EPA-600/2-76-233
October 1976
Environmental Protection Technology Serie
DESIGN CRITERIA FOR
SWINE WASTE TREATMENT SYSTEMS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental 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-600/2-76-233
October 1976
DESIGN CRITERIA FOR SWINE WASTE TREATMENT SYSTEMS
By
Frank J. Humenik
Michael R. Overcash
Biological and Agricultural Engineering Department
North Carolina State University
Raleigh, North Carolina 27607
Grant No. R-802203
Project Officer
Lynn R, Shuyler
Agricultural Wastes Section
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOBffiNT
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.
-------
ABSTRACT
Coordinated laboratory, field pilot-, and farm-scale lagoon studies were
conducted to define relationships between loading intensity and frequency
based on treatment performance, sludge accumulation, and odor potential.
Surface aeration of field pilot units and farm-scale lagoons was also
investigated to evaluate aeration levels required for odor control and
the effect of surface aeration on nitrogen and organic transformations.
Laboratory studies were designed to elucidate basic chemical, physical,
and biological mechanisms important in explaining and modeling lagoon
performance. Long-term mass balance studies were conducted to define
the fate of waste input and thus total constituent loss from the system.
Predictive and interpretive relationships for lagoons based on constant
batch loading and continuous loading were derived to describe the super-
natant concentration of unaerated lagoons. Methods for determining
steady-state concentrations and first-order reaction rate constants
for oxygen demand, organic carbon, and nitrogen were developed and com-
pared with laboratory and field pilot-scale data.
Lagoon liquid from a farm-scale unit was irrigated to nine 9.24 m x
9.24 m Coastal Plain soil-Bermuda grass plots at nitrogen loading
rates of 300, 600, and 1,200 kg N/ha./year. Mass balance data were
collected to determine the fate of applied waste constituents.
Analytical technique evaluations lead to recommendation of the 15-
minute digestion period for the standard chemical oxygen demand (COD)
test. The suitability of simplified portable laboratory methods for
nitrogen and phosphorus and selective electrode procedures for nitrate
and ammonia nitrogen were evaluated by comparison with results according
to standard wet chemistry procedures.
This report was submitted in fulfillment of Contract Number R-802203,
Program Element 1D2072, by North Carolina State University, Biological
and Agricultural Engineering, under the partial sponsorship of the
Environmental Protection Agency. Work was completed as of December 1,
1975.
-------
CONTENTS
Sections
I Conclusions 1
II Recommendations 10
III Introduction 13
General 13
Project Objectives 14
Scope and Purpose 15
IV Literature Review 18
Swine Production Trends 18
Properties of Swine Waste 18
Overview of Lajoons 20
Overview of Land Application 29
V Lagoon Studies 33
Sampling and Analytical Procedures 33
Definition of Reference Loading Rate 40
Swine Waste Characterization for 40
Experimental Studies
Laboratory Scale Experiments 4b
Field Pilot-Scale Experiments 105
Farm-Scale Lagoon 153
VI Predictive and Interpretive Relationships 163
for Lagoons
Batch Loading Approach 164
Continuous Loaded Approach - Organics 171
Continuous Loaded Approach - Nitrogen 184
VII Land Application Studies 190
Experimental Procedures 190
Field Results and Discussion 197
v
-------
CONTENTS (Continued)
Sections
VIII References 244
IX List of Resultant Publications 252
X Appendices 256
Laboratory Data 257
Lagoon Data 279
Land Application Data 287
VI
-------
LIST OF FIGURES
Number Page
1 Schematic of shear-type blender head located inside 35
blending container utilized for raw waste sample
preparation, stainless steel construction
2 Comparison of ammonia analysis by ion specific 38
electrode and standard distillation techniques for
swine and poultry waste
3 Schematic of sludge management experiment with one-liter 47
Imhoff cones loaded with swine waste
4 Supernatant COD concentration changes in Imhoff cones 49
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw swine
waste (first experimental set - once/week)
5 Supernatant TOG concentration changes in Imhoff cones 50
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw swine
waste (first experimental set - once/week)
6 Supernatant TKN concentration changes in Imhoff cones 51
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw
swine waste (first experimental set - once/week)
7 Supernatant o-PO^-P concentration changes in Imhoff cones 52
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw swine
waste (first experimental set - once/week)
8 Supernatant COD concentration changes in Imhoff cones 53
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw swine
waste (first experimental set - thrice/week)
9 Supernatant TOG concentration changes in Imhoff cones 54
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw swine
waste (first experimental set - thrice/week)
vii
-------
LIST OF FIGURES (Continued)
10 Supernatant TKN concentration changes in Imhoff cones 55
begun with swine waste lagoon supernatant with or
without sludge as inoculum and loaded with raw
swine waste (first experimental set - thrice/week)
11 Supernatant o-POA-P concentration changes in Imhoff 56
cones begun with swine waste lagoon supernatant with
or without sludge as inoculum and loaded with raw
swine waste (first experimental set- thrice/week)
12 Supernatant COD concentration changes in Imhoff cones 60
begun with tap water and loaded with raw swine waste
(second experimental set - reference rate)
13 Supernatant TOC concentration changes in Imhoff cones 61
begun with tap water and loaded with raw swine
waste (second experimental set - reference rate)
14 Supernatant TKN concentration changes in Imhoff cones 62
begun with tap water and loaded with raw swine waste
(second experimental set - reference rate)
15 Supernatant COD concentration changes in Imhoff cones 63
begun with tap water and loaded with raw swine
waste (second experimental set - four times reference)
16 Supernatant TOC concentration changes in Imhoff cones 64
begun with tap water and loaded with raw swine waste
(second experimental set - four times reference)
17 Supernatant TKN concentration changes in Imhpff cones 65
begun-with tap water and loaded with raw swine waste
(second experimental set - four times reference)
18 Supernatant TKN and NHo-N changes associated with 71
removal of the evaporative loss reduction cover from
the Imhoff cones
19 Microbial population and supernatant COD concentration 73
changes immediately after loading laboratory reactors
20 Schematic of 14-1 laboratory reactor loaded with swine 77
wastes
Vlll
-------
LIST OF FIGURES (Continued)
21 Supernatant COD concentration changes in 14-1 laboratory 78
reactors loaded with swine waste at various
frequencies
22 Supernatant TOG concentration changes in 14-1 laboratory 79
reactors loaded with swine waste at various
frequencies
23 Supernatant TKN concentration changes in 14-1 laboratory 80
reactors loaded with swine waste at various
frequencies
24 Supernatant o-PO^-P concentration changes in 14-1 81
laboratory reactors loaded with swine waste at
various frequencies
25 Supernatant COD concentration changes in laboratory 86
14-1 reactors with the nominal reference loading
rate of swine waste on a continuous or batch basis
26 Supernatant TOG concentration changes in laboratory 87
14~1 reactors with the nominal reference loading
rate of swine waste on a continuous or batch basis
27 Supernatant TKN and O-P04-P concentration changes in 88
laboratory 14-1 reactors with the nominal reference
loading rate of swine waste on a continuous or batch
basis
28 Supernatant COD and TOC concentration changes in 89
laboratory 14-1 reactors with four times the nominal
reference loading rate of swine waste on a continuous
or batch basis
29 Supernatant TKN concentration changes in laboratory 90
14-1 reactors with four times the nominal reference
loading rate of swine waste on a continuous or batch
basis
30 Supernatant o-P04~P concentration changes in laboratory 91
14-1 reactors with four times the nominal reference
loading rate of swine waste on a continuous or batch
basis
IX
-------
LIST OF FIGURES (Continued)
31 Supernatant COD and TOC concentration changes in 93
laboratory 14-1 reactors with 0.6 times the
reference loading rate of swine waste and once
per week frequency
32 Supernatant TKN concentration changes in labora- 94
tory 14-1 reactors with 0.6 times the reference
loading rate of swine waste and once per week
frequency
33 Supernatant o-PO^-P concentration changes in 95
laboratory 14-1 reactors with 0.6 times the
reference loading rate of swine waste and once
per week frequency
34 Supernatant COD and TOC concentration changes in 96
laboratory 14-1 reactors with 0.3 times the
reference loading rate of swine waste and once
per week frequency
35 Supernatant TKN concentration changes in laboratory 97
14-1 reactors with 0.3 times the reference loading
rate of swine waste and once per week frequency
36 Supernatant o-PO^-P concentration changes in labora- 98
tory 14-1 reactors with 0.3 times the reference
loading rate of swine waste and once per week
frequency
37 Supernatant TOC concentration changes in laboratory 100
14~1 reactors with 10.8 times the reference rate
of swine waste
38 Supernatant TKN and O-PO/-P concentration changes 1Q1
in laboratory 14-1 reactors with 10.8 times the
reference rate of swine waste
39 Schematic of pilot-scale lagoon research site 106
40 Photograph of pilot-scale lagoon research site 107
41 Supernatant COD and TOC concentration changes (two- 110
week averages) in field pilot-scale lagoons loaded
once per week at four times the reference rate for
swine wastes
-------
LIST OF FIGURES (Continued)
42 Supernatant TKN and o-PO^-P concentration (two-week 111
averages) in field pilot-scale lagoons loaded once
per week at four times the reference rate for swine
wastes
43 Supernatant COD and TOG concentration changes (two- 112
week averages) in field pilot-scale lagoons loaded
once per week at the reference rate for swine wastes
44 Supernatant TKN and o-PO,-P concentration changes (two- 113
week averages) in field pilot-scale lagoons loaded
once per week at the reference rate for swine wastes
45 Supernatant COD and TOC concentration changes (two-week 114
averages) in field pilot-scale lagoon receiving
effluent from lagoon loaded once per week at reference
rate for swine wastes
46 Supernatant TKN concentration changes (two-week 115
averages) in field pilot-scale lagoon receiving
effluent from lagoon loaded once per week at reference
rate for swine wastes
47 Supernatant COD and TOC concentration changes (two- 116
week averages) in third lagoon of three-unit series
with the first lagoon of this field pilot-scale series
loaded once per week at the reference rate for swine
waste
48 Supernatant TKN concentration changes (two-week averages) 117
in third lagoon of three-unit series with the first
lagoon of this field pilot-scale series loaded once
per week at the reference rate for swine waste
49 Supernatant TKN and o-PO^-P concentration changes (two- 118
week averages) in field pilot-scale lagoons loaded
once per week at 0.5 times the reference rate for
swine waste
50 Supernatant TKN and o-PO^-P concentration changes (two- 119
week averages) in field pilot-scale lagoons loaded once
per week at 0,5 times the reference rate for swine
waste
XI
-------
LIST OF FIGURES (Continued)
51 Supernatant COD and TOC concentration changes (two- 120
week averages) in field pilot-scale lagoons loaded
once per week at 0.25 times the reference rate for
swine wastes
52 Supernatant TKN and o-PO^P concentration changes 121
(two-week averages) in field pilot-scale lagoons
loaded once per week at 0.25 times the reference
rate for swine wastes
53 Supernatant COD and TOC concentration changes (two- 122
week averages) in field pilot-scale lagoons loaded
once per week at 0.125 times the reference rate for
swine wastes
54 Supernatant TKN and o-PO^-P concentration changes 123
(two-week averages) in field pilot-scale lagoons
loaded once per week at 0,125 times the reference
rate for swine wastes
55 Supernatant COD and TOC concentration changes (two- 124
week averages) in field pilot-scale lagoons loaded
once per week at 0,0625 times the reference rate
for swine wastes
56 Supernatant TKN and o-PO,-P concentration changes 125
(two-week averages) in field pilot-scale lagoons
loaded once per week at 0.0625 times the reference
rate for swine wastes
57 Supernatant COD and TOC concentration changes (two- 126
week averages) in field pilot-scale lagoons
loaded once per week at 0.031 times the reference
rate for swine wastes
58 Supernatant TKN and o-PO^-P concentration changes 127
(two-week averages) in field pilot-scale lagoons
loaded once per week at 0.031 times the reference
rate for swine wastes
59 Liquid TOC and TKN concentration as a function of 130
depth in an unaerated, anaerobic, pilot-scale field
lagoon loaded at the reference rate with swine
waste
xii
-------
LIST OF FIGURES (Continued)
60 Schematic of aerated pilot-scale lagoon with water 143
level control for fixed surface aerator
61 Supernatant COD concentration for field pilot-scale 144
lagoons with surface aeration loaded once per week
at 2 times the reference rate for swine wastes
62 Supernatant TOG concentration for field pilot-scale 145
lagoons with surface aeration loaded once per week
at 2 times the reference rate for swine wastes
63 Supernatant TKN concentration for field pilot-scale 146
lagoons with surface aeration loaded once per
week at 2 times the reference rate for swine wastes
64 Supernatant o-PO -P concentration for field pilot-scale 147
lagoons with surface aeration loaded once per week
at 2 times the reference rate for swine wastes
65 Depth profile of orthophosphorus concentration for 150
• : pilot-scale swine lagoons receiving different rates
of surface aeration
66 Supernatant TKN concentration and contributory 156
liveweight changes for an on-farm swine waste
lagoon loaded at once per day or more frequency
(Clayton)
67 Supernatant COD concentration changes for an on-farm 157
lagoon for swine wastes loaded at once per day or
more frequency as waste material runs into lagoon
68 Supernatant TOG concentration changes for an on-farm 158
lagoon for swine wastes loaded at once per day or
more frequency as waste material runs into lagoon
69 Schematic of actual lagoon supernatant concentration 165
with batch loading operation
70 Steady-state supernatant COD and TOG concentrations 170
for laboratory reactors (14 liters) loaded once per
week with swine waste (44,000 mg COD/1, 15,000 mg
TOC/1)
xiii
-------
LIST OF FIGURES (Continued)
71 Parametric plot of Equation (20) showing linear 177
dependence of steady-state concentration on loading
rate at high reaction rate constants
72 Steady-state supernatant COD concentrations for pilot- 182
scale swine lagoons receiving various loading rates -
experimental and predicted results (Equation 21)
73 Steady-state supernatant TOG concentration for pilot- 183
scale swine lagoons receiving various loading rates -
experimental and predicted results (Equation 21)
74 Steady-state supernatant TKN concentration for pilot- 188
scale swine lagoons receiving various loading rates -
experimental and predicted results (Equation 26)
75 Schematic diagram of facility developed to evaluate 191
lagoon pretreatment - land application system for
swine waste (Clayton)
76 Schematic of experimental plots for land application 193
of swine lagoon effluent
77 Schematic of soil and water sampling profile in 195
experimental plots for land application of swine
lagoon effluent
78 Soil profile total Kjeldahl nitrogen concentrations 224
for first-year application of swine lagoon effluent,
1973, at 1,200 and 300 kg N/ha.
79 Soil profile nitrate concentrations for first-year 225
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
80 Soil profile calcium concentrations for first-year 227
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
81 Soil profile magnesium concentrations for first-year 228
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
xiv
-------
LIST OF FIGURES (Continued)
82 Soil profile phosphorous concentrations for first-year 229
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
83 Soil profile potassium concentrations for first-year 230
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
84 Soil profile sodium concentrations for first-year 231
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
85 Soil profile chloride concentrations for first-year 232
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
86 Soil profile manganese concentrations for first-year 234
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
87 Soil profile iron concentrations for first-year 235
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
88 Soil profile copper concentrations for first-year 236
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
89 Soil profile zinc concentrations for first-year 237
application of swine lagoon effluent, 1973, at 1,200
and 300 kg N/ha.
xv
-------
LIST OF TABLES
Numbe r
1 Anaerobic Lagoon Loading Rates 24
2 Representative Summary of Various Parameter 34
Concentrations for Blended and Unblended Samples
3 Variability of COD Analyses for Raw Waste Samples 36
Using Shear Blender, Sample Weighing and Dilution
for This Experimental Study
4 Reference Loading Rate Utilized in Anaerobic Lagoon 41
Experiments Expressed in Units Commonly Reported
5 Mean Concentrations of Raw Swine Waste as Used For 43
Input to Laboratory and Field Pilot Scale
Experiments
6 Waste Generation from Under Slat Pit Receiving Swine 44
Waste
7 Chemical Parameter Ratios of Various Sources of Swine 45
Waste
8 Sludge or Settled Solids Recovery for Several Sludge 57
Management Techniques from Imhoff Cones Receiving
Swine Waste Inputs - First Experiment
9 Sludge or Settled Solids Recovery for Several Sludge 59
Management Techniques from Imhoff Cones Receiving
Swine Waste Inputs - Second Experiment
10 Initial Settling of Swine Wastes Evaluated from 67
Imhoff Cone Anaerobic Reactors (No Sludge)
11 Relative Supernatant and Sludge Removal Levels for 69
Imhoff Cones with Accumulated Sludge and Loaded
with Swine Waste
12 Samples Taken at Several Depths in Imhoff Cone and 74
14-1 Laboratory Reactors Loaded with Swine Wastes
xvi
-------
LIST OF TABLES (Continued)
13 Overall Mass Balance Results for Anaerobic Treatment 75
of Swine Wastes - Iinhoff Cone Reactors with Various
Sludge Managements
14 Amount of Waste Parameters Remaining in Sludge Zone 82
of Anaerobic 14-Liter Laboratory Reactors Loaded at
Different Frequencies
15 Comparison of the Amounts of Various Pollutional 83
Parameters in the Sludge Zone (Bottom 2.8 cm) and
the Sludge Plus Lower Supernatant Zone (Bottom 6.4
cm)
16 Sludge Accumulation for Swine Waste Input to Anaerobic 84
Laboratory Reactors
17 Overall Mass Balance for 14-1 Anaerobic Laboratory 85
Reactors Loaded with Raw Swine Waste at Various
Frequencies
18 Removal Efficiencies (Equation 2) of Various Pollutional 102
Parameters from 14-1 Laboratory Reactor Loaded with
Swine Wastes
19 Removal Efficiencies of Various Pollutional Parameters 103
from Imhoff Cone Laboratory Reactors Loaded with Swine
Waste
20 Uniformity of Raw Swine Waste During Agitation and 108
Loading of Field Pilot-Scale Lagoons
21 Swine Waste Loading Rate of Field Pilot-Scale Units 109
22 Steady Supernatant Concentration of Field Pilot-Scale 129
Lagoons Loaded at Various Rates with Swine Wastes
23 Liquid Concentration at Several Depths for a Field 131
Pilot-Scale Lagoon Loaded with Swine Wastes at Four
Times the Reference Rate
24 Dependence of Panel-Rated Odor Rank on Sampling 134
Location and Loading Rate for Field Pilot-Scale
Units with Swine Waste
25 Sludge Depth Determination for Anaerobic Swine Lagoons 135
Loaded at Different Rates
xvii
-------
LIST OF TABLES (Continued)
26 Removal Efficiencies (Equation 1) and Effluent 137
Concentration of Various Parameters from Field
Pilot-Scale Reactors Loaded with Swine Waste
27 Design Considerations for Naturally Aerobic Lagoons 138
and Pilot-Scale Lagoon Performance
28 Comparison of Biochemical and Chemical Oxygen Demand 139
of Various Field and Laboratory Anaerobic Reactors
Loaded with Swine Waste
29 Oxidation-Reduction Potential Measurements at Mid-Depth 141
in Field Pilot-Scale Anaerobic Swine Lagoons
30 Gas Composition from Field Pilot Units Receiving 152
Different Aeration Intensity
31 Oxygen Mass Balance for Fixed Aerator Operating in 154
Field Pilot-Scale Reactor Loaded Once Per Week at
Two Times the Reference Rate for Swine Wastes
32 Average Supernatant Concentration Values (January- 159
April 1974) for Farm-Scale and Field Pilot-Scale
Lagoons Receiving Swine Waste
33 On-Farm Lagoon Supernatant Concentration Distribution 160
in Vertical and Horizontal Directions (4/4/74)
34 Final Concentration and Batch Reaction Rate Constant 168
for Laboratory Anaerobic Swine Reactors (14 1)
35 First-Order Reaction Rate Constants for Chemical 179
Oxygen Demand (COD) in Anaerobic Swine Lagoon Units
36 First-Order Reaction Rate Constants for Total Organic 180
Carbon (TOG) in Anaerobic Swine Lagoon Units
37 Mass Transfer Coefficients for Total Kjeldahl Nitrogen 186
(TKN) as Lost from Anaerobic Swine Lagoon Units,
Equation 26
38 Concentration and Areal Loading Rate for Swine Lagoon 199
Effluent Reaching Plots Receiving High, Medium,
and Low Rates of Application
xvi 11
-------
LIST OF TABLES (Continued)
39 Comparison of Irrigated Swine Lagoon Effluent and 200
Liquid Reaching Plot Surface - Night Irrigation
40 Comparison of Land Application Rate Based on Lagoon 202
Concentration and Actual Material Reaching the
Plant-Soil System
41 Bulk Density Variation with Depth in Coastal Plains 203
Experimental Plots (Norfolk Sandy Loam)
42 Average Initial Soil Content in Upper 75 cm of 204
Various Waste Parameters for Plots Receiving-
High, Medium, and Low Application Rates
43 Rainfall Runoff Volumes from Experimental Plots 205
Evaluated in 1973
44 Effluent Irrigation Runoff Volumes from Experimental 206
Plots Evaluated in 1973
45 Concentration of Rainfall Runoff from Experimental 209
Plots Receiving Swiae Lagoon Effluent
46 Dry Matter Yields (kg/ha.) and Percentage Dry Matter 211
of Coastal Bermuda grass for First-Year Application
of Three Loading Rates of Swine Lagoon Effluent
(1973)
47 Mineral Concentrations of Coastal Bermuda grass 213
(Average of Three Replications) for First-Year
Application of Three Loading Rates of Swine
Lagoon Effluent (1973)
48 Mineral Concentrations of Coastal Bermuda grass 214
(Average of Three Replications) for First-Year
Application of Three Loading Rates of Swine Lagoon
Effluent (1973)
49 Crop Removal Rates for First Year of Swine Lagoon 215
Effluent Application to Coastal Bermuda grass (1973)
50 In Vitro Dry Matter Disappearance (IVDMD) and Total 217
Nitrogen Concentration of Coastal Bermuda grass Hays
for First-Year Application of Three Loading Rates of
Swine Lagoon Effluent, 1973, (Mean Values of 3
Replications)
xix
-------
LIST OF TABLES (Continued)
51 Animal Intake and Hay Composition for Coastal Bermuda grass 218
for First-Year Application of Swine Lagoon Effluent,
1973
52 Soil Accumulation in Upper 75 cm Beneath Plots Receiving 221
Swine Lagoon Effluent After One Season, September,
1973
53 Soil Accumulation in Upper 75 cm Beneath Plots Receiving 222
Swine Lagoon Effluent After One Season, December,
1973
54 Overall Mass Balance for First-Year Application of Swine 239
Lagoon Effluent at the High Rate to a Coastal Bermuda
grass - Norfolk Soil Plot System (Average of Three
Replicates)
55 Overall Mass Balance for First-Year Application of 240
Swine Lagoon Effluent at Medium Rate to a Coastal
Bermuda grass - Norfolk Soil Plot System (Average
of Three Replicates)
56 Overall Mass Balance for First-Year Application of 241
Swine Lagoon Effluent at the Low Rate to a Coastal
Bermuda grass -Norfolk Soil Plot System (Average of
Three Replicates)
57 Percentages of Various Waste Constituents Removed by 243
the Major Plant-Soil System Pathways for the Coastal
Bermuda-Coastal Plains Experiment Receiving One Season
of Swine Lagoon Effluent
xx
-------
Acknowledgements
The continuous and enthusiastic support of the North Carolina Agricul-
tural Experiment Station and Extension Service was fundamental for the
expansion of the research objectives and goals of this project. The
direction, encouragement, and perseverance of Mr. Lynn Shuyler, project
officer, and associated EPA staff have been extremely helpful in the
completion of this Environmental Protection Agency grant, number R802203.
Expressions of gratitude are in order for more Individuals than can
be reasonably expressed in an acknowledgement section because of the•
scope of this project in terms of interdisciplinary effort and magnitude
of supportive assistance. At the risk of unintentionally overlooking
most valuable and helpful inputs, the following contributions are noted
by work unit and some key individuals.
At the University Unit 2 Research Farm, Mr. H. V. Marshall, Superintendent
in Charge of University Research Farms, and Mr. R. Jc Williams, Unit 2
Farm Superintendent; Dr. 0. W. Robison, Dr. R. F. Behlow, and Dr. A. J.
Clawson, Animal Science; and at the Central Crops Research Station,
Mr. W. C. Allsbrook and Mr. W. R. Baker, Superintendents; Mr. J. R.
Woodard and Dr. J. R. Jones of Animal Science are thanked. Cooperators from
Soil Science included Dr. G. A. Cummings, Dr. B. L. Carlile, and Dr. J. W.
Gilliam. Dr. J. C. Burns, Crop Science, and Dr. Lemuel Goode, Animal
Science, were essential for the work on yield and nutritive value of
experimental plot cover crop. Dr. R. W. Skaggs was most helpful with
the soil-water considerations for the land irrigation work, and both
Mr. L. B. Driggers and Dr. R. G. Holmes assisted in the total planning
and operational analyses, as well as many other colleagues in Biological
and Agricultural Engineering, and in particular Mr. R. B. Greene,
Supervisor of the research shop, and his staff.
Special appreciation is expressed to Dr. G. A. Cummings, Soil Science,
and Dr. R. E. Sneed, Irrigation Specialist, for their hard work in
assisting with installation of field plots. Field plot operation and
maintenance, as well as the analyses and interpretation of the soils
data, were coordinated by Dr. G. A. Cummings. Dr. Burns coordinated
plot crop activities in conjunction with the feeding trials conducted
by Dr. Goode. The resources, continuous practical advice, and physical
labor of Mr. H. V. Marshall and Mr. R. J. Williams at the research
farm were superlative. Selective thanks also go to Dr. A. Stewart
for swine management activities.
xxi
-------
The continuous advice and stimulation of Dr. G. J. Kriz, Assistant
Director of the Experiment Station, Mr. D. H. Howells, Director of
the Water Resources Research Institute of the University of North
Carolina, and especially Dr. F. J. Hassler, Department Head of Biological
and Agricultural Engineering, in conjunction with the technical assistance
of Dr. P. Westerman and Dr. J. C. Barker, departmental faculty members
in agricultural waste management research and extension, served to
reinforce and augment project activity.
The majority of the hard work in the field and laboratory was supervised
by Mr. S. I. Britt, Mr. J. F. Koon, Ms. June Preston, and Mr. J. D.
Olson, and executed by a well respected but anonymously diligent staff.
Gratitude is expressed for the dedication and associated graduate
student activity of Mr. Howell, who was responsible for the lagoon
work, and Mr. T, Miller for surface aeration work.
Brenda Butts and Jan Jackson are thanked for their patience and excellence
in producing the final report.
xxii
-------
SECTION I
CONCLUSIONS
A. LABORATORY SCALE LAGOONS
1. Supernatant concentrations were related to loading rate. At
extremely heavy loading or about 0.21 m3/45-kg hog, supernatant
TOG, TKN, and orthophosphorus concentrations were approximately
equal to raw waste values. Data for operation at 2.3 m^/45-kg
hog or more resulted in 90-95 percent removals of input COD,
TOG, TKN, and o-P04~P.
2. Approximately 70 .percent of the input TKN, TOC, and COD for
laboratory batch or continuous loading was not recovered as
output or accumulation -due to organic stabilization and
nitrogen volatilization.
3. Results for one- and 14-liter laboratory units loaded at
frequencies between once/2 weeks and three times/week for units
with 2.3 nH and 0.6 m3/45-kg h°o showed that loading frequency
had little effect on:
a. supernatant quality.
b. sludge amounts and mass balances.
4. A slight reduction in supernatant quality resulted by using
continuous loading as compared to batch loading when considering
COD, TOC, and TKN at 2.3 nH/45-kg hog or 0.9 kg COD/m^/wk.
Little difference was evidenced for orthophosphorus. At the
heavier loading of 0.6 m-^/45-kg hog or 3.6 kg COD/m^/wk, there
did not appear to be any supernatant quality differences
between once per week and continuous loading for COD, TOC,
TKN, and O-P04-P.
5. Approximately the same amount of time (about 12 weeks) was
required between start-up and steady-state conditions for
batch type laboratory reactors seeded with lagoon sludge and
supernatant when compared with those begun with tap water.
6, Response and achievement of steady-state times for laboratory
and field lagoons were generally longest for TKN. COD concen-
tration response was less rapid than TOC and was generally
more variable as an indicator of supernatant quality, while
minimal response for orthophosphorus was noted.
1
-------
B. LABORATORY SCALE LAGOONS WITH DIFFERENT SLUDGE MANAGEMENT
1. Three sludge management reactors were studied: a) sludge
removal 6-8 hours after loading, b) average sludge detention
of 4-5 weeks, and c) accumulation of all settled material
with an average sludge residence time of 10-15 weeks. There
were no significant differences for supernatant concentrations
of COD, TOC, TKN, and o-PO^P among these reactors loaded at
2.3 m3/45-kg hog. Thus, it is concluded that the presence of
a sludge layer and subsequent interfacial transport was less
important in determining supernatant quality than biostabiliza-
tion and other loss mechanisms ongoing in the supernatant.
2. Postulated reasons for similar supernatant concentrations
regardless of sludge management were that the supernatant
had a high level of microbial activity which with once-per-
week loading was very probably underutilized. Thus, these
microorganisms effectively used sludge by-products, as well
as the raw waste input as substrate masking any differences
in interfacial transfer between reactors with no sludge,
controlled sludge, or accumulated sludge. Supernatant
bacterial data supported these conclusions.
3. Sludge mass balance differences between laboratory reactors
with no sludge or total solids removal and accumulated sludge
or no solids removal were dramatic. About 40-50 percent of
the COD, TOC, and TKN which settled initially in the no-sludge
cones was ultimately stabilized in the accumulated sludge
cones. This partially explains the slow rate of sludge
buildup in actual lagoons. As expected, phosphorus com-
pounds were conservative; and thus, the same percentage
amounts were present in all alternative sludge management
studies.
4. The percentage.of input constituents remaining in the sludge
for three laboratory experiment duration periods (19 to 56
weeks) was very similar indicating a steady-state decay level
for accumulated bottom sludge. The percentage of swine
waste input COD, TOC, and TKN which remained in the sludge
zone was approximately 30 percent. Sludge buildup rates in
laboratory reactors indicated that a lagoon would fill with
sludge in about 1,000-3,000 days. Field experience had indi-
cated that a slower buildup occurs; therefore, caution should
be exercised in transferring laboratory sludge accumulation
data to actual field conditions. Factors such as compaction
under greater liquid head, soil incorporation, effluent carryover,
-------
and long-term mixed culture biochemical stabilization could
reduce field lagoon sludge buildup rates below that measured
in shorter duration laboratory studies.
C. FIELD PILOT-SCALE UNAERATED LAGOONS
1. Lagoon supernatant concentrations for laboratory or field
lagoons having 2.3 nH/45-kg hog or more were found to be
uniform with depth throughout the total supernatant.
2. Steady-state concentrations, for the pilot-scale lagoons were
higher than laboratory units at comparable loading rates.
3. Supernatant COD, TOG, TKN, and orthophosphorus concentrations
were found to decrease in;value as loading rate Decreased or
as the lagoon volume increased for both laboratory and field
units. Supernatant TOG and TKN concentrations were directly
proportional to the loading rate at about, 2.3 m^/45^kg hog
or more. . ,
4. The percentage of input COD, TOG, TKN,and orthpphosphorus in
the effluent of unaerated lagoons at 203 m3/45~kg hog or more
was very low. Correspondingly, 90 percent and often more than
95 percent removals were obtained. At higher loading rates,
the phosphate and nitrogen removal became considerably lower.
The nitrogen removal at a loading rate of 0.6 m3/45-kg hog was
only about 36-65 percent of the input; and thus, this type of
lagoon would be more appropriate for nitrogen conservation^,
5. ...After 120 weeks of operation, the three pilot-scale field
units had a sludge buildup of approximately 12 percent of
the waste input volume, whereas sludge buildups in laboratory
units were about 25 to 30 percent of the waste input volume.
6. Sludge COD, TOG, O-PO/-P, and TKN concentrations were similar
for all laboratory and field experiments conducted., No
conclusive evidence of concentration profiles within the well
- defined sludge zone were found. Sludge was generally 35,000
to 60,000 mg/1 COD; 12,000 to 18,000 njg/1 TOC; and 2,000 to 3,000
mg/1 TKN. These sludge values are close to raw swine waste
concentrations. Sludge orthophosphorus concentrations of 2,000
to 3,000 mg/1 were 2 to ;4 times the. raw waste value indicating
settling and accumulation of phosphorus which is a conservative
element. ' .••'•• .. > ;
7, Detectable levels of dissolved oxygen near the lagoon surface
were not consistently found except for units at 37 to 74
m3/45-kg hog and the third lagoon in series in which the
-------
initial unit had 2.3 m^/45-kg ho^:. No dissolved oxygen was
found at depths greater than 10 cm below the surface even
for units at 37 and 74 m^/45-kg hog which transcended design
criteria for unaerated aerobic lagoons for swine waste.
Hence, design criteria for unaerated aerobic lagoons were not
supported by this study.
8. Based upon aperiodic field observation and odor panel rankings,
it was concluded that there was a discernible odor threshold
for swine waste lagoons loaded at approximately 9.2 to 18.4
tn3/45-kg hog. Odor was not manure-like nor was an odor always
detectable for lagoons with more than 18.4 nr/45-kg hog. For
lagoons with less than 9.2 m3/45-kg hog, odor was not always
detectable, but when found it was characteristic of swine
manure and hence was deemed offensive.
9. Individual consensus indicated that the frequency or pro-
bability of odor detection when visiting the site was 80 percent
for the unit at 0.6 m3/45-kg hog, 60 percent for 2.3 m3/45-kg
hog, 20 percent for 4.6 m3/45-kg hog, and little odor for units
with 9.2 m^/45-kg hog or greater.
D. FIELD PILOT-SCALE AERATED LAGOONS
Results from aerated lagoon experiments were impacted by bottom
scour resulting from the aerator-reactor size ratio employed.
Therefore, the magnitude of the resultant conclusions are
somewhat unique to the investigated reactor conditions.
1. Supernatant organic and nitrogen concentrations for aerated
units without bulk phase dissolved oxygen are lower than
similarly loaded unaerated units. No dissolved oxygen
was found in the aerated units, even in the surface layers.
Additionally, reduced odor potential existed for aerated
units at the same loading intensity as similar unaerated
reactors.
2. Supernatant COD and TOG concentrations increased with increased
aeration rates from 37 to 120 watts and associated greater
bottom scour, while TKN concentrations showed only a modest
increase indicating that the greatest impact of surface
aeration is on nitrogen reduction by ammonia volatilization.
3. Allowing a quiescent period after raw waste loading for
settling (24 hours) did not result in improved supernatant
quality for aeration at the 60-watt level.
-------
4. Aeration strategy employed was to accomplish odor and scum
. control through complete surface agitation by horizontal
pumpage which requires a minimum of 1 hp/93 m2 (1 hp/1,000 ft2)
surface area for utilized equipment. This operation also results in
substantial nitrogen reduction by ammonia volatilization.
5. Mass balance calculations indicated an oxygen transfer of 1.7-
1.8 kg 02/hr/kw or about 80 percent of manufacturer's rating
of 2.1 kg 02/hr/kw.
E. FARM SCALE LAGOON
1. The investigated farm-scale lagoon had lower effluent concen-
trations than comparatively loaded field pilot units. The ex-
planation for these lower steady-state conditions for the farm
scale lagoon may have included the daily loading frequency and
the cyclic nature of the total waste load. Also, the overall
loading increased slowly as the hog population built which would
allow development of good biological populations. Additionally,
because the housing unit was more open to the atmosphere than
totally enclosed houses, there was added opportunity for waste
degradation prior to lagoon input. Thus, growing unit configura-
tion and waste management techniques can have significant impact
on waste degradation prior to lagoon loading and thus influence
the performance of treatment units.
2. Sampling of producer scale field lagoons at various locations
verifies laboratory and pilot field conclusions that supernatant
quality is relatively homogeneous. This uniformity of supernatant
concentration for analyzed constituents in experimental units
ranging in size from 1 liter to over 850 m^ was an unexpected
result. Calculations indicated that diffusion represented a
minor contribution to supernatant uniformity. Daily temperature
cycles and thus thermal induced currents were not considered
as the principle explanation because supernatant uniformity also
occurred with laboratory reactors in a constant temperature
environment. Postulated mechanisms were a) high level of active
biomass and b) mixing effects of microorganisms and liberated gas.
3. Supernatant TKN concentrations for pilot-scale and full scale la-
goons evidenced an annual variation, about 250 mg/1 for the
spring-summer period and 350 mg/1 for the fall-winter period.
This difference in nitrogen concentration reflected the shift in
equilibrium between ammonium ion and ammonia and the decreased
volatilization associated with lower temperatures.
4. Sludge buildup for laboratory units loaded at 2.3 m^/45-kg
hog was from 15 to 35 percent of the lagoon volume per year.
-------
Field pilot-scale units had a sludge buildup of 10 to 15
percent of lagoon volume per year at the same loading. Sludge
buildup rates for producer-scale field units have not been as
great because of long-term compaction and degradation. How-
ever, it was concluded that sludge buildup for lagoons loaded
at about 2.3 m3/45-kg hog would require cleanout at a 10-year
interval or greater.
F. PREDICTIVE AND INTERPRETIVE RELATIONSHIPS FOR LAGOONS
1. The batch loading model developed to describe lagoon perfor-
mance could predict the effect of periodic loading. The major
disadvantage was that transient conditions which had an
impact for longer than one loading period such as a change
in loading concentration or volume could only be determined
by a number of successive weekly calculations. This limita-
tion prevented rapid prediction of steady-state conditions
associated with lagoon management options.
2. Predicted steady-state supernatant concentrations for storage
lagoons by the continuous modeling technique indicated that
when the supernatant COD concentration was above 30,000 mg/1
the lagoon was not functioning biologically,,
3. On an overall, basis, the use of an average reaction rate
constant based on a continuous loading model as determined
in the laboratory and temperature corrected or from a field
experiment for a single loading rate gave a good approximate
value for the supernatant COD and TOG concentrations. First-
order rate constants for COD and TOG were 0.55 week-1 for
la'boratory-scale (T=25° C) and 0.15 week"! for field-scale
units. The prediction of supernatant TOG concentration was
' better than COD.
4, Heaetiom rate constants for COD and TOG removal calculated
from! laboratory data and adjusted to field temperatures
predicted1 Lagoon sjpernatant values that were somewhat
higher than actual recorded data, indicating that the field
units were slightly more efficient than laboratory reactors
when put on the; same temperature basis.
5. The internal consistency of first-order'TKN mass transfer
coefficients derived from the continuous loading model for
various lagoon loading rates was better than first-order
reaction constants for COD and TOG. Rate constants for
COD and TOG varied by a factor of 3 while TKN removal constant
values varied by less than a factor of 2 over the same loading
-------
range, 2.3 to 74 m3/45-kg hog. This consistency may be
partly attributed to the physical mechanism for TKN removal
(ammonia volatilization) as opposed to the microbial dependent
stabilization of organics.
6. Developed models verified the slower supernatant TKN response
to changed lagoon operation and that supernatant nitrogen
levels were determined by surface area independent of volume.
The larger the surface area-to-volume ratio the faster the
approach to steady-state TKN conditions indicating a surface
volatilization mechanism instead of a bulk reaction removal.
This areal removal dependence for TKN was contrasted to the
microbial base volumetric dependence for COD and TOG.
G. LAND APPLICATION OF SWINE WASTEWATER
1. Concentration data for irrigation of lagoon supernatant
showed that for a lagoon loaded at 2,3 m^/45-kg hog approximately
25 percent of the nitrogen and 10 percent of the COD are lost
during night irrigation. This irrigation loss data cannot
be directly extrapolated to larger farm-type sprinklers, but
higher losses would be anticipated for daytime irrigation
because of higher temperatures and wind velocities.
2, Preliminary runoff data for the first year application of
300, 600, and 1,200 kg of nitrogen/ha./year indicated that
runoff volume was approximately 15 percent of the rainfall.
3. For COD application rates of 1,525, 3,050, and 6,100 kg/ha., the
percentage loss in runoff decreased with increasing loading
rate from 2.0 to 0.5 percent. Although there was a slightly
greater rainfall runoff volume for the high nitrogen appli-
cation plots, the liquid concentration of organics (COD) lost
could not be directly correlated to waste application rate
because neither the high nor low rate plots consistently
yielded the greatest amount of COD runoff. Runoff concentra-
tion similarities implied that the effect of different loading
rates was minimal and that the amount lost as a percentage of
that applied decreased with an increasing loading rate.
4. Concentrations for direct irrigation runoff were roughly
50-80 percent of the irrigated wastewater levels; but within
0.3 to 1.5 m of the receiver plot, complete infiltration of
irrigation runoff occurred.
-------
5. Although irrigation runoff volumes were small compared to the
total applied, waste constituent transport was significant when
compared to rainfall runoff especially for the high nitrogen
rate plots. The expected high concentration of pollutants in
this liquid emphasized the need to preclude this type of runoff.
However, continued long-term monitoring as sod develops more
fully may show that these runoff volumes become significantly
reduced.
6. Order of magnitude calculations indicated that less than 5
percent of all applied waste constituents appeared in rainfall
runoff.
7. Sheep acceptability evaluations of Coastal Bermuda hay from
the terminal swine waste irrigation plots indicated no evidence
of reduced hay palatability. The hay intake per unit body
weight was not significantly different for the effluent
land loading rates evaluated although crude protein content
of the hay increased with increased loading rate.
8. Waste plot Coastal Bermuda grass dry matter contents were lower
at the 600 and 1,200 kg of nitrogen/ha./yr rates than for the
300 kg of nitrogen/ha./yr demonstrating the greater water
uptake and top growth associated with excess nitrogen conditions,
Comparison of the dry matter yield among the three effluent
treatments demonstrated increased growth with increased
effluent loading rate. However, use of kg of nitrogen/
ha did not significantly increase dry matter yield over
the 600 kg of nitrogen/ha./yr rate indicating the plateau
region for response to nitrogen. The amount of N, P, and K
applied at the highest rate was more than double the amounts
found to produce maximum yields of Coastal Bermuda grass in
North Carolina. In general, the increase in grass dry matter
concentration of applied elements was significant between
various application rates.
9. Summarizing the soil accumulation results after 1 year of
application for the total 75 cm profile, four waste consti-
tuents, NC-3-N, K, Na, and Cl, were deduced to have increased
significantly between low and high rate waste loading above
the control or initial soil levels on an overall mass balance
basis. The other soil parameters, TKN, Ca, Mg, P, Cu, Zn,
Fe,and Mn, were not significantly affected by loading rate
although some were at higher levels than control cores or
had slightly elevated surface concentrations. Both Mg and
Ca evidence a surface accumulation, but levels returned to
control concentrations about 3 months after irrigation
termination. This second group of constituents were either
-------
not applied at high rates compared to initial levels, were
not reliably detected, or were leached from the upper 75 cm
of soil and hence were not recorded as constituents that
accumulated in the soil profile.
10. Crop removals as a percentage of applied material increased
as the effluent application decreased as was expected because
plant requirements were below applied levels. The actual
amount of various waste materials measured in the harvested
crop increased by nearly twofold from the low to the medium
rate and only slightly from the medium rate to the high appli-
cation. Material unaccounted for by crop uptake, accumulated
in the soil profile, or transported in runoff was about the
same for the low and medium rate plots except for nitrogen
which increased from 100 kg/ha, to 250 kg/ha. The high rate
plots had nearly a twofold increase in unaccounted materials
including nitrogen over the medium rate plots. Attributing
these losses conclusively to leaching after this first-year
study was not possible because the anticipated relative free-
dom of movement or mobility of these constituents was not
verified. That is, K, Na, and Cl should have been much more
mobile in the soil than Ca, Mg, or P. However^ leaching
losses as a percentage of the amount applied or initially
present were not significantly greater for K, Na, and Cl as
compared to P, Ca, and Mg. Thus, other factors yet unsub-
stantiated prevented indepth conclusions regarding mass balances
or pathways for removal of waste constituents. Several years
of data would be required to reduce the effect of annual varia-
bility of crop uptake and rainfall runoff as related to the
total system mass balance.
-------
SECTION II
RECOMMENDATIONS
A. ANALYTICAL TECHNIQUES
1. Results from the 15-minute and 2-hour digestion for the
chemical oxygen demand test as outlined in Standard Methods
show no statistical difference for all swine waste samples.
Thus, this shortened procedure represents the most cost-
effective manner to determine oxygen demand for waste char-
acterization, aeration requirements and unit performance
evaluations.
B. LAGOON TREATMENT
1. Animal waste is quite concentrated and contains all the
microflora needed for adequate anaerobic stabilization.
Therefore, seeding new lagoons to enhance treatment has
marginal benefit in tiost cases. These recommendations
do not, however, conflict with suggestions to provide
waste dilution during lagoon start-up.
2. Engineered systems to control or remove settled solids
to improve lagoon supernatant quality would not be warranted
based upon recorded insignificant differences in supernatant
quality due to sludge management techniques ranging from
no solids removal to elimination of all solids attendant
to no bottom sludge.
3. For maximum destructive pretreatment, it may be best to
allow all solids to enter the primary lagoon rather than
provide selective removal in view of the excellent bottom
sludge stabilization.
4. Greater supernatant COD, TOG, and TKN reduction occurred
for both laboratory and field units loaded on a continuous
or nearly continuous basis at 2.3 nr/45-kg hog or more.
This type of waste input could be accomplished by a con-
tinuous manure pit overflow, frequent flushing, or daily
scraping and cleaning.
10
-------
5. The best parameter to evaluate lagoon operation and
achievement of steady-state conditions is TKN because
response time is longer than for TOG and COD and the
importance of nitrogen content for terminal land
application.
6. A. sample taken at an intermediate depth is representative
of average lagoon supernatant conditions.
7. Although approximately 90 percent removals of input GOD, TOG,
TKN, and orthophosphorus were achieved at a lagoon loading
rate of 2.3 m3/45-kg hog, effluent levels for these para-
meters were 1,500, 500, 400,and 40 mg/1, respectively. This
resultant poor effluent quality emphasized justification for
regulatory criteria specifying no-discharge for anaerobic
swine lagoons.
8. Lagoon sizing criteria should be based upon the intended
function of pretreatment in the overall waste management
strategy. If odor or aesthetic nuisances are critical, then
lagoon sizing based upon odor threshold of about 9 to 18
m3/45-kg hog should be followed. If maximum nitrogen
conservation is desired, then heavier loading rates such as
2.3 m3/45-kg hog or less, even up to about 0.6 m3/45-kg hog,
could be used. Surface aeration may be employed in heavily
loaded units to counteract odor potential and allow smaller
sized :lagoons. However, surface aeration results in increased
nitrogen volatilization and thus is more appropriate as
degradative pretreatment for land limited situations. Lagoon
sizing will vary as a function of climatic conditions and
! moisture relationships. Regardless of management scheme,
which may emphasize nitrogen conservation and thus heavy
loading, or minimum nuisance conditions and correspondingly
lighter loading, lagoons are to be regarded as pretreatment-
storage devices and thus only one unit process in the overall
waste management system pursuant to terminal land application.
9. Lagoon supernatant TKN, COD, and TOG concentrations can be
estimated for lagoons sized at 2.3 m3/45-kg hog or more" by
developed modeling techniques to give order-of-magnitude
values for preliminary design of land application systems.
10. Design criteria for unaerated, aerobic, lagoons are often
counterproductive because bulk phase dissolved oxygen was
not recorded for lagoons at commonly specified sizes of
37 to 74 m3/45-kg hog and effluent quality was not suitable
for discharge.
C. LAND APPLICATION
I. Land application of anitfal waste to pasture or hay crops
11
-------
represents one of the most cost-effective terminal waste
management systems. AIL field data collected indicate no
hazard associated with applying waste at recommended fer-
tilizer rates based upon nitrogen, phosphorus, and potassium
for the moisture excess southeast. Actually, a margin of
safety exists for short periods because when twice the
recommended fertilization rate for Coastal Bermuda grass or
1200 kg of nitrogen/ha./year was added, little effect on crop
quality and runoff transport was noted over the initial one-
year period. However, soil nitrate, potassium, sodium, and
chloride did show increases between the low and high-rate
loadings and, in general, were above initial soil levels on
an overall mass balance basis. Other soil parameters
including copper and zinc were not significantly affected
by the high loading rate and both magnesium and calcium
which evidenced a surface accumulation returned to control
concentrations about 3 months after Irrigation termination.
No evidence of reduced hay palatability in sheep feeding
trials was recorded and intake per body weight was not
significantly different among the loading rates studied,
although the crude protein content increased with increased
loading. Therefore, all data would indicate that virtually
no environmental hazard results from the irrigation of animal
wastewater at fertilizer rates over long-term periods, and
that in fact, the environment can assimilate short-term,
high-rate loads. However, indications are that long-term,
high-rate applications would lead to soil accumulation,
increased runoff transport, and excessive soil-water nitrate.
(However, before final conclusions can be made, waste
receiver plots would have to be studied for several years
to reduce variabilities associated with start-up and annual
cycles as well as to obtain data on a mature plot that
would represent long-term,steady-state conditions.)
2. Runoff during or as a result of irrigation at excessive
rates or antecedent moisture conditions must be avoided.
A buffer zone can provide excellent attenuation by infiltra-
tion and overland flow stabilization and thus should always
be provided to further reduce rainfall runoff transport.
3. Night irrigation substantially reduced nitrogen lasses and
mist nuisances without causing any noted liabilities, and thus
should always be considered.
12
-------
SECTION III
INTRODUCTION
GENERAL
Recommendations directing terminal Land application of animal waste
have been verified by the no-discharge criteria specified in the final
Effluent Guidelines and Limitations for the Livestock Industry
published in the February 14, 1974, Federal Register. Land
requirements for the terminal application of animal waste are
contingent upon the degree of pretreatment provided and the fate of
waste materials applied to a particular soil-plant receiver system.
Limiting waste materials of primary importance at present are total
nitrogen, total salts, potassium, or feed additives depending upon
relative concentrations in the applied waste and geoclimatic
conditions.
The most common pretreatment-storage device for swine waste in the
southeast is a single unaerated lagoon. Therefore, the thrust of this
research project has been to define kinetics and mechanisms for waste
material degradation or removal for various lagoon loading conditions in
conjunction with efforts to define relationships between loading
intensity and frequency, sludge accumulation, odor potential-, and
treatment performance.
Design, operational, and economic information for aerated lagoons have
also become a high priority request as attention is being directed to
this alternative method for odor reduction from these pretreatment-
storage devices prior to terminal land application. Additionally,
surface aeration can augment nitrogen removal for cases where
sufficient land is not available to assimilate the nitrogen load
associated with application of raw waste or even excess lagoon liquid.
The major goal of these studies was to develop design and operational
criteria for pretreatment units and terminal land receiver plots that
are responsive to particular management strategies, such as maximum
component destruction or maximum nutrient conservation. Odor potential
13
-------
was a most important consideration for lagoons. Criteria for any land-
based system must also take into account health and nuisance require-
ments, cover crop management, geoclimatic conditions, equipment, and
producer preference in conjunction with environmental impact.
Ideally, recommendations would be developed to define loading and
operational criteria for maximum nitrogen or organic conservation,
with or without regard for odor control, that could be utilized for
producers interested in this management strategy. Correspondingly,
other producers could adopt criteria that would either exclusively or
by selective compromise achieve goals desired for a particular
management program.
Over-treatment of waste results in unnecessary and burdensome expense
and work input, whereas inadequate treatment represents a misuse of
facilities and a negligent effort. The importance of determining
relationships between loading and operation characteristics of
treatment facilities was readily apparent because many unaerated and
aerated lagoons were being constructed by rule-of-thumb estimates based
on many different and possibly irrelevant parameters. Correspondingly,
research and field experience pursuant to design and, operational
criteria responsive to a particular waste management strategy and
terminal disposal or utilization scheme represented a high priority
work area.
PROJECT OBJECTIVES
The specific objectives of this research project outlined in the
original proposal and pursued throughout the study period were:
1. To provide specific information on the performance of field
pilot-scale lagoons and farm-scale systems for swine
waste with the view of evaluating or refining the design
criteria in the southeast region for:
a. unaerated lagoon loading rates ranging from 1.15 cubic
meter (cu m) to 73.6 cu m per 45-kilogram (kg) hog;
b. two and three lagoon series systems consisting of
unaerated facultative or aerobic lagoons following
an initial anaerobic lagoon;
c. surface aeration strategies for minimum lagoon oxygenation
for odor control and increased nitrogen removal by
ammonia volatilization and nitrification-denitrification
in the aerated unit; and
14
-------
d. land irrigation of liquid from a full-scale unaerated
lagoon to Coastal Plains, Coastal Bermuda grass plots
characteristic of the southeast.
2. To evaluate and correlate with contemporary standard analytical
methods additional tests necessary for the development of
sound criteria for evaluation and design of animal waste
treatment facilities.
SCOPE AND PURPOSE
Field Pilot-Scale Lagoons
During the first portion of this study, a comparative experiment of
the following unaerated lagoon treatments was performed based on a
reference loading rate of 2.3 cu m of lagoon volume per 45-kg hog
which was the proposed .criterion for the region:
Unit 1 - four times reference loading
Unit 2 - reference loading
Unit 3 - 1/2 reference loading
Unit 4 - 1/4 reference loading
Unit 5 - 1/8 reference loading
Unit 6 - 1/16 reference loading
Unit 7 - 1/32 reference loading
Unit 8 - overflow from Unit 2
Unit 9 - overflow from Unit 8
This series of loading rates ranged from greater than any recommendations
to lower than criteria for unaerated aerobic units. In addition, Units
2, 8, and 9 represented a series lagoon .system. The recommendation
for naturally aerobic lagoons was bracketed between Units 6 and 7 so
the validity of recommendations for unaerated aerobic lagoons was .
tested. All lagoons were ,3.5 meters (m) in diameter and 1.8 m deep,
except for the second and third lagoons comprising the series system
for Unit 2, which were 1.2 m deep.
Surface aeration by variable speed, 185-watt, fixed aerators was
employed in two 3.5 m diameter and 1.8 m deep reactors. Relationships
between loading intensity and aeration input were evaluated to
determine minimum energy requirements for odor control. Operational
strategies to increase ammonia volatilization and nitrification-
denitrification in a single reactor operated without bulk phase
oxygen excess were investigated pursuant to increased nitrogen removal
prior to land application In limited areas.
15
-------
Laboratory-Scale Lagoons
Model field unit investigations were augmented by laboratory studies
with 14-liter (1) reactors and Imhoff cones to further elucidate
operational parameters and basic mechanisms with the aid of controlled
and definable conditions. Data derived from these laboratory reactors
were most helpful in analyzing and evaluating performance of both the
field full-scale and model reactors.
Farm-Scale Lagoons
The producer- scale lagoon used as a wastewater source for the land
application studies was monitored in addition to other producer
lagoons to document performance of typical farm-scale lagoons in the
southeast. The lagoon at the land application study site was built
in 1961 according to Midwest Plan Service recommendations of 0.12
cu m lagoon volume per kg hog live-weight. Waste was either washed
or scraped daily from the solid concrete floors into a collection
trough for lagoon input. After an identical housing unit and lagoon
were added in 1971, the two lagoons were operated in series with
all waste initially discharged to the original lagoon and overflow
going to the new lagoon. Wastewater for the land application studies
was irrigated from the initial lagoon. The swine research facility
has cyclic population densities; and thus, waste lagoon inputs
continuously vary with time at this site and at other farm-scale
lagoons typical of actual producer operations. Data from these
lagoons serve to evaluate predictive and interpretive relationships.
Land Application Studies
The land application investigations were conducted to determine the
effects of various nitrogen loading rates for swine waste lagoon
liquid on crop productivity and environmental quality. The three
loading rates studied were 300, 600, and 1,200 kg of nitrogen/ha./year
by the employment of three replicate plots for each application.
The middle rate approximates the nitrogen requirements for optimum
yield of Coastal Bermuda glass which was used as the assay crop.
Data were collected to determine the amount of nitrogen which leaves
the irrigated plots in rainfall runoff. The control area for soil
accumulation comparisons was adjacent to the Coastal Bermuda pasture
and received annual maintenance fertilization of 127 kg nitrogen/ha.,
55 kg P/ha., and 127 kg K/ha.
Analytical Studies
Analytical techniques were evaluated throughout the duration of this
study in an effort to either verify suitability of contemporary tests
16
-------
or develop more suitable procedures. Extensive experimentation to
determine the best analysis for oxygen demand evaluation was conducted.
Selective electrode and instrumental techniques were compared with
contemporary wet chemical procedures for nitrogen species and chloride.
Material balances on conservative elements were routinely made to check
techniques and augment laboratory quality control efforts. Procedures
required for accurate characterization of sludge components for field
and laboratory units pursuant to mass balance and mechanistic analyses
were developed.
Technology Transfer
Factual information on the performance of swine waste lagoons in this
area was conflicting when this project was proposed. Therefore, it
was difficult to evaluate design criteria or make competent recommen-
dations. Additionally, the waste load for a typical swine production
unit was not adequately documented. Thus, this project was undertaken
to obtain baseline data, determine lagoon performance, investigate basic
mechanisms, and develop design and operational criteria responsive to
desired unit and overall system management goals.
Results of this study have been used to develop technical papers and
extension publications. Extension publications on waste characterization
and management alternatives have served as basis for state-wide training
sessions. Additionally, regional workshops on the treatment of agri-
cultural waste pursuant to land application were held for state
regulatory personnel and consulting engineers in the southeastern
United States. Results and information derived from this research
have set a basis for continued development and transfer of technical
data at all user levels as opportunity and need direct.
17
-------
SECTION IV
LITERATURE REVIEW
SWINE PRODUCTION TRENDS
Swine production has changed drastically in the United States in recent
years. The unit where a small number of hogs were grown on open pasture
has largely been replaced by a production site where large numbers are
grown in confinement. While the number of producers has declined
somewhat, the size of operations and the amount of investment have
steadily increased. For example, in North Carolina in 1972 there
were 72 percent fewer producers with 12 sows or less than there were
in 1966, and during this same period, the number of producers with
100 sows or more increased by 450 percent (Jones-*-).
Production intensity of hogs at the national level usually follows
cycles over a relatively short number of years that are responsive to
market conditions. As of December 31, 1974, hogs and pigs on U.S.
farms were estimated at 55.1 million head, ten'percent less than a
year earlier and seven percent below December 1, 1972, for the lowest
December 1 since 1965 (NGDA ). The 14 states that make quarterly
estimates, which include the ten Corn Belt states plus the southeastern
states of North Carolina, Georgia, Kentucky and Texas, account for
about 85 percent of this total.
These swine production changes and increased public concern for
environmental quality magnified waste management problems. The
Secretary of Agriculture in a 1969 report to the President of the
United States emphasized that animal wastes exceed the waste from any
other segment of our agricultural-industrial-commercial-domestic
complex. As a result of a study on the role of animal wastes in
agricultural land runoff, Robbins et a1.^ concluded that the direct
discharge of untreated or partially pretreated swine waste was
totally unacceptable and that land application is very effective in
minimizing water pollution.
PROPERTIES OF SWINE WASTE
According to Muehling manure properties have been classified primarily
18
-------
as physical, chemical, and biological. These properties may be
affected by, at least, physiology of the animal (size, sex, breed,
activity), the feed ration (digestability and the protein fiber content),
and the environment (temperature and humidity). The quality of the
feed influences the amount hogs eat, conversion efficiency, and
ultimately the quantity and chemical composition of waste.
Copper is a required trace diet supplement in the range of 5-10 parts/
million (ppm), and feed additive levels of 125 to 250 ppm are used by
some producers. Taiganides5 estimated that 80 percent of swine feed
copper was excreted in the manure. Data of Arial6 revealed about
80 percent carry-over of ingested copper, 70 percent of the zinc,
and miniscule levels of antibiotic chlortetracycline due to rapid
degradation.
The total amount of .liquid manure to be handled is largely influenced
by the type of operation and particularly wastage from waterers and
foggers. Muehling^ reported that the most generally accepted estimate
of swine waste quantity has been five percent to eight percent of live
weight/day consisting of about 15 percent dry matter. Humenik et a1.'
in a literature review concluded that the most reliable weight and
volume of raw waste/45-kg hog/day was 3.8 kg and 3.8 1, respectively.
Manure fertilizer and soil conditioner constituents have recently been
investigated more extensively than any of the other chemical and
biological characteristics, but considerable value variation for all
constituents still exists in the literature. Values judged most
reliable in a recent annotated literature review (Humenik et al.')
were 0.32 kg chemical oxygen demand (COD), 0.13 kg five-day
biochemical oxygen demand (BOD5), 0.022 kg total nitrogen (TN),
0.063 kg total phosphorus, and 0.094 kg potassium/45-kg hog/day. The
total organic carbon (TOG) was evaluated at 0.09 kg/45-kg hog/day
(COD/TOG = 3.5). Based on literature references and calculations, it
was judged that the raw waste COD and BOD5 concentrations are about
80,000 milligrams/liter (mg/1) and 35,000 mg/1, respectively. The
concentration of total solids (TS) in raw waste was evaluated to be
about 80,000 mg/1 for a corresponding moisture content about 92
percent.
Total waste volume in underfloor pits with partial or total slats and
no water washing is approximately 3,8 l/day/45-kg hog, based upon data
collected at N. C. State University (Humenik e_t a_l.') and other
researchers (ScholzS, Van Arsdall9). Therefore, the parameter
concentrations for the waste in a storage pit with no overflow are
about 1/2 the level of the defecated raw waste based on a total
waste load of 3.8 1/45-kg hog/day.
19
-------
OVERVIEW OF LAGOONS
Unaerated Aerobic Lagoons
Aerobic bacteria utilize waste as substrate, breaking down part of
the organic portion into the more basic compounds of water and carbon
dioxide in the presence of free oxygen. Some nitrogen is released
to the atmosphere, but much of it is converted into nitrites and
nitrates. However, the aerobic process is not 100 percent efficient
in removing organic matter. According to Muehling^ there is usually
only 40 percent to 50 percent degradation of volatile solids (VS).
McKinney-'-0 also states that there will be a large mass of solids
in terms of generated biomass requiring disposal.
The major advantage of aerobic decomposition is that the entire
process is essentially odorless. According to Dale other benefits
of aerobic decomposition are (a) partial decomposition of volatile
(organic) solids into odorless gases, (b) destruction of most
pathogenic organisms, (c) reduction in the pollutional characteristics
of the wastes, and (d) concentration of minerals which may be more
readily applied to the land.
Unfortunately, the use of unaerated aerobic lagoons for animal waste
is generally not feasible because of the excessive surface area
requirements. Hart and Turner concluded that unaerated aerobic
lagoons cannot be practically used because the high concentration
of organics would require excessive amounts of dilution water to
develop naturally aerobic conditions. McKinney stated that the
use of oxidation ponds for swine waste required far too much land
area and that for such a large pond, mixing would be a problem
because a 0.40 hectare (ha.) oxidation pond 1.2 m deep would be
required for a 200-head hog operation. Most of the early livestock
lagoons were expected to function as an aerobic lagoon but usually
became anaerobic because loading was too heavy.
> was one Of the first researchers to make recommendations for
the design, operation, and management of lagoons for disposal of
livestock wastes. Design criteria were based on specifications for
aerobic municipal sewage lagoons, calling for a shallow pond (.9 to
1.5 m deep) and allowing 125 to 500 pigs/ha, of lagoon depending on
the climatic conditions. Eby -* also characterized a properly
functioning aerobic lagoon as one that is not offensive in appearance
or odor .
Clark ' concluded that lagoons for livestock waste must be aerobic
if they are to be satisfactory for the producer and community. He
recommended approximately one ha. of lagoon for each 680 hogs at
40° North latitude (about 14.7 square meters (sq m) of surface per
20
-------
hog). A loading variation of 15 percent/each 2^° variation in latitude
was suggested. According to Muehling? recent studies show that a
lagoon will not remain aerobic at this loading without some method of
restricting solids loading. Dale^ recommended the volume of an
aerobic lagoon for the Midwestern states as 0.31 cu m/kg of hog on
feed. This was determined by using a BOD5 loading of 50 kg/ha, of
lagoon. He also concluded that aerobic lagoons must be cleaned after
several years and weeds kept under control.
Unaerated Anaerobic Lagoons
Anaerobic decomposition has become one of the most common treatment
alternatives for the swine producer. The anaerobic process usually
results in undesirable odors, but the major benefit is the large
flexibility to degrade organic naterial. Dornbush19 has described
the anaerobic process as follows: (a) In the initial stage, the
complex materials such as carbohydrates, proteins, and fats are
biologically converted and fragnented by hydrolysis and fermentation
to the simpler organic end products including aldehydes and alcohols
but principally fatty organic acids; and (b) during the second or methane
formation stage the organic acids produced during the initial breakdown
are converted by the me thane-forming organisms to gaseous end products,
principally methane and carbon dioxide. Waste stabilization or organic
removal is directly proportional to the methane produced.
According to Muehling^ the gas produced in a properly operating
anaerobic digester is about 60 percent methane with the remainder of
the gas being carbon dioxide and small quantities of various intermediate
products. Muehling further judges that it is the intermediate products,
less than one percent of the gas produced, which are responsible for
odors.
Oswald2^* conducted extensive studies of lagoons in California and
concluded that to develop an anaerobic lagoon that will be odor-free,
the design must result in environmental conditions favorable to
continuous methane formation.
Loehr21 stated that in general the main purpose of anaerobic lagoons
is the removal, destruction,and stabilization of organic matter and
not water purification. He concluded that anaerobic lagoons offer
considerable potential for handling and treating concentrated animal
waste. Loehr22 later concluded that anaerobic lagoons are practical
only when used prior to further treatment and disposal.
Summarizing the status of livestock lagoons, TaiganideS23 stated that
reports on the variable success of anaerobic lagoons reflect the
variety of criteria and standards used to evaluate effectiveness. He
felt that even though considerable research has been done in the past
21
-------
15 years, it is still not clear how to design and operate an anaerobic
lagoon so it will not occasionally give off undesirable odors.
Loading Rate Recommendations for Anaerobic Lagoons
Loading recommendations for anaerobic lagoons differ widely within the
United States depending upon the loading parameter used and geocliraatic
conditions. Loading recommendations range from about 1 to 24 cu m of
lagoon volume/45-kg hog depending on the reference used and the particu-
lar part of the United States being considered (Muehling^).
Early recommendations given by Ricketts2^ and Jedele and Hansen25 called
for about 1.4 sq m of surface area/hog and a depth of 0.9 to 1.5 m.
These lagoons were expected to function as municipal aerobic lagoons,
but did not because of overloading.
Eby2(> divided the United States into regional areas and based recom-
mendations on municipal lagoons which varied from 3,750 pigs/ha, of lagoon
in the northern United States to 15,000 pigs/ha, of lagoon in the southern
United States. The 1970 interim specifications for animal waste lagoons
prepared by the Soil Conservation Service2? were based on work by Eby
but incorporated a 25-percent safety factor. Loading rates are specified
as minimum surface area/animal for four different climatic zones within
the United States. For the Southeast (Zone B), the loading rate is
1680 kg of BOD5/surface ha. If the minimum specified depth of 1.8 m is
utilized, this is equivalent to 2.2 cu m of lagoon volume/45-kg hog.
The latest engineering standard for disposal lagoons from the Soil
Conservation Service28 specifies 12.8, 9.6, 6.4, and 3.2 kg of VS/1,000
cu m of lagoon volume/day for the four climatic zones of the United
States for swine. This is equivalent to about 1.9 cu m/45-kg hog
for the Southeast (Zone B).
Many reserachers now feel that anaerobic lagoons should be designed
on a volume basis rather than on surface area. Hart and Turnerl2 varied
their loading rates to determine the effect of different loading inputs
on small test lagoons 1.22 m in diameter and 2.1 m deep in California.
They compared lagoons loaded once/week at the rate of 3.5, 1.9, and 1.3
cu m of lagoon/45"kg hog and found relatively no difference in the
appearance of the three lagoons. They felt that a loading rate of 1.9
cu m or even 1.3 cu m of lagoon/45~kg hog did not result in a parti-
cularly odorous lagoon and that 3.5 cu m/45-kg hog was very satisfactory
for California. From these studies they concluded that anaerobic lagoons
should be designed on a volumetric basis.
29
Willrich reported studies with several small experimental micro-
lagoons on the Iowa State University farm. He recommended a minimum of
0.124 cu m of lagoon/kg of hog and about 0.3 cu m/hog additional volume/
year for sludge storage.
22
-------
Based on their field investigations, Dornbush and Anderson30 reported
that the design of anaerobic lagoons should be on a volumetric basis.
For South Dakota they suggested a satisfactory rule of thumb to be in
the range of 3.7 to 4.8 cu m of lagoon/hog. Curtis31 also studied
existing lagoons in South Dakota and concluded that tentative design
criteria should provide a volume of 2.1 to 2.8 cu m/hog.
A volume of 0.124 cu in/kg hog should be provided according to the
Midwest Plan Service32 Also Included is a recommendation for increased
volume because swine wastes will cause sludge to accumulate in a lagoon
at a rate of about 0.34 cu m/year/hog. Swine waste contains about 0.08
kg fixed or inert solids/45-kg hog/day or about 29 kg/year. Depending
upon the sludge moisture content the corresponding volume will be pro-
portionally larger than the 0.001 cu m which 1.0 kg of water occupies.
33
Lynn studied the use of lagoons for treatment of hog wastes in South
Carolina and compared the loading rates of one market-size hog per 1.7,
3,4, 5.1,. and 6.8 cu m of lagoon. The 6.8 and 5.1 cu m lagoons were more
efficient in removing BODs. The 1.7 and 3.4 cu m lagoons had highly
offensive odors at times. From these studies he concluded that the
quality of lagoon effluent was significantly affected by the loading
rate and recommended a minimum of 5.1 cu m of lagoon/market-size hog
and also 1.4 additional cu m of space to prolong cleaning time up to
five years.
23
According to Taiganides, design loading rates for anaerobic lagoons
should be 0.016 to 0.238 kg of VS matter/day/cu m lagoon volume.
From the standpoint of odor acceptability, the recommended loading
rate is 0.029 kg of VS/day/cu n,, which is equivalent to about 6.5 cu m
lagoon volume/45-kg hog.
Miner34 surveyed much of the available research and recommended a lagoon
loading rate of 0.016 to 0.16 kg vS/cu m lagoon volume/day for the United
States. This ten-fold range for the Midwestern climates 0.08 kg VS/
cu m/day, which is equivalent to 3.54 cu m lagoon volume/45-kg hog, was
recommended.
For Louisiana,Barr e_t aJL.35 recommended that a swine lagoon should be
1.5 m deep, with 3.7 sq m surface area and 5.1 to 5.7 cu m capacity/45-
kg animal.
Baldwin and Nordstedt36 divided the state of Florida into two areas and
recommended 2.38 cu m/45-kg hog for south Florida and 3.34 cu m/45-kg
hog for north Florida.
Hermanson and Watson37 recommended .that a swine waste lagoon should be at
le-st 2.4 m deep and have a capacity of 0.093 cu m/kg of hog. State of
23
-------
o o
Tennessee Criteria also recommended that a swine lagoon should have a
capacity of 0.093 cu m/kg of hog with a minimum acceptable liquid depth
of 2.7 m.
An engineering practice for anaerobic lagoon design (ASAE33) is being
developed for endorsement by the American Society of Agricultural Engineers
which incorporates a loading adjustment factor for various geographical
areas in the United States. This load factor ranges from 0.6 to 1.1,
and criteria are based on 96 kg of VS/1,000 cu m/day. For a load factor
of one, this corresponds to 2.26 cu m/45-kg hog. Recommended unaerated
lagoon loading rates are summarized chronologically in Table 1 with the
most recent listed last.
Table 1. ANAEROBIC LAGOON LOADING RATES
Reference
Ricketts and
Jedele and
Hansen2^
Eby26
Dornbush and
O f\
Anderson
Hart and
Turner
29
Willrich
T1
Curtis
Midwest Plan
Service
Barr35
Loading recommendation
1.4 sq m/hog- -1.5 m deep
(Missouri and Illinois)
3,750 (far northern U.S.) to
15,000 (far southern U.S.)
hogs/ha.; 1.5 to 3 in deep
3.7 to 4.8 cu m/hog (South
Dakota)
3.5 cu m/45-kg hog
(California)
Uniform loading--0.062 cu
m/kg of hog Intermittent
loading--0.057 cu m (Iowa)
2.1 to 2.8 cu m/hog
(South Dakota)
0.124 cu m/kg of hog (North
Central Region)
3.2 to 3,7 sq m/hog--
1.5 m deep (Louisiana)
Cu m
745
1.3
1.5
1.1
3.0
2.1
3.7
3.5
2.8
2.1
5.6
4.9
of lagoon
kg hog
to 2.1
m deep --
to 4.1,
m deep --
to 8.2
to 4.8
to 2.8
to 5.6
24
-------
Table 1 (continued). ANAEROBIC LAGOON LOADING RATES
Reference
Loading recommendation
Cu m of lasoon
//,
Taiganides23
Hermanson and
Watson
37
Lynn
33
34
Miner
Soil Conservation
Service^?
Soil Conservation
Service^
Baldwin and
Nordstedt
36
State of Tennessee
Dept. of Public
Health
38
ASAE
39
0.016 to 0.238 kg VS/day/cu m
of lagoon (Variation for U. S.)
0.093 cu m/kg of hog
(Alabama)
5,1 cu m/market hog
(South Carolina)
0.016 to 0.16 kg VS/day
cu m of lagoon
(Variation for U. S.)
0.00008, 0.00012, 0.00016
and 0.00036 surface ha/hog for
climatic zones A, B, C, and D
respectively--!.8m deep
254, 193, 128, and 64 kg
VS/1000 cu m of lagoon/day
respectively for climatic
zones A, B, C, and D
0.111 kg VS/day/cu m of lagoon
(South Florida)
0.080 kg VS/day cu m of lagoon
(North Florida)
0.093 cu m/kg of hog--
2.7 m deep minimum
95.4 kg VS/1,000 cu m of lagoon/
day times a correction factor
of 0.6 to 1.1 (Variation for
U. S.)
1.6 to 24.0
4.2
2.8
2.4 to 24.0
(3.5 for Midwest)
1.5 to 6.<
1.4 to 5.(
2.4
4.2
1.3 to 2.5
Nordstedt and Barth40 noted that there are two loading rates which
must be considered, the organic and hydraulic loading, and in most
cases the organic loading rate will be the limiting factor. Correspond-
25
-------
in^ly, the following conclusions were presented. First, a very wide
range of anaerobic lagoon loading rates will produce satisfactory
results. Second, the organic and hydraulic loading rates affect the
rate of sludge accumulation and influence future sludge removal-main-
tenance costs. Third, the hydraulic loading rate directly affects
costs for terminal land disposition.
Series Lagoons
The use of series lagoons has attracted attention; however, the
efficiency of removal generally decreases as the number of lagoons
increases. Nordstedt et ajL. studied a multi-stage lagoon system for
treatment of dairy waste and reported that BOD5 reduction in the first,
second, and third lagoons was 88.4 percent, 59.7 percent, and 22.6
percent, respectively.
* in his early lagoon studies observed that when an existing lagoon
was too small, 'the first step was to consider one or more connecting
lagoons. He also concluded that two or more small lagoons in series
are more efficient than one large lagoon with the same surface area
since the first will be anaerobic and successive lagoons may be aerobic.
Lagoon Performance
While reviewing the state-of-the-art of anaerobic lagoons, Dornbushl9
concluded that the following areas need more research: (a) the thermal
environment throughout lagoons, (b) the nature of the organics in both
the lagoon influent and effluent, (c) the actual detention time of
organics within the lagoon, and (d) the relative importance of sludge
accumulations as a major source of nuisance odors.
42
Barsom conducted an extensive review on the performance of municipal
lagoons and discussed the following lagoon problems: (a) Poor effluent
quality due to short-circuiting or reduction of designed detention
time, (This can be due to poor design of inlet and outlet structures,
thermal stratification, or hydraulic overloading), (b) presence of
odors and other aesthetic failures due to production of hydrogen sulfide
and growth of algae which remains suspended in effluent, and (c) water
losses due to percolation from lagoon bottom into groundwater. He
also found that sufficient data were not available for many lagoons on
the quality and quantity of influent and effluent so that actual per-
formance could be evaluated.
The Missouri Basin Engineering Health Council43 conducted a review on
the current state-of-the-art of wastewater lagoons. The following are
some of the conclusions for anaerobic lagoons. (a) Anaerobic lagoons
are useful in obtaining up to 80 percent BOD5 reduction for concentrated
organic wastes. (b) Anaerobic lagoons must be followed by aerobic
treatment for high quality effluent. (c) Anaerobic lagoons have been
26
-------
most successful in treating meat packing wastes, (d) the future of all
TiTr^ ^aS°°ns dePends upon proper design and operation in relationship
witn the fundamental biochemistry of the microbes in the various systems.
Lagoon Loading Frequency
Early designers of municipal waste treatment systems recognized the need
tor very frequent loading in order to maintain a viable microbial
population. For example, the Water Pollution Control Federation, Manual
of Praetxce No. 11^ states, "Organisms in a digester are most efficient
when food is furnished in small quantities at frequent intervals."
McGhee investigated the effect of intermittent sludge feeding and the
consequent fluctuations in volatile acid concentration on digester
performance. He concluded that better digestion could be achieved by
regular and more frequent feedings.
Willrich recommended 0.124 cu m of lagoon volume/kg of hog for inter-
mittent loading, but felt that uniform loading was such an advantage that
only 0.062 cu m of lagoon volume/kg of hog was needed. Miner34 recommended
a loading rate of 0.08 kg of VS/cu m of lagoon volume/day for continuous
loading in the moderate midwestern climates.
Lagoon Aeration
Dale^" and Loehr ' in reviewing the status of animal waste treatment
techniques emphasized that it is generally not feasible to design and
operate aerated systems for direct effluent discharge to receiving
waters. However, aeration for odor control and additional pretreatment
prior to terminal land application has been supported as a technically
feasible approach by many workers.
Aerobic treatment of swine waste is generally accomplished by an in-
house oxidation ditch or an extramural lagoon with a floating surface
aerator. Reports (Jones e_t aj^.4^) On oxidation ditch performance have
indicated BOD5 reduction of about 90 percent and about 40 to 50 percent
reduction of VS, along with complete odor elimination. Generally,
gravity overflow to a lagoon must be provided because excess sludge
and surplus water must be handled. Potential disadvantages of the in-
house oxidation ditch are the impact on animal environment and energy
requirements.
Design and operational criteria for aerated lagoons vary greatly in the
literature dependent upon treatment strategy and equipment. Aeration
recommendations for odor control range from satisfying the BOD5 to 1/3
that value and aeration rates for complete treatment are reported to be
from one to two times the daily BOD5 input. Size requirement recommenda-
tions for aerated swine waste lagoons are from 0.0025 cu m to 0-062 cu m
of lagoon volume/kg of hog.4' 10> 48' 49
27
-------
Aeration can be employed to conserve nitrogen by conversion to nitrate
or enhance nitrogen removal by ammonia volatilization and nitrification-
denitrification depending upon unit management. Although biological
denitrification generally requires sequential reactors for nitrogen
oxidation to nitrite followed by an anaerobic unit for biological
conversion of nitrate to nitrogen gas, nitrification-denitrification
can be accomplished in a single reactor providing that both the
oxidizing and reducing environment is present. An oxidation ditch
which becomes anaerobic before the liquid recycles to the rotor can
provide the aerobic-anaerobic sequential environment for denitrification
(McKinney e_t_ a_l..^0). A diphasic lagoon in which aeration is restricted
to the surface area can also provide the aerobic-anaerobic interface
required for biological denitrification (McCalf and Eddy^l). These
types of units require a delicate balance to achieve a nitrification-
denitrification cycle for continuous or batch generation of nitrates
and then subsequent conversion to nitrogen gas.
Lagoon Evaluation by Oxidation-Reduction Potential (ORP) Measurements
The ORP has been used in anaerobic systems to measure the degree of
reduced conditions. A definite relationship exists between ORP and the
aerobic-anaerobic condition of wastewater. Eca]_ is the ORP measured
with a platinum electrode and a saturated calomel electrode, whereas
Eh stands for measurements with bright platinum electrodes corrected
to coincide with a standard hydrogen electrode. Microorganisms present
will also be influenced by reactor ORP, and thus corresponding
metabolic end products. Microbes operate in an Ecaj_ range of +400
to -200 millivolts (mv) with anaerobes in a range of +50 to -400 mv
and facultative bacteria in the range of +50 to -100 mv. ^
Luddington" Was one of the first to use ORP to control the amount of
aeration employed in investigating odor threshold between aerobic and
anaerobic treatments of liquid chicken manure. VS reduction, hydrogen
sulphide production, and time after termination of aeration before
hydrogen sulphide production for different ORP levels were studied.
54
Converse et al. reported that the En potential varied from +400mv
for aerobic to -250mv for highly anaerobic liquid waste treatments.
Odor was found to increase dramatically below a threshold level of
-50mv to -lOOmv. Relationships between odor and ORP were suggested
as the most reasonable way to predict nuisance potential of an anaerobic
treatment unit.
28
-------
OVERVIEW OF LAND APPLICATION
Characterization of
A prediction of the effects of land disposal of animal wastes on soil
properties and plant growth requires a foreknowledge of waste composition
and recexver system characteristics. Generally, the composition range
tor waste constituents is so great that analysis of the specific residue
is an absolute necessity before land application judgments or recommenda-
tions can be competently made.
Nitrogen
Nitrogen has received primary attention in judging the effects and, thus,
allowable^rates for land application or disposal of manure. Ammonia
and organic nitrogen can be easily transformed by soil microorganisms
to nitrate-nitrogen. Nitrate-nitrogen leaches easily through soil
because it is an anion that has low sorptive capabilities and does
not form insoluble precipitates. Excess nitrate can contribute to
the stimulation of algal blooms and present potential health hazards
in potable water supplies if concentrations exceed 10mg/l as nitrogen.
Several researchers have measured increases in total soil nitrogen
after heavy application of beef waste and dairy-manure slurries. However,
applications of animal waste will generally not increase steady-state
soil nitrogen unless large quantities are continuously applied.•"> -*"
The accumulation and movement of nitrate-nitrogen in soils resulting
from animal waste applications have been investigated by many workers.
Mineralization of organic nitrogen into nitrate is most rapid during
the first year following application and steadily declines in subsequent
years. The decay series concept of nitrogen mineralization developed
by Pratt et_ al. describes this declining rate of nitrate-nitrogen
production with time. These rate relationships for various manures
can be used to estimate the amount of nitrate-nitrogen that will then
become available for plant uptake or leaching into groundwater.
Nitrate-nitrogen in excess of crop needs or root sorptive capabilities
is generally leached to lower soil zones. In some cases little if any
nitrate movement was recorded even when large quantities of waste were
applied. Much of this variability can be explained by differences in
soil-water relationships. Soils that are well drained usually have a
greater potential for nitrate movement and for applications of animal
waste. Soils with restricted drainage will usually have a greater
potential for nitrate transport because of insufficient leaching volume
or anaerobic conditions which can lower nitrification and augment
denitrification potentials. Therefore, nitrate generation and movement
in soil profiles is governed by many factors requiring detailed and
coordinated evaluation.
29
-------
Nitrogen losses occur during land application and after deposition on
the plant-soil receiver system. Kolliker and Miner^S reported an
unaccountable nitrogen loss of 2,307 kg/ha, in soil receiving anaerobic
lagoon effluent by sprinkler irrigation. Olsen et al^.59 found that
20 to 76 percent of the nitrogen added to soil by dairy manure was lost
through volatilization. Wallingford"^ and Meek et al_. measured
unaccountable nitrogen losses ranging from 6,7 to 100 percent for soils
receiving beef feedlot manure. These .losses were attributed to
denitrification and illustrate that denitrification can significantly
lower the potential for nitrate leaching after land disposal of
manure. However, caution must be exercised so that assumed nitrogen
reduction by denitrification does not lead to excessive and thus,
negligent land loading criteria.
Phosphorus
Verifications of the fact that phosphates are immobilized in most
soils are abundant in the literature. Phosphorus transport is
effectively restricted by incorporation of manure into the soil, and
thus, erosion control is generally sufficient to protect water.
quality.
Inorganic Salts
Animal waste can improve soil fertility by the addition of inorganic
salts, such as potassium, calcium, and magnesium. However, excessive
accumulations of these salts and sodium result in increased salinity
and thus, reduced fertility. The form of the ion and regional moisture
conditions greatly affect the accumulation and movement of these
constituents in a plant-soil receiver system. Generally, salt buildup
as a result of even heavy manure applications does not reduce soil
fertility or impede infiltration in the humid southeast and thus
does not control application rates.
Heavy Metals and Trace Elements
Concern has developed that land-applied swine waste could be toxic
to plant growth or grazing animals because of feed copper additions
up to 250 ppm as a dietary supplement. Research on plant and soil
accumulation of copper under various geoclimatic conditions for
applications characteristic of feed additive levels of 125 ppm or soil
ammendment applications up to 3.4 kg/ha./yr have shown from little effect
(Hedges et a_1.62 and Chaney°3) to toxic conditions in cover grass
(Humenik64).
30
-------
Quantity and Quality
Miner and Willrich65 concluded several years ago that "although runoff
trom teeding areas confining animals other than cattle may be expected
co oe high m.. organic matter, .no data are currently available concerning
these sources." Research directed at collecting data or developing
models to access pollutional potential of runoff from land areas which
have been used for terminal disposal of swine waste has just recently
been initiated. Early work by Robbins et aJL.66 showed that the pollutional
potential of runoff from watersheds where swine waste was land spread,
hogs were on drylot, and even where swine had stream access was similar
to the natural pollutional load on streams draining agricultural lands
devoid of farm animals. Mass balances for a watershed with 200 sows
on a 1.2 ha, drylot plus wastes from. 300 confined hogs spread on 2 ha.
show that .69 percent, 1.66 percent, and 3 percent of the defecated
BOD5, TOG, and total nitrogen, respectively, were present in the stream
draining the studied watershed. Their conclusion included that even in
cases where disposal sites are poorly located or swine grown on drylot,
the amount of pollutants (natural plus animal waste) that reach receiving
streams were less than 10 percent of the raw waste deposited in the
watershed.
McCaskey et a_L. evaluated runoff from grassland to which three rates
of dairy manure were applied by irrigation, tank wagon, and dry
spreading. Runoff data variations were attributed to either the type
of waste, rate of application, or differences in soil character and
vegetation. For some plots the amount of pollutant runoff was the
same as for the controlled plots. Thus, few conclusive and quantitative
deductions were drawn from this study.
Additional information about runoff as a diffuse source can only be
inferred from beef feedlot runoff which represents a probable upper
limit and hydrological studies of sediment, pesticide, and fertilizer
transport. Upper estimates on the faction of waste transported in
rainfall runoff from beef feedlots is less than 10 percent of the
defecated load.68' 69> 7°' 71 Madden and Dornbush'0 concluded that
potentially only 5 percent of the total beef waste generated would
leave a feedlot in surface runoff, and this could be reduced to less
than 2 percent if minimum detention facilities, diversion of foreign
drainage, and reduction of runoff velocities were provided.
Concentration ranges for runoff from unpaved cattle feedlots are:
COD from about 1,300 to 8,247 mg/1 to 10,900 - 286,000 mg/1 and nitrate
from 0-17 mg/1 to 0-31/mg/l (Gilbertson et al.6^). Total nitrogen
ranges from 50 to 540 mg/1 (Miner et al,68) to 1,500 to 10,000 mg/1
(Gilbertson et al.69) . Robbins et al..06 data showed similar stream
concentrations for watersheds which had animals on drylot or pastures
and were also used for terminal waste disposal as the control watershed
31
-------
of 0.1-70 mg/1 BOD5, 1-140 mg/1 TOC, 1,012 mg/1 ammonia plus organic
nitrogen, 0.1-9.9 mg/1 nitrate, and 0.1-17 mg/1 orthophosphate
(o-P04-P).
It is clear from the preceding summary that additional research is
required to secure representative order of magnitude values for rainfall
runoff quantity and quality. These data have become more urgent as
attention is currently being directed to non-point source impacts on
water quality. Any universal relationship to define quantity
and quality of runoff from terminal application plots is not
now possible because of the many unique and unpredictable factors
involved. Regional geoclimatic conditions which profoundly affect
runoff and solids transport very considerably. Additional work on
handling methods, land management, cover crops, and environmental
impact are necessary to expedite implementation of land application
techniques which are most effective to achieve the no discharge goal
for the livestock industry. Conjunctively, such data would provide
competent direction for developing non-point source criteria as it
pertains to runoff from areas utilized for terminal application of
animal waste according to best recommended practices.
32
-------
SECTION V
LAGOON STUDIES
SAMPLING AND ANALYTICAL PROCEDURES
Sample Collection
Liquid samples from the model laboratory reactors and the Imhoff cones
were taken at mid-depth by using a pipette and then refrigerated. This
method was satisfactory except for one very heavily loaded unit in which
hindered settling and sludge buildup were pronounced. This reactor was
subsequently sampled with an enlarged tip pipette 2-4 cm below the
liquid surface.
Liquid samples from the field pilot-scale lagoons were taken at mid-
depth in order to avoid collecting scum and surface algal growth.
Sampling was accomplished by means of the APHA-type sampler recommended
in Standard Methods^ for collection of Dissolved Oxygen (DO), 8005 , and
other samples from ponds. Samples were then refrigerated at approxi-
mately 4° C until the specified analyses were performed.
A simple device constructed for collecting sludge samples from lagoons
consisted of a 1.25 cm I.D. aluminum tube with a 90° bend approximately
15 cm from one end. To collect sludge samples, a stopper was placed in
the end of the tube with the 90° bend. The tube was then submerged to
the desired depth and the stopper removed from the bottom of the tube
my means of a cord to allow sample entrance. The tube was then
stoppered at the top and removed from the lagoon. After withdrawal
the stopper was removed and the sludge sample poured into a sample
bottle.
Samples were also taken from several on-farm lagoons. In some cases,
"grab samples" were taken as far out as one could reach at the surface
with a long-handled dipper. In other cases, samples were taken from
a boat with the APHA-type sampler.
33
-------
Sample Preparation
When necessary the samples from the laboratory reactors, Imhoff Cones,
and field lagoons were diluted on a volumetric basis in order to bring
the concentration of the particular parameter involved within the
range of the analysis being performed. It was necessary to blend
concentrated samples with large particulate matter such as raw waste
and sludge, A high-speed blender with a shear-type head design,
Figure 1, was used for 2 minutes at 18,000 rpm to reduce the size of
suspended solids. Particle size reduction is very important when
using instrumental analyses such as for total organic carbon where a
syringe with a small orfice is used for sample measurement and injection.
This prevents screening out of large particles and assures a more
representative sample. Measured parameter variations for characteristic
blended and unblended samples shown in Table 2 indicates that blending
did not alter sample concentrations but only made handling more reliable.
Table 2. REPRESENTATIVE SUMMARY OF VARIOUS PARAMETER CONCENTRATIONS
FOR BLENDED AND UNBLENDED SAMPLES
Multiple or
Fraction of
Reference Rate
4
1
1/2
4
1
1/2
4
Parameter
COD
COD
COD
TOG
TOC
TOC
NH3
Concentration, mg/1
Blended Unblended
3,372
1,396
831
1,050
310
195
1,187
3,686
1,412
831
1,150
325
215
1,268
A more rigorous procedure was used for raw waste samples. These samples
were weighed initially because it is extremely difficult to pipette a
correct volume of unblended raw swine waste or sludge. The sample
bottle including the sample was weighed on an analytical balance to the
nearest tenth of a gram. Then the sample was placed in a 500-ml
volumetric flask, followed by washing out the sample bottle with
distilled water and adding enough distilled water to bring the flask
up to volume. The contents of the flask were then placed in the high-
speed, shear-type blender and blended at 18,000 rpm for approximately
34
-------
SLOTS FOR
LIQUID
FLOW
SHEARING
ZONE
ROTATING
HEAD
STATIONARY
""MEMBER
TOP VIEW OF BLENDING HEAD
Figure 1. Schematic of shear-type blender head located inside
blending container utilized for raw waste sanvnle
preparation, stainless steel construction.
35
-------
2 minutes. The blender contents were then washed into a 1,000-ml
volumetric flask with distilled water and brought to the correct
volume. The diluted sample was then placed in a 1,000-ml Erlentneyer
flask and agitated with a magnetic stirrer because it was observed
that mixing under similar conditions was incomplete in a volumetric
flask. A 50-ml aliquot was then transferred with an enlarged tip
volumetric pipette, to a 500-ml volumetric flask and brought to
correct volume with distilled water. The contents of this flask
were then placed into a 600-ml breaker and agitated with a magnetic
stirrer. A portion was then transferred back into the sample bottle
by using a large tip volumetric pipette. The percent dilution was
then obtained by subtracting the weight of the empty sample bottle
from the weight of the sample bottle plus sample and multiplying by
the correction factor of 0.01. The above procedure gave the best
results of the several different methods investigated to obtain the
most accurate description of the raw swine waste used to load lagoons
and all laboratory reactors. The expected precision of analytical
results for raw waste sample replicates prepared using this procedure
on four, successive, identical samples are indicated in Table 3.
Table 3. VARIABILITY OF COD ANALYSES FOR RAW WASTE SAMPLES
USING SHEAR BLENDER, SAMPLE WEIGHING'AND DILUTION
FOR THIS EXPERIMENTAL STUDY.
Date
2/15/74
COD, mg/1
59,860
61,112
60,084
65,322
Date CODi mg/1
2/18/74 67,263
65,054
57,666
58,243
Analytical. Procedures For Samples
Analyses performed on all samples were generally done according to
procedures described in Standard Methods except for some minor
modifications described herein. All analyses described were not
necessarily performed on every sample but tests performed on a
particular sample were chosen to accomplish the study objectives.
Chemical Oxygen Demand (COD) -
The COD of all samples was obtained by using the procedure outlined
in Standard Methods^ for a 10-ml sample size with the modification
that the 2-hour digestion time was reduced to 15 minutes. The validity
of this modification for animal wastewater samples had been verified
by the work of Overcash, et al.?3
36
-------
TOP r ?°ntent of al1 samPles was obtained with a Beckman Model
1UC Analyzer. The amount of total carbon was determined by injecting
THP (l ^ ^ * Syri^e int° a 950°-C catalytic (Cobalt-
ated, asbestos packing) combustion tube. The sample was then
vaporized and the carbonaceous material completely oxidized to carbon
dioxide and water in the presence of a cobalt catalyst. Zero grade
carrier gas transported the generated carbon dioxide to this infrared
analyzer for measurement. The carbon dioxide detected was directly
proportional to the total carbon of the sample. The actual concentra-
tion was determined from the peak recorded on a strip chart, which was
compared to a calibration curve.
Inorganic carbon was determined by injecting an identical microsample
into a 150 -C combustion tube that contained quartz chips wetted
with 85 percent phosphoric acid. This temperature was below the value
at which organic matter oxidizes. All inorganic carbon was converted
to carbon dioxide which was measured by the infrared analyzer. Inorganic
carbon valves were then determined by comparing the peak recorded on a
strip chart to a calibration curve. The sample TOG was then the
difference between the total carbon and total inorganic carbon.
Biochemical Oxygen Demand (BOD5) -
Several investigators (Clark , Humenik and Overcast^, Busch75) have
reported the inappropriateness of using the BOD5 test to characterize
animal wastes. For this reason primary reliance was placed on TOC
and COD values. However, a number of BOD5 values were obtained
for correlation purpo-ses. Most BOD tests were five-day, 20°-C
unseeded, determinations according to procedures outlined in Standard
Methods. Pond water was used as dilution water for a few samples.
Ammonia and Organic Nitrogen (NH.3N and 0-N) -
Ammonia concentrations for many of the samples were determined by the
Kjeldahl distillation procedure outlined in Standard Methods'^ fOr
concentrated waste. Ammonia was collected in indicating boric acid
solution and the amount determined by titration with dilute (.02 N)
sulfuric acid. During the course of the study an Orion model 95-10
ammonia electrode was purchased and correlation of the ammonia values
for these two methods proved that results were very similar, Figure 2.
Organic nitrogen concentrations were usually obtained by the difference
between total Kjeldahl and nitrogen (TKN) and ammonia values. These
correlations were continued and after a year, the electrode method
became unreliable. Electrode poisoning was postulated as the reason
37
-------
00
B
Q
§
H
2,000
1,600
g 1,200
PS
H
U
w
O
H
U
PM
W
a
o
800
400
POULTRY
170 + 1.07 X
O POULTRY
D SWINE
I
I I
I I
400 800 1,200 1,600
DISTILLATION METHOD, mg/1
2,000
Figure 2. Comparison of ammonia analysis by ion specific
electrode and standard distillation techniques
for swine and poultry waste.
38
-------
, electr°de dependability, Overcash et al.73 After breakdown
of the direct correspondence of the two tests ,"o"nl^ the standard
distillation was used.
Total Kjeldahl Nitrogen (TKN) -
Total kjeldahl nitrogen, the sum of the free ammonia plus organic
nitrogen compounds was determined according to recommendations in
Standard Methods /z for concentrated waste. Procedures for test
efficiency improvement were made frequently, Overcash jet al. 73
Orthophosphate (
The orthophosphate concentration of all samples was obtained by using
the stannous chloride method as outlined in Standard Methods.
Total Phosphate (t-PC^-P) -
Verification was made early in the study that the total phosphate
concentration was approximately equal to the orthophosphate concen-
tration in virtually all samples except raw waste and sludge. There-
fore, in most cases only orthophosphate values were obtained. When
total phosphate concentrations were determined the per-sulfate
digestion and stannous chloride method outlined in Standard Methods
were used.
Bacteria Densities -
Total coliform and fecal coliform densities were determined for a
select number of samples, by employing the membrane filter technique
(Taylor et al. 7^) . Total coliform (TC) counts were made using M-Endo
medium and incubation at 35° C for 20 hours (Kabler and Clark77) .
Fecal coliform (FC) determinations were made using M-FC broth and
incubation at 44° C for 24 hours (Geldreich £t al.78). Values were
reported as colonies/100 ml.
Dissolved Oxygen (DO) -
Periodic dissolved oxygen checks were made of the liquid profile in the
pilot field lagoons. A Weston-Stack dissolved oxygen meter equipped
with an A- 10 D.O. probe and a motorized sampler for submergence was
used and calibrated against standards before and during each test
period.
pH Measurements -
Frequent checks were made of the pH in the lab reactors and the pilot
field lagoons. Determinations were made with a Corning Model 7 pH meter
and a Fisher Standard combination electrode.
39
-------
DEFINITION OF REFERENCE LOADING RATE
At the initiation of the anaerobic Lagoon studies, a reference loading
rate was established as a basis for comparison. This reference rate was
developed on a per time and per volume of reactor basis to allow several
size-scales of experimental devices. In an effort to operate in the
range of common producer usage, the loading rates of several state and
government agencies were consulted.
Units for the reference lagoon loading were chosen to be cubic meters
of reactor volume per 45-kg hog, which by using the waste per week for
a 45-kg hog could be converted to waste quantity per week per cubic
meter of lagoon volume. The reference rate was chosen to be 2.3 m->
volume per 45-kg hog which approximated the Soil Conservation Service
recommendation for lagoons. The chemical oxygen demand was chosen as
the basis for reactor loading and where feasible, the raw waste concen-
tration was held constant at a level characteristic of swine operations.
This concentration was 40,000 mg COD/1 which was roughly equivalent to
the raw waste of a 45-kg hog in a volume of manure, urine, and normal
wash water of 7.5 liters per 45-kg hog per day.
Combining swine waste characterization with reactor volume requirements,
various parametric means of defining the reference loading rate were
determined to compare to other loading rates which have appeared in the
literature, Table 4.
In the experimental plan of this study, various multiples or fractions
of the reference rate were used so.that it would be easy to calculate
the actual loading rates by a multiplicative factor of the reference
rate. The term reference rate referred only to the average amount of
waste per unit reactor volume, m , and per unit time, week. It did not
imply that the frequency of loading was once per week although per week
appears in the units of reference loading. The frequency of loading
varied with the experiments conducted.
SWINE WASTE CHARACTERIZATION FOR EXPERIMENTAL STUDIES
Included as an integral part of the lagoon pretreatment and land appli-
cation investigations was the effect of the swine housing or production
unit upon the raw waste load generated. The waste used to load
laboratory and field pilot-scale experiments was from partially
slatted houses (25 percent slats) with two pits per house and 16 pens
per pit. The pigs maintained in these houses were characteristic of
a finishing operation with constant purchasing and marketing to
40
-------
Table 4. REFERENCE LOADING RATE UTILIZED IN ANAEROBIC LAGOON
EXPERIMENTS EXPRESSED IN UNITS COMMONLY REPORTED
Equivalent Expressions for Reference Raf-p
Metric System
0
2.3 nr per 45-kg hog
f\
24 nr raw waste per week per
1,000 m3 reactor volume (at
7.5 1 per 45-kg hog per day)
0.91 kg COD per week per m3
reactor volume
0.82 kg TS week per m3
reactor volume
0.67 kg VS per week per m3
reactor volume
960 kg BOD^ per day per ka of
surface area
English System
80 ft3 per 100- Ib hog
o
24 ft raw waste per week per
1,000 ft3 reactor volume (at
2 gal per 100- Ib hog per day)
58 Ib COD per week per 1,000 ft3
reactor volume
52 Ib TS per week per 1,000 ft3
reactor volume
43 Ib VS per week per 1,000 ft3
reactor volume
860 Ib BOD5 per day per acre of
surface area
maintain relatively constant liveweight. Production unit management
was directed to achieve the goal of minimizing water usage to reduce
waste volume. Hence, scraping for sanitation and thermostatically-
activated fans and foggers were used so that when the temperature
was above the comfort level the foggers operated 2 minutes out of
every 10 minutes.
Generally pigs were maintained over three pits at a time, so that at
least two pits had sufficient liveweight for the experiments underway.
These two pits were dumped alternate weeks so that the waste in each
nit was collected over a two-week period. All pits were connected via
underground 25 cm pipe to a flume and a flow splitter. Usually no
flow splitter was used so all the waste from a pit went to the 7,500-1
mixing tank used to load the experimental lagoons. Prior to opening
the pit flap valve and releasing pit contents, the liquid depth at the
middle of each pit length was measured. If the pit volume was greater
than experimental needs, the flow splitter was adjusted to divert a
41
-------
fraction of the waste to the on-farm Lagoon. A flume and stage recorder
were also installed so that total flow or waste volume could be determined.
Total pit emptying took considerable time because final slurry drainage
was very slow. When emptying was complete, the solids level was measured
at the middle of each pit length. If there was no flow splitting, the
height of liquid in the mixing tank was measured to verify volume drained.
After each pit was emptied approximately 2,000 I of precharge water
was added to reduce manure adhesion to the pit floor. Over the length
of these experiments, it was observed that when such precharging was
practiced, a pit slurry with no discernible settled or packed solids was
characteristic. However, after draining a bottom solids layer remained
indicating that the floor slope was not sufficient to maintain an
adequate cleaning velocity. Nevertheless, it was observed that over
many weeks this solids layer did not build and was resuspended when the
pit precharge was added. Thus when a steady filling and draining
pattern was established, easy evaluation of the characteristic swine
waste load was possible.
After the swine waste entered the mixing tank, it was agitated for 15
to 30 minutes and then a representative sample was taken for analysis
and loading of laboratory units. After the analyses were performed,
the amount of water needed to dilute to approximately 40,000 mg COD/1
was added. Mean data for raw waste analyses and diluted waste input
for study units are listed in Table 5. In addition swine feces were
collected and mixed with water to give a waste concentration of
approximately 40,000 mg COD/1 for certain laboratory studies. At
various times, feces and urine have been analyzed for various para-
meters separately. This broad range of raw waste samples allows a;
more complete evaluation of swine waste and slurry, from a slatted
floor-manure pit production unit.
Two types of swine waste evaluations were derived from this study. The
first was the quantities per 45-kg hog per day of waste volume, organic
load (COD and TOC), phosphorus, and total Kjeldahl nitrogen (TKN) for
an underfloor pit system. Secondly, the ratios of.various parameters
were analyzed. This second evaluation was less subject to the consid-
erable variations often encountered in quantifying the generated waste.
These parameter ratios also allowed expansion of data limited to one or
two parameters to better characterize waste loads. Additionally, these
parameter ratios could be used as a tool for analytical quality control.
42
-------
Table 5. MEAN CONCENTRATIONS OF RAW SWINE WASTE AS USED
FOR INPUT TO LABORATORY AND FIELD PILOT-
SCALE EXPERIMENTS
Experiment
First Imhoff cones
Second Imhoff cones
Laboratory 14-1 reactors
Field pilot-scale units
11/73-5/74
Mean raw waste concentration
3/73-5/74 before dilution
Parameter Concentration.,
COD
31,000
40,000
34,000
40,000
74,000
TOG
9,000
12,600
11,000
13,500
21,000
TKN
1,800
2,800
2,300
3,100
4 , 600
mg/1
o-P04-P
500
700
600
1,000
The total generated waste volume was determined by two methods. ' One
was to measure the change in pit liquid height each week as gallons of
input. The second was to measure the total volume drained from a pit
with appropriate substraction of the precharge volume. From these two
methods, the waste volume for a unit with a manure pit below partially
slatted floors where manure is scraped and minimal water used was
determined to be between 3.8 and 5.7 liters per 45-kg'hog per day.
This was close to the waste volume for feces plus urine of 3.0 to 3.8
liters per 45-kg" hog per day as extensively documented in the literature
(Overcash, et al.^").
The produce of the drained pit volume and constituent concentration
yielded the waste generation data presented in Table 6 in conjunction
with the valve judged most reliable in a recent literature review
(Humenik et'aJL.'). All chemical parameters determined for the waste
at this unit were lower than the reference literature values.
43
-------
Table 6.
WASTE GENERATION FROM UNDER SLAT PIT RECEIVING
SWINE WASTE
Parameter
H20
TKN
o-P04-P
t-P04-P using
ortho:total=.7
COD
TOC
Waste Amount, kg/d/45-kg hog
9/9/74-
12/31/74
3.5
0.017
0.0035
0.0050
0.28
0.080
1/1/75-
3/31/75
4.2
0.019
0.0038
0.0054
0.28
Literature values for
fresh swine wastes
3.8
0.022
___
0.0064
0.32
0.091
The ratios of various waste constituents were calculated, for the various
sources of raw waste, raw waste diluted to about 40,000 mg COD/1, and
feces plus water, Table 7. The COD:TOC, NH3-N:TKN, and the o-PC^-P:
t-P04-P ratios have an upper theoretical numerical value. The COD:TOC
ratio as defined in Standard Methods?-^ has an upper limit of 5.3 which
would indicate the most reduced organic carbon compound, methane, and a
lower limit of 0 representing the most oxidized organic carbon compound,
carbon dioxide. In this study, the CODrTOC ratio fluctuated around
3-3.5 but occasionally values above the upper theoretical limit of 5.3
were recorded indicating analytical or sampling errors. Literature
value ratios of 5.3 or more were considered suspicious and not included
for comparison in Table 6.
The fraction of the total nitrogen load which was in the more readily
available ammonia (ammonia or ammonium ion) form was approximately 55
percent. The ortho: total phosphorus ratio indicated the fraction
which was loosely bound and for raw swine waste this was about 70
percent. Thus a relatively large portion of the nitrogen and phosphorus
in swine wastes was in the readily available form.
44
-------
Table 7. CHEMICAL PARAMETER RATIOS OF VARIOUS SOURCES OF SWINE WASTE
Source
Agitated contents
of gutter receiving
once/day scraping
Underslat pit
emptied once/ two
weeks-aliquot
collected after
emptying and mixing
Underslat pit
contents emptied
once /two weeks
and diluted to
approximately
40,000 mg COD/1
Fresh feces mixed
with water to
approximately
40,000 mg COD/1
Ratio
COD:
TOG
2.8
3.9
3.5
2.6
TKN:
TOG
0.22
0.27
0.31
0.18
M3:
TKN
0.52
0.51
-
o-P04-P:
TKN
0.21
0.26
0.24
o-P04-P:
t-P04-P
0.72
0.83
-
In routine analyses of swine waste as well as a number of other char-
acteristic samples, it was useful to have a range of acceptable values
to screen for errors. Absolute constituent concentrations could
fluctuate and still be acceptable for evaluations conducted because
concentrations of raw waste often changed due to dilutional effects.
Thus the ratio of concentrations and not the concentration itself
was deemed a better analytical error screening tool. Ratios of the four
most used parameters, COD, TOG, TKN, and O-P04-P, (Table 7), would indicate
errors in any constituent such as TKN because the COD:TOC ratio would
be in an acceptable range but the TKN:TOC and TKN:COD would be outside
the range indicating the need for TKN reanalysis. Similar arguments
45
-------
could be made for other parameter errors. When two errors may cancel
then reliance of the 0-PC>4:TK.N and other ratios was needed. This proved
to be a useful first screening test even though it was not foolproof
when several analytical errors in one sample were made. When several
errors were indicated a complete rerun was instituted.
LABORATORY SCALE EXPERIMENTS
Imhoff Cones
Loading Frequency and Sludge Management Study-
Two sets of one-liter Imhoff cones with three cones per set were begun
as the first experimental series. The differences among the three cones
in a given set were in the management of the sludge which settled from
the raw swine waste input. These operational differences were designed
to indicate the magnitude of settling for various waste constituents
and interfacial transfers between sludge and supernatant.
For the first set of three cones the loading frequency was once per week,
an interval utilized in most of the laboratory and field experiments,,
The loading rate was the reference level corresponding to 2.3 m3 of
lagoon volume per 45-kg hog. Thus for these cone reactors the input
was 25 ml of raw swine waste at a controlled concentration of 40,000
mg COD/1.
The first cone was managed such that after the weekly raw waste loading
the settled solids were drained from the bottom through a stopcock.
Thus except for the 6 to 8 hours required for initial settling this
reactor operated with no sludge hence the designation 1NS1 (reference
or unity loading, no sludge, one load per week)., A second cone was
loaded as the first but only enough sludge was drained to maintain
about a 12-cm bottom sludge blanket. This scheme allowed interfacial
contact between the supernatant and a controlled amount of settled
material which represented the interfacial area or upper layer of the
total sludge blanket. The designation for the second cone is 1CS1
(reference load, controlled sludge, one load per week). Finally the
third cone, loaded as the previous two, was allowed to accumulate
sludge. This cone served as a control or simulator of usual lagoon
management and was designated 1AS1 (reference load, accumulated sludge,
one load per week).
The second set of cones was different from the first in that the same
reference waste input was added but at a frequency of three times per
week. Thus 8.3 ml of raw swine waste were added on Monday, Wednesday,
and Friday. Sludge management was the same as outlined for the first
set of cones. The cone designations were 1NS3, 1CS3, and 1AS3. These
cone experiments are depicted in Figure 3.
46
-------
COVER
PLATE
LOADING
POSITION
I LITER
SAMPLING LOCATION
16 CM. FROM U MARK
NO SLUDGE
CONTROLLED
SLUDGE
ACCUMULATED
SLUDGE
Figure 3. Schematic of sludge management experiment with one-liter Imhoff
cones loaded with swine waste.
-------
Procedurally, the cones were filled with 990 ml of supernatant from a
field lagoon. The four cones 1CS1, 1AS1, 1CS3, and 1AS3 received 10
ml of sludge from the same field lagoon. The two no-sludge cones
received an additional 10 ml of field lagoon supernatant so that all
reactors began with 1 liter of material. The use of lagoon supernatant
and sludge was to provide an inoculum of microorganisms in both reactor
phases. Loosely fitted cover plates were put over each cone to reduce
evaporation losses.
Prior to each loading event, whether once or three times per week, a
20-ml sample of the supernatant was taken with a pipette at the 250-ml
level from each cone. To evaluate the various processes present in an
anaerobic lagoon, several parameters were monitored. Organic stabili-
zation was indicated by changes in COD and TOG. The breakdown and
subsequent losses of nitrogen were very important because nitrogen is
generally the constituent in animal waste which limits terminal land
application. Hence, total Kjeldahl nitrogen (TKN) was measured and
phosphorus, being a conservative element, was analyzed to check mass
balance experiments. Usually orthophosphorus (o-PO/-P) was determined
because a high percentage of the total phosphorus was found to be ortho-
phosphorus and because of the greater simplicity and speed of the ortho
test.
The raw waste was added by gently releasing the pipette input onto the
liquid surface at the center of each cone (Figure 3). After 6 to 8
hours, the settled solids were drained from the no-sludge cones as
previously described. Controlled sludge units were drained of sludge
periodically because of difficulty in clearly seeing the sludge height.
Sludge sample volumes and COD, TOG, TKN, and o-PO/-P concentrations
were determined. The cover plates on each cone were not completely
successful in preventing evaporation so the volume of the reactors
decreased during the experiment. The final volumes are given in
Appendix Al along with all of the sample and sludge analyses. The
change in volume was taken into consideration in data analyses.
Concentrations of four waste constituents in the supernatant of the two
cone sets are shown in Figures 4-11. Steady-state supernatant concen-
trations were reached after eight to twelve weeks, and the experiment
was terminated after about nineteen weeks. The slight increase in
parameter concentrations near the end of the experiment was due to
evaporative losses. The supernatant was removed from each cone and the
volume and constituent concentrations determined. Then the sludge was
removed, with volume and concentrations being measured. Finally, the
cone walls were washed and this cleanout liquid was evaluated. With
these termination measurements, the feed and initial charge measurements,
and the volume and concentration of all the samples removed, mass
balances for TOG, COD, TKN, and o-P04~P were performed. The amounts
of each parameter and the percentages of raw waste input which were re-
covered as settled material or sludge are listed in Table 8. The
48
-------
5,00{3 _
60
e
o
M
H
U
3
O
O
Q
O
O
O
S3
8
3
o
M
§
4,000
3,000
2,000
1,000
ALL CONES LOADED AT REFERENCE RATE
AND ONCE PER WEEK FREQUENCY
NO SLUDGE, 1NS1
CONTROLLED SLUDGE, 1CS1
ACCUMULATED SLUDGE, 1AS1
0 12 24 36 48 60 72 84 96 108 120 132
TIME, days
Figure 4. Supernatant COD concentration changes in Imhoff cones begun with swine
waste lagoon supernatant with or without sludge as inoculum and loaded
with raw swine waste (first experimental set).
-------
o
rH
E?
O
II
H
H
a
w
o
23
O
o
G
o
H
^~^
S
O
FQ
Pi
2,0001
1,800(
1,600
1,400
1,200
1,000
u
H
O
Pi
O
H
O
H
ALL CONES LOADED AT REFERENCE RATE
AND ONCE PER WEEK FREQUENCY
a NO SLUDGE, 1NS1
O CONTROLLED SLUDGE, 1CS1
ACCUMULATED SLUDGE, 1AS1
800
600
400
200
84
96 108 120 132
TIME, days
Figure 5. Supernatant TOG concentration changes in Imhoff cones begun with
swine waste lagoon supernatant with or without sludge as inoculum
and loaded with raw swine waste (first experimental set).
-------
1,000
60
e
O
M
H
H
a
w
o
§
H
M
S3
iJ
P
H-3
W
P
EH
O
H
800
ALL CONES LOADED AT REFERENCE RATE
AND ONE PER WEEK FREQUENCY
D NO SLUDGE, 1NS1
O CONTROLLED SLUDGE, 1 CS1
A ACCUMULATED, SLUDGE. 1AS1
600 —
400 —
200 —
0 12
Figure 6.
36 48
96 108 120 132
60 72 84
TIME, days
Supernatant TKN concentration changes in Imhoff cones begun with
swine waste lagoon supernatant with or without sludge as inoculum and
loaded with raw swine waste (first experimental set).
-------
Ul
H
*Z
W
O
O
P-i
I
st
O
PH
O
w
i
pj
CO
s
O
EC
O
ALL CONES LOADED AT REFERENCE RATE
AND ONCE PER WEEK FREQUENCY
a
o
NO SLUDGE, 1 NS1
CONTROLLED SLUDGE, 1CS1
ACCUMULATED SLUDGE, 1AS1
120
40
96 108 120 132
TIME, days
Figure 7. Supernatant o-PO^-P concentration changes in Imhoff cones begun
with swine waste lagoon supernatant with or without sludge as
inoculum and loaded with raw swine waste (first experimental set)
-------
5,000
01
u>
oo
e
2
o
H
z
w
o
o
p
o
o
o
w
p
o
o
o
P5
O
ALL CONES LOADED AT REFERENCE RATE AND
THRICE PER WEEK FREQUENCY.
D NO SLUDGE, INS 3
O CONTROLLED SLUDGE, 1CS3
ACCUMULATED SLUDGE, IAS3
4,000
3,000
2,000 —
1,000 —
0 12 24 36
48
108 120 132
60 72 84
TIME, days
Figure 8. Supernatant COD concentration changes in Imhoff cones begun with swine
waste lagoon supernatant with or without sludge as inoculum and loaded
with raw swine waste (first experimental set).
-------
rH
M)
g
a
o
M
H
H
W
O
S3
O
O
u
o
E-H
s
o
M
/vl
2,000
1,800^
1,600
1,400
1,200
1,000
800
1
e
0
hr
l
T
4
0
j
O
M
O
«
O
O
H
600
400
200 —
ALL CONES LOADED AT REFERENCE RATE
AND THRICE PER WEEK FREQUENCY
D NO SLUDGE, INS3
O CONTROLLED SLUDGE, 1 CSS
A ACCUMULATED SLUDGE, 1AS3
0 12 24 36 48 60 72 84 96 108 -120 132
TIME, days
Figure 9. Supernatant TOC concentration changes in Iirihoff cones begun with swine
waste lagoon supernatant with or without sludge as inoculum and loaded
with raw swine waste (first experimental set).
-------
1,000 —
a
o
H
H
S3
W
U
U
S3
W
O
§
H
M
O
H
ALL CONES LOADED AT REFERENCE RATE
AND THRICE PER WEEK FREQUENCY
800
O
O
A
NO SLUDGE. INS3
CONTROLLED SLUDGE, 1CS3
ACCUMULATED SLUDGE, IAS 3
600
400 —
200 —
12 24
Figure 10.
36
108 120 132
48 60 72 84
TIME, days ' .
Supernatant TKN concentration changes in Imhoff cones begun with
swine waste lagoon superantant with or without sludge as inoculum
and loaded with raw swine waste (first exoerimental set).
-------
60
a
o
H
H
3
H
25
W
o
o
FM
-------
volumes, concentrations, and other data used in these calculations are
listed in Appendix A2. The discussion of these data and overall mass
balances are included at the end of the Imhoff Cone Loading Rate and
Sludge Management Section.
Table 8. SLUDGE OR SETTLED SOLIDS RECOVERY FOR SEVERAL SLUDGE
MANAGEMENT TECHNIQUES FROM IMHOFF CONES RECEIVING
SWINE WASTE INPUTS
First experiment:
operation
Reference loading
rate, no sludge,
once per week
frequency, 1NS1
(19 weeks)
Reference loading
rate, controlled
sludge, once per
week frequency,
1CS1 (19 weeks)
Reference loading
rate, accumulated
sludge, once per
week frequency,
1AS1 (19 weeks)
Reference loading,
rate, no sludge,
thrice per week
frequency, 1NS3 ,
(19 weeks)
Reference loading
rate, controlled
sludge, thrice per
week frequency
1CS3, (19 weeks)
Reference loading
rate accumulated
sludge, thrice per
week frequency
1AS3, (19 weeks)
Sludge Recovered
Parameter
TKN
TOC
COD
o-P04-P
TKN
TOC
COD
o-P04-P
TKN
TOC
COD
o-P04-P
TKN
TOC
COD
o-P04-P
TKN
TOC
COD
o-P04-P
TKN
TOC
COD
o-P04-P
Amount,
me
360
3,087
8,037
225
283
1,967
7,392
242
206
1,554
4,061
229
492
3,087
9,368
301
304
1,960
6,298
216
227
1,650
4,844
220
As percent
of feed
44
79
58
98
34
50
54
106
25
40
30
100
60
79
68
131
37
50
46
94
28
42
35
96
As percent of
feed plus
precharge
20
55
42
66
15
34
38
71
11
27
21
67
27
54
49
89
16
34
32
64
12
29
25
65
57
-------
Loading Rate and Sludge Management Study-
A second experiment was begun with one-liter Imhoff cones to investigate
a much higher raw waste loading rate with the same three sludge manage-
ment schemes previously used. Two sets of cones were used, both loaded
once per week. The first set was similar to the earlier cones in that the
reference loading rate was used. These cones were thus 1'NSl, 1CS1, and
1AS1. The second set was loaded at four times the reference rate which
meant that the cones received 100 mg of raw swine waste (40,000 mg
COD/1) per week corresponding to 0.58 nH reactor volume per 45-kg hog.
Designation for these cones was 4NS1, 4CS1, and 4AS1.
All cones were initiated with one liter of tap water to allow comparison
with the previously seeded cones. The tops were covered, again to
minimize water vapor loss. The sampling procedure was the same as that
previously described with the sample volume still being 20 ml to
minimize volume loss. The sludge management was the same: as the first
Imhoff cone experiment.
Supernatant concentrations of COD, TOG, and TKN reached steady condition
in ten to twelve weeks, Figures. 12-17. At the completion of the experi-
ment, the supernatant and sludge volume and constituent measurements
were made and the corresponding sludge recovery results are shown in
Table 9, (detailed data in Appendix A2).
Imhoff Cone Reactor Discussion-
In the first cone experiment, there did not appear to be any differences
between the supernatant concentration of COD, TOG, TKN and o-PO^-P in
the no sludge, controlled sludge, and accumulated sludge cones. Also,
comparison of Figures 5 and 9 indicates that the more frequent loading
(3/week) did not result in a different quality supernatant than the
once per week loading. Thus, no concentration difference due to sludge
management or loading frequency was evidenced for this investigation.
58
-------
Table 9. SLUDGE OR SETTLED SOLIDS RECOVERY FOR SEVERAL SLUDGE
MANAGEMENT TECHNIQUES FROM IMHOFF CONES RECEIVING
SWINE WASTE INPUTS
Second experiment:
operation
Reference loading
rate, no sludge,
once per week
frequency, 1NS1,
(29 weeks)
Reference loading
rate, controlled
- sludge, once per
week frequency,
1CS1, (29 weeks)
Reference loading
rate, accumulated
sludge, once per
week frequency
1AS1, (23 weeks)
Four times the
reference loading
rate, no sludge,
once per week
frequency, 4NS1,
(1.6 weeks)
Four times the
reference loading
rate, controlled
sludge, once per
week frequency,
4CS1, (16 weeks)
Four times the
reference loading
irate, accumulated
sludge, once per
week frequency,
4AS1, (16 weeks)
Sludge Recovered
Parameter
TKN
TOC
COD
TKN
TOC
COD
TKN
TOC
COD
TKN
TOC
COD
TKN
TOC
COD
TKN
TOC
COD
Amount , mg
549
2,787
10,332
280
1,939
5,617
306
1,538
5,325
1,573
9,028
33,742
1,455
8,126
32,075
1,288
5,183
16,923
As percent of feed
27
30
36
14
21
19
18
20
22
38
38
57
35
34
54
31
22
28
59
-------
5,000
O
M
H
W
O
Q
O
4,000
3,000
ALL CONES LOADED AT REFERENCE RATE
AND ONCE PER WEEK FREQUENCY
D NO SLUDGE, 1NS1
O CONTROLLED SLUDGE, 1CS1
A ACCUMULATED SLUDGE, 1AS1
53
w
O
2,000
o 1,000
§
Pd
u
12
Figure 12.
36 48 60
108 120 132 144 156 168 180
72 84 96
TIME, days
Supernatant COD concentration changes in Imhoff cones begun with
tap water (second experimental set) and loaded with raw swine waste.
-------
§
H
H
S3
w
u
3
o
o
o
H
is
o
pq
23
u
u
o
erf
o
O
H
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
200
ALL CONES LOADED AT REFERENCE RATE
AND ONCE PER WEEK FREQUENCY
D NO SLUDGE, 1NS1
O CONTROLLED SLUDGE, 1CS1
& ACCUMULATED SLUDGE, 1AS1
0 12 24 36
48
60 72 84 96 108 120 132 144 156 168 180
TIME, days
Figure 13. Supernatant TOG concentration changes in Imhoff cones begun with
tap water and loaded with raw swine waste (second experimental set).
-------
1,000
60
e
o
M
H
H
a
w
o
o
H
o
Pi
H
H
S3
3
w
H
O
H
800
600
400
200
ALL CONES LOADED AT REFERENCE
RATE AND ONCE PER WEEK FREQUENCY
D NO SLUDGE, 1NS1
O CONTROLLED SLUDGE, 1CSI
A . ACCUMULATED SLUDGE, 1AS1
0 12 24 36 48 60 72 84 96 108 120 132 144 156 168 180
TIME, days
Figure 14, Supernatant TKN concentration changes in Imho£f cones begun with tap water
and loaded with raw swine waste (second experimental set).
-------
5,000 I—
U)
4,000
ALL CONES LOADED AT FOUR TIMES THE REFERENCE
RATE AND ONCE PER WEEK FREQUENCY
NO SLUDGE, 4NS1
CONTROLLED SLUDGE, 4CS1
ACCUMULATED SLUDGE, 4AS1
3,000 —
2,000
1,000 —,
0 12
Figure 15.
36
84 96 108 120 132
48 60 72
TIME, days
Supernatant COD concentration changes in Imhoff cones begun with
tap water and loaded with raw swine wastes (second experimental set)
-------
2,000
Os
a
o
H
3
W
U
§
o
o
H
S5
O
PQ
Prf
-------
ON
Ln
OC
B
§
U
§
o
w
1
H
M
sa
1-3
Q
.-4
W
O
H
ALL CONES LOADED AT FOUR TIMES THE
REFERENCE RATE AND ONCE PER WEEK FREQUENCY
2,000 1—
D
O
NO SLUDGE, 4NS1
CONTROLLED SLUDGE, 4CS1
ACCUMULATED SLUDGE, 4AS1
1,500
1,000
500 —
12 24
Figure 17.
48 60
108 120 132
72 84 96
TIME, days
Supernatant TKN concentration changes in Imhoff cones begun with
tap water and loaded with raw swine waste (second experimental set).
-------
In the second cone experiment there were occasional periods, for both
the reference and the four times reference rate, in which the no-sludge
supernatant concentration was lower than the controlled or accumulated
sludge cones. There were also periods in which the supernatant concen-
tration for the accumulated sludge cone was higher than the other two
sludge management schemes. However in comparing the three waste
parameters (COD, TOG, & TKN) over the entire length of both cone exper-
iments the differences in supernatant quality were not significant.
That is, results for these tests indicate the presence of a sludge layer
and subsequent interfacial transport was a less significant factor in
determining supernatant quality than biostabilization and other loss
mechanisms ongoing in the supernatant. Thus differences in supernatant
quality due to sludge management would not be large enough to warrant
an engineered system to control or remove settled solids.
The initial settling of the raw swine waste (40,000 tng COD/1) in an
anaerobic supernatant fluid was estimated by the amounts of the
various waste parameters recovered from the sludge fraction of 1NS1,
1NS3, and 4NS1. Sludge removal after initial settling or 6 to 8
hours after loading from the no-sludge cones minimized degradative
breakdown of the settled solids. A conflicting factor is that the
settled material drained from the no sludge cones contained some super-
natant because of the difficulty in detecting the exact sludge depth
due to solids bridging and other visibility or operational problems.
This increased the volume of material drained, but because the super-
natant concentration is only about 5 percent of the sludge concentration
the inclusion of supernatant did not overly influence the amount of
chemical parameters drained from these cones.
An exact percentage estimate of organic or oxygen demanding consti-
tuents which settle initially could not be obtained because of the
variability between the two cone experiments, Tables 8 and 9. The
average percent consituent initially settled for all cone experiments
was 55 percent - COD, 56 percent - TOC, and 42 percent - TKN, Table 10.
The second experiment was better controlled with less evaporation and
more uniform waste input. Thus typical COD and TOC removals with initial
settling may be closer to the results of the second experiment ( 30
percent - 40 percent TOC and 35 percent - 55 percent COD); however,
these conclusions are only tentative. In related swine waste experiments
by Jett, et a_l.80 with .5, 1, and 2 percent total solids mixtures
settling was essentially complete after 60 minutes with from 60 percent
to 80 percent total solids reduction. Settled solids contain inert
materials hence total solids removal would be expected to be higher
than COD or'TOC removals (10 percent to 80 percent versus 30 percent to
50 percent).
66
-------
Table 10. INITIAL SETTLING OF SWINE WASTES EVALUATED FROM IMHOFF
CONE ANAEROBIC REACTORS (NO SLUDGE)
First Experiment
Second Experiment
Average
1NS1
1NS3
1NS1
4NS1
Percentage of raw waste input which
settled after 6 to 8 hours
TKN
44
60 ,
27
38
42
TOG
79 ,
79
30
38
56
COD
58
68
36
57
(
I
55
o-P04-P
98
131
114
Initial settling values for.total Kjeldahl nitrogen were lowest, Table
10, which was expected because about 50 percent - 60 percent of the raw
swine waste nitrogen was in the ammonia or ammonium .form, Table 6 and
thus about 40 percent - 50 percent organic nitrogen. If the settled
TKN was all organic it would mean ,80+ percent settling for organic
nitrogen. It can be assumed that some ammonia was removed by entrap-
ment, precipitation, or another mechanism because organic settling as
measured by .COD and TOC phase separation of the TKN was only 55 to
65 percent. From the. sludge characterization, Appendix B6, sludge,
nitrogen was found to be 65 percent organic nitrogen and 35 percent
ammonia. The raw waste or input was about 40 percent organic
nitrogen and 60 percent ammonia.
Data showed that..about 90 percent - 1.00 percent of raw waste total
phosphorus settled even though the raw waste was about 70 percent
orthophosphorus. Possible mechanisms for this removal were entrapment
or precipitation. The latter was more probable since large amounts of
settling solids would be required to remove such a high percent
0-P04-P by physical entrapment. Precipitation has been proposed by
Smith et aj^.° and these cone experiments lend further evidence
to the importance of the phosphate solubility limit in a swine lagoon
system. Various researchers have postulated ammonium magnesium phos-
phate (NH^MgPO^) as the precipitated, salt which has a gray flake
structure. Problems with pump or pipe occlusion have been attributed
to this precipitation especially in areas with high water hardness or
magnesium content.
67
-------
It was very difficult to maintain the sludge at a constant level
in the controlled sludge cones due to the resulting opaque conditions.
Thus, the solids residence time varied throughout the experiment and it
was difficult to draw any quantitative conclusions about percent removals
or degradation. In the first set of experiments, the average sludge
residence time was five weeks for the controlled sludge cones (CS) as
compared with ten weeks for the accumulated sludge cones (AS) and 6-8
hours for the no sludge (NS) cone tests. In the second experiment,
average solids residence time for the controlled sludge cones was four
weeks versus fifteen weeks and 6-8 hours for the accumulated and no
sludge cones, respectively. The settled solids residence times in the
controlled sludge cones, although variable, were intermediate to the
residence times for the accumulated and no sludge cones. The amount of
COD, TOC, and TKN remaining in the controlled sludge cones was also
intermediate to the other two sludge schemes. However, the exact
decomposition could not be estimated because of residence time differences
and other difficulties described above and verified by the erratic data
summarized in Appendix Al.
The accumulated sludge cones had an average sludge residence time of
10-15 weeks (one half the total experiment span). Sludge recoveries as
a percentage of the raw waste input were 20 percent to 35 percent COD,
20 percent to 40 percent TOC, 20 percent to 30 percent TKN, and 95 percent
to 100 percent o-P04~P. These percentages were fairly uniform for both
sets of cone experiments. The differences between no sludge and accu-
mulated sludge cones were dramatic. About 40 percent to 50 percent of
the COD, TOC and TKN which settled initially had been stabilized or
liberated and thus was not present in the sludge zone in the accumu-
lated sludge cones. This partially explained the slow rate of sludge
accumulation in actual lagoons. As expected, phosphorus compounds
were conservative; and the same percentages were present in all three
sludge management alternatives.
The microbial activity in the supernatant zone relative to the sludge
zone were calculated from the results of the cone experiments. From the
initial settling data, the percent of the raw waste entering the super-
natant and the sludge layers was calculated. The initial assumption
was made that there was no transfer between supernatant and sludge
zones. For the COD, the partition was approximately 60 percent to
the sludge and 40 percent to the supernatant. The expected concen-
tration of the supernatant portion of the raw waste input (40,000 mg/1)
was then:
Hypothetical COD
of the supernatant = (40,000 mg COD/1)(.40)(feed volume)
rraction or raw (feed volume)(1-fraction of feed volume which
swine waste . .. .. , . . , .
is settled solids)
68
-------
This equation would give the hypothetical supernatant concentration of
COD if the raw waste was allowed to settle.
The fraction of the feed volume considered settleable solids varies
considerably depending on the measurement technique or the experiment
performed. However from the literature the fraction which is settled
solids ranged from a maximum of 50 percent to a minimum of 10 percent and
thus the hypothetical supernatant input COD would be between 18,000
rag COD/1 to 32,000 mg COD/1, equation 1. Comparing this input concentra-
tion range with the steady-state supernatant concentration measured
in this study, Figures 4 and 8, the percentage of non~settled waste
input remaining in the supernatant was 3 percent - 10 percent.
The reduction of sludge COD was estimated by the amount left in the
accumulated sludge divided by the amount drained from the no-sludge cones
or approximately 50 percent - 60 percent. For the COD, a comparison
of the amount remaining in the supernatant phase versus the sludge
phase as a percent of the amount entering, those phases indicated a much
greater level of stabilization in supernatant versus the sludge (3
percent - 10 percent remaining versus 50 percent - 60 percent remaining).
This ratio indicated a higher level of tnicrobial activity and hence
higher potential for waste stabilization in the supernatant than in the
sludge zone. The relative supernatant and sludge levels of removal for
TKN, "TOG, and COD are given in Table 11. The ratio of percent remaining
in sludge to supernatant was large for the carbonaceous parameters, but
slightly less for nitrogen. Thus a restriction in the removal of
supernatant nitrogen, a volatilization process, was indicated since
the sludge removals for carbon and nitrogen compounds were nearly equal,
Table 11.
Table 11. RELATIVE SUPERNATANT AND SLUDGE REMOVAL LEVELS FOR IMHOFF
CONES WITH ACCUMULATED SLUDGE AND LOADED WITH SWINE WASTE
Parameter
Total Organic
Carbon, TOC
Chemical Oxygen
Demand, COD
Total Kjeldahl
Nitrogen, TKN
Percent of amount
entering sludge
which remained
55
52
60
Percent of amount
entering superna-
tant which remained
3-10
3-10
7-15
Ratio of percent
remaining in
sludge to that of
the supernatant
6-18
5 - 17
4 - 8
69
-------
Further evidence that there was a partial restriction on nitrogen loss
from these loosely covered Itnhoff cones was obtained near the end of
the second experiment. After steady-state concentration levels of TKN
and NH3~N were achieved the lid was removed and the supernatant concen-
tration monitored, Figure 18. There were slight reductions in the
NH3~N and correspondingly the TKN level indicating increased surface
volatilization. However, the final supernatant concentrations were not
greatly different from those with the lid present as compared to the
input raw waste concentrations. The 2 to 4 cm of freeboard in the cones
was probably as restrictive of air flow over the liquid surface as the
loosely fitting lids.
The comparison of the no-sludge, controlled sludge, and accumulated
sludge cones was designed to clarify the mass transfer exchange across
the sludge-supernatant interface. The major removal mechanism for
organic carbon or nitrogen compounds is gasification and subsequent
volatilization from the supernatant surface. Thus some sludge decom-
position products will diffuse into the supernatant phase. The no-
sludge cone had an immediate sludge-supernatant separation; hence,
relatively no opportunity for sludge decomposition product transfer to
the supernatant existed. Therefore a lower supernatant concentration
was expected. The comparison of the controlled sludge, representing the
upper surface of sludge blanket with the accumulated sludge cones would
indicate whether the surface layer or the entire sludge zone has the
greatest impact on the supernatant. The controlled and accumulated
sludge cone should have interfacial transfer and hence higher superna-
tant concentrations than the no-sludge cone.
Examination of the sludge recovery data, Tables 8 and 9, provided
secondary evidence of the interfacial mass transfer. Constituents
reduced in the sludge can be inferred as entering the supernatant
directly or as metabolic byproducts. Regardless of parameter (COD,
TOG, TKN) the total mass drained from the accumulated sludge (1AS1,
1AS3, 4AS1) and controlled sludge cones was 45 percent - 67 percent
and 65 percent - 95 percent of the amount drained from the no-sludge
cones (6-8 hours after loading), respectively.
However the supernatant concentration in cones with different sludge
management were not significantly different, Figures 4-17. Postulated
reasons were that the supernatant contained a high level of microbial
activity which with once per week loading was very probably underutilized,
Thus these microorganisms effectively used sludge byproducts as well as
the raw waste input as substrates masking any differences in interfacial
transfer. Another mechanism which would diminish any supernatant
differences would/be the direct liberation of breakdown products from
the sludge to the atmosphere. The bubbles carrying carbon dioxide
and methane, which evolve from anaerobic reactors, were evidence of
this direct sludge atmosphere loss. Obviously more precise experiments
70
-------
I
w
s
H
o
H
160
120
80
40
I II
10
20
30
40
50
40
70
80
TIME , days
Figure 18. Supernatant TKN and NHg-N changes associated with removal of the evaporative
loss reduction cover from the Imhoff cones.
-------
will be needed to quantify the sludge-supernatant exchange inferred in
these experiments.
These data indicating sludge loss of the carbonaceous compounds corrobor-
ated the conclusions drawn earlier about the level of supernatant microbial
activity. If some or all of the sludge breakdown products entered the
supernatant the hypothetical fraction of raw waste COD in the
supernatant would be increased, Equation 1. Supernatant concentrations
remaining at a steady state would then become a smaller percentage of
the total hypothetical COD entering the supernatant. Thus the amount
of nutrients remaining in the supernatant would be even smaller in
comparison to the sludge than calculated earlier and thus the super-
natant biochemical activity even larger.
In an effort to further refine the measurements of swine waste settling
and biological activity, total coliform microbial counts were performed
immediately after loading on supernatant samples. Typical data over a
5-hour period are shown in Figure 19. Observation of the large popu-
lation present (1 x 10° to 10" total coliform/100 ml) makes the reasons
obvious why lagoon supernatant is not suitable for direct stream dis-
charge. Stream standards for total coliform are on the order of 1 x
10^- coliform per 100 ml. The time for bacteria population doubling is
on the order of .5 to 1 hour; and although the coliform density data
were somewhat erratic, an increase in total coliform can be seen over
the initial five-hour time span-. Qualitative conclusions were that
the microbial population maintained a steady viability throughout the
latter part of the week after the initial organic material was meta-
bolized. Possibly the mass transfer from the sludge helped maintain
this large viable population. Then after loading, a rapid response
occurred resulting in organic stabilization. Thus the full microbial
capacity of the system was probably not utilized with once per week
loading.
Occasionally when animal producers install anaerobic lagoons, designers
have recommended sludge seeding for faster start-up response. This is
also a common practice for digesters at municipal waste treatments.
However comparison of Figures 4 and 5 with Figures 12 and 13 showed
that for the reference loading rate at a once per week frequency, the
time to achieve steady state supernatant concentration for these two
sets of cones was 12-15 weeks. Yet one set of cones had seeded super-
natant and sludge while the other was begun with tap water. This
accentuates the fact that animal waste is quite concentrated and con-
tains all the microflora needed for adequate anaerobic stabilization.
These data do not, however, contradict suggestions to provide waste
dilution during lagoon start-up.
72
-------
Q
O
CJ
2,200
2,000
1,800
O
O
M
01
50
40
X! 30
CO
4J
53 §
S 8
20
10
2
6/15
i 012
6/22
TIME AFTER LOADING, Hours
Figure 19.
Mlcrobial population and supernatant COD concentration changes
immediately after loading laboratory reactors.
-------
Potential existed for significant concentration gradients with depth in
both small-scale and field-scale units and thus a single sample could
have inadequately portrayed average supernatant concentrations. The
concentration of four waste parameters at the top, middle, and bottom
of the reactor fluid are listed in Table 12. Results for the bottom
sample were somewhat higher due to hindered settling and side wall
effects at the cone apex. In general, there appeared to be no large
gradients especially in the bulk of the supernatant (the middle and
top). This uniformity of supernatant except for top algal or.sludge
mats was found in many laboratory reactors and field units and is
discussed in a later section.
Table 12. SAMPLES TAKEN AT SEVERAL DEPTHS IN IMHOFF CONE AND 14-1
LABORATORY REACTORS LOADED WITH SWINE WASTES
Reactor
1NS1
1NS3
1NS1
14 1 - 3
loads per
week
Height above
bottom, cm
42
26
1.2
42
26
1.2
42
33
26
19
14
25
17
12
7
Parameter concentration, rng/1
TOG
380
315
485
330
345
525
400
400
450
575
575
--
--
--
COD
820
930
1,383
805
883
915
__
--
--
--
--
1,635
1,600
1,770
1,920
TKN
97
92
91
87
83
114
112
125
146
174
145
380
290
360
360
o-POA-P
31
57
63
56
6>
71
54
56
-56.
57
60
For each cone the overall mass balance was calculated by comparing the
inputs versus the sum of outflow plus accumulation for four waste con-
stituents, COD, TOC, 0-P04-P, and TKN. Detailed data*for these four
components are listed in Appendices Al and A2. The input included
the raw waste feed plus initial seeding. Output was the sum of all the
samples taken plus any drained sludge. Accumulation included the
remaining supernatant, sludge and any material adhering to the cone
wa11s.
74
-------
The difference between total input and accumulation plus outflow was
termed the amount unaccounted for; which expressed as a percent of the
total input, is shown in Table 13. This material was lost from the total
system as either volatilization or conversion to gaseous end products.
For the COD and TOG this loss was primarily as carbon dioxide and methane
evolution while for TKN it would be ammonia volatilization. The total
lost from the accumulated sludge units loaded at the .reference rate was
recorded to be 55 percent - 65 percent for COD, 40 percent - 70 percent
for TOG, and 50 percent - 65 percent for TKN. This loss was the same
for the once-per-week and three-per-week loading frequencies. The cones
loaded at four times the reference rate (4AS1) has about the same COD
and TOG losses but a much lower TKN loss. Thus the sludge microbial
decomposition of organics was approximately the same at these two very
different loading rates and frequencies.
Table 13. OVERALL MASS BALANCE RESULTS FOR ANAEROBIC TREATMENT OF
SWINE WASTES - IMHOFF CONE REACTORS WITH VARIOUS SLUDGE
MANAGEMENTS .
First Experiment
1NS1
• 1NS3
1CS1
1CS3
1AS1
1AS3
Second Experiment
1NS1
4NS1
1CS1
4CS1
1AS1
4AS1
Percent of cone input which was not recovered
as output or accumulation in cone
TKN
58
47
63
53
63
56
66
35
68
28
50
22
TOG
20
31
.39
i 38 .
.42 .
44
64
55
55
58
62
71
COD
43
38
.42
47
57
55
60
36
64
38
61
64
o-PO/i-P
0
0
0
0
0
0
-
-
-
-
-
-
75
-------
14-1 Reactor Experiments
An assumption, based on the Itnhoff cone studies, was that the anaerobic
microbes were underutilized at the once-per-week input of raw waste.
A series of open-top cylindrical reactors with a tap water volume of
14 1 were initiated to evaluate biological response to different
nutrient inputs, Figure 20. The reactors were sampled with a wide-tip
pipette at mid-depth and then tap water was added to offset evaporative
and sampling losses.
Frequency of Loading Studies-
The raw swine waste loading rate in terms of grams of COD per week, was
held constant, but the frequency varied. The raw waste concentration
was also held constant at 40,000 mg COD/1 so that the volume input per
week was the same. For example, a unit loaded twice a week received
one half the weekly charge at each loading event. The range of loading
frequencies employed was once per two weeks, one per week, two per week,
and three per week. Because of a calculation-related error, these units
were actually loaded at 1.2 times the reference loading or 1.1 kg COD/
week/rn^ of reactor, which was thus considered as a nominal reference
loading rate. During the initial 26 weeks the raw waste input was
lower than 40,000 mg COD/1j therefore, no feed control was exercised
until after the 26th week when the raw waste was maintained at about
40,000 mg COD/1.
The weekly concentrations of o-PO^P, COD, TOG, and TKN for these four
reactors are presented in Figures 21-24. The cyclic behavior of these
units was a manifestation of feed strength variations. The supernatant
concentration showed no long-term, significant dependence on loading
frequency for the range used. The concentration levels of the various
units relative to each other changed many times both during start-up
and the steady operational periods. The final concentrations of COD,
TOG, TKN and o-PO^P for all loading frequencies evaluated were about
1,300, 500, 400, and 25 mg/1, respectively. Because of the 20 percent
76
-------
SAMPLING
POSITION '
SLUDGE
LAYER
/\
LOADWG
POSITION
14 LITER VOLUME
REACTOR
LIQUID
Scale I-cm = 4cm
Figure 20. Schematic of 14-1 laboratory reactor loaded with
swine wastes.
77
-------
LOADING FREQUENCY
OD
1,600 —
1,200
800
400
• ONCE PER TWO WEEKS
A ONCE PER WEEK
+ TWICE PER WEEK
X THRICE PER WEEK
12 16
TIME, fortnights
Figure 21. Supernatant COD concentration changes in 14-1 laboratory reactors loaded with
swine waste at various frequencies.
-------
—I
VO
rH
e
O
M
H
a
o
a
o
H
o
-PQ
Pi
'<
O
o
M
O
PS
o
H
O
H
700
600
500
400
300
200
100
LOADING FREQUENCY
A
t-
X
ONCE PER TWO WEEKS
ONCE PER WEEK
TWICE PER WEEK
THRICE PER WEEK
20
24
TIME, days
Figure 22. Supernatant TOC concentration changes in 14-1 laboratory reactors loaded
with swine waste at various frequencies.
-------
oo
o
360
2
O
H
2
w
o
a
o
270
is 180
w
o
H
M
53
d
1 90
1-4
o
H
T.OADING FREQUENCY
• ONCE PER TWO WEEKS
A ONCE PER WEEK
+ TWICE PER WEEK
X THRICE PER WEEK
12
16
20
24
28
Figure 23.
TIME, fortnights
Supernatant TKN concentration changes in 14-1 laboratory reactors loaded with swine
waste at various frequencies.
-------
100
oo
w
H
W
U
z
o
U
CM
I
-d-
o
PL.
o
W
H
PL.
cn
O
§
80
LOADING FREQUENCY
O ONCE PER TWO WEEKS
A ONCE PER WEEK
+ TWICE PER WEEK
X THRICE PER WEEK
60 —
40
20
Figure 24,
TIME, fortnights
Supernatant o-PO^-P concentration changes in 14-1 laboratory reactors loaded
with swine waste at various frequencies.
-------
higher reactor loading rate, the levels of oxygen demand and organic
carbon were 10-20 percent higher than cones loaded at comparable rates
thus showing the effluent concentration dependence on loading rate.
However, the TKN level in the cones was only about 40 percent of the
TKN steady-state concentration of 400 mg/1 for these 14-1 reactors.
The final orthophosphate (o-PC^-P) level of these 14-1 reactors was
the same as for comparable cones, 30-40 mg/1.
After 56 weeks of operation, these four 14-1 reactors were stopped and
the volume and concentration of both the supernatant and the sludge
determined. The amount of COD, TOG, TKN, and o-P04-P remaining in the
sludge as a percent of the raw waste input are shown in Table 14 and
Appendix A3. The 20 percent to 30 percent COD and 20-25 percent TOC
remaining compared favorably with the accumulated sludge cone data,
Tables 8 and 9. The TKN remaining was about 10 percent less than from
similarly loaded Imhoff cones. The 14-1 reactors were operated for 56
weeks as compared to 23 weeks for the second cone experiment and 19
weeks for the first cone experiment. Because the percentages of the
feed which remained in the sludge was nearly the same over these three
experiment durations, there appears to be a steady-state decay level
for accumulated bottom sludge. The corresponding percentage of swine
waste input COD, TOC and TKN which remained in the sludge for the
Imhoff cones and 14-1 reactors was 20 percent to 35 percent and 10
percent to 30 percent, respectively.
Table 14. AMOUNT OF WASTE PARAMETERS REMAINING IN SLUDGE ZONE OF
ANAEROBIC 14-LITER LABORATORY REACTORS LOADED AT
DIFFERENT FREQUENCIES
Loading rate -
1. 15 kg COD/week/m3
Loading frequency
One time per two weeks
One time per week
Twice per week
Three times per week
Amount remaining
a percent of the
TOC
25
21
25
21
COD
29
27
30
21
in sludge as
feed
TKN
12
15
13
13
The sludge depth in these 14-1 reactors was about 2.8 cm when the experi-
ment was completed. In addition to sludge determinations, the bottom
6.4 cm of sludge and liquid was sampled and the corresponding analysis
82
-------
then included the sludge, interfacial region, and bottom supernatant.
The comparison of the bottom 6.4 cm with the sludge zone (2.8 cm),
Table 15, indicated that the supernatant added very little to the sludge
portion. This observation further strengthens earlier estimates that
sludge material drained from the Imhoff cones while larger in volume
than the actual sludge was close to the correct sludge amount of the
various pollutional parameters.
Table 15. COMPARISON OF THE AMOUNTS OF VARIOUS POLLUTIONAL PARAMETERS
IN THE SLUDGE ZONE (BOTTOM 2.8 cm) AND THE SLUDGE PLUS LOWER
SUPERNATANT ZONE (BOTTOM 6.4 cu)
Reactor
14-1 laboratory
reactor 1 load
per 2 weeks
14- 1 laboratory
reactor 1 load
per week
14-1 laboratory
reactor 2 loads
per week
14-1 laboratory
reactor 3 loads
per week
Sludge zone
Parameter amount, &
COD
249
229
256
176
TOC
72
61
58
60
TKN
7.3
8.8
7.8
7.5
Sludge plus lower
supernatant zone
Parameter amount, g
COD
252
232
259
176
TOC
73
62
72
61
TKN
8.2
9.5
8.4
8.3
The accumulated sludge volumes from the cone and cylindrical reactor
experiments expressed as milliliters per week of experiment are given in
Table 16. This volume as a percent of the raw waste was about 10 percent
25 percent as also shown in Table 16. At the reference loading rate,
about 2.3 nP or 2250 1 per 45-kg hog were provided in the lagoon. The
raw waste strength used in this study corresponded to a waste output
of 7.5 l/d/45-kg hog. Multiplying the percent of the feed which became
sludge times 7.5 I/day gave a sludge build up of .75 to 1.9 l/d/45-kg
hog (10 percent - 25 percent accumulation). The available lagoon volume
per 45-kg hog divided by this laboratory reactor buildup rate indicated
that from 1,000 - 3, 000 days would be needed to fill the lagoon with
sludge (2.7 - 8.2 years). Experience has indicated that a much slower
83
-------
buildup occurs under field conditions. Thus caution should be exercised
in transferring sludge accumulation data to actual field lagoons. Factors
such as compaction under greater liquid head, soil incorporation, and
long-term, mixed culture, biochemical stabilization could reduce the field
lagoon sludge buildup below that measured in the lab.
Table 16. SLUDGE ACCUMULATION FOR SWINE WASTE INPUT TO ANAEROBIC
LABORATORY REACTORS
Reactor
First Imhoff
Cone experiment
1AS1
1AS3
Second Imhoff
Cone experiment
1AS1
4AS1
14-1 reactors
Once per two weeks
Once per week
Twice per week
Thrice per week
Sludge accumulation,
ml/week
6.5
6.5
6.4
16
45
41
41
39
Total sludge volume
as a percentage of
input volume
26
26
26
16
11
10
10
9
The overall mass balance for COD, TOC, TKN and o-PO^-P calculated as
described earlier with the Imhoff cones are given in Table 17 and Appendix
Al and Appendix A2. The amount of COD, TOC, and TKN lost or not recovered
from the 14-1 reactors was greater than that determined for the Imhoff
cone experiments and no agreement between the cones and reactors is
discernible concerning the amount of o-PO^-P lost. At this time no
84
-------
reason, except the possible contribution of analytical error or reactor
configuration differences can be postulated for these differences,
especially for phosphorus.
Table 17. OVERALL MASS BALANCE FOR 14 1 ANAEROBIC LABORATORY REACTORS
LOADED WITH RAW SWINE WASTE AT VARIOUS FREQUENCIES
Loading Rate: 1.1 kg COD/week
per nr of reactor
volume
Loading Frequency
Once per two weeks
Once per week
Twice per week
Thrice per week
Percent of reactor input which
was not recovered as output or
accumulation
Parameter
TKN
77
76
77
76
TOC
71
74
71
74
COD
67
72
67
76
o-P04-P
71
56
71
70
To put the loading frequencies of these four units in perspective, the
once per two weeks was not infrequent enough to present a shock load to
the reactor. Shock loads are characteristic of manure pit management
in which the lagoon receives a large charge two or taree times per
year. Such shock inputs can cause imbalance or overloading which can
lead to lagoon failure. 9 Such infrequent loadings were not investi-
gated in this study.
The most frequent loading would be continuous input; hence another
comparative study was undertaken to expand the frequency studies. Two
reactors (14-1, open top) were established with tap water and provisions
for quasi-continuous loading. The loading rates were a) four times the
nominal reference load or 4.8 times the reference rate and b) the nominal
loading rate. The raw waste input was refrigerated and inputs were once
per 4 hours at 10 ml/pumping and once per hour at 10 ml/pumping for the
1.2 and 4.8 times units, respectively. These rates corresponded to 1.1
and 4.4 kg COD/day/m of volume. Two other identical reactors were run
at these same loading rates, but with a once per week loading frequency.
The supernatant values for these reactors are plotted for the duration
of the experiment in Figures 25-30.
The heavier loaded reactors evidenced no significant difference in super-
natant quality (COD, TOC, TKN, and O-P04-P) between quasi-continuous and
85
-------
00
B
S3
o
§
H
53
W
U
Z
O
o
p
o
o
o
25
W
O
s
o
1,400
1,200
1,000
800
600
400
200
LOADING FREQUENCY
O ONCE PER WEEK
& CONTINUOUS
280
Figure 25. Supernatant COD concentration changes in laboratory 14-1 reactors with
the nominal reference loading rate of swine waste on a continuous or
batch basis.
-------
00
60
a
O
M
H
H
a
S3
O
u
O
O
H
O
«
Pi
-------
00
e
00
00
2
O
w
u
2
O
u
w
H
PL,
CO
O
sc
a
w
o
O
PS
H
1
.J
w
H
O
H
350 —
300 —
250
200 —
150
100
50
LOADING FREQUENCY
O ONCE PER WEEK - TKN
O CONTINUOUS - TKN
ONCE PER WEEK - o-PO^.-P
CONTINUOUS - o-POA-P '
40
Figure 27.
120
TIME, days
Supernatant TKN and o-PO^-P concentration changes in laboratory 14-1 reactors
with the nominal reference loading rate of swine waste on a continuous or
batch basis.
-------
23
O
M
H
23
W
O
M
e>
etf
o
H
O
H
W
Q
W
0
Pd
o
8,000
6,000 —
4,000 —
2,000
LOADING FREQUENCY
• COD - ONCE PER WEEK
A COD - CONTINUOUS
O TOC - ONCE PER WEEK
O TOC - CONTINUOUS
120 144
TIME, days
168
192
216
Figure 28.
Supernatant COD and TOC concentration changes in laboratory 14-1 reactors
with four times the nominal reference loading rate of swine waste on a
continuous or batch basis.
-------
60
&
§ 2,000
H
U
S 1,500
H
O
8 1,000
M
25
Q
,-5
500 —
H
O
H
LOADING FREQUENCY
• ONCE PER WEEK
O CONTINUOUS
120
144
168
192
216
240
Figure 29.
TIME, days
Supernatant TKN concentration changes in laboratory 14-1 reactors with four times
the nominal reference loading rate of swine waste on a continuous or batch basis.
-------
LOADING FREQUENCY
ONCE PER WEEK
CONTINUOUS
vo
H
Pd
O
2;
o
o
PM
I
O
PM
H
H
PM
c/D
§
P-J
o
Pd
160 ,—
r 140
120
100
80 —
60 —
40 —,
20
216
240
Figure 30. Supernatant o-PO^-P concentration changes in laboratory 14-1 reactors with four times
the nominal reference loading rate of swine waste on a continuous or batch basis.
-------
weekly batch loading. In addition the response time to achieve a
steady supernatant concentration was the same for both reactors. However,
at the nominal reference loading rate the quasi-continuous unit had
consistently lower supernatant concentrations than the weekly batch
reactor except for ortho-phosphates which were the same, Figures 25-27.
In terms of biological activity these factors indicated that at the heavier
rate there were sufficient nutrients added so that even with batch type
loading the microorganisms could maintain a high viability between loading
events. Thus when the next loading event occurred the viable population
was sufficient to stabilize or reduce the organic matter to the same
level as for the continuous loaded unit. However, at the lower loading
rate the nutrient level was not sufficient to maintain the population in
the batch reactor required to stabilize organic matter to the same level
as with continuous reactor at the same loading rate. Experiments are
underway to see if this trend is more pronounced at lower nutrient
levels such as one-fourth of the nominal reference loading.
Tentative conclusions were that in a producer situation there are ad-
vantages in loading anaerobic lagoons on a continuous or nearly continu-
ous manner. This conclusion is corroborated by data from both the
laboratory and field studies. The advantages of better stabilization
and improved supernatant quality may be greater if the producer is
operating his lagoon with 2.3 nP of volume per 45-kg hog or more. This
type of loading could be accomplished by having a manure pit overflow,
a frequent flush system, or daily scraping or cleaning.
Loading Rate Studies-
Having determined the relative performance of 14 1 laboratory reactors
loaded at 4.8 and 1.2 times the reference loading rate (2.3 m3 per 45-
kg hog) additional 14 1 reactors at 0.6 and 0.3 tin.es the reference
rate were begun. These rates corresponded to 3.9 and 7.8 m3 of reactor
volume per 45 kg hog. The reactors were loaded once per week and all
operational and sampling procedures were the same as the previously
described laboratory units. After a period of five to six weeks the
TOC and COD supernatant concentrations reached the steady-state level
and these data are shown in Figures 31 and 34. The total Kjeldahl
nitrogen achieved a steady-state supernatant concentration in eight to
nine weeks for both units, Figures 32 and 35. Supernatant phosphate
concentrations are shown in Figures 33 and 36. The reduced loading
rate produced effluent with a lower concentration of all parameters
measured. This trend of effluent with a lower concentration being
inversely dependent on loading rate was also shown with the units
loaded at 4.8 and 1.2 times the reference rate.
To check sampling errors attributable to gradients which may have
existed, liquid was removed from several vertical locations in the
92
-------
50
6
53
o
H
1
w
o
is
o
CJ
O
PQ
o
Pi
o
o
H
W
n
55
w
CD
o
M
O
1,000 —
A COD
O TOG
800
600
400
200
24
Figure 31.
48
72
96
120
TIME, days
144
158
192
216
Supernatant COD and TOC concentration changes in.laboratory 14-1 reactors with
0.6 times the reference loading rate of swine waste and once per week frequency.
-------
60
320
a
o
H
z;
w
240
o
o
H
M
2
I
O
w
H
O
H
160
80
32
64
96
128
160
192
224
TIME, days
Figure 32. Supernatant TKN concentration changes in laboratory 14-1 reactors with
0.6 times the reference loading rate of swine waste and once per week
frequency.
-------
00
Ul
a
o
M
H
H
O
/ s
PH
o
p-l
w
H
P-t
CO
§
PM
O
tfi
60
40
20
32
Figure 33.
64 96 128
TIME, days
160
192
224
Supernatant o-P04~P concentration changes in laboratory 14-1 reactors
with 0.6 times the reference loading rate of swine waste and once cer
week frequency.
-------
vD
A COD
O TOC
216
240
sa
u
TIME, days
Figure 34. Supernatant COD and TOC concentration changes in laboratory 14-1 reactors with
0.3 times the reference loading rate of swine waste and once per week frequency,
-------
60
a
S3
o
160
120
O
o
o
§
H
w
2
3
O
80
40
24
48
72
96 120
TIME, days
144
168
192
216
Figure 35. Supernatant TKN concentration changes in laboratory 14-1 reactors with 0.3
times the reference loading rate of swine waste and once per week frequency.
-------
VO
00
W)
6
55
O
w
o
o
CJ
CM
I
-------
14~1 reactor. These concentrations are given in Table 12. As with the
Imhoff cones there was a good concentration uniformity with the exception
of the sludge blanket region, Researchers describing swine waste settling^
and dairy lagoon performance''41 have observed an interfacial zone of
flocculated type material. This zone was not readily detected with
employed physical sludge measurements but was probably the reason for
the higher chemical concentrations near the sludge. However, the super-
natant zone was uniform; hence a mid-depth sample was representative.
An extremely heavy loading rate was selected to complete the upper loading
range for the 14 1 laboratory reactors. This rate was 10.8 times the
reference loading which is equivalent to a volume of 0.21 m~> (210 liter)
per 45-kg hog for a waste strength of 40,000 rng/1 COD characteristic for
a waste volume of 7.5 1 per 45-kg hog per day. Thus the residence time
was approximately 28 days. The supernatant concentrations of TOC, TKN,
and o-PO^-P, Figures 37 and 38, were approximately equal to the raw waste
values. These levels indicated that there was only minimal settling
and biological activity and that the anaerobic system served primarily
as a storage vessel. As discussed in a later section, a unit with such
a heavy loading rate would be expected to reach the approximate feed
concentration in about 12-16 weeks as verified by Figures 37 and 38
showing that in fact steady conditions were reached in about 11 or 12
weeks. The raw waste-like characteristics of tne supernatant were also
verified by visual observations and the high analytical variability,
resulting in part due to sampling difficulties with these thick type
slurries.
After initial data were taken on this reactor, a sealed top was installed
with a tube leading to a gas collection device. This unit then simulated
a methane producing digester at constant laboratory temperature (25° C).
Steady-state gas generation was 3.6 I/day with a composition of 50 percent
CHA and 40 percent COp. Gas evolution represented 22 percent of the input
carbon based upon mass balances for carbon content of collected gas
and TOC of waste input (55.6 g/wk).
Efficiency of a lagoon could be calculated in several ways depending
upon the assumptions used. If one assumes that sludge buildup is very
slow and thus will 'not have to be removed, then lagoon efficiency is
related to the ratio of effluent concentration or supernatant quality
to the influent quality. This efficiency can be expressed for each
chemical parameter. The assumption of slow sludge buildup is based on
general experience:and appeared to be reasonable for swine. For example,
the sludge depth of about 0.6 m in one North Carolina State University
swine lagoon operating near the reference loading rate for 13 years
has been judged to represent a very slow buildup rate.
The lagoon efficiency for the four loading rates in the laboratory
reactors are thus calculated for each parameter as follows:
99
-------
o
o
o
M
H
w
u
o
u
u
o
H
z;
o
H
O
H
4,000 —
40
80 120
TIME, days
160
200
240
Figure 37. Supernatant TOC concentration changes in laboratory 14-1 reactors with
10.8 times the reference rate of swine waste.
-------
60
e
a
o
H
S3
W
u
CO
g
o
33
P*
t/3
O
33
PM
O
PC
o
o
O
§
H
M
S3
1
a
o
H
O TKN
• o-PO^-P
I 4,000 —
3,000 _
2,000 —
1,000 —
40
120
TIME, days
160
200
240
Figure 38. Supernatant TKN and o-PO^-P concentration changes in laboratory 14-1 reactors
with 10.8 times the reference rate of swine waste.
-------
Table 18.
REMOVAL EFFICIENCIES (EQUATION 2) OF VARIOUS POLLUTIONAL
PARAMETERS FROM 14-1 LABORATORY REACTOR LOADED WITH
SWINE WASTES
Input to all reactors:
COD
TOC
TKN
o-P04-P
40,000 mg/1
15,000 mg/1
2,500 mg/1
800 mg/1
14'1 reactors
Loading rate
(fraction or
multiple of
reference rate)
.25
.5
1
1
4
4,
Loading
Frequency
once per
week
once per
week
once per
week
continuous
continuous
once per
week
Parameter
COD
TOC
TKN
o-P04-P
COD
TOC
TKN
o-P04-P
COD
TOC
TKN
o-P04-P
COD
TOC
TKN
o-P04-P
COD
TOC
TKN
o-P04-P
COD
TOC
TKN
o-P04-P
Effluent,
mg
350
160
100
35
500
275
200
30
1,200
500
400
30
1,000
400
275
25
4,500
2,200
1,500
90
4,000
2,000
1,700
110
Removal
efficiency, 70
99
99
96
96
99
98
92
96
97
97
84
96
98
97
89
97
88
85
40
89
90
87
32
86
102
-------
Table 19. REMOVAL EFFICIENCIES OF VARIOUS POLLUTIONAL PARAMETERS
FROM IMHOFF CONE LABORATORY REACTORS LOADED WITH SWINE
WASTE
Input to all reactors:
COD
TOG
TKN
o-P04-P
40,000 mg/1
15,000 mg/1
2,500 mg/1
800 mg/1
Imhoff Cones
Loading rate
(multiple of
reference rate)
1
1
1
4
Loading
frequency
once per
week
thrice per
week
once per
week
once per
week
Parameter
COD
TOG
TKN
o-P04-P
COD
TOG
TKN
o-P04-P
COD
TOG
TKN
COD
TOG
TKN
Effluent,
mg/1
1,250
500
150
60
1,400
500
150
60
1,000
450
200
2,300
900
900
Remova 1
Efficiency, %
97
97
94
90
97
97
94
90
98
97
92
94
94
64
efficiency = influent: concentration - effluent concentration (2)
influent concentration
Results for the 14-1 reactors are shown in Table 18 and for the Imhoff
cones in Table 19.
103
-------
It was quite evident that when calculated in this manner, the removal
efficiencies for all parameters were quite high with the highest effici-
encies being for the organic carbon species. The percentage of TOC and
COD in the effluent increased at the higher loading rates with the
removals being only 85 - 95 percent at the reference loading rate or
lower. However, at the four times reference rate, the nitrogen removal
was much lower, 35 - 65 percent.
Summarizing removal efficiencies based on laboratory experiments, it was
clear that, at the reference loading rate or lower, the percentage of
input COD, TOC, TKN, and o-PO^P in the effluent was very low. Corres-
pondly, 90 percent and often more than 95 percent removals were obtained
in these anaerobic systems. At higher loading rates, the phosphate and
nitrogen removal became considerably lower. However, it should be noted
that even with about 95 percent removals in the laboratory tests, the
effluent levels of COD, TOC, TKN, and o-P04-P were 1,500, 500, 400, and
40 mg/1, respectively. In comparison to discharge criteria and even
certain municipal and industrial raw waste streams, this effluent was
very strong. This emphasized the basis for regulations specifying no
discharge of effluents from anaerobic swine lagoons.
104
-------
FIELD PILOT-SCALE EXPERIMENTS
Unaerated Anaerobic Reactors
Field pilot-scale experimentation was initiated to allow verification
of laboratory results and screening of variables under actual field
conditions without the large costs associated with full-scale studies.
Field pilot-scale lagoons were constructed with 2 mm thick steel in a
cylindrical shape 3.5 meters in diameter and 2.5 meters high. Overflow
pipes with external gate valves were set at the 1.85 m height providing
a volume of 17.5 cubic meters. The inside and outside of all the pilot
lagoons were painted with primer and an epoxy paint. After two years
of operation, paint peeled in some small areas where the sides were
alternately exposed to air and anaerobic lagoon supernatant but no
peeling has been found in continuously submerged regions. It was
determined that there was little or no seepage from the reactors. The
schematic and photographic representations of the field site are shown
in Figures 39 and 40.
The field procedure involved once-per-week loading of raw swine waste
from a manure pit under a partially slatted floor described in detail
in the lagoon studies field pilot-scale experiments section. The raw
waste was mixed in a 7,500-1 holding tank by pump recirculation until
homogenous and, if necessary, diluted with water to a 40,000 mg COD/1
level. During the weekly reaction period, the overflow gate valve on
each lagoon remained closed. Prior to loading, the volume of material
above the overflow level was measured and samples of the supernatant
taken at the ,93-m depth with the APHA-type sampler. If the lagoon
level was below the overflow pipe level, water was added to restore
unit standard volume. These effluent or water addition volume measure-
ments, along with on-farm precipitation and temperature records,
allowed mass balance evaluation of various inputs and outflows from
these pilot lagoons. After supernatant sampling, each gate valve
was opened for overflow draining and then closed.
Next each lagoon received the waste input which was a constant volume
since the input concentration was held relatively constant. This pro-
cedure was repeated once each week.
The reliability of complete tank mixing before and during loading and
effectiveness of loading by hose discharge and manure loading tanks were
determined. Results for the COD analysis, shown in Table 20, indicated
that the agitation and loading system was adequate even though some
unexplained slightly higher concentrations would periodically occur.
105
-------
FROM MANURE
I STORAGE PITS
STEVENS RECORDER
AERATION FOR TREATMENT OF
DISSOLVED SOLIDS (FACULATIVE
UNIT). THREE UNIT SERIES FOR
USE IN VARIOUS TREATMENT
STRATEGES.
o
Figure 39.
FLUME
PROPORTIONAL DIVIDER
WATER
RESERVOIR
ELECTRIC
PANEL
MODEL LAGOONS
AT INDICATED
REFERENCE LOADING
AERATED UNIT
FOR ODOR CONTROL
EFFLUENT
COLLECTION
TANKS
MANURE
LOADING VESSEL
-—IRRIGATION PUMP
TO
LAND IRRIGATION
MANURE
LOADING
VESSEL
Schematic of pilot-scale lagoon
research site.
-------
Figure 40. Photograph of pilot-scale lagoon research site,
107
-------
Table 20. UNIFORMITY OF RAW SWINE WASTE DURING AGITATION AND LOADING
OF FIELD PILOT-SCALE LAGOONS
GOD concentration,
Sample description
43,100
41,100
36,700
39,400
48,900
45,600
hose discharge recycling to mixing tank
hose discharge to first unit
loading tank discharge to heaviest loaded
reactor, start of loading
loading tank discharge to heaviest loaded
reactor, end of loading
hose discharge to next to the last pilot
lagoon
hose discharge to last pilot lagoons
Loading of the field pilot-scale or model lagoons began on July 11,
1972, with swine wastes stored in nearby manure pits and fresh manure
was used since October 31, 1972. Raw waste was continuously character-
ized and loading rates adjusted as necessary to conform with the desired
pre-experiment rates. The loads to each reactor were based on multiples
or fractions of the reference rate (Table 4) of 2.3 m^ of lagoon volume
per 45-kg hog, Table 21. Additionally, two pilot-scale tanks were
established to give series lagoon treatment to the effluent from the
unit receiving the reference loading rate. These were referred to as
the second lagoon in series (1A) and the third lagoon in series (IB).
For the period January 16, 1973, to April 1, 1973, the waste material
was fairly constant at 40,000 mg COD/1, the expected value for a total
daily waste volume of 7.5 1 from a 45-kg hog.
However, during the spring and summer, the increased use of malfunctioning
foggers and waterers prevented the maintenance of a high concentration
waste. This period was characterized by a waste in the range of 10,000 -
25,000 mg COD/1. The small lagoon freeboard available prohibited loading
at a constant magnitude of COD per week when the waste input was so
dilute due to excessive water wastage. Thus, a constant feed volume was
added but it was impossible to maintain a uniform waste concentration
during this period. Finally in the fall of 1973, a source of concen-
trated waste was secured and, as needed, diluted to approximately 40,000
mg COD/1.
108
-------
Table 21. SWINE WASTE LOADING RATE OF FIELD PILOT-SCALE UNITS
Multiple or
fraction of
reference rate
4
1
1A
IB
1/2
1/4
1/8
1/16
1/32
Swine waste
volume added
per week, 1
840
420
overflow from 1
overflow from 1A
210
105
52
26
13
kg COD/week/m3
(40,000 mg COD/1)
3.64
0.31
--
--
0.45
0.23
0.12
0.055
0.028
o
m lagoon volume
per 45-kg hog
0.6
2.3
--
--
4.6
9.2
18.4
36.8
73.6
Lagoon Supernatant Concentrations
The supernatant concentrations for the model units are plotted in Figures
41-58. These data are averages of the supernatant concentration over
successive two-week intervals after attainment of the more concentrated
waste (October, 1973). Hence, the lower initial concentration reflected
the antecedent period of dilute waste input.
After the attainment of the more uniform concentrated raw waste (October,
1973) all of the pilot-scale lagoons attained a steady-state concentration
during the late winter and early spring period. The judgement of the
steady-state value was difficult because lagoon temperature changes with
climate cycles affected reaction dynamics and thus supernatant concentra-
tion. The steady-state or uniform concentration level was more
109
-------
O COD
D TOC
JAN FEE
TIME, months
MAR
APR
MAY
Figure 41.
Supernatant COD and TOC concentration changes (two-week averages) in field pilot-scale
lagoons loaded once per week at four times the reference rate for swine wastes.
-------
(JO
„ 3,600 i—
JS
O
H
a
w
w
3,200 —
2,800
2,400
2,000
H
0 1,600
Q
§1,200
o
o
pi
tj 800
33
<
Q
iJ
W
H
O
H
400
O TKN
D 0-P04-P
OCT
Figure 42.
NOV
DEC JAN FEE
TIME, months
MAR
APR
MAY
Supernatant TKN and o-PO^-P concentration (two-week averages) in field pilot-scale
lagoons loaded once per week at four times the reference rate for swine wastes.
-------
GO
e
o
M
S 3600
OS
H
KJ
W
•s 3200
o
§ 2800
w 2400
^*.*
o
«
0 2000
H
O
H 1600
o
g 1200
Q
800
0 400
w
o
O COD
D TOC
PC
o
OCT
NOV
DEC
JAN FEE
TIME, months
MAR
APR
MAY
Figure 43. Supernatant COD and TOC concentration changes (two-week averages) in field pilot-
scale lagoons loaded once per week at the reference rate for swine wastes.
-------
25
o
M
H
H
is
W
O
IS
o
w
C/)
§
O
EC
H
-------
CxO
e
o
M
H
§1,200
u
^
o
o
is
o
M
CJ
CJ
O
ft
o
O
H
Q
W
O
PC
u
900
600
300
O COD
D TOC
OCT
NOV
DEC
JAN, 74 FEE
TIME, months
MAR
APR
MAY
Figure 45. Supernatant COD and TOC concentration changes (two-week averages) in
field pilot-scale lagoon receiving effluent from lagoon loaded once
per week at reference rate for swine wastes.
-------
400
o
M
H
H
S3
W
u
z
o
u
300
Ul
w
o
§
H
M
200
P
i-i
Pd
100
H
O
H
OCT
NOV
DEC
JAN, 74
FEE
MAR
APR
MAY
TIME, months
Figure 46. Supernatant TKN concentration changes (two-week averages) in field pilot-scale
lagoon receiving effluent from lagoon loaded once per week at reference rate
for swine wastes.
-------
800 r-
60
6 600
o
M
H
H
S
z
o
u
u
o
§
U
400
200
O COD
D TOC
OCT
Figure 47.
NOV
DEC
JAN FEE
TIME, months
MAR
APR
MAY
Supernatant COD and TOC concentration changes (two-week averages) in third lagoon
of three-unit series with the first lagoon of this field pilot-scale series
loaded once per week at the reference rate for swine waste.
-------
120
60
e
z.
o
W
O
a
o
s
w
o
o
PS
90
60
30
w
o
H
OCT
Figure 48.
NOV
DEC
JAN, 74
FEB
MAR
APR
MAY
TIME, months
Supernatant TKN concentration changes (two-week averages) in third lagoon of three-
unit series with the first lagoon of this field pilot-scale series loaded once per
week at the reference rate for swine waste.
-------
2,000
oo
e
Si, 600
H
H
a
w
o
^1,200
O
CQ
o
O
&
O
O
H
800
400
O COD
D TOG
§
CD
B
O
3
u
OCT
Figure 49.
NOV
DEC
JAN
FEE
MAR
APR
MAY
TIME, months
Supernatant TKN and o-P04~P concentration changes (two-
week averages) in field pilot-scale lagoons loaded once
per week at 0.5 times the reference rate for swine wastes.
re
o
-------
O
M
H
W
CJ
23
O
o
w
H
cn
O
FM
O
PC
H
PS
O
800 -
O TKN
D o-P04-P
600
400
la
w
o
200
I
H
O
H
OCT
Figure 50.
NOV
DEC
JAN FEE
TIME, months
MAR
AP.R
MAY
Supernatant TKN and o-PO^-P concentration changes (two week averages) in field pilot
scale lagoons loaded once per week at 0.5 times the reference rate for swine wastes.
-------
,2,000
60
E
S3
O
Hi, 600
a
w
O
CJ
o
M
< 1,200
o
M
O
800
H
O
H
400
Q
2!
W
O
o
u
M
s
W
W
u
O COD
D TOC
OCX
Figure 51.
NOV
DEC
JAN
FEE
MAR
APR
MAY
TIME, months
Supernatant COD and TOC concentration changes (two week averages) in field pilot
scale lagoons loaded once per week at 0.25 times the reference rate for swine wastes.
-------
Q
W
Q
>J
W
H
O
H
400
O
a
o
w
^ 300
PU
o
PH
O
ffi
H
200
100
O TKN
D o-P04-P
OCT
Figure 52.
NOV
DEC
JAN
FEE
MAR
APR
MAY
Supernatant TKN and o-PO.-P concentration changes (two week averages) in field
pilot scale lagoons loaded once per week at 0.25 times the reference rate for
swine wastes.
-------
60
6
a
o
-
w
o
S3
o
u
o
PQ
u
o
H
O
O
H
o
H
is
w
o
800
O COD
D TOC
600
400
200
CJ
EC
O
OCT
Figure 53.
NOV
DEC
JAN FEE
TIME, months
MAR
APR
MAY
Supernatant COD and TOC concentration changes (two week averages) in field pilot scale
lagoons loaded once per week at 0.125 times the reference rate for swine wastes.
-------
oo
e
§
M
i
W
U
a
O
CJ
w
a
j
Q
W
O
H
160.
120
Pd
CO
O
nd
CM
O
Pd
I 80;
w
a
§
40 —
O TKN
a o-po4-p
-G-
-o
OCT
Figure 54,
NOV
DEC
JAN
FEE
MAR
APR
MAY
Supernatant TKN and o-PO,-P concentration changes (two week averages) in field pilot
scale lagoons loaded once per week at 0.125 times the reference rate for swine wastes,
-------
60
B
~T 800
^-j
o
n
H
H
53
W
C_>
g 600
a
o
o
o
400
O
Pd
O
H
O
H
200
O COD
D TOC
OCT
NOV
DEC
JAN FEE
MAR
APR
MAY
TIME, months
Figure 55. Supernatant COD and TOC concentration changes (two week averages) in
field pilot scale lagoons loaded once per week at 0.0625 times the
reference rate for swine wastes.
-------
o8G
M
H
W
U
a
o
u
w
H
60
CM
C/}
O
§40
H
W
| 20
H
M
E3
,-J
Q
W
O TKN
n o-po4-p
OCT
NOV
DEC
JAN
FEE
MAR
APR
MAY
H
O
H
Figure 56. Supernatant TKN and o-PO,-P concentration changes (two week averages) in field
pilot scale lagoons loaded once per week at 0.0625 times the reference rate for swine
wastes.
-------
oo
S
2
O
w
u
o
o
o
o
erf
o
H
O
H
P
W
Q
W
O
400
O COD
D TOC
300
200
100
o
OCT
NOV
DEC
JAN
FEE
MAR
APR
MAY
33
O
Figure 57. Supernatant COD and TOC concentration changes (two week averages) in field pilot-
scale lagoons loaded once per week at 0.031 times the reference rate for swine waste.
-------
40
O
D
TKN
o-P04-P
30
20
10
OCX
NOV
DEC
JAN
FEE
MAR
APR
MAY
TIME, months
Figure 58. Supernatant TKN and o-P04~P concentration changes (two week averages) in field
pilot scale lagoons loaded once per week at 0.031 times the reference rate for
swine wastes.
-------
pronounced in certain units, but in general by late April and early
May, the four parameters monitored were close to a steady-state value.
Shortly after this time, supernatant concentration values decreased
with the warmer summer temperatures. Thus steady-state supernatant
values derived from Figures 41-58 were for the winter-early spring
period.
The temperature of the lagoon supernatant 1 meter below the surface was
found to equal weekly average air temperature determined by averaging
the daily maximum and minimum temperatures throughout a week period,
Appendix Bl. This temperature agreement continued throughout the annual
cycle.
Temperature readings at the top (.15 m below surface) middle and
bottom (within sludge blanket if present) were made on several occasions
during the year, Appendix Bl. For a given pilot-scale reactor tempera-
ture uniformity was within one to two degrees Celsius throughout the
1.8 m depth. The more heavily loaded unit evidenced a consistent
one to three degree Celsuis higher temperature than the more lightly
loaded units, particularly at the bottom. Heat of anaerobic microbial
reactions or better solar heat entrapment may have been factors in
the higher reactor temperature.
Afcer the late winter to early spring period the supernatant concen-
trations began to decrease as the lagoon temperature began to increase.
The responses of the various lagoons to the seasonal temperature change
were quite different in that the time and rate of concentration de-
crease was not uniform or did not consistently seem to depend on
1 jading rate. The explanation of seasonal lagoon supernatant changes is
not presently available and these lagoons are being continued through an
eatire annual cycle.
Response and achievement of steady-state times were generally longest
f:vr TKN. COD was also less rapid than TOG in concentration response
and ivas generally more variable as an indicator of supernatant quality
wtille minimal response for ortho phosphorus was noted. These trends
follow those noted in the laboratory units.
The steady supernatant levels of COD, TOC, TKN,and o-P04~P are listed
in Table 22 for all loading rates investigated. The supernatant con-
c-itration of COD, TOC, TKN, and O-P04-P were found to decrease in
value as the loading rate (kg COD/m /week) was decreased or as the
lagcon volume (m3/45-kg hog) increased. This trend x^as also similar
t? that observed with the laboratory units. However the steady-state
concentration of the pilot-scale lagoons was considerably higher than
iaDoratory units loaded at a comparable rate. Experimental reactor
size and environmental factors could contribute to these differences.
128
-------
Table 22. STEADY SUPERNATANT CONCENTRATION OF FIELD PILOT-SCALE LAGOONS
LOADED AT VARIOUS RATES WITH SWINE WASTES
Loading rate
(fraction or multiple
of reference rate)
4
1 (first of series)
LA (second of series)
IB (third of series)
.5
.25
.125
.0625
.0312
Concentration, mg/1
COD
18,000
3,000
1,000
500
2,100
1,600
900
500
250
TOG
7,000
1,200
400
250
700
500
300
190
90
TKN
2,700
1,000
300
90
600
350
150
70
30
0-P04-P
450
90
__
—
75
50
50
25
15
The lagoon supernatant concentration for field pilot-scale lagoons was
found to be homogeneous for the reference rate unit, Figure 59, and
also for those units loaded at fractions of the reference rate. A
sample taken at an intermediate depth thus was representative of lagoon
supernatant, as also verified earlier for laboratory scale units.
The three lagoon units in series were originally connected directly by
overflow pipes, Figure 39. In a recent EPA report it was emphasized
that short circuiting commonly occurred in lagoons when the inflow
and outflow ports were not properly situated. With these small units,
it was observed that during loading some of the input flowed directly
into the next tank. Thus the connecting pipe between the first two
units was sealed and excess supernatant was pumped with subsurface dis-
charge each week from the first to the second unit. However, free flow
between second and third lagoons was maintained. This more effectively
compartmentalized the series system. However, supernatant quality data
for the three lagoon system showed no noticeable effect due to manage-
ment loading change. Thus, for this size unit no conclusive infor-
mation regarding short circuiting was obtained. Additional supernatant
concentration reductions did occur in the field pilot-scale series
129
-------
u>
o
L, I/
£ 1.86
0)
e
1.55
S
o
H
H
§ 1.24
o
0
< 0.93
w
§ 0.62
w
u
5 0.31
H
C/5
0
- 0 O TKN
-
--1
—
3i
<
n TOC
>
If 8
1 1 1 1 1 I
0 3,000 6,000 9,000 12,000 15,000 18,000
SUPERNATANT TOTAL ORGANIC CARBON AND KJELDAHL NITROGEN CONCENTRATION, mg/1
Figure 59. Liquid TOC and TKN concentration as a function of depth in an unaerated
anaerobic, oilot-scale field lagoon loaded at the reference rate with swine
waste.
-------
reactors. At the steady-state conditions for the winter-spring period
the effluent concentration of COD and TOG were approximately one-third
and one-sixth of first lagoon levels for the second and third stage
lagoons respectively. The TKN level for the second and third series
units were one-third and one-tenth of the first lagoon effluent concen-
tration. Thus, a somewhat greater percent nitrogen removal was recorded
for the third unit compared with reduction of organics (COD and TOG).
Lagoon Surface and Physical Properties-
Among the field pilot-scale lagoons receiving different loading rates
the unit receiving four times the reference rate had much different
physical characteristics and correspondingly, chemical composition.
Supernatant samples evidenced a characteristic thick fibrous slurry
not found with lower loading rates. In addition the supernatant con-
centration of phosphate was quite high (450-500 mg/1 o-P04~P). As
discussed earlier in the laboratory cone experiments an upper limit of
50-80 mg/1 o-P04~P according to a solubility limit appeared to exist
for phosphate. The elevated phosphate level and supernatant composi-
tion observations indicated considerable hindered settling at this high
loading rate. Samples were taken at several depths in the four times
reference rate lagoon and waste parameter concentrations with depth are
given in Table 23. The effects of the surface mat and the increase in con-
centrations with greater depths were evident. The other pilot-scale
units did not evidence these concentration gradients with depth which
further indicated hindered settling in the heaviest loaded unit. Because
the supernatant concentration leveled off well below raw waste levels
for this reactor, the increased levels of suspended material were in
part offset by biological activity apparent from visual examination of
gas bubble production and total bacteria assays.
Table 23. LIQUID CONCENTRATION AT SEVERAL DEPTHS FOR A FIELD PILOT -
SCALE LAGOON LOADED WITH SWINE WASTES AT FOUR TIMES THE
REFERENCE RATE
Distance
above bottom
(m)
1.85 (surface)
0.92
0.62
0.31
0
COD
11, .000
6 , 000
49,000
67,000
70,000
Concentration, mg/1
Parameter
TKN
3,500
1,900
4,600
5,400
5,800
o-PO/.-P
420
230
2,500
3,900
6,800
131
-------
Visual observations and physical evaluation of these pilot-scale reactors
were made continuously as the experiment progressed. The surface of
the most heavily loaded unit (4 times reference rate) had a pronounced
scum layer which during hot periods was quite dry and mat-like. How-
ever, after loading and during periods of pronounced gaseous eruptions and
sludge rising, the surface mat would break apart. The more lightly loaded
lagoons (1, 1/2, and 1/4 times the reference rate) had a more liquid type
of surface with variable scum amounts and location depending on wind or
other conditions. The least heavily loaded units (1/8, 1/16, and 1/32
times the reference rate) had frequent algal blooms and thus a very
different appearance; but when algal growth was not present, surface
conditions were more like farm ponds than lagoons.
Field Pilot-Scale Lagoon Gas and Odor Generation-
The gas generation rate of the most heavily loaded unit was estimated
by the time required to fill an inverted quart jar. The rate was about
210 liters of gas per m^ of surface area per day or 113 liters per
cubic meter of lagoon volume per day. For comparison, total gas
production for high rate sewage digesters is about 4,000 liters per
cubic meter of volume per day. The approximate concentration of lagoon
gas was recorded to be about 70 percent methane and 25 percent carbon
dioxide.
Odor from these lagoons varied considerably with climate conditions
and loading rate. On many days with average humidity and wind condi-
tions, offensive odor was not recorded for any of the units. However,
on other days when odors were prevalent, a very rough judgment based
on observations of researchers and students indicated that the unit
loaded at 1/8 the reference loading rate was below the odor threshold.
Individual consensus indicated that the frequency or probability of
odor detection when visiting the site was 80 percent for the 4 times
unit, 60 percent for the reference unit, 20 percent for the 1/2 unit,
and little odor for units receiving 1/4, 1/8, or a lower fraction of
the reference rate. For experimental reactors located in such close
proximity as in this study, certainly odor interferences affect field
evaluations.
In order to further relate odor threshold to loading rate, an odor panel
approach was begun. Samples of lagoon supernatant, from the surface
and from the one-meter depth were put in 200-ml wide mouth jars. A
cover was placed on each jar and each panelist would remove the lid
and waft the gases toward himself to rate the odor. The panel of 7 to
8 persons consisted of secretaries, faculty members, and laboratory
personnel who were not heavy smokers. The rating scale was: 1 -
strongly object, 2 - object, 3 - would not object, and 4 - not offensive.
Colored samples of tap water were incorporated as a control. An odor
rating sheet is presented in Appendix B2.
132
-------
The odor ranking was the most offensive (1.14+0.56) for the surface of
the unit receiving the heaviest loading rate and as loading decreased to
the lowest rate, the odor ranking became less offensive (3.58+0.67),
Table 24. The nature of the odor became questionable at the lower lagoon
loading rates. Odor differences for all units between the surface
sample and that from one meter below the surface were within one
standard deviation and thus not significantly different which further
indicated some homogeneity in lagoon supernatant.
The difficulty of an odor ranking system such as used here is that no
odor characterization is included. Most natural water bodies and even
some processed drinking water have an "odor" but generally typed as
algal decay, mineral, sulfur, etc., rather than manure-related. Thus,
the relative ranking by an odor panel was not an entirely valid indicator
of an offensive or nuisance-causing lagoon loading rate. Further
refinements of odor sensing and inclusion of field climatological
conditions are necessary for the best characterization of the odor
potential for different lagoon loading or design size.
Nevertheless, from the less refined aperiodic field observations and
the results of the odor panel ranking the initial conclusion was that
there was a discernible odor threshold at approximately 0.25 to 0.125
times the reference loading rate. Below this threshold, odor was not
manure-like nor was an odor always detectable. Above 0.25 times the
reference rate (9.3 m^/45-kg hog) odor was not always detectable but
when found it was characteristic of swine manure and hence, was deemed
more offensive.
133
-------
Table 24. DEPENDENCE OF PANEL-RATED ODOR RANK ON SAMPLING LOCATION
AND LOADING RATE FOR FIELD PILOT-SCALE UNITS LOADED WITH
SWINE WASTE
Lagoon loading
(multiple or fraction
of reference rate)
4
4
1
1
0.5
0.5
0.25
0.25
0.125
0.125
0.062
0.062
0.031
0.031
Control
Sampling position
(meters below surface)
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Odor rank
(Total scale:
4 - would not offend
at all
1 - strongly object)
Standard
Mean Deviation
1.14 0.56
1.34 0.52
1.47 0.79
1.39 1.04
2.00 0.70
1.63 0.87
1.85 0.60
1.98 0.77
2.67 0.68
2.44 0.81
3.29 0.71
2.74 0.82
3.58 0.67
3.17 0.82
3.6 0.46
134
-------
Lagoon Sludge-
After about 22 months of operation the sludge depth in these pilot-scale
reactors was determined. The first technique used was a hollow tube,
stoppered at the bottom, which was lowered to successive lagoon depths.
At a given depth the stopper was removed and a sample taken. Another
aliquot was taken at a lower depth until a number of depths were repre-
sented by samples. Then visual observations were made to determine the
depth at which a consistency change was noted and this was termed the
sludge depth. The results for this visual method of analysis for several
lagoon loading rates are given in Table 25.
Table 25. SLUDGE DEPTH DETERMINATION FOR ANAEROBIC SWINE LAGOONS LOADED
AT DIFFERENT RATES.
Unit loading rate
(multiple or fraction
of reference rate)
4
1
0.5
0.25
0.125
Sludge depth_, m
Visual
Method
__
0.33
0.28
0.21
--
Concentration
change method
0.77-0.92
0.44
0.36
0.14
0.13
Sludge volume
percent of
input volume
9.5
16
12
24
In order to further refine the visual technique for determining sludge
depths, a number of samples were taken at several vertical positions with
the hollow tube device, previously described and analyzed for TOG, O-P04-P,
NH3-N, and TKN (Appendix B3). A typical plot of these data, Figure 59,
shows the sudden concentration increase marking the change from super-
natant to sludge zone. The distance from the bottom to this concentra-
tion demarcation was noted as the sludge depth and these visual and
chemical measurements for several lagoon loading rates are included in
Table 25- The two methods for determining sludge depths gave similar
results;hence, the simpler visual method was preferred, given the
approximate nature of measuring techniques.
The sludge depths did not increase linearly with increased loading
rates indicating that either compaction or biological activity reduced
accumulation rate in swine lagoons. Sufficient numbers of samples
from these field units have not been analyzed to determine the nature
and factors affecting rate of sludge build up. Long-term sludge buildup
135
-------
was especially difficult to predict because of the variance in
accumulation rates with the various laboratory, pilot-scale and field
lagoon experiments. A the reference loading rate, which provided
2.3 m^ of lagoon volume per 45-kg hog, the sludge buildup rate deter-
mined for the first two years of operation would result in lagoon
filling after about 9 years. These pilot-scale field reactors have
an impermeable bottom thus maximizing sludge buildup. Continued
monitoring of sludge buildup for 5 to 10 years would be necessary
to determine compaction and other sludge accumulation variables
contributory to much slower lagoon filling noted in full scale units.
Lagoon Treatment Efficiency
The removal efficiency as defined earlier for the laboratory reactors
was based on the influent and effluent concentrations and the assump-
tion that the sludge accumulated at a slow rate. Calculated removal
efficiencies for all units are summarized in Table 26. For the heaviest
loading, 4 times the reference rate, removal efficiencies were quite
low with organics being 40 to 50 percent, orthophosphorus being 25
percent, and no total nitrogen removal. These field unit removals
were less than fifty percent as effective as the corresponding
laboratory units. However, the relative efficiencies for the laboratory
units were still the highest for carbon, medium for phosphate, and
lowest for nitrogen.
The removal efficiencies improved with decreased loading rate so that
at or below one-half the reference rate the organ carbon removal was
95+ percent. At one-eighth of the reference rate or less, the phosphate
and nitrogen removals were 94+ percent. Thus, the conclusion, verified
by laboratory results, is that a high removal efficiency for organic
carbon (COD and TOC), nitrogen (TKN), and phosphorus (o-PO^-P) was
attained at realistic anaerobic swine lagoon loading rates. However,
the quality of the lagoon effluent was still very poor and not suitable
for stream discharge.
Miscellaneous Supernatant Quality Measurements
In addition to the conventional waste constituents used as performance
measures of anaerobic swine lagoons, several other parameters were
monitored on an occasional basis. Detectable levels of dissolved
oxygen near the lagoon surface were not consistently found except for
the 1/16 and 1/32 times the reference rate and the third series lagoon.
Dissolved oxygen and supernatant COD and TOC data for these three
field pilot-scale units included in Appendix B4 showed considerable
variation although dissolved oxygen was consistently present. High
levels of surface dissolved oxygen were found even when supernatant
levels were as high as 600 rag COD/1 (300 mg TOC/1). No dissolved
oxygen was detected at the mid depth or bottom of any of the pilot-scale
136
-------
Table 26. REMOVAL EFFICIENCIES (EQUATION 1) AND EFFLUENT CONCENTRATION OF VARIOUS PARAMETERS FROM
PILOT-SCALE FIELD REACTORS LOADED WITH SWINE WASTE
Unit
4
1
1/2
1/4
1/8
1/16
1/32
Effluent
COD,mg/l
20,000
2,500
2,000
1,650
900
500
250
Efficiency,
7<>
38
94
95
96
98
99+
99+
Effluent
TOG, mg/1
7,000
1,200
700
550
300
175
90
Efficiency,
7o
53
92
95
96
98
99
99+
Effluent
TKN,mg/l
2,700
1,050
600
350
160
70
32
Efficiency,
70
0
58
76
86
94
97
99
Effluent
o-PO^-P
600
75
60
60
50
30
15
Efficiency,
%
25
91
92
92
94
96
98
U)
--4
Input concentrations:
COD
TOG
TKN
o-P04-P
40,000 mg/1
15,000 mg/1
2,500 mg/1
800 mg/1
-------
units regardless of Loading rate. Thus it was concluded that other
factors beyond bulk supernatant concentrations controlled the presence
or absence of surface dissolved oxygen. These phenomena were not
fully developed.
27
The unit loaded at 1/32 of the reference rate was designed to be
naturally aerobic based on Soil Conservation Service recommendations.'
Loading and design recommendations for naturally aerobic lagoons are
shown in Table 27, along with the operation criteria for the 1/16
and 1/32 unit. After the waste input was stabilized at 40,000 mg COD/1,
no dissolved oxygen was found at depths greater than 10 cm below the surface
in either the 1/16 or 1/32 times the reference rate units ; hence
referenced design criteria for unaerated aerobic lagoons for swine
waste was not supported by this study. However, since effluent quality
from even aerobic ponds is not sufficient to allow stream discharge
the presence of bulk dissolved oxygen is relatively unimportant.
Lagoons are only pretreatment units prior to land disposal; and since
odor thresholds are at higher loading rates than specified for
unaerated aerobic units, regulations or efforts to achieve aerobic
ponds are counterproductive.
Table 27. DESIGN CONSIDERATIONS FOR NATURALLY AEROBIC lAGOONS AND
PILOT-SCALE LAGOON PERFORMANCE
Source
Soil Conservation Service
North Carolina
1/32 reference rate -
pilot scale
1/16 reference rate -
pilot scale
Surface area, m2
per 45-kg hog
27.5
27.5
41.2
20.8
kg BOD5 per
ha, per day
70 - 130
-
37
70
The second parameter monitored as a correlation to the chemical oxygen
demand (COD) was the five-day biochemical oxygen demand (6005). Results
for COD and BOD5 analysis on samples from the pilot-scale units during
a steady operation period in May, 1974, are given in Table 28. The
variability of the BOD5-COD ratio further verified the many difficulties
associated with the BODC test.
138
-------
Table 28. COMPARISON OF BIOCHEMICAL AND CHEMICAL OXYGEN DEMAND OF
VARIOUS FIELD AND LABORATORY ANAEROBIC REACTORS LOADED
WITH StfLNE WASTE
Sample
source
Pilot
reactors
Lab
reactor
1
1
2
2
3
3
4
4
Raw
waste
Multiple or fraction
of reference loading
4
4
1
1/2
1/4
1/4
' 1/8
1/8
1/16
1/16
1/32
1A
1A
IB
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
BOD5
(ms/1)
4,600
2,520
386
250
460
503
210
41?
95
149
54
190 •
110
22
113
693
104
223
117
269
98
240
3,800
8,900
COD
(mg/1)
21,500
18,500
2,000
1,100
825
825
625
62.5
550
550
200
900
900
475
1,515
1,570
1,439
1,439
1,511
1,515
1,476
1,476
6,718
24,120
BOD «
COD'
21
14
19
23
56
61
34
66
17
27
27
21
12
5
8
44
7
16
8
18
7
16
57
37
139
-------
Because the dissolved oxygen level of the majority of the lagoons was
zero, a more sensitive measure of anaerobic levels for the various
loading rates was needed. The oxidation - reduction potential (Eh)
has been used in anaerobic systems to measure the degree of reduced
conditions. Converse^ found that the Eh potential varied from -+400
mV for aerobic systems to -250 mV for highly anaerobic conditions.
Below a threshold Eh level of -50 mV to -100 mV odor was found to
increase dramatically; hence Converse concluded that the presence
of these lower, more reduced conditions was a good measure of the
odor potential associated with anaerobic treatment of animal waste.
Additionally, the odor associated with a waste system was more accurate-
ly evaluated by the oxidation-reduction potential (Eh) than the dissolved
oxygen (D. 0.) level.
It was attempted to use the laboratory approach of Converse to evaluate
the odor threshold of these pilot-scale units. However, the oxidation-
reduction potential of a given lagoon varied considerably over the
3-week data period, Table 29. Thus no opportunity existed for a
reliable unit characterization by E, potential. There was the expected
trend toward more reduced conditions in the more heavily loaded units,
but the week-to-week variability has so far limited the use of this
parameter for lagoon studies.
The pH of the laboratory and field units were measured over a wide
variety of temperature and loading conditions, Appendix B5. Lagoon
supernatant was fairly constant ranging between 6.5 and 8.5 with the
majority of the pH values between 7.3 and 7.8. The laboratory
units were slightly more basic but no trend existed for changes in
pH with loading rate or reactor concentration. The exception so far
unexplained is that the second and third series lagoons characteristically
had higher pH values than the other lagoons even though supernatant
concentrations were similar to laboratory and field units at a fraction
of the reference loading rate.
Pilot-Scale Aeration
The mixed culture microbial activity of a anaerobic lagoon breaks
down the long-chain organic compounds characteristic of animal feeds
and waste products to short-chain alcohols, amines, sulfide compounds,
as well as other more stabilized gaseous products including methane
and carbon dioxide. These short-chain molecules have characteristic
odors which even at low levels are identified as nuisances. Obser-
vations of lagoons which have little or no odor indicate that the
upper surface zone is aerobic or at least not highly anaerobic even
though the lower zones are strictly anaerobic. Such lagoons have been
termed facultative or diphasic and have little or no manure odor.
140
-------
Table 29. OXIDATION-REDUCTION POTENTIAL MEASUREMENTS AT MID-DEPTH IN
FIELD PILOT-SCALE ANAEROBIC SWINE LAGOONS
Lagoon loading
(multiple or fraction
of reference rate)
4
1
1A
IB
1/2
1/4
1/8
1/16
1/32
Raw waste
4
1
1A (second series lagoon)
IB (third series lagoon)
1/2
1/4
1/8
1/16
1/32
4
1
1A (second series lagoon)
IB (third series lagoon)
1/2
1/4
1/8
1/16
1/32
Raw waste
Date
3/20
3/20
3/20
3/20
3/20
3/20
3/20
3/20
3/20
3/20
3/23
3/23
3/23
3/23
3/23
3/23
3/23
3/23
3/23
3/27
3/27
3/27
3/27
3/27
3/27
3/27
3/27
3/27
3/27
Eh
Reading
-340
-120
11
54
-22
-27
51
151
41
-232
-285
-183
-105
-110
-160
-208
-139
-135
-86
-290
-200
-165
-142
-186
-230
-200
-161
-106
-272
141
-------
An engineered equivalent of these facultative lagoons is achieved by
the use of a mechanical device for adding oxygen to the upper zone by
promoting mixing and pumpage across the entire lagoon surface. These
surface aerators of which there are a number of commercial types
increase oxygen transfer by various mixing or surface agitation
patterns. In order to test horsepower requirements, degree of stabili-
zation of waste constituents, and level of odor control associated with
surface aeration two additional pilot-scale units were installed at
the swine research site, Figure 40. As with the other pilot-scale
units, these were 2.5 meters deep and 3.5 meters in diameter. Metal
bottoms were fabricated to prevent any seepage. A catwalk superstruc-
ture spanning the tank diameter at the top of the lagoon was constructed
to support a fixed aerator, Figure 60. The fixed aerator was chosen
because of the small horsepower requirements for this pilot-scale
volume (maximum 250 watts). A 187 watt, variable speed, mixer-
aerator was selected. The impeller was specified to be just completely
submerged below the liquid surface. The constant liquid level required
to maintain proper impeller submergence was achieved with an overflow
stand-pipe and a water reservoir controlled by a mechanical float
valve to counterbalance rainfall and evaporation, Figure 60.
The first field pilot-scale unit was operated at an impeller rotational
speed of 65 rpm while the second unit aerator speed was 110 rpm. These
speeds required 37 and 60 watts, respectively. Each reactor was
charged with two times the reference loading rate at a once-per-week
frequency resulting in 840 liters per week of raw swine waste at a
constant concentration 40,000 mg COD/1. The operation and sampling
were the same as the field pilot-scale unaerated anaerobic lagoons.
Phase One Experiment-
i
The surface aeration experiments were divided into two phases with the
first beginning after the swine waste input was held constant at ,
40,000 mg COD/1 (October, 1973). For the first experimental phase
(power input 37 and 60 watts, respectively), the supernatant concen-
trations averaged over two-week periods are presented in Figures
61 - 64. Steady-state concentrations for COD, TOG, and O-P04-P were
achieved more quickly than for TKN, 20 weeks versus 25 weeks from
initiation of uniform waste input. The relative rate steady
condition attainment among the various parameters was the same as that
observed for unaerated anaerobic units, both field pilot-scale and
laboratory reactors. Steady conditions were reached at about the same
time for both aeration rates indicating temperature, chemical transfer
or microbial reactions were as important in governing supernatant
concentrations as oxygen transfer or mixing intensity.
142
-------
WATER
STORAGE
SURFACE MIXER WITH
RHEOSTAT FOR VARIABLE
SPEED CONTROL
FLOAT VALVE FOR
WATER LEVEL
CONTROL
>~ OVERFLOW
Figure 60. Schematic or aerated pilot-scale lagoon with water level
control for fixed surface aerator.
-------
20,000
I? 16,000
z
o
55
O
CJ
Q
O
O
W
O
53
pa
o
U
M
rc
u
12,000
8,000
4,000
Phase I
60 watts, 7 days/week
- - - 37 watts, 7 days/week
Phase II
60 watts, 6 days/week
- - - 120 watts, 7 days/week
Phase I Phase II
7/10/73
Figure 61. Supernatant COD concentration for field pilot-scale lagoons with surface
aeration loaded once per week at 2 times the reference rate for swine wastes.
-------
8,000
6,000
u
§
o
CJ
o
H
55
O
O
O
4,000
3 2,000
03
O
o
H
Phase I
60 watts, 7 days/week
37 watts, 7 days/week
Phase II
60 watts, 6 days/week
120 watts, 7 days/week
Phase I Phase II
7/10/73
Figure 62.
40
50
60
TIME, weeks
Supernatant TOC concentration for field pilot-scale lagoons with surface
aeration loaded once per week at 2 times the reference rate for swine wastes,
-------
Phase II
00
H
2
W
O
&
O
u
2:
w
o
§
H
H
Q
.J
W
H
O
H
1,600
800
Phase •!
60 watts, 7 days/week
- - - 37 watts, 7 days/week
Phase II
60 watts, 6 days/week
- - - 120 watts, 7 days/week
7/10/73
10
20
TIME, weeks
Figure 63. Supernatant TKN concentration for field pilot-scale lagoons with surface
aeration loaded once per week at 2 times the reference rate for swine wastes,
-------
r-t
60
O
M
H
H
55
W
u
53
O
o
CM
i
o
pa
H
en
O
as
o
I
o
800
600
400
200
Phase I
60 watts,
37 watts,
7 days/week
7 days/week
Phase II
60 watts, 6 days/week
120 watts, 7 days/week
Phase I Phase II
7/10/73
Figure 64.
10
30
40
TIME, weeks
50
60
Supernatant o-PO^-P concentration for field pilot-scale lagoons with surface
aeration loaded once per week at 2 times the reference rate for swine wastes,
-------
The final supernatant concentration in these surface aerated units
was affected by a complex combinatJ~ . of particulate settling, surface
agiu tion and evaporation, and aerobic and anaerobic reactions. Overall,
rh^ r ipernatant concantraf:'.or. of these units loaded at two times the
reference rate was preate^ than pilot scale units discussed earlier
loaded at the reference rate and was usually less than the anaerobic unit
loaded at four times the reference rate. Thus, on a very rough basis
the supecnav.'nt concentration appeared to depend on the swine waste
loading rate in a manner consistent with other unaerated anaerobic
units, although there were a number of conflicting factors.
The a_ priori predicted levels of organics in aerated systems are depen-
dent on input waste concentration (COD), reactor volume, and aerator
horsepower. For a given waste type, continuous aerator operation, and
a fixed reactor size, the greater the horsepower the greater the oxygen
transferred and hence the more organic stabilization and lower COD
expected. Manufacturers estimated oxygen transfer rate for most
surface aerators is about 2 kg oxygen per hour per kilowatt. Calculations
based on the COD load (2 times reference rate) and the manufacturer's
oxygen transfer rate for continuous operation indicated that the unit
with the higher aeration input received about 70 percent of the oxygen
needed for complete stabilization and the lower one about 40 percent.
Subsequently, an approximate oxygen balance to verify the manufacturer's
oxygen transfer rate was performed during the second phase of experi-
mentation. Thus, the pilot-scale reactor with the higher power setting
was anticipated to have a lower COD supernatant concentration because
of greater oxygen transfer. However, it was found that the higher
aerated unit had greater COD, TOC>and orthophosphate concentrations
but about the same Kjeldahl nitrogen level as the lower aerated unit,
Figure 61-64. To explore reasons for this concentration trend,
procedures were initiated to more completely evaluate the employed
surface aeration.
The degree of mixing was first determined at a number of aerator speeds
with a OTT-Current Meter which measured the fluid velocity and direction
and thus could be used to determine the depth of aerator influence
or significant flow patterns. From this method it was found that at
the surface for 37 watt power input the fluid velocity decreased from
20 cm/sec at 15 cm from the impeller to 6.4 cm/sec at 45 cm from the
wall. For the 60-watt setting these spatial velocities decreased from
25 to 11 cm/sec over the same distances from the aerator. Fluid
velocities below the surface should be detectable to the lower limits
of the current meter (0.025 m/sec). However, it was found that the
mixing zone could not be adequately defined by this current meter approach;
therefore, samples were taken to determine the variation of parameter
concentrations with depth in both aerated units.
148
-------
The orthophosphate concentration at several depths for the high and low
aeration rate units are given in Figure 65. Both units had the same
raw waste input, loading frequency, and management but different power
inputs. Depth profiles showed a uniform upper zone and then an abrupt
concentration increase, indicative of the sludge blanket. However, for
the high aeration unit, the sludge depth was less and supernatant
phosphate concentration was higher (by a factor of three, 350 versus
125 mg/1) than for the low aeration unit. This resulted from the greater
resuspension of sludge and hindered settling associated with the greater
aeration rate. Both aerated units had supernatant concentrations above
the levels of 40 - 80 mg/1 phosphate found in the unaerated field pilot
scale units. These facts indicated that the depth of aerator influence
was greatest for the 60 watt aerator settling (1.7 m) and that this
mixing caused an abnormally high supernatant concentration despite the
greater oxygen input obtained at this aerator setting. The lower aera-
tion reactor (37 watt) had a smaller mixing zone depth of 1.4 m. It
should be noted that the field pilot unaerated units also had uniform
supernatant concentrations. However, because of the higher levels
of orthophosphate, it was concluded that the uniform supernatant was
evidence of aerator agitation and not the inherent homogenity phenomena
associated with comparable unaerated units.
This investigation thus suggested that the higher organic levels were
due to the increased agitation and thus suspension of sludge solids.
Because the phosphate level was about three times greater at the higher
aeration rate, Figure 64, and the organic level (COD and TOG) for this
unit was only twice as high, Figures 61 and 62, the expected higher
level of oxygen addition and organic stabilization for the greater
horsepower input was realized. There were no distinct aerobic zones
at either aeration rate and no dissolved oxygen was found even in the
surface layer of these lagoons.
The supernatant ammonia concentration was 1,250-1,300 mg/1 at 60-watt
power input and 900-950 mg/1 at the 37-watt level. Thus there was
only about 40-50 percent difference in ammonia between these units.
The supernatant TKN levels in the reactors were about 1500 mg/1 for
both aeration rates. Compared to phosphate or COD, the nitrogen
concentrations were more similar between units showing that the
greatest impact of aeration was on ammonia volatilization. Because
ammonia loss is gas phase limited, the augmentation in gas phase
turbulence and the increased surface renewal of reactor liquid due to
aeration increased ammonia volatilization potential. Thus, the expected
resuspension of sludge at higher aeration input was counteracted par-
tially by increased volatilization.
149
-------
0.0
W
O
PS
OT
1
rd
H
P-I
w
Q
1.0
U Q
Q 37 watts AERATION
° 60 watts AERATION
I
1
2.0 400 800 1200 1600 2000 2400 2800
ORTHO PHOSPHATE (o-PO^P) CONCENTRATION, mg/1
Figure 65. Depth profile of ortho phosphorus concentration for
pilot scale swine lagoons receiving different rates
of surface aeration.
-------
Gas and Odor Generation-
Observations regarding odor were made on a number of different occasions
after these pilot-scale units were being operated at steady-state condi-
tions. These observations were by no means rigorous but on the average
the more intensively aerated unit had less (sometimes even no odor) than
did the lightly aerated unit. The higher horsepower input caused the
liquid to be thrown into the air in droplets while the lower horsepower
input only disturbed the liquid in a wave-like manner. The relative
effects of oxygenation, volatilization, and surface mixing due to the
employed surface aerator are not exactly known; hence, a complete
qualitative and quantitative explanation of mechanisms for odor control
has not been developed. The power in terms of watts/m of volume were
3.3 and 2.1 for the two reactors. The required power input per unit
volume to achieve a consistent, high level of odor control depends on
the lagoon depth and shape as well as a number of other factors that
could not be specifically determined from this field pilot-scale study.
When the aerator in either reactor was stopped, a large number of bubbles
were observed to evolve from the liquid surface. The investigation of
this phenomenon covered two different premises. The first was that in
the surface layer there was a population of microorganisms which
nitrified the ammonia present, and as the liquid was drawn into the
lower zones containing anaerobic microorganisms denitrification
occurred. The liberated nitrogen was then released as bubbles.
The second premise assumed that entrapped air lost oxygen content
as it passed through the anaerobic liquid and then the remaining gas
which would be primarily nitrogen was liberated in surface bubbles.
Gas samples were taken from these reactors at the surface and analyzed
for various gases, Table 30. There was considerable variability but
significant levels of nitrogen were present as well as oxygen. The
presence of methane indicated anaerobic activity but the nitrogen
source or conversions could not be inferred from these data since both
premises allowed nitrogen as a constituent of the bubbles formed. The
bubbling observed after the aeration was stopped appeared to be
sustained although the exact duration was not measured. However,
because of the surface disturbances during aeration, it could not be
determined if the bubbles were liberated continually. Nitrate analysis
of this high COD supernatant was difficult because of analytical
interferences and the potential for denitrification. Taking these
factors into consideration, the nitrate level in both aerated reactors
was in the range of 0 to 15 mg N03-N/1. This was low compared to
oxidation ditch levels but does not rule out the generation of signi-
ficant amounts of nitrates. That is, if nitrification and denitrifi-
cation were occurring simultaneously, the nitrate level might in fact
remain low while significant steady-state nitrogen losses occurred.
Unfortunately, with these aerated field reactors, a mass balance on
151
-------
nitrogen would not indicate nitrification-denitrification losses because
of the confounding effect of ammonia volatilization. Hence, further
supportive experiments are needed to refine these initial aeration
results and proposed mechanisms.
Table 30. GAS COMPOSITION FROM FIELD PILOT UNITS RECEIVING DIFFERENT
AERATION INTENSITY
Date
7/16/73
7/19/73
8/1/73
7/2/74
7/11/74
Aeration,
watts
60
60
37
120
120
Parameter volume percentage
°2
2.6
2.0
4.0
2.0
4.0
N2
30.2
35.6
91
96
76
CH4
67.2
61.0
4.0
0.0
18
C02
trace
1.4
1.0
2.0
2.0
Phase II Experiment-
After 5 to 10 weeks of operation at the steady-state supernatant concen-
trations attained in phase I, the management of these two reactors was
changed as phase II of these experiments. It had been concluded that
both units had hindered settling or enhanced sludge suspension. The
management of the high aeration reactor was changed so that after raw
waste loading the aerator was stopped for 24 hours. This time period
had been shown to be more than adequate for solids settling so that
any aerator effect which hindered initial settling would be overcome.
After 24 hours the aerator was restarted. The expected result would be
lower supernatant concentrations due to enhanced settling. The other
reactor initially at the lower power setting was increased to 137 rpm
or 120 watts to determine the impact of greater aeration input. Both
units remained at the same swine waste loading rate and frequency.
The supernatant concentration changes for these new management variables
are shown in Figures 61-64. As expected, concentrations for the reactor
with the power change from 37 to 120 watts immediately increased above
those for the 60-watt unit, due to the deeper mixing zone and thus
greater bottom sludge scour. Unexpectedly, TKN showed only a modest
increase. A possible explanation is that the increased volatilization
compensated for the greater part of the expected concentration increase.
152
-------
That ammonia volatilization was greater for increased aeration levels
was indicated by the lower supernatant TKN concentration for the 120-
watt versus 60-watt unit (after the 52nd week), Figure 63.
The decline of COD, TOG, and TKN in both aerated units after the 52nd
week (July 15, 1974) roughly followed that of unaerated units and was
due to greater biological activity and volatilization at warmer liquid
temperatures. Because of this warmer temperature interference, direct
quantitative comparison of the various management schemes was not
possible. Qualitatively, the 120 watt input did not proportionally in-
crease the COD, TOG, TKN, or o-PO^-P concentrations as would be indicated
by concentration changes between the 60 and 120 watt operation after
initial concentration increased due to bottom scour. The decline in
orthophosphate values (56-61 weeks) was not explainable although the
unaerated pilot-scale lagoon loaded at 4 times the reference rate
behaved similarly. If the supernatant phosphate removal mechanism is
a precipitation, this would not be improved at higher temperatures on
a solubility basis. The heavily loaded unaerated unit was found also
to have hindered settling; thus, it may be possible that the decline in
phosphate concentration was associated with improved settling character-
istics with higher liquid temperatures. The complete explanation is
not yet clear.
Allowing a quiescent period (24 hours) for settling does not appear to
improve the supernatant quality as evidenced by the similarity of con-
centrations in the 60-watt reactor immediateley before and after the
initiation of the six-day-per-week aeration. It was felt that the
subsequent decline in supernatant concentration was due to the increased
reactor liquid temperature during the summer. The reasons for the
elevated supernatant concentrations at the 60-watt over the previous
37-watt setting was due to greater bottom sludge scour rather than
hindered initial settling.
After 77 weeks of operation, a mass balance on the two aerated reactors
was made accounting for COD input, effluent, and accumulation within the
lagoon. The difference between input and the effluent plus accumulation
was divided by the number of hours of operation and power rating to give
the oxygen transfer rating of these aerators, Table 31. The calculated
transfer wrs 1.7-1.8 kg 02/hr/Kw which was about 80 percent of the
manufacturer's rating of 2.1 kg 02/hr/Kw.
FARM SCALE LAGOON
Full scale field lagoons were of necessity studied under conditions
more closely following producer management. The producer lagoon
system used in this study was jointly operated by North Carolina State
University and the North Carolina Pork Producers Association and was
153
-------
Table 3 1. OXYGEN MASS BALANCE FOR FIXED AERATOR OPERATING IN FIELD PILOT-SCALE REACTOR LOADED
ONCE PER WEEK AT TWO TIMES THE REFERENCE RATE FOR SWINE WASTES
Reactor X
Reactor Z
Ul
Operation period
Aeration input based on
manufacturer's rating
2.1 kg 02/hr/kw, kg 02
Oxygen demand in
input, kg 02
Oxygen demand in
effluent, kg 0»
Oxygen demand in
reactor, kg 02
Oxygen demand
unaccounted for, kg 02
Oxygen transfer rate
based on unaccounted COD
a) 47 weeks, 7 days/week
60 watts.
b) 20 weeks, 6 days/week
60 watts
1,380
1,820
490
220
1,120
1.8 kg 02/hr/kw
a) 47 weeks, 7 days/week
37 watts.
b) 20 weeks, 7 days/week
120 watts.
1,530
1,820
330
280
1,120
1.7 kg 02/hr/kw
-------
located near Raleigh. Lagoon operation was begun in 1961 in conjunction
with a single concrete floor, totally roofed swine house. Wastes were
flushed from the floor of each pen (1.4 m wide by 4 m long) into a gutter
which sloped along the length of the house and emptied into the lagoon.
The wastes were flushed to the gutter with a high pressure hose system
once per day resulting in about 38 1 of wastewater/d/45-kg hog. Swine
were raised from approximately 20 kg to 100 kg in about 6-month intervals
and then marketed much like a typical production unit. From 1961 to
1970 the average steady-state population was 160 hogs.
In 1971, an identical house and a second lagoon were constructed and the
lagoons operated in series with the oldest lagoon being the first unit.
From 1971 through 1972, swine production was similar to the previous
years except that the average steady-state population was about 220 head.
After January, 1973, the production unit was converted to a boar testing
station. The upper one meter length of each pen was segregated and filled
with shavings to reduce boar foot damage. Except for some spillage,
these shavings were removed for land disposal after each testing cycle.
The manure was removed primarily by scraping resulting in a volume
reduction to about 11-15 l/d/45-kg hog. In terms of waste management,
the hog population expressed as pounds of liveweight increases, peaks,
and then declines on a twice-per-year basis.
The first lagoon is about 28 x 26 meters (730 nr) with an original
average liquid depth of about 1.2 m. This lagoon had been in operation
since 1961, and had an average sludge depth of about 0.46-0.62 meters. Thus,
there were about 362 nr of lagoon liquid volume. The daily live-
weight contributing to this lagoon and the corresponding value of the
supernatant TKN was evaluated, Figure 66, as well as the concentration
pattern for COD and TOC over a fourteen-month period, Figures 67 and 68.
During both hog population cycles, the lagoon TKN supernatant concentra-
tion responded in rough sychronization with population liveweight and
the response times were fairly rapid. The average weight present in
these cycles was calculated by integrating the areas under the live-
weight curve (Figure 66) and dividing by the total number of days.
For the spring-summer period, the average liveweight was equivalent to
230-45-kg hogs while the fall-winter was equivalent to 220-45-kg hogs.
Thus, the average lagoon loading is about 1.61 m-^ per 45-kg hog based
on the current supernatant volume of about 1.4 times the reference
rate used in this study. Based on the original lagoon volume (804 nP),
this waste loading rate was 3.6 m per 45-kg hog or .6 times the
reference loading rate (2.3 m /45-kg hog).
The supernatant concentration of TKN for the warmer spring-summer cycle
was about 250 mg/1 while for the fall-winter period about 350 mg/1.
The median ambient air temperature for these periods was about 24°- C
(spring-summer) and 10°-C (fall-winter). The difference in nitrogen
concentration reflected the shift in equilibrium to ammonium ion and
the decreased volatilization associated with lower temperatures. An on-
farm lagoon monitored in Iowa evidenced a cold-warm shift in lagoon
155
-------
21.5
00
B
E5
O
W
O
O
u
a
w
O
§
H
W
H
O
H
500 I—
D LIVEWEIGHT
O TKN
400
300
200
100
^ /
' . f\
3/73
Figure 66.
7/73 8 9 10 11 12/73 1/74 2
TIME, months
5/74
Supernatant TKN concentration and contributory liveweight changes for an on-farm
swine waste lagoon loaded at once per day or more frequency (Clayton).
-------
2,000
o
M
H
H
3
W
u
Q
O
o
S3
W
o
o
M
s
w
pa
u
1,000
3/73 4
Figure 67.
7/73 8
10
11
12/73 1/74 2
TIME, months
Supernatant COD concentration changes for an on-farm lagoon for swine wastes
loaded at once per day or more frequency as waste material runs into laeoon.
-------
00
60
8
O
M
H
w
O
CJ
u
O
e.
O
M
O
1,200 _.
1,000
800
600
400
200 r-
O
erf
O
H
O
H
3/73
I I I
6 7/73 8
I I I
I I I
10 11 12/73 1/74 2
Figure 68.
TIME, months
Supernatant TOG concentration changes for an on-
farm lagoon for swine wastes loaded at once per day
or more frequency as waste material runs into lagoon.
-------
o o
concentration of TKN and NHo-N . The reported winter to summer change of
650 mg/1 TKN to 200 mg/1 TKN for the Iowa situation was larger than the
investigated North Carolina lagoon, probably due to the much colder Iowa
winter period.
The farm-scale lagoon supernatant concentrations were compared to several
field pilot units using the concentrations obtained during the winter-
spring period 1974, Table 32. The farm-scale lagoon was comparable
in supernatant concentration to the pilot unit loaded at 0.5 times the
reference rate. This meant that the loading based on the original
lagoon volume (0.6 times the reference) was similar to the pilot scale
unit loaded at 0.5 times the reference rate. Thus, in terms of available
volume, the farm-scale lagoon was similar to and had lower effluent con-
centration than comparably loaded field pilot units. The explanation
for these lower steady-state concentrations for the farm-scale lagoon
may have included the frequency of loading and the cyclic nature of the
total waste load. From the laboratory experiments described earlier,
it was determined that continuous loading at the lower input rates
resulted in lower supernatant concentrations. This field lagoon was
loaded daily and often the urine and some feces flowed continuously
into the lagoon, thus resulting in nearly continuous loading. Also,
the overall waste addition built up slowly as the hog population
increased which would allow accumulation and development of a good
biological population pursuant to lower effluent concentrations.
Another important factor was the degradation which occurred prior to
the waste being scraped into the lagoon. Because these production
houses were more open to the air than enclosed houses over pits,
there was added opportunity for waste degradation. This indicates the
significant effect growing unit configuration and management can have
on the swine lagoon waste load and thus supernatant concentrations.
Table 32. AVERAGE SUPERNATANT CONCENTRATION VALUES (JANUARY-APRIL,
1974) FOR FARM SCALE AND FIELD PILOT-SCALE LAGOONS
RECEIVING SWINE WASTE
Lagoon
Farm scale
Field pilot scale
reference rate
0.5 times reference
rate
Lagoon volume
m3/45-kg hog
3.6
2.3
4.6
Concentration, mg/1
TKN
400
1,000
500
COD
1,900
2,600
2,000
TOG
1,000
1,000
800
0-P04-P
50
80
1 40
159
-------
To test the possibility of concentration gradients in this lagoon,
supernatant samples were taken at four circumferential positions, 3
meters from the shoreline and one in the lagoon center. At three of
these sites, samples were taken at several points (Table 33). The
variability was larger than for similar gradient studies in pilot
scale lagoons; but the data verified that on the whole, the lagoon
supernatant was reasonably uniform. Gradient conclusions for these
test lagoons has been verified by recent sampling at other producer
lagoons.
Table 33. ON-FARM SWINE LAGOON SUPERNATANT CONCENTRATION DISTRIBUTION
IN VERTICAL AND HORIZONTAL DIRECTIONS (4/4/74)
Position
Direction
North
East
South
West
Center
Depth below
surface, cm
15
15
15
15
15
15
15
Parameter concentrationj mg/1
COD
1,470
1,540
1,300
1,740
1,350
1,820
1,410
TOC
700
660
700
900
680
660
740
TKN
320
210
300
330
320
300
300
Since the initiation of waste management studies at this swine growing
unit, the sludge depths have been measured on an occasional basis.
Initially, sludge height was determined by the resistance offered to
a flat object lowered into the lagoon or by the sludge adhering to a
rod which was dipped into the lagoon. A boat was used to sample or take
these rather approximate sludge measurements at various lagoon locations.
Sludge depth was estimated by the average of five different sludge measure-
ments taken around the lagoon. Overall sludge buildup from lagoon
installation in 1961 until March, 1972, was about 2.5 cm per year. Since
the depth measurements were begun on a more frequent basis (1972), the
buildup appeared to be about 15 cm per year. Utilization of shavings
for foot protection could have contributed to this increased rate. The
15 cm/yr rate would mean an 8-10 year period to fill a lagoon. This
was longer than the 3-8 year period indicated by the laboratory experiment
160
-------
but similar to the filling time of 8-10 years projected for pilot-
scale units loaded at the reference rate. Tentatively, then sludge
'ibuildup with swine lagoons loaded at the reference rate probably will
lead to filling in 8 or more years. However, as noted, the primary
field lagoon reported on herein had been in operation about 13 years
and the present sludge depth was about 60 cm which indicated excellent
sludge stabilization and compaction over long periods.
Anaerobic Lagoon experiments were run in the laboratory, on a field
pilot scale, and for a single, farm-scale unit. Direct transfer of
results would not be feasible because of differences in temperature,
size, and loading strategy; however, comparison of all experiments is
useful for determining trends and lagoon characteristics. In general,
laboratory experiments with Imhoff cones and 14 1 reactors produced
similar results in terms of supernatant quality. Comparisons of steady
state supernatant concentrations for similarly loaded units during the
winter-spring period showed that the COD and TOC concentrations of
laboratory reactors were about one-third of the field pilot unit values.
The TKN laboratory concentrations were about one-fifth of the field
values; while the laboratory orthphosphates were about one-third to
one-half of field units. Thus, direct transfer of effluent concentrations
was not possible without some corrections for differences between experi-
ments .
The uniformity of supernatant concentrations for various pollutional
parameters in laboratory and field units ranging in size from one to over
850,000 liters was an unexpected result for which several possible
explanations are available. The contribution of diffusion was
calculated for point source diffusion into an infinite medium assuming
the waste input was placed in one side or region of a reactor so that
disturbances of total supernatant were reduced. This was reasonably
true of the actual experimental procedure both in the laboratory and
the field. From the standard chemical gradient diffusion models, the
approximate time to attain uniform concentration by diffusion across a
given distance is given by the ratio of the square of the diffusion
distance divided by the diffusivity coefficient. The diffusion co-
efficient for organic molecules in aqueous solution is
10"° cm^/sec or smaller. Such diffusion across a meter of reactor
would require 150 - 450 weeks. Thus, diffusion in these experiments
represented a minor contribution to supernatant uniformity.
Daily temperature cycles and thus thermal induced currents were not
considered as the principle explanation for uniformity since super-
natant uniformity also occurred with laboratory reactors in a constant
temperature environment.
Other mechanisms thus were postulated. Two reasonable explanations
were a) high level of active biomass and b) mixing effects of micro-
organisms and liberated gas. As calculated earlier, the supernatant
161
-------
had large biological populations; thus, reactor gradients after
loading could be rapidly reduced by microbial activities.
The continual mixing action of gas bubbles liberated with waste
stabilization and supernatant microorganism movement could contribute
considerably to uniformity of concentration. Micro mixing, as well as
the gaseous eruptions commonly seen in large anaerobic reactors, would
agitate the liquid and certainly promote uniformity.
Both laboratory and field pilot units showed the same trend of decreased
supernatant COD, TOC, and TKN concentration with reduced swine waste
loading. Supernatant TKN and TOC concentrations were directly propor-
tional to loading rate at 2.3 m3/45-kg hog or more for laboratory and
field experiments. Thus, the effluent concentration for swine lagoons
can only be estimated for a given loading rate.
Sludge COD, TOC, o-PO,-P, and TKN concentrations were similar for all
the experiments conducted, Appendix Al, A2,and A3. COD levels were 35,000
45,000 mg/1 on an as-is basis in field and Imhoff cone studies with higher
concentrations of 60,000-70,000 mg/1 found at the higher loading rates.
No conclusive evidence of concentration profiles within the well defined
sludge zone were found as trends were conflicting in different reactors.
Sludge TOC was generally 12,000-18,000 mg/1 with TKN being 2,000-3,000
mg/1. These sludge COD, TOC, and TKN levels are close to the raw swine
waste values. Sludge orthophosphate concentrations of 2,000-3,000 mg/1
which were two to four times the raw waste value indicated settling
and accumulation of phosphorus consistent with removal mechanisms
discussed previously. Because the sludge values for the conservative
constituent phosphorus were much higher than other parameters, it would
be consistent to assume sludge decomposition and stabilization of
organics and liberation of ammonium.
The sludge buildup for laboratory units loaded at the reference rate
was from 15 percent to 35 percent of the lagoon volume per year. The
field pilot-scale units had a sludge buildup of 10 percent to 15 per-
cent of the lagoon volume per year for the reference loading. Build-
up rate for the field unit was at best an estimate because of the
variable input since startup. Effects of long-term compaction under
controlled conditions for the pilot-scale units must be studied over a
longer time period. However, at this time, it was concluded that sludge
buildup for a lagoon loaded at the reference rate would necessitate
lagoon cleanout at ten-year intervals or greater.
162
-------
SECTION VI
PREDICTIVE AND INTERPRETIVE RELATIONSHIPS FOR LAGOONS
The aggregate data from the laboratory, pilot-scale, and on-farm
experiments conducted in this study allowed certain simplifying assump-
tions to be made as a part of the development of predictive modeling
relationships for lagoon performance and effluent quality. Models were
useful in both interpreting data obtained from this study and allowing
comparison of various experiments on a common basis. Modeling the
investigated anaerobic reactors required certain assumptions to be made,
mostly on a physical basis. The first assumption was that lagoons
consisted of two distinct zones, the supernatant and the sludge. When
material was put into a lagoon operating at steady state, distribution
occurred within a certain time period between both zones. Additionally,
during this time period and prior to the next loading event, some of
the sludge material undergoing microbial reaction was liberated as
by-products from the sludge to the supernatant. Therefore, in consider-
ing the two zones the initial input, transformations, and interfacial
transfer between loading events could be used in mass balance equations.
In terms of these lagoon studies, the net amount of material entering
the supernatant was taken to be the initial unsettled fraction of the
raw swine waste plus the resuspension of sludge components while the
ultimate material remaining in the bottom zone was considered sludge.
Lagoons may be operated with no continuous overflow; or, if part of
a series treatment system, lagoon overflow into another unit may exist.
No continuous discharge was the lagoon management scheme used for the
majority of these laboratory and field experiments. Based on the
absence of substantial supernatant concentration gradients in
laboratory, pilot-scale, and field units (excepting the very heavy
loading of four times the reference rate), the second assumption was
that the lagoon supernatant zone was well-mixed and hence at uniform
concentration. Reasons contributing to lagoon supernatant uniformity
included loading procedures, thermal and wind mixing, gas evolution
and bubbling, and microbial activity. The mechanisms and magnitude of
these phenomena varied considerably between experiments but for predic-
tive purposes the lumped effect was that the
-------
The no discharge lagoon criterion necessitates liquid removal in
moisture excess regions because lagoons have a finite liquid capacity.
Usually a pump-irrigation or a tank wagon system is employed to remove
supernatant liquid on a periodic basis. Over long time periods, the
average overflow will equal the in-flow minus any evaporative or other
losses. Under these operating conditions, the residence time in the
lagoon closely approximated the design equation:
Tr = V/Q (3)
where ^ = residence time, weeks
V = lagoon or reactor volume, liters
Q = waste input rate, liters/week
If continuous or semi-continuous overflow existed, there would be
potential for nonuniform conditions or incomplete mixing42 due to short
circuting or surface streaming. This would lead to shorter and more
variable residence times. However, if influent and effluent pipes
were located as far apart as possible or if baffle construction was
used, then short circuiting would be prevented and lagoon residence
would approximate the theoretical retention time. The well-mixed
assumption would then be more valid and the lagoon supernatant more
uniform.
With these two primary assumptions the standard constant-stirred tank
reactor analysis could be applied to anaerobic lagoons. A continuous
and a batch loading approach was considered since the lagoon operation
had elements of both schemes.
It should be noted that the employed models ,are predominately empirical
in nature and that as data on annual cycles become available models
can be expanded to better predict lagoon effluent quality.
BATCH LOADING APPROACH
The actual operation of the pilot-scale units and most of the laboratory
reactors was batch loading, usually on a once-per-week basis. Prior to
each loading, samples were taken and material was drained to restore
the constant reactor design volume; e.g., 14 liters for the cylindrical
laboratory reactors. These experimental units were nonoverflow, and
immediately after loading the supernatant concentration would rise
reflecting the kilograms of parameters added and the dilution effect of
the lagoon liquid, Figure 69. Settling distribution would then occur
followed by the various anaerobic reaction processes operational over
164
-------
W
CJ
W
Fn
o
g
H
§
U
"Z
o
u
INITIAL CONCENTRATION,'
DOUBLE INPUT
CONCENTRATION
7
TRACE OF-
CONCENTRATIONS
PRIOR TO LOADING,
WASTE LOADING EVENT
TIME, weeks
Figure 69. Schematic of actual lagoon supernatant concentration with batch loading operation.
-------
the period between loading. Thus, a mass balance on any constituent, i,
which remains in the supernatant after initial settling was:
dq
dt
- -kRC,V (4)
with C^ = C? when t = o
where C. = concentration of species i in the supernatant, mg/1
C° = initial concentration of species i in supernatant, mg/1
k^ = removal rate constant, batch system, weeks
t = time, weeks
The term on the left is the rate of change of the amount of species i
with respect to time, which for a constant volume system is actually
the concentration change, grams or gram moles per unit time per unit
volume. The term on the right represents the disappearance.of species
i, assumed to occur at a rate proportional to the amount present in the
supernatant, VC.; i.e., the more material present the faster the rate of
disappearance. This is referred to as first-order kinetics and is often
substantiated by natural systems. The proportionality constant, kg,
will vary primarily with reaction species and temperature.
Hue first-order reaction and the batch-reaction constant, kg, can be
interpreted as the lumped summation of the various anaerobic pathways
by which long-chain organic compounds characteristic of raw swine
waste are degraded. This breakdown can be indicated by a change in
COD, the oxygen equivalent required to convert a given carbon compound
to complete stabilization end products of carbon dioxide and water.
In general a decrease in COD would indicate a conversion to more
oxidized and shorter-chain compounds, but it does not necessarily
indicate a loss of organic carbon from the system. The total organic
carbon (TOC) does directly represent carbon content, excluding in-
organic forms and thus changes in the TOC do reflect actual system
losses, primarily as carbon dioxide and methane. If organic carbon
converted to carbon dioxide stayed in solution because the liquid was
unsaturated, there would be an apparent loss of total organic carbon
but actually only a change in chemical component form. However, the
166
-------
amount of inorganic carbon dissolved and measured in all lagoon samples
was small compared to the organic carbon concentration indicating that
TOG reductions actually indicated supernatant carbon losses. Therefore,
both TOG and COD are useful parameters which change value much the
same as the concentration of the actual chemical reactants participating
in the ongoing anaerobic reactions.
Equation 4 was integrated under steady operation conditions to yield
the change in supernatant concentration with respect to time between
one loading event and another. The integrated form was:
C± = C9 expC-kgt) (5)
where t varied from zero immediately after loading up to the time of
the next loading event; e.g., one week. The initial concentration,
C.£, was calculated from a mass balance immediately after loading. The
concentration of species i just prior to loading, the volume and con-
centration of the raw waste input, and the volume of the reactor were
the needed parameters. Any liquid volume input which was not matched
by evaporative losses was drained prior to loading to maintain a
constant volume over long periods of operation. This mass loss
equaled the effluent volume, E, times the supernatant concentration,
C|, with E ranging from zero to the feed volume. The effluent concen-
tration was quite small compared to the input so without a large error
the effluent and input volumes were set equal. Thus, the initial
concentration of species i was:
VG*: H
i
- a,
fr-
CiF
V
r,O _ f,\
C± -- - - (6)
where Ce = supernatant concentration of species i at the end of
the reaction period just prior to waste loading, mg/1
C. = concentration of species i in raw waste input, mg/1
F = waste input and effluent volume, liters
a. = net fraction of waste input settled
Consider first the case where the loading frequency was constant; e.g.,
once per week. During steady operation with constant waste input,
167
-------
concentration and volume, the initial concentration (C.^) and final
concentration (C?) would remain the same from batch to batch. To
evaluate kg under these conditions, Equation 6 was substituted into
Equation 5 using t = 1 week so that C± = c|. Additionally, g± = l-a±
or represents the fraction of waste input remaining in the supernatant,
The resulting equation was:
Ci =[Ci + BiCi ~ Ci f
W6ek)
(7)
where &. = 1-a.
after rearranging Equation 7 becomes
exp(k_ " 1 week) = 1 +
B
exp(k ' 1 week) - 1 =
B
B
P
iCi-
Cl
rf
i i
Ci F
V
re
Li F
Ce V
(8)
(9)
The value of the batch reaction rate, kg, was calculated for TOG and COD
for laboratory reactors (14-1) loaded at several input rates during the
steady-state operational period, Table 34.
Table 34. FINAL CONCENTRATION AND BATCH REACTION RATE CONSTANT FOR
LABORATORY ANAEROBIC SWINE REACTORS (14 1)
Fraction or
multiple of
reference rate
4.8
1.2
.6
.3
Average
ce
TOC
2,100
400
275
100
mg/1
COD
4,000
L,000
500
350
TOC
.50
.67
.53
.48
.54
weeks"
COD
.69
.74
.74
.59
.69
168
-------
The constants, kg, showed relatively little change over the sixteen-fold
range of loading rate. The laboratory average ambient temperature was
220 c + 1.5°C.
From these data it was concluded that the various loading rates in
themselves did not affect the microbial activity in these small reactor
units. A change in reaction constant could indicate some toxic effect
or for reactors on the border of anaerobic and aerobic conditions, a
change in microbial populations, metabolic kinetics, and end products.
Therefore, with kg constant the implicit dependence of effluent supernatant
concentration on loading rate was:
- C| v[-l + expCkg ' 1 week)]
(10)
Because the raw swine waste was very concentrated, C| « 3^0, and thus:
eCF
C =
i = V[-l + exp(kB ' 1 week)]
The demoninator was constant for a given species as was C. for a steady-
state waste input so that the effluent TOC or COD concentration was
linearly dependent on the feed volume which agreed with data, Figure 70,
for the laboratory swine units. The effluent concentration for another
waste type under similar reactor conditions can be calculated by
equation 11 if the reaction constant (kg) is known.
The batch-loading model describes the actual weekly operation of these
anaerobic lagoon experiments. It has some capability to predict the
effect of aperiodic loading because the time between inputs can be
changed in a straightforward manner in Equation 5. Additionally,
variable input volume and concentration can be taken into account.
There were, however, several disadvantages in this batch-loading
approach. The first limitation was that the net settling effect had
to be included in a circumlocutory manner by means of the hypothetical
initial phase separation condition, a^Ci. A second disadvantage
was that transient conditions which had an impact for longer than one
loading period such as a change in loading concentration or loading rate
and changes in volume associated with large lagoon drawdowns, could only
169
-------
l,200r
t>o
2
o
M
H
H
5S
W
O
u
O
PQ
O
O
H
P
a
w
Q
w
o
X!
O
w
K
O
1,000
800
600
400
200
WEEKLY
FLUCTUATIONS
ABOUT MEAN
REFERENCE RATE
I
I
100 200 300 400
LOADING RATE, ml/week
Figure 70. Steady-state supernatant COD and TOC concen-
trations for laboratory reactors (14 liters)
loaded once per week with swine waste (44,000
mg COD/1, 15,000 mg TOC/1).
170
-------
be determined by a number of successive weekly calculations. The
transient condition resulting from a changed loading volume depicted
in Figure 69 is descriptive of the successive calculations needed. This
latter limitation prevented rapid prediction of steady-state conditions
associated with management options in lagoon operation. Some of these
disadvantages could be overcome with a continuous-stirred tank reactor
modeling approach.
CONTINUOUS LOADED APPROACH - ORGANICS
The development of a continuous model required one assumption in addition
to the two primary assumptions made earlier. This assumption was that
when the overall month by month operation was considered, the overall
loading rate was simply (C? • Q) kg per week (where Q is the waste input
in 1/wk), regardless of the actual time span of loading. In other words,
over months of operation there was little difference in how or over what
period of time the loading occurred as long as the time period between
loading was short in comparison to the time period to be modeled. This
assumption was substantiated by the laboratory reactor data discussed
earlier showing little difference in supernatant concentration over a
wide range of batch loading frequencies, Figures 25 and 26. The slight
supernatant concentration difference observed in units which actually
were loaded continuously would simply change some of the equation rate
constants but not affect the general model.
The continuous loading equation describing the supernatant of an anaerobic
reactor included parameter accumulation, inflow, outflow, removal via
settling, and removal or generation by microbial or chemical reaction.
The inflow, outflow, and accumulation were standard terms. The initial
settling removal was assumed to be rapid in comparison to the overall
time available for reaction and accumulation. This appeared valid
because several workers including Jett _e_t al.80 found settling of various
raw swine waste constituents was essentially complete in one hour, and
investigated loading frequencies were on the order of once per week.
It was further assumed that for each species i, of a given waste
stream, the amount settled was a constant fraction (a.) of the raw
waste input. If the microbial populations had acclimated, then the
net settled fraction could be used; i.e., initially settled component
minus the amount remineralized from the sludge to the supernatant.
The constant net settling assumption appears to be verified by the
uniform amount of raw waste parameters found in sludge over various
lengths of storage.
Component removal via reation was assumed to follow first order kinetics
as with the previous model; i.e., the reaction occurred at a rate propor-
tional to the amount of species i present. Correspondingly, the chemical
171
-------
oxygen demand and total organic carbon were the parameters used to
reflect the stabilization and loss of organic material and were assumed
to be removed at a first-order rate.
A mass balance for species i under unsteady conditions for a uniformly
mixed reactor with balanced inflow and outflow yields:
dC. , f
v —i- = QC - QC - k C V - o.QC. (12)
dt ii ii 11
with C. = C. when t = o
where C. = reactor supernatant concentration of species i, mg/1
C? = supernatant concentration at the start of the modeling
period, (t = o) , mg/1
Q = flowrate, liter /week
k. = reaction or removal constant, week
The stabilization or reaction constant k± would be dependent primarily on
the parameter being considered as well as the temperature and type of
waste. The net fraction settled, a., was as defined above in the
discussion of sludge accumulation.
More sophistication could have been included in Equation 12 by adding
terms which represent additional loss or generation mechanisms. For
example, the sludge-supernatant transfer could be added as a pathway
for increasing the amount of material entering the supernatant between
batch loading events to more closely describe actual lagoon operation.
To do this, the initial rather than net fraction settled would be
represented as ct^. This supernatant addition from the sludge was not
included because it involved the evaluation of two unknowns, the initial
settling and the sludge-supernatant interfacial exchange rates. Overall,
the net settling and the coefficient a. take both of these phonemena
into consideration. Evaporation or rainfall volume change effects on
concentration were not included because of the small losses or gains
found as compared to the raw waste inputs. More detailed reaction
models reflecting the microbial nature of waste stabilization (e.g.,
Michaeles-Menten kinetics) were not used because routine monitoring did
172
-------
not include microbial counts and the complexity of such a model would
exceed control possibilities for the investigated swine lagoons.
Certain anaerobic lagoon performance characteristics can be evaluated
by considering the limiting cases of Equation 12. The first such case
involved the variation of concentration with time for lagoons if the
lagoon was simply a body of water used for waste storage and there were
no microbial or chemical removal mechanisms. In colder regions there
are periods when storage is in fact the only lagoon function because
anaerobic activity and liquid volatilization are essentially zero.
If a lagoon were serving only for storage, either because low temperature,
some toxic material or shock waste input have prevented microbial
activity, then this implies that the reaction rate constant k^ is zero.
Assuming no microbial activity, initial settling would still take place,
but there would be no further sludge-supernatant transfer; thus, a-
would represent both the initial and total fraction settled. Integrating
equation 12 under these assumptions and subject to a constant initial
condition yields:
Ci = BiCi + C ~ fcexpC-Qt/V) (13)
where 8. = 1 - a.
Several consequences of this solution were noteworthy. The first was
the time required to attain steady-state supernatant concentration,
defined as the time required to reach 99 percent of the steady-state
concentration. The second term on the right in Equation 13 decreased
exponentially as time increased; hence it was a transient term. Initial
concentration in these studies was zero or very low when compared with
the raw waste input. The time at which the supernatant concentration,
C±, was 99 percent of the steady-state, &±C±, was calculable by solving
for t:
99 R C — 8 C
1 1 f X 1 = exp(-Qt/V) (14)
.01 = exp(-Qt/V)
4.6 ? = t (15)
173
-------
At the reference loading rate (2,300 1 reactor volume/45-kg hog) and a
swine waste volume of 7.5 l/day/45 kg hog, solving for t yields:
t = 2,300 I/hog . 4>6 = 1;41D days~ 3.9 years
7.51/hog/day
Thus, a long transient or start-up period of almost four years would be
expected for a storage-only lagoon. Correspondingly, if the waste input
concentration was held constant for lagoons loaded at different rates
over the time periods necessary to achieve steady-state, then this
transient or start-up time should be inversely proportional to the
loading rate, Q, Equation 15. These times would range from 3.8 to
122 years as the loading varied from the reference to one thirty-second
of the reference rate used in this study. Comparing these deductions
with the pilot scale lagoon data, Figures 41 to 58, it follows that
transient times are from 10 to 1000 times more rapid than could be
accounted for by settling alone. Also, for most of the loading rates
investigated the transient times were the same; i.e., independent
rather than inversely dependent on the loading rate.
A second consequence of Equation 13 is that at steady-state, with the
transient term insignificant, the supernatant concentration is:
where C?s = steady-state concentration of species i, mg/1
The ultimate or net fraction of COD settled in the laboratory studies
was about 25 percent, so the steady supernatant concentration should be
(1-.25) ' C * or 30,000 mg/1 for a feed strength of 40,000 mg COD/1.
Thus, all the units loaded at different rates should attain the same
steady-state supernatant concentration, equation 16. Supernatant
concentrations from the laboratory and field reactor data were much
lower than 30,000 mg COD/1 and evidenced a decrease in concentration
as the loading rate was decreased.
Differences between these experimental results and the predicted values
of transient times and steady-state supernatant concentrations for storage
lagoons indicated that much more was involved in an actively functioning
animal waste lagoon than included in postulated models; and thus, another
loss mechanism must be the significant factor in lagoon performance. Also
174
-------
if the COD supernatant concentration for a swine waste lagoon was
above 30,000 mg/1, then it could be concluded that the lagoon was
not functioning biologically. This then could be used as a criterion
for lagoon failure based on supernatant COD concentration.
Integration of Equation 12 developed earlier with a first-order reaction
loss term and a constant initial condition, C? yielded:
A f n O
x Q ft*- i / /•»*•' w or
,• ~ ~z—r~~;— p,-1-1-: + v'-'- ~ ~z—r~~,— p.-1-
Q + k±
where k. = k. • V, liters/week
The reaction rate constant l^is specific for the species monitored, i.
Two undetermined constants g^, the settling component, and k^ the first-
order removal component existed. Since gj_ did not appear in the time
dependent decay term, exp [-(Q + k.)t/v], then data for both the transient
response and the steady-state operation would be needed to determine
8^ and k^.
The solution for the supernatant concentration, C^, was composed of a
steady-state and transient term, the first and second terms on the
right of Equation 17, respectively. The transient term was an exponen-
tial decay function where the rate of decay was given by the term
(Q + k^)t/V. Expanding and rearranging the exponential power gives:
The reaction constant k. has units of reciprocal time and can be
visualized as the inverse of a characteristic time for the reaction
involving the gain or loss of species i. The exponential power was
thus:
r rxn
where T = V/Q = bulk fluid residence time, weeks
175
-------
T = 1/k!, weeks
rxn i
The larger this exponential power expression in Equation 19, the more
rapidly the system approaches steady-state. As shown with the storage
model, the ratio of real time to the characteristic residence time
(t/T ) was a small term since Tr was on the order of years for the
investigated swine lagoons. Therefore, it was expected and subse-
quently verified that the ratio of real time to characteristic reaction
time (t/Trxn) was large since the total time from start-up to steady-
state was about 10-20 weeks for the investigated anaerobic lagoons
or reactors, Figures 4-17. Physically, the interpretation of the
exponential constants is that the microbial reactions take place over
short intervals, especially in comparison with common lagoon residence
times.
The steady-state supernatant concentration derived from Equation 17
was related to the input concentration as follows:
<»>
The predicted steady-state supernatant concentration was dependent on
the loading rate with a lower concentration expected at the lower
loading rate. A parametric graph of this relation is given in Figure
71 for constant B^C? and increasing values of k^.
A plot of laboratory reactor (14 1) COD supernatant values versus
loading rate, Figure 70, indicated a nearly linear relationship. The
value of k-^ was calculated from Equation 20, Tables 35 and 36, after
first determining B.J_ which is equal to 1 - a^ where a^ is the net
fraction settled or remaining in the sludge. From the laboratory
Imhoff cones and cylindrical 14 1 reactors, the ultimate sludge
accumulation of the total waste input for the 10-30 week periods of
this experiment was about 30 percent for COD and 25 percent for TOC.
Thus:
"COD " '" "
3TOC = 1 - .25 = .75
176
-------
1.0
w
P-.
a
o
H
Z «H
W
U
z o
o ^
o en
w
O
o
M
H
Z
w
U
Z
O
W
52
H
W
I
H
CO
O
O
H
U
.8 -
.6 -
.4
.2
k± = 375 1/wk
375
1,500
Figure 71.
750 1,125
LOADING RATE, I/week
Parametric plot of Equation (20) showing linear
dependence of steady-state concentration on loading
rate at high reaction rate constants.
1,875
-------
With these values the reaction rate constants, k,, were calculated,
Tables 35 and 36. The units of k^ are liters per week which when
compared to the input loading rate, Q, shows that Q « k^. With this
inequality equation 20 simplifies to:
QCf
Css =r2- 6 Cf = -2- g.Cf = -4 (21)
so that the ratio of k./3i which is still a constant, can be redefined
as K.j_ and is thus the only unknown to be calculated. Because &± was
determined from separate measurements in these experiments, there was no
inherent advantage in using either K^ or k.^, but for other situations the
use of KJ_ should be more advantageous. Both constants are listed in
Tables 35 and 36.
From the comparison of the rate constants for chemical oxygen demand
and the total organic carbon, both representative of lagoon organic
matter, it was seen that indeed the reaction rate constants were very
similar, Tables 35 and 36. This reemphasized the fact that either
parameter could be substituted for the other in a monitoring or modeling
scheme. The parameter chosen then would be dictated by the analytical
facilities available.
The very high reaction rate constant in comparison to the input volume
emphasized that reaction times are small compared to lagoon residence
times. Thus, there were large reductions in raw waste concentrations.
Under these conditions the error in 3^, when 3^ is near 1.0, was not
great. Calculations showed that when g- varied between 0.7 to 1.0,
the predicted ratio of effluent concentration to influent concentration
ranged from 97.0 to 95.7 percent. Therefore, the net amount settled
with raw swine waste was so low that the effect was insignificant. With
another waste type yielding a greater settled fraction, the errors in
measuring ct^ may be sufficiently large that the use of the lumped
parameter K = k^/3 would be warranted over the coefficient k.
Values of the reaction rate constant for laboratory units should be
capable of scale-up into a predictive equation for the pilot- scale field
units because the raw waste input was the same for both experiments.
However, a correction for temperature must be included. The traditional
relationship of doubled biological activity or microbial reaction rate
with a 10° rise in temperature was used because kinetic data at several
temperatures were not taken. The laboratory units were controlled at
26 C ± 1.5° C while the field lagpon supernatant values were evaluated
during February to April, 1974, when the average lagoon temperature was
178
-------
Table 35. FIRST-ORDER REACTION RATE CONSTANTS FOR CHEMICAL OXYGEN DEMAND (COD) IN ANAEROBIC
SWINE LAGOON UNITS
Fraction of
Reference
Loading Rate
1
.5
,25
.125
.0625
.031
kCOD
Laboratory (14
• (B-.7)
COD
equation equation
(20), (21),
1/wk week"!
8.6
8.0
7.0
.61
.57
,50
1)
KCOD
equation
(21),
1/wk
12.2
11.4
10.0
Pilot scale
kCOD
equation
(20),
1/wk
6,000
3,830
2,400
2,250
2,000
2,000
• (B-.
KCOD
equation
(21),
week"1
.34
.22
,14
.13
.11
.11
87)
if
COD
equation
(21),
1/wk
6,638
4,295
2,733
~2,538
2,148
2,148
Extrapolated
from laboratory
to field units
' (B-.87), 11° C
kCOD
equation
(21),
week~l
.27
.26
.22
-------
Table 36. FIRST-ORDER REACTION RATE CONSTANTS FOR TOTAL ORGANIC CARBON (TOG) IN ANAEROBIC SWINE
LAGOON UNITS
Fraction of
Reference
Loading Rate
1.0
.5
.25
.125
.0625
.031
Laboratory
"roc >4oc"--75>
equation equation
(20), (21)
1/wk week 1
9.0 .64
7.8 .55
6.5 .47
KTOC
equation
(21),
1/wk
11.9
10.3
8.8
Pilot scale
^TOC
equation
(20),
1/wk
4,440
3,600
2,500
2,200
1,900
1,600
^-.84)
equation
(21).
week"1
.25
.20
.14
.12
.11
.094
KTOC
equation
(21),
1/wk
5,050
4,040
2,830
2,430
2,220
1,900
Extrapolated
from laboratory
to field units
k^QC(6=.84) 11° C
equation
(21).
.26
.22
.19
00
o
-------
11 C ± 5.5 C, Appendix Bl. This temperature differential translated
into a difference in rate factor of 2.8 between laboratory and field
units. The reaction constants k± and K± calculated from the laboratory,
laboratory adjusted to field temperature, and the measured pilot-scale
field conditions are given in Tables 35 and 36. The predicted and actual
supernatant concentrations of the field units as a function of the
weekly loading rate are given in Figures 72 and 73. The predicted
concentrations were too high indicating that the field units were
slightly more efficient than the laboratory units when put on the same
temperature basis. However, in light of the large number of assumptions
about kinetic temperature dependence, settling values, and validity
of a direct scale-up approach, the agreement between experiments was
not unreasonable.
The value of the kinetic rate constants for the field units was less
consistent among the various loading rates than the laboratory data,
especially when comparing the higher and the lower loading extremes.
This inconsistency emphasized that the model was only an approximate
approach and that the actual lagoon functioning was more complex.
Another laboratory-field difference was the fraction settled, a^. After
about 120 weeks the three pilot-scale field units measured for sludge
depths, Table 25, had approximately 12 percent of the input volume which
remained settled whereas the laboratory units tested had a value of about
25 to 30 percent. The reasons for these differences are not known at
this time, but one factor may be the longer duration of the field units.
Comparison of the rate constants derived from the laboratory and the
field units, Tables 35 and 36, indicated that the field data were less
consistent over the range of loading rates than the laboratory units.
Only the field units with the lower loading rates of .25 to .031 times
the reference had rather similar and consistent rate constants. If the
average removal rate from these lower loading inputs was used to predict
the COD and TOC supernatant concentration, Figure 72, then the measured
concentrations would be lower than predicted. This may indicate some
greater microbial response at higher loading rates which possibly could
be predicted with more detailed biochemical data and a more sophisticated
model. The predicted supernatant concentrations for the various average
k-£ values measured are also shown in Figures 72 and 73.
On an overall basis, the use of an average reaction rate constant as
determined in the laboratory (and temperature corrected) or from a
field experiment at a single loading rate gave a good approximate
order-of-magnitude value for the supernatant concentration of organic
matter as represented by COD and TOC. The prediction of supernatant
TOC concentration was better than that for COD for reasons not yet
fulfyexplained. The extensive range of oxidation states for organic
181
-------
, 000
60
e
C/3
3
O
H
W
U
53
O
O
W
Q
a
W
O
O
PC
PREDICTED FROM AVERAGE
it' .. OF PILOT SCALE DATA
2,400
1,800
1,200
600
PREDICTED FROM
LABORATORY k.!
CORRECTED FO
TEMPERATURE
ACTUAL PILOT SCALE
FIELD DATA
1
I
20
PREDICTED FROM
UNCORRECTED
LABORATORY k'
I
I
80
100
120
Figure 72,
40 60
LOADING RATE, gal/week
Steady-state supernatant COD concentrations for pilot-scale swine
lagoons receiving various loading rates - experimental and
predicted results (Equation 21).
-------
00
OJ
O
M
w
u
O
H
a
o
o
Pi
o
O
H
1,500
PREDICTED FROM AVERAGE
PILOT SCALE DATA
1,200
900
600
300
ACTUAL PILOT SCALE
FIELD DATA
PREDICTED FROM
LABORATORY k'
CORRECTED FOR
TEMPERATURE
I
PREDICTED FROM
UNCORRECTED
LABORATORY k'
J. v*\j
I
20
100
120
Figure 73.
40 60 80
LOADING RATE, gal/week
Steady-state supernatant TOC concentration for pilot-scale swine
lagoons receiving various loading rates-experimental and
predicted results (Equation 21).
-------
compounds before total removal from the system as CC^ may be an impor-
tant reason. Despite these limitations, it should be possible with
other waste types to do a single experiment at one loading rate
and to then determine the supernatant response to changes in input
waste concentration and rate.
CONTINUOUS LOADED APPROACH - NITROGEN
The same modeling mass balance approach was used with total Kjeldahl
nitrogen as with the organic parameters (COD and TOC). The discussion
of the batch loading model would be equally as valid and contain the
same disadvantages as that described for COD and TOC, so that analysis
was not redeveloped.
The major loss mechanism for nitrogen is surface volatilization of
ammonia, which is assumed to obey the standard relationships derived
for other ammonia solutions?^ The rate of ammonia loss is proportional
to the difference in gas phase and liquid phase concentrations when both
are expressed on the same basis. This common basis is achieved by
multiplying the gas phase concentration by the gas-liquid solubility
coefficient (Henry's law coefficient). The volatilization expression
then is:
R^ (Ci - h±Pi) (22)
where R = rate of volatilization, mg or g moles per hour
h.^ = Henry's law coefficient, mg/l/atm
C. = liquid concentration of species, i, mg/1
P£ = partial pressure of species i in the gas phase, atm
It was assumed that under most swine lagoon field conditions the gas
phase concentration times the Henry's law constant was quite low so
that the driving force for ammonia loss was approximately equal to the
liquid phase concentration, C^. This assumption was partially cor-
roborated by field data of Miner^2. Deviations from this assumption
would generally lead to a lower rate of volatilization because the
driving force would be lowered.
The expression for nitrogen loss was thus proportional to the driving
force, C±, and to the interfacial area between lagoon and atmosphere,
184
-------
A, so the overall mass balance yielded:
dC
V -T-i = OC - OC - h Ar - rvnrif (23)
(24)
where bi = proportionally constant for ammonia volatilization,
cm/week
A = surface area, cm^
H^ = overall mass transfer coefficient, cnr/week
The ammonia loss was also a first-order term with the rate of loss propor-
tional to the concentration present in the supernatant and mathematically
was comparable to Equation 12.
From a mechanistic approach there were several additional assumptions
included in Equation 24. First, the parameter chosen to be modeled was
total Kjeldahl nitrogen (TKN) instead of ammonia. While nitrogen
losses from the lagoon system were primarily as ammonia, there were also
gains or losses of ammonia which occurred by microbial breakdown of long
chain molecules and deamination or incorporation into cell mass by
synthesis. Thus, using ammonia would involve several other source and
disappearance terms. Because the swine raw waste and lagoon supernatant
contained a high fraction of ammonia and were uniform, it was assumed
that anaerobic deamination and the mixing-diffusional process were not
rate-limiting. The liquid to gas transfer of ammonia was assumed to be
the rate-limiting step; hence, only the volatilization term was included
in Equation 24. This assumption, although very reasonable for swine
waste, warrants further verification when another waste type is involved.
Total Kjeldahl nitrogen losses represented predominantly ammonia losses
so the use of either parameter would reflect total system losses.
Volatilization of organic nitrogen compounds occurred but were insigni-
ficant when compared to volatilization of the smaller ammonia molecule.
Conversions between organic and ammonia nitrogen would not appear as
a TKN loss. Furthermore, it had been found that preservation of samples
for ammonia analysis was much more difficult than for TKN, again because
185
-------
of microbial interconversions. For these reasons the total Kjeldahl
nitrogen was used to model nitrogen losses from anaerobic swine lagoons,
The supernatant TKN concentration derived from integrating Equation
24
was:
TKN Q + H
TKN
QBCf - (Q + H^C^ ^ [_(Q + HTKN)t/V](25)
Q + H
1TKN
The steady-state concentration was given by:
-,ss _
"TKN
Q3C1
QgC1
Q + HTKN Q + b
(26)
TKN
Steady-state results from the laboratory (14 1) and pilot-scale reactors
were used to calculate the overall mass transfer coefficient E
TKN
from
Equation 26, Table 37. The overall transfer constant HTKN was the
product of the actual mass transfer coefficient, bTIQj, and the reactor
surface area, A. Therefore, dividing HTKN by the respective areas of
laboratory and pilot-scale reactors yields the coefficient b^KN' Table
37.
Table 37. MASS TRANSFER COEFFICIENTS FOR TOTAL KJELDAHL NITROGEN (TKN)
AS LOST FROM ANAEROBIC SWINE LAGOON UNITS, EQUATION 26
Fraction of
reference
loading rate
1.0
.5
.25
.125
.0625
.031
Average
Laboratory
HXKN
I/week
2.8
2.3
1.9
bTKN
cm/week
5.2
4.4
3.5
4.4
Pilot Scale
HTKN
I/week
520
650
690
780
930
1,060
bTKN
cm/week
5.5
6.9
7.3
8.3
9.9
11.2
8.2
186
-------
Predicted supernatant TKN concnetrations based upon uncorrected laboratory
and average pilot-scale data for bTRN are shown in Figure 74. Predicted
values based upon uncorrected laboratory data were 30 to 50 percent above
the measured values. Differences in temperature, wind conditions, and
reactor size contributed to the deviations between measured and predicted
values. The laboratory reactors had a sevenfold larger surface area to
lagoon volume ratio than the pilot scale units and yet the mass transfer
coefficients were fairly similar. The dependence of loading rate on
nitrogen supernatant concentration for the pilot-scale field reactors
more closely followed values predicted by field-scale data than
predicted from uncorrected laboratory data, Figure 74.
The internal consistency of first-order TKN mass transfer coefficients
was better than the first-order reaction constants found for COD and TOG.
The k values for COD and TOC varied by a factor of three while the HTKN
varied by less than a factor of two over the same loading rate range,
.031 to 1.0 times the reference rate. This consistency may be partially
attributable to the physical mechanism for TKN removal (ammonia vola-
tilization) as opposed to the microbially dependent loss of organics.
Analyses for COD and TOC verified that the reaction constant k. was much
greater than the loading rate Q and thus predictive Equation 21 could be
simplified because kj » Q. However for TKN, the overall mass transfer
coefficient H^-,^ was of the same magnitude as Q so no further simpli-
fication was possible. Mechanistically, the relative magnitude of
ETKN and Q was related to lagoon response time. Writing the exponent
part of the transient term (as done earlier in Equation 18 for COD and
TOC) yielded:
£ HTKN _1_ bTKNA
V + V = T + V (27)
The second term, the inverse characteristic mass transfer time for total
Kjeldahl nitrogen was the same magnitude as the inverse residence time.
The inverse residence time as shown earlier in the COD discussion was
relatively small. Thus, a slower supernatant TKN response to some
change in the anaerobic swine lagoon operation such as increased input
waste concentration would be expected. The discussion of the results
for laboratory reactors and Imhoff cones pointed out that indeed a
slower TKN response was observed as compared to COD and TOC.
Further examination of Equations 26 and 27 showed the relationships which
exist among surface area, volume, and supernatant nitrogen concentration
for swine lagoons. The response time, to achieve steady-state after an
operational change, was affected by the volume, Equation 27, as the
187
-------
l,500r-
00
OO
bO
E
O
M
H
H
55
W
U
a
o
S3
W
O
H
M
S3
hJ
9
w
o
H
1,200
900
PREDICTED FROM
UNCORRECTED
LABORATORY
bTKN
600
300
ACTUAL PILOT
SCALE FIELD
DATA
PREDICTED FROM AVERAGE
TKN
PILOT SCALE DATA
1
20
80
100
120
Figure 74.
40 60
LOADING RATE, gal/week
Steady-state supernatant TKN concentration for pilot-scale swine lagoons
receiving various loading rates-experimental and predicted results (Equation 26)
-------
ratio of surface area to volume. Thus, the larger the area to volume
ratio, the faster the approach to steady-state conditions. The absolute
value of the effluent nitrogen level, Cffjg, was determined by the surface
area, independent of volume, Equation 26. Larger surface area lead
to lower supernatant TKN concentrations which indicated the surface
volatilization mechanism as opposed to a bulk reaction removal. This
areal removal dependence for TKN was contrasted to the microbial based
volumetric dependence for COD and TOG, Equation 21.
In review, results of the continuous loading model for nitrogen
appeared to give reasonable agreement to measured values when labora-
tory data or data from a single field unit were used to predict the
effects of lagoon loading rate, climatological changes, and raw waste
concentration. A few of the assumptions used in the model should be
investigated further to provide a more precise predictive capability.
The use of volatilization as the limiting step in conversion of organic
nitrogen to ammonia which was lost from the system and the effect of
loading rate on surface properties which may restrict volatilization
should be reconsidered in future work.
189
-------
SECTION VII
LAND APPLICATION STUDIES
EXPERIMENTAL PROCEDURES
Results of lagoon studies for swine waste indicated that while these
reactors were efficient in terms of percent removal, effluent quality was
not sufficient to allow stream discharge. Thus, lagoons must be consid-
ered as only pretreatment-storage units. The terminal process which
provides for the most economically available achievement of the national
goal of zero discharge for the animal production industry is the plant-
QO
soil receiver system?0
In order to evaluate the various facets of a lajoon pretreatment - land
application system, a field-scale study was undertaken at the North
Carolina State University Boar Testing facility described in detail
earlier, Figure 75. The overall objectives were to investigate the
effect of wastewater applications on crop quality, runoff potential,
soil accumulation and soil-water migration, and to demonstrate the
implementation of pretreatment - land application system for the swine
industry.
Site Description
The total system included the Boar Testing Station buildings and a two-
lagoon series. In order to reasonably obtain the full range of nitrogen
applications desired, the effluent from the first-stage lagoon was used
as the soil-applied liquid.
Prior to this study, the application site had been extensively used
for forage crops. In May, 1972, the entire field was fumigated with
methyl bromide. The field site sloped at a uniform 1-3 percent grade
from the farm road toward the woods, an east-to-west orientation.
Thus, the plots and monitoring devices were located in this direction.
There was negligible north-south slope, thus limiting substantially
the runoff and soil water flows to one direction, west to east.
An interception drainline to cutoff soil water interflow from above
the study site was installed at the 1- to 1.5-m depth. This line
190
-------
SECONDARY
LAGOON
rS
PRIMARY
LAGOON ^
»RIGATION PU
MANURE INPUT TO
•FULL-SCALE SERIES
LAGOON SYSTEM
P AUTOMATIC
.
i
E
CM
CM
E
fli
op
CM
1
1
1^
1
1
1
1
1
1
1
1
i-
i
I
I
1
i
*
1
1
1
I
1
^IRRIGATION
CONTROLLER
LTILE DRAIN .XV DIRECTION
i C °F
IRRIGATION PIPE FLOW
t^
AREA OF WELLS
AND POROUS CUPS SURFACE RUNOFF
N •
^
//COLLECTION BARRELS
O O CT_^ RUNOFF SUBPLOT O O O O
._)•* ---PLOT FLANGES
NOZZLES
REP #3 V|L7 DRA1N
ATYPICAL LOCATION
f OF POROUS CUPS - A
0
CM
^- REP* 2
o
kTING
B
K>
CM
\CKQf
uu
0
-------
sloped toward the southwest corner and then down the side of the study
area, Figure 75. In order to divide the study area into two sections
to increase the number of plots, an interception drainline was put
across the field at 35 m in the downslope direction. Each drainline
trench was back filled above the plastic drain tile with number 7
crushed stone.
An experimental plan was established to investigate three plot loading
rates representing approximately 336,672, and 1,344 kg of nitrogen per
hamper growing season, with three replicates of each loading rate.
Because all wastewater was irrigated from the same lagoon, these rates
meant also that there was a 1:2:4 ratio of liquid volume applied
with these different nitrogen rates. The field was divided into three
areas or replicates of three plots each (plots 1, 2 and 3; 4, 5 and 6;
7, 8 and 9). Within each replicate, the plots were assigned loading
rates on a random selection basis (1, 4 and 7 high rate; 3, 5 and 9
medium rate; 2, 6 and 8 low rate). Each 9.24-m by 9.24-m plot was
isolated from surrounding area runoff by 23-cm wide galvanized flashing
buried 13 cm deep giving a 10-cm above-ground isolation barrier.
To achieve uniform effluent distribution, the irrigation nozzles were
located at the corner of each plot and were operated as one-quarter
turn sprinklers having a radius equal to the plot dimension, 9.24 cm.
The wetted area extended outside the 9.24-x 9.24-m area by approximately
0.6 m to assure total plot coverage. The irrigation cycle was adjusted
so that the total amount of a given parameter applied to the 9.24-x
9.24-m area was equal to the experimental plan. The various applica-
tion rates were achieved by varying the irrigation time.
Irrigation equipment included a 3. 8 cm main header from the lagoon, a
2 horsepower centrifugal pump (Stayrite DHHG), laterals to each plot,
electric control valves to regulate flow to sprinkler nozzles, and a
timer controller. The layout for each plot was identical and is shown
in Figure 76.
Experiment and Monitoring Description
The experimental site was prepared and sprigged to certified Coastal
Bermuda grass in the Spring of 1972 according to recommended practices.
The area was irrigated as needed after sprigging and during the summer
of 1972. A low rate of ammonium nitrate was applied after the sprigs
began to produce new shoots and thereafter during the summer at appro-
ximately 30-day intervals. The experimental area had almost complete
grass cover by fall and residue (10-15 cm) was left on the plots during
the winter to reduce possible loss of stands from winter injury.
Application of swine effluent, consistent with the three loading rates
based on nitrogen application, was initiated April, 1973, shortly after
192
-------
3.2cm PVC MAIN HEADER
tt 12 UNDERGROUND
WIRE TO
AUTOMATIC CONTROLLER
"I I
—OJ
PVC
2.5cm
HEADER"
ALUMINUM
FLASHING
AREA FOR
SOIL CORES
o o o
POROUS CUPS
,SPRINKLER NOZZLE'•
9.24 m RADIUS
23° TRAJECTORY
.32 cm NOZZLE
D
RUNOFF
SUBPLOT
O--
RUNOFF COLLECTION
BARREL
Figure 76. Schematic of experimental plots for land application
of swine lagoon effluent.
193
-------
the emergence of new growth. Irrigation was controlled by an automatic
timer and was applied once every seven days on the same day of the week
at night.
From the state maps and further testing, it was determined that the soil
was a Norfolk sandy loam (Typic Palendult: fine-loamy, siliceous,
thermic). This is characteristic of much of the Coastal Plains region of
North Carolina and the Southeast. The profile contained an A horizon of
loamy sand, with weak fine and medium granular structure material above
the B horizon of sandy clay loam with a weak medium angular blocky
structure, Figure 77. Hydrolojically, this resulted in a well-drained or
permeable layer (hydraulic conductivity, K = 6 - 9 cm/hr) on top of a
tight or less permeable zone (K = 0.8 - 2 cm/hr) (Lutz ).
Soil water and movement of effluent constituents were monitored by
sampling with 1 x 105 pascal ceramic cups. The sampling grid for each
plot included one cluster of cups at the plot center and two at the
downslope locations of 6.2 and 12.4 meters from the lower plot edge,
Figure 75. A cluster consisted of cups at 23 cm (plow sole layer in
A horizon); 56 - 71 cm for plots 1 through 6 and 30.5 - 46 cm for the
slightly more eroded plots 7 to 9 (A-B horizon interface), and about
91 cm (25.5 cm below the surface of the B horizon), Figure 77. In
addition to the in-place porous cups, three soil cores (1.6 cm diameter)
were taken from each plot in the general area depicted in Figure 76.
These were divided into 5-cm increments over the upper 25 cm and then
into 10-cm increments until the 75-cm depth, Figure 77. Core holes
were repacked with top soil and successive cores were tsken uD-slope
of previous samples. Control cores to evaluate soil changes were
taken in an area not receiving lagoon effluent about 23 m down-
slope from plots 2 and 3. This area was in Coastal Bermuda grass and
received a very low level of ammonium nitrate for crop survival. Each
week after irrigation the in-plot, soil-water cups were sampled.
At locations outside the plot, sampling was more correlated with rain-
fall events because of sampling difficulty under unsaturated conditions.
The soil cores were taken once every three months.
To evaluate irrigation losses and actual plot loading, samples were
collected at the ground level with 7.5~cm diameter beakers (400 ml)
for volume and constituent analysis. Two collectors were placed in
each plot and then composited for analysis. Difficulty arose when
rain fell between the start of an irrigation event and when the
samples were composited. This happened infrequently and thus such
samples were discarded. The amount of runoff from this type of plant-
soil system was measured from a sub-area within the 9.24 m x 9.24 m
study plot, Figure 76. This sub-area was 2.25 m wide and 4.5 m long
and, of course, received the same effluent application as the rest of
the plot. At the lower edge, a funnel and tubing arrangement channelled
the runoff into a submerged 55-gal (208 1) drum. After each rainfall event,
194
-------
PORUS CUPS FOR
1
I
20-30 cm
Ap HORIZON
TO PLOW
SOLE LAYER
~^~1
76cm
DEPTH OF
SOIL CORES
ri
•C
-
-------
the total runoff volume was determined, a sample taken of the resulting
liquid and the barrels pumped out with the liquid being dispersed just
below the runoff plots as would be the case for the runoff.
Each set of three plots receiving a given effluent rate was harvested
when forage reached the hay stage (30 - 35 cm of growth). Because the
nitrogen applied was a variable, the forage grew at different rates
for each treatment. Consequently, harvest dates varied with effluent
application rates.
Two swaths (0.5 x 8.3 m) were harvested at random from the 9.24-m by
9.24-m plots with a Jari mower leaving a 5-cm stubble. The forage
was weighed green and subsampled for dry matter and chemical determina-
tions. The subsamples were dried in a forced air oven and dry matter
yield was determined by multiplying green weight by the percentage dry
matter and then expressed as dry matter per unit area (kg/ha.). The
dried samples were ground through a screen with 1-mm openings and
stored in plastic bags until subjected to chemical analyses.
After sampling, the remaining forage from all replications of one
treatment was clipped and tied into small, loose bales. These bales of
fresh forage were placed into a bin type forced-draft dryer and held
at 65° C until dry. Thereafter, the forage was rebaled into approximate-
ly 18-kg bales and stored in an open shed for later use in a feeding
trial. Coastal Bermuda grass fertilized with 67 kg N/ha. of ammonia
nitrate (no swine lagoon effluent) was harvested in midsummer at
approximately the same height as the experimental plots, air cured,
and designated as the control treatment.
Sample Collection and Analyses
Sampling in regard to plant and soil effects of effluent application
to Coastal Bermuda grass consisted as follows: irrigated wastewater
was collected at ground level within the plot during each effluent
application in 400-ml beakers. All runoff resulting from effluent
application or rainfall was funneled into a submerged 55-gal. (208 1) drum
measured and samples taken for chemical analysis. Soil samples were
taken in 15-cm increments prior to initiation of the experiment and
every 3 months starting in September, 1973,in 5-cm increments to a 25-
cm depth and every 10 cm from 25 to 75 cm. Soil water samples from
one-bar ceramic cups were taken prior to effluent application and the
day following the weekly application of effluent. Lack of soil
moisture especially during late summer greatly limited the number of
solution samples outside of plots .and in plots where effluent was
•replied at low rates. Yield estimates of forage were obtained for each
'•invest and subsamples of forage were taken for dry matter and mineral
determinations and nutritive evaluations. Effluent, soil, runoff, and
196
-------
soil water samples were frozen until analyzed. Forage samples were dried
in a forced-air oven at 70° C and ground prior to analyses.
Effluent, runoff, soil water, soil, and forage samples were analyzed
for total N, P, K, Ca, Mg, Na, Cl, and Cu. In addition, effluent was
analyzed for COD, TOG, pH, N03, NH3, Fe, Mn, and soils for pH, N03,
NH3, Fe, Zn, and Mn, and initially for exchangeable Al, organic matter,
cation exchange capacity, percent clay; forage for Mn, Zn, and Fe;
runoff for pH, COD, TOC, NIL,, and NO-,; and soil water samples for pH,
TOG, N03, NH3, and COD
After routine analysis for COD, TOC, pH, N03, and NH-, several
different methods of sample preparation were compared for analysis of
runoff, effluent, and soil water samples. A large portion of the ele-
ments, especially in the effluent samples, were lost by filtering or
centrifuging. Results from dry ashing of the samples also proved
to be more variable than determinations carried out on the original
sample after adequate dilution. Hence, all results of runoff, soil
water, and effluent samples were from unaltered samples.
After extraction with water for Cl and N03 determinations, soil samples
were air dried and extracted with 0.05 N HC1, 0.025 N tUSO^ for
exchangeable K, Ca, Mg, Na, Cu, Fe, Mn, NH3, and P. Soil pH was
determined on a 1:1 soil-water ratio. After extraction with weak acid,
an aliquot was taken for determination of NEU and the remaining liquid
was dry ashed, concentrated HC1 was added, redried, and resuspended
with 0.5 N HC1. On all soil samples, N was determined by Kjeldahl;
K and Na by flame photometer; Ca, Mg, Cu, Mn, Zn, and Fe by atomic
absorption; Cl by a chlorimeter; N03 by direct scan on UV spectro-
photometer, NH» by NH~ electrode and P by a vanadate molybdate procedure.
Forage samples were dry ashed, concentrated HC1 added, redried and
resuspended in 0.5 N HC1. Chemical analysis of forage after dry ashing
were by the same precedures used for soil analysis. Liquid samples for
soil water and runoff were analyzed according to procedures outlined
for lagoon samples.
FIELD RESULTS AND DISCUSSION
The field equipment and monitoring system were installed to ascertain
the feasibility of determining the significant pathways for swine
lagoon wastewater in a plant-soil system. The resultant field data were
used to establish a mass balance on these plots for several of the waste
constituents. These data along with the soil profile characteristics
and soil-water movement would be inputs into modeling systems and if
successful could be utilized to determine environmental impacts of
wastewater land application. The various mechanisms for removal or
movement evaluated were crop uptake, surface runoff, soil accumulation,
and soil-water interflow. Waste inputs were also delineated.
197
-------
Irrigation Swine Lagoon Supernatant
The single-stage lagoon serving a Boar Testing facility with periodically
varying total liveweight was the source of the wastewater irrigated.
The first year of irrigation was late April through late September,
1973. Lagoon supernatant concentrations, hog populations, and complete
facility description have been presented in the Farm-Scale Lagoon Portion of
the Lagoon Studies and the Predictive and Interpretive Relationships Sections.
Calculation of the parameter which limits or determines the acres
required per unit volume of lagoon effluent indicated that nitrogen was
rate-limiting for this experiment. Nitrogen application generally limits the
loading intensity for most animal waste in the moisture excess southeast, so
the variable chosen for loading and replication in the test plots was nitrogen
application rate. Nitrogen applications chosen were a) below minimum
uptake of Coastal Bermuda grass, b) approximately equal to maximum uptake,
and c) in large excess of crop uptake rates. These translated to 336,
672, and 1344 kg N/ha. over the seasonal growing period for North Carolina
(May to September). There was no irrigation durin0 the remainder of
the year, thus these rates are 336, 672, and 1,344 kg N/ha./yr with
first irrigation on April 20 and last on September 17, 1973.
The actual amounts of chemical constituents and water applied for each
rate were as shown in Table 38 with each rate being replicated three
times as different field plots. The actual amounts applied were
slightly less than projected because of concentration variations due
to lagoon changes and irrigation losses. However, the 1:2:4 ratio
representing low, medium, and high was maintained. Actual applications
were 300, 600, and 1,200 kg N/ha. Since the wastewater for application
to all plots came from the same lagoon, the higher nitrogen rate plots
also received higher water application. This farm utilized a dietary
copper addition at about the 140-ppm level and thus the environmental
impact of heavy metals in swine feed could be evaluated (Overcash ).
However, the first stage lagoon reduced these metal levels so that the
lagoon supernatant had a low copper and zinc concentration.
Experience from initial irrigation events was used to redirect the work
plan. During the daytime period, the wind and resultant spray drift
were significant so that actual plot application rates would be erratic.
From experience with other irrigation systems, it was determined that
late night or early morning application would result in less drift
and thus the automated control system was programmed for this opera-
tion. The agreement between calculated volume of liquid needed to
achieve weekly nitrogen loads and actual volume of liquid collected
throughout the plots verified the superiority of night application.
Initially, the high nitrogen plots (1,200 kg N/ha./yr) were to receive one
hour of application at the sprinkler nozzle rate of 1.25 cm/hr. Several
198
-------
Table 38. CONCENTRATION AND AREAL LOADING RATE FOR SWINE LAGOON
EFFLUENT REACHING PLOTS RECEIVING HIGH, MEDIUM, AND
LOW RATES OF APPLICATION
Parameter
COD
TOG
TKN
NH3-N
P
K
Ca
Mg
Na
Cl
Cu
Zn
Mn
Fe
H20
Average
concentration
tng/1
1,140
375
224
148
49
220
52
62
86
139
.49
.56
.41
1.58
--
Application rate, Kg/ha . /yr
High
6,100
2,000
1,200
800
270
1,200
280
340
470
760
2.7
3.0
2.2
8.6
55 cm/yr
Medium
3,050
1,000
600
400
135
500
140
170
235
380
1.35
1.50
1.1
4.3
28 cm/yr
Low
1,525
500
300
200
70
300
70
85
120
190
.68
.75
.55
2.1
14 crn/yr
199
-------
attempts at this operation indicated that irrigation runoff occurred
for this 2-3 percent sloped Coastal Plains soil. Therefore, this high
level irrigation was reduced to two periods of thirty minutes spaced
2 to 3 hours apart. Wastewater irrigation runoff was reduced with
this 30-minute, 1.25 cm per hour rate of application, but direct
runoff occurred occasionally under wet conditions or when a sprinkler
failed to rotate.
By comparing the concentration of lagoon supernatant prior to irriga-
tion and that of the liquid falling onto the test plot surface,
(collected in several 400-ml beakers and composited) the volatiliza-
tion losses were determined, Table 39. However, these irrigation
losses were only indicative of night application.
Table 39. COMPARISON OF IRRIGATED SWINE LAGOON EFFLUENT AND LIQUID
REACHING PLOT SURFACE - NIGHT IRRIGATION
Parameter
COD
TOC
TKN
NH3-N
Average concentration,
niR/1
Lagoon
1,274
387
269
195
Plot surface
1,140
375
224
148
Percent loss
I.
3
17
24
The irrigated liquid samples were collected on approximately the same
day as the lagoon samples so the comparison of the two concentrations
could be done directly. It was found that analytical errors for
individual samples made the daily ratios of these two concentrations
erratic. Therefore, averages for all dates with complete analyses
(fifteen data points) were compared. It was felt that although there
was some change in the lagoon supernatant concentration due to hog
liveweight changes and operational fluctuations, this average was
meaningful for detecting irrigation losses.
Comparison of lagoon supernatant concentrations and liquid concentra-
tions reaching the plot grass level, Table 39, indicated that for the
organic parameters, there was an approximately ten-percent drop in
chemical oxygen demand but little or no change in total organic carbon.
The concentration difference between lagoon supernatant and wastewater
reaching the plot was significantly different from zero for COD (at
the 95-percent confidence level) while the TOC difference was not
200
-------
significantly different (at the 90-percent confidence level). These
facts would indicate that some oxidation of the wastewater occurred
but little or no volatilization of organics. Ammonia, a more volatile
compound, showed a greater concentration drop of approximately 50 mg/1
from an initial level of about 270 mg/1. The total Kjeldahl nitrogen,
containing organic and ammonia components, evidenced a similar 50
mg/1 drop in concentration indicating that ammonia volatilization was
the primary nitrogen loss mechanism. The TKN and NH3-N concentration
difference between the lagoon and plots was significantly different
from zero at the 99-percent confidence. Thus, from a lagoon loaded
at the reference rate approximately 25 percent of the nitrogen and 10
percent of the COD are lost during night irrigation. It should be
emphasized that the sprinkler heads were small (9.2 meter radius:
.32 cm nozzle) so the irrigation losses cannot be directly extrapolated
to larger farm-type sprinklers. Higher losses would be anticipated
for daytime irrigation because of higher temperatures and wind velo-
cities.
Research at Maryland with poultry processing wastewater (Larson^?)
has shown that 5-10 percent of the ammonia falling on grass and soil
surfaces is lost by volatilization. While these percentages would not
be expected to hold for this waste, these facts do indicate potential
additional volatilization losses beyond those measured for just irriga-
tion.
The input to the plant-soil system could be expressed by two methods
depending on the information available or needed by the designer.
The first was to describe original lagoon liquid and the corresponding
amounts of chemical constituents applied. This method would represent
the parameters available to the designer of a land application system
from lagoon sampling. The second method was based on the actual amounts
of material reaching the plant-soil system. Application rates for both
methods are given in Table 40. For purposes of mass balances, the
second method was more useful; hence, the rates referred to in the
discussion section are based on actual amounts reaching the plot
surface. For purposes of the second method, irrigation can be viewed
as an additional pretreatment resulting in material losses, Table 40.
Initial Soil Conditions
The preparation of the research plots and Coastal Bermuda grass cover
is given in the land application experimental procedures section.
Prior to application of swine lagoon effluent, soil cores were taken
from all plots for detailed sectioning and analysis to determine
baseline conditions. During the study additional cores were taken
outside the application area and upslope from the plots to verify
background soil concentrations.
201
-------
Table 40. COMPARISON OF LAND APPLICATION RATE BASED ON LAGOON
CONCENTRATION AND ACTUAL MATERIAL REACHING THE
PLANT-SOIL SYSTEM
Parameter
COD
TOG
TKN
P
High rate land application, kg/ha. /hr
Lagoon basis
6,830
2,060
1,440
270
Field basis
6,100
2,000
1,200
270
Initial conditions for the receiver plots were defined as the amount
of the various constituents of interest found in the upper 75 centi-
meters of the soil profile. These amounts were expressed on a kilo-
gram per hectare basis while the chemical analyses were obtained in
kilogram of species per kilogram of soil. To convert between these
measures, the plot area (97 m^ = .0097 ha.) and the bulk density profile
over the upper 75 cm of soil were needed. The bulk densities, Table
41, were the average of several core samplings and data indicated no
abnormalities. A slightly higher density was found in the 20 to 36
cm zone which was the A2 zone. This density profile was then used
with the concentrations (ppm) for all amount determinations in kg/ha.
by means of standard calculations.
The background soil values were measured either from initial cores
or subsequent control cores for nitrogen, phosphorus, potassium,
calcium, magnesium, sodium, chloride, copper, zinc, iron, and manganese,
Table 42. Comparison of the values for the upper 75 cm indicated a
fair uniformity between the randomly selected plots for high, medium,
and low application rates for parameters tested and thus indicated
homogenity for all soil constituents. Differences were partly due to
sampling and analytical variation. Complete data for individual plots
and sampling dates are listed in Appendix Cl. It was assumed that
over the long term, the control cores would approximate initial condi-
tions especially with respect to buildup of effluent applied consti-
tuents .
202
-------
Table 41. BULK DENSITY VARIATION WITH DEPTH IN COASTAL PLAINS
EXPERIMENTAL PLOTS (NORFOLK SANDY LOAM)
Depth, cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
Bulk density, g/cw?
1.56
1.76
1.65
1.68
1.90
1.82
1.63
1.61
1.60
1.52
203
-------
Table 42. AVERAGE INITIAL SOIL CONTENT IN UPPER 75 cm OF VARIOUS WASTE
PARAMETERS FOR PLOTS RECEIVING HIGH, MEDIUM, AND LOW
APPLICATION RATES
Parameter
TKN
P
Ca
Mg
K
NO -N
Na
Cl
Cu
Zn
Mn
Fe
Initial Concentration, ppm
High rate plots
(1,4,7)
2,887
145
814
149
262
-
-
-
-
-
-
-
Medium rate plots
(3,5,9)
2,253
142
820
118
262
-
-
-
-
-
-
-
Low rate plots
(2,6,8)
2,385
144
710
129
315
-
-
-
-
-
-
-
Control
clots
Concen-
tration,
ppm
-
47
700
170
204
4.0
119
176
10.9
40
36
410
Plot Runoff
Rainfall Runoff Volume-
The impact of rainfall on the movement of waste constituents in the
plant-soil system included rainfall-runoff, dilution, and soil perco-
lation. The amount of waste material removed in runoff was measured
from liquid volume and concentration measurements of the material col-
lected in submerged barrels from 1/8 of the total plot, Figure 76. Run-
off was found to occur as a result of rainfall as well as occasional
effluent irrigation so these data were separated, Table 43 and 44.
These preliminary runoff data represented fourteen events between start
204
-------
Table 43. RAINFALL RUNOFF VOLUMES FROM EXPERIMENTAL PLOTS EVALUATED IN 1973
ho
O
Ul
Date
4/27/73
5/4
5/18
5/21
5/29
6/18
6/25
7/16
7/30
9/19
11/30
12/10
12/20
Total
Plot rainfall runoff, I/plot
High a
1
495
45
75
709
1,163
79
1,598
1,174
844
979
120
--
75
7,356
splication plots
4
248
--
15
765
1,178
38
1,714
724
608
1,174
101
8
45
6,618
7
1,748
--
64
1,095
1,725
45
1,721
1,406
761
1,744
--
--
60
10,369
Medium application plots
3
278
--
38
652
983
38
1,391
435
293
671
--
34
34
4,847
5
131
--
--
221
649
105
1,185
165
75
1,035
79
56
53
3,754
9
1,095
64
45
780
1,380
191
1,875
911
623
1,665
49
--
41
8,719
Low application plots
2
1,166
--
109
776
1,316
56
1,553
330
559
671
150
79
98
6,863
6
52
--
--
128
315
--
143
19
86
221
94
--
--
1,058
8
994
--
60
919
1,609
191
1,721
1,084
1,058
863
--
--
--
8,499
-------
Table 44. EFFLUENT IRRIGATION RUNOFF VOLUMES FROM EXPERIMENTAL PLOTS EVALUATED IN 1973.
O
ON
Date
4/20/73
5/1
5/8
5/15
5/22
6/6
6/13
6/19
6/26
7/3
7/12
7/17
7/24
8/14
9/18
Total
Plot effluent irrigation runoff, I/plot
High application plots
1
— _
—
—
86
514
90
--
611
608
—
240
322
90
__
98
2,659
4
128
540
566
319
199
34
109
15
101
52
—
15
--
79
56
2,213
7
..
—
—
82
1,208
__
—
-_
720
68
—
799
34
—
52
2,963
Medium application plots
3
_„
—
41
—
251
--
—
308
—
--
—
45
—
--
--
645
5
__
—
--
—
109
--
--
34
--
—
-_
—
45
—
—
188
9
86
109
278
218
281
146
86
45
52
—
--
334
—
—
—
1,635
Low application plots
2
..
--
98
—
98
49
--
26
—
--
—
49
—
—
—
319
6
._
—
—
—
--
--
—
--
--
--
—
—
—
__
—
—
8
..
--
--
—
—
86
—
-_
--
—
--
360
—
-_
--
446
-------
up (5/73) through the end of waste application (9/73) and to the end of
December, 1973. Comparing all plots, Table 43, there were only five storm
events in which not every plot had some runoff.
Summation, of all runoff volumes, scaled up by a factor of eight to be
represented as total plot runoff, are also given in Tables 43 and 44.
The extremely low volumes produced from plot 6 and to some extent,
plot 5, opened questions as to the bias effect of these plots on the
evaluation of material lost in rainfall-runoff. Possible explanations
were inadequacy of liquid collection facilities or that the soil proper-
ties and crop conditions allowed greater infiltration. In any event,
this entire replicate of plots (4,5,6) were eliminated from further
calculations or conclusions pertaining to runoff. Results from these
plots were, however, included in other aspects of this field study.
Rainfall-runoff relationships from these plot data should be taken as
preliminary because the largest single event volume is 20 to 30 percent
of the total runoff volume, Table 43, thus making determinations very
error sensitive. Data should be taken over several" years to obtain
conclusive information. The third replicate (plots 7,8,9) periodically
had greater runoff volume than the first replicate (plots 1,2,3) but
within each replicate group there was about the same amounts of runoff
even though each plot had a different nitrogen application and thus
hydraulic load. Considering the soil profiles, Figure 77, the third
replicate had 18 to 25 cm less topsoil above the less permeable B
horizon; that is, it was more eroded. This reduced storage capacity
would result in more runoff during large rainfall events thus leading
to slightly higher total runoff volumes. However, despite such
limitations, there appeared to be a degree of uniformity among the other
loading rate replicate (plots 1, 2, 3) and thus some definable trends
became apparent. Comparisons between plots receiving animal waste
and plots under normal pasture conditions or in other types of land use
should be made to better relate these data to baseline runoff from
predominantly rural areas.
For the first (1,2,3) and third replicates (7,8,9) the plot receiving
the higher loading had the highest and rather similar runoff volume.
Increase runoff especially immediately following irrigation for the high
rate plots (1,200 kg N/ha./yr) due to wetter conditions could explain
recorded runoff volume differences. Rainfall data was not available
to. correlate storm intensity or volume to runoff volume for so few
events. Averaging the runoff volume for the two replicates used over
the period shown in Table 43 yielded 8,100 1 per plot. Rainfall over this
period from a farm weather station 1.25 Km from the site was 51,800 1
per plot for a runoff percentage of approximately 15 percent.
Irrigation Runoff Volume-
Runoff occured as a result of irrigation events on some occasions for
207
-------
even the low waste loading plots, Table 44. Again, eliminating replicate
2 (plots 4, 5, and 6) the average irrigation runoffs were calculated.
A more pronounced dependence on liquid loading rate was evidenced
for irrigation related runoff than for rainfall runoff with the
highest irrigation runoff occurring at the 1,200 kg N/ha./yr application
rate. The runoff as a percentage of the applied liquid was 4.3 per-
cent, 3.7 percent, and 1.5 percent for 1,200, 600, and 300 kg N/ha./yr,
respectively. The liquid application rate was 1.3 cm per hour with
varying duration irrigation used to give the required nitrogen rates.
This irrigation runoff although small compared to the total applied,
was significant compared to rainfall runoff especially for the high
nitrogen rate plots. The expected high concentration of pollutants
in this liquid emphasized the need to preclude this type of runoff.
However, continued long-term monitoring may show that these runoff
volumes become significantly reduced as the received plot matures.
Some of the irrigation runoff seemed directly related to prior rain-
fall. Runoffs from the hig., waste plots on 5/22 and 6/26 were quite
large compared to other events and on 5/21 and 6/25 there were heavy
rainfalls as evidenced by the large rainfall runoff. On 6/18, there
was a light rainfall-runoff event and the irrigation runoff was
correspondingly lower on 6/19. At lower waste application rates, the
link between prior rainfall and irrigation runoff was less conclusive
than at the high rates. Certainly, irrigation immediately after rain-
fall events will require more control if runoff is to be minimized.
Also a buffer distance to allow infiltration of irrigated area runoff
would reduce the impact of the irrigation runoff found for the appli-
cation conditions of this study.
Waste Constituent in Runoff-
Analysis of samples taken from the runoff barrels was performed for
a wide variety of constituents. However, experimental difficulties
prevented analysis of all samples, thus a complete assessment of the
amount of material lost from the plots by means of runoff was not made.
Chemical oxygen demand was the parameter analyzed most consistently.
Using the subplot area and runoff volumes for the 8 samples collected
representing 60 percent - 70 percent of the total runoff volume, the
kg/ha.of COD lost was calculated for the two replicates of the three
loading rates, Table 45. Although there was a slightly greater runoff
volume for the high nitrogen plots, the liquid concentration of organics
(COD) lost could not be directly correlated to waste application rate,
Table 45. Neither the high nor low rate plots consistently yielded
the greatest amount of COD runoff. Again 30 percent - 50 percent
of the total COD was lost during the largest individual runoff
event so that normal errors could easily have distorted runoff evalua-
tions for this single year of data. The COD applied was 6,100, 3,050,
208
-------
TABLE «5. CONCENTRATION OF RAINFALL RUNOFF FROM EXPERIMENTAL PLOTS RECEIVING SWINE LAGOON
EFFLUENT,
5/8
5/21
5/29
6/18
6/25
7/16
7/17
7/30
11/30
7/24
8/14
••"-•—-« rarameter concentration, mg/l
High application plots Medium application plots
1473 59
COD
TKN
NO^-N
COD
TKN
NO--N
Ca
Mg
K
Na
Cu
COD
TKN
NO -N
o-PO -P
Ca
Mg
K
Na
Cu
COD
TOC
TKN
NO -N
o-PO -P
4
COD
TOC
TKN
NO--N
o-P04-F
COD
TOC
TKN
NO -N
o-lo^-P
COD
TOC
TKN
NO..-N
Ca3
Mg
K
Na
Cu
COD
TOC
TKN
NO -N
o-PO P
Ca
Mg
K
Na
Cu
COB
TOC
TKN
NO -N
o-PO.-P
4
Ca
Mg
K
Na
Cu
Ca
Mg
K
Na
Cu
116
9
4
35
2
.3
2.2
.86
4.8
5.5
-Oil
61
6
.20
2
3
1.33
6.8
1.1
.011
176
32
21
8
5.5
34.5
10
2.5
1
3
39
10
2.5
1
9
298
195
25
5
23.7
5
107
26.4
.34
100
40
5
1
9.9
3.4
22.2
4.6
.29
108
-
6
39
-
-
2.7
.98
4.7
5.5
.033
49
3
0
2
2.7
1.05
18.6
7.7
.033
111
32
9.5
6
9
19.5
5.5
2
.5
3
39
10
2.5
3
8
96
65
15
5
13.3
4.16
64.9
10.5
.10
39
12
4
3
7.1
2.9
19.5
2.4
.11
105
7
-
109
2
1
1.5
.71
5.2
9.9
-
34
2
0
1
2.0
.96
6,7
1.7
.022
123
30
11
11
3
27
5
2
1
1
31
11
4
2
8
27
20
11
2
9.3
4.0
16.8
.0
.099
31
17
3
3
7.0
3.0
12.4
1.2
.077
105 No runoff
_
„
_
-
29.3
5.1
215.6
58.3
.143
No runoff
"
"
"
"
25
19
54
3
No runoff
'»
"
"
"
"
"
"
30.2
4.5
204.6
55
.594
25 No runoff
4.9
177
45
.24
"
11
"
"
136 No
13
27
2
0
1.7
.69
5.8
5.5
.011
68
6
0
1
2.3
.87
4.8
6.6
.011
188
62
13
9
5
31
5
2
2
1
39
10
3
3
6
169 No
90
12
15
9.7
3.6
26.3
24.2
.24
35
14
2
2
6.5
2.3
12.5
1.9
0
No runoff
"
"
"
"
No runoff
"
"
"
"
No runoff
"
"
"
"
cunof f
'i
"
39
6
0
3.3
25.0
10.7
2.3
-Oil
91
4
.1
1
1.9
.82
17.7
1.3
.022
84
32
16
5
3
27
5
2
1
.9
54.5
30
15
6
8
runoff
"
"
"
"
11
"
"
"
73
23
6
10
7.9
3.0
30.8
6.2
.08
70
25
5
10
0
116
_
-
113
3
.55
2.1
.95
5.9
7.7
.011
106.5
6
.6
1
1.3
.67
8.8
1.5
.033
84
25
7
2
3
4
5
2
1
.9
35
10
4
1
3
21
18
5
1
4.2
2.2
12
1.5
.099
54
15
5.4
2
21.3
2.4
11.0
2.0
.11
31
40
4
21
0
Lou
2
101
-
117
6
2.
21
8.
38
5.
1.
8.
3.
100
20
6
6
2
27
5
2
2
58
25
7
3
3
117
73
8
4
9.
application plots
6 8
No runoff
"
43
45 0
.4 1.7
.83 .66
5.3
.3 1.1
.044
152
5 10
03 .15
12 2
3 3.0
82 1.03
8 14.0
1 9.9
132 .022
No runoff
"
"
M
"
35
5
2.5
1.5
7 .1
117
10
4
7
3
No runoff
"
"
"
3
3.5
38.
13.
73
32
6
2
5
2
23.
2
-
-
-
-
18.4 No runoff No
4.5
61.6
15.4
.154
No runoff
"
"
.5 "
.9
.088
324
57
11
13
.5 7.1
2.6
,1 25,0
.3 4.4
. 099 . 10
70
25
5
10
0
128
3
101
3
.70
1.1
.47
4.7
6.6
.022
148
11
.55
3
1.3
.63
8.4
2.2
.022
84
25
6
3
2.5
12
5
1
1
.4
27
9
3
1
3
35
24.5
11
1.5
4.1
2.2
13.2
2.2
.077
58
14
3
1
2.9
1.1
9.9
.66
.11
No runoff
runoff No runoff No
" "
„
11
t]
runoff
tt
,,
209
-------
and 1,525 kg/ha, for the high, medium and low rate plots, respectively,
so that the percentage lost in runoff decreased with increasing loading
rate from 2 percent to .5 percent.
The rainfall runoff concentrations for the remainder of the parameters
evaluated were included in Table 45. Conclusions based on the measured
amount of material lost in runoff were not made; instead trends were
deduced from concentration values and only approximate order of mag-
nitude values for runoff losses were determined.
For the four events listed as receiving complete analysis, Table 45,
there appeared to be uniformity of concentration values for a single
event with differences in concentration between rainfall events.
This similarity of concentration would imply (as with the COD) that
there was little differences among loading rate and that the amount
lost as a percentage of that applied decreased with increasing loading
rate. This uniformity of runoff amount was only a preliminary trend
based on approximate concentration and volume values. Further data
over a number of years will be needed for conclusive results. Order
of magnitude calculations for rough average runoff concentration
values and total runoff volumes indicated that less than five percent
of applied constituents appeared in the rainfall runoff.
Runoff from irrigation events was analysed for two dates, 7/24/73
and 8/14/73, Table 45. The concentrations for the majority of the para-
meters was roughly 50 percent to 80 percent of the irrigated liquid
concentration. Observations of the irrigation runoff from the
remainder of the plot showed that within 0.3-1.5 m the liquid has
soaked into the soil. Therefore, it was concluded that the impact of
irrigation runoff would be insignificant for this type of plant
soil system provided a small buffer strip was utilized.
Crop Uptake and Utilization
Dry Matter Yield-
The Coastal Bermudagrass was managed as a hay crop with the plots for
each loading rate harvested at the 30-35 cm height, detailed in the
Land Application Experimental Procedures Section. Evaluation of data
uniformity indicated that the second replicate (plot 4, 5, and 6) was
not significantly different from the first and third replicates. The
second replicate, while not used in runoff determinations, was included
in crop uptake calculations. Dry matter yields and percentage dry
matter for the three high, three medium, and three low rate plots were
determined, Table 46.
210
-------
Table 46. DRY MATTER YIELDS (kg/ha.) AND PERCENTAGE DRY MATTER OF COASTAL
BERMUDA GmSS FOR FIRST-YEAR APPLICATION OF THREE LOADING RATES
OF SWINE LAGOON EFFLUENT (1973)
Loading
rate
Low
Harvest
date
6/13
7/16
8/16
9/21
Dry matter yield
average of 3 replicates
(kg/ha.)
1.955
3,212
3,865
2,257
Percentage
dry matter
27.9
30.9
39.2
30.7
Total
11,290
Mean
32.2
Medium
6/13
7/11
8/7
9/10
Residue
after frost
Total
High
6/13
7/6
8/1
8/31
Residue
after frost
Total
2,125
4,308
2,747
4,525
900
14,605
2.393
3,392
3,773
4,623
1,970
16,151
Mean
without
residue
Mean
without
residue
2205
25.4
22.7
30.5
84.2
25.37
22.2
22.9
23.3
28.7
69.9
24.3
The dry matter contents were lower at the 600 and 1,200 kg N/ha./yr rates
than for the 300 kg N/ha./yr rate demonstrating the greater water uptake
and top growth associated with excess nitrogen conditions (Doss ).
The residue samples were high in dry matter because they were taken
after autumn frosts but low in yield and thus were used only to deter-
mine total parameter amounts removed. Comparison of the dry matter
211
-------
yield among the three effluent treatments demonstrated increased growth
with increased effluent loading rate. However, only the dry matter
yield difference between the 300 and 600 kg N/haJyr rates and 300 and
1,200 kg N/ha./yr rates were significant (P<.05). The use of 1,200
kg N/ha./yr did not significantly increase dry matter yield over the
600 kg N/ha. rate indicating the plateau region for response to
nitrogen. The amount of N, P, (Woodhouse89) and K (Woodhouse90)
applied at the highest effluent application rate was more than double
the amounts found to produce maximum yield of Coastal Bermuda grass in
North Carolina.
Weekly observations of grass growth and appearance did not evidence
any adverse effects for any of the experimental plots. The only
visible forage symptom was a temporary chlorotic condition at the high
rate application which developed in the regrowth during September, 1973,
Foliar analysis indicated no nutrient abnormalities and no permanent
damage occurred with the loading rates used. Thus, the crop yields
reflected response under good growing conditions.
Nutrient and Trace Mineral Uptake-
Analysis results for each grass harvdst is presented in Tables 47 and
48 on a dry matter basis. These were the first year results and con-
clusions represented initial trends. In general, the increase in grass
dry matter concentration of the elements tested was significant between
the various effluent application rates. Zinc represented the exception
to the trend of increased concentration with increased loading rate.
However, using the incremental increase between the application of
300 and 600 kg N/ha,/yr as a reference, further increases in plant
composition commensurate with the increase from 600 to 1,200 kg N/ha./yr
were found only for calcium, manganese, and iron. The other parameters
increased only slightly with increased effluent application. Reasons
for the first year information showing larger Ca, Mn, and Fe increases
at the high rate were not fully defined but may have involved initial
soil deficiencies or alterations in soil exchange capacity with varying
water content or pH conditions.
Conversion of the plot yields and grass composition to the amount of
material removed by the crop was completed and shown in Table 49.
Increased effluent application resulted in increased crop removal
since both yield and concentration increased with loading rate. How-
ever, for none of the parameters did the removal as a percentage of
the applied material increase with larger waste input, Table 49.
212
-------
Table 47. MINERAL CONCENTRATIONS OF COASTAL BERMUDA GRASS (AVERAGE OF
THREE REPLICATIONS) FOR FIRST-YEAR APPLICATION OF .THREE
LOADING RATES OF SWINE LAGOON EFFLUENT (1973)
Loading
rate
Low
Medium
High
Harvest
date
6/13
7/16
8/16
9/21
Mean
6/13
7/11
8/7
9/10
Mean
6/13
7/6
8/1
8/31
Mean
N
2.07
1.54
1.81
1.92
1.84
2.21
2.21
2.71
2.09
2.31
2.46
2.75
2.90
2.50
2.65
P
0.190
0.175
0.177
0.183
0.180
0.203
0.227
0.243
0.193
0.215a
0.207
0.227
0.243
0.217
0.225a
K
% D
1.81
1.66
2.06
1.84
1.84
2.14
2.43
2.41
2.28
2.32
2.41
2.72
2.64
2.47
2.56
Ca
.M
0.34
0.27
0.25
0.28
0.29
0.32
0.38
0.39
0.35
0.36
0.38
0.43
0.50
0.46
0.44
Mg
0.18
0.15
0.15
0.17
0.16
0.20
0.22
0.23
0.24
0.22
0.21
0.24
0.28
0.29
0.26
Cl
0.763
0.697
0.833
0.823
0.778
0.887
0.867
0.793
0.970
0.878
0.883
0.717
0.663
0.627
0.722
Means with same letter are not significantly different at a 95-
percent confidence level.
213
-------
Table 48. MINERAL CONCENTRATIONS OF COASTAL BERMUDA GRASS (AVERAGE OF
THREE REPLICATIONS) FOR FIRST-YEAR APPLICATION OF THREE
LOADING RATES OF SWINE LAGOON EFFLUENT (1973)
Loading
rate
Low
Medium
High
Harvest
date
6/13
7/16
8/16
9/21
Mean
6/13
7/11
8/7
9/10
Mean
6/13
7/6
8/1
8/31
Mean
Na
910
1055
840
955
940
1200
1800
1780
1750
1640
1400
1410
1910
1610
1580
Cu
11.67
11.00
10.00
10.33
10.75
13.33
17.00
15.00
13.00
14.58a
18.00
15.66
16.66
12.00
15.58a
Mn
ppm
57.0
44.0
38.0
38.3
44.3
64.0
59.6
59.0
45.0
56.9
69.0
73.0
84.3
64.3
72.65
Zn
24.0
23.0
24.6
25.3
24. 2a
25.3
26.0
26.6
21.6
24.9a,b
26.6
24.3
30.0
24.6
26. 4b
Fe
262.3
202.3
097.6
112.3
168.7
377.0
283.7
236.3
119.3
254. Oa
370.3
287.3
249.6
132.3
259. 4a
Means with same letter are not significantly different at 95-
percent confidence level.
214
-------
Table 49. CROP REMOVAL RATES FOR FIRST YEAR OF SWINE LAGOON EFFLUENT APPLICATION TO
COASTAL BERMUDA GRASS (1973)
Parameter
TKN
P
K
Ca
Mg
Na
Cl
Cu
Zn
Mn
S
Fe
Crop uptake, kg/ha.
Low rate
207
20
201
32.5
18.5
10.6
88
.12
.27
.50
22.2
1.9
Medium rate
336
31
340
52.5
32.7
23.9
128
.21
.36
.83
28.8
3.7
High rate
428
36
413
71.2
41.6
25.6
116
.25
.42
1.18
33.6
4.1
Crop uptake percentage of applied
Low rate
70
30
69
46
22
10
46
18
34
90
--
89
Medium rate
57
23
56
37
20
10
33
16
25
82
--
87
High rate
36
13
34
25
12
5
15
9
14
55
--
48
KJ
I—'
Ul
-------
Hay Quality and Animal Acceptability-
CoastaL Bermuda grass from the plots receiving swine lagoon effluent and
control plots receiving only ammonium nitrate were used to assess
Coastal Bermuda grass quality as an animal feed. Grass preparation was
described in the Land Application Experimental Procedures Section. Hay
quality was evaluated with the two-stage Tilley and Terry in vitro pro-
cedure and the Kjeldahl nitrogen analysis as an expression of crude
protein content. The former technique estimated the _in vivo dry matter
digestibility of forages by means of in jLn vitro dry matter disappearance
(IVDMD) under controlled conditions
Average IVDMD (four harvest dates) were about 57 percent to 58 percent
of initial dry matter and were quite similar for the high, medium
and low application rates, Table 50, with the greatest differences
occurring in harvest-to-harvest measurements. For comparison, the
control was 46 percent IVDMD while standard forage values for Alfalfa
were 49 percent IVDMD (50 percent in vivo) and bromegrass 67 percent
IVDMD (62 percent in vivo). Thus compared to the Bermuda grass grown
under standard fertilizer conditions and alfalfa grass, the IVDMD
for the crop from the effluent plots was 7 percent to 12 percent
higher.
The total N (TN) concentrations for these forages increased with effluent
nitrogen loading rate. However, the fact that at even the high rate,
the total nitrogen was less than three percent was unexpected compared
to forage values of up to five percent obtained under greenhouse condi-
tions (Burtony ). As expected, the correlations of IVDMD and TN were
positive and highly significant (Table 50) for the low and medium rate
plots. At higher rate, there was a lower correlation level possibly
because of other limiting factors besides nitrogen.
Crop yields for individual harvests were insufficient for hay intake
trials so composite samples were made, i.e., for the low, medium, and
high rates, grass was mixed together from 7/16 and 8/16, 7/11 and 8/7,
and 7/6 and 8/1, respectively. Thus, the four forage sources, ground
through a 3.8-cm screen in a Davis mill, available for feeding, were
a) control (inorganicly fertilized forage), b) low rate of effluent,
c) medium rate of effluent and d) high rate of effluent.
Sheep were chosen for hay acceptability testing because sufficient
forage was not produced for cattle feeding. Based on previous work in
the Animal Science Department, it was felt that if sheep would consume
the field hay then acceptability by cattle would be assured. A total
of 16 ewes (8 Dorset, 8 grade Suffolk) were weighed, placed in individual
pens and fed control hay ad libitum for an 11-day standardization period.
216
-------
Two animals refused to eat and were removed from the experiment. After
the animals were accustomed to the Bermuda grass hay, they were randomly
assigned, within breed groups, to the four hay treatments. Hays were fed
ad_ libitum for a 13-day experimental period. Hay intake was determined
from pounds of hay offered minus pounds of weighback. Samples of each
hay were obtained daily throughout the experimental period. These samples
were composited and analyzed for crude protein (percent total N x 6.25)
and dry matter.
Table 50. IN VITRO DRY MATTER DISAPPEARANCE (IVDMD) AND TOTAL NITROGEN
CONCENTRATION OF COASTAL BERMUDA GRASS HAYS FOR FIRST-YEAR
APPLICATION OF THREE LOADING RATES OF SWINE LAGOON EFFLUENT,
1973, (MEAN VALUES OF 3 REPLICATIONS)
Loading
rate
Low
Medium
High
Harvest
date
6/13
7/16
8/16
9/21
Mean
6/13
7/11
8/7
9/10
Mean
6/13
7/6
8/1
8/31
Mean
IVDMD
(%)
58.9
51.6
58.7
58.3
56.9
55.7
55.0
61.4
55.0
56.8
53.4
63.2
58.5
58,3
58.4
Total N
a)
2.07
1.54
1.81
1.92
1.84
2.21
2.21
2.71
2.09
2.31
2.46
2.75
2.90
2.50
2.65
Correlation (r)
between total
N and IVDMD
0.89
0.98
0.58
217
-------
There was no evidence of reduced hay palatability due to the use of
lagoon effluent during crop production when compared to the control,
Table 51. The hay intake per unit body weight was not significantly
different among the loading rates studies, Table 51, although the crude
protein content increased with increased loading. Within the animal
types the grade Suffolk lines consumed more hay in relation to body
weight than Dorset ewes, .02 versus 0.15 kg hay/kg body weight (P<.05).
These trends in palatability, crude protein level, hay composition,
and animal roughage intake by breed and forage species found in the
first year study need to be continued for further optimization of
crop utilization. Long-term effects of lagoon effluent irrigation on
hay quality must also be considered.
Table 51. ANIMAL INTAKE AND HAY COMPOSITION FOR COASTAL BERMUDA GRASS
FOR FIRST-YEAR APPLICATION OF SWINE LAGOON EFFLUENT, 1973.
Loading
rate
Low
Medium
High
Control
No. of
ewes
3
3
4
4
Hay intake
kg per
head
1.21
1.35
1.27
1.34
kg per kg
body weight
.018
.017
.016
.018
Hay comoosition, %
Dry
matter
91.9
92.2
92.4
91.7
Crude protein
As feed
10.2
13.8
17.0
8.2
Dry
Basis
11.1
15.0
18.6
8.9
Soil Accumulation
The soil buildup after one year of swine lagoon effluent application
was evaluated from analyses of soil cores taken of the top 75 cm of
the soil solum. These cores were sectioned to provide profile as
well as the total 75 cm core data. Three cores were taken from each
plot and composited at each depth prior to analysis as outlined in the
Land Application Procedure Section. All subsequent core samples have
been taken in the same general vicinity but upslope to eliminate any
leaching effects associated with the refilled holes. The check or
control cores were taken in triplicate downslope from plots 2 and 3
at a distance equivalent to the upslope edge of plots 7, 8, and 9,
Figure 75 . The region of the control cores had an established stand
of Coastal Bermuda grass and received maintenance level fertilization.
218
-------
Initial soil cores were taken at 15-cm increments because the upper 12-
15-cm was plowed prior to the experiment and thus assumed to be uniform.
Later samples were taken at 5-cm segments until the 25-cm depth and then
10-cm segments to the 75-cm depth. Thus, soil profile comparisons were
not one-to-one. The difference was most prevalent in the surface area
where concentrations changed the most; hence concentration profile
comparisons were approximate. However, this difference in sampling
procedure did not affect overall determinations in the upper 75 cm
of these plots.
A large number of pathways or mechanisms for removal or attenuation
of the applied waste constituents exist in the field situation. The
three most significant pathways in this study were movement or loss
associated with lateral or vertical soil water flows, uptake by the
growing crop, and the incorporation of chemical constituents into the
soil matrix. This latter mechanism represented accumulation or buildup
in the experimental situation. If the rate of buildup was greater than
loss rates to percolation or crop uptake, there would be a positive
accumulation; if the rate of buildup was less, there would be a negative
accumulation or a soil depletion.
Materials such as potassium or calcium which accumulate in a soil
system are retained with varying binding energy levels. Traditional
definitions refer to three reservoirs, listed according to decreasing
binding energy as 1) tightly bound in the lattice of secondary
minerals, 2) moderately held as at the outer edges of the clay lattice,
and 3) that in soil solution or retained by the soil exchange capacity.
Heavy or extended wastewater applications could cause components to
enter any of these reservoirs and thus accumulate. As the energy
required to remove these compounds or elements decreases, the likeli-
hood of detecting them with the tests employed in this study increases.
Temperature, soil water levels, and microbial populations can also
influence the concentration of an element as these factors alter the
-i r\ OQ
amount of exchangeable material (Murrmann- , Hunt-7-*). Other elements
such as chlorides evidence little accumulation in soils similar to
those of this study because of the low soil anion exchange capacity.
Thus the amount of soil accumulation was affected by the particular
element or compound, the ability to detect all of a material present,
and the levels present prior to the initiation of the swine lagoon
effluent study. The presence of these complexities in determining
parameter soil levels required some simplfying assumptions. Primarily,
it was assumed that for each increment of a material added in the
effluent regardless of loading rate, approximately the same fraction
of the amount accumulating in the soil was detected Dy chemical analyses.
Also over several years of application, it was assumed that the build-
up of materials determined from the soil core analyses would better
approximate the actual soil conditions. The first year data would thus
be observed for trends.
219
-------
Concentration data for the initial cores; those taken from effluent
plots in September, 1973, after one season of irrigation; those taken
from the same plots after three more months with no waste input, December,
1973; and September and December, 1973, cores from control areas were
tabulated in Appendix Cl. Extensive comparisons of these data showed
that results for plots 4, 5, and 6 (removed from discussion of runoff
data) were not significantly different from the first and third replicates,
and hence, were included in further discussions. From these nine sets
of data, averages for plots receiving the same waste load were made
to that the results appeared as high, nedium, and low rates corresponding
to the three plots receiving 1,200, 600, and 300 kg N/ha./yr, respectively.
The soil concentration in September and the accumulation over the first
irrigation season in the upper 75 centimeters of soil, expressed as kg
per ha.for the waste constituents analyzed were listed in Table 52.
Nitrogen data indicated the Kjeldahl nitrogen level increased at the
high rate as did the nitrate nitrogen levels. For the medium and low
rates, there were slight increases in nitrate levels over the control
areas, but little TKN accumulation. T.ie other u.ajor nutrients, P and
K, showed increases at the high rate (23 and 174 kg/ha. , respective-
ly) but evidenced only slight accumulations and depletions at the medium
and low rates.
Cations Mg and Na were determined to have increased slightly in the
upper 76 cm of soil while Ca levels decreased even at the high rate.
Other micro nutrients and heavy metals varied in the levels of
accumulation or depletion, Table 52.
These data presented a mixed picture with respect to soil accumulation
in the entire upper 75 cm of soil. In general, the higher rates
plots had higher accumulated levels of the applied waste constituents
while there was little difference between loading rates for some para-
meters.
The soil levels and accumulation over the initial soil conditions for
the cores taken three months after irrigation termination in December,
1973, were tabulated in Table 53. Comparison of September and December
data indicated the potential variability in these large average values
for the upper soil zone, since some levels such as TKN show as increase
after this three-month time period. Despite the absolute soil level
disparities recorded for the December, 1973, and September, 1973, data
the trend of higher accumulation at the higher application rates was
evident for both samples.
220
-------
Table 52. SOIL ACCUMULATION IN UPPER 75 cm BENEATH PLOTS RECEIVING SWINE LAGOON EFFLUENT
AFTER ONE SEASON, SEPTEMBER, 1973
Parameter
TKN
P
K
Ca
Mg
Cl
Na
Zn
Cu
Fe
Mn
N03
Soil concentration, kg/ha.
High rate
3,063
174
436
768
222
603
130
28.7
17.7
298
43.9
81.6
Medium rate
2,265
140
306
796
179
510
138
29.0
25.5
319
46.8
12.2
Low rate
2,374
109
218
652
140
492
81.8
21.3
14.3
291
64.1
12.9
Increase in soil amount between initial
or control plot conditions and September
conditions, kg/ha.
High rate
176
28
174
-45.9
72.8
427
11.2
-11.2
6.83
-113
8.06
77.3
Medium rate
11.2
-2.3
43.7
-23.5
61.6
334
19.0
-11.2
14.7
-91.8
11.0
8.18
Low rate
-11.2
-37
-96.3
-58.2
11.2
316
-37.0
-19.0
3.47
-120
28.2
8.85
-------
Table 53. SOIL ACCUMULATION IN UPPER 75 cm BENEATH PLOTS RECEIVING SWINE LAGOON EFFLUENT
AFTER ONE SEASON, DECEMBER, 1973
Parameter
N
P
K
Ca
Mg
Cl
Na
Zn
Cu
Fe
Mn
NO
Soil concentration, kg/ha
High rate
3,189
255
716
736
162
393
237
37.0
18.9
556
48.0
21.4
Medium rate
2,811
177
642
790
160
388
174
28.2
10.9
486
50.5
10.4
Low rate
2,866
131
325
766
143
358
168
28.3
12.8
496
53.3
7.17
Increase in soil amount between initial
or control plot conditions and December
conditions, kg/ha
High rate
301
110
454
-78.4
13.4
217
119
-3.36
8.06
144
12.2
17.4
Medium rate
558
34.7
380
-30.2
42.6
212
54.9
-12.1
0
75.0
14.7
6.38
Low rate
480
-13.4
10.1
56
14.6
183
47.0
12.0
1.90
85.1
17.5
3.14
NJ
K3
-------
Calcium, chloride, and nitrates showed consistent decreases in the total
amounts present between September and December. All of the waste
constituents were expected a_ priori, to decrease during this three-
month dormant period when only rainfall leaching was occurring. However,
there was only 7.5 cm of rainfall during these three months which was
only about 30 percent of the expected normal. However, some parameters
still appeared to increase possibly representing variability in this
testing procedure.
Treating September and December data as two estimates of soil buildup
and averaging the data the conclusion that greater accumulation occurred
at the heavy loading rate remained valid. The variability in amount of
material accumulated per unit area prohibited exact determinations of
buildup rates after a single season of application.
More detailed data analyses indicated some of the reasons for this large,
short-term variability. One or two incorrect analyses or contamination
of one or two bottom se.gments could represent 20 percent to 30 percent
of the total parameter quantity. Additionally, for several components,
the amount applied at the medium and low rates was small compared to
the initial amount in the soil, e.g.,Ca, P, Mg and N. Of more importance
for many of the parameters such as K and Fe was the high levels of
primary and secondary minerals in the soil. These were often thousands
of kilograms per hectar as compared to levels evidenced in analyses in
the order of hundreds of kg/ha. Thus, small shifts in available levels
of these minerals would have introduced data variability.
Soil profile analyses of chemical constituents were more useful in
observing initial trends because surface effects could be identified
and erroneous points more easily elucidated. These data were averages
of values in Appendix C3 for replicated plots receiving the same waste
input applications, high, medium or low. Only the profiles for high,
low, and either initial or control cores were plotted to demonstrate
differences associated with loading rates.
Total Kjeldahl nitrogen levels did evidence a large increase at the
surface zone commensurate with the amount of nitrogen added, Figure 78.
By March, 1974, the concentration profile was about the same as in
September, 1973 at irrigation termination except for a modest nitrogen
decrease in the upper 5"cm zone. Soil nitrate from oxidation of applied
TKN was evident for the higher loading rates throughout the waste
application profile in comparison with the control cores, Figure 79.
Decreases in the soil profile nitrate levels between September and
December, 1973, probably represented leaching losses. Nitrate levels
were a very small percentage of the TKN detected.
223
-------
15
NJ
ho
js
4J
cx
o
45
September, 19^3
Concentration (PPM)
200 400
600
751-
Figure 78.
15
30
g
u
ex
01
Q
45
60
751-
March, 1974
Concentration (PPM)
200 400
600
HIGH RATE
LOW RATE
INITIAL AVERAGE
Soil profile total Kjeldahl nitrogen concentrations for first-year application of
swine lagoon effluent, 1973, at 1200 and 300 kg N/ha.
-------
l-o
01
g
u
a,
0)
Q
45
60
September, 1973
Concentration (PPM)
3 6
S
u
ft
4J
P.
P
45
60
December, 1973
Concentration (PPM)
3 6
r
—O— HIGH RATE
—Q— LOW RATE
CONTROL CHECK
75 »- 75
Figure 79. Soil profile nitrate concentrations for first-year application of swine lagoon
Effluent, 1973, at 1200 and 300 kg N/ha.
-------
Soil calcium concentrations remained relatively constant from before
application through one season (both September and December, 1973,
profiles) of loading for the low and high application rates, Figure 80.
The September control sample evidenced a lower surface zone concen-
tration possibly due to crop uptake in this unirrigated area. Magnesium,
another divalent cation, was increased for both the low and high effluent
plots in the surface zones by September, 1973, Figure 81. Cores taken
in December, 1973, were lower in magnesium concentration at the surface
but low and high rate concentrations were still somewhat above control
levels below the 75-cm depth. This trend of downward leaching was found
for both the low and high rate plots and thus by December, 1973, surface
concentrations returned to levels existing prior to waste irrigation.
Phosphorus, largely removed from the raw waste by the lagoon treatment
mechanisms described earlier, was found to accumulate mostly for the high
rate of application and then basically in the surface zones, Figure 82.
Between September and December, there was a slight increase in the
surface zone concentration of phosphorus indicating analytical variations.
As with calcium, the control area was lower in phosphorus than the
original samples presumably due to Coastal Bermuda grass growth.
Three monovalent and more mobile ions were measured through the upper
75-cm profile, K, Na, and Cl. Potassium was increased over the initial
soil levels throughout the whole profile for the high rate, Figure 83.
At the low rate, no significant increases were found. The profiles
for both low and high rates had similar potassium concentrations at the
0-to 5-cm depth with the low rate soil levels decreasing to background
levels at about 20~cm. The high rate remained elevated over background
down to the 75-cm depth. These data indicate that the exchange capacity
was satisfied for the surface layer even at the low application rate after
which downward movement resulted. The December soil potassium profile
was higher than September for the high rate plots. Explanations for
this increase were not totally available but the effect of season on
exchangeable levels of K has been reported. The low rate plots
showed the anticipated leaching loss of K to lower soil layers, Figure
83, as evidenced by the lower surface concentrations and slightly
higher levels in the lower profile for the December data.
The soil sodium profiles were about the same or slightly elevated over
control cores, Figure 84, for low and high application rates. Sodium
in the control cores, the low rate plots and the high application plots
for December was above levels measured from the same areas in September.
Chloride soil concentrations were approximately the same for the low
and high plots in the upper 36 cm and above control core levels, Figure
85. The heavy application plots had greater chloride levels from the 36- to
75-cm depth indicating that some movement out of the upper soil profile
226
-------
Deptember, 1973
Concentration (PPM)
100 200
T
300
Decmeber, 1973
Concentration (PPM)
100 200
300
15
30
o
ex
0)
Q
45
60
75
HIGH RATE
LOW RATE
--A-- INITIAL AVERAGE
Soil profile calcium concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
KJ
N>
ro
O
September, 1973
Concentration (PPM)
25 50
151-
a)
Q
15 h-
30 h
45 h
60 L
December, 1973
Concentration (PPM)
25 50
75
I
-£r-
HIGH RATE
LOW RATE
CONTROL CHECK
o
Figure 81. Soil profile magnesium concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
September, 1973
Concentration (PPM)
30 60
90
December, 1973
Concentration (PPM)
30 60
90
15
30
S
CJ
s—'
fX
0)
n
45
60
15
75
Figure 82.
30
0)
Q
45
60 h
75 L
HIGH RATE
LOW RATE
INITIAL AVERAGE
Soil profile phosphorous concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
ISJ
OJ
o
September, 1973
Concentration (PPM)
40 80
T
(X
0)
o
15
30
45 H
60 h-
December, 1973
Concentration (PPM)
40 80
T
120
HIGH RATE
LOW RATE
INITIAL AVERAGE
75 «-
Soil profile potassium concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
September, 1973
Concentration (PPM)
10 20
15
30
ho
u>
-------
September, 1973
Concentration (PPM)
2Q 40
N)
OJ
o
ex
OJ
n
December, 1973
Concentration (PPM)
20 40
60
15
30
ex
0)
(=>
45 _
60
75
HIGH RATE
LOW RATE
CONTROL CHECK
Figure 85. Soil profile chloride concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
had already occurred by September and both high and low rate plots were
two-to three-fold higher in chloride concentration than control cores.
During the following three winter months chloride leaching occurred
with the December profiles nearly equal for the high and low application
plots. However, December control core soil chloride levels remained
lower than those in the effluent plots.
Trace elements Fe and Mn in the applied waste were also monitored along
with the heavy metals Cu and Zn. For manganese, the high and low rate
plots had nearly the same soil concentrations, Figure 86. These levels
were greater than the control plots in the upper 50 cm of soil. There
was little shift in soil Mn concentrations between September and December,
1973. Iron profiles in September were the same for all the waste applied
and control plot cores, Figure 87. The same uniformity among cores was
present in December but the absolute levels were about twice those
found in September for reasons not apparent.
Copper levels reduced by 90 percent from the high growth stimulant
levels in the swine feed on a ppm basis by lagoon pretreatment were
only higher by 0.2 ppm in the upper 20 cm of soil in the heavy rate
plots over the low rate plots, Figure 88. Zinc also evidenced a
slightly higher surface zone accumulation for the high rate as com-
pared to the low application rate, Figure 89, for both the September
and December profiles. Control core data for copper and zinc were
somewhat erratic. These trace elements were applied at quite low
rates so that accumulation or leaching trends indicated after one year
of data must be verified by further study.
The efficacy of detailed profile analysis in comparison to lumped
analysis of the upper 75 cm of soil profile are demonstrated by com-
paring results in Tables 52 and 53 and Figures 78 - 89. Greater
sensitivity in detecting initial trends was available with the soil
concentrations profile.
Summarizing the soil accumula tion results and conclusions for the total
75-cm profile, four waste constituents, N03-N, K, Na, and Cl were deduced
to have increased significantly between low and high rate waste loadings
above the control or initial soil levels on an overall mass balance.
The other soil parameters, TKN, Ca, Mg, P, Cu, Zn, Fe and Mn were not
significantly affected by loading rate although some were at higher
levels than initial or control cores or had slightly elevated surface
concentrations. Both Mg and Ca evidenced a surface accumulation but
levels returned to control concentrations about three months after
irrigation termination. This second group of constituents were either
not applied at high rates compared to initial levels, were not reliably
detected, or were leached from the upper 75 cm of soil and, hence, were
not recorded as constituents that accumulated in the soil profile.
233
-------
15
30
o
f,
01
Q
45
60
September, 1973
Concentration (PPM)
5 10
75 *-
Figure 86.
15
I
o
CU
Q)
O
15
30
45
60
75
December, 1973
Concentration (PPM)
5 in
15
HIGH RATE
LOW RATE
CONTROL CHECK
Soil profile manganese concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha. ^guon
-------
15
30
0)
Q
45
60
75 -L,
Figure
September, 1973 :
^Concentration (PPM)
25 50
T
75
Cu
0)
n
15
30
45
60
December, 1973
Concentration (PPM)
25
50
75
A
—O- HIGH RATE
—D— LOW RATE
CONTROL CHECK
87.
75
Soil profile iron concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
t-o
UJ
September, 1973
Concentration (PPM)
1 2
15
30
A-
QJ
O
60
75
Figure
30
0)
a
45
60
75 *-
December, 1973
Concentration (PPM)
1 2
D
-A-
HIGH RATE
LOW RATE
CONTROL CHECK
88. Soil profile copper concentrations for first-year application of swine lagoon
effluent, 1973, at 1200 and 300 kg N/ha.
-------
4-1
cx
a>
15
30
45
60
75
September, 1973
Concentration CPPM)
2 4
6
D-
-------
Mass balance evaluations for control core Cu and Zn verified analytical
difficulties and emphasized that these initial results should only be
regarded as trends.
Field Plot Mass Balance
The mass balance on selected waste constituents consisted of comparison
of inputs and those measured pathways or reservoirs for removal for the
studied plant-soil system. The reference compartment or volume used
in the mass balance was the upper 75 cm of soil volume beneath the
triplicate plots receiving high, medium, and low levels of waste appli-
cation. The common calculative basis was kilogram of parameter per
hectare. Input consisted of the liquid which fell on the experimental
plot area. The entire system balance from lagoon to plot to removal
pathway was not used since a discussion of irrigation losses was covered
earlier. The defined removal mechanisms were crop harvest, rainfall
runoff, and soil accumulation. Differences between input and accounted
for removals were labelled as unaccounted material. '
The mass balance results were tabulated for each effluent loading rate,
Tables 54 to 56. Rainfall runoff transport for 40 percent of the total
rainfall runoff volume was found to represent approximately .5 percent to
2.5 percent of the total material applied with little difference between
low and high application rates. Correcting these values by a ratio
of total to analyzed runoff volume theoretically adjusted this loss
mechanism to about 1 percent to 5 percent. During this initial study,
runoff from surrounding plots not receiving effluent was not measured
as a control or background assessment. Nonetheless, the absolute amounts
of rainfall runoff transport were small. Additionally, the low and
medium rates resulted in about the same rainfall runoff parameter amounts
as measured with the high rate.
Crop removals as a percentage of applied material increased as the
effluent application decreased. This was an expected trend since plant
requirements are below applied levels and thus as the amount of applied
nutrients decreased, the amount in the crop as percentage of that
applied increased. The actual kg of the various waste materials measured
in the harvested crop increased by nearly two-fold from the low to the
medium rate and only slightly above the medium rate with the high
application rate. These responses were similar to those described in
the crop section for nitrogen reflecting the shape of the growth-
response curves for these components.
Accumulation in the upper 75 cm of soil was treated as a removal or
storage mechanism. Measurement accuracy depended on the ratio of applied
to existing amounts of material and low rate additions of K, Na, and
Cl were approximately equal to the initial amounts. Thus detection of soil
accumulation was difficult for these components. At the high rate P and
Mg, as well as K, Na, and Cl were applied at amounts greater than the
initial soil contents, hence the reliability of these soil analyses was
increased in the high rate plots.
238
-------
Table 54. OVERALL MASS BALANCE FOR FIRST-YEAR APPLICATION OF SWINE LAGOON EFFLUENT AT THE HIGH RATE
TO A COASTAL BERMUDA GRASS - NORFOLK SOIL PLOT SYSTEM (AVERAGE OF THREE REPLICATES)
Parameter
TKN
P
Ca
Mg
K
Na
Cl
Cu
Zn
Mn
Fe
Input, output, and accumulation, kg/ha
Initial
amount
2,890
145
814
149
262
119
176
10.9
40
36
411
Applied
effluent
1,180
268
282
332
470
762
2.7
3.0
2.1
8.1
Crop
uptake
428
36
72
42
413
26
116
0.25
0.43
1.18
4.1
Rainfall
runoff3
2.2
2.6
2.2
0.9
6.2
2.1
7.8
0.034
--
--
Accumulation,
September, 1973
176
28
-46
73
174
11
427
6.8
--
--
Unaccounted
for
579
204
254
216
612
431
212
-4.5
__
aEvaluated for events totalling about 40 percent of total runoff
-------
Table 55. OVERALL MASS BALANCE FOR FIRST YEAR-APPLICATION OF SWINE LAGOON EFFLUENT AT MEDIUM RATE
TO A COASTAL BERMUDA GRASS - NORFOLK SOIL PLOT SYSTEM (AVERAGE OF THREE REPLICATES)
Parameter
TKN
P
Ca
Mg
K
Na
Cl
Cu
Zn
Mn
Fe
Input, output, and accumulation, kg/ha
Initial
amount
2,264
142
820
118
262
119
176
11
40
36
411
Applied
effluent
592
134
141
166
603
235
382
1.3
1.5
1.0
4.0
Crop
uptake
336
31
53
32
339
24
128
0.21
0.36
0.83
3.7
Rainfall
runoff3
1.3
0.8
1.3
0.4
2.5
1.3
4.5
0.01
__
--
Accumulation,
September, 1973
11.2
2.2
-24
62
44
19
334
15
--
__
Unaccounted
for
243
104
110
71
217
190
-84
-13.9
--
__
4>
O
-------
Table 56. OVERALL MASS BALANCE FOR FIRST-YEAR APPLICATION OF SWINE LAGOON EFFLUENT AT THE LOW RATE
TO A COASTAL BERMUDA GRASS - NORFOLK SOIL PLOT SYSTEM (AVERAGE OF THREE REPLICATES)
Parameter
TKN
P
Ca
Mg
K
Na
Cl
Cu
Zn
Mn
Fe
Input, output, and accumulation, kg/ha
Initial
amount
2,386
144
710
129
315
119
176
11
40
36
411
Applied
effluent
297
67
71
83
301
118
190
0.67
0,78
0.56
2.0
Crop
uptake
207
20
33
18
207
11
87
0.12
0.27
0.50
1.9
Rainfall
runoff3
2.1
1.3
1.1
0.46
4.5
1.4
3.4
0.028
Accumulation,
September, 1973
-11
-37
-58
11
-96
-37
316
3.5
Unaccounted
for
96
83
94
53
184
141
-216
-2.9
-------
The difference between input and accounted for amounts of the various
parameters was termed unaccounted naterial, Tables 54, 55, and 56.
The unaccounted material was attributed to leaching losses from the
studied upper 75 cm of soil. However, in the case of phosphorus,
which is not highly mobile in these soils, the unaccounted material
certainly contained losses to soil states not evaluated by the analyti-
cal tests used. The leaching losses of other ions could not be
separated between vertical losses through the B horizon and more hori-
zontal or downslope losses with water flow along the interface of the
A and B horizons.
Unaccounted material was about the same for the low and medium rate
plots except for nitrogen which increased from 100 kg/ha, to 250 kg/ha.
The high rate plots had nearly a two-fold increase in unaccounted
materials, including nitrogen, over the medium rate plots. Attributing
these losses conclusively to leaching after this first year study was
not possible because the anticipated relative freedom of movement or
mobility of these constituents was not verified. That is, K, Na, and
Cl should have been much more mobile in the soil than the Ca, Mg, or
P. However, the leaching losses as a percentage of the amount applied
or applied plus initially present were not significantly greater for
K, Na, and Cl as compared to P, Ca, and Mg, Table 57. Thus other
factors, yet unsubstantiated, prevented in depth conclusions regarding
mass balances or pathways for removal of waste constituents. Extension
of these studies over several years would allow greater application
levels so that accumulation could be more easily documented. Also,
several years of data would reduce annual variability of crop uptake
and rainfall runoff as related to the total system mass balance.
242
-------
Table 57. PERCENTAGES OF VARIOUS WASTE CONSTITUENTS REMOVED BY THE MAJOR PLANT-SOIL SYSTEM
PATHWAYS FOR THE COASTAL BERMUDA-COASTAL PLAINS EXPERIMENT RECEIVING ONE SEASON
OF SWINE LAGOON EFFLUENT
Crop removal,
percent of input
Rainfall, runoff,
percent of input3
Unaccounted for,
percent of input
Unaccounted for,
percent of input
plus initial
Parameter High Medium Low
TKN 36 57 70
P 13 23 30
Ca 25 37 46
Mg 12 20 22
High Medium Low
0.5 0.6 1.8
1.5 3.5
2.5 4
0.75 0.5 1.4
1.2 1.0 3.8
1.0 1.2 2.8
2.5 3.0 4.5
3.0 2.0 10
High Medium Low
49 41 33
75 78 123
90 79
65 43
51 36
92 81
28 -22
-167 -1,000 -433
High Medium Low
14 8 4
48 38 39
NJ
*-
U)
K
Na
Cl
Cu
34
5
15
9
56
10
33
16
69
10
46
18
2.5
2
134
65
62
121
113
433
23
45
42
73
22
-33
12
25
25
54
-15
-111
12
25
30
60
-59
-25
aScaled up from values measured for 40 percent of runoff by means of a proportional ratio
-------
SECTION VIII
REFERENCES
1. Jones, J. R. Estimated Swine Situation Reports by County
Extension Workers. N. C. State University, Raleigh, N. C. 1972.
2. N. C. Department of Agriculture, USDA Federal-State Crop
Reporting Service. December 31, 1974.
3- Robbins, J. W. D., D. H. Howells, and G. J. Kriz. Role of Animal
Wastes in Agricultural Land Runoff. Water Pollution Control
Research Series. EPA Report 13020 DGX. August 1971. 114 p.
4. Muehling, A. J. Swine Housing and Waste Management. A Research
Review, AEnd-873. Department of Agricultural Engineering,
University of Illinois, Urbana - Champaign. August 1969. 91 p.
5. Taiganides, E. P. Characteristics and Treatment of Wastes From
a Confinement Hog Production Unit. Unpublished Ph.D. Thesis,
Department of Agricultural Engineering, Iowa State University,
University Microfilm, Ann Arbor, Michigan. 1963.
6. Arial, J. D. The Influence of Feed Additives on the Biochemical
Oxygen Demand Analysis for Swine Wastes. M.S. Thesis, N. C.
State University, Raleigh, N. C. 1970.
7. Humenik, F. J., M. R. Overcash, and L. B. Driggers. Swine
Production Industry - Waste Characterization and Management.
Department of Biological and Agricultural Engineering, N. C.
State University, Raleigh, N. C. 1973. 56 p.
8. Scholze, H. G. Systems for the Dehydration of Livestock Wastes,
A Technical and Economical Review. Proc. Int'l. Symp. on
Livestock Wastes. ASAE PROC-271:27-29. 1971.
9. Van Arsdall, R. N. The Economic Value of Manure from Confinement
Finishing of Hogs. University of Illinois Agricultural Experiment
Station Bulletin 687, Urbana, Illinois. May 1962.
10. McKinney, Ross E. Manure Transformations and Fate of Decomposition
Products in Water. Proceedings of the Conference Concerning the
Role of Agriculture in Clean Water, Iowa State University, Ames,
Iowa. 1970.
244
-------
11. Dale, A. C. Aerobic Treatment of Animal Wastes. ASAE Paper No.
67-927A. 1967.
12. Hart, S. A. and M. E. Turner. Lagoons for Livestock Manure.
J. Water Pol. Cont. Fed. 37:1578-1596, November 1965.
13. Eby, H. J. Design Criteria for Manure Lagoons. Paper presented
at the Annual Meeting of American Society of Animal Meat Producers,
Chicago, Illinois. 1961.
14- Eby, H. J. Manure Disposal Lagoons. USDA, ARS 42-75. 1963.
15. Eby, H. J. Correcting Malfunctioning Lagoons. USDA, Agricultural
Research Service, Agricultural Engineering Research Division. Mimeo.
1964. 8 p.
16. Clark, C. E. How Are Lagoons Working on Hog Farms? ASAE Paper
No. 64-419. 1964.
17. Clark, C. E. Hog Waste Disposal by Lagooning. J. San. Eng.
Div., ASCE. 91:27-41, June 1965.
18. Dale, A. C. Oxidation Lagoons. In: Farm Animal Wastes, 1969.
Regional Bulletin NC-69. 1969.
19. Dornbush, J. N. State of the Art - Anaerobic Lagoons. 2nd
International Symposium for Waste Treatment Lagoons, Kansas City,
Missouri. 1970. pp. 382-387.
20. Oswald, W. J. Advances in Anaerobic Pond Systems Design. Advances
in Water Quality Improvement, University of Texas Press, Austin,
Texas. 1968. p. 409.
21. Loehr, R. C. Pollution Implications of Animal Wastes - A Forward -
Oriented Review. U. S. Dept. Int., Fed. Water Pol. Control Adm.,
R. S. Kerr Water Res. Center. Ada, Oklahoma. 1968.
22. Loehr, R. C. Anaerobic Lagoons: Considerations in Design and
Application. Transactions ASAE. 11:320-330. March 1968.
23. Taiganides, E. P. Theory and Practice of Anaerobic Digesters
and Lagoons. 2nd National Poultry and Litter Waste Management
Seminar Proceedings, College Station, Texas, 1968. p. 200-221.
24. Ricketts, R. L. Lagoons for Disposal of Wastes from Hog Feeding
Floors. ASAE Paper No. 60-402. 1960.
25. Jedele, D. G. and E. L. Hansen. Handling Hog Manure As a Liquid.
University of Illinois College of Agriculture Circular 820. 1960.
26. Eby, H. J. Anaerobic Lagoons - Theory and Practice. 2nd National
Symposium on Poultry Industrial Waste Management Proceedings,
Lincoln, Nebraska. 1964. p. 77-91.
245
-------
27. Soil Conservation Service. Interim Specifications - Disposal
Lagoons, Advisory ENG-5, USDA, Washington, B.C. 1970.
28. Soil Conservation Service. National Engineering Standard,
Disposal Lagoons. 1971.
29. Willrich, T. L. Primary Treatment of Swine Wastes by Lagooning.
Proc. Nat'l. Symp. on Animal Waste Management. ASAE Pub.
SP-0366:70-74. 1966.
30. Dornbush, J. N. and J. R. Anderson. Lagooning of Livestock in
South Dakota. Proc. 19th Ind. Waste Conference, Purdue University.
1964. p. 317-325.
31. Curtis, D. R. Design Criteria for Anaerobic Lagoons for Swine
Manure Disposal. Proc. Nat'l. Symp. on Animal Waste Management.
ASAE Pub SP-0366:75-80. 1966.
32. Midwest Plan Service. AED-1. Ames, Iowa. August 1969. 2 p.
33. Lynn, H. P. The Effects of Loading Rates on the Design and
Operation of Anaerobic Swine Lagoons. Summary of Unpublished
M.S. Thesis, Clemson University, Clemson, South Carolina. 1968.
34. Miner, J. R. Anaerobic Treatment of Animal Wastes. In: Farm
Animal Wastes, 1969. Regional Bulletin NC-69. 1969.
35. Barr, H. T., B. A. Tower, C. D. Steelman, and L. J. Cabes, Jr.
Farm Lagoons in Louisiana. Louisiana State University
Agricultural Experiment Station Bulletin No. 638. 1969.
36. Baldwin, L. B. and R. A. Nordstedt. Design Procedures for
Animal Waste Treatment Lagoons in Florida. Paper No. SE 71-214
presented at the Southeast Regional Meeting of ASAE, Jacksonville,
Florida. 1971.
37. Hermanson, R. E. and H. Watson. Manure Disposal Lagoons.
Agricultural Experiment Station Circular R-6, Auburn University,
Auburn, Alabama. 1968.
38. State of Tennessee. Engineering Standard, Animal Waste Disposal
System. Department of Public Health, Nashville, Tennessee 1973.
39. American Society of Agricultural Engineers. Tentative Guidelines
For Lagoons. Development being coordinated for SE-412 Committee
by C. L. Barth, Clemson University, Clemson, South Carolina.
40. Nordstedt, R. A. and C. L. Barth. Animal Waste Lagoons—Research
and Design in the Southeast. Paper presented at the Southeast
Regional Meeting of ASAE, Memphis, Tennessee. 1974.
246
-------
41. Nordstedt, R. A., L. B. Baldwin, and C. C. Hortenstine. Multistage
Lagoon Systems for Treatment of Dairy Farm Waste—Progress Report.
Presented at the Southeast Regional Meeting of ASAE, Jacksonville,
Florida. Paper No. SE 71-403. 1971.
42. Barsom, George, Lagoon Performance and the State of Lagoon
Technology. Environmental Protection Technology Series EPA-R2-
73-144. 1973.
43. Missouri Basin Engineering Health Council. Waste Treatment
Lagoons - State-of-the-Art. Water Pollution Control Research
Series, Project 17090 EHS, EPA, Washington, D. C. July 1971.
44. WPCF Manual of Practice No. 11. Operation of Wastewater
Treatment Plants. Water Pollution Control Federation, Washington,
D.C. 1970.
45. McGhee, T. J. Volatile Acid Concentration in Batch Fed
Anaerobic Digesters. Water and Sewage Works. 118:130. May 1971.
46. Dale, A. C. Status of Dairy Cattle Waste Treatment and
Management Research. Proceedings National Symposium on Animal
Waste Management, Warrenton, Virginia. 1971. p. 85-95.
47. Loehr, R. C. Alternatives for Treatment and Disposal of Animal
Wastes. J Water Pol. Cont. Fed. 43:668-678. 1971.
48. Jones, D. D., D. 0- Day, and A. C. Dale. Aerobic Treatment of
Livestock Wastes. Experiment Station Bulletin 737, University
of Illinois, Urbana - Champaign, Illinois. 1970.
49. Irgens, R. L. and D. L. Day. Aerobic Treatment of Swine Wastes.
Proc. Nat'l. Symposium on Animal Waste Management. ASAE Pub.
SP-0366:58-30. 1966.
50. McKinney, R. E. and R. Bella. Water Quality Changes in Confined
Hog Waste Treatment. Kansas Water Resources Res. Inst. Report
No. 24, University of Kansas, Lawrence. December 1967.
51. McCalf and Eddy, Inc. Wastewater Engineering. McGraw-Hill,
New York. 1972.
52. Grune, N. N. , R. E. Cossitt, and R. H. Philp. An Aerobic Process
Automation by Orp Conductivity and Gas Chromatography. Purdue
University Engineering Extension Service No. 92-95. 1957. p.
604-635.
53. Ludington, D. C. Oxidation-Reduction Potential, Control by
Aeration and Influence on Odor Production and Degradation of
Chicken Manure. Unpublished Ph.D Dissertation, Purdue University
Lafayette, Indiana. 1968.
247
-------
54. Converse, J. C. K. L. Day, J. T. Pfeffer, and B. A. Jones, Jr.
Aeration with ORP Control to Supress Odors Emitted from
Liquid Swine Manure Systems. Proc. Int'l. Symposium on
Livestock Wastes. ASAE PROC-271:267-271. 1971.
55. Adriano, D. C., P. F. Pratt, and S. E. Bishop. Fate of Organic
Forms of N and Salt from Land-Disposed Manures from Dairies.
Livestock Waste Management and Pollution Abatement, Proc.
Int. Symposium on Livestock Wastes, ASAE, St. Joseph, Michigan.
1971. p. 243-246.
56. Mathers, A. C., B. A. Stewart, J. D. Thomas, and B. J. Blair.
Effects of Cattle Feedlot Manure on Crop Yields and Soil
Conditions. USDA Southwestern Great Plains Research Center,
Technical Report #11, Symposium on Animal Waste Management,
Bushland, Texas. 1973.
57. Pratt, P. F., F. E. Broadbent, and J. P. Martin. Using Organic
Wastes as Nitrogen Fertilizers. California Agriculture.
June 1973.
58. Koelliker, J. K. and J. R. Miner. Use of Soil to Treat
Anaerobic Lagoon Effluent Renovation as a Function of Depth
and Application Rate. Transactions of the ASAE. 13:496-499.
1970.
59. Olsen, R. J., R. F. Hensler, and 0. J. Attoe. Effect of Manure
Application, Aeration, and Soil pH on Soil Nitrogen
Transformations and on Certain Soil Test Values. Soil Science
Society Amer. Proc, 34:222-225. 1970.
60, Wallingford, G. W. Effects of Solid and Liquid Beef Feedlot
Wastes on Soil Characteristics and on Growth and Composition
of Corn Forage. Ph.D. Thesis, Kansas State University, Manhattan,
Kansas. 1974.
61. Meek, B. D., A. J. MacKenzie, T. J. Donovan, and W. F. Spencer.
The Effect of Large Applications of Manure on Movement of Nitrate
and Carbon in an Irrigated Desert Soil. J Envron. Quality.
3:253-258, 1974.
62. Hedges, J. D., E. T. Kornegy, and D. C. Marteno. Effect of Applying
Swine Feces on Soil and Plant Mineral Levels. Virginia Polytechnic
Institute and State University, Blacksburg. J. Animal Sci.
37:281. 1973.
63. Chaney, R. L. Crop and Food"Chain Effects of Toxic Elements in
Sludges and Effluents. Proc. Conf. on Recycling Municipal Sludges
and Effluents on Land. EPA, Champaign, Illinois. 1973.
248
-------
64. Humenik, F. J. Swine Waste Characterization and Evaluation
of Animal Waste Treatment Alternatives. Water Resources Research
Institute of University of North Carolina, Report No. 61.,
Raleigh, N. C. 1972.
65. Miner, J. R. and T. L. Willrich. Livestock Operations and Field-
Spray Manure as Sources of Pollutants. In: Agricultural
Practices and Water Quality, Iowa State University, Ames, Iowa.
1970. p. 231-240.
66. Robbins, J. W. D., D. H. Howells, and G. J. Kriz. Stream
Pollution from Animal Production Units. J. Water Pol. Cont.
Fed. 44:1536-1544, August 1972.
67. McCaskey, T. A., G. H. Rollins, and J. A. Little. Water Pollution
by Dairy Farm Wastes as Related to Method of Waste Disposal.
WRRI Bui. 18, Auburn University, Auburn, Alabama. 1973.
68. Miner, J. R., L. R. Fina, J. W. Funk, R. I. Lipper, and G. H.
Larson. Stormwater Runoff from Cattle Feedlots. Proc. Nat'l.
Symposium on Animal Waste Management. ASAE Pub. No. SP-0366:
23-27. 1966.
69. Gilbertson, C. B., T. M. McCalla, J. R. Ellis, 0. E. Cross, and
W. R. Woods. The Effect of Animal Density and Surface Slope
on Characteristics of Runoff Solid Wastes and Nitrate Movement
on Unpaved Beef Feedlots. University of Nebraska Agr. Exp.
Sta. Pub. SB508. 1970. p. 23.
70. Madden, J. M. and J. N. Dornbush. Measurement of Runoff and
Runoff Carried Waste from Commercial Feedlots. Proc. Int'l.
Symposium on Livestock Wastes. ASAE PROC.-271:44-47. 1971.
71. McCalla, T. M., J. R. Ellis, C. B. Gilbertson, and J. R. Woods.
Chemical Studies of Solids, Runoff, Soil Profile and Ground-
water from Beef Cattle Feedlots at Mead, Nebraska. Proc.
Cornell Agr. Waste Management Conference. 1972. p. 211-233.
72. APHA. Standard Methods for the Examination of Water and Waste-
Water. 12 ed. American Public Health Association, Inc., New
York. 1965.
73. Overcash, M. R., D. L. Reddell, and D. L. Day. Chemical Analysis.
ASAE Paper No. 74-4546. 1974.
74. Humenik, F. J. and M. R. Overcash. Discussion of Paper, Evaluation
of Methods for the Analysis of Physical, Chemical and Biochemical
Properties of Poultry Wastewater, by Pfakasam, et al. ASAE
Workshop, Chicago, 111. December 1972.
249
-------
75. Busch, A. W. Aerobic Biological Treatment of Waste Waters. Houston,
Oligodynamics Press, 1971. 416 p.
76. Taylor, E. W., N. P. Burman, and C. V. Oliver. Membrane Filtration
Technique Applied to the Routine Bacteriological Examination of
Water. Journal of the Institute of Water Engineers. 9:248-254.
January 1955-
77. Kabler, P. W. and H. F. Clark. Coliform Group and Fecal Coliform
Organisms as Indicators of Pollution in Drinking Water. Journal
American Water Works Association. 52:1577-1579. February 1960.
78. Geldreich, E. E., H. F. Clark, C. B. Huff, and L. C. Best. Fecal
Coliform Organism Medium for the Membrane Filter Technique. Journal
American Water Works Association. 57:208-214. February 1965.
79. Overcash, M. R., F. J. Humenik, and L. B. Driggers. Swine Production
and Waste Management: State of the Art0 Proc. Third International
Symposium on Livestock Waste Management, In Press. Urbana-Champaign,
Illinois. April 1975.
80. Jett, S. C., H. E. Hamilton, and I. J. Ross. Settling Characteristics
of Swine Manure. ASAE Trans., In Press.
81. Smith, R. J., C. V. Booram, and T. E. Hazen. Some Chemical and
Physical Aspects of Phosphate Precipitation from Anaerobic Liquors
Derived from Animal Waste Treatment Lagoons. ASAE Paper No. 73-4522.
(Presented at ASAE Meeting. Chicago, Illinois. 1973).
82. Miner, J. R. and J. I. Koelliker. Resorption of Ammonia From Anaerobic
Lagoons. Iowa Agric. and Home Economics Experiment Station, Journal
Paper No. J-6873. 1973.
83. EPA. Development Document for Proposed Effluent Limitations
Guidelines and New Source Performance Standards for the Feedlots
Point Source Category. Office of Air and Water Programs, Washington,
D.C. August 1973. 302 p.
84. Moore, J. A., M. R. Overcash, and T. B. S. Prakasam. Animal Waste
Sample Preservation and Preparation. ASAE Paper No. 74-4545.
(Presented at ASAE Meeting. Chicago, Illinois. 1974).
85. Lutz, J. F. Movement and Storage of Water in North Carolina Soils'.
N. C. Agricultural Experiment Station Sorts Information Series No.
15. April 1970.
86. Overcash, M. R., J. W. (Jilliam, and F. J. Humenik. Lagoon Pretreat-
ment: Selected Heavy Metals and Cation Removals, In Press.
250
-------
87. Larsen, V., J. H. Axley, and G. L. Miller. Agricultural Wastewater
Accommodation and Utilization of Various Forages. Univ. of Ind.
Water Resources Research Center. Technical Report No. 19. 1972.
88. Doss, B. D. , D. A. Ashley, 0. L. Bennett, and R. M. Patterson.
Interactions of Soil Moisture, Nitrogen, and Clipping Frequency
on Yield and Nitrogen Content of Coastal Bermuda Grass. Agron. J.
58:510-512. 1966.
89. Woodhouse, W. W., Jr. Long-Term Fertility Requirements of Coastal
Bermuda II Nitrogen, Phosphorus, and Lime. Agron. J. 61:251-256.
1969.
90. Woodhouse, W. Wo, Jr. Long-Term Fertility Requirements of Coastal
Bermuda I. Potassium. Agron. J. 60:508-512. 1968.
91. Burton, G. W., W. S. Wilkinson, and R. L. Carter. Effect of Nitrogen,
Phosphorus, and Potassium Levels and Clipping Frequency on Forage
Yield and Protein, Carotene, and Xanthophyll Content of Coastal
Bermuda Grass. Agron. J. 61:60-63. 1969.
92. Murrmann, R. P. and F. R. Koutz. Role of Soil Chemical Processes
in Reclamation of Wastewater Applied to Land. In: Wastewater
Management by Disposal on Land, Corps of Eng. Cold Region
Research and Eng. Lab., Hanover, N. H. 1972.
93. Hunt, P. Microbial Responses to the Land Disposal of Secondary-
Treated Municipal-Industrial Wastewater. In: Wastewater Management
by Disposal on Land, Corps of Eng. Cold Region Research and Eng.
Lab., Hanover, N. H. 1972.
94. Burd, J. S. and J. C. Martin. Secular and Seasonal Changes in the
Soil Solution. Soil Science 18, 151-167, 1924.
251
-------
SECTION IX
PUBLICATIONS ASSOCIATED WITH PROJECT RESULTS
Humenik, F. J., R. W. Skaggs, C. R. Willey, and D. Huisingh.
Evaluation of swine waste treatment alternatives. Proceedings
Cornell University Conference on Agricultural Waste Management.
pp 341-352. 1972.
Humenik, F. J. and M. R. Overcash. Analyzing Physical and Chemical
Properties of Liquid Wastes. ASAE Proceedings SP-0275, Standardizing
Properties and Analytical Methods Related to Animal Waste Research.
pp 114-182. St. Joseph, Michigan, 1975.
Humenik, F. J. The latest in swine waste disposal. Proceedings North
Carolina Pork Producers Conference, pp 43-46. January 1973.
Humenik, F. J., M. R. Overcash, and L. B0 Driggers. Swine production
industry - waste characterization and management. 56 pp.
Department of Biological and Agricultural Engineering, North
Carolina State University, Raleigh, N. C. 1973.
Swine Waste Management Alternatives. North Carolina Agricultural
Extension Service Circular 569. Raleigh, N. C. September 1973.
Dairy Waste Management Alternatives. North Carolina Agricultural
Extension Service Circular 568. Raleigh, N. C. September 1973.
Poultry Waste Management Alternatives. North Carolina Agricultural
Extension Service Circular 570. Raleigh, N. C. September 1973.
Beef Cattle Waste Management Alternatives. North Carolina Agricultural
Extension Service Circular 571. Raleigh, N. C. March 1974.
Regulatory Criteria for Animal Waste Management. North Carolina
Agricultural Extension Service Leaflet 1910 Raleigh, N. C.
May 1974.
Overcash, M. R. Economics of alternative wastewater treatment systems.
Proceedings Workshop on Land Disposal of Wastewaters. Report No.
91. pp 69-74. Raleigh, N. C, November 1973.
Humenik, F. J. Agricultural Waste Management Techniques. Treatment
systems for animal, agricultural, and municipal wastes. Proceedings
Workshop on Land Disposal of Wastewaters. Water Resources
252
-------
Research Institute of the University of North Carolina. Report
No. 91. pp 91-104. Raleigh, N. G. November 1973.
Humenik, F. J. Effect of effluent discharge guidelines on the
producer. Proceedings North Carolina Irrigation Conference.
pp 11-20. Raleigh, N. C. November 1973.
Overcash, M. R. Research to comply economically with environmental
guidelines. Proceedings North Carolina Irrigation Conference.
pp 21-32. Raleigh, N. C. November 1973.
Huraeaik, F. J. New developments in swine waste management. Proceedings
from Virginia Pork Industry Conference, pp 1-14. Sandston,
Virginia. December 1973.
Humenik, F. J., M. R, Overcash, L. B. triggers, and G. J. Kriz.
Cleaning the animal farm environment. Environmental Science and
Technology 8(12):984-989. 1974.
Humenik, F. J. Animal waste as a source. Proceedings U.S. EPA and
Extension Committee on Organizational Policy Workshop on Non-
point Source Water Pollution Control, pp 113-123. Washington,
D. C. September 1974.
Moore, J. A., M. R. Overcash, T. B. S. Prakasam. Animal waste sample
preservation and preparation. ASAE Paper No, 74-4545. Chicago,
Illinois. December 1974.
Overcash, M. R., D. L. Reddell, D. L. Gay, and A. G. Hashimoto.
Chemical analyses. ASAE Paper No. 74-4546. Chicago, Illinois.
December 1974.
Sneed, R. E., F. J. Humenik, and M. R. Overcash. Design considerations
for land application waste treatment systems. ASAE Paper 74-2556.
Chicago, Illinois. December 1974.
Howell, E. S., M. R. Overcash, and F. J. Humenik. Unaerated swine
waste lagoon response to loading intensity and frequency. ASAE
Paper 74-4514. Chicago, Illinois. December 1974.
Humenik, F. J. An overview of land application of agricultural waste
in the Southeast, Proceedings North Carolina Irrigation
Conference, pp 1-10. Charlotte, N. C. December 1974.
Humenik, F. J., and J. C. Barker. New developments in swine waste
management. Proceedings North Carolina Pork Producers Conference.
pp 24-27. January 1975.
Humenik, F. J. System design considerations - process. Proceedings
Southeastern Workshop on Land Application of Municipal, Industrial,
and Agricultural Wastewater. North Carolina Agricultural
Extension Service and Agricultural Experiment Station; Water
253
-------
Resources Research Institute of the University of North Carolina.
pp 4.1-4.14. Raleigh, N. C. January 1975.
Humenik, F. J. Animal waste land application systems. Proceedings
Sprinkler Irrigation Association, Atlanta, Georgia, February 1975.
Overcash, M. R. and F. J. Humenik. Concepts for pretreatment-land
application for small municipal systems. Second National
Conference on Complete Water Reuse. AICHE. Chicago, Illinois.
May 1975.
Overcash, M. R. and F. J. Humenik. Agriculture related pretreatment
land application systems. Second National Conference on Complete
Water Reuse. AICHE. Chicago, Illinois. May 1975.
Axtell, R. C., D. H. Rutz, M. R. Overcash and F, J. Humenik. Mosquito
production and control in animal waste lagoons. Third International
Symposium on Livestock Wastes, Univ. of 111., Amer. Soc, Agr. Eng.,
PROC - 257:15-18, 21. Managing Livestock Wastes.
Overcash, M. R., F. J. Humenik and L. B. Driggers. Swine production
and waste management. State-of-the-Art, Managing Livestock Wastes,
Proc. Third International Symposium on Livestock Wastes, Univ.
of 111., Amer. Soc. Agr. Eng., PROC - 257:154-159, 163. April 1975.
Humenik, F. J., R. E. Sneed, M. R. Overcash, J. ,C. Barker and G. D.
Wetherill. Total waste management for a large swine production
facility. Managing Livestock Wastes. Proc. Third International
Symposium on Livestock Wastes, Univ. of 111. Amer. Soc. Agr.
Eng. PROC - 257:168-173, April 1975.
Cummings, G. A., J. C. Burns, R. E. Sneed, M. R. Overcash and F. J.
Humenik. Plant and soil effects of swine lagoon effluent
applied to coastal bermuda grass. Managing Livestock Wastes,
Proc. Third International Symposium on Livestock Wastes, Univ.
of 111., Amer. Soc. Agr. Eng., PROC - 257:598-601, April 1975.
Humenik, F. J., M. R. Overcash and R. Miller. Surface aeration:
design and performance for swine lagoons. Managing Livestock
Wastes, Proc. Third International Symposium on Livestock Wastes,
Univ. of 111. Amer. Soc. Agr. Eng. PROC - 257:568-571, April
1975.
Howell, E. S. Design criteria for swine waste lagoons. Ph.D Thesis.
North Carolina State University, Raleigh, N. C. In preparation.
Miller, T. D. Aeration of animal waste lagoons for odor control and
stabilization. M.S. Thesis. North Carolina State University,
Raleigh, N. C. In preparation.
254
-------
Overcash, M. R., J. W. Gilliam, F. J. Humenik and P. W. Westerman. Lagoon
pretreatment: selected heavy metal and cation removals. Submitted
to J. Water Pollution Control Federation.
Overcash, M. R., E. S. Howell, F. J. Humenik and P. W. Westerman. Swine
waste lagoon response to loading frequency and intensity. Submitted
to Transactions ASAE.
255
-------
SECTION X
APPENDICES
Page No.
Al Imhoff Cones - First Experiment 257
A2 Imhoff Cones - Second Experiment 266
A3 14"! Reactors - Mass Balance 275
Bl Pilot-Scale Lagoon Temperature Data-°C 279
B2 Odor Rating Sheet 280
B3 Typical Lagoon Profile Data Obtained 281
with Curved Tube Technique
B4 Field Pilot-Scale Reactor Temperature, Dissolved 282
Oxygen at Surface and Bulk Supernatant TOG
B5 Hydrogen Ion Concentration Data 283
Cl Soil Concentration Profiles 287
256
-------
APPENDIX AI. IMHOFF CONES - FIRST EXPERIMENT
I.
Reference JLoading rate, no sludge,
once per week loading frequency,
1NS1
TKN
Parameter
TOC COD
o-P04-P
A. Inputs
1) Initial load
Volume, liters
Concentration, mg/1
Amount, g
% of feed
7» of total input
2) Feed
Volume, liters
Concentration, mg/1
Amount, g
% of feed
% of total input
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
% of feed
7o of total input
2) Drained Sludge
Volume, liters
Concentration, mg/1
Amount, g
% of feed
% of total input
C. Accumulation
1) Final Supernatant
Volume, liters
Concentration, mg/1
Amount, g
% of feed
70 of total input
.450
1,833
.825
100
45.3
.450
424
.191
23.2
10.5
.353
1,020
.360
43.6
19.8
.670
225
.151
18.3
8.3
1
1,750
1.75
44.8
30.9
.450
8,667
3.9
100
69.1
.450
931
.419
10.7
7.4
.353
8,754
3.09
79.2
54.7
.670
781
.523
13.4
9.3
1
5,504
5.5
39.9
28.5
.450
30,612
13.78
100
71.5
.450
2,055
.925
6.7
4.8
.353
22,776
8.04
58.3
41.7
.670
1,251
.838
6.09
4.3
.450
510
.229
100
67.6
.450
67
.030
13.1
8.8
.353
637
.225
98.3
66.4
.670
97
.065
28.4
19.2
257
-------
APPENDIX Al (Continued)
Parameter
TKN TOG COD o-PC^-P
1
45
.045
19.6
13.3
II.
2) Final Cleanout of Walls
Volume, liters
Concentration, mg/1
Amount, g
% of feed
% of total input
1
63
.063
7.6
3.5
Reference loading rate, controlled sludge
once
1CS1
A.
per week loading frequency
Inputs
1) Initial load
(a) Supernatant
Volume, liters
Cone. , mg/1
Amount, g
J
.990
995
.985
7> of feed 119.4
7o of total input
(b) Sludge
Volume, liters
Cone. , mg/1
Amount, g
% of feed
7o of total input
2) Feed
Volume, liters
53.5
.010
1
460
.460
11.2
8.1
level,
.990
1,750
1.733
44.4
30.0
.010
3,200 14,500
.032
3.9
1.7
.450
Concentration, mg/1 1,833
B.
Amount, g
7<, of feed
7o of total input
Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amoun t , g
7o of feed
7» of total input
.825
100
44.8
.500
370
.185
22.4
10.0
.145
3.7
2.5
.450
8,667
3.9
100
67.5
.500
842
.421
10.8
7.3
].
1,117
1.12
8.1
5.8
.990
5,504
5.449
39.5
27.7
.010
46,900
.469
3.4
2.4
.450
30,612
13.78
100
69.9
.500
2,386
1.193
8.6
6.1
.990
110
.109
47.6
32. 1
.010
200
.002
.87
.59
.450
510
.229
100
67.4
.500
60
.030
13.1
8.8
258
-------
APPENDIX Al (Continued)
Parameter
TKN TOG COD o-PO^-P
2) Drained sludge
Volume, liters .181 .181 _181 >lgl
Concentration, mg/1 1,554 10,867 40,840 1,337
Amount, g .283 1.967 7.392 .242
% of feed 34.3 50-4 53_6 iQ5^
7, of total input 15.4 34^ 37.5 71^2
C. Accummulation
1) Final supernatant
Volume, liters .735 ,735 ^735 735
Concentration, mg/1 148 484 1,687 67
Amount, g .109 .356 1.240 .049
% of feed 13.2 9.1 9.0 21.4
7o of total input 5.9 6.2 6,3 14.4
2) Final cleanout of walls,
etc.
Volume, liters 1 1-1 1
.Concentration, mg/1 104 ., 780 1,563 77
Amount, g •, .104 .780 ,1.563 .077
% of feed 12.6 20,0 ,11.3 33.6
% of total input 5.6 13.5 7.9 22.6
III. Reference loading rate, accumulated sludge,
once per week loading frequency,
1AS1
A. Inputs
, 1) Initial .load
(a) Supernatant
Volume, liters .990 .,990 .990 .990
Cone., mg/1 995 1,750 5,504 110
Amount, g . .985 1.733 5^.449 .109
. 7» of feed 119.4,, 44.4 39,5 47.6
% of -total input 53.5 30.0 27.7 32.1
Sludge;
Volume, liters .010 .010 .010 .010
Cone., mg/1 3,200 14.,500 46,900 200
Amount, g .032 .145 .469 .002
% of feed 3.9 3.7 3.4 .87
70 of total input 1.7 2.5 2.4 .59
259
-------
APPENDIX Al (Continued)
Parameter
2) Sludge
Volume, liters
Concentration, mg/1
Amount, g
7, of feed
7, of total input
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
?o of feed
70 of total input
C. Accumulation
TKN
.450
1,833
.825
100
44.8
.580
417
.242
29.3
13.1
1) Final sludge
Volume, liters .105
Concentration, mg/1 1,962
Amount, g .206
% of feed 25.0
% of total input 11.2
2) Final supernatant
Volume, liters .650
Concentration, mg/1 157
Amount, g .102
% of feed 12.4
% of total input 5.5
3) Final Cleanout of walls,
etc.
Volume, liters 1
Concentration, mg/1 125
Amount, g .125
% of feed 15.2
% of total input 6.8
TOC
.450
8,667
3.9
100
67.5
.580
819
.475
12.2
8.2
.105
14,800
1.554
39,8
26.9
-650
720
.468
12.0
8.1
1
860
.860
22.1
14.9
COD
.450
30,612
13.78
100
69-9
.580
2,298
1.333
9.7
6.7
.105
38,676
4.061
29.5
20.6
.650
1,442
.937
6.8
4.8
I
2,207
2.207
16.0
11.2
o-P04-P
.450
510
.229
100
67.4
.580
62
.036
15.7
10.6
.105
2,180
.229
100.0
67.4
.650
85
.055
24.0
16.2
1
115
.115
50.2
33.8
260
-------
APPENDIX Al (Continued)
Parameter
TKN TOC COD o-PO,-P
4
IV. Reference loading rate, no sludge,
3 times per week loading frequency,
1NS3
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters 1 111
Cone., mg/1 995 1,750 5,504 110
Amount, g .995 1.75 5.5 .110
7o of feed 121.0 44.8 39.9 48.0
7. of total input 54.7 30.9 28.5 32.4
2) Feed
Volume, liters .450 .450 .450 .450
Concentration, mg/1 1,833 8,667 30,612 510
Amount, g .825 3.9 13.78 .229
7o of feed 100 100 100 100
7o of total input 45.3 69.1 71.5 67.6
B. Outputs
1) Samples
Volume, liters .780 .780 .780 .780
Concentration, mg/1 506 669 1,963 44
Amount, g .395 .522 1.531 .034
7» of feed 47.9 13.4 11.1 14.8
% of total input 21.7 9.2 7.9 10.0
2) Drained sludge
Volume, liters .492 .492 .492 .492
Concentration, mg/1 1,000 6,262 19,041 612
Amount, g .492 3.081 9.368 .301
70 of feed 59-6 79.0 68.0 131.4
7o of total input 27.0 54.5 48.6 88.8
C. Accumulation
1) Final Supernatant
Volume, liters .295 .295 .295 .295
Concentration, ing/I 140 675 2,437 54
Amount, g -041 .199 .719 .016
% of feed 5.0 5.1 5.2 7.0
% of total input 2.3 3.5 3.7 4.7
261
-------
APPENDIX Al (Continued)
Parameter
TKN TOG COD o-PC^-P
2) Final cleanout of walls,
etc.
Volume, liters 1 111
Concentration, mg/1 40 100 360 40
Amount, g .040 .100 .360 .040
% of feed 4.8 2.6 2.6 17.5
% of total input 2.2 1.8 1.9 11.8
V. Reference loading rate, controlled sludge level,
3 times per week loading frequency
1CS3
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters
Cone. , mg/1
Amount , g
% of feed
70 of total input
(b) Sludge
Volume, liters
Cone. , mg/1
Amount, g
7o of feed
7o of total input
2) Feed
Volume, liters
Concentration, mg/1
Amount, g
7> of feed
7o of total input
.990
995
.985
119.4
53.5
.010
3,200
.032
3.9
1.7
.450
1,833
.825
100
44.8
.990 .990 .990
1,750 5,504 110
.1733 5.449 .109
44.4 39.5 32.1
30.0 27.7 32.1
.010 .010 .010
14,500 46,900 200
.145 .469 .002
3.7 3o4 .87
2.5 2.4 .59
.450 .450 .450
8,667 30,612 510
3.9 13.78 .229
100 100 100
67.5 69.9 67.4
262
-------
APPENDIX Al (Continued)
Parameter
TKN TOC COD o-P04-P
B. Outputs
1) Samples
Volume, liters .580 .580 .580 .580
Concentration, mg/1 731 855 2,391 71
Amount, g .424 .496 1.387 .041
% of feed 51.4 12.7 10.1 17.9
% of total input 23.0 8.6 7.0 12.1
2) Drained sludge
Volume, liters .158 .158 .158 .158
Concentration, mg/1 1,924 12,405 39,861 1,367
Amount, g .304 1.960 6.298 .216
% of feed 36.8 50.3 45.7 94.3
% of total input 16.5 33.9 32.0 63.5
C. Accumulation
1) Final supernatant
Volume, liters .708 .708 .708 .708
Concentration, mg/1 68 705 1,938 25
Amount, g .048 .499 1.372 .018
% of feed 5.8 12.8 10.0 7.9
% of total input 2.6 7.8 6.9 5.3
2) Final cleanout of walls,
etc.
Volume, liters 1 111
Concentration, mg/1 88 600 1,446 78
Amount, g .088 .600 1.446 .078
% of feed 10.7 15.4 10.5 34.1
% of total input 4.8 10.4 7.3 22.9
VI. Reference loading rate, accumulated sludge,
3 times per week loading rate,
1AS3
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters .990 .990 .990 .990
Cone., mg/1 995 1,750 5,504 110
Amount, g .985 1.733 5.449 .109
% of feed H9.4 44.4 39.5 47.6
% of total input 53.4 30.0 27.7 32.1
263
-------
APPENDIX Al (Continued)
Parameter
TKN TOC COD o-PO^-P
(b) Sludge
Volume, liters .010 -010 .010 .010
Cone., tng/1 3,200 14,500 46,900 200
Amount, g .032 .145 .469 .002
% of feed 3.9 3.7 3.4 .87
7» of total input 1.7 2.5 2.4 .59
2) Feed
Volume, liters .450 .450 .450 .450
Concentration, mg/1 1,833 8,667 30,612 510
Amount, g .825 3.9 13.78 .229
% of feed 100 100 100 100
7o of total input 44.8 67.5 69.9 67.4
B. Outputs
1) Samples
Volume, liters .800 .800 .800 .800
Concentration, mg/1 498 598 1,951 45
Amount, g .398 .478 1.561 .036
% of feed 48.2 12.3 11.3 15.7
7, of total input 21.6 8.3 7.9 10.6
C. Accumulation
1) Final Sludge
Volume, liters .100 .100 .100 .100
Concentration, mg/1 2,270 16,500 48,440 2,200
Amount,g .227 1.650 4.844 .220
% of feed 27.5 42.3 35.2 96.1
% of total input 12.3 28.6 24.6 64.7
2) Final Supernatant
Volume, liters .548 .548 .548 .548
Concentration, mg/1 159 541 1,411 135
Amount, g .087 .351 .773 .074
% of feed 10.5 9.0 5-6 32.3
7, of total input 4.7 6.1 3.9 21.7
264
-------
APPENDIX Al (Continued)
Parameter
TKN TOC COD o-PC^-P
3) Final cleanout of walls,
etc.
Volume, liters 1 111
Concentration, mg/1 105 740 1,660 83
Amount, g .105 .740 1.660 .083
% of feed 12.7 18.9 12.0 36.2
% of total input 5.7 12.8 8.4 24.4
265
-------
APPENDIX A2. IMHOFF CONES - SECOND EXPERIMENT
Parameter
TKN TOG COD
I. Reference loading rate, no sludge,
once per week loading frequency,
1NS1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters 111
Concentration, mg/1 000
Amount, g 0 0 0
% of feed 000
7o of total input 000
2) Feed
Volume, liters .725 .725 .725
Concentration, mg/1 2,812 12,604 40,189
Amount, g 2.039 9.138 29.137
7» of feed 100 100 100
7» of total input 100 100 100
B. Outputs
1) Samples
Volume, liters .580 .580 .580
Concentration, mg/1 138 305 734
Amount, g .080 .177 .425
7» of feed 3.9 1.9 1.5
7« of total input 3.9 1.9 1.5
2) Drained Sludge
Volume, liters .466 .466 .466
Concentration, mg/1 1,178 5,981 22,172
Amount, g .549 2.787 10.332
7, of feed 26.9 30.5 35.5
7» of total input 26.9 30.5 35.5
C. Accumulation
1) Final Supernatant
Volume, liters .810 .810 .810
Concentration, mg/1 81 420 894
Amount, g .066 .340 .724
7» of feed 3.2 3.7 2.5
7o of total input 3.2 3.7 2.5
.266
-------
APPENDIX A2 (Continued)
Parameter
TKN TOC COD
II. Reference Loading rate, controlled sludge level,
once per week loading frequency,
1CS1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters 111
Concentration, mg/1 000
Amount, g 0 0 0
% of feed 000
% of total input 000
(b) Sludge
Volume, liters 000
Concentration, mg/1 000
Amount, g 0 0 0
% of feed 000
% of total input 000
2) Feed
Volume, liters .725 .725 .725
Concentration, mg/1 2,812 12,604 . 40,189
Amount, g 2.039 9.138 29.137
% of feed 100 100 100
% of total input 100 100 100
B. Outputs
1) Samples
Volume, liters .560 .560 .560
Concentration, mg/1 202 363 859
Amount, g .113 .203 .481
% of feed 5.5 2.2 1.6
?„ of total input 5.5 2.2 1.6
267
-------
APPENDIX A2 (Continued)
Parameter
TKN TOG 0-P04-P
C. Accumulation
1) Sludge
Volume, liters .160 .160 .160
Concentration, mg/1 1,750 12,119 35,106
Amount, g .280 1.939 5.617
% of feed 13.7 21.2 19.3
7» of total input 13.7 21.2 19.3
2) Final supernatant
Volume, liters .845 .845 .845
Concentration, mg/1 134 705 1,215
Amount, g .113 .596 1.027
% of feed 5.5 6.2 3.5
7o of total input 5.5 6.2 3.5
3) Final cleanout of walls,
etc.
Volume, liters 111
Concentration, mg/1 137 1,350 3,294
Amount, g .137 1.350 3.294
7. of feed 6.7 14.8 11.3
7o of total input 6.7 14.8 11.3
III. Reference loading rate, accumulated sludge,
once per week loading frequency,
1AS1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters 111
Concentration, mg/1 000
Amount, g 000
% of feed 000
70 of total input 000
268
-------
APPENDIX A2 (Continued)
Parameter
TKN TOC o-P04-P
(b) Sludge
Volume, liters 0 0 0
Concentration, mg/1 000
Amount, g 0 0 0
% of feed 000
% of total input 000
2) Feed
Volume, liters .600 .600 .600
Concentration, mg/1 2,820 13,093 40,765
Amount, g 1.692 7.856 24.459
% of feed 100 100 100
% of total input 100 100 100
B. Outputs
1) Samples
Volume, liters .460 .460 .460
Concentration, mg/1 191 341 852
Amount, g .088 .157 .392
7» of feed 5.2 2.0 1.6
7o of total input 5.2 2.0 1.6
C. Accumulation
1) Final Sludge
Volume, liters .147 .147 .147
Concentration, mg/1 2,084 10,465 36,227
Amount, g .306 1.538 5.325
% of feed 18.1 19.6 21.8
70 of total input 18.1 19.6 21.8
2) Final Supernatant
Volume, liters .700 .700 .700
Concentration, mg/1 232 445 1,402
Amount, g .163 .312 .981
7o of feed 9.6 4.0 4.0
70 of total input 9.6 4.0 4.0
-------
APPENDIX A2 (Continued)
TKN
Parameter
TOG
COD
3) Final cleanout of walls,
etc.
Volume, liters
Concentration, mg/1
Amount, g
% of feed
7o of total input
IV. 4 times reference loading rate, no sludge,
once per week loading frequency,
4NS1
1
297
.297
17.6
17.6
1
965
.965
12.3
12.3
1
2,717
2.717
11.1
11.1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters
Concentration, mg/1
Amount, g
% of feed
% of total input
2) Feed
Volume, liters
Concentration, mg/1
Amount, g
7o of feed
7, of total input
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
7, of feed
% of total input
1
0
0
0
0
1.6
2,624
4.198
100
100
.300
670
.201
4.8
4.8
1
0
0
0
0
1.6
14,826
23.722
100
100
.300
760
.228
.96
.96
1
0
0
0
0
1.6
37,243
59.589
100
100
.300
1,697
.509
.85
.85
270
-------
APPENDIX A2 (Continued)
2) Drained sludge
Volume, liters
Concentration, mg/1
Amount, g
% of feed
% of total input
C. Accumulation
TKN
.701
2,244
1.573
37.5
37.5
Parameter
TOC
.701
12,879
9.028
38.1
38.1
1
880
.880
20.9
20.9
1
940
.940
4.0
4.0
1
2,039
2.039
3.4
3.4
1) Final supernatant
Volume, liters
Concentration, ing/1
Amount, g
% of feed
70 of total input
2) Final cleanout of walls,
etc.
Volume, liters
Concentration, mg/1
Amount, g
% of feed
7o of total input
V. 4 times reference loading rate, controlled sludge level,
once per week loading frequency,
4CS1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters
Concentration, mg/1
Amount, g
7» of feed
% of total input
1
0
0
0
0
1
0
0
0
0
COD
.701
48,134
33.742
56.6
56.6
1
66
.066
1.6
1.6
1
410
.410
1.7
1.7
1
1,250
1.25
2.1
2.1
1
0
0
0
0
271
-------
APPENDIX A2 (Continued)
Parameter
TKN TOG COD
(b) Sludge
Volume, liters 000
Concentration, mg/1 000
Amount, g 0 0 0
% of feed 000
7« of total input 000
2) Feed
Volume, liters 1.6 1.6 1.6
Co, centration, mg/1 2,624 14,826 37,243
Amount, g 4.198 23.722 59.589
% of feed 100 100 100
% of total input 100 100 100
B. Outputs
1) Samples
Volume, liters .320 .320 .320
Concentration, mg/1 665 747 1,847
Amount, g .213 .239 .591
7o of feed 5.07 1.0 .99
7» of total input 5.07 1.0 .99
2) Drained sludge
Volume, liters .647 .647 .647
Concentration, mg/1 2,249 12,559 49,575
Amount, g 1.455 8.126 32.075
7. of feed 34.6 34.3 53.8
7. of total input 34.6 34.3 53.8
C. Accumulation
1) Final supernatant
Volume, liters .940 .940 .940
Concentration, mg/1 1,036 1,110 3,061
Amount, g .974 1.043 2.878
7o of feed 23.2 4.4 4.8
70 of total input 23.2 4.4 4.8
-------
APPENDIX A2 (Continued)
Parameter
TKN TOC COD
2) Final cLeanout of walls,
etc.
Volume, liters 111
Concentration, mg/1 53 260 688
Amount, g .053 .260 .688
% of feed 1.3 1.09 1.15
% of total input 1.3 1,09 1.15
3) Effluent
Volume, liters .500 .500 .500
Concentration, mg/1 684 802 1,826
Amount, g .342 .401 .913
7» of feed 8.1 1.7 1.5
7, of total input 8.1 1.7 1.5
VI. 4 times reference loading rate, accumulated sludge,
once per week loading frequency,
4AS1
A. Inputs
1) Initial load
(a) Supernatant
Volume, liters 111
Concentration, mg/1 000
Amount, g 000
% of feed 000
7o of total input 000
(b) Sludge
Volume, liters 000
Concentration, mg/1 0 0 0
Amount, g 000
7o of feed 0 0 0
7o of total input 0 0 0
2) Feed
Volume, liters 1.6 1.6 1.6
Concentration, mg/1 2,624 14,826 37,243
Amount, g 4.198 23.722 59.589
% of feed 100 100 100
% of total input 100 100 100
273
-------
APPENDIX A2 (Continued)
TKN
Parameter
TOG
COD
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
7, of feed
7o of total input
C. Accumulation
1) Final sludge
Volume, liters
Concentration, mg/1
Amount, g
7> of feed
7, of total input
2) Final supernatant
Volume, liters
Concentration, mg/1
Amount, g
7» of feed
7o of total input
3) Final cleanout of walls,
etc.
Volume, liters
Concentration, mg/1
Amount, g
7o of feed
7o of total input
4) Effluent
Volume, liters
Concentration, mg/1
Amount, g
7, of feed
7o of total input
.320
847
.271
6.5
6.5
.320
834
.267
1.1
1.1
.320
2,034
.651
1.1
1.1
.240
537
1.288
30.7
30.7
.240
21,596
5.183
21.8
21.8
.240
70,513
16.923
28.4
28.4
.725
1,389
1.007
24.0
24.0
.725
950
.689
2.9
2.9
' .725
2,399
1.739
2.9
2.9
1
109
.109
2.6
2.6
1
210
.210
.88
.88
I
734
.734
1.2
1.2
1.025
599
.614
14.6
14.6
1.025
588
.603
2.5
2.5
1.025
1,408
1.444
2.4
2.4
274
-------
APPENDIX A3. 14 1 REACTORS - MASS BAIANCE
I. 1.2 times reference loading rate,
once per 2 weeks loading frequency
TKN
A. Inputs
1) Initial load
14 1 liters of water
2) Feed
Volume, liters
Concentration, rag/1
Amount, g
% of feed
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
% of feed
25.54
2,323
59.328
100
2.25
267
.600
1.0
Parameter
TOC COD
o-PO.-P
25.54 25.54 25.54
11,052 33,574 713
282.26 857.48 18.199
100 100 100
2.25
377
.848
.30
2.25
993
2.235
.26
2.25
28
.063
.35
Accumulation
1) Final supernatant
Volume, liters 11.487
Concentration, mg/1 419
Amount, g 4.813
% of feed 8.1
11.487 11.487 11.487
680 2,402 28.5
7.811 27.592 .327
2.7 3.2 1.8
2) Final sludge
Volume, liters 2.513
Concentration, mg/1 2,914
Amount, g 7.325
% of feed 12.4
3) Final cleanout of walls,
etc.
Volume, liters 1
Concentration, mg/1 778
Amount, g .778
% of feed 1.3
2.513 2.513 2.513
28,618 98,982 1,800
71.917 248.741 4.523
25.5 29.0 24.9
1
1,705
1.705
.60
1
5,894
5.894
.69
1
356
.356
1.9
275
-------
APPENDIX A3 (Continued)
TKN
II. 1.2 times reference loading rate,
once per week loading frequency
A. Inputs
1) Initial load
14 liters of water
Parameter
TOC COD
o-PO -P
4
2) Feed
Volume, liters 25.54 25.54
Concentration, mg/1 2,323 11,052
Amount, g 59.328 282.26
% of feed 100 100
25.54 25.54
33,574 713
857.48 18.199
100 100
B. Outputs
1) Samples
Volume, liters 3.65
Concentration, mg/1 261.4
Amount, g .954
70 of feed 1.6
3.65 3.65 3.65
372.3 1,075.7 27.b
1.36 3.93 0.101
.48 .46 .56
C. Accumulation
1) Final supernatant
Volume, liters
Concentration, mg/1
Amoun t, g
7o of feed
2) Final sludge
Volume, liters 2.299
Concentration, mg/1 3,826
Amount, g 8.796
% of feed 14.83
3) Final cleanout of walls,
etc.
Volume, liters 1
Concentration, mg/1 630
Amount, g .530
7» of feed ]_.l
9.48
386
3.659
6.2
9.48
875
8.30
2.9
9.48
1,867
3.66
0.43
9.48
17.2
0.163
0.89
2.299 2.299 2.299
26,465 99,508 3,150
60.843 228.769 7.242
21.60 26.6 39.79
1
1,900
1.900
.67
1
3,851
3.851
.45
1
456
.456
2.5
276
-------
APPENDIX A3 (Continued)
TKN
III. 1.2 times reference loading rate,
two times per week loading frequency
A. Inputs
1) Initial load
14 liters of water
2) Feed
Volume, liters 25.54
Concentration, mg/1 2,323
Amount, g 59.328
% of feed 100
B. Outputs
1) Samples
Volume, liters 4.35
Concentration, mg/.l 260.9
Amount, g 1.135
7, of feed 1.9
C. Accumulation
1) Final supernatant
Volume, liters 10.526
Concentration, mg/1 378
Amount, g 3.979
% of feed 6.7
2) Final sludge
Volume, liters
Amount, g
7, of feed
2.299
Concentration, mg/1 3,391
7.796
13.1
3) Final cleanout of walls,
etc.
Volume, liters 1
Concentration, mg/1 572
Amount, g .572
% of feed -96
Parameter
TOG COD
1
1,310
1.310
.46
1
3,297
3.297
.38
o-P04-P
25.54
11,052
282.26
100
4.35
320.5
1.394
.49
10.526
780
8.210
2.9
25.54
33,574
857.48
100
4.35
956.6
4.161
.49
10.526
1,781
18.747
2.2
25.54
713
18.199
100
4.35
32.4
.141
.77
10.526
22.8
.240
1.3
2.299 2.299 2.299
31,137 111,475 1,980
71.583 256.281 4.552
25.3 29.8 25.0
1
270
.270
1.5
277
-------
APPENDIX A3 (Continued)
TKN
IV. 1.2 times reference loading rate,
three times per week loading frequency
A. Inputs
1) Initial load
14 liters of water
Parameter
TOG COD
o-P04-P
2) Feed
Volume, liters 25.54 25.54
Concentration, mg/1 2,323 11,052
Amount, g 59.328 28.226
% of feed 100 100
25.54 25.54
33,574 713
857.48 18.199
100 100
B. Outputs
1) Samples
Volume, liters
Concentration, mg/1
Amount, g
7» of feed
6.4
260.9
1.67
2.8
6.4
334.7
2.142
.76
6.4
997.5
6.384
.74
6.4
41.4
.265
1.4
C. Accumulation
1) Final Supernatant
Volume, liters 10.540
Concentration, mg/1 406
Amount, g 4.274
% of feed 7.2
10.540 10.540 10.540
845 1,914 26.0
8.894 20.147 .274
3.2 2.3 1.5
2) Final sludge
Volume, liters 2.168
Concentration, mg/1 3,473
Amount, g 7.529
% of feed 12.7
3) Final cleanout of walls,
etc .
Volume, liters 1
Concentration, mg/1 557
Amount, g .557
I of feed .94
2.168 2.168 2.168
27,577 79,695 2,130
59.787 176.343 4.618
21.2 20.6 25.4
1
1,435
1.435
.51
1
3,563
3.563
.42
1
220
.220
1.2
278
-------
APPENDIX Bl. PILOT SCALE LAGOON TEMPERATURE DATA-°C
Date
1 1/16
Top of Reactor
Loading Rate - Fraction of Reference Rate
1/32 b I 1/16 1/32 k 1 1/16 1/32
Middle of Reactor Bottom of Reactor
9/15/72
9/19/72
9/22/72
9/27/72
9/29/72
10/4/72
10/6/72
11/2/72
11/16/72
11/24/72
12/1/72
1/18/73
2/1/73
2/16/73
3/14/73
7/6/73
10/18/73
11/13/73
11/29/73
1/8/74
2/12/74
3/19/74
4/9/74
5/17/74
6/25/74
7/2/74
7/16/74
8/6/74
8/22/74
10/8/74
2/7/75
25
25.5
24
25
25
21.5
23.5
23
11
7
7.5
7
8
7
16.5
31
19
8
16
18
26.5
26
28.8
25.5
28
20
9
26
26
25
24
27
23
22.5
23
11
8.5
6.5
10.5
8
4.5
18
27.2
19
8
15
17.5
24
26
25.3
23
26.5
17
7.5
29
30.5
28
27.5
27.5
25
23
22.5
12.5
10
7.5
8.5
9
5
15
34
20
15.5
15.5
8
6.5
15.5
17
25
24
29
26.3
23
29.5
16
6.5
28.5
30
27
27.5
27
24
23
23
12
10.5
8
9
9
7
21
32.5
20
16
15
8.5
8
15
17.5
28.5
24.5
30
27
24
30.5
17
7
25
25
23
24
23.5
20
21
21
11
7.5
7
4
6
6
15
28
17
7.5
15.5
16
25.5
26
28.3
25
28
19.5
8
22
22.5
22
22
23.5
19.5
20
21
10.5
9
7
5
5.5
4.5
12
26
16
7
15
15.5
23.5
23.2
24.3
23.5
25.5
16
6.5
25.5
26
23
23.5
23.5
20
20
20
12.5
10
7.5
5
5.5
5
10
21.5
16.5
15
14.5
7.5
15
15
15.5
22.5
21.8
24
23
25
15.5
7
26
26
23
24.5
25.5
21.5
21.5
21
11
10.5
8
5
5.5
5
12
24
17.5
14.5
14
7
14.5
14.5
17
24
23.3
24
23.5
25
16.5
7
25
25
23
24
23.5
20
20
20
11
8.5
7
4
6
6
11
27
17
7
15
15
25
26
28
25
27.5
19
8
22
21.5
21
21.5
22
18.5
19.5
19.5
10.5
9
7
5
5
5
10
26
16
7
14.5
14.5
23.5
24
24.3
23.5
25.5
16
8
24
24
21.5
22
22.5
20
20
19
12.5
10
7
4.5
5
6
9
20
16.5
14.5
14
6.5
14.5
14
14.5
21.5
21
22.8
22
24
15.5
7
24.5
23
22.5
23.5
22.5
20.5
20.5
19
11
10.5
8
5
5
5
9
22
17
13.5
13.5
6.5
7
14
14
16
23
22
22.5
23.5
24.5
16.5
7
-------
Appendix B2. ODOR RATING SHEET
Name Date
If you encountered this odor which one of the following would must
closely describe your attitude?
1. Strongly object
2. Object
3. Would not object
4. Would not offend you at all
Lagoon designation Description (1-4)
280
-------
Appendix B3
TYPICAL LAGOON PROFILE DATA OBTAINED WITH
CURVED TUBE TECHNIQUE
Lagoon
Position
above
bottom, m
1.49
1.18
0.873
0.565
0.257
0.206
0.155
0.104
0.053
0.
1.04
0.924
0.616
0.536
0.459
0.308
0.153
0.076
0.
Concentration, m£
COD
1,860
1,800
2,000
1,700
2,200
2,800
1,560
1,500
10,300
58,300
1,550
1,740
3,600
5,400
35,700
41,500
45,300
43,500
42,400
TOC
70
180
170
250
305
285
320
305
1,100
4,000
700
750
1,600
2,560
14,800
16,800
19,000
18,200
16,200
o-P04-P
65
56
56
53
59
65
52
52
240
2,470
93
109
112
290
2,200
701
2,540
2,400
2,750
5/1
TKN
370
360
370
360
380
780
190
350
380
3,460
850
760
750
790
990
1,060
1,230
1,620
950
NH -N
--
--
--
--
--
--
--
--
--
240
840
1,010
4,330
3,300
3,570
3,600
3,500
3,900
281
-------
APPENDIX B4. FIELD PILOT-SCALE REACTOR TEMPERATURE, DISSOLVED OXYGEN
AT SURFACE AND BULK SUPERNATANT TOC
Date Loading rate Temperature D.O COD TOC
(fraction of °C mg/1 mg/1 mg/1
reference rate)
10/18/73 1/16 20 15+ 240 100
3rd series lagoon 19.5 15+ 340 140
1/32 20 5.3 160 60
11/13/73 1/16 15.5 11.6 420 120
3rd series lagoon 15 9.5 310 120
1/32 16 15+ 140 75
11/29/73 1/16 15.5 12.6 380 160
3rd series lagoon 15 1.6 290 130
1/32 15 6.4 200 90
1/8/74 1/16 8 7.6 350 160
3rd series lagoon 9 3.4 490 160
1/32 8.5 9.5 205 95
2/12/74 1/16 6.5 0 470 180
3rd series lagoon 6 0 500 200
1/32 8 1.3 235 75
3/19/74 1/16 15.5 13.6 500 180
3rd series lagoon 16 8.6 580 240
1/32 15 15 295 220
4/9/74 1/16 17 8.6 550 220
3rd series lagoon 18 3.6 600 290
1/32 17.5 11.9 390 160
282
-------
Appendix B5. HYDROGEN ION CONCENTRATION DATA
Sample source
Pilot-scale reactors
Multiple or fraction
of reference loading
4
4
4
4
4
4
4
4
4
4
1
1
1
1
1
1
1
1
1
1
1/2
1/2
1/2
1/2
1/2
1/2
1/2
1/2
1/2
1/2
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
Date
pH
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
7.30
7.65
7.60
7.43
7.49
7.60
7.75
7.65
7.80
7.40
7.49
7.88
7.67
7.50
7.63
7.56
8.02
7.70
7.70
7.45
7.57
7.71
7.98
7.70
7.70
7.59
7.53
7.90
7.75
7.40
7.45
7.71
7.55
7.50
7.48
7.27
7.27
7.75
7.65
7.60
283
-------
Sample source
Pilot-scale reactors
APPENDIX B5. (continued)
Multiple or fraction
of reference loading
1/8
1/8
1/8
1/8
1/8
1/8
1/8
1/8
1/8
1/8
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/16
1/32
1/32
1/32
1/32
1/32
1/32
1/32
1/32
1/32
1/32
1A
1A
1A
IA
1A
Date
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/X7/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
7/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
7.37
7.24
7.54
7.18
7.45
7.23
7.20
7.80
7.60
7.90
7.34
7.12
7.31
7.18
7.19
7.00
6.91
7.85
7.70
7.40
7.34
7.12
7.31
7.18
7.19
7.00
6.91
7.85
7.70
7.80
7.38
7.10
7.00
6.71
7.06
6.46
6.80
7.55
7.30
8.40
7.60
7.90
7.65
7.76
7.62
284
-------
Sample source
Pilot-scale reactors
Lab reactors
APPENDIX B5. (continued)
Multiple or fraction
of reference loading
1A
1A
1A
1A
1A
IB
IB
IB
IB
IB
IB
IB
IB
IB
IB
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
Date
Raw waste
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
3/27/73
4/3/73
4/10/73
4/17/73
5/2/73
5/15/73
10/16/73
10/19/73
3/21/74
3/23/73
4/5/73
4/11/73
4/18/73
5/2/73
5/16/73
3/23/73
4/5/73
4/10/73
4/17/73
5/1/73
5/16/73
3/23/73
4/5/73
4/11/73
4/18/73
5/2/73
5/16/73
3/23/73
4/5/73
4/11/73
4/18/73
5/2/73
5/16/73
3/20/73
3/22/73
2/22/73
4/3/73
7.70
7.52
8.00
7.95
8.10
7.70
7.58
7.70
7.39
7.60
7.50
7.40
8.40
8.30
8.40
7.63
7.59
8.31
8.32
8.26
8.18.
7.83
8.00
8.48
8.36
8.23
8.18
8.00
7.96
8.37
8.21
8.06
8.31
7.56
7.45
8.13
8.01
7.91
8.12
7.43
7.78
7.20
7.62
285
-------
APPENDIX B5. (continued)
Multiple or fraction
Sample source of reference loading Date pH
Raw waste 4/17/73 7.68
5/2/73 7.63
5/15/73 7.46
286
-------
APPENDIX Cl. SOIL CONCENTRATION PROFILES
txi
OO
—1
September, 1973
Depth,
era
0- 5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
1-H
484
348
288
290
265
312
272
294
234
248
2-L
441
625
374
364
210
168
120
87
98
107
3-M
288
377
231
205
205
276
154
123
158
66
Plot
4-H 5-M
326 188
237 273
182 245
245 364
182 186
276 101
133 105
213 102
140 78
186 136
6-L
231
156
175
224
216
151
95
41
102
136
Total Kjeldahl Nitrogen (TKN) , ppm
7-H
439
467
367
139
151
147
209
151
142
-
8-L
534
391
290
205
228
161
119
146
104
103
9-M
388
478
296
147
147
153
216
150
64
-
1-H
280
280
280
340
480
220
170
140
170
360
2-L
390
360
360
310
360
200
200
110
110
200
3-M
450
280
220
200
310
220
200
140
200
200
4-H
390
310
280
250
250
140
140
140
110
220
March, 1974
Plot
5-M 6-L
390 220
220 250
220 250
200 220
200 200
200 170
110 140
110 140
110 110
220 140
7-H
530
450
340
200
220
250
280
310
220
360
8-L
480
500.
390
250
280
250
220
220
220
220
9-M
420
390
280
220
220
200
140
310
280
250
Nitrate-Nitrogen
(N03-N)
, PP".
September, 1973
Depth,
Plot
December, 1973
Plot
0- 5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
5.
1.
2,
5.
8.
10
8
13
11
14
1
.9
.8
I
.0
.2
.5
.1
.0
.0
.8
1.0
.3
.2
.6
.8
1.9
.9
3.5
1.3
.8
.7
.6
,3
.5
.8
.7
.8
.7
.5
4.0
3.0
2.1
2.1
2.4
4.6
3.9
4.4
8.0
7.8
.5 .9
.9 .7
.5 .4
.5 .4
.4 .7
.8 .5
1.3 1.2
.8 .6
.7 ,7
.6 .9
2.8
3.2
4.4
6.9
5.8
7.4
4.6
6.5
3.0
3.5
1.8
1.5
1.2
1.4
1.0
1.2
.6
.8
1.0
.9
.9
.5
1.2
1.5
.9
1.5
2.9
2.4
1.9
1.6
1.7
1.3
1.2
1.0
1.2
1.3
1.0
1.7
2.5
4.8
.8
.6
.6
.6
.6
.8
.8
1.0
.8
2.1
1.3
1.0
.6
.9
.9
1.0
1.0
1.2
1.7
2.5
2,
1,
1,
1,
1,
1,
4.
,2
.9
.8
,9
.0
.2
.0
.2
.4
.2
1.0
.8
.6
.9
.8
.6
.6
.8
.8
2.3
1.0
.5
.5
.5
.5
.5
.4
.5
.4
.5
1.2
.6
.6
1.5
1.0
1.6
1.7
1.7
3.1
2.3
.1
.1
1.2
.1
0
0
0
0
0
1.3
0
0
0
0
0
.6
.5
.3
.4
.3
Depth,
cm
0- 5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
September, 1973
Plot
254
206
98
75
44
29
27
30
35
43
209
205
107
121
85
40
22
21
20
26
251
206
115
80
61
41
39
32
46
63
247
155
85
82
62
28
34
17
45
69
179
123
91
69
79
39
34
19
20
29
182
87
60
54
50
41
28
24
16
40
194
150
57
21
16
22
57
87
61
62
161
101
42
30
45
37
26
25
29
44
Calcium
189
127
51
26
28
28
52
82
88
68
(Ca) , ppm
199
138
93
59
38
26
33
20
20
25
December, 1973
Plot
165
173
121
73
60
67
41
18
22
30
169
127
88
83
97
65
39
22
34
58
184
130
69
53
40
31
46
27
35
52
186
101
82
54
47
67
31
20
20
31
238
201
90
43
46
31
35
17
29
23
155
102
43
22
26
50
80
71
67
59
117
131
28
27
46
51
63
58
61
27
174
91
28
27
22
40
61
65
84
81
-------
APPENDIX Cl. (continued)
ho
00
OO
Sodium (Na) , ppm
September, 1973
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
1-H 2-L
8 6
9 8
13 7
10 8
11 7
9 6
8 6
9 7
10 5
12 6
3-M
10
13
15
17
15
12
9
6
9
11
4-H
9
10
9
9
10
10
9
8
12
15
5-M
7
11
13
13
10
9
8
8
7
7
6-L
5
6
10
8
4
7
4
4
4
7
7-H
10
12
11
9
7
12
16
15
9
14
8-L
9
19
10
9
10
7
6
4
4
6
9-M
10
14
10
11
10
11
14
21
19
12
1-H
16
9
9
12
9
7
6
8
8
12
2-L
9
10
13
10
11
10
7
7
6
16
3-M
2
3
3
7
11
11
9
7
15
22
4-H
6
6
10
13
9
11
12
12
19
31
December, 1973
5-M
4
4
9
10
10
12
9
9
10
15
6-L
9
9
11
11
13
9
10
8
17
11
7-H
14
13
15
15
19
31
48
34
28
34
8-L
17
23
11
10
12
15
22
18
26
14
9-M
8
11
12
17
19
19
25
31
30
22
Copper (Cu), ppm
September, 1973 December, 1973
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
1.8
1.0
.9
1.4
2.1
1.1
1.4
1.8
1.4
1.2
1.6
1.0
1.2
1.3
1.0
.9
.8
1.4
1
1
2,
1
1
2,
3,
.2
.9
.6
.1
.1
.8
.5
.8
.9
1.1
1.6
1.8
1.4
1.4
1.4
1.5
2.9
1.9
2.4
2.9
, .6
1.1
.9
1.1
1.5
1.6
10.3
1.10
.9
.8
.6
.9
.9
.6
.5
1.5
1.1
1.4
1.1
1.5
.9
.9
.9
1.3
1.6
1.2
.9
1.4
1.1
2.5
1.0
1.3
1.4
2.0
1.4
.9
1.4
1.1
2.1
.9
.9
1.8
1.4
3.2
,
3.0
3.0
4.0
2.2
2.0
1.5
1.2
1.8
19
. i.
1.0
.9
.8
.9
.8
.9
.8
1.2
1.5
1 .
1.
1.
1.
.7
.7
.6
.9
.8
\ f\
i . U
.8
1.1
.8
.7
.7
.7
.8
.8
1 1
1.1
.9
.7
1.1
.7
.9
.7
.7
.6
7.6 1.0
37 in
, i 1 . U
.9 2.2
.9 1.3
.7 1.1
.4 .8
.6 .9
.7 .7
.7 .8
.4 1.2
.8
1 7
1 . £.
1.1
1.0
1.0
.7
.6
.6
.8
.8
1,3
1 1
1.1
1.1
1.3
1.5
.7
.7
.6
.7
.7
Zinc (Zn), ppm
September, 1973 December, 1973
Depth,
cm
0-5'
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
5,
1,
I (
1,
7,
1,
19.
1.
,2
,3
,8
9
,3
,3
,8
,1
.1
.3
1.6
1.6
1.2
1.4
1.3
1.4
1.1
2.4
1.5
1.3
1
2.
2
1
6
1
1,
1,
2,
1,
.9
.3
.6
.1
.0
.0
.5
.9
.3
.7
3.
2.
1
1.
1
1,
1,
2.
.0
.0
.9
.1
.6
.1
,3
.2
.0
.9
4
2.
2.
2,
2,
2.
1,
1,
5.
,9
.2
.3
.0
.0
.7
,9
,0
.1
.1
1.6 2,
.9 1.
.7 1,
3.6 1,
.6 1.
.9 2.
.6 3,
2.1 2,
2.6 2.
3.0 1.
.6
.5
.5
.0
.1
.2
,5
,1
.0
,4
1.9 1.8
1.6 1.9
3.6 2.0
1.1 2.0
1.3 2.6
2.6 1.6
1.7 1.2
2.8 7.7
1.3 1.1
1.5 3.6
7,
7,
7
10.
6
2
2
4
4
2
.3
.3
.7
.6
.9
.8
.1
.6
.0
5.
3.
2.
3.
3.
2.
1.
1.
2.
7
6
3
5
9
3
7
3
2
1
4,
2.
1.
2,
2.
1
2
2
1
1
.4
.9
.3
.9
.5
.4
.8
.8
.6
2
2
2.
2
1
1
1
1
2
.8
.9
.0
.6
.6
.6
.5
.9
.5
1
2
1
2
1
1
1
1
1
.2
.9
.0
.9
.1
.7
.2
.7
.3
.6
5,
3
2
2.
2
1
1
1
1
1
.9
.9
.3
.0
.9
.2
.9
.1
.1
.6
3.
2,
2.
3.
2,
1,
1,
1.
1
1.
.2
,8
.8
,1
.2
,3
,7
.1
.6
,5
2.6
2.9
2.2
2.0
2.8
2.0
1.5
1.1
1.8
1.9
4,
2,
1
2.
2
1
3
4
2
1
.0
.3
.9
.5
.0
.9
.4
.2
.6
.4
-------
APPENDIX Cl. (continued)
NJ
00
September, 1973
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
1-H
51
37
32
26
36
18
18
13
19
15
2-L
35
34
33
37
28
24
15
8
15
19
3-M
36
34
28
38
30
21
23
20
22
38
4-H
38
31
25
29
29
21
14
12
18
21
5-M
39
31
36
30
28
25
16
14
19
23
6-L
29
26
26
35
29
19
19
13
9
20
7-H
46
36
29
25
24
19
37
33
33
18
8-L
33
28
27
32
30
27
25
29
19
24
Iron (Fe),
9-M
36
36
29
19
18
25
7
26
27
22
1-H
66
69
67
56
63
37
36
30
20
23
ppm
2-L
43
57
56
50
52
37
38
28
24
27
December, 1973
3-M
38
48
53
48
36
39
31
20
14
30
4-H
55
46
47
47
54
36
41
30
31
35
5-M
52
49
41
45
39
39
30
28
30
29
6-L
59
55
49
46
51
35
31
24
34
30
7-H
80
60
63
52
62
38
38
41
38
56
8-L
66
58
48
47
46
37
32
32
46
29
9-M
73
51
59
60
58
41
40
39
30
40
Depth,
September, 1973
Manganese (Mn), ppm
December, 1973
0-5
5-10
10-15
15-20
20-25
25-35
35-45
55-65
65-75
9.6
8.8
11.8
6.1
5.5
2.3
1.8
2.3
1.2
11,
11,
11.
11,
8,
4
2,
1
1
.6
.4
,1
.2
.3
.7
.2
.6
.0
11.0
9.3
8.1
7.2
5.5
4.0
1.7
1.1
1.5
B.5
6.6
4.7
6.4
5.2
2.3
1.5
1.7
1.2
7.1
5.6
6.4
4.8
5.7
2.9
1.3
.8
.9
5.4 10.
3.1 9.
4.3 6.
4.6 8.
3.2 2.
2.5 1.
1.9 2.
1.2 1.
1.7 1,
,9
,4
.9
.5
.9
,6
,1
,6
.3
12.3
12.5
9.9
10.9
14.5
17.8
3.0
.8
1.4
11.9
10.9
9.1
4.9
1.9
1.5
1.8
1.2
1.3
9.5
9.6
9.4
9.1
6.8
3.4
2.8
1.2
1.8
10.6
22.8
11.8
9.1
9.7
5.1
3.2
1.7
1.1
6.6
8.0
9.2
10.2
6.8
5.1
3.4
1.2
1.5
7.2
5.9
6.7
5.7
5.2
2.4
2.5
1.5
2.0
7.0
5.1
5.9
5.1
2.5
3.7
2.0
1.3
1.3
5.9
6.2
4.5
3.0
3.7
1.8
1.7
1.1
1.2
9.8
9.2
7.7
5.0
2.7
1.7
1.9
1.6
1.9
11.5
9.2
7.9
5.5
4.3
2.3
1.9
1.6
1.1
10.8
13.5
7.0
8.4
4.4
3.9
2.9
1.8
2.3
September, 1973
December, 1973
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
2.5 3.6
2.2 3.4
2.2 1.6
1.6 2.8
2.4 2.3
3.2 3.4
1.5 1.9
1.8 1.9
3.6 2.3
2.2 2.0
2.4
2.2
2.0
2.0
2.9
2.2
2.2
1.2
2.8
4.7
6.9
2.6
1.7
2.5
2.6
2.4
2.0
1.9
2.0
0.0
2.9
3.2
1.8
7.8
1.5
2.9
2.1
2.0
2,7
2.2
3.0
1.7
1.5
4.8
2.5
2.4
3.6
8.8
4.4
1.9
3.0
2.5
3.7
0.0
3.6
3.2
3.1
3.4
4.1
3.6
3.4
3.1
2.5
2.7
2.2
2.2
2.4
2.8
2.2
3.0
4.7
4.5
2.5
2.3
1.4
1.9
5.7
3.4
3.4
3.1
7.3
3.3
2.4
2.7
2.6
2.1
2.5
2.3
1.9
1.8
1.8
1.5
1.6
1.5
1.6
1.2
1.2
1.1
1.2
1.1
1.4
1.3
1.5
1.4
1.4
1.5
1.6
1.3
1.9
2.2
2.7
2.5
2.2
1.9
1.6
1.5
1.8
1.4
1.2
1.0
1.6
1.5
1.5
1.4
1.9
1.4
1.3
1.2
1.8
1.7
4.9
3.3
1.7
1.4
1.2
1.2
1.3
1.1
1.8
1.4
2.1
2.1
1.8
.6
.6
.6
.5
.9
.5
.5
2.1
2.3
2.4
2.2
2.0
1.7
1.6
1.5
1.9
1.3
2.3
2.0
2.2
2.3
2.2
3.2
1.8
1.6
1.8
1.3
-------
Magnesium (Mg), ppm
N3
^£>
O
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
September, 1973
Plot
1-H
73
60
23
14
9
7
10
8
6
Q
2-L
57
64
20
14
7
3
3
2
2
3-H
68
57
30
14
7
3
5
2
3
4-H 5-M 6-L
77
34
22
18
17
7
8
6
6
51
21
18
11
7
3
5
2
2
54
21
12
6
4
3
4
2
1
7-H
56
38
21
11
12
10
15
20
17
1 7
1 /
8-L
54
32
12
7
9
9
8
5
6
9-M
54
36
22
11
10
10
12
19
19
i ft
10
1-H
32
26
21
13
9
6
8
6
5
2-L
29
27
21
9
8
4
5
2
4
5
3-H
30
25
19
18
20
12
8
5
6
4-H
31
28
19
16
11
7
10
7
8
11
Decc
5-M
32
23
17
12
9
8
7
5
4
5
:mber ,
Plot
6-L
37
35
23
12
8
7
5
3
5
5
1973
7-H
30
26
16
12
10
10
13
15
14
14
8-L 9-M
29 32
25 24
11 13
9 15
14 9
11 11
14 14
15 13
15 15
5 15
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
AC ee
*t j~jj
55-65
65-75
117
52
20
15
16
7
9
6
3
49
18
13
19
15
8
6
5
5
72
30
20
18
11
8
16
5
5
September
100
30
14
20
11
5
5
6
2
, 1973
65
16
20
17
18
8
7
3
3
Phosphorus
(D,
ppm
December,
39
12
11
12
11
6
5
2
2
47
15
15
5
4
3
5
4
5
12
5
6
4
8
8
3
7
1
5
4
26
8
5
3
3
5
7
A
**
3
3
158
102
62
42
26
14
13
9
7
50
32
22
18
30
25
10
11
4
58
41
26
22
14
11
11
6
3
94
41
27
24
22
10
13
9
9
82
y
27
15
7
13
7
A
2
3
55
17
15
8
12
3
6
7
4
1973
48
16
6
5
3
3
2
4
4
12 85
7 48
2 8
2 4
3 7
1 1
2 3
If.
*t
2 1
4 2
Depth,
en
0-5.
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
101
99
78
47
26
23
22
23
19
17
103
46
25
17
9
10
8
7
8
18
129
85
35
20
14
9
9
10
27
23
September
113
96
60
46
36
20
17
19
14
15
, 1973
107
64
36
18
9
6
8
10
7
18
103
31
11
8
5
6
7
7
7
19
121
81
49
32
28
36
28
12
16
13
105
16
8
10
8
8
18
9
9
11
Potassium
107
55
21
11
13
10
16
14
12
11
(K),
132
132
101
85
56
40
39
34
29
34
ppm
116
74
28
20
21
11
14
9
14
30
90
88
63
43
74
29
47
39
48
53
155
135
89
87
48
48
61
61
73
102
Dece
130
113
51
61
30
25
18
20
60
45
mber.
77
56
36
23
24
9
17
10
32
26
1973
97
78
43
31
23
19
23
13
20
31
68 76
48 61
15 29
11 48
21 22
18 24
14 29
13 14
19 18
28 18
-------
APPENDIX Cl. (continued)
Depth,
cm
0-5
5-10
10-15
15-20
20-25
25-35
35-45
45-55
55-65
65-75
1-H
5.
5.
5.
4
4,
4.
It.
4
3
/
8
.3
.2
.2
.4
.0
.1
.8
2-L
5.9
5.4
5.2
4.6
4.3
4.1
4.1
4.1
3-M
6.0
5.6
4.9
4.6
4.2
4.1
4.2
September, 1973
4-H 5-M 6-L 7-H
6.3
5.6
5.1
4.4
4.0
4.1
6.1
5.6
4.8
4.6
4.2
4.1
5.5
4.9
4.6
4.7
4.2
4.2
5.5
4.8
4.4
4.4
4.4
4.4
8-L
5.
4.
4.
4.
4,
4
.a
i
6
.5
,4
.3
,4
9-M
5.9
5.4
5.2
4.6
4.5
4.7
4.6
pH n
1-H
5,
1
2
2
0
4
4,
.9
.8
,5
,2
.7
.5
2-L
6.4
6.3
5.9
.5
.3
.3
.3
.0
.2
4.7
3-M 4-H
6.1 6.4
6.1 6.4
5.8 6.0
.8 5.5
.7 .3
.4 .2
.5 .3
.1 .0
.0 4.9
5.0 4.8
December,
5-M 6-L
6.4
6.3
6.0
5.7
.8
.6
.5
.2
4.9
4.9
6.7
6.5
6.2
5.5
.4
.2
.3
.1
4.6
4.7
1973
7-H
5.9
.8
.4
.2
.1
.2
.1
.4
.3
5.1
8-L
5.8
.7
.3
.1
.2
.2
.1
.3
.3
4.7
9-M
6,
5.
0
1
2
2
3
1
5.
.1
.6
,2
1
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-76-233
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
DESIGN CRITERIA FOR SWINE WASTE TREATMENT SYSTEMS
5. REPORT DATE
October 1976 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Frank J. Humenik and Michael R. Overcash
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
Biological and Agricultural Engineering Department
North Carolina State University
Raleigh, North Carolina 27607
10. PROGRAM ELEMENT NO.
1HB617
11. CONTRACT/GRANT NO.
R-802203
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 Report (6/71-5/74)
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
•IACT
Coordinated laboratory, field pilot-, and farm-scale lagoon studies were con-
ducted to define relationships between loading intensity and frequency based on
treatment performance, sludge accumulation, and odor potential. Surface aeration
of field pilot units and farm-scale lagoons was also investigated to evaluate
aeration levels required for odor control and the effect of surface aeration on
nitrogen and organic transformations.
Laboratory studies were designed to elucidate basic chemical, physical, and
biological mechanisms important in explaining and modeling lagoon performance. Long-
term mass balance studies were conducted to define the fate of waste input and thus
total constituent loss from the system.
Predictive and interpretive relationships for lagoons based on constant batch
loading and continuous loading were derived to describe the supernatant concentration
of unaerated lagoons. Methods for determining steady-state concentrations and first-
order reaction rate constants for oxygen demand, organic carhon, and nitrogen were
developed and compared with laboratory and field pilot-scale data.
Lagoon liquid from a farm-scale unit was irrigated to nine 9.24 m x 9.24 m
Coastal Plain soil-Bermuda grass plots at nitrogen loading rates of 300, 600, and
1,200 kg N/ha./year. Mass balance data were collected to determine the fate of
applied waste constituents.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN-ENDED TERMS
c. cos AT I Field/Group
Swine; Agricultural Wastes
Treatment Processes;
Aerated Lagoons
Q2/A, C, E
8. DISTRIBUTION STATEMENT
RELEASE UNLIMITED
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
314
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
292
*USGPO: 1977 - 757-056/5460 Region 5-1
------- |