.
IV
-------
CONTENTS
Foreword ill
Abstract iv
Figures vi
Tables viii
Abbreviations and Symbols x
Acknowledgment xi
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Experimental Procedures 6
5. Results and Discussion . . 16
6. Economics 59
References 62
Bibliography 64
-------
FIGURES
Number Page
1. Location of vegetative filter systems studied 7
2. Alternative configurations for vegetative filters used as
a treatment for feedlot runoff 8
3. System 1 - runoff collection and vegetative filter configu-
uration 9
4. System 2 - runoff collection and vegetative filter config-
uration 10
5. System 3 - runoff collection and vegetative filter config-
uration 11
6. System 4 - runoff collection and vegetative filter config-
uration 12
7. Rainfall-runoff relationship for a paved dairy feedlot
(System 1) 17
8. Nitrogen concentration changes with overland flow (System 1). 13
9. COD and solid concentration changes with overland flow
(System 1) 20
10. Nitrogen concentration changes with channelized flow
(System 3) 24
11. COD and solid concentration changes with channelized flow
(System 3) 25
12. Nitrogen concentration changes with overland flow—curvi-
linear regression (System 1) 26
13. Nitrogen concentration changes with channelized flow—curvi-
linear regression (System 3) 27
14. Nitrogen concentration changes with channelized flow for an
individual storm (System 4) 30
15. COD and solid concentration changes with channelized flow for
an individual storm (System 4) 31
VI
-------
Number Page
16. Dry matter yields at System 1 33
17. Nitrogen uptake by the forage on System 1 36
18. Phosphorus uptake by the forage on System 1 37
19. Potassium uptake by the forage on System 1 38
20. Nitrate nitrogen in the soil profile at System 1 39
21. Available phosphorus in the soil profile at System 1 40
22. Available potassium in the soil profile at System 1 41
23. Nitrate nitrogen in the soil profile at System 3 at a
30-meter (m) flow distance 44
24. Nitrate nitrogen in the soil profile at System 3 at a 120-m
flow distance 45
25. Nitrate nitrogen in the soil profile at System 3 at a 275-m
flow distance 46
26. Nitrate nitrogen in the soil profile at System 3 at a 425-m
flow distance 47
27. Total nitrogen in the soil profile at System 3 at a 30-m
flow distance 48
28. Total nitrogen in the soil profile at System 3 at a 120-m
flow distance 49
29. Total nitrogen in the soil profile at System 3 at a 275-m
flow distance 50
30. Total nitrogen in the soil profile at System 3 at a 425-m
flow distance 51
31. Approximate required channelized flow distances for various
slopes and contact times 54
VII
-------
TABLES
Number Page
1. Constituent Concentrations in the Settling-Basin and
Vegetative-Filter Effluent (System 1) 19
2. Constituent Removal During the Study Period on a Mass-
Balance Basis by Vegetative Filter Treatment of Feedlot
Runoff (System 1) 19
3. Constituent Concentrations in the Settling-Basin and Vege-
tative Filter Effluent (System 2) 22
4. Effluent Constituent Concentrations in the Vegetative Filter
at Various Distances From the Settling Basin Discharge
(System 3) 23
5. Percent Constituent Reduction in the Basin Effluent at
Various Locations in the Vegetative Filter (System 3).... 28
6. Constituent Removal on a Mass-Balance by Vegetative
Filter Treatment of Feedlot Runoff for Three Storms (System 3) 28
7. Constituent Concentrations in the Settling-Basin and the
Vegetative Filter Effluent After a Flow Distance of 148 m
(450 ft.) (System 4) 29
8. Nutrient Uptake at System 1 at Various Cuttings 34
9. Nitrogen Balance Sheet for the Vegetative Filter at System 1. 42
10. Ground Water Quality at the Vegetative Filter at System 1. . 52
11. Minimum Contact Times for Vegetative Filters Utilizing
Channelized Flow for Various Feedlot Sizes 55
12. Minimum Flow Lengths for Vegetative Filters Utilzing Overland
Flow and Having Various Slopes 56
13. Recommended Overland Flow Filter Areas with Various Soil Types
(climatic conditions should be similar to those of central
Illinois 57
viii
-------
Number Page
14. Total Investment Costs, Operating Costs, and Percentage
Difference Between the Vegetative Filter and Zero-Discharge
Systems for Six Illinois Demonstration-Research Sites. ... 60
IX
-------
ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
BOD
COD
IEPA
ac
cm
et al.
ft
ft2
ft3
gal
g
ha
hr
SYMBOLS
pH
l.ON KC1
N03-N
NH3-N
°C
°F
biochemical oxygen demand
chemical oxygen demand
Illinois Environmental
Protection Agency
acre
centimeter
and others
foot, feet
square feet
cubic feet
gallon
gram
hectare
hour
positive hydrogen ion
concentration
1.0 normal potassium
chloride
nitrate-N, nitrate nitrogen
ammonia-N, ammonia nitrogen
degrees Celsius
degrees Fahrenheit
kg
Ib
1
m
m2
m3
mg
ml
mm
mt
ppm
sec
T
in
$
/
P
K
N
Cl
ymhos
y
to
kilogram
pound
liter
meter
square meter
cubic meter
milligram
milliliter
millimeter
metric ton
parts per million
second
ton
inch
dollar
per
phosphorus
potassium
nitrogen
chloride
micromhos
percent
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge additional financial support for
this research from the Illinois Institute for Environmental Quality and the
Illiois Beef Industry Council. Sampling equipment and laboratory services
were provided by the Illinois Environmental Protection Agency. James F.
Frank, agriculture adviser with the Illinois Environmental Protection
Agency, was particularly supportive in initiating and conducting the re-
search. Close communication and cooperation were maintained throughout the
study with Mr. Frank and with Southern Illinois University staff members
who were conducting a complementary study on vegetative filters.
The help of Wayne Seifert in managing the University of Illinois Dairy
system and Steve Maddock and other field and laboratory technicians who
participated in this research is gratefully acknowledged.
Soil Conservation Service personnel assisted in planning and designing
the experimental systems.
Finally, the authors especially appreciate the assistance and coopera-
tion of the livestock producers and landowners involved in this study:
Kendrick Fesler, Barry; Phil Bradshaw, Griggsville; Roger Nordman, Oregon;
Merwin Ness, Big Rock; and Lawrence Strope, Aurora.
xi
-------
SECTION 1
INTRODUCTION
Many areas of the country have significant numbers of livestock feed-
lots not subject to the National Pollutant Discharge Elimination System
(NPDES) permit program. Most are small feedlot facilities, but many of them
may have a potential water pollution problem due to uncontrolled runoff from
open lot areas. Installation of a zero-discharge runoff-control system is
one alternative for solving the pollution threat at each of these locations.
This approach may be economically prohibitive for many of the smaller opera-
tions, however, even though the zero-discharge system is required by regula-
tion in several states. A second alternative is to install a vegetative
filter system to adequately control the runoff so that a violation of water
quality standards will not occur in the case of storm runoff. This alter-
native has the advantages of controlling the runoff at a lower cost than the
conventional zero-discharge system and at the same time requiring less man-
agement .
The vegetative filter can be described simply as a system in which a
vegetative area such as pasture, grassed waterway, or even cropland is used
for treating runoff by settling, filtration, dilution, absorption of pollu-
tants, and infiltration. Most early systems have been designed on the prem-
ise that all or a major portion of the feedlot runoff that does not infil-
trate into the ground in certain situations will be treated to such a degree
that it can be allowed to ultimately enter surface watercourses. Pretreat-
ment of the feedlot runoff by some method such as settling is advisable.
While systems of this type are certainly not adaptable or practical for
every situation, they could provide successful, low-cost runoff control for
many feedlots.
Much of the early use of the vegetative filter method has been for the
disposal of wastes from the canning industry. Mather (1969) reported re-
moval of biochemical oxygen demand (BOD) from cannery wastes of 94 to 99
percent during overland flow in a disposal area, although Bendixen et al.
(1969) reported only 66 percent BOD removal. Nitrogen removals of 61 to 94
percent and phosphorus removals of 39 to 81 percent have also been reported
in these two studies.
Some research has been conducted using vegetative filters for treatment
of livestock wastes. McCaskey et al. (1971) found a renovating effect for
waste water traveling over a grassed surface in a thin layer but did not de-
termine the effect on a quantitative basis. In Ohio, Edwards et al. (1971)
measured significant reductions in the nutrient content of feedlot runoff
after the runoff traversed a grassed waterway. They attributed this reduc-
-------
tion to the deposition of solids in the waterway and to the dilution of
feedlot runoff by surface water from nearby cropland. Research by Kramer
et al. (1974) in Kansas indicated that, for the beef feedlot studied, the
spray-runoff system of removal of BOD5 and total suspended solids was pos-
sibly adequate for discharge but nutrient levels could still be too high for
discharge to be practical.
Sievers et al. (1975) used a grassed waterway type of vegetative fil-
ter to treat anaerobic swine lagoon effluent. Willrich and Boda (1976) also
treated swine lagoon effluent with sloping grass strips. Open feedlot run-
off-treatment systems have been reported by Sutton et al. (1976) and Swan-
son et al. (1975). While the degree of treatment observed varied, all
these studies indicated that this method—vegetative filters—was effective
and potentially a satisfactory treatment method. No uniform design criteria
have evolved from these studies, however, and variable performance has made
environmental authorities hesitate to give blanket approval to this concept.
A study was begun in Illinois in 1975 to evaluate vegetative filter
systems and, if feasible, to develop design criteria for them. This work
was designed to determine whether the vegetative filter system would ade-
quately control the feedlot runoff so that a violation of water quality
standards would not occur in the case of storm runoff. The objectives of
the research reported here were as follows:
a. To determine whether vegetative filters are a feasible alternative
for management of feedlot runoff.
b. To identify the vegetative filter system configuration most likely
to be successful for the range of conditions encountered.
c. To develop design standards and management recommendations for
successful vegetative filter systems.
The research was conducted on four full-scale field systems installed
on beef, dairy, and swine lots. Two types of systems were studied, the
channelized flow system and the overland flow system. The channelized flow
system can have various configurations such as a graded terrace channel,
grassed waterway, or something else. In general, it is a system in which
the flow is concentrated in a relatively narrow channel. One of the chan-
nelized flow systems studied was a graded terrace system traversing the
hillside several times in a serpentine fashion. The other channelized flow
system had one section of graded terrace channel followed by a section of
grassed waterway.
The overland flow system refers to a situation in which the flow is
not concentrated but rather occurs as shallow sheet flow over a relatively
wide area.
Several grass varieties and various configuration of pretreatment and
distribution systems were tested. The study continued for over two years
and was conducted year-round.
-------
SECTION 2
CONCLUSIONS
Vegetative filters are an effective means of removing nutrients, sol-
ids, and oxygen-demanding materials from feedlot runoff before discharge.
The observed reductions of these constituents by the filter systems under
study were over 80 percent on a concentration basis and over 95 percent on
a mass-balance basis.
Bacteria levels in feedlot runoff are not greatly reduced by using the
vegetative filter. High levels of fecal coliform and fecal streptococcus
were found both in the effluent from the filter areas and in the effluent
from control areas, on which no runoff or manure had been applied.
Effluent discharged from vegetative filters during large runoff events
may not meet current standards for stream quality. However, the discharge
rates are usually relatively low, occur during periods of high stream flow
and, therefore, have high dilution rates, which results in negligible effects
on stream quality, especially as compared with the uncontrolled discharges
from open feedlots.
Overland flow is more effective than channelized flow for removal of
pollutants from runoff. This is reflected in the recommended design crite-
ria, which specify much greater contact times for channelized flow systems
to achieve equivalent treatment equivalent to that of overland flow systems.
Vegetation kill in the channel bottom may become a problem with channelized
flow systems.
Some increase in nutrient levels in soil and ground water under vegeta-
tive filter areas was observed. This increase was variable and generally
small, although additional study is needed to determine if this could be a
problem on a long term basis.
Vegetative filters in the study produced high yields of forage which
was used for livestock feed.
A comparison of orchardgrass, smooth bromegrass, and reed canarygrass
showed that all were satisfactory in overland flow systems, and only small
differences were observed in yields and effectiveness as a cover. Since
reed canarygrass is somewhat better in yield, tolerance, to both wet and
dry conditions and adaptability to high fertility, it appears to be the best
grass for most situations.
-------
The acceptance of the vegetative filter system by farmers is much bet-
ter than their acceptance of the runoff control systems having a holding
pond. Thus, the vegetative filter is likely to be adopted by smaller feed-
lots much more readily than the conventional systems, resulting in a reduc-
tion of pollution problems associated with feedlot runoff.
-------
SECTION 3
RECOMMENDATIONS
Vegetative filters should be designed so that the runoff from most
smaller precipitation events will infiltrate into the soil. To accomplish
this, soil type and filter area should be major design considerations.
Vegetative filters should be designed to adequately treat runoff from
large precipitation events so that it can be safely discharged. Flow con-
tact time (time of travel) should be the major design variable for this fac-
tor.
Since vegetative filter systems may not adapt well to every situation,
when planning a system, both the conventional holding-pond system and the
vegetative filter system should be considered, and the selection should be
based on site-specific factors.
When planning a vegetative filter system, especially where daily load-
ing (such as from a milking center) is anticipated, establishing a second
filter area should be considered. This would allow alternating use and give
a chance for periodic system recovery and drying.
Because of excessive sizes of vegetative filters required and a lack of
monitoring data for large systems, these systems cannot yet be recommended
for feedlots with a capacity of approximately 500 beef animal units or more.
Vegetative filters have been reported successful on larger feedlots in other
geographic areas and may later be recommended for large feedlots in this
area if proved satisfactory by additional research.
Before widespread installation of vegetative filter systems can be re-
commended, the concept of discharge from these systems will need to be re-
viewed and approved by pollution control agencies. Since this and other
recent studies have reported favorable results, approval is recommended.
Additional research is needed to verify the results reported for other
conditions and for long term operation under a wide variety of conditions.
This would also permit refinement of the design criteria. Specific areas of
study should include a comparison of trapezoidal (flat bottom) and parabolic
channels, especially with regard to reducing vegetation kill in the channel;
a study of the long-term effects on soil and ground water; and the develop-
ment of reliable distribution systems.
-------
SECTION 4
EXPERIMENTAL PROCEDURES
FIELD INSTALLATIONS
Four feedlots in which the vegetative filters adapt well to the physi-
cal situation and appear to have a reasonable chance of providing a success-
ful method for managing feedlot runoff were selected for study. One of
these systems was installed on the University of Illinois dairy farm, where
construction and management could be carefully controlled and observed and
where data could easily be collected. The other three systems were located
adjacent to commercial livestock production facilities. The locations of
the vegetative filter systems studied are shown in Figure 1. While some
advisory control was exercised in these situations, the management was pri-
marily up to the cooperating farmers. This arrangement provided the research
team with an opportunity to evaluate the manageability of these on-farm
systems as well as the efficiency of both the on-farm and the carefully con-
trolled system at the University of Illinois.
The basic system consists of a settling facility, a distribution com-
ponent, and the vegetative filter area, as shown in Figure 2. The runoff
from each storm event went directly to the filter area. Similar concrete
settling basins were used at all four locations, but each vegetative filter
was quite different (Figures 3-6).
At the University of Illinois dairy facility (System 1), effluent from
the settling basin was pumped by an automatic pump (controlled by the water
level) through a gated irrigation-pipe distribution system, spreading the
effluent on three field plots. One grass species was seeded on each plot.
Three different grass species were evaluated—reed canarygrass, bromegrass,
and orchardgrass. Each plot was surrounded by a berm to prevent any outside
drainage water from entering the plot area and to keep any applied effluent
and rainfall from escaping at any point other than the controlled plot out-
let. The control plot, planted to bromegrass, received no effluent applica-
tions. The 12 m wide by 91 m long (40 ft by 300 ft ) plots have a relative-
ly low slope, approximately 0.5 percent. The flow over the plots was intend-
ed to approximate sheet or overland flow. The ratio between the vegetative
filter area and the feedlot area is approximately 1:1.
System 2 in northwest Illinois, was also an overland flow type and was
installed to control the runoff from a beef feedlot with a capacity of ap-
proximately 450 cattle. Due to the total size of the operation, an NPDES
permit written to allow the use of a vegetative filter area was obtained.
This was strictly a gravity-flow system, with runoff passing through the
settling basin and distributed across the upper end of a sloping vegetated
-------
FESLER
SWINE
System 4
NORDMAN
BEEF
System 2
STROPE
BEEF
System 3
CHAMPAIGN-
URBANA
U Of I
Dairy
System I
Figure 1. Location of vegetative filter systems studied.
-------
oo
SETTLING
BASIN
GRASSED AREA
GRASSED TERRACE
CHANNELIZED FLOW
OVERLAND FLOW
X
Figure 2. Alternative configurations for vegetative filters used
as a treatment for feedlot runoff.
-------
Figure 3. System 1 - runoff collection and vegetative filter configuration.
-------
Figure 4. System - 2 runoff collection and vegetative filter configuration.
-------
Figure 5. System 3 - runoff collection and vegetative filter configuration.
11
-------
Figure 6. System 4 - runoff collection and vegetative filter configuration.
12
-------
area. At first, runoff was distributed through perforated plastic pipe 15.2
cm (6 in ) in diameter; later, rigid plastic pipe was split to form a weir.
The vegetative filter area was seeded predominantly to a fescue mixture.
Since the soil in the filter is sandy, a filter, lot area ratio of 0.7:1 was
used. The constructed flow length was 61 m (200 ft ).
System 3 was located in northern Illinois on a beef feedlot with a
capacity of 500 cattle. The runoff was directed through a concrete settling
basin and then to a vegetative area of the channelized flow (graded terrace)
type, patterned after the serpentine waterway system studied by Swanson et
al. (1975). The terrace channel was approximately 564 m (1,850 ft ) long
and had a parabolic cross-section with a top width of 8.5 m (28 ft ) and a
depth of 0.5 m (3 ft ). The channel slope was 0.25 percent.
System 4, in western Illinois, was on an uncovered swine-finishing fa-
cility with a capacity of 480 animals. The runoff from the feedlot passed
through a concrete settling basin and entered a vegetated terrace channel
seeded with garrison grass. The runoff traversed 152 m (500 ft ) of terrace
channel and 457 m (1,500 ft ) of grassed waterway before reaching a defined
watercourse. The terrace channel slope was 0.25 percent. The waterway
slope was approximately 2 percent.
EXPERIMENTAL PROCEDURES
A recording rain gage collected rainfall data at each site. For System
1, the quantity of runoff applied to plots was calculated from records of
elapsed pumping time and pump calibration curves. The amount of applied run-
off in System 3 was measured by means of an H-flume and a water-stage recor-
der located where the settling basin effluent enters the terrace channel.
The applied runoff quantities were estimated for Systems 2 and 4 by using
rainfall data and previously developed rainfall-runoff relationships for
feedlots in Illinois, as reported by Dickey and Vanderholm (1977).
Each study site was equipped with automatic sampling devices capable of
taking 24 discrete 550 ml (1 pint) samples. In addition, three composite
type automatic samplers were used in sampling the vegetative filter effluent
at System 1. Each of these samplers had a back-flushing cycle between each
sampling event. At each automatic sampler location, H-type flumes with
stage recorders were used to measure the flow rate of effluent at the sam-
pling site At each automatic sampler location, H-type flumes with stage
recorders were used to measure the flow rate of effluent at the sampling
site.
Flow-activated automatic water samplers were used to obtain filter out-
flow samples whenever discharge occurred. These units were usually set to
take a sample of 500 milliliter at intervals of 45 minutes as long as the
discharge continued. Whenever possible during runoff events, the automatic
samples were augmented by grab sampling at several points along the length
of the flow. Grab samples of the runoff entering the filters were also
taken periodically. Samples were later analyzed for chemical and microbio-
logical quality.
13
-------
Extensive soil sampling was conducted at all four sites during the
study to determine the effects of the runoff application on the location of
soil nutrients. On System 1, each plot was divided into four equal-length
segments, and six soil cores were composited to make one sample from each
segment on three occasions during the two-year treatment period: April 1976,
before the treatment began; November 1976, the end of the first treatment
period; and November 1977, the end of the experiment. Soils of System 1
were sampled at the following depths: 0-5 cm (0-2 in ), 5-10 cm (2-4 in ),
10-20 cm (4-8 in ), 20-30 cm (8-12 in ), 30-46 cm (12-18 in ), and 46-61 cm
(18-24 in ). Samples were analyzed for pH, total N, NC^-N, available P and
K, and conductivity. Groundwater samples were taken periodically and ana-
lyzed for chemical quality.
Soil samples from System 3 were taken in April and November 1976 and
November 1977 from four depths: 0-15 cm (0-6 in ), 15-30 cm (6-12 in ),
30-46 cm (12-18 in ), and 46-61 cm (18-24 in ). Sampling was conducted at
four positions over the length of the channel: 30 m (100 ft ), 120 m (400
ft ), 275 m (900 ft ), and 425 m (1,400 ft ) from the effluent discharge
point. At each sampling position, four cores were composited from the bot-
tom of the channel to form one sample, and three were composited from each
side of the channel about 1.5-1.8 m (5-6 ft ) from the centerline to form
another sample.
At all locations, forage from the filters was harvested as hay or hay-
lage and used for cattle feed. On System 1, a forage yield measurement was
taken on each segment before harvesting. To estimate yields, an area 0.9 x
6.1 m (3 x 20 ft ), representative of the segment, was harvested. Forage
green weights for each sampled area were obtained in the field and 500-1,000
gram (g) (18-35 oz ) samples were taken to determine moisture percentage and
nutrient content. These samples were oven dried at 66°C (centigrade)
(150°F) until constant weights were attained; these weights were used with
the sample green weights to determine field dry matter. Samples were then
ground and analyzed for total organic N, P, and K. Samples were extracted
with 1.0 N KC1 and analysis procedures were as given by Bremner (1965) and
Peck, T.R. (1978). (Personal Communication Concerning Liquid Fire Digestion
Technique Used in the Agronomy Soil Testing Lab, University of Illinois at
Urbana-Champaign.)
Water samples from System 1, which were obtained both by manual sampling
and automatic sampling, were taken directly to the laboratory for refrigera-
tion. Samples to be used for bacteria and BOD determination were obtained
manually in sterile bottles furnished by the Illinois Environmental Protec-
tion Agency (IEPA) and taken directly to the IEPA laboratory in Champaign,
which is only a short distance from the research site.
Samples obtained at outlying sites were collected by part-time local
assistants immediately after storm runoff events. Either immediately or
after a short period of refrigerated storage, these samples were shipped in
styrofoam cartons by United Parcel Service to the Agricultural Engineering
Department Laboratory, where they were analyzed for chsrr.ical quality only.
Normal time in shipment was one day.
14
-------
Water samples from all sites were analyzed for ammonia-N, total
Kjeldahl N, solids, chloride,COD, total P and K, and conductivity. Filter
influent and effluent samples from System 1 were analyzed for fecal coliform,
fecal streptococcus, and BOD. All chemical analyses of water samples were
according to procedures outlined in Methods for Chemical Analysis of Water
and Waste (USEPA, 1974), except that nitrogen analyses were determined by
the method described by Bremer and Keeney (1965). Analyses for bacteria
were according to the methods of the American Public Health Association
(1971).
15
-------
SECTION 5
RESULTS AND DISCUSSION
RAINFALL AND RUNOFF
Rainfall data, elapsed pumping times, and pump calibration curves were
used to calculate the rainfall-runoff relationship for System 1, as shown in
Figure 7. Supportive data obtained on System 3 were used for comparative
purposes but were not adequate for inclusion in this report. The rainfall-
runoff relationship found in this study compares quite favorably with pre-
vious results reported by Dickey and Vanderholm (1977), although those data
were from paved beef lots rather than a paved dairy lot. The calculated re-
gression line intercepts the abscissa at 3.11 millimeters (mm) (0.12 in ),
which indicates that runoff would be expected after approximately 3.11 mm
(0.12 in ) of rainfall.
The rainfall-runoff data were substituted into the equation used by the
Soil Conservation Service for estimating runoff volume (Schwab et al., 1966)
to obtain an appropriate runoff curve number for concrete-paved dairy feed-
lots in Illinois. That equation is
Q = (I-0.2S)2/ (I + 0.8S)
where Q = direct surface runoff, in
I = storm rainfall, in
N = arbitrary curve number varying from 0 to 100
S = (1000/N)-10
The average N value, calculated from the data from the 19 events where run-
off did occur, was 96.7 and the range was 95.0 to 99.9. This indicates that
the selection of a runoff curve number near 97 would be appropriate when
using the Soil Conservation Service method to estimate runoff volumes from
paved dairy lots in regions having climatic conditions similar to those in
central Illinois. Similarly, a runoff curve number of 90 (Dickey and Vander-
holm, 1977) would be appropriate for paved beef feedlots. The smaller curve
number, and thus lower runoff volumes, for beef feedlots is due to the fact
that dairy lots are normally cleaned more frequently than beef lots. With a
lower frequency of lot cleaning, the beef lots will accumulate a manure pack
that retains a larger amount of rainfall, thus resulting in lower runoff
volumes.
FILTER TREATMENT EFFICIENCY
Evaluation of the overland flow vegetative filter at System 1 indicates
good performance in treatment of runoff. During the monitoring period
(April 1976 to September 1977) there were 19 effluent discharges from the
filter area. The average ammonia-N concentration in the vegetative filter
effluent was 18.5 mg/1(Figure 8) and the average concentration of total
16
-------
70
60
ac
id
I- 50
Ul
40
o
oc
30
20
10
.5
RAINFALL- INCHES
1.0 1.5 2.0
2.5
RUNOFF (mm) = [ RAIN (mm) X .9031 ]- 2.805
RUNOFF (in) = [ RAIN (in ) X .9Q3l] - .1105
r = 0. 975
19 events
10 20 30 40 50 60
RAINFALL - MILLIMETERS
3.0
2.5
20 (O
LU
X
o
1.5 I
1.0 CC
0.5
70
Figure 7. Rainfall-runoff relationship for a paved dairy feedlot
(System 1).
17
-------
ui
o
o
o:
2001—
OVERLAND FLOW
SYSTEM I
TOTAL KJELDAHL-N • [ DIST. (m) * -1.423] + 145.5
A A = [PIST. (ft.) It - .4338J + 145.5
r = 0.947
AMMONIA -N * [ DIST. (m ) » - .65681 + 56.8
O r *
* [ DIST. (ff) * - .2002] +56.8
A r « 0.880
meters
feet
200
300
FLOW DISTANCE
Figure 8. Nitrogen concentration changes with overland flow (System 1.)
18
-------
solids was 996 mg/l(Figure 9). Concentrations of these and other constituents
are shown in Table 1. In general, the concentrations measured in the filter
effluent represented a reduction of about 80 percent in the constituent con-
centrations present in the settling-basin effluent as applied to the filter
area. However, the quantity of filter effluent was considerably less than
the quantity of basin effluent, primarily because of the amount of infiltra-
tion that occurred in the filter area. The filter effluent volume was
413 m3 (14,576 ft3), while the filter area received 2,453 m3 (86,666 ft3)
of feedlot runoff. On a mass-balance basis, the vegetative filter reduced
the amount of constituents applied in the basin effluent by about 96 percent
as shown in Table 2. Ammonia-N had the greatest reduction, showing a remov-
al of 97.7 percent; total solids had the least reduction, a removal of 95.5
percent.
TABLE 1. CONSTITUENT CONCENTRATIONS IN THE SETTLING-BASIN
FILTER EFFLUENT
(SYSTEM 1)
AND VEGETATIVE-
--
Concentrations (mg/1)
Constituent
NH3-N
Total kjeldahl nitrogen
Total solids
COD
P
K
Settling basin
effluent
134
300
3,697
4,224
64.1
665
Vegetative
filter effluent
18.5
59.6
996
616
14
168
Percent
reduction
86.2
80.1
73.1
85.4
78.2
74.7
Number of samples 33
TABLE 2. CONSTITUENT REMOVAL DURING THE STUDY PERIOD ON A MASS-BALANCE
BASIS BY VEGETATIVE FILTER TREATMENT OF FEEDLOT RUNOFF (SYSTEM 1)
Effluent quantity
Constituent
NH3-N
Total kjeldahl
Total solids
COD
P
K
Settling
basin
kg
329
nitrogen 736
9,069
10,361
157
1,631
Ib
725
1,622
19,993
22,843
347
3,596
Vegetative
filter
kg Ib
7.62 16.8
24.6 54.2
411 906
254 561
5.76 12.7
69.4 153
Percent
removal
97.7
96.7
95.5
97.5
96.3
95.7
Effluent volumes
19
-------
OVERLAND FLOW
SYSTEM I
3,000
TOTAL SOLIDS
A A
= [OIST. (m) * -22.70] + 2,483
[OIST. (ft) * - 6.920]+ 2,483
r » 0.959
Ul
Q
UJ
O
X
O
u
I **
i -
o E
O
z
V)
Q
_l
O
to
apoo
m»t«rs
feet
1.000 —
100 200
FLOW DISTANCE
300
Figure 9. COD and solid concentration changes with overland flow
(System 1).
20
-------
As determined by the limited number of BOD measurements, the BOD levels
in the basin effluent were reduced by 86 percent. The filter discharge had
an average BOD concentration of 165 mg/1.
Illinois stream standards (Illinois Pollution Control Board, 1973)
specify an upper limit of 1.5 mg/1 for ammonia and 1,000 mg/1 for total dis-
solved solids. The average ammonia-N concentration discharged from System 1
exceeded this level. Discharge rates from System 1 were low, however, aver-
aging 1.70 I/sec (0.45 gal/sec or 0.06 ft3/sec) with a maximum observed dis-
charge of 10.8 I/sec (2.84 gal/sec or 0.38 ft3/sec). Since this flow rate
is quite small relative to many receiving stream flow rates during storm
events, it is likely that adequate dilution would occur in most situations
so that stream standards would not be violated.
Samples obtained for bacterial analysis from the vegetative filter at
System 1 averaged 5.75 x 10^ fecal coliforms per 100 ml in the control-plot
effluent receiving no waste, 1.05 x 10 per 100 ml in the treated plot
effluent, and 1.25 x 10^ per 100 ml in the applied lot runoff. Fecal strep-
tococcus averaged 1.8 x 103 per 100 ml from the control plot, 1.1 x 105 per
100 ml in the treated-plot effluent, and 1.6 x 106 per 100 ml in the applied
lot runoff. From this we see that the bacteria levels are quite high in the
vegetative filter discharge but also are quite high in the discharge from
the control plot, on which no waste had been applied. This is consistent
with a previous study (Dornbush et al., 1974), which found high levels of
fecal coliforms and fecal streptococcus in runoff from agricultural land to
which no animal waste had been applied.
Figures 8 and 9 clearly show decreases in constituent concentrations as
the basin effluent traversed the vegetative filter at System 1. The data
points on Figures 8 and 9 are averages of grab samples obtained during seven
different runoff events. The figures also illustrate that a flow distance
of 104 m (340 ft ) would reduce the average constituent concentrations in
the effluent from the vegetative filter at System 1 to levels approaching
zero.
Performance of the vegetative filter at System 2, also an overland flow
type, was similar to that of System 1; the quality of the vegetative filter
effluent from System 2 would also not meet Illinois Stream Standards for
ammonia. Several factors contributed to the relatively high constituent
concentrations in the vegetative filter effluent from System 2. Concentra-
tion differences between Systems 1 and 2, shown in Tables 1 and 3, were pri-
marily due to differences in animal populations and frequency of lot clean-
ing. The animal population in System 1 averaged 100 dairy cows, while Sys-
tem 2 had approximately 500 beef cattle. The lot size in System 1 was about
1.4 times larger than that of System 2, so System 2 had an animal density
approximately 7 times that of System 1. In addition, since System 1 was a
dairy, it was cleaned frequently—daily, when possible. The beef feedlot in
System 2 was thoroughly cleaned every three or four months. Thus, there were
much higher constituent concentrations in the feedlot runoff entering the
settling basin at System 2 than in System 1.
21
-------
Another factor causing higher nutrient concentrations in the filter ef-
fluent of System 2 is related to the operation of the settling basin. Gen-
erally, the setting basin at System 2 was cleaned infrequently, which meant
both a loss of settling capacity overall and almost no effective settling
capacity during some storm events. Because of later expansion of the lot,
the constructed capacity of the settling basin was about 20 percent below
current design recommendations. All of these factors contributed to exces-
sive concentrations of constituents in the settling basin effluent for Sys-
tem 2. As a result, the upper end of the vegetative filter at System 2 be-
came a shallow but effective settling area, trapping large amounts of manure
solids.
Representative samples of the settling basin effluent at System 2 were
not obtained. Consequently, the effluent from the settling basin at System
2, after traversing the first few feet of filter, was assumed to be similar
to the effluent at System 3, also a beef feedlot. The relative percentage
of constituent concentration reductions obtained at System 2 are shown in
Table 3. The constituent concentrations in the vegetative filter effluent
of System 2 generally represent about a 70 percent reduction of the concen-
trations in the settling basin effluent.
Using the relationships between concentrations and distances developed
for System 1 (Figures 8 and 9) and the 61 m (200 ft) flow distance of System
2, the projected concentration reduction for constituents in the settling
basin effluent after traversing System 2 would be about 65 percent. This
projected concentration reduction after 61 m (200 ft ) of flow is close to
the 70 percent reductions shown in Table 3. The comparison between the con-
centration reductions at Systems 1 and 2 indicates that overland flow vege-
tative filters can achieve substantial and consistent concentration reduc-
tions under a wide range of management conditions.
Although a mass-balance of measurement of nutrient removal was not con-
ducted on System 2, there were a large number of rainfall events for which
there was no vegetative filter discharge, indicating that removal of constit-
ents in the settling basin effluent on a mass-balance basis would be greater
than the 70 percent reduction on a concentration basis.
TABLE 3. CONSTITUENT CONCENTRATIONS IN THE SETTLING-BASIN AND VEGETATIVE-
FILTER EFFLUENT (SYSTEM 2)
Concentrations (mg/1)
Constituent
NH3-N
Total kjeldahl nitrogen
Total solids
COD
Settling basin
effluent
608
1,122
12,777
14,288
Vegetative
filter effluent
173
324
4,710
2,691
Percent
reduction
71.5
71.1
63.1
81.1
Number of samples 3* 69
*Taken from System 3, also a beef feedlot.
22
-------
Because the vegetative filter at System 3 was of the channelized flow
type, sample location was flexible, and automatic sampler locations were
changed occasionally to evaluate the vegetative filter performance at various
flow distances. During the 1976 sampling period, the automatic sampler lo-
cation was 305 m (1,000 ft ) downslope from the settling basin discharge.
In 1977, two samplers were positioned at 229 m (750 ft ) and at 381 m (1,250
ft ) from the basin discharge until mid summer, after which time the sampler
at 229 m (750 ft ) was moved to 533 m (1,750 ft ) from the discharge.
The average constituent concentrations of the effluent at the four dif-
ferent sampler locations at System 3 are shown in Table 4. As illustrated
in Figures 10 and 11 there is a linear decrease in the constituent concentra-
tions as the settling basin effluent traverses the vegetative filter at Sys-
tem 3. The percentage of concentration reduction at the System 3 sampling
points are listed in Table 5. Comparing these percentage reductions with
TABLE 4. EFFLUENT CONSTITUENT CONCENTRATIONS IN THE VEGETATIVE FILTER AT
VARIOUS DISTANCES FROM THE SETTLING BASIN DISCHARGE (SYSTEM 3)
Distance from basin discharge
meters : 0
feet
Constituent
NH3-N
Total Kjeldahl
Total solids
COD
P
: 0
229
750
305
1,000
Concentration
608
nitrogen 1,222
12,777
14,288
-
362
566
7,699
7,413
115
226
439
5,248
5,661
105
381
1,000
(mg/1)
218
379
5,603
4,660
64.3
533
1,750
101
190
2,590
2,001
-
Number of samples 3 79 69 121 48
those in Systems 1 and 2 (Tables 1 and 3) shows that a vegetative filter
utilizing channelized flow must be considerably longer than an overland flow
system to achieve the same reduction. For example, the overland flow sys-
tems have about a 70 percent concentration reduction after 91 m (300 ft ) of
flow, while the channelized flow system requires about 427 m (1,400 ft ) of
flow distance to achieve a similar reduction.
Using the regression equations shown in Figures 10 and 11 and assuming
the filter discharge should meet the current stream quality standards for
Illinois (1.5 mg/1 NH3~N and 1,000 mg/1 solids), we can calculate that the
length of the vegetative filter at System 3 should be 648 m (1,975 ft ).
The assumption is this calculation is that the concentration reduction will
continue at the same rate regardless of length; however, this is not entirely
correct. Figures 12 and 13 illustrate the nitrogen decrease with flow dis-
tance for both Systems 1 and 3, respectively. A curvilinear regression
was used to fit the nitrogen data. With both systems, the fit obtained with
the curvilinear regression is better than the straight-line regression, as
indicated by the r values. In practice, this means that the pollutant con-
centrations approach a background level (and the stream standards) asympto-
tically, and excessive flow distances would be required to meet the stan-
dards. As noted previously however, the discharge rates are small and only
minimal dilution would be required to meet the standards.
23
-------
1,200
CHANNELIZED FLOW
SYSTEM 3
TOTAL KJELDAHL - N « [ DIST. (m) *-1.7461+1,045
* J
A • [DIST. (ft) * -.5322]* 1,045
r = 0.976
AMMONIA - N * [ DIST. (m ) * - .96461+ 582
——O
= [DIST. (ft)* - .2940J+582
r « 0.981
0>
E
UJ
e>
o
oc.
800
fett
FLOW DISTANCE
Figure 10. Nitrogen concentration changes with channelized flow
(System 3).
24
-------
o
<
2
UJ
o
I
X
o
12,000
epoo
o
(O
o
H-
4,000
m«t«rrt
fMt
CHANNELIZED FLOW
SYSTEM 3
TOTAL SOLIDS
A A
* [OIST. (m) *- 18.93] +12263
"A « [OIST. (ft) *- 5.769]+12,263
r » 0.979
COD • [DIST.(m) * - 23.08]+13,483
O O . [oiST. (ft) * - 7.02 9]+ 13483
r • 0.984
A
O
100
I
I
600 1200
FLOW DISTANCE
1,800
Figure 11. COD and solid concentration changes with channelized flow
(System 3).
25
-------
OVERLAND FLOW
SYSTEM I
TOTAL KJELDAHL-N « ( I 60 M .98253 )DIST (m)
200
A A
• (160 )( .99464
r * .983
AMMONIA -N « (63.4) (.97648 )°'ST(m)
8 (63.4){.99277 )
r « .971
DIST(ft)
100
200
FLOW DISTANCE
3OO
Figure 12. Nitrogen concentration changes with overland flow—curvi-
linear regression (System 1).
26
-------
CHANNELIZED FLOW
SYSTEM 3
1,200
TOTAL KJELDAHL-N = <||7I ) ( .99676)'
A A
DIST (m)
= (M7I ) ( .99901 )
r = .993
AMMONIA-N = (667 H.9967O )
DIST (ft)
DIST(m)
800
UJ
e>
o
a:
400 —
meters
feet 0
= (667 ) (.99899 )
r = .980
OIST(ft)
600 i^QO
FLOW DISTANCE
1*00
Figure 13. Nitrogen concentration changes with channelized flow—curvi-
linear regression (System 3).
27
-------
Mass-balance studies were conducted for three relatively large rainfall
events at System 3 during 1977. The three storms totaled 17.4 cm (6.84 in )
of rainfall. Using the average concentrations presented in Table 4 and the
flow volumes measured for each storm, mass balances were calculated for four
constituents. The total quantities of constituents applied to the vegeta-
tive filter at System 3 during the three storms are listed in Table 6. Ta-
ble 6 also lists the percentage of constituent removals at various flow dis-
tances from the settling basin. Generally, about 30 percent of the constit-
uents were removed in the first 229 m (750 ft ) of flow, with the next 152 m
(500 ft ) removing an additional 50 percent. The last 152 m (500 ft ) of
TABLE 5. PERCENT CONSTITUENT REDUCTION IN THE BASIN EFFLUENT AT VARIOUS
LOCATIONS IN THE VEGETATIVE FILTER (SYSTEM 3)
Distance from
meters :
feet:
229
750
305
1,000
basin discharge
381
1,250
533
1,750
Constituent
NH3-N
Total Kjeldahl nitrogen
Total solids
COD
P
40.5
49.6
39.2
49.2
-
62.9
60.9
59.0
60.4
16.0
64.2
66.3
56.2
67.4
48.6
83.4
83.1
79.7
86.0
—
TABLE 6. CONSTITUENT REMOVAL ON A MASS-BALANCE BASIS BY VEGETATIVE FILTER
TREATMENT OF FEEDLOT RUNOFF FOR THREE STORMS (SYSTEM 3)
Constituent
NH,-N
Total Kjeldahl
Total solids
COD
Effluent
quantity of
basin mete
feet
kg
200
nitrogen 370
4,212
4,710
9
10
Flow distances
rs:
Ib
441
815
,283
,380
229
750
24
35
23
34
.3
.8
.4
.0
381
1,250
Average
80
81
75
81
533
1,750
percent removal
.0 92.3
.2 92.2
.6 90.7
.8 93.5
vegetative filter removed approximately 12 percent of the constituents, so
that the resulting total constituent removal for System 3 was about 92 per-
cent on a mass balance basis.
The low removal rates at the upper end of the vegetative filter at Sys-
tem 3 reflect an inherent problem with a channelized flow system that has a
parabolic cross section. Even during large runoff events, the flow width in
the waterway seldom exceeds 1.5 m (5 ft ), primarily because of the control-
led outflow from the settling basin. Since the basin effluent contains a
large amount of nutrients and solids, the grass in the waterway bottom has
28
-------
been killed in a 0.3 - 0.9 m (1-3 ft ) width for about 90 m (295 ft ). Be-
yond the killed area, the vegetation has been stunted for approximately
another 150 m (492 ft ). Nutrients, solids, and water from most small run-
off events are deposited or infiltrated in the waterway segment where the
vegetation is killed or stunted. A waterway with a larger flow width(such as
a flat bottom) might help distribute the basin effluent more evenly and per-
haps alleviate the vegetation kill that is due to excessive nutrients and
water in the narrow channel bottom.
The mass-balance calculations indicate that the removal rates at System
3 are somewhat lower than the removal rates of System 1 (Tables 2 and 6).
Extending the vegetative filter to the previously calculated length of 648 m
(1,975 ft ) should result in constituent removals of near 97 percent. Even
though discharges from the vegetative filter at System 3 also do not meet
Illinois stream standards for ammonia and total dissolved solids, the mass-
balance calculations show that a high degree of treatment occurs. The mass
balances were calculated for only three storm events. However, during the
17-month study period (May 1976-October 1977) only about 10 storm events
resulted in discharges from the vegetative filter at System 3. Effluent
from the waterway traversed another 150 m (492 ft ) of nearly level cropland
before reaching a receiving stream. For the three storm events mentioned,
only 15.4 kg (34 Ib) of ammonia-N was discharged from the vegetative filter.
Assuming this number is representative of the other severi^raj.nfall events
(which were of about the same magnitude), the total annual ammonia-N dis-
charge onto adjacent cropland from System 3 would he about 51.3 kg (113 Ib).
The channelized flow vegetative filter at System 4 performed somewhat
better than at System 3. As indicated in Table 7, the average constituent
concentration reduction after 148 m (450 ft ) of flow distance was about 86
percent. System 3 however, had only a 45 percent reduction after 229 m
(750 ft ) of flow. The lowest concentration reduction observed with System
4 was for total solids, which was reduced only 78.7 percent. As with the
other systems, the quality of discharge from the vegetative filter does not
meet existing standards for discharge into receiving streams. However, at
System 4, the graded terrace discharges into an existing grass waterway.
Figures 14 and 15 show the constituent concentration decrease for System 4
as the effluent traverses the vegetative filter, including the additional
grass waterway. These decreases, although for only one runoff event, are
similar to the decreases observed for Systems 1 and 3 (Figures 8-13). For
TABLE 7. CONSTITUENT CONCENTRATIONS IN THE SETTLING BASIN AND VEGETATIVE
FILTER EFFLUENT AFTER A FLOW DISTANCE OF 148 M (450 FT.) (SYSTEM 4)
Concentrations(mg/1)
Settling basin Vegetative Percent
Constituent effluent filter effluent reduction(%)
NH3-N
Total Kjeldahl nitrogen
Total solids
COD
478
1,081
7,010
11,063
70.6
120
1,492
871
85.2
88.9
78.7
92.1
Number of samples
12
115
29
-------
CHANNELIZED FLOW
SYSTEM 4
STORM OF 3/12/77
160
120
O>
6
Id 80
O
O
tr
40
A A TOTAL KJELDAHL-N
AMMONIA -N
meters
100
200
300
400
feet 0
300
600 900 I.2OO 1,500
FLOW DISTANCE
Figure 14. Nitrogen concentration changes with channelized flow for an
individual storm (System 4).
30
-------
30001—
UJ
o
LU
(D
X
O
O
2000
CO ipQO
o
CO
meters •
feet 0
CHANNELIZED FLOW
SYSTEM 4
STORM OF 3/12/77
-A TOTAL SOLIDS
COD
300
600 900 1^00
FLOW DISTANCE
1,500
Figure 15. COD and solid concentration changes with chamelized flow for
an individual storm (System 4).
31
-------
the storm of March 12, 1977 on System 4, the approximate concentration re-
duction after 152 m (500 ft ) of flow was about 55 percent; after 320 m
(1,050 ft ) of flow, the constituent concentrations were reduced by nearly
90 percent.
The data for Systems 3 and 4 indicate that for equivalent treatment
longer flow lengths are needed when channelized flow is used than when over-
land flow is used.
FORAGE AND SOILS
Dry matter yields of System 1 are illustrated in Figure 16 for each
plot for the two-year period of the study. Production ranged from slightly
greater than 2 MT/ha per year (0.95T/ac/yr) in 1976 to greater than 16MT/ha
(7.48T/ac/yr) in 1977. Total annual production was greater in 1977 than in
1976 for the effluent-treated plots; the grass stand at the beginning of the
test period was less than optimum, and only two cuttings were taken during
the first year to allow dense stand development before winter. Total annual
yields for both years were greater in the treated plots than in the untreat-
ed (control) plot. This would be expected, considering the large quantity
of nutrients applied to the treatment plots in contrast to none in the check,
and the greater amount of water that the treated plots received (runoff from
the feedlot plus normal rainfall). Yields from treated plots were high re-
lative to results of variety tests (Graffis et al., 1978). Analyses of var-
iance were conducted on dry matter production data for each year. The four
treatments-orchardgrass, smooth bromegrass, reed canarygrass, and the con-
trol, smooth bromegrass - were compared, as were the four plot-segment pos-
sitions relative to the distribution pipe. The position of plot segment had
no significant effect on dry matter production in either year.
Single-degree-of-freedom comparisons in the 1976 analysis indicated that
the treated smooth bromegrass plot had significantly higher yields than the
control, that orchardgrass was similar to smooth bromegrass, and that reed
canarygrass did not differ from orchardgrass and smooth bromegrass in dry
matter production (Figure 16). The 1977 analysis of variance indicated
that, again, the treated smooth bromegrass plot was superior to the check
and that orchardgrass was not different from smooth bromegrass; however,
reed canarygrass had significantly greater production than orchardgrass and
smooth bromegrass (Figure 16).
Reed canarygrass, although not as widely used in Illinois as orchard-
grass and smooth bromegrass, has attributes that make it more suited to the
environment in this experiment; its natural habitat is poorly drained, wet
areas; nevertheless, it is also one of the more drought-tolerant grasses and
can utilize high fertility. Although somewhat less palatable than the other
species considered here, it appears to be theoretically and practically the
best grass to use in a situation where there is occasional waterlogging with
high-nutrient solution.
32
-------
ORCHARDGRASS
SMOOTH BROMEGRASS
REED CANARYGRASS
CONTROL- SMOOTH
BROMEGRASS
ORCHARDGRASS
SMOOTH BROMEGRASS
REED CANARYGRASS
CONTROL- SMOOTH
BROMEGRASS
DRY MATTER YIELD -
o
10
mt/ha
15
T
1976
1977
tl
I
I
I
02*68
DRY MATTER YIELD- t/ac
Figure 16. Dry matter yields at System 1.
33
-------
Levels of nutrient uptake as determined by forage analysis are present-
ed in Table 8. Grasses in treated plots removed 4 to 6 as much as N, P, and
K as did the check plot over the two-year experimental period. Percentage
nitrogen in forage from the treated plots averaged 2.7, whereas the check
was about half that level at 1.4. Crude protein level for the check, 8.8
percent, is somewhat low; 16.9 percent in the treatment is considered a
reasonable amount. The variation in pounds of N removed between cuttings is
largely a factor of dry matter production differences. Phosphorus uptake
differences between cuttings (Table 8) were also dependent on yield varia-
tions. Average phosphorus compositions were 0.29 and 0.23 percent in the
treated and check plots, respectively, neither of which was abnormally low.
Variations in forage yield were reflected in potassium uptake at differ-
ent cuttings as with the other two nutrients discussed above. Check plot
levels averaged 2 percent. Concentrations in treated plots averaged 3.4
percent, ranging up to 4.6 percent. Although potassium levels of this
nature are not unusual, they do indicate what is termed luxury consumption,
the tendency of plants to absorb amounts of a nutrient far in excess of
actual nutritional requirements. Heath et al. (1973) mentioned that a po-
tassium level of two to three percent is generally adequate for the physio-
logical processes in forages; anything above this is of no value nutrition-
ally to animal consumers, and, except in very high yield situations, any
increment above this should be considered wasteful absorption by the plant.
TABLE 8. NUTRIENT UPTAKE AT SYSTEM 1 AT VARIOUS CUTTINGS
1976
Nutrient 1
Treatment
Check
Treatment
Check
N
P
K
N
P
K
N
P
K
N
P
K
44.8
4.0
48.4
26.5
3.8
32.5
40.0
3.6
43.2
23.6
3.4
28.9
2
kg/ha
97.4
11.8
124.5
17.7
2.4
21.9
Ib/ac
86.9
10.5
111.1
15.8
2.1
19.5
1977
1
165.1
17.5
230.7
3.5
2.8
27.0
147.8
15.7
205.8
3.1
2.5
24.1
2
91.8
10.1
117.5
18.0
1.8
16.1
81.9
9.0
104.9
16.1
1.6
14.4
3
78.2
9.0
87.7
8.7
1.1
8.9
69.8
8.0
78.2
7.8
1.0
7.9
Total
uptake
478.0
52.5
608.9
74.4
11.9
106.3
426.4
46.8
543.3
66.4
10.6
94.8
34
-------
In this project, however, the goal was to increase nutrient uptake as much
as possible and luxury consumption of any nutrient was advantageous.
Nutrient uptake in relation to distance from the effluent discharge
pipe is shown in Figures 17-19. The check plot consistently had consider-
ably less uptake of all nutrients, N, P, and K. A similar trend of uptake
is evident in all three graphs: an increase from the first to second plot
segment, a decrease at the third, and another increase to the fourth. In
the treatment plots orchardgrass displayed lower uptakes of N and P than did
smooth bromegrass and reed canarygrass and had a generally lower uptake of
K. An overall decrease in orchardgrass nutrient uptake was observed as dis-
tance from the distribution pipe increased. The smooth bromegrass treatment
plot had higher uptakes of all three nutrients in the third segment of the
plot with lower uptake nearer the distribution pipe and in the fourth seg-
ment. The reason for this is not obvious. Smooth bromegrass generally had
higher uptakes of the three nutrients than orchardgrass and lower uptakes
than reed canarygrass. In addition to generally taking up more N, P, and
K than the other species, reed canarygrass displayed a generally decreasing
trend of N and K uptake with increasing distance but an increasing trend of
P removal. It is clear that the application of runoff with high nutrient
concentrations on the treatment plots increased total nutrient removal of
the grasses. Generally, a slight decrease in nutrients removed was percep-
tible from the beginning to the end of the filter.
Soil analyses for System 1 are presented in Figures 20-22. These
figures, based on the average value of the three treatment plots jand the
actual control plot data, indicate the respective levels of the various soil
constituents at various depths in the soil profile. Data for each of the
three sampling dates are also presented.
Nitrate is a very mobile soil nutrient whose concentration at any given
time is closely related to biotic and abiotic components of the soil ecosys-
tem; for this reason, much variation is expected in soil nitrate levels.
Soil nitrate-N data (Figure 20) indicate that at the first sampling the
treatment plots had higher nitrate concentrations than the control plot,
possibly because of leveling, past history, etc. This is a minor problem
since the primary concern is the possible accumulation of nutrients in the
soil profile over time. There are several observations of importance, how-
ever, including differences between treatment and check. Treated plots had
greater nitrate concentrations at all levels than did the check in the fall
of both 1976 and 1977. This would be expected, considering the large
amounts of nitrogen applied in the runoff. Fall 1976 levels were generally
lower than the other two sampling periods in both treatment and control
plots. In the check, nitrate levels were fairly constant throughout the soil
profile, but in the treatment plots they continually decreased to the maxi-
mum depth sampled. This indicates that nitrates are being leached throughout
the soil profile. The concentration rapidly approached the check at the 60-
cm (24-in.) level, with leaching probably responsible for the reduced concen-
tration at lower levels.
35
-------
600
o>
I
400
UJ
^^^
^^r
H
Q.
UJ
O 200
Q?
FLOW DISTANCE
150
- ft
O ORCHARDGRASS
A BROMEGRASS
* REED CANARYGRASS
D CONTROL
I
I
30 60
FLOW DISTANCE - m
300
O
O
400
I
UJ
200
o
CL
90
Figure 17. Nitrogen uptake by the forage on System 1.
36
-------
a:
O
a.
en
o
FLOW DISTANCE - ft
0 150
60
40
20
O ORCHARDGRASS
A BROMEGRASS
REED CANARYGRASS
D CONTROL
300
60
40
20
0 30 60 90
FLOW DISTANCE - m
Figure 18. Phosphorus uptake by the forage on System 1,
u
o
£
I
LU
CO
tr
o
o.
8
37
-------
600
I
Ld
*
h- 400
Q_
CO
O
Q. 200
FLOW DISTANCE - ft
150
(637)
O ORCHARDGRASS
A BROMEGRASS
* REED CANARYGRASS
D CONTROL
30 60
FLOW DISTANCE - m
300
600
400
o
O
.a
I
UJ
V)
200
O
90
Figure 19. Potassium uptake by the forage on System 1.
38
-------
SOIL NITRATE NITROGEN - PPM
0 5 10
20
o 40
I
UJ
o
Jf
oc
Q.
g 20
40
60
I
15
TREATMENT
6
12
18
UJ
o
* SPRING, 1976
0 FALL, 1976
A FALL, 1977
CHECK
J
6
o
cr
0.
UJ
o
12
18
0 5 10
SOIL NITRATE NITROGEN - PPM
24
IS
Figure 20. Nitrate nitrogen in the soil profiLe at System 1.
39
-------
60
AVAILABLE PHOSPHORUS
Ib/ac
100
* SPRING 1976 —
0 FALL 1976
A FALL 1977
200
AVAILABLE PHOSPHORUS
kg /ha
Figure 21. Available phosphorus in the soil profile at System 1.
40
-------
20
O 40
I
O
Jf
«*
O
o:
z
h-
Q.
60
20
40
60 Is
AVAILABLE POTASSIUM
Ib/oc
1000
2000
0
TREATMENT
12
18
* SPRING, 1976
0 FALL, 1976
A FALL, 1977
CHECK
24
0
9 E
a
12
18
J-J24
0 1000 2000
AVAILABLE POTASSIUM
kg/ha
Figure 22. Available potassium in the soil profile at System 1.
41
-------
Figure 21 represents the phosphorus soil-analysis data. It is apparent
that phosphorus levels in both the check and treatment plots are quite high.
Given the close proximity of the dairy feedlot and the fact that phosphorus
is relatively unleachable and not luxuriently consumed, these concentration
levels can be justified. With these characteristics of phosphorus in mind,
it seems likely that the differences in concentration between the treatment
and check below 10 cm (4 in ) are due to initially nonuniform experimental
areas. This is evident when the two lines for spring 1976 are contrasted.
Available phosphorus levels in the first 60 cm (24 in) of the soil did not
increase in the treated plots; no evidence for phosphorus movement in the
soil profile is apparent.
Soil concentrations of available potassium are high in both treatment
and check plots(Figure 22). The trend in treatment plots is one of decreasing
available K with increasing soil depth, indicating that some K is moving
through the profile. Potassium is generally leached less than nitrate and
more than phosphorus.
A total nitrogen balance sheet is presented in Table 9. Total N in the
top 60 cm (24 in) of soil was highest in the third segment of the treatment
plots and check plots in the spring of 1976. Fall 1977 data followed this
same trend in the treatment plot but not in the control. N removal by the
plants was generally constant down the length of the plots, with the treat-
TABLE 9. NITROGEN BALANCE SHEET FOR THE VEGETATIVE FILTER AT SYSTEM 1
Flow
distance
m
11.4
34.3
57.2
80.0
ft
37.5
112.5
187.5
262.5
m
11.4
34.3
57.2
80.0
ft
37.5
112.5
187.5
262.5
Soil N,
spring 1976
12,478
14,934
18,717
18,313
11,131
1.3,322
16,695
16,336
8.175
10,957
17,705
17,465
7,293
9,774
15,794
15,580
N
applied in
basin effluent
kg/ha -
2,2000
2,200
2,200
2,200
Ib/a
1,963
1,963
1,963
1,963
kg/ha -
Ib/a
Soil N,
fall 1977
treatment
13,145
13,394
19,109
18,602
11,726
12,840
17,046
16,594
control
7,339
6,689
7,840
11,112
6,547
5,967
6,994
9,913
N
removed
by plants
485
441
529
458
433
393
472
409
95
126
90
115
85
112
80
103
N
unaccounted
for
1,048
2,300
1,280
1,453
935
2,052
1,142
1,296
741
4,142
9,775
6,237
661
3,695
8.720
5,564
42
-------
ment plots removing around four times as much as control. The last column
gives the amounts of N that are unaccounted for. These losses are probably
the result of a combination of factors, including leaching, denitrification,
and volatilization. More variation and higher values in this parameter are
noted in the control plot than in the treatment plots.
Soil nitrate-nitrogen data for System 3 are presented in Figures 23-26.
The data for the first sampling position, 30 m (100 ft ) from the effluent
discharge point, are shown in Figure 23, the second position, 120 m (400
ft ), in Figure 24; and so on to the last position, 425 m (1,400 ft ), near
the end of the channel. The upper graph in each figure represents the soil
samples taken from the center of the channel, and the lower graph, samples
from the sides; the three lines in each graph correspond to the three samp-
ling dates.
Nitrate levels in the channel bottom show decreasing trends with in-
creasing depths through the soil profile at two sampling positions on two
sampling dates, that is 30 m (100 ft ) and 275 m (900 ft ) during the fall
of 1976 and 1977. These exceptions indicate nitrate is being leached through
the profile. There is no such trend in the other graphs. A few general ob-
servations, however, can be made: fall 1976 levels of nitrate-nitrogen are
consistently higher than at the other sampling dates and are considerably
higher at the 425 m (1,400 ft ) position in both the channel side and bot-
tom. Nitrate is a notoriously unstable soil component; its concentration is
dependent on interactions of abiotic and biotic factors as mentioned above.
The variance in the fall of 1976 data from that at the other two is probably
a product of these interacting facets of the soil ecosystem.
A trend in the fall 1976 channel side data, however, should be pointed
out: nitrate-nitrogen levels tend to be higher in the mid-part of the soil
profile samples than in the extremes. This trend could perhaps be attributed
to lateral movement of nitrate nitrogen from the major part of the channel.
With the exception of the channel bottom at 30 and 275 m (100 and 900 ft )
mentioned above, spring 1976 and fall 1977 data are very similar; their con-
centrations are low and vary little through the profile. Apparently with
nitrate varied considerably among sampling dates, an inspection of total N
in the soil profile would perhaps be more informative.
Total nitrogen at different soil levels, at four sampling positions,
and at the channel bottom and side is shown in Figures 27-30. With a few ex-
ceptions, total nitrogen concentrations decrease from the soil surface to 60
m (24 in ) in both the channel bottom and sides. The similarity between the
spring 1976 and the fall 1977 samplings in most cases appears to imply the
absence of N movement or accumulation during the course of the experiment.
Increased levels of total N are noted in the fall 1976 sampling at
the 275m and 425 m (900 - 1,400 ft ) distances (Figures 29 and 30);
these increases were not found in either of the other samplings. Erratic
fluctuations in fall 1976 data at 30 m (100 ft ) are also apparent and again
are not repeated. Total N levels are noticeably higher at the 30 and 425
m positions than at either the 120 or 275 m positions in both the channel
bottom and sides. Explanations for this might center around the site itself;
43
-------
SOIL NITRATE NITROGEN - ppm
10
20
30
40
E
o
UJ
O
Si
CO 60
O
tr.
U.
0.
UJ
o
20
4O
60
T?
O SPRING, 1976
A FALL, 1976
* FALL, 1977
CHANNEL BOTTOM
6
12
18
24
CHANNEL SIDE
I
I
I
I
LJ
O
S.
or
CO
o
a:
UJ
o
12
18
24
Figure 23.
0 10 20 30 40
SOIL NITRATE NITROGEN - ppm
Nitrate nitrogen in the soil profile at System 3 at
flow distance.
a 30 m
44
-------
E
o
I
LJ
O
tr.
O
QL
U.
ft
O
SOIL NITRATE NITROGEN - ppm
0 10 20 30 40
20
40
so
40
60 fc=
I
f
ANNEL BOTTOM
12
18
I
LJ
O
CHANNEL SIDE
O SPRING, 1976
A FALL, 1976
*FALL, 1977
I
-------
SOIL NITRATE NITROGEN - ppm
CHANNEL BOTTOM
O SPRING. 1976
A FALL, 1976
FALL, 1977
Figure 25.
601=
0 10 20 30 40
SOIL NITRATE NITROGEN - ppm
Nitrate nitrogen in the soil profile at System 3 at a 275 m flow
distance.
46
-------
SOIL NITRATE NITROGEN - ppm
E
o
UJ
O
2
(T
30
20
40
60
T
90
120
0
\
I
>*
60 t.
0
o
o:
u.
£j 20
40
60
\
\
CHANNEL BOTTOM
12
18
UJ
tr.
24 CO
CHANNEL SIDE
O SPRING, 1976
A FALL, 1976
* FALL, 1977
I
I
I
Q.
UJ
O
12
24
Figure 26.
0 30 60 90 120
SOIL NITRATE NITROGEN - ppm
Nitrate nitrogen in the soil profile at System 3 at a 425 m
flow distance.
47
-------
20
40
I
Ld
o:
Z) 60
cn
2
§
X
Q_
LJ
Q
40
60
TOTAL NITROGEN - %
.1 .2 .3
I
I
I
CHANNEL BOTTOM
12
18
UJ
O
if
O SPRING, 1976
A FALL, 1976
I FALL, 1977
CHANNEL SIDE
I I
24
0
•
o
(£
U_
Q.
LJ
12
18
24
0 .1 .2 .3 .4
TOTAL NITROGEN - %
Figure 27. Total nitrogen in the soil profile at System 3 at a 30 m flow
distance.
48
-------
£
o
U
O
1
o
(T
L_
Q_
LJ
O
TOTAL NITROGEN - %
• 10 .20 .30 .40
20
40
60
0
20
40
60
I
I
I
CHANNEL BOTTOM
12
18
O SPRING, 1976
A FALL, 1976
•» FALL, 1977
CHANNEL SIDE
I I
UJ
O
24 (7)
0 2
O
U.
• £
Ou
< 2
12
18
.1 .2 .3
TOTAL NITROGEN - %
.4
Figure 28.
Total nitrogen in the soil profile at System 3 at a 120 m
flow distance.
49
-------
E
O
I
UJ
O
CO
2
O
or
a.
LJ
o
TOTAL NITROGEN - %
.10 .20 .30 .40
20
40
60
20
40
60
I
CHANNEL BOTTOM
12
18
24
I
LJ
O
2
CK
CO
O SPRING, 1976
A FALL, 1976
* FALL, 1977
CHANNEL SIDE
I
O
a:
CL
UJ
o
12
18
24
.1 .2 .3
TOTAL NITROGEN - %
.4
Figure 29. Total nitrogen in the soil profile at System 3 at a 275 m
flow distance.
50
-------
E
O
I
UJ
O
O
-------
that is, the fact that the system was constructed on a hillside. And soil
types, N composition, etc. are bound to vary from one spot to another. Soil
displacement resulting from the channel construction also could contribute
to total N variation. No major changes in total N appear to have occurred.
GROUNDWATER QUALITY
PVC pipes, 1.91 cm (0.75 in ) in diameter, were installed in auger holes
in the vegetative filter at System 1. The auger holes were approximately
2.3m (7.5 ft ) deep with 3 holes being located in each of the four plots at
System 1. Sand was packed around the bottom of the pipe, which had been
slotted, and the hole was backfilled with tight clay. Groundwater samples
obtained with a vacuum pump were analyzed by the methods previously outlined.
Depth to groundwater was not measured.
Constituent concentrations in groundwater under the vegetative filter
at System 1 are listed in Table 10. Generally, there are increased nitrogen
concentrations in the groundwater beneath both the plots having feedlot run-
off applications and the check plot, which did not receive any runoff. The
total salt content of the groundwater, as indicated by electrical conducti-
vity measurements, also generally increased on both the test and check plots
during the study period. However, the chloride concentration in groundwater
under the test plots increased during a period of unusually low rainfall
events and then decreased to previous levels as rainfall increased during
the fall of 1977. The chloride concentration in the check plots remained
nearly constant throughout the sampling period.
For the constituents measured, there was a trend of increased concen-
trations in the groundwater beneath the vegetative filter at System 1 as
more feedlot runoff was applied. This increasing trend was observed in both
the check plot and test plots, indicating a possible movement of groundwater
from the treatment plot to the check plot area. A check plot located farther
away from the area receiving feedlot runoff would possibly have not reflected
the concentration changes occuring in groundwater beneath the vegetative fil-
ter at System 1. Although the increased concentrations could indicate a
possible contribution from the applied runoff, it should be noted that in a
TABLE 10. GROUNDWATER QUALITY AT THE VEGETATIVE FILTER AT SYSTEM 1
Date
Constituent
NH3-N
(mg/1)
N03-N
(mg/1)
Cl
(mg/1)
Conductivity
(umhos /cm)
6/76
test
check
test
check
test
check
test
check
0.
0.
6.
2.
44.
45.
1,065
895
54
98
40
08
3
3
7/76
1.
1.
5.
0.
65.
51.
602
522
47
05
82
86
9
2
8/76
0.
0.
214
57.
746
665
32
13
-
-
5
7
3
116
60
664
665
5/77
.54
.38
-
-
.5
10/77
7
5
40
32
50
45
4,803
2,471
.59
.22
.3
.5
.3
.3
52
-------
previous study by Kendrick (1977) nitrogen levels in outlets from tiles
draining nearby cropland had NC^-N levels ranging from 12 to 40 mg/1. The
observed nitrogen levels in the groundwater below the vegetative filter were
of the same general magnitude as found in local field tile drainage from land
receiving commercial fertilizer.
DESIGN CRITERIA FOR VEGETATIVE FILTERS
Based on calculated flow velocities and verified by observation, approx-
imately two hours are required for the basin effluent to travel the 91 m
(300 ft) flow distance for the System 1 overland flow vegetative filter
during large runoff events. About five hours are required for the effluent
to traverse the channelized flow vegetative filter at System 3. With vege-
tative filters, the major nutrient removal mechanisms are thought to be set-
tling, filtration by the vegetation, and absorption on soil and plant mate-
rials. For these to be effective removal mechanisms, flow velocity is an
important variable affecting pollutant removal. Thus, a major design cri-
terion affecting the quantity of pollutants removed is the flow time, or
contact time, required for applied runoff to travel the length of the filter.
This contact time is a direct function of flow distance, flow velocity,
slope, and other factors.
Data from both the overland flow and channelized flow vegetative filters
suggest that pollutant removal efficiencies above approximately 95 percent
may not be practical to achieve with these systems because of the excessive
filter size requirements beyond that level. The two hour contact time
associated with System 1 is adequate to remove slightly over 95 percent of
the pollutants, while the five-hour contact time associated with System 3
should remove about 92 percent of the pollutants. Although mass removals
were not developed for System 4, the calculated 1.5 hour contact time for
the 148 m (450 ft ) flow distance was sufficient to remove about 86 percent
of the pollutants on a concentration basis.
Given these pollutant removals and the associated contact times, a con-
tact time of two hours is recommended as the minimum for any vegetative fil-
ter system. For the channelized flow vegetative filters, contact times must
be increased as the size of the feedlot increases. On the basis of the data
from Systems 3 and 4, the minimum two hour contact time would be appropriate
for System 4, but System 3 would need a contact time of approximately six
hours to have a comparable reduction in pollutants. The size of the feed-
lot at System 3 is 2,508 m2 (27,000 ft 2) , while the lot size at System 4
is approximately 836 m2 (9,000 ft 2). Thus, it appears that for each addi-
tional 465 m2 (5,000 ft 2) of lot area an additional hour of contact time
is required. Table 11 lists various lot sizes and the contact times re-
quired for vegetative filters utilizing channelized flow.
Manning's equation as described by Schwab et al. (1966), and the mini-
mum contact times were used to calculate minimum flow lengths for channel-
ized flow vegetative filters having various slopes; these flow lengths are
shown in Figure 31. As illustrated, the flow lengths for a vegetative fil-
ter utilizing channelized flow would be very large on lot sizes larger than
0.4 ha (1 ac). It should be noted that the contact times shown are for a
53
-------
6000
bJ 4000
CO
O
o
H- 2000
'1500
1000
500
oi- o
s «
*• E
1.0 2.0
SLOPE - PERCENT
Figure 31. Approximate required channelized flow distance for various
slopes and contact times.
54
-------
specific design flow, which was relatively high; at lower flows, velocity
would be lower and contact time higher.
The values shown in Figure 31 were calculated using a design flow
depth of 15.2 cm (6 in ) and a parabolic channel shape. The somewhat ar-
bitrary selection of this flow depth was based on the assumption that a
15.2 cm (6 in ) flow depth is about the maximum at which any filtration by
channel vegetation would be effective. In the systems studied, peak flow
from a one year recurrence interval, two hour duration design storm would
normally exceed this flow depth, but temporary storage in the settling basin
and restricted basin outlet flow resulted in no channel flow depths of over
15.2 cm (6 in ) during the study period. For larger feedlots with higher
peak flows, exceeding the design channel flow depth can be avoided by pro-
viding temporary storage and controlled discharge with a settling basin or
by widening the channel sufficiently to handle larger peak flows without
excessive depths.
TABLE 11. MINIMUM CONTACT TIMES FOR VEGETATIVE FILTERS UTILIZING CHANNEL-
IZED FLOW FOR VARIOUS FEEDLOT SIZES
Lot
m2
929
1,394
1,858
2,323
size
ft ^
10,000
15,000
20,000
25,000
Minimum
hr
2
3
4
5
contact time
Because of uncertainities in predicting infiltration in a channelized
flow situation, infiltration has not been included as a design variable.
However, it was commonly observed on Systems 3 and 4 that runoff from smal-
ler storms was completely infiltrated. This benefited total system perfor-
mance in that the total quantity of nutrients discharged was zero in these
events. As contact times become larger with the larger lot sizes, infiltra-
tion and dilution influence system performance; since larger lots were not
observed in this study, however, these additional effects were not evaluated.
The design criteria presented may be adequate for large lots also, but with-
out further study the recommendations in this report must be limited to lots
in the size range observed.
Overland flow vegetative filters do not appear to require longer con-
tact times as the feedlot size increases, although total filter size is de-
pendent upon lot area, as the following design procedures indicate. Addi-
tional contact time is probably helpful, however. Therefore, the two-hour
contact time is the recommended criterion for determining minimum filter
length for overland flow filters. Again using Manning's equation (Schwab et
al., 1966) and the two-hour contact time we developed a set of minimum flow
lengths for overland flow vegetative filters with various slops. These
lengths are presented in Table 12. Because of low velocities, leveling, and
maintenance problems, slopes of less than 0.5 percent should be used with
55
-------
TABLE 12. MINIMUM FLOW LENGTHS FOB VEGETATIVE FILTERS UTILIZING OVERLAND
FLOW AND HAVING VARIOUS SLOPES3
Slope
%
0.5
0.75
1.0
1.5
2.0
3.0
4.0
Flow
m
91.4
113
131
160
185
227
262
length
ft
300
372
430
526
608
744
860
aDesign flow depth is 1.3 cm(0.5 in ).The assumed Manning's roughness co-
efficient is 0.3.
caution. Slopes of more than 4 percent should not be used because of high
velocities, reduced filter effectiveness, and possible erosion. The minimum
recommended length for any vegetative filter using overland flow is 91.4 m
(300 ft.).
Infiltration, settling, filtration, and adsorption are important in re-
moving pollutants in the overland flow vegetative filters. Thus, the second
phase in the design of overland flow vegetative filters is to develop the
total size criterion.
The recommended procedure for this is based on allowing runoff from most
small storms to completely infiltrate the soil in the vegetative filter area,
resulting in no discharge. Runoff from larger storms would be allowed to
discharge. The infiltration rate and soil type are the factors that deter-
mine how much runoff could be handled by infiltration during a given time,
so the recommended filter area is partly a function of soil type.
The required overland flow filter area is also a function of storm size.
If filters can be allowed to discharge several times annually, the size of
the infiltration area should be designed in terms of a storm size having a
short recurrence interval. From our initial experience, a one-year recur-
rence interval seems suitable. Since the filter length should provide for a
minimum contact time of two hours, selecting a two-hour storm duration is
also recommended. This allows the runoff to flow over the complete length
of the filter before rainfall ceases. Storm events larger than the one year-
two hour event or storms occurring when the vegetative filter is saturated
would result in a discharge. The two-hour contact time would provide
adequate treatment so that the filter discharge would not cause a signifi-
cant pollution hazard for the receiving stream.
For central Illinois the one year-two hour rainfall event is 40.6 mm
(1.6 in ). A typical medium-textured silt loam soil in central Illinois
(Drummer silt loam, maximum cover) has an infiltration rate of 38.1 mm/hr.
(1.5 in /hr ). Using the one year-two hour storm event and typical infiltra-
tion rates, the overland flow vegetative filter area required to handle both
56
-------
the direct rainfall on the filter and the feedlot runoff from System 1 would
be 0.44 ha (1.09 ac ). The approximate ratio of required filter area to
feedlot area for System 1 is 1:1. Thus, when sizing filters in areas with
rainfall and soil characteristics similar to those of System 1, the overland
flow vegetative filter should be about the same as the feedlot area. Table
13 lists the minimum overland flow filter area to lot area ratios for various
soil types when climatic conditions are similar to those in central Illinois.
With the two-hour contact time dictating the flow distance and with a
ratio of 1:1 for the filter area to feedlot area, the general vegetative fil-
ter configuration is thus specified. One other recommended criterion is a
minimum flow width. Observations and management practices indicate that a
vegetative filter utilizing overland flow should be at least 6.1 m (20 ft )
wide. Although there is no maximum width, the distribution of the basin
effluent across the top of the filter area could become a problem at widths
greater than 30.5 m (100 ft ) unless pressure distribution systems are used.
TABLE 13. RECOMMENDED OVERLAND FLOW FILTER AREAS WITH VARIOUS SOIL TYPES
(CLIMATIC CONDITIONS SIMILAR TO THOSE OF CENTRAL ILLINOIS)
Soil
Silty clay loam
Silt loam
Sandy loam
Infiltration
mm/hr
30.5
38.1
43.2
Rate
in /hr
1.2
1.5
1.7
Minimum filter area
1.6
1.0
0.7
x lot
x lot
x lot
area
area
area
The following example illustrates use of the proposed design criteria
for overland flow vegetative filters. Assume a central Illinois paved dairy
lot of approximately 0.2 ha (0.5 ac) with a capacity of 50 animals. The
adjacent field area has a slope of one percent. The soil is a silty clay
loam with an infiltration rate of 30.5 mm/hr (1.2 in /hr ). (information on
infiltration rates can usually be found in state irrigation guides and soils
handbooks for local areas.) The rainfall for the one year-two hour storm is
40.6 mm (1.6 in ).
Step 1. Find the required flow distance.
From Table 10, the required minimum distance should be 131 mm
(430 ft ).
Step 2. Find the required filter area
From Figure 2 lot runoff = (40.6 mm x .9031) - 2.805
= 33.86 mm (1.33 in )
Runoff volume = 0.2 ha x 33.86 mm = 6.77 ha-mm (0.65 ac-in)
The filter's infiltration capacity (1C) must equal or exceed the volume
to be infiltrated (VR) for proper filter operation. So:
Volume to be infiltrated (VR) = lot runoff volume + rainfall
on the filter area
57
-------
Filter infiltration capacity (1C) = infiltration rate x storm
duration x infiltration
area
(Recall that 1C = VR )
30.5 mm/hr x 2 hr x filter area = 6.77 ha.mm + (filter area x
40.6 mm)
61 mm x filter area - 40.6 mm x filter area = 6.77 ha-mm
20.4 mm x filter area = 6.77 ha-mm
filter area = 6.77 = 0.33 ha (0.8 ac)
20.4
Step 3. Specify filter area dimensions.
Filter length x width = area
Use minimum length of 131 m (300 ft )
131 m x width = 0.33 ha = 3,238 sq m
width = 24.7 m (81 ft )
Thus, the required minimum overland flow filter size for this example
is 24.7 m (81 ft ) wide by 131 m (430 ft ) long. If desired, the filter
width could be reduced and the length increased to obtain the same area, as
long as a minimum filter width of 6.1 m (20 ft ) is maintained. The total
filter size may be increased, too, if specific site conditions make a higher
degree of treatment advisable.
58
-------
SECTION 6
ECONOMICS
There was no attempt in this study to develop a comparison of investment
and operating costs between vegetative filter systems and conventional zero-
discharge systems. However, in a cooperating study, Lybecker (1977) utilized
cost data from the four systems included in this study and also two other
systems to make a cost comparison. For the vegetative filter systems,
actual cost data were available for all the systems. For the zero-discharge
systems, actual cost data were used, as well as cost estimates from the
Soil Conservation Service for zero-discharge systems at the vegetative filter
locations.
Table 14 contains a summary of the cost information developed by Lybeck-
er. For the dairy and beef systems, the vegetative filter system investment
costs ranged from 75 to 90 percent of the zero-discharge system costs. Since
management requirements are minimal for vegetative filters, the comparison
on that basis was even better. For the systems at hog facilities, a large
difference was shown, but there seems to be some question as to the accuracy
of the estimated cost for the zero-discharge systems.
It is clear from these figures that vegetative filter systems are less
expensive to construct and maintain. In many situations, farmers can con-
struct much of the vegetative filter systems themselves, which would result
in additional savings over those shown.
OPERATOR EVALUATION
The success of any pollution control system is highly dependent upon
the attitude of the operator. If the operator likes the system initially,
he is more willing to provide the necessary management and maintenance to
make it work properly. Without exception, the private operators involved
in the study preferred the vegetative filter concept to a system with holding
pond and pumping equipment. Their reasons for this included:
1. The land is not removed from production completely; instead it still
produces a useful forage.
2. A grassed area is preferable in appearances and offers less odor poten-
tial than holding pond.
3. Labor and equipment to empty a holding pond are not required.
The research sites are frequently visited by other livestock producers,
and these attitudes were found to be quite common. After the first year of
the study, the operator of System 4 installed an identical vegetative filter
for a new livestock facility he was building. Because of reactions of this
type, it is apparent that the vegetative filter concept is acceptable to
livestock producers. Widespread adoption of this method can be expected if
it is approved by state environmental authorities and if uniform design and
59
-------
TABLE 14. TOTAL INVESTMENT COSTS, OPERATING COSTS AND PERCENTAGE DIFFERENCE
BETWEEN THE VEGETATIVE FILTER AND ZERO DISCHARGE SYSTEMS FOR SIX
ILLINOIS DEMONSTRATION-RESEARCH SITES
Feedlot
Vegetative
filter ($)
costs
Zero discharge
($)
costs
Zero discharge costs
as a percent of vege-
tative filter costs
SIU-C~Dairy
85 head
Investment costs 8,302
Operating costs 1,103
U of I—Dairy
83 head
Investment costs 6,746
Operating costs 844
Strope—Cattle
700 head
Investment costs 8,960
Operating costs 958
Nordman—Cattle
425 head
Investment costs 7,920
Operating costs 874
Fesler—Hogs
450 head
Investment costs 2,617
Operating costs 299
Bradshaw—Hogs
800 head
Investment costs 5,986
Operating costs 648
9,190
1,386
8,103
1,111
9,823
1,656
10,725
1,333
13,055
1,916
11,453
1,882
111
125
120
132
110
173
134
153
500
640
191
290
Source: Lybecker (1977)
construction criteria are available.
RECOMMENDED MANAGEMENT
After operating and observing the systems in the study for over two
years, we developed a set of recommended management criteria. The following
criteria are relatively simple and should help maintain system performance
as well as prolong system life:
60
-------
Clean the accumulated solids from settling basin frequently so that settling
effectiveness is not impaired and outlet clogging problems will be reduced.
Frequency of lot cleaning will affect settling basin cleaning requirements.
If necessary, clean accumulated solids from the filter area near the inlet
area annually.
If possible, cut and remove forage from the filter area at least once a
year or more. By thus removing nutrients, this helps in reducing the rate
of nutrient accumulation in the filter area.
Avoid cutting and harvesting when the filter area is very wet, so that
equipment traffic will not form cuts that interfere with flow and develop
into wet spots.
Do not remove forage late in the fall. Instead, cut early enough that the
filter will go into winter with a good forage growth to aid in treating win-
ter and spring runoff.
If forage growth or soil analysis indicates excessive levels of salt or
other constituents after several years of operation, consider changing the
filter location to allow recovery of the filter to full productivity.
61
-------
REFERENCES
Bendixen, T.W., R.D. Hill, F.T. DuByne and G.G. Robeck. Cannery Waste
Treatment by Spray Irrigation - Runoff. Journal Water Pollution
Control Federation 41{3): 385-391, 1969.
Bremner, J.M. Total Nitrogen. In: Methods of Soil Analysis, Part 2 Chemi-
cal and Microbiological Properties. Agron. No. 9, ASA Madison,
Wise.,1965. pp. 1149-1178.
Bremner, J.M. and D.R. Keeney. Steam Distillation of Ammonium, Nitrate, and
Nitrite. Analytica Chemica Acta 32: 485-495
Dickey, E.G. and D.H. Vanderholm. Feedlot Runoff Holding Ponds-Nutrient
Levels and Related Management Aspects. Jour, of Env. Qual. 6, (3):
307-312, 1977.
Dornbush, J.N., J.R. Andersen, and L.L. Harms. Quantification of pollutants
in Agricultural Runoff. EPA-660/2-74-005. U.S. Enviornmental
Protection Agency, Washington D.C.,1974. 149 pp.
Edwards, W.M., F.W. Chichester and L.L. Harrold. Management of Barnlot Run-
off to Improve Downstream Water Quality, In: Proceedings of the
Int. Symposium on Livestock Wastes. ASAE, St. Joseph, Michigan,1971.
pp. 48-50.
Graffis, D.W., D.A. Miller, J.J. Faix, and E.R. Hawkins. April, 1971. For-
age Crops Variety trials, Illinois, 1977. Illinois Agric. Exp.
Stn. Publication AG-2015,1978.
Heath, M.E., D.S. Metcalfe, and R.F.Barnes. Forages: The Science of Grass-
land Agriculture, 3rd edition. Iowa State Univ. Press. Ames,
Iowa,1973.
Kendrick, R.F. Nitrate and Chloride Movement in the Soil Profile Due to Land
Application of Dairy Manure. unpublished M.S. Thesis, Library.
Univ of II. at Urbana-Champaign, 1977. 122 pp.
Kramer, J.A., D.E. Eisenhauer, R.I. Lipper and H.L. Manges. A Spray-Runoff
System with Recirculation of Treating Beef Cattle Feedlot Runoff.
ASAE Paper No. MC-74-301, ASAE, St. Joseph, Michigan, 1974.
Lybecker, D. Comparative Surface Runoff Control System Investment and Oper-
ting Costs for Six Illinois Demonstration-Research Sites. Paper
presented at Southern II. Univ. Liquid Livestock Waste Disposal
Field Day, Dept. of Ag. Ind. Southern II. Univ., Carbondale, 1977.
62
-------
Mather, R.J., Editor. An Evaluation of Cannery Waste Disposal by Overland
Flow Spray Irrigation. Publications in Climatology, C.W. Thorn-
thwaite Associates, Vol. XXII, No. 2.,1969.
McCaskey, T.A., G.H. Rollins, and J.A. Little. Water Quality of Runoff from
Grassland Applied with Liquid, Semi-liquid and Dry Dairy Waste. In:
Proceedings of the Int. Symposium on Livestock Wastes. ASAE, St.
Joseph, Michigan, 1971. pp.239-242.
Methods for Chemical Analysis of Water and Wastes. EPA-625/6-74-003. U.S.
Environmental Protection Agency, Washington D.C.,1974. 298 pp.
Schwab, G.O., R.K. Frevert, T.W. Edminster, and K.K. Barnes. Soil and Water
Conservation Engineering. John Wiley and Sons, Inc., New York,
2nd edition, 1976. 683 pp.
Sievers, D.M., G.B. Garner, and E.E. Picket. A Lagoon-Grass Terrace System
to Treat Swine Waste. In:Proceedings of the 3rd Int. Symposium on
Livestock Wastes. ASAE Publ. PROC-275. ASAE, St. Joseph, Michi-
gan, 1975. pp. 541-543.
American Public Health Association. Standard Methods for the Examination of
Water and Wastewater. New York, 13th edition, 1971. 769 pp.
Sutton, A.L., D.D. Jones, and M.C. Brumm. A Low Cost Settling Basin and In-
filtration Channel for Controlling Runoff From an Open Swine Feed-
lot. ASAE Paper No. 76-4516, ASAE, St. Joseph. Michigan, 1976.
Swanson, N.P., C.L. Linderman, and L.N. Mielke. Direct Land Disposal of
Feedlot Runoff. In:Proceedings of the 3rd Int. Symposium on Live-
stock Waste, ASAE Pub. PROC-275. ASAE, St. Joseph, Michigan, 1975.
pp. 255-257.
63
-------
BIBLIOGRAPHY
Butler, R.M., E.A. Myers, J.N. Walter, and J.V. Husted. Nutrient Reduction
in Wastewater by Grass Filtration. ASAE Paper No. 74-4024, ASAE,
St. Joseph, Michigan, 1974.
Eisenhauer, D.E. Treatment and Disposal of Cattle Feedlot Runoff Using a
Spray-Runoff Irrigation System. Unpublished M.S. thesis, Manhat-
tan, Kansas. Kansas State University Library, 1973
Eisenhauer, D.E., R.I. Lipper and H.L. Manger. Experience With a Spray-
Runoff System for Treating Beef Cattle Feedlot Runoff. ASAE Paper
No. MC-73-302, ASAE, St. Joseph, Michigan, 1973.
Feedlots Point Source Category - A development Document for Effluent Limita-
tions Guidelines and New Source Performance Standards. EPA-440/1-
74-004-a. U.S. Environmental Protection Agency, Washington D.C.»
1974 302 pp.
Koelliker, J.K. and J.R. Miner. Use of Soil to Treat Anaerobic Lagoon Ef-
fluent: Renovation as a Function of Depth and Application Rate.
Transactions of the ASAE. 13:469-499, 1970.
Kreis, R.D., M.R. Scalf, and J.F. McNabb. Characteristics of Rainfall Run-
off from a Beef Cattle Feedlot. EPA-RZ-72-061. U.S. Environmen-
tal Protection Agency, Washington D.C., 1972. 43 pp.
Law, J.P., R.E. Thomas, and L.H. Myers. Nutrient Removal From Cannery
Wastes by Spray Irrigation on Grassland. Water pollution Control
Research Series. 16080—11/69. Federal Water Pollution Control
Administration. U.S. Department of the Interior, Washington D.C.>
1969.
Lindley, J.A. Grass Filtration Beds for Disposal of Milking Center Waste.
Univ. of Connecticut Cooperative Extension Service Bulletin WWD-.
Storrs, Connecticut, 1973.
Nye, J.D., D.D. Jones and A. Sutton. Settling as a Method for Pretreating
Feedlot Runoff. InrProceedings of the 1974 Purdue Agricultural
Waste Conference, West Lafayette, Indiana, 1974.
64
-------
Overcash, M.R. and F.J. Humenik. Agriculturally related pretreatment Land
Applications Systems. In:Proceedings Second National Conference
on Complete Wateruse. American Institute of Chemical Engineering,
Chicago, Illinois, 1975.
Sopper, W.E. Crop Selection and Management Alternatives-Perennials. In
Proceedings of the Joint Conference on Recycling Municipal Sludges
and Effluent on Land. Champaign, Illinois, 1973. pp. 143-154.
Thomas, R.E. Feasibility of Overland Flow Treatment of Feedlot Runoff.
EAP-660/2-74-062. U.S. Environmental Protection Agency, Washing-
ton D.C., 1974. 28 pp.
Thomas, R.E. K. Jackson, and L. Penrod. Feasibility of Overland Flow for
Treatment of Domestic Wastewater. EPA-660/2-74-087. U.S.
Environmental Protection Agency, Washington D.C., 1974. 78 pp.
Wastewater Treatment and Reuse by Land Application. Volume II. Washington
D.C., 1973. 249 pp. U.S. Environmental Protection Agency.
Willrich, T.L., and J.O. Boda. Overland Flow Treatment of Swine Lagoon
Effluent. ASAE Paper No. 76-4515, ASAE, St. Joseph, Michigan,
1976.
65
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-143
3. RECIPIENT'S ACCESSION" NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
LIVESTOCK FEEDLOT RUNOFF CONTROL BY VEGETATIVE
FILTERS
August 1979 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHORisiDale Vanderholm, Elbert C. Dickey, Joseph A.
Jackobs, Roger W. Elmore, Sidney L. Spahr
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Departments of Agriculture, Engineering, Agronomy,
Dairy Science, University of Illinois, Urbana,
Illinois 61801
10. PROGRAM ELEMENT NO.
1BB770
11. CONTRACT/GRANT NO.
R804341-01-1
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final 2/9/76 - 5/8/78
14. SPONSORING AGENCY CODE
EPA-600-15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Vegetative filters were installed to treat runoff from two beef feedlots, one dairy
lot, and one swine feedlot in central Illinois. Two configurations were used-chan-
nelized flow and overland flow. Runoff underwent settling for partial solids re-
moval and was then applied directly to vegetative filter area. Runoff from most
smaller rainfall events infiltrated completely, resulting in no discharge. Runoff
from larger events partially infiltrated and partially discharged. Discharge sample
analysis indicated a removal of over 95 percent of nutrients and oxygen demanding
materials on a mass balance basis and over 80 percent reduction on a concentration
basis when compared to runoff applied to the filter area. Discharge rates were very
low and minimal dilution was necessary to meet state water quality standards. Design
criteria were developed for overland flow and channelized flow systems. The proposec
criteria would completely infiltrate runoff from small storm events and provide ad-
equate treatment for discharge during larger events.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
water quality
runoff
manur e
feedlot runoff
vegetative filter
animal wastes
68D
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
78
2O. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
66
*U.S. OOVtRHMENt PBIKIIHO OFWf 1979 -657-060/5374
------- |