v>EPA
United States
Environmental Protection
Agency
Health Effects Research
Laboratory
Cincinnati OH 45268
EPA-600/1-80-01 5
February 1980
Research and Development
The Evaluation of
Microbiological
Aerosols
Associated with the
Application of
Wastewater to Land
Pleasanton, California
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9 Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL HEALTH EFFECTS RE-
SEARCH series This series describes projects and studies relating to the toler-
ances of man for unhealthful substances or conditions This work is generally
assessed from a medical viewpoint, including physiological or psychological
studies In addition to toxicology and other medical specialities, study areas in-
clude biomedical instrumentation and health research techniques utilizing ani-
mals — but always with intended application to human health measures
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
THE EVALUATION OF MICROBIOLOGICAL AEROSOLS
ASSOCIATED WITH THE APPLICATION OF
WASTEWATER TO LAND: PLEASANTON, CALIFORNIA
by
D. E. Johnson, D. E. Camann, J. W. Register,
R. E. Thomas, C. A. Sorber, M. N. Guentzel,
J0 M. Taylor, H. J. Harding
Southwest Research Institute
San Antonio, Texas 78284
Contract No. DAMD 17-75-C-5072
Interagency Agreement No0 IAG-D7-0701
Project Officers
Stephen A0 Schaub
U.S. Army Medical Bioengineering Research and Development Laboratory
Fort Detrick, Frederick, Maryland 21701
Herbert R. Pahren
Health Effects Research Laboratory
Cincinnati, Ohio 45268
This study was conducted
in cooperation with
U.S. Army Medical Research and Development Command
Fort Detrick, Frederick, Maryland 21701
HEALTH EFFECTS RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
U.S. t:-.\ • "
r- •'-,.- '•'
-------
DISCLAIMER
This report has been reviewed by the Health Effects Research Laboratory, U.S. Environmental Protec-
tion Agency, and the U.S. Army Medical Research and Development Command, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the views and policies of the U.S. Envi-
ronmental Protection Agency nor the U.S. Army, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
UC r~"rr ;''••" -"-.--•-
4O. L-i 1 a .;'..!.;> - .
ii
-------
FOREWORD
Research and development is that necessary first step in problem solution and it involves defining the
problem, measuring its impact, and searching for solutions. To that end, the Environmental Protection
Research Division of the U.S. Army Medical Bioengineering Research and Development Laboratory con-
ducts comprehensive basic and applied research in support of the Surgeon General's responsibilities in envi-
ronmental protection to include air, land, and water pollution control and disposal of hazardous wastes and
pesticides, and in occupational health associated with exposure to chemicals. The primary mission of the
Health Effects Research Laboratory is to provide a sound health effects data base in support of the regulatory
activities of the U.S. Environmental Protection Agency. HERL conducts a research program to identify,
characterize, and quantitate harmful effects of pollutants that may result from exposure to chemical, physi-
cal, or biological agents found in the environment. In addition to the valuable health information generated
by these activities, new research techniques and methods are being developed that contribute to a better
understanding of human biochemical and physiological functions, and how these functions are altered by
low-level insults.
This report describes a joint research effort by the two laboratories. An in-depth microbiological evalu-
ation was made at a site where treated municipal wastewater was spray irrigated. Special emphasis was given
to microbial transport by aerosols.
111
-------
ABSTRACT
The purpose of this study was to determine the extent that individuals near spray irrigation sites are
exposed to microorganisms in wastewater aerosols. This report reviews a monitoring effort of a spray irriga-
tion site utilizing unchlorinated secondarily-treated wastewater from biofiltration treatment processes. Objec-
tives included an in-depth pathogen screen of wastewater, establishing the relationship between pathogen
levels and traditional indicator organisms, monitoring microorganisms in air within 600 meters of the spray
source, and development/validation of a microbiological dispersion model for predicting aerosol pathogen
concentrations. Effluent was monitored for microbiological, chemical, and physical characteristics and exten-
sive microorganism and dye aerosol samples were collected (77 aerosol runs). Enteroviruses were detected in
air, but at a very low density. Conclusions: There is considerable underestimation of aerosol pathogen levels
when using traditional indicators to predict human exposures. A microbiological, dispersion model may be
used with minimal monitoring to estimate exposure. There is little correlation between wastewater levels of
traditional indicators and pathogens. Aerosols containing microorganisms are generated by spray irrigation
of wastewater; they do survive aerosolization and can be transported to nearby populations. Until dose-
response relationships are developed, neither the levels of aerosolized microorganisms that constitute a
hazard nor the degree of required wastewater disinfection can be specified.
This report was submitted in fulfillment of Contract DAMD 17-75-C-5072 by Southwest Research Cor-
poration under the sponsorship of the U.S. Environmental Protection Agency and the U.S. Army. This
report covers a period from June 1975 to March 1978.
IV
-------
EXECUTIVE SUMMARY
Application to land is perhaps the oldest method of disposing of wastewater, and an attractive modern
alternative. The advantages include re-use of the water, avoiding contamination of bodies of surface water,
and return of nutrients in the wastewater to the soil. The most frequent means of application is spray irriga-
tion, which can be applied to a wide variety of land sites where the topography of the land would make other
methods of irrigation infeasible.
Aerosols containing pathogenic microorganisms are created during wastewater spraying and the orga-
nisms can be transported to populated areas by prevailing winds. The principal purposes of this program were
to determine to what extent individuals living near sites practicing spray irrigation are exposed to these micro-
organisms and to gain insight into the potential health effects. Gastro-intestinal and respiratory illnesses such
as dysentery, typhoid fever, and infectious hepatitis might be spread by spraying poorly-treated, undisin-
fected wastewater.
The study was conducted at a spray irrigation site associated with a sewage treatment plant located in
Pleasanton, California. Here, treatment plant effluent was utilized to irrigate grazing lands. The wastewater
is secondarily treated but not chlorinated by a process called contact biofiltration, and approximately 1.4 mil-
lion gallons per day are sprayed onto the fields. The program was designed with three potential phases, Phase
I was to be a site characterization, Phase II to be extensive aerosol monitoring effort, and Phase III an epi-
demiology study of the exposed population. This report covers the Phase II monitoring effort, which was
conducted over the period from May 1976 to April 1977.
There were two distinct efforts performed in Phase II. The sewage treatment plant receives a large input
from the Alameda County Fair during a month-long period each summer and the sewage effluent at this time
is not typical of the remainder of the year. The decision was made not to monitor during this period, but to
monitor both before and after the fair. The two efforts were conducted with differing sets of objectives, and
these were designated as Pre-Fair and Post-Fair.
The principal objectives of the Pre-Fair study were to perform an in-depth pathogen screen of the waste-
water, to establish the relationship in wastewater between pathogen levels and levels of the traditional indica-
tor organisms (total and fecal coliform and standard bacterial plate count), to determine microorganism lev-
els in air within 100 meters of the spray source, and to begin the assessment of factors thought to affect the
levels of pathogenic organisms collected in aerosol samples, including aerosolization efficiency, pathogen sur-
vival upon becoming airborne (impact), and microbiological die-off with time (viability decay). These factors
were to be used to begin development of a predictive model of pathogen concentration to estimate the degree
of exposure of the nearby populations.
The objectives of the Post-Fair study were primarily oriented toward the development and validation of
a predictive model. To accomplish this goal, air sampling was to be conducted to 600 meters downwind, the
factors affecting microbiological aerosols were to be identified and quantified over a wide range of meteoro-
logical conditions, and sufficient aerosol runs were to be completed to permit model development. The model
developed from the Pre-Fair and Post-Fair data was then to be validated using data from studies at Fort Hua-
chuca and Deer Creek and some Pleasanton data not usable in model development.
To accomplish these objectives, it was necessary to monitor the effluent for its chemical and physical
characteristics as well as microbiological constituents and to obtain large volumes of the wastewater for pa-
thogen screening. Extensive aerosol samples were to be collected downwind and upwind of the spray line to
determine the concentration of both traditional microorganism groups and of pathogens, and to compare
v
-------
these concentrations with the expected levels based on the spray rate and the microorganism levels in the ef-
fluent. Additionally, samples were taken after the injection of dye into the wastewater to allow estimation of
the proportion of the sprayed effluent that became aerosolized.
Routine monitoring of the wastewater for chemical, physical, and microbiological parameters was ac-
complished during Pre-Fair by taking a composite sample from the aeration basin during the hours of spray-
ing. Chemical and physical analyses included total and free chlorine, pH, total organic carbon (TOC), total
solids, and total suspended solids (TSS). In addition, one-half of the composite samples were tested for bio-
chemical oxygen demand (BOD), chemical oxygen demand (COD), total phosphorus, hardness, and the ni-
trogen series, (nitrite, nitrate, ammonia, and organic nitrogen). Microbiological analyses run on all wastewa-
ter samples included total and fecal coliform, standard bacterial plate count, coliphage, and assays for
selected pathogens, (Klebsiella, Pseudomonas, fecal streptococci, Clostridium perfringens, and 3- and 5-day
enteroviruses).
To perform the wastewater pathogen screen, eight large-volume (20L) samples of effluent were taken at
intervals throughout the Pre-Fair period and at the beginning of the Post-Fair period. These were sent to the
UTSA-CART laboratory in San Antonio for a semi-quantitative screen to determine those microorganisms
appearing with frequency in the effluent and to assist in selection of organisms for routine assay of wastewa-
ter and aerosol samples.
In the Post-Fair study routine chemical/physical analyses of the wastewater included pH, TOC, TSS
and conductivity. Three samples were collected over the Post-Fair period for more detailed analyses, which
included BOD, COD, total phosphorus, and the nitrogen series. Microbiological analyses conducted on all
wastewater and aerosol samples included total coliform, coliphage, standard bacterial plate count, and se-
lected pathogens. The pathogens sought in the Post-Fair period were limited to fecal streptococci and myco-
bacteria. The wastewater analyses were conducted on a composite sample taken in conjunction with each mi-
crobiological aerosol run.
Microbiological aerosol monitoring during both the Pre- and Post-Fair studies for microbiological aero-
sols was conducted using large-volume (1000 1/min) = 1 mVmin) electrostatic precipitator samplers. These
samplers were selected because the large volume of air sampled over a 30-minute period increases the sensitiv-
ity for the microbiological assay. Twenty-one successful aerosol runs were made during the Pre-Fair study
and an additional 29 in the Post-Fair study. The collection and transfer medium selected was brain-heart infu-
sion broth with 0.1 percent Tween 80® , which was shown to be adequate for sample concentration and for
preservation of the microorganisms. The samples were analyzed for the same microbiological parameters as
the wastewater, with the exception of one run for which the collecting fluids from all samplers were pooled
for conduct of a pathogen screen.
A minimum of eight samplers was specified for each run and these were deployed along predesignated
configurations to obtain the necessary information to perform the mathematical modeling. The distances of
the samplers from the spray line were selected based upon sampling protocols and prevailing meteorological
conditions. One sampler was used at a remote, upwind location to ascertain background concentrations of the
organisms sought.
All-glass impingers were used to collect the aerosols from the dye runs, to determine the wastewater
aerosolization efficiency of the sprinklers. Seven dye aerosol runs were conducted in Pre-Fair and an additio-
nal ten in Post-Fair.
In the Pre-Fair study, it was determined that virus levels in air consistently fell below the detection limit
of the method and that special procedures would be required to obtain the necessary sensitivity. Two special
virus runs were conducted in the Post-Fair study which increased the sensitivity and allowed estimation of
impact factors for the enteroviruses. These two runs were conducted with all available samplers operating
VI
-------
close to the spray line under meteorological conditions expected to result in high virus aerosol concentrations.
The sampler collection medium was changed every 30 minutes and the samplers run for a total of about three
hours. The collecting fluids were pooled and concentrated for analysis so that the results were based upon a
total of over 5000 m3 of air.
An explicit model for predicting downwind concentrations of pathogens was developed by expanding
more general mathematical dispersion models. The model adds factors for microorganism impact, viability
decay, and aerosolization efficiency, to the standard diffusion model estimate of pathogen concentration
based on source strength. The distributions of aerosolization efficiency and the impact and decay values for
each organism were determined and these were used to allow evaluation of the model using monitoring data
from other sites.
The study was supported by an extensive quality assurance program conducted primarily during the Pre-
Fair portion of Phase II. Chemical, physical, and microbiological methods used were subjected to accuracy
and precision studies, and alternative laboratories were used, where feasible, to verify the results. Certain
aerosol runs were made in both Pre- and Post-Fair to allow the determination of the precision of the microbi-
ological assay procedures and the estimation of any sampler collection efficiency bias.
The Phase II study yielded several important conclusions. From the wastewater monitoring, the conclu-
sion was reached that wastewater quality as measured by chemical and physical parameters was unrelated to
the generation or transport of microbiological aerosols. In addition, little correlation could be found in the
wastewater between levels of total coliform, fecal coliform, standard bacterial plate count, and coliphage (the
traditional indicator organisms), with the levels of the pathogens which they are intended to indicate.
Results obtained from the aerosol studies indicate that use of the traditional indicator organisms to pre-
dict human population exposure results in extreme underestimation of pathogen levels. The pathogens stud-
ied survived the wastewater aerosolization process much better than did the indicator organisms. Based upon
the results of this study, fecal streptococci may be an appropriate indicator due to ease of assay, levels rou-
tinely seen in wastewater, and the similarity of their hardiness upon impact and viability decay rate to those of
the pathogenic organisms of interest.
Large-volume samplers of the type used in this study are most useful for obtaining the sensitivity re-
quired for assay for bacteria in aerosol samples, especially at background and far downwind locations. Sam-
pling and analyses for enteroviruses in wastewater aerosols requires even greater volumes of air and only a
special effort such as that performed here can be expected to provide the necessary sensitivity to allow their
detection.
The dispersion model developed in this study was validated. It was shown to produce satisfactory results
when used to predict aerosol concentrations at three sites. Most of the predicted results fell within a factor of
five of the measured concentrations when non-chlorinated effluent was being sprayed. The use of such a
model with minimal monitoring is a viable alternative to extensive aerosol monitoring, and is significantly less
costly.
The overall conclusion of Phase II of this program is that microbiological wastewater aerosols are gen-
erated by spray irrigation, do survive aerosolization, and can be transported to nearby populations. The most
reliable means of reducing a potential health hazard from pathogenic aerosols is by disinfecting the wastewa-
ter before spraying. Until the necessary dose-response relationships are developed, neither the level of aero-
solized microorganisms that constitute a hazard nor the degree of required disinfection can be specified.
VII
-------
TABLE OF CONTENTS
Page
FOREWORD Hi
ABSTRACT iv
EXECUTIVE SUMMARY v
LISTOFFIGURES xii
LISTOFTABLES xiii
I. SUMMARY AND CONCLUSIONS 1
A. Wastewater 1
B. Methodology 1
C. Aerosol and Wastewater Microbiology 2
D. General 4
II. RECOMMENDATIONS 5
III. STUDY DESCRIPTION 6
A. Statement of the Problem 6
B. Study Background 10
C. Phase I Summary 11
D. Phase II Objectives and Design 12
1. Phase II Objectives 12
a. Pre-Fair Objectives 12
b. Post-Fair Objectives 12
2. Phase II Design 12
E. Participating Organizations and Principal Personnel 15
IV. STUDY SITE 17
A. Site Description 17
1. Sunol Sewage Treatment Plant 17
2. Treatment Plant Process 19
3. Sunol STP Spray Effluent Quality 19
B. Spray Irrigation Operations 25
C. On-Site Facilities 28
D. General Meteorological Conditions 28
1. Pre-Fair 28
2. Post-Fair 29
V. METHODS AND MATERIALS 32
A. Sample Collection and Handling Methods 32
1. Meteorological Measurements and Instrumentations 32
2. Wastewater Sampling Methods 32
a. Daily Composite Samples 32
b. Grab Samples 32
3. Aerosol Sampling Methods 32
a. High-volume Samplers For Microorganism Aerosols 32
b. All Glass Impinger (AGI) Samplers For Dye Aerosols 34
c. Rotorod Samplers For Fluorescent Particle (FP) Tracer 34
ix
-------
TABLE OF CONTENTS (cont'd)
Page
B. Analytical Methods 34
1. Chemical Analysis of Wastewater Samples 34
2. Microbiological Analyses 34
a. Wastewater 34
b. Aerosols 34
C. Quality Assurance 38
D. Aerosol Sampling Protocols 38
1. Microbiological Aerosol Runs 38
2. Dye Aerosol Runs 40
3. Quality Assurance Runs 40
4. Special Enteric Virus Aerosol Runs 40
E. Data Flow, Processing and Analysis Methods 40
1. Sample Identification and Labeling 40
2. Data Forms and Reporting System 43
3. Aerosol Data Processing 43
4. Computational Techniques 43
5. Statistical Approach 43
VI. RESULTS 45
A. Wastewater Characteristics 45
1. Chemical Data and Patterns 45
2. Microbiological Data and Patterns 45
a. Daily Composite Microbiological Data 45
b. Distributional Characteristics 49
c. Wastewater Analysis Variability 51
d. Equivalence of Composite and Pond Grab Samples 55
e. Relationship of Pathogen Levels to Indicator Organism Levels 58
/. Microbial Characterization 65
g. Respiratory Virus 71
B. Aerosol Run Data Characteristics 72
1. Meteorological and Sampling Conditions 72
a. Meteorological Conditions 72
b. Spray Line and Sampler Configurations 83
2. Sampled Concentration Data 83
a. Dye Runs 83
b. Microbiological R uns 86
c. Quality Assurance Runs 96
d. Virus Runs 104
3. Nature of Aerosol Data 110
a. Distributional Characteristics 110
b. Relative Prevalence Ill
c. Systematic Sampler Differences 112
4. Aerosol Measurement Precision from Quality Assurance Program 117
a. Dye Aerosol Concen tra tions 119
b. Microbiological Aerosol Concentrations 120
x
-------
TABLE OF CONTENTS (cont'd)
Page
5. Particle Size Distributions 126
6. Aerosol Microbial Characterization 131
C. Aerosol Data Analyses 132
1. Microbiological Dispersion Model 132
a. Model Derivation 132
b. Effect of Each Model Factor ]35
2. Diffusion Model Concentration D 137
a. Approach 137
b. Source Inputs 137
c. Calculation Procedure and Model Concentrations 138
3. Aerosolization Efficiency Factor E 138
a. Dye Run Aerosolization Efficiency Estimates 138
b. Microbiological Run Aerosolization Efficiency Predictions 142
4. Impact Factor I and Aerosol Viability Decay Rate A 144
a. Estimation Procedures for I and A 144
b. Impact Factor 1 147
c. Viability Decay Rate A 149
5. Prediction Using the Microbiological Dispersion Model 155
a. Usage Considerations 155
b. Examples v..» 156
6. Preliminary Evaluation of Distance and Solar Radiation Factors 159
a. Analysis of Variance 159
b. Source and Distance Analysis 159
7. Preliminary Assessment of Factors Affecting Microbiological Aerosol Levels 159
D. Evaluation of the Microbiological Dispersion Model 161
1. Evaluation Data 161
2. Accuracy of Model Predictions 165
3. Precision of Model Predictions 169
VII. DISCUSSION OF MICROBIOLOGICAL DISPERSION MODEL 177
A. Model Components 177
1. Aerosolization Efficiency E 177
2. Impact Factor 1 178
3. Viability Decay Rate 179
B. Validity of the Model and its Predictions 180
C. Microbiological Inferences Derived from the Model 181
1. Interpretation of Impact Factors Exceeding One 182
2. Relative Aerosol Survival Hardiness of Microorganism Groups 184
D. Model Applications 184
XI
-------
LIST OF FIGURES
Figure Number Page
IV.A-1. Schematics of Study Site 18
IV.A-2. Plant Layout at Pleasanton, California 20
IV.A-3. Contact Bio-Filter Process (CBF) 21
1V.A-4. Solids Handling Facilities 22
IV.A-5. Plant and Disposal Facility Flow Schematic 23
IV.B-1. Spray Pattern for Fields During Pre-Fair Study 26
IV.D-l. Percent Distribution of Wind Directions and Wind Speeds 30
V.A-1. Site Map with Environmental Measurement Locations 33
V.D-1. Aerosol Sampling Configuration for Pre-Fair 39
V.D-2. Primary (Microbiological) Aerosol Sampling Configuration for Post-Fair 41
V.D-3. Dye Aerosol Sampling Configuration For Post-Fair 42
VI.B-1. Plots of Aerosol Dye Concentration with Downwind Distance 92
VI.B-2 Footnotes for Unusual Events Ill
VI.C-1. Schematic of Aerosol Transport Downwind of a Spray Line 134
VI.C-2. Schematic of Effects of Model Factors 136
VI.C-3. Concentration Isopleths of Diffusion Model D for A Typical Dye Run 139
VI.C-4. Normalized Isopleths of Diffusion Model D for A Typical Microbiolobical Aerosol Run
Having Two Spray Line Contributions 140
VI.C-5 Normalized Isopleths of Diffusion Model D for a Microbiological Aerosol Run Having Two
Spray Line Contributions 141
Xll
-------
LIST OF TABLES
Table Number „
Page
III.D-1 Summary of Phase II Design 13
III.E-1 Participating Organizations and Principal Personnel 16
IV.A-1 Summary of Sunol Secondary Effluent Parameter Analysis Monthly Average 24
IV.B-1 Plant Flow Data 25
IV.B-2 Spray Volume—Field B 27
IV.B-3 Characteristics of Pre-Fair Spray Fields 28
V.B-1 Analytical Methods for Pre-Fair Wastewater Chemical Analyses 35
V. B-2 Analytical Methods for Post-Fair Wastewater Chemical Analyses 36
V.B-3 Microorganisms Routinely Assayed in Wastewater 37
V.B-4 Microbial Types Sought in Pathogen Screen 37
V.E-1 Automated Computing Procedures 44
VI.A-1 Daily Composite Effluent Concentrations of Microbiological Indicator Parameters 46
VI.A-2 Daily Composite Effluent Concentrations of Coliphage and Selected Pathogenic Bacteria
and Viruses 47
VI.A-3 Distributional Characteristics of the Weighted Daily and Large-Volume Effluent Sample
Concentrations of the Indicator and Pathogenic Microbiological Parameters 50
VI.A-4 Microbiological Indicator Quality Assurance Precision Study—Analytical Results 51
VI. A-5 Microbiological Quality Assurance Precision Study—Precision Estimates 52
VI.A-6 Precision Quality Assurance Study for Coliphage 53
VI.A-7 Precision Quality Assurance Study for Bacteria 53
VI.A-8 Precision Quality Assurance Study for Virus 54
VI.A-9 Estimated Replication Error for Pathogenic Analyses 54
VI.A-10 Comparative Total Coliform 56
VI.A-11 Comparative Fecal Coliform Data 57
VI.A-12 Comparative Standard Bacterial Plate Count Data 58
VI.A-13 Comparative Coliphage Data 59
VI.A-14 Unweighted and Weighted Sample Correlations of the Natural Log Transformed Effluent
Concentrations of the Indicator and Pathogenic Microbiological Parameters 61
VI.A-15 Microbiological Constituent Notation 62
VI.A-16 Summary of Best Multiple Regression Equations for Predicting Pathogen Effluent Concen-
tration from the Indicator Effluent Concentrations 63
VI.A-17 Canonical Correlation of the Pathogen Effluent Concentration Set with the Indicator
Effluent Concentration Set 65
VI.A-18 Effluent Concentrations of Usual Microbiological Constituents in Large-Volume Samples
Taken for Microbial Characterization 66
VI.A-19 Summary of Bacteria Identified—Large-Volume Samples 68
VI.A-20 Groups of Ogansisms from Direct Platings and Enrichments for Enterics 69
VI.A-21 Species of Enterobacteriaceae Identified—Large Volume Aerosol Samples 70
VI.A-22 Analysis of Enterobacteriaceae 71
VI.A-23 Quantitative Microbiological Screen of Post-Fair Sample (11-29-76) 72
VI.B-1 Aerosol Run Meterological and Source Data Summary
a. Temperature and Relative Humidity 73
b. Wind Direction, Velocity, Stability and Solar Radiation 77
xiii
-------
LIST OF TABLES (cont'd)
Table Number page
VI.B-2 Sampling Conditions for Microbiological Aerosol Runs 84
VI.B-3 Sampling Conditions for Dye Aerosol Runs 85
VI.B-4 Sampling Conditions for Quality Assurance Aerosol Runs 86
VI.B-5 Sampling Conditions for Virus Aerosol Runs 86
VI.B-6 Pre-Fair Dye Aerosol Run Concentration Data 87
VI.B-7 Post-Fair Dye Aerosol Run Concentration Data 89
VI.B-8 Smoothed Standard Bacterial Plate Counts by Sampler Distance from Microbiological
Aerosol Runs , 97
VI.B-9 Smoothed Total Coliform Concentrations by Sampler Distance from Microbiological
Aerosol Runs 98
VI.B-10 Smoothed Fecal Coliform Concentrations by Sampler Distance from Microbiological
Aerosol Runs 99
VLB-11 Smoothed Coliphage Concentrations by Sampler Distance from Microbiological Aerosol
Runs 100
VI.B-12 Smoothed Fecal Streptococci Concentrations by Sampler Distance from Microbiological
Aerosol Runs 101
VLB-13 Smoothed Pseudomonas Concentrations by Sampler Distance from Microbiological Aerosol
Runs 102
VI.B-14 Smoothed Clostridium perfringens Concentrations by Sampler Distance from
Microbiological Aerosol Runs 103
VLB-15 Smoothed Mycobacteria Concentrations by Sampler Distance from Microbiological Aerosol
Runs 103
VI.B-16 Standard Bacterial Plate Counts from Quality Assurance Aerosol Runs 105
VI. B-17 Total Coliform Concentrations from Quality Assurance Aerosol Runs 106
VI. B-18 Fecal Coliform Concentrations from Quality Assurance Aerosol Runs 107
VI.B-19 Coliphage Concentrations from Quality Assurance Aerosol Runs 108
VI.B-20 Fecal treptococci Concentrations from Quality Assurance Aerosol Runs 108
VI.B-21 Pseudomonas Concentrations from Quality Assurance Aerosol Runs 109
IV.B-22 Mycobacteria Concentrations from Quality Assurance Aerosol Runs 109
VI.B-23 Klebsiella Concentrations from Aerosol Quality Assurance Aerosol Runs 110
VI.B-24 Microbiological Concentrations on Virus Aerosol Runs 110
VI.B-25 Identification of Confirmed Enterovirus Isolates from Aerosol Samples 112
VI.B-26 Distributional Characteristics of the Natural Log Transformed Microorganism Group
Concentrations 113
VI.B-27 Geometric Means and Ratios of Wastewater and Aerosol Concentrations of Microorganism
Groups 115
VI.B-28 Relative Prevalence of Microorganism Groups 116
VI.B-29 Mean Normalized Flavobacterium Counts, Adjusted for Flow Rate (CFU/1) 116
VI.B-30 Summary of F-Ratios 117
VI.B-31 Dye Accuracy Result Summary, Simulated Aerosol Samples 119
VI.B-32 Dye Accuracy Result Summary, Simulated Effluent Samples (ng/1) 120
VI.B-33 Dye Aerosol Concentration Precision 121
xiv
-------
VI.B-34 Standard Bacterial Plate Count Aerosol Concentration Precision 122
VI.B-35 Total Coliform Aerosol Concentration Precision 123
VI.B-36 Fecal Coliform Aerosol Concentration Precision 123
VI.B-37 Fecal Streptococci Aerosol Concentration Precision 124
VI.B-38 Coliphage Aerosol Concentration Precision 124
VI.B-39 Pathogen Aerosol Concentration Precision 125
VI.B-40 Aerosol Concentration Precision Summary 126
VI.B-41 Aerosol Particle Size Distribution from Two-Stage Andersen Samplers Total Coliform
(No./m3) 127
VI.B-42 Aerosol Particle Size Distribution—Total Count (No./m3) 128
VI.B-43 Aerosol Particle Size Distribution—Total Coliform (Percent) 129
VI.B-44 Aerosol Particle Size Distribution—Total Count (Percent) 130
VI. B-45 Aerosol Concentrations for the Aerosol Microbial Characterization Run 132
VI. C-1 Estimates and Predictions of Dye Run Aerosolization Efficiency E 142
VI.C-2 Distribution of Aerosolization Efficiency Values E 143
VI.C-3 Regression Prediction of Aerosolization Efficiency E for the Microbiological Aerosol Runs.. 145
VI.C-4 Run Estimates of Microorganism Group Impact Factor I and Standard Error SE(I) 147
a. Pre-Fair Runs 147
b. Post-Fair Runs 148
VI.C-5 Reliability of Impact Factor Estimates I 149
VI.C-6 Distributions of Aerosol Impact Factor, I 150
VI.C-7 Run Estimates of Microorganism Group Viability Decay Rate A and Standard Error SE(A), s~'
a. Pre-Fair Runs 151
b. Post-Fair Runs 152
VI.C-8 Percentage of Indeterminate Viability Decay Rate Estimates (A = x) 153
VI.C-9 Reliability of Negative Viability Decay Rate Estimates A < 0 153
VI.C-10 Distributions of Viability Decay Rate A, s'1 154
VI.C-11 Prediction of Typical Nighttime Microorganism Aerosol Levels Entering Pleasanton
Residential Area 157
VI.C-12 Prediction of Typical Midday Microorganism Aerosol Levels Entering Pleasanton
Residential Area 158
VI.C-13 Prediction of Microorganism Aerosol Level Extremes Entering Deer Creek Lake Campsite... 160
Vl.D-1 Number of Data Pairs for Model Evaluation 163
VI.D-2 Comparison of Model Predictive Ability of Chlorinated and Unchlorinated Effluent 164
VI.D-3 Analysis of the Accuracy of Standard Bacterial Plate Count Model Predictions 166
VI.D-4 Analysis of the Accuracy of Total Coliform Model Predictions 167
VI.D-5 Analysis of the Accuracy of Fecal Coliform Model Predictions 168
Vl.D-6 Analysis of the Accuracy of Coliphage Model Predictions 169
VI. D-7 Analysis of the Accuracy of Pathogenic Microorganism Model Predictions at Pleasanton 170
VI.D-8 Summary of Model Prediction Accuracy for All Sites and Methods 170
VI.D-9 Analysis of the Precision of Standard Bacterial Plate Count Model Predictions 171
IV.D-10 Analysis of the Precision of Total Coliform Model Predictions 172
xv
-------
VI.D-11 Analysis of the Precision of Fecal Coliform Model Predictions 173
VI.D-12 Analysis of the Precision of Coliphage Model Predictions 173
VI. D-13 Analysis of the Precision of Fecal Streptococci Model Predictions 174
VI.D-14 Analysis of the Precision of Pathogenic Microorganism Model Predictions at Pleasanton 175
VI.D-15 Summary of Model Prediction Precision for All Sites and Methods 176
VII.B-l Meteorological Conditions on Pleasanton Aerosol Runs 181
VII.C-l Aerosol Survival Hardiness of Microorganism Groups 185
xvi
-------
I. SUMMARY AND CONCLUSIONS
A. Wastewater
1. Throughout the Pre-Fair and Post-Fair portions of this study there were no abrupt changes in waste-
water quality which appeared to have an adverse effect on the results of the study. In general, the wastewater
effluent applied was of relatively consistent day-to-day quality. No significant changes were observed during
the conduct of a single aerosol run.
2. In general the quality of the irrigated wastewater was typical of an undisinfected, secondarily treated
domestic wastewater. Mean values during the Pre-Fair studies were: BOD-18.7 mg/L, COD-99.5 mg/L,
TOC-33.0 mg/L, pH 8.4, hardness 235.2 mg/L, TSS033.0 mg/L, total phosphorus-5.6 mg/L, and nitrite, ni-
trate, ammonia and organic nitrogen-0.15 mg/L, 0.06 mg/L, 23.9 mg/L, and 5.6 mg/L, respectively. How-
ever, there were periods when the wastewater quality decreased to that of a poor quality undisinfected second-
arily treated domestic wastewater.
3. Neither the wastewater quality nor the slight changes in the wastewater quality as measured by tradi-
tional chemical/physical parameters appeared to have an impact on the generation or transport of microbio-
logical aerosols in this study.
4. A strong relationship was observed among TOC, COD, and BOD in determining the overall quality
of wastewater at the site. The significance level of the correlation between TOC and BOD was 0.006, and for
the other pairs was less than 0.001.
5. Pathogenic bacteria and viruses were found consistently in the effluent samples, and coliphage were
found in all effluent samples. A wide range of levels of these microbial components was found. Concentra-
tion levels routinely varied by one order of magnitude and variation often approached two orders of magni-
tude.
B. Methodology
1. The quantitative evaluation of microbiological wastewater aerosols during field studies requires high-
volume sampling equipment, competent personnel and extensive laboratory and logistical resources.
2. Studies conducted on the aerosol collection media, the temperature at which the samples are shipped,
and the total time from collection to analysis were examined in detail in the laboratory. The results led to the
design of adequate methods for sampling and analysis such that pathogenic organisms were found consis-
tently.
3. Some difficulties were encountered in contamination of the high-volume aerosol samplers between
aerosol runs. This problem appeared primarily in the standard bacterial plate count and Pseudomonas assays.
Special care must be taken to adequately decontaminate high-volume aerosol samplers between aerosol runs.
4. The microbiological aerosol data varied substantially in quality and informational content. Accord-
ingly, a suitable aerosol data weighting procedure was employed, according to consistent rules, in conducting
the aerosol factor analyses.
5. There is substantial imprecision using the methods employed in this study for measuring microbiolog-
ical concentrations in aerosol samples. The aerosol measurement coefficients of variation were 17°7o for dye,
50% for total coliform and standard bacterial plate count, 58% for fecal coliform and Pseudomonas, 60%
for Clostridium perfringens, 73% for coliphage, 74% for Klebsiella, 77% for fecal streptococci, and 81% for
mycobacteria. While the microbiological aerosol variation due to field sampling sources is considerable, even
more variation was caused by analytical sources such as sample processing, shipping, and laboratory proce-
dures. Relatively little of the analytical variability is reflected in replicated analyses, which is the usual manner
of reporting analytical variation.
-------
6. A special study of respiratory viruses in wastewater found confirmed viruses in five of the forty roller
tubes cultured. Typing disclosed that four of the five tubes contained ECHO virus 6, while the other viral
isolate could not be identified. Echoviruses 6 may occur as either a respiratory-tract virus or as an enteric
virus. The failure to isolate purely respiratory viruses in the Pleasanton wastewater confirmed our suspicion
that the likelihood of finding respiratory viruses in wastewater is very small.
7. Reliable enterovirus aerosol concentrations can be detected by the methods employed in the two spe-
cial virus aerosol runs conducted at Pleasanton. The measured enterovirus aerosol concentrations obtained,
0.011 pfu/m3 and 0.017 pfu/m3, were 1-1/2 orders of magnitude higher than was expected based on the mea-
sured wastewater concentrations during these runs.
8. At sites with aerosol source strengths similar to the Pleasanton site and with sampling and assay meth-
ods currently available, it is generally not advisable to conduct microbiological aerosol sampling at distances
beyond 200 meters from a wastewater aerosol source.
9. In the quality assurance aerosol runs for systematic sampler differences, it was concluded that after
correcting for the air flow rates, there was no systematic bias in microbiological collection efficiency among
the high-volume samplers evaluated.
10. An acceptable state-of-the-art procedure extending previous models <'-2'3'4' has been developed for the
estimation of microorganism aerosol concentrations in wastewater aerosol downwind from a spray irrigation
site. The microbiological dispersion model developed permits the prediction of downwind aerosol concentra-
tions of specific pathogen and indicator microorganism groups emanating from sprayed wastewater aerosols.
This multiplicative model, P = D«E«I»eia, incorporates a diffusion factor D for which any applicable stan-
dard atmospheric dispersion model can be used; a wastewater aerosolization efficiency factor E that depends
upon atmospheric and operating conditions; and microbiological impact and viability decay (die-off) factors I
and eia which depend upon the microorganism group and atmospheric conditions. If allowance is made for
the imprecise nature of microbiological aerosol data, this multiplicative model appears adequate to represent
microbiological dispersion.
11. Prediction by the microbiological dispersion model of the pathogenic microorganism concentrations
from wastewater aerosol sources to which downwind workers and nearby residents are exposed is the most
promising method of determining their level of pathogen exposure. However, the model has been validated
only to downwind distances of about 100 meters.
C. Aerosol and Wastewater Microbiology
1. Prior to selecting the test organisms for a study of this nature (or for the microbiological monitoring
at a spray irrigation site) it is essential that several site specific screens of a wide variety of organisms (both
pathogens and indicators) be conducted.
2. Over the ranges of the effluent microorganism concentrations obtained during sampling at Pleas-
anton (from one order of magnitude for coliphage and total coliform to well over two orders of magnitude
for Pseudomonas and Clostridium perfringens), there are only the most tenuous of relationships between
some pathogenic organisms and some indicator organisms. For all practical purposes the use of indicator or-
ganisms as a measure of specific pathogen levels in wastewater is invalid.
3. The geometric mean aerosol concentrations obtained at 50 meters downwind of the wetted spray area
were:
standard bacterial plate count 460/m3
total coliform 2.4 MFC/m3
fecal coliform 0.37 MFC/m3
coliphage 0.38 PFU/m3
-------
fecal streptococci 0.61 CFU/m3
Pseudomonas 34.CFU/m3
Klebsiella <5.CFU/m3
Clostridium perfringens 0.9 CFU/m3
mycobacteria 0.8 CFU/m3
enteroviruses (3 and 5 day) 0.014 PFU/m3
Individual aerosol measurements frequently differed by more than an order of magnitude from these mean
values.
4. Limited particle size data obtained with two-stage Andersen samplers showed a substantial portion in
the respirable range. The median percent respirable particle values downwind of the spray line were 44 percent
for total count and 74 percent for total coliform. In general, there was a higher percentage of respirable parti-
cles at close downwind distances (5 to 25 meters), than at background and farther downwind distances. These
meager data is in general agreement with more thorough particle size studies performed at other sites'5-6'71.
Particle size was not considered in the mathematical modeling.
5. The use of most of the traditional organisms for monitoring wastewater aerosols results in a gross
underestimation of pathogen levels. Total coliform, fecal coliform, coliphage, and standard bacterial plate
count, which are commonly used as indicators of wastewater pathogens, do not survive wastewater aerosoliza-
tion nearly as well as do the pathogens studied.
6. One of the better indicators for wastewater aerosol monitoring may well be fecal streptococci due to
the relative ease of the assay, the levels found in the wastewater, its relative hardiness during aerosolization,
and its relatively low viability decay rate. However, an apparent problem was the occasional presence of fecal
streptococci in aerosols due to non-wastewater sources.
7. Although Klebsiella was relatively prevalent in the wastewater, it was far less prevalent in the waste-
water aerosol. It appears that Klebsiella die off rapidly during the aerosolization process. This finding is in
contrast to data seen in the literature which consistently report Klebsiella as the predominant pathogen found
in the air near spray irrigation sites and near sewage plants. More confirmation steps were used in this study
than in earlier studies. If the confirmation steps had been stopped at the point used by other investigators,
more values would have been reported as Klebsiella when, in fact, they were primarily other organisms of the
mucoid type.
8. There was no significant difference in the coliform or coliphage concentration in corresponding ef-
fluent samples taken from a spray head during the aerosol runs and from the effluent composite samples at
the pond pump. The standard bacterial plate count, however, was significantly higher in the spray field sam-
ples. The correlations of the spray field and pond composite microorganism concentrations were generally
significant, but not adequate for prediction.
9. The median aerosolization efficiency E obtained for the Rainbird® impact sprayers over 17 dye runs
during Phase II at Pleasanton was 0.33%. There was over an order of magnitude of variation in aerosoliza-
tion efficiency estimates from the tenth percentile (0.09%) to the ninetieth percentile (1.8%). Eighty percent
of this variation in aerosolization efficiency at Pleasanton appears to result from changes in meteorological
conditions (air temperature, wind velocity, and solar radiation) that affect the evaporative capability of the
air.
10. The median impact factor estimates I for the microorganism groups studied were 0.13 for fecal col-
iform (13% survive aerosol impact), 0.16 for total coliform, 0.21 for standard bacterial plate count, 0.34 for
coliphage, 0.89 for mycobacteria, 1.2 for Clostridium perfringens, 1.7 for fecal streptococci, 14 for Pseudo-
monas, ca. 10 for three-day enteroviruses (mostly polioviruses), and ca. 40 for all (3-day and 5-day) enterovi-
ruses. Most individual impact factor estimates were quite imprecise, reflecting the imprecision of the microbi-
-------
ological aerosol concentration measurements. Since the middle range of impact factor values (fortieth to
sixtieth percentiles) for each microorganism group were quite consistent, they are considered to be character-
istic of the microorganism groups' typical survival through aerosol impact.
11. As indicated by impact factors exceeding 1.0, the enteroviruses and some hardy bacterial pathogens
are frequently found in wastewater aerosols at higher concentrations than could be expected based on their
wastewater concentrations. The survival, masking, and mechanical splitting hypotheses discussed later may
collectively explain this phenomenon.
12. The indicator microorganism groups, especially total coliform and fecal coliform, experience more
consistent and rapid die-off with aerosol age than do the pathogenic bacteria evaluated. The viability decay
rates A for total coliform and fecal coliform were more rapid, more reliable, and more frequently detectable
than those of the other microorganism groups. The median viability decay rates were -0.032s'1 for total col-
iform and -0.023s-' for fecal coliform. Viability decay was less rapid for coliphage, Clostridium perfringens,
and standard bacterial plate count and its effect could only be ascertained within 100 meters on about half the
runs. Viability decay could seldom be ascertained for fecal streptococci, mycobacteria, and Pseudomonas.
No attempt was made to determine the viability decay of enteroviruses.
13. The range of impact factor estimates for each microorganism group was broad, generally covering
two orders of magnitude from the tenth percentile to the ninetieth percentile. The detectable viability decay
rates of each microorganism group also covered a wide range. Preliminary analyses suggest ambient condi-
tions such as low relative humidity, high wind velocity, and a large temperature differential between wastewa-
ter and air may reduce the initial survival, i.e. produce low impact factor values I. Viability decay may be
more rapid with high solar radiation, high temperatures, and middle or low relative humidity. Research is
needed to quantify the relationships of impact factor variation and viability decay rate variation to the am-
bient atmospheric conditions for the specific microorganism groups.
14. The accuracy and precision of microbiological dispersion model predictions have, in general, been
validated to 100 meters downwind of spray sources of unchlorinated wastewater aerosols. Most model predic-
tions (e.g., 77 percent for standard bacterial plate count, 71 percent for total coliform, and 80 percent for
coliphage) were within a factor of five of the net measured aerosol concentrations evaluated. Considering the
imprecision and cost of measuring microorganism aerosol concentrations from spray irrigation by field sam-
pling, using predictions of the microbiological dispersion model supplemented with minimal field sampling
does appear to be a preferable alternative to extensive field sampling when the sprayed wastewater does not
contain residual chlorine.
D. General
Results of this study indicate that pathogenic organisms do survive as aerosols from the spray irrigation
of wastewater and can be transported downwind into nearby populated areas. Until such time as more infor-
mation is developed on the infectious dose for a particular organism, effective control of pathogens must be
accomplished prior to spray irrigation to minimize this risk. The most reliable means of reducing potential
health hazards from pathogenic aerosols is by disinfecting the wastewater before spraying. Until the necessary
dose-response relationships are developed neither the level of aerosolized microorganisms that constitute a
hazard nor the degree of required disinfection can be specified.
-------
II. RECOMMENDATIONS
1. Dose-response relationships need to be developed for the pathogenic microorganisms prevalent in
wastewater aerosols. The available techniques include epidemiological studies of human populations, sentry
animal studies, laboratory animal studies, clinical studies, and/or professional judgment. Epidemiological
studies may be the best technique. The evaluation of spray irrigation as a viable means of disposing of waste-
water will remain incomplete until such dose-response relationships can be established.
2. A reliable procedure is needed for selected values of the model parameters 1 and A when predicting
microorganism aerosol concentrations using the microbiological dispersion model. There are preliminary in-
dications in our data supported by considerable published data that the I and A values for a microorganism
group depend highly upon ambient atmospheric conditions. A careful multivariate analysis of the existing va-
lues of I, A, and the meteorological conditions for each Pleasanton run should be conducted. The resulting
relationships of microorganism I and A values as a function of ambient meteorological conditions would pro-
vide a substantive basis for their selection in model prediction applications.
3. To assess the predictive capacity of the microbiological dispersion model, a thorough evaluation of
the model should be conducted through a limited field sampling program using the Pleasanton study methods
at other sites. The enhanced model to be evaluated should incorporate the I and A selection procedure dis-
cussed in the preceding recommendation. The model evaluation conducted herein needs to be extended be-
cause the field and sampling methods differed from the Pleasanton methods, pathogens were not assayed at
sites other than Pleasanton, and no adequate I and A parameter selection procedure existed.
-------
III. STUDY DESCRIPTION
A. Statement of the Problem
Application of wastewater to land is perhaps the oldest method of disposing of wastes by man. The
more recent trend has been to discharge treated and untreated wastes into streams, lakes, and oceans. During
the past several years, there has been increased interest in applying treated wastewater to land as an alternative
to discharging into surface waters, in order to avoid the contamination of these bodies of water. The Environ-
mental Protection Agency (EPA) has in recent years required that applicants for federal construction grants
(Section 201) for wastewater treatment facilities show in their request that they have considered the applica-
tion of wastewater to land as an alternative.(8) More recently, the EPA has announced plans to press vigor-
ously for the recycling of wastewater via land application.(9) There are several advantages to applying treated
wastewater to land, including re-use of the water, avoiding contamination of surface waters, and return of
nutrients to the soil. Many small communities primarily in arid regions, have long employed wastewater land-
application systems as a means of water conservation and waste disposal. During the past several years, a
large metropolitan area, Muskegon County (Michigan), has completed a project involving biological treat-
ment, storage lagoons, disinfection, and spray irrigation for disposal of wastewater from a population of
more than 170,000.
Land application of wastewater can be accomplished by several methods, which can be categorized as
overland flow, rapid infiltration, and slow infiltration. Spray irrigation is perhaps the most popular method
of wastewater application because it can be applied to a wide variety of land sites and the irrigation apparatus
can be moved readily from one location to another. For many municipalities, both small- and medium-sized,
spray irrigation is the most attractive means for land application of treated wastewater. Recreational areas
such as parks, golf courses, and highway right-of-ways, can be irrigated. Irrigation of land sites immediately
adjacent to the waste treatment facility for growing of grass cover is also an economically and environmen-
tally attractive use of the wastewater. The applicability of spray irrigation to a much larger segment of waste
treatment facilities will necessitate that these spray sites be located adjacent to the facilities and, therefore,
probably near populated areas. Certainly, application of wastewater to recreational areas by spray irrigation
has a high potential for contact between the wastewater aerosol and individuals, both the spray applicators
and surrounding populations.
A number of investigators have raised questions regarding the health and hygienic aspects of application
of wastewater to land, especially by spray application. These investigators include Wellings, et a/.,(10) Sorber
and Guter,*1» Parson, cra/.,(12> Katzenelson and Teltch,<13> and Elliott and Ellis<14>.
Wellings, et al., cautioned the utilization of spray irrigation of wastewater because of the many un-
knowns regarding the fate of pathogenic bacteria and especially viruses. These authors' studies have been pri-
marily directed at possible contamination of ground water by viruses following spray application of treated
wastewater. Sorber and Guter examined the literature regarding the health aspects of land application of
wastewater by spray irrigation and concluded that there was a probability of humans inhaling pathogenic
aerosols near a spray irrigation site. They recommended that research be performed to determine the viability
of human pathogenic microorganisms present in biological aerosols from spray irrigation of wastewater.
They also suggested that there be a determination of the type of spray distribution system, nozzles, and asso-
ciated operating pressures necessary to minimize the health hazards from biological aerosol formations.
Earlier work by investigators such as Ledbetter and Randall"5* showed that activated sludge treatment
facilities operations were sources of aerosolized bacteria, some of which might be pathogenic. These bacterial
microorganisms were from aerosolized wastewater in the aeration basins present in the treatment facilities. It
is, therefore, quite probable, as pointed out by Sorber and Guter, that spray application of treated wastewater
6
-------
with various types of spray irrigation equipment and with different spacing and type of nozzles would signifi-
cantly influence the quantities of aerosols formed. The aerosols generated by the sprinklers might contain pa-
thogenic bacteria and viruses, and they might consist of particles in the size range of 0.01 to 50 microns."6*
From the data seen in the literature, it appears that spray irrigation of treated wastewater will result in
formation of aerosols which may contain pathogenic microorganisms. Inhalation of pathogenic microorga-
nism may lead to infection. This is a major source of the public health concern for this method of wastewater
disposal and the reason for conducting this research. From limited data, it appears that some of these orga-
nisms will remain viable for extended periods of time and may be carried by winds for considerable distances
from the spray irrigation sites.
An environmental monitoring effort was performed by Katzenelson and Teltch near spray irrigation
fields in Israel. The effluent sprayed on these fields was from partially-treated undisinfected municipal waste-
water and levels of coliform bacteria in the effluent were approximately the same as those seen in raw waste-
water present in the United States. Coliforms were measured in aerosols and a portion of these aerosol sam-
ples was further examined for the presence of Salmonella. They found elevated levels of coliform bacteria 300
meters downwind from the spray irrigation site. In one of the aerosol samples collected at 600 meters down-
wind, they found an isolate of Salmonella. They calculated that an individual 100 meters downwind from the
wastewater sprinkler line would inhale approximately 36 coliform bacteria in a period of 10 minutes. These
authors concluded from these preliminary results that there may be a relatively neglected potential danger to
agricultural workers and neighboring settlements from the use of spray irrigation wastewater. They stated
that proper disinfection of wastewater for irrigation purposes may be the most effective means of reducing
this risk.
Sorber, et a/.,(l7) conducted an extensive environmental monitoring effort at a field site in Arizona
where chlorinated secondary municipal effluent was used to irrigate a golf course. These authors performed
field testing of both chlorinated and unchlorinated effluent to enumerate bacteria present in the effluent, to
determine the fraction of wastewater entering the aerosol state, and to determine the survival in aerosols of
total aerobic, indicator, and selected pathogenic bacteria. Klebsiella was the most commonly found pathogen,
with fecal streptococci found in some samples. When chlorinated effluent (versus unchlorinated effluent) was
sprayed, much lower levels of pathogenic bacteria were found in the aerosol samples, although the reduction
was less than expected from comparison of pre- to post-chlorinated wastewater bacteria levels. Elevated levels
of bacterial aerosols were measured out to 200 meters downwind of the spray line. It was estimated that bacte-
rial levels would be present above background up to 500 to 1800 meters for unchlorinated wastewater, de-
pending on meteorological conditions. A prediction model was developed in the course of this study for esti-
mating concentrations of microbiological constituents downwind of the spray line. Components of the model
included meteorological and other measurements taken downwind as well as the concentration of the micro-
organism in the effluent. In the discussion on microbial aerosols from the spray irrigation of wastewater,
these authors point out that the factors of prime interest are: (1) the biological aerosol concentration at any
distance from the source, (2) the buffer zone required to reduce such aerosols to near background levels, and
(3) the effect of terminal disinfection or other polishing wastewater treatments on aerosol strength. The aero-
sol levels of viable microorganisms are dependent upon the levels of these organisms in the wastewater, the
proportion of wastewater that is aerosolized, the volume of water sprayed per unit time, the aerosol decay
rate and the atmospheric stability, wind speed, and other meteorological parameters. These authors defined
aerosolization efficiency as the proportion of wastewater that becomes divided into droplets sufficiently small
to remain in an airborne state. They found a statistically significant increase in the mass median diameter of
viable bacteria-bearing particles with distances downwind of the spray irrigation sites. They postulated that
this may indicate a lower rate of decay for bacteria associated with the larger particle size material. An alter-
-------
native mechanism suggested was that the droplets receive pre-existing clumps or aggregates of many bacteria
cells which tend to form larger solid nuclei. There was no indication that the particle size distribution for a
viable airborne bacteria recovered from chlorinated effluent aerosols differed from those recovered in experi-
ments without chlorination. They suggested that terminal disinfection is probably a more practicable and cost
effective means of limiting the problems with pathogenic aerosols than are buffer zones.
Little information is available concerning the possible health effects following exposure to the low levels
of bacteria and possible viruses that are present in ambient air near spray irrigation facilities. It is known that
the inhalation of specified quantities of pathogenic bacteria or viruses can infect persons, but epidemiological
data concerning a definable health effect following exposure to aerosols from a wastewater treatment facility
are quite limited. Clark, et al. (18) reviewed the literature concerning the possible health effects of persons ex-
posed to municipal wastewater via physical and aerosol routes of exposure. The review covered microbial
aerosols from activated sludge, trickling filter, and spray irrigation of wastewater. They found that no corre-
lation had been made between specific microorganism levels and the incidence of selected diseases. They re-
ported on a 3-year prospective epidemiology-serology study to be performed in their laboratory involving
sewage maintenance workers and an appropriate control group. This study has been expanded to include sew-
age plant operators and appropriate controls for study of effects of the exposure to aerosols from wastewater
plant operations. Overall, the prospective study is aimed at determining the possible health effects from expo-
sure to relatively high levels of wastewater via physical and oral ingestion of wastewater for the sewage work-
ers and to inhalation of aerosols for the plant operators. These studies include collection of symptomatology
data, blood, urine, throat swabs, and stools for examination for viral, bacterial, and parasitic isolation, as
well as a detailed serological analysis. The results of this study should provide indications of possible health
effects at relatively high levels of exposure to adult male populations.
A retrospective epidemiological study of communicable diseases associated with spray irrigation of
wastewater was performed near settlements in Israel by Katzenelson, et a/.('9). These authors examined the
incidence of enteric communicable diseases in 77 kibbutzim (agricultural settlements) practicing spray irriga-
tion with partially-treated, nondisinfected wastewater. These data were compared with that from 130 kibbut-
zim not practicing spray irrigation. They found that the incidence of Shigellosis, Salmonellosis, typhoid fever,
and infectious hepatitis were two to four times higher than for those communities not practicing spray irriga-
tion. They also noted that for the months in which there was no spray irrigation, i.e., the winter months, there
were no differences between the study populations for these enteric diseases. The study populations (positive)
lived from 100 to 3000 meters from the spray irrigation field. This retrospective study provides some epidemi-
ological evidence for an increased risk for enteric communicable diseases among populations living near sites
spray irrigating with municipal wastewater. The study does not identify the transmission route, that is, via
physical or aerosol exposure, nor does it directly relate spray irrigation with these elevated diseases. The oper-
ators of the spray irrigation facilities live in these kibbutzim; thus, one pathway might be via physical expo-
sure to the wastewater irrigation site workers and via the clothes of the irrigation site workers when they re-
turn home. These authors point out the need for caution in the utilization of poorly-treated wastewater via
spray irrigation near residential areas. A follow-up study is underway.
A prospective environmental epidemiological study by Carnow, et al., at the University of Illinois (EPA
Grant R-805003-01) is nearing completion. This environmental epidemiological study was performed near an
existing activated sludge treatment facility located in Skokie, Illinois. This effort has similarities to the study
(to be discussed below) conducted by Southwest Research Institute in the type of monitoring performed.
However, a major difference is that it was located near an existing wastewater treatment facility, whereas the
study conducted by Southwest Research was designed for a new plant with a before-operational monitoring
effort and a monitoring effort after the plant was in operation. The objective of Skokie study was to collect
-------
health effects data relative to the operation of the plant and to examine these data with regards to health ef-
fects related to the distance the population lives from the plant. Environmental monitoring was performed
and an assessment of the bacterial microorganisms present around this plant was made. This study had 269
households living from about 600- to 800-meters from the center of the wastewater treatment plant.
An environmental monitoring and a prospective epidemiological study was performed by Johnson, et
a/ (20) Of Southwest Research Institute for a new activated sludge treatment plant located near Chicago, Illi-
nois. The purpose of this study was to identify possible health effects which might be attributed to the opera-
tion of an activated sludge wastewater treatment plant. The program was to involve three independent modes
of investigation of a new activated sludge treatment plant prior to its operation and after its initial operation:
• Environmental monitoring in the vicinity of the site to determine the source and transport of
indicator and pathogenic microorganisms and of trace metals.
• A health survey of a cross section of the households located within a 5-km radius of the plant
aeration basin to examine the incidence of respiratory and intestinal diseases and symptoms.
• Analysis of clinical specimens from more than 200 participants residing within 3.5 km of the
plant to determine viral antibody tilers, to isolate pathogenic bacteria, viruses, and parasites,
and to measure trace metal concentrations.
Each of the above was designed to determine whether the data followed a pattern that might implicate
the wastewater treatment plant as a health hazard. The results showed that for the trace metals studied (cad-
mium, lead, mercury, and zinc), only mercury appeared to be elevated as a result of the wastewater treatment
plant and that mercury elevation was not transported away from the plant. Also, the levels of these trace met-
als were not elevated in the soil and water samples collected in neighborhood residential areas relative to back-
ground levels. The plant did appear to be a source of indicator bacteria, coliphage, pathogenic bacteria, and
enteroviruses emanating from its aeration basins. However, the levels of these microbial agents in air, soil,
and water samples in the neighboring residential areas were not distinguishable from background levels. From
the patterns observed in the household health survey, there was an increased incidence of skin disease and in
the symptoms of nausea, vomiting, general weakness, diarrhea, and pain in chest on deep breathing among
residents living close and in prevalent downwind directions from the wastewater treatment plant. Viral anti-
body tests and attempted isolation of many pathogenic bacteria, parasites, and viruses, however, yielded no
evidence of an adverse wastewater treatment plant effect. The findings overall obtained in this study did not
detect a public health hazard for populations living beyond 400 meters of this well-operated wastewater treat-
ment plant. The lack of sufficient participants living close to the wastewater treatment plant precluded an as-
sessment of the possible hazard near the plant. The study does confirm results seen in previous studies of
wastewater treatment plants in that the plant is a source of pathogenic bacteria and possible viruses; however,
levels above background were not obtained at distances beyond approximately 300 meters. Of the pathogenic
bacteria monitored in aerosol samples (fecal streptococci, Salmonella, Shigella, Pseudomonas, and Kleb-
siella), only Pseudomonas, fecal streptococci, and Klebsiella were found. Pseudomonas was present in virtu-
ally all aerosol samples, while fecal streptococci and Klebsiella were found in only a few. One aerosol sample
taken 300 meters downwind was positive for enteroviruses (poliovirus type III).
A recent review by Akin, et al. ,(2I) of the health hazards associated with the treatment and disposal of
wastewater effluents and sludge, reports that in the absence of adequate epidemiological data to evaluate the
potential health hazard from pathogens applied to soil, the monitoring for the occurrence of the pathogens in
the environment must be the primary public health measure. These authors also pointed out the possible
health hazards of parasites in the application of treated wastewater and sludge to land. In particular, the pro-
tozoan of greatest health interest in the past several years has been Giardia lamblia. Cysts from this protozoan
-------
are able to survive many of the wastewater and sludge treatment processes and only a small number of cysts
are necessary to infect humans. However, because of their size, it is questionable whether the cysts can travel
as aerosols for large distances.
From the review of the literature on environmental monitoring, it is apparent that treated wastewater
can contain significant quantities of microorganisms and pathogenic bacteria and viruses. It also appears that
various processes involved in wastewater treatment facilities can create aerosols of these wastewaters con-
taining viable microorganisms. Spray irrigation appears to offer the greatest potential for aerosolization for
bacterial organisms. Some evidence is indicated that bacteria are present in aerosols from these sites. Bacterial
organisms are present in air around spray irrigation facilities at distances up to at least 300 meters downwind
and model projections indicate that these aerosols could be present above background out to several kilome-
ters. The studies also indicate that various meteorological conditions can have a significant impact on the vi-
ability of these organisms in air following aerosolization. In general, high wind conditions, high humidity,
and low solar radiation tend to enhance the viability of bacterial aerosols.
B. Study Background
This research effort was jointly funded by the U.S. Army Medical Research and Development Com-
mand, Fort Detrick, Maryland, and the Health Effects Research Laboratory of the Environmental Protection
Agency, Cincinnati, Ohio. Each of these governmental organizations needs definitive data regarding the pos-
sible health effects associated with spray irrigation of municipal wastewater. The Army has a number of in-
stallations throughout the United States and foreign countries where disposal of municipal wastes is required.
Application of treated wastewater to land, especially via spray irrigation, is an attractive means of final dispo-
sal and serves as an alternative to discharge into a watercourse. This permits the reuse of the wastewater for
irrigation purposes and avoids the discharge of wastes into rivers, lakes, and oceans. As stated earlier, the
Environmental Protection Agency has been an advocate of land application of treated wastewater as an alter-
native to discharging into surface waters. During recent times, the EPA has announced that it intends to press
vigorously for the use of construction grant funds for wastewater treatment plant facilities to be directed at
application of treated wastewater to land. The funds obligated.for construction of new wastewater facilities in
the United States are in the billions of dollars and the EPA has both the authority and the responsibility to
insure that these facilities are constructed to effectively treat wastewater in the most economical and cost-ef-
fective means possible. In addition, the EPA must insure that the health of people in communities near these
facilities is protected and that surface and ground waters are not significantly contaminated by chemical or
biological pollutants. As discussed earlier, spray application of wastewater is one of the most attractive land
application methods from an engineering standpoint. It does, however, offer the possibility that aerosols,
formed during spraying operations, can be transported to nearby human populations and that some portion
of the population will be infected by wastewater-associated pathogenic microorganisms. The initial objective
of this study was to collect information regarding the types and quantities of microbial organisms emanating
from a spray irrigation facility and to study these concentrations downwind of the spray facility into a pop-
ulated area. A potential extension of this effort would examine the health status of a population adjacent to
the spray fields as compared with a suitable control population. The initial monitoring study was designed
such that sufficient numbers of samples would be collected and analyzed, along with a sound quality control
program so that statistically valid data would be obtained. The overall study would provide both the U.S.
Army and the EPA with much of the data necessary to provide design criteria for construction and perfor-
mance of spray irrigation facilities. This is extremely important since it is apparent that the utilization of
spray irrigation of treated municipal wastewater could have numerous advantages over conventional waste-
water treatment for a wide section of the United States.
This project was performed in three phases. At the completion of each phase, an analysis of the data
10
-------
was made and a report prepared, with the final design of the next phase of study dependent upon the results
of the previous phase. Thus, there was considerable flexibility during the conduct of this effort to adjust the
program's emphasis on the basis of gained knowledge.
Phase I was designed to select a suitable site for the conduct of the study and to develop the optimum
methods for sampling and analysis of wastewater and aerosol samples for various types of bacteria, viruses,
and chemical constituents. Phase II was designed to perform extensive environmental monitoring of the se-
lected spray site to cover a period of some eight months. This phase would accurately measure the quantities
and types of pathogens and other constituents under a variety of meteorological conditions and a range of
source strengths of these materials in the wastewater. Phase II was divided into two parts: one conducted
prior to a county fair held near the selected site, labeled "Pre-Fair", and the second performed afterward,
labeled "Post-Fair". In general, the objectives of the Pre-Fair portions of Phase II were aimed at characteri-
zation of the aerosols for bacterial and viral microorganisms within 100 meters of the spray fields. Environ-
mental monitoring during the Post-Fair phase emphasized sampling and analysis of selected pathogenic mi-
croorganisms, with aerosol monitoring conducted at distances up to 600 meters from the spray fields,
extending into the populated areas. Phase III of this program is an optional phase to be directed at examining
the potential human health effects of the spray irrigation facility. The study would examine a population liv-
ing near the spray irrigation facility and compare their health status with a suitable control population. A de-
cision to conduct Phase III has not been made but will follow examination of the data obtained during Phases
I and II. The findings for the Phase I study are summarized in the Phase I report published in December
1975.122)
C. Phase I Summary
Following a telephone survey of known wastewater spray irrigation facilities in the United States and an
on-site survey of two locations, the wastewater land treatment system in Pleasanton, California, was selected
for evaluation. The site selection criteria are listed in the Phase I report, but one of the important require-
ments, and one of the most difficult to satisfy, was that a suitable population had to live within one mile of
the spray irrigation facility. The Pleasanton, California site met most of the desired criteria, including an ad-
equate study and control population.
The Phase I results indicated that some of the initially-selected methods for sampling, sample transport,
and analysis were inadequate to accurately determine the levels of pathogenic bacteria and viruses in both ef-
fluents and aerosols. Measurement of indicator organisms and other chemical constituents showed that the
wastewater at this site was of a quality to be expected of a plant not practicing disinfection.
It was apparent from the preliminary aerosol sampling conducted during Phase I that the micrometeo-
rology of the site would complicate the overall air sampling protocol and the interpretation of the data. No
significant daily or hourly changes were seen in the chemical or biological constituents measured in the ef-
fluent, apparently due to the utilization of two aeration ponds at the end of the treatment plant that appeared
to dampen cyclic changes which might have occurred through the plant. This made the environmental mon-
itoring study easier since a rather uniform quality of wastewater was being sprayed.
One of the important findings of Phase I was that high-volume aerosol samplers were essential for mea-
surement of pathogenic microorganisms present in the air downwind from the spray irrigation site. Samplers
such as the Andersen and AGI, although perhaps having similar sampling efficiencies, would not sample suf-
ficient quantities of air to provide the necessary sensitivity. It was concluded that aerosols generated from the
adjacent secondary wastewater treatment plant itself should not complicate the study of aerosols from the
spray irrigation fields because the head work, aeration chamber, and trickling filter were all covered to con-
trol odors. There was a possibility that the aeration ponds, also adjacent to the spray fields, could generate
aerosols and confound the field study results.
11
-------
D. Phase II Objectives and Design
1. Phase II Objectives
As stated above, Phase II was divided into two subtasks, Pre-Fair and Post-Fair. The Pre-Fair
monitoring effort was performed so it would be completed prior to the beginning of a large county fair held in
Pleasanton, California in the summer months. During the fair, a large influx of wastewater from animal
holding facilities produced a typical wastewater effluent and no monitoring was to be performed during the
fair or immediately afterwards. The Post-Fair monitoring effort was conducted after the effluent had re-
turned to more typical characterization.
The objectives of Phase II are listed below for both Pre-Fair and Post-Fair.
a. Pre-Fair Objectives
The following is a list of the primary objectives to be accomplished during the Pre-Fair activ-
ities of Phase II.
Microbial Aerosol Runs
• begin evaluation of factors affecting microbiological aerosol levels within 100 me-
ters of the spray source.
Dye Runs
• determine the aerosolization efficiency range of the sprinkler irrigation machinery
used at the Pleasanton site.
Effluent Samples
• determine in-depth pathogen screen
• assess validity of using the common measures of wastewater microbiological quality
(standard bacterial plate count, total coliforms, fecal coliforms) as indicators of the
pathogen levels of the effluent.
• examine relationships between the microbiological and chemical water quality con-
stituents.
Quality Assurance
• determine accuracy and precision of laboratory analyses
• determine if there are systematic differences in high-volume sampler collection
efficiency.
b. Post-Fair Objectives
• identify and quantify the factors affecting microbiological aerosol levels over a wide
range of meteorological conditions.
• develop a general microbiological dispersion/die-off model for appropriate micro-
organism groups that can be applied at other wastewater spray irrigation sites.
• predict the downwind microbiological aerosol concentration in the residential areas
adjacent to the spray fields, relative to background levels.
2. Phase II Design
The types of samples obtained in the Phase II study (Pre- and Post-Fair) are summarized in Table
III.D-I.
12
-------
Table III.D-1.
SUMMARY OF PHASE II DESIGN
Pre-Fair
Post-Fair
Wastewater Samples
Dai ly
Composites
Microbiols'gcal Dye Run
Run Composites Grabs
Chemical Analyses
Total Chlorine X
Free Chlorine x
pH X
Total Organic Carbon CTOC) X
Total Solids X
Total Suspended Solids (TSS) X
Biochemical Oxygen Demand (BOD) X*
Chemical Oxygen Demand (COD) X*
Total Phosphorus X*
Nitrite Nitrogen X*
Nitrate Nitrogen X*
Ammonia Nitrogen X*
Organic Nitrogen X*
Conductivity
Rhodamine Dye
fficrobiological Analyses
Standard Bacterial Plate Count X
Total Coliform X
Fecal Coliform X
Coliphage X
Fecal Streptococci
Pseudomonas
Klebsiella
Clost.ridi.um pertringens
Mycobacteria
3-day enterovirus
5-day enterovirus
Pathogen screen
500 ml
Grabs
Large Volume
(201) Grabs
Microbiological
Run Composites
X
X
X
X
X*
X*
X*
X*
X*
X*
X
Dye Run
Grabs
Large Volume
(201" Grabs)-
-------
Table III.D-1 (cont'd)
Pre-Fair
Aerosol Samples
Microbiologi cal Analyses
Standard Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Fecal Streptococci
Pesudomonas
Klebsiella
Cltfstrji.dium perfjjngens
Mycobacteria
3-day enterovirus
S-d'ay enterovirus
Pathogen screen
Rhodamine Dye Analyses
Microbiological
X
X
X
X
X
X
X
X
X
X
Quality T
Dye Assurance
5
5
5
2
2
2
2
2
X
Large '
Volume
X
X
X
X
X
Hicrobioloigcal Dye
Quality
Assurance
Special
Virus
* not run on all samples
f various analyses/, n'juber of runs indicated
^pooled collection fluid from all samplers
-------
E. Participating Organizations and Principal Personnel
The research effort documented in this report has been conducted by Southwest Research Institute with
the significant support and contributions of a number of other organizations. The following is a listing of
participating organizations:
1. Southwest Research Institute (SwRI)
San Antonio and Houston, Texas
2. The University of Texas at San Antonio (UTSA)
3. The Pacific Environmental Laboratory (PEL)
San Francisco, California
4. Naval Biosciences Laboratory (NBL)
Oakland, California
5. Environmental Quality Analysts (EQA)
San Francisco, California
6. Dugway Proving Ground
Dugway, Utah
7. Manpower, Incorporated
Hayward, California
8. Sunol Sewage Treatment Plant
Pleasanton, California
9. H. E. Cramer, Incorporated
Salt Lake City, Utah
The study site is located at the Sunol Plant, listed above, and a detailed description of the site and the
principal personnel supporting this study at the site is presented in the following section of this report. Details
regarding area of responsibility and principal participating personnel for organizations other than the Sunol
Plant are presented in Table III.E-1. The organization directing the study is Southwest Research Institute and
all other organizations listed are acting in support of SwRI in this study. All personnel listed are permanent
professional staff at the various institutions, with the exception of those listed for Manpower, Incorporated.
The personnel listed for Manpower, Incorporated are temporary employees hired for the field surveys. The
field survey team, though temporary staff rather than permanent professional staff, exhibit the same high
degree of professionalism as the permanent staff.
15
-------
Table III.E-1.
PARTICIPATING ORGANIZATIONS AND PRINCIPAL PERSONNEL
Organization
SwRI
UTSA
PEL
NBL
EQA
Dugway
Manpower,
Inc.
Responsibility
of Organization
Project direction and final
results
Analysis of coliphage
pathogenic bacteria
Chemical analysis, indicator
microbial parameters
Consultant organization
Alternate laboratory used in
quality assurance study
Meteorological support
Temporary employment
firm used to hire field
survey team
Responsibility
Participants
D. E. Johnson, Ph.D.
J. W. Register, Jr.
D. E. Camann
R. E. Thomas
R. J. Prevost
J. L. Gulinson
J. M. Taylor
H.J.Harding
J. Salinas
J. Trevino
J. Paulk
B. P. Sagik, Ph.D.
C. A. Sorber, Ph.D.
M. N. Guentzel, Ph.D.
B. E. Moore
T. Nakamura
N. Harper
M. A. Chatigny
H. Wolochow, Ph.D.
J.Tyler
Primary Responsibility
of Participant
Project director
Direction of field activities
Statistician
Statistician
Subcontracting and reports
Meteorology
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Direction of analysis
Direction of analysis
Analytical methods
Virology
Supervising chemist
Bacteriologist
Consultants in aerobiology
Consultants in aerobiology
Laboratory director
H. E. Cramer,
Inc.
Aerosol modeling and
additional statistical
support
C. Spendlove, Ph.D.
E. Rengers
J. Scudiri
T. Rooney
P. Anderson
J. DeNicola
G. Langlois
D. Lewis
R. Purdie
R. Stover
B. Pruett
D. Gaines
J. Graham
N. Houlding
C. March
G. Murdock
E. Sternstein
B. McLeod
M. Sturgis
L. Harrison
R. Menzimer
D. Wheaton
J. Synder
M. Mitchell
M. Krause
K. Dumbauld
A. Anderson
Aerosol sampling
Meteorology
Meteorology
Chemical technician
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Air sampling
Equipment maintenance
and air sampling
Direction of aerosol
modeling
Aerosol modeling
16
-------
IV. STUDY SITE
A. Site Description
1. Sunol Sewage Treatment Plant
As documented in the Phase I report, a water reclamation plant in the City of Pleasanton, Califor-
nia, was selected as the study site: the City of Pleasanton Sunol Sewage Treatment Plant.
A schematic of the study area is shown in Figure IV. A-l. A population with middle-class socioe-
conomic characteristics is located within one mile to the east/southeast of the plant. This population is lo-
cated in a recently completed subdivision off Mission Drive. Mission Drive runs east-west, and the street be-
gins on Sunol opposite the treatment plant. The prevailing winds in this area are from the southwest to
northwest quadrant; thus, this inhabited area would be downwind of the spray fields. There is a population in
this subdivision to conduct an epidemiological study, and there are also suitable control populations in Pleas-
anton with middle-class socioeconomic characteristics located more than 2000 meters from the spray fields.
The people in the subdivision near the treatment plant live from 600 to 1000 meters southeast of the edge of
the spray fields.
The Pleasanton water reclamation plant is under the control of the city government of Pleasanton,
California. City personnel involved in the study are as follows:
City Manager Clayton E. Brown
Assistant City Manager Alan B. Campbell
Director of Public Works Field Services H. Arnold Eaton
Public Works Field Superintendent Arthur N. Monaco
Plant Foreman John Wayneberg
Treatment Plant Operator George Oxsen
Laboratory Analysis Jerry Taylor
Arthur Monaco is directly responsible for the operation of the water reclamation plant, including
the spray irrigation system. The city had no immediate (within one year) plans to change the operation of this
treatment plant with respect to the spray irrigation of wastewater. The plant was modified just prior to the
Pre-Fair study by the addition of an activated biofilter process (ABF) following the trickling filter to enhance
the treatment system's biochemical oxygen demand (BOD) removal efficiency. The spray operation is nearly a
"break-even" cost operation for the city because of the income it derives from the leasing of the irrigated
pastures to cattlemen. The beef cattle which graze on the grass (44% alta tall fescue, 33% Ariki perennial rye,
and 23% Potomac Grass Orchard) appear to grow well without supplemental food. The grazing of dairy
cattle on the spray fields is prohibited. Beef cattle must wait seven days before grazing on fields that have been
spray irrigated. The cattle are moved ahead of the sprayers and the normal rotation of the spayers through the
fields ensures that a sufficient period of time has elapsed.
An average of 1.4 million gallons per day (MOD) of sewage is treated by trickling filtration and is
stored in aeration ponds with a total retention volume of three million gallons. Approximately 600 gal./min.
is recycled during irrigation from pond number 2 outlet to pond number 1 inlet to promote further oxidation.
Pumping into the irrigation system from pond number 2 begins daily between 8 and 9 A.M. - and continues
for a period of 16 to 18 hours depending on the early morning level of the pond, anticipated inflow, and pre-
cipitation. For optimum operation, the pond level is kept between 2.4 and 3.0 ft., with the most desirable level
being 2.6 to 2.7 ft. The objective is to spray daily until about one-half of the wastewater present in the two
ponds has been sprayed.
There are four major industrial waste sources in the area:
• Cheese Factory: The cheese factory waste probably has the greatest effect on the overall BOD input to the
plant. Data obtained by the Kennedy Engineers, Incorporated for the City of Pleasanton indicate that the
17
-------
Figure IV.A-1.
SCHEMATIC OF STUDY SITE
<, ^ i
<<;>: \ i
*•"» S*W\ "OKI- ' i
OCATE STREET
ADRANT
18
-------
BOD level discharge waste from the cheese factory is approximately ten times that of normal domestic
sewage. The cheese factory discharges approximately 0.016 MOD.
• Research Center: There is sizeable input (0.12 MOD) from the Kaiser Research Center, but available test
data indicate it to be a waste of normal strength.
• Winery: The Villa Armondo Winery contributes the majority of its discharge (0.01 MOD) during the
crushing season (fall and early winter). Data on this discharge have not been evaluated as of this date.
• Fair: The Alameda County Fair contributes a major portion of the industrial flow (0.25 MOD) during its
month-long operation each summer. Available data indicate the possibility of high strength wastes.
2. Treatment Plant Process
The City of Pleasanton Sunol Sewage Treatment Plant (STP) utilizes physical and biological pro-
cesses in the treatment of its sewage flow. This STP is unique in that it combines two biological waste treat-
ment systems, fixed film and fluidized culture. The fixed film is conveniently termed "trickling filter" and the
fluidized culture is termed "activated sludge". Additionally, the STP has aerated ponds which serve as pol-
ishing and equalization for the land application phase. Provisions have been made for odor control, such as
lime addition, partial chlorination, and off-gas ozonation.
The biological treatment system, called contact biofiltration (CBF), appears to be a modification
of a system described by Owen and Slechta in Water and Sewage Works, November 1975. The Sunol STP
utilizes a variation of the combined trickling filter and activated sludge systems by using the system of contact
stabilization. Figures IV. A-2 through IV. A-5 present the general plant layout flow scheme for liquid wastes
and solids handling facilities.
The flow scheme proceeds as follows. After the raw influent passes through the chemical addition,
pre-aeration, and primary sedimentation basins, the wastewater is combined with biological solids returned
from the sludge aeration unit and biofiltration unit (trickling filter) recycled in the bio-filter sump to form a
mixed liquor (similar to contact stabilization). This mixture (mixed liquor) is pumped from the bio-filter sump
to the trickling filter, where it is distributed over a horizontal redwood slat filter media, reportedly creating an
attached microorganism growth. The trickling filter underflow (effluent) proceeds to a second bio-filter sump
where provisions exist for recycle of some portion of the flow back to the first bio-filter sump. The liquid that
is not recycled proceeds to the secondary sedimentation basin where the biological solids are separated from
treated wastewater. The biological solids are removed from the secondary sedimentation basins and some
portions are wasted to the sludge digesters. The remaining sludge (biological solids) proceeds to the bio-unit
sludge aeration tank where air is blown through the mixture. This stabilized sludge then proceeds to the first
bio-unit sump, where it is mixed with fresh incoming clarified wastewater to undergo treatment in the trick-
ling filter.
The liquid effluent from the secondary sedimentation basin is discharged to two aerated ponds in
series. Water is pumped from the second pond to the spray irrigation fields.
3. Sunol STP Spray Effluent Quality
In general, the effluent quality as measured by Total Organic Carbon (TOC) and BOD values ap-
peared to show significant seasonal variability. The average BOD and TOC concentrations during August-
September 1975 were 35 and 48 mg/1, respectively, while during May-June 1976 the BOD average was 19 mg/1
and the TOC was 33 mg/1. From October to December, 1976, the BOD values indicated effluent quality simi-
lar to the 1975 data, with monthly averages from 35 to 50 mg/1. After the first of the year, however, the ef-
fluent quality again improved, giving average BOD's from 19 to 24 mg/1.
The effluent quality as measured by organic parameters [TOC, COD (Chemical Oxygen Demand),
BOD] and suspended solids appeared to improve during the Pre-Fair sampling period. The values showed
slight declines during the first three weeks, then a marked decline from the third to the fourth week, and re-
mained at this lowered level until the end of the sampling period.
19
-------
K)
O
UOIND
WASTEWATER SAMPLING STATION
ODOR SAMPLING STATION
PRIMARY SEDIMENTATION TANKS
ACTIVATED CARBON ODOR
'CONTROL UNIT
PRIMARY
DIGESTER
SECONDARY
DIGESTER
ABANDONED
SLUDGE THICKENER
BIO-FILTER SUMP
OZONE GENERATOR
ODOR CONTROL UNIT _..
ACTIVATED CARBON
ODOH CONTROL UNIT
PRE-AERATION CHAMBER
SECONDARY
SEDIMENTATION
TANK
BIO-FILTER
(DOME)
BIO-UNIT
SLUDGE
AERATION
TANK
BIO-UNIT EXHAUST
GASES CONVEYED
TO FILTER DOME
FOR TREATMENT
LIME STORAGE^ FEED BLOG
LIME ADDITION FOR
pH ADJUSTMENT AND
ODOR CONTROL
AW INFLUENT SEWAGE
SEWAGE
1 ^CHLORINE
I ADDITION
FOR ODOR
CONTROL
INFLUENT
PUMP STATION
J®.
SUNOL BLVO
Figure IV .A-2.
PLANT LAYOUT AT PLEASANTON, CA
-------
LIME ADDITON FOR
pH ADJUSTMENT
STABILIZED SECONDARY
SLUDGE RETURN -
RAW
SEWAGE
i
TO FLOW
EQUALIZA
PONDS
V-fr-i
0 * ^ •*
Y INFLUENT
... . I ...... WASTEWATE
1 I COMMlNUT'lb
INFLUENT
PUMP STATION
SECONDARY SI
RETURN
t 1
|t i~J-
SECONDARY SEDIMENTATION
TION
..
I
R PRE-AERATION
N "*"V
PRIMARY SLUDGE -*
TO DIGESTER
| SLL
.UOGE — H
1
1
J
| ' f t V
ri' ' ^
- — J PRIMARY SEDIMENTATION j
1 1
*
JDGE AERATION
V
*.
1
AERATION UNI1
SUMP
__ ** f FILTER SUMP
WASTE SLUDGE I
TO DIGESTER j
^ J— FILTER RECYCLE
— , ^ — ^i ^ ^ , . REDWOOD SLAT
: cr: — TT^-T- FILTER MEDIA
Y
TRICKLING FILTRATION
Figure IV.A-3.
CONTACT BIO-FILTER PROCESS (CBF)
-------
DIGESTER
COMMINUTION ft PflE-AERATI
ELUENT { ¥
MIY
WENT
1
9N
PRIMARY SEDIMENTATION i
PRIMARY
SLUDGE —
SOLIDS
L
\
*
i
J
' i
!
J
j ,
' (T
I
i
*>_ PRIMARY
SKIMMINGS
f~*l
1 '
1
u.
t
BIO-FILT
j
PL____WASTE GAS
U BURNER
1
RATION
1
V ^
SL
I
'
q
T
*
>-
1
BIO-UNIT
L/DGE AERATION
SECO
^ WASTE DIGESTER GAS
*—
•*—
NDARY StDIM
j
— - *-
\
7>— '
T^« —
1
• m*m
t
•••
"^EFFLUENT DISCHARGE
TO OXIDATION PONDS
SECONDARY
DIGESTER
PRIMARY
DIGESTER
-*•
DIGESTED SLUDGE DRYING BEDS
•€L
OO
OIGESTEO SLUDGE TANKER
DRY SLUDGE USED
AS A SOIL AMENDMENT
WASTE ACTIVATED
SLUOGi
•SLUDGE HEATER
Figure IV.A-4.
SOLIDS HANDLING FACILITIES
-------
STABILIZED SECONDARY
SEWAGE
K>
LIME ADDITION FOR
HH ADJUSTMENT
tr^
INFLUENT
£-
WASTEWATER FflE-AERATION
COMMINUTION
INFLUENT
PUMP STATION
r&'
SECONDARY _
SLUDGE RETURN
t-^L-J
i^ ^ i
SECONDARY SEDIMENTATION
BIO-UNIT
SLUDGE AERATION
BIO-FILTRATION
RECLAIMED WASTEWATER
IRRIGATION PUMPING
FLOW EQUALIZATION AND OXIDATION PONDS
RUNOFF WATER RECYCLED
INTO IRRIGATION SYSTEM
»
r\
PASTURE IRRIGATION WATER
RUNOFF CONTAINMENT RESERVOIR
HIGH PRESSURE SPRAY IRRIGATION SYSTEM
*.
IRRIGATED CATTLE FORAGE PASTURE LAND
Figure IV.A-5.
PLANT AND DISPOSAL FACILITY FLOW SCHEMATIC
-------
The weekly average values for BOD (during Pre-Fair) met the standard for secondary treatment as
defined by EPA (30mg/l BOD), but exceeded the standard in the October through December, 1976 period.
After this time, the levels decreased to be consistent with EPA guidelines. During the period of the fourth
week until the seventh week of the Pre-Fair, the weekly average values for total suspended solids (TSS) also
met the standard for secondary treatment of 30 mg/I TSS.
During the Pre-Fair sampling period, one sampling date, May 27th, coincided with sampling of
the aerated pond effluent by plant operators. A comparison is made below:
Sunol STP
SwRI Data Plant Data
Total Solids (mg/1) 618 631
Total Suspended Solids (mg/1) 12 15
BOD (mg/1) 11 (May 28) 10
These values for BOD and TSS represent reasonably good quality effluent.
The Sunol STP plant operators sampled the effluent from the secondary sedimentation tanks. This
sampling point does not represent overall plant effluent but represents influent to an additional treatment step
-aerated ponds. The values for May, June, and July are presented in Appendix A, along with October
through April analyses. A summary of these results is presented in the following table.
TABLE IV. A-l. SUMMARY OF SUNOL SECONDARY
EFFLUENT PARAMETER ANALYSIS
MONTHLY AVERAGE
Month/Year BOD5 (mg/1) TSS (mg/1) COD (mg/1)
15 93
40
42
24
24
36
100
15
18
15
Pre-Fair
May, 1976
June
July
Post-Fair
October
November
December
January, 1977
February
March
April
27
29
26
34
>34*
48
42
27
21
22
*one value reported as 54 mg/1.
These values suggest that some removal of organic material (suspended or dissolved) is taking place in the
aerated ponds.
A listing of operational modifications by date is presented in Appendix B. These operational
changes were primarily made as controls of the suspended solids in the mixed liquor. The addition of digester
solids was made to increase the mixed liquor suspended solids (MLSS) concentration in order to maintain ap-
proximately 3,000 mg/1 MLSS applied to the bio-filter. The wastage of the return activated solids (RAS) to
the primary sedimentation tanks was made for the same purpose.
24
-------
B. Spray Irrigation Operations
The spray equipment is designed so that a volume of water equal to the amount of incoming raw sewage,
less evaporation losses, is sprayed onto cattle pastures daily. This is attained by keeping the level in the final
effluent pond No. 2 (see Figure IV. B-l) receiving the secondary effluent as constant as possible by pumping
more or less water to the irrigation fields. This is accomplished by adjusting the daily irrigation time from 12
to 18 hours. Two pumps, 100- and 75-HP, respectively, with a third standby pump, are used to move the
water through 10-in. mains to the fields.
Table IV. B-l lists the daily incoming raw sewage volumes versus the volume sprayed for the months of
May and June, 1976. The volume of wastewater sprayed daily is estimated from the water levels in ponds 1
and 2.
TABLE IV. B-l PLANT FLOW DATA
(thousands of gallons)
MONTH
MAY
DAY INFLOW
1 1,310
2 1,293
3 1,301
4 1,398
5 1,310
6 1,298
7 1,296
8 ,377
9 ,276
10 ,261
11 ,357
12 ,331
13 ,328
14 1,310
15 1,357
16 1,291
17 1,291
18 ,400
19 ,316
20 ,403
21 ,306
22 ,365
23 ,368
24 ,394
25 ,535
26 ,391
27 ,332
28 ,357
29 ,384
30 ,256
31 ,176
TOTAL: 41,368
AVG.: 1,334
JUNE
OUTFLOW INFLOW OUTFLOW
1,873 1,371
1,981
1,878
1,410
1,398
1,676
1,657
1,756
1,741
2,137
1,911
1,028
1,410
1,357
1,291
1,791
1,600
1,016
1,103
1,206
1,465
1,568
1,494
1,335
1,491
1,232
1,457
1,884
,386
,366
,400
,183
,384
,406
,476
,480
,417
,482
,046
,400
,390
,376
,332
,408
,326
,342
,332
,331
,520
,452
,333
,382
,475
,372
,486
,466
,300
,183
,368
,476
,406
,580
,317
,582
,400
,390
,356
,476
,432
,108
,426
,392
,332
,331
,620
,252
,433
,282
,475
,172
,611
,361 1,594
1,256 1,494 2,001
1,076 1,501 1,008
1,076
46,554 41,734 42,273
1,502 1,391 1,409
25
-------
.•/*•/,^<- ' •• :/#
City of PleasantoiT -
Reclaim Water Spray Fields
Scale 1 200'
Figure IV.B-1.
SPRAY PATTERN FOR FIELDS DURING PRE-FAIR STUDY
-------
Two major parcels and one smaller parcel of land are subdivided into smaller fields for the purposes of
cattle management. The two larger areas are a 62-acre field labeled A,B,C, and D and a 100-acre field labeled
0,1,2, 3, and 4. The smaller fields across Interstate 680 are labeled E,F,G,H, and I.
The water is applied to these fields through Rainbird No. 30 sprinkler heads which have been drilled
to have 7/32- and 11/64-in. orifices. The spray heads , on 2-ft. high risers of 1-in. iron pipe, are connected to
3-in. aluminum irrigation pipes. The sprayers are located every 30 ft. along the irrigation pipe. Each spray
wets an area approximately 18 meters in diameter and reaches a maximum height of approximately 5 meters.
The transit time of the water in the pipes ranges from 4 minutes, 40 seconds to the center sprays in field 3
to 24 minutes, 40 seconds to the furthermost sprays in field D. These transit times were measured by observ-
ing the appearance of dye at the sprays after its introduction at the pumps. Transit time to the fields across the
Interstate was not measured, since no aerosol sampling was planned for that area.
The rate of application of water from individual sprayers depended primarily on their location along a
line of sprayers. Measured rates ranged from a high of 86 1/min at individual sprayers next to the main water
line (usually in the center of the line of sprayers) to a low of 30 1/min for the sprayers at the ends of sprayer
line. These rates varied depending on the field and number of sprayers used.
As an example, Table IV. B-2 presents the flow measurements made in field B at every other spray start-
ing from the northwest side of the field. These flows were determined by measuring the time required to fill a
20-1 container directly from the individual spray heads. The entire flow from both nozzles on each sprayer was
directed into a measuring container by two short lengths of tubing slip-sealed over the nozzles. The time re-
quired to fill the 20-1 container was measured using a standard stopwatch.
TABLE IV. B-2. SPRAY VOLUME—FIELD B
(Every other Spray, Second Day Setting)
Spray Head No. Liter/Minute Flow
1 35.3
3 35.8
5 36.7
7 40.5
9 39.9
11 42.9
13 43.6
15 46.2
17 50.0
19 53.6
21 51.1
23 69.8
25 74.1
Main irrigation supply line
27 78.9
29 63.5
31 32.3
33 44.1
35 53.1
37 46.0
39 45.3
41 42.1
43 39.9
45 39.2
47 31.0
49 38.7
51 42.0
27
-------
Effluent is simultaneously sprayed onto one field in each of three parcels of land (i.e., three separate
fields) every day of the year. The dashed lines in Figure IV. B-l represent the first position the spray lines
would be in for each field if irrigation was scheduled for that particular field. Note that some fields have more
than one spray line. The number of days each field is irrigated and the number of sprayers used is indicated in
Table IV.B-3. It should also be noted that some fence lines were changed between Pre- and Post-Fair sam-
pling periods.
TABLE IV.B-3. CHARACTERISTICS OF PRE-FAIR SPRAY FIELDS
No. of Days No. of Spray
Field No. or Letter Irrigated Heads
0 7 70-90
1 6 66-69
2 6 68
3 4 59-69
4 5 70
A 7 35-51
B 7 51-53
C 7 52
D 7 52
E 6 13-17
F 7 17
G 8 18
H 8 18-19
I 7 19
The number of days irrigated and the number of sprayers used depends on the size, shape, and contour
of the field being irrigated. The normal sequence of irrigation begins with fields 4, A, and E, with the spray
lines being moved 60 ft. daily in the direction of the arrows as shown in Figure IV. B-l. Once any field (4, A,
or E) has been irrigated, the spray lines are moved to the first position of the next field to be irrigated (i.e., 3,
B, or F). This sequence is continued daily, always keeping the cattle in the dry field just ahead of the irriga-
tion.
Runoff from the spray fields drains through steel pipe to a holding pond (Lake Monaco) located south-
west of Interstate 680 (Figure IV. B-l). This pond, as well as West Lake Field, is used for additional holding
capacity. This additional capacity is needed during rainy periods and as a reservoir for holding of effluent if
problems occur with the spray operations. The plant may also irrigate from Lake Monaco when necessary.
C. On-Site Facilities
On-site facilities consisted of a large storage area and a mobile office trailer with laboratory space. The
storage building was used as an assembly area for the maintenance and storage of all portable field equip-
ment. The laboratory space in the mobile office trailer was utilized for sample preparation and on-site analy-
sis. Various equipment in the laboratory essential to the operation of the Pre-Fair and Post-Fair studies are in
Appendix D.
D. General Meteorological Conditions
1. Pre-Fair
May and June, 1976, were, from a meteorological standpoint, abnormally warm and dry in the
general area of the sampling site and, indeed, in most of California. The period was marked by a persistent
surface high pressure area centered in the Pacific Ocean well off the West Coast of the United States. This
high pressure extended over California.
28
-------
Aloft at 500 mb (about 18,000 ft), a low pressure area persisted over or in the vicinity of the Gulf
of Alaska. Occasionally, this low pressure area would elongate generally north-south in the Pacific Coast
area.
These synoptic features resulted in weak maritime cold fronts moving southeastward from the vi-
cinity of the Gulf of Alaska and the area immediately to the south of the Gulf down over the Pacific Coast at
intervals ranging from every one to three days. The weaker cool outbreaks tended to move primarily across
the northern tier of states, with their western extremities moving through or close to the sampling site with an
east-west orientation. Such fronts tended to dissipate in the general area of the sampling site. The stronger
cool outbreaks tended to continue moving southeastward across the Rockies to the central United States. The
accompanying western extension of the frontal zones moved through the sampling site with a northeast-
southwest orientation. However, as the cool push continued into Utah, Arizona, and New Mexico, the trail-
ing or western extremity of the front tended to pass through the sampling site and then stagnate with a north-
south orientation paralleling the coastal hill ranges. This situation resulted in a temporary trough of low pres-
sure lying near the sampling site, or moving eastward across the site and then regressing to the west.
Precipitation with the frontal passages was mainly confined to the states north of California. Very
seldom during May and June did precipitation reach California and, when it did, it was generally confined to
extreme coastal northern California.
The windrose pattern (percentage of time wind blows from a certain direction at several ranges of
velocity) measured for the period of May and June, 1976 is presented in Figure IV. D-l. This pattern is signifi-
cantly different than that measured for August and September, 1975. Calm winds were prevalent 20% of the
time during the later period versus 70% of the time during August and September, 1975.
2. Post-Fair
The period of December 1976 through January 1977 was marked by a persistent surface high pres-
sure area centered over the Great Basin to the east of Pleasanton. The 500-mb flow was predominantly from
the northwest influenced by a quasi-permanent ridge just off the west coast. The result was a generally weak
surface wind flow at the site from the east quadrant.
The period of February through April, 1976, for the most part, came under the influence of a sur-
face high pressure area centered over the Pacific Ocean well off the west coast of the United States. This pres-
sure system extended its influence over California and brought surface winds to the Pleasanton area from the
northwest and north quadrants. The flow at 500 mb was also generally from the northwest, with ridging per-
sisting over the Pacific Ocean. There were brief interludes of frontal passages during this period which were
preceded by surface winds from the southwest in the Pleasanton area.
A set of synoptic weather maps is presented in Appendix C, in order to better depict the various
synoptic regimes which occurred during each of the sampling trials conducted during the period of December
1976 through April 1977. The surface winds observed during this period were characterized by three distinct
synoptic regimes as follows:
Synoptic Regime No. 1
A high pressure cell is centered over the Great Basin to the east or northeast of
Pleasanton. The mean surface wind direction observed during the trial periods was
from the east quadrant.
Synoptic Regime No. 2
A cold front approaches the Pleasanton area from the west or northwest. The
mean surface wind direction observed was from the south quadrant.
29
-------
Figure IV.D-1.
PRECENT DISTRIBUTION OF WIND DIRECTIONS AND WIND SPEEDS
330'
300
210'
(CALM) < 3 ,MPH
3-5 MPH
5-10MPH
10-15 MPH
15-20 MPH
30°
150°
60°
120°
LOCATION:
PERIOD:
NO. of OBSERVATIONS: 863
FREQUENCY. 1 hour
Pleascinton, Calif.
Wastewater Treatment Plant
May 1-June 16, 1976
30
-------
Synoptic Regime No. 3
A post-frontal high pressure system lies to the west or northwest of Pleasanton.
The mean surface wind direction observed during the trial periods was from the north-
west and north quadrants.
During the period of testing, the surface winds over the site were generally light. Under this situa-
tion, the wind direction was quite variable because of the influence of local topography. Pleasanton is situ-
ated in a valley bounded on the near west by a low range of hills. Another range of hills lies to the southeast
and is oriented northeast-southwest. The south opening of the valley leads to San Jose, and the terrain to the
north slopes gradually upward. Thus, when the surface winds are light, there is an ill-defined flow generally
from a northwesterly direction or its reciprocal broken by a sporadic flow from the northeast or southwest. A
well-organized flow from the southeast occurs when a front is approaching and becomes northwesterly after
the frontal passage.
The orientation of the spray line (northwest-southeast) is such that when the flow is well organized
it parallels the spray line. Conversely, under light wind conditions the surface wind direction is variable, and
the flow perpendicular to the spray line occurs only intermittently.
31
-------
V. Methods and Materials
Note to Reader: Complete documentation of the methods and materials is necessary to establish the credibil-
ity of this study's results and to serve as a guide to future investigators. Inclusion of this lengthy section in the
body of the report, however, hampered readability and was not necessary to permit reader comprehension of
the study's findings. Thus, the complete Materials and Methods section has been included as Appendix D.
The following synopsis of that section is provided to indicate what procedures, measurements and evaluations
were performed and their reason for inclusion in this research effort. Appendix D follows the same outline
format as this section to aid the reader requiring greater detail.
A. Sample Collection and Handling Methods
1. Meteorological Measurements and Instrumentations
On-site meteorological measurements were required to: 1) permit appropriate placement of the
aerosol sampling equipment and (2) provide necessary input to the microbiological dispersion model. Mea-
surements made during the study period included: wind velocity and direction, temperature, relative humid-
ity, precipitation, and solar radiation. In addition, estimates of the atmospheric stability category during each
aerosol run were made using meteorological data and observations to provide additional input to the microbi-
ologial dispersion model. A site map giving meteorological measurement locations is presented in Figure V.
A-l.
2. Wastewater Sampling Methods
A wastewater sampling and analysis program was conducted to document the quality of effluent
during the study period and explore the possible relationships between effluent quality and aerosol concentra-
tions.
a. Daily Composite Samples
During Pre-Fair studies only, daily composite effluent samples were collected from the wast-
ewater being pumped from pond 2 to the spray fields. These samples underwent analysis for pH, chlorine
(free and total), total organic carbon, solids (suspended and total), nitrogen series (nitrate, nitrite, ammonia,
and organic nitrogen), phosphorus, BOD, COD, hardness and microbiological analyses.
b. Grab Samples
Three types of wastewater grab samples were collected during the study period.
1. Pond-chlorine. A single 500 mL grab sample was collected daily near the pond 2
pumping station for analyses of free and total chlorine (Pre-Fair only).
2. Pond-pathogen screen. A number of 20-liter large-volume grab (LVG) samples were
collected from pond 2, in the vicinity of the spray field pumps, for detailed microbiological
characterization. This permitted identification of prevalent pathogens in the effluent as can-
didates for later more routine analyses.
3. Spray line. To permit direct comparisons between effluent quality and aerosol con-
centration levels, wastewater samples were collected at the spray line during each aerosol
sampling run.
3. Aerosol Sampling Methods
a. High-volume Samplers For Microorganism Aerosols
A total of 16-18 high-volume aerosol samplers were available during the study and usually
32
-------
LEGEND
[^Meteorological Instrument Station
(medsurement heights!
[T] Meteorological Tower (4 m, 8 m, 16 m, 32 m)
[2] Effluent Pond (2m, 10m)
[3] Sprdy Field EdQrf (3 ml
(Tj Just Upwind (2 m)
[jf] Just Downwind (2 mi
MEASUREMENTS
D Wind Direction
H Relatvvp Humtdtty
R Solji Radution
T Air Temperature
V Wind V.-iocitv
Figure V.A-1.
SITE MAP WITH ENVIRONMENTAL MEASUREMENT LOCATIONS
33
-------
about eight were used simultaneously in most microbiological sampling efforts. Several were LEAP samplers,
Model 3440, manufactured by Environmental Research Corporation, St. Paul, Minnesota. Most samplers
were the Litton Model M Large-Volume Air Sampler, manufactured by Applied Science Division, Litton Sys-
tems, Inc., Minneapolis, Minnesota. Both samplers are designed to collect airborne particles by electrostatic
attraction to a rotating disk on which they are concentrated into a thin, moving film of liquid. The nominal
sampling rate of both instruments is 1000 liters per minute. Nominal sampling time per aerosol run was 30
minutes.
Appendix D provides extensive detail on description of samplers, selection of the liquid col-
lection medium, sampler calibration, sampler disinfection procedures and sample handling.
b. All Glass Impinger (AGI) Samplers For Dye Aerosols
A 20-percent solution of Rhodamine WT dye was injected into the spray line and resultant
aerosols sampled using 50-mL graduated all-glass impingers manufactured by Scientific Glass and Instru-
ments, Houston, Texas. The purpose of these dye aerosol runs, conducted during a variety of meteorological
conditions, was to determine the percent of the sprayed wastewater that leaves the site as aerosol (aerosoliza-
tion efficiency).
c. Rotorod Samplers For Fluorescent Particle (FP) Tracer
During Post-Fair microbiological aerosol sampling runs, a fluorescent particle (FP) aerosol
was generated near the spray line and sampled downwind, as a quasi-quantitative tracer, to document wind
(and, consequently, aerosol plume) direction. A total of 40 rotorod samplers (Barber Colman type BYQM
2020) were available for sampling FP aerosols, and about 20 were used during each run.
B. Analytical Methods
1. Chemical Analysis of Wastewater Samples
Chemical analyses of wastewater samples fell under two categories, "routine" and "selective",
based on the frequency these analyses were conducted during both Pre-Fair study periods. Tables V. B-l and
V. B-2 identify the parameters measured and reference the method used during both study periods.
In order to justify the shipment of TOC water samples to SwRI laboratories in San Antonio for
subsequent analysis (maximum 14 days holding time) a holding time study was performed.
Fluorescent particles (FP), collected on rotorod samples during Post-Fair aerosol sampling, were
analyzed microscopically. The rotorod FP samples were analyzed by comparison to six standard samples pro-
vided by Dugway Proving Grounds.
2. Microbiological Analyses
a. Wastewater
Most microbiological analyses of wastewater samples were conducted on a routine basis.
Table V. B-3 lists the microorganisms that were routinely assayed from daily composite of spray line grab
samples during the Pre-Fair and Post-Fair study periods. The enumeration of these microorganisms in waste-
water samples provided a basis for evaluating their survival when aerosolized.
Most large-volume grab wastewater samples, collected from pond 2 primarily during Pre-
Fair, underwent analysis for the 20 microbial types listed in Table V. B-4. This pathogen screen was intended
to identify prevalent microorganisms in the effluent as improved candidates for routine analysis during later
study periods. One large-volume sample was frozen prior to shipping and analyzed for respiratory viruses (ad-
enoviruses, mumps virus, REO viruses, herpesviruses, cytomegalovirus, and measles virus).
b. Aerosols
Aerosol samples were submitted for analysis of the same microorganisms identified for wast-
34
-------
Table V.B-1
ANALYTICAL METHODS FOR PRE-FAIR WASTEWATER CHEMICAL ANALYSES
Analysis
Routine Parameters
pH
Free chlorine
Total chlorine
Total Suspended Solids
(TSS)
Total Solids
Total Organic Carbon
(TOC)
Selective Parameters
Chemical Oxygen Demand
(COD-total)
Biochemical Oxygen Demand
(BOD)
Total Phosphorus
Nitrate Nitrogen (NO3—N)
Nitrate Nitrogen (NO2—N)
Ammonia Nitrogen (HH3—N)
Organic Nitrogen
(Organic N)
Analytical
Laboratory*
Method
Ref+ Page
SwRI-PL 1
SwRI-PL 3
SwRI-PL
PEL
PEL
SwRI-SA
SwRI-SA
PEL
SwRI-SA
SwRI-SA
PEL
SwRI-SA
SwRI-SA
SwRI-PL
1
1
276
129
268
270
237
495
489
523
461
215
222,244
244
179
Detection Limit
mg/L
NA
0.05
0.1
10
10
1
1
0.1
0.1
0.01
0.1
0.1
Remarks
Glass electrode and Fischer
Accumet pH meter.
FACTS method. Optimal
density determined by Hach
DR-EL/2.
DPD ferrous colormetric
method. Optical density
determined by Hach DR-EL/2.
Gooch crucible with glass fiber
filtration and dry ness at 103 to
105°C.
Evaporation at 103 to 105°C.
Analysis performed with a
modified Beckman Carbon
Analyzer after acidification and
N2 stripping.
Dichromate reflux followed by
ferrous ammonium sulfate
titration.
5-day incubation at 20°C.
Weston and Stack D.O. probe.
Persulfate digestion, stannous
chloride color development.
Brucine method.
Diazo dyemethod.
Distillation followed by
Nesslerization.
Determined on residue from
ammonia distillation. Digestion,
distillation and Nesslerization.
EDTA Titration.
Total Hardness
(CaCOj)
* Analytical laboratories were:
SwRI-SA - Southwest Research Institute, San Antonio laboratories
SwRI-PL - Southwest Research Institute, Pleasanton, California facility
PEL - Pacific Environmental Laboratory, San Francisco, California
+ References and footnotes for analytical tests were:
(1) AWWA, APHA, WPCF, Standard Methods for the Examination of Water and Wastewater, Thirteenth edition,
American Public Health Association, Washington, D.C. 1971.
(2) Methods for chemical Analysis of Water and Wastes, U.S. Environmental Protection Agency, Washington, D.C.
1974.
(3) Cooper, W. J., Sorber, C.A. and Meier, E.P., "A Rapid Specific Free Available Chlorine Test with Syringaldazine
(FACTS), Journal of American Water Works Association, 67(1), 34-39, January 1975.
35
-------
Table V.B-2
ANALYTICAL METHODS FOR POST-FAIR WASTEWATER CHEMICAL ANALYSES
Analysis
Routine Parameters
PH
Total Organic Carbon
(TOC)
Total Suspended Solids
(TSS)
Conductivity
Selective Parameters
Chemical Oxygen Demand
(COD)
Nitrate Nitrogen
(NO,—N)
Ammonia Nitrogen
(NHj-N)
Organic Nitrogen
Nitrate Nitrogen
(NO2—N)
Total Phosphorus
(Total) P
Total Solids
Biochemical Oxygen
Demand (BOD5)
Analytical
Laboratory*
SwRl-PL
UTSA
SwRl-HOU
UTSA
SwRl-PL
PEL
PEL
PEL
PEL
PEL
PEL
PEL
PEL
Method
Ref+ Page
460
237
Detection Limit
mg/L
NA
94
71
550
427
417
437
215
473
91
543
10
5 mho/cm
0.1
0.1
0.1
0.01
0.1
10
1
Remarks
Glass electrode and Fischer
Accumet pH meter.
Samples acidified upon
collection. N2 stripped before
analysis. Beckman Model 915A
with Model 865 Infrared
Analyzer.
Modified Beckman Model315
with Model 865 Infrared
Analyzer.
Glass fiber filtration and
drynessat 103 lo 105°C.
Hach DR-EL/2 conductivity
probe.
Dichromate reflux followed by
ferrous ammonium sulfate
titration.
Brucine method
Distillation followed by titration
with standard acid.
Determined on residue from
ammonia distillation. Digestion,
distillation and titration with
standard acid.
Diazo dye method.
Persulfate digestion, stannous
ehloride color development.
Evaporation at 103 to 105"C.
5-day incubation at 20°C.
Weston & Stack D.O. Probe.
* Analytical laboratories were:
SwR[-SA - Southwest Research Institute, San Antonio laboratories
SwRI-PL - Southwest Research Institute, Pleasanton, California facility
PEL - Pacific Environmental Laboratory, San Francisco, California
f References and footnotes for analytical tests were:
(1) AWWA, APHA, WPCF, Standard Methods for the Examination of Water and Wastewater, Thirteenth edition,
American Public Health Association, Washington, D.C. 1971.
(2) Methods for chemical Analysis of Water and Wastes. U. S. Environmental Protection Agency, Washington, D.C.
1974.
36
-------
Table V.B-3
MICROORGANISMS ROUTINELY ASSAYED IN WASTEWATER
Microorganism Pre-Fair Post-Fair
Standard bacterial plate count D,R R
Total coliform D,R R
Fecal coliform D,R
Coliphage D,R R
Fecal streptococci D R
Pseudomonas D
Klebsiella D
Clostridium perfringens D
Mycobacteria R
Enteroviruses-3 day count D
Enteroviruses-5 day count D
D—daily composite sample
R—spray line grab sample with each aerosol run
Table V.B-4
MICROBIALTYPES SOUGHT IN PATHOGEN SCREEN
Quantitative:
Klebsiella
Pseudomonas
Clostridium perfringens
Fecal Streptococci
Semi-quantitative:
Staphylococcus aureus
Mycobacteria
Leptospira
Shigella
Salmonella (including Arizona)
Enterobacter
Serratia
Edwardsiella
Escherichia
Citrobacter
Proteus
Providencia
Yersinia
Neisseria (pathogenic)
Aeromonas and other oxidase-positive fermenters
37
-------
ewater samples (10 assays on Pre-Fair samples, 5 assays on Post-Fair samples). Comparison of the wastewater
and aerosol concentration data permitted evaluation of the survival properties of each microbial type. This
included survival during Ihe aerosolization process and, by comparion of concentration data at various down-
wind locations, long-term survival while in the aerosol state.
Several of the specific assay methods were different from those used for wastewater samples.
This was largely due to the smaller volume of aerosol collection fluid available for assay.
C. Quality Assurance
A major goal of this research was to determine if wastewater measurements could be used as a basis for
predicting the microbiological quality of aerosols from spray irrigation sites. This necessitated a definition of
the accuracy and precision of all methods used. To accomplish this a quality assurance program was estab-
lished which had two primary objectives:
(1) To determine the accuracy and precision of laboratory analyses, and
(2) To determine if there were systematic differences in high-volume aerosol sampler collection
efficiency.
Quality assurance tasks for wastewater analyses included the use of spiked reference samples, the com-
parison of replicate samples sent to more than one laboratory and the comparison of replicate samples sent to
a single laboratory. Quality assurance aerosol runs were conducted by placing all samplers at the same dis-
tance from the spray line on approximately one meter centers. Analyses of the sampler collection fluid were
performed by several laboratories in a manner that permitted relative comparison of between-samplers and
within-sampler precision.
D. Aerosol Sampling Protocols
1. Microbiological Aerosol Runs
Eight (when available) high-volume aerosol samplers were utilized during each aerosol sampling
run. One was located well upwind to provide a measure of aerosol background levels. The remaining seven
samplers were placed downwind to measure elevated aerosol levels due to operation of the spray source.
During Pre-Fair studies emphasis was placed on characterization of source strength and identifica-
tion of pathogen survival characteristics upon aerosolization. The sampling configuration designed for the 20
Pre-Fair aerosol sampling runs is shown in Figure V. D-l. Following the eighth aerosol run, however, the
five-meter station was eliminated and moved to 30 meters. This was due to the high frequency at which this
station was hit with spray during wind gusts. Late in the Pre-Fair study period, it was determined that more
duplicate samples at a given distance were required to better estimate source of variability. Thus, one of the
following configurations was used during the last five aerosol runs:
Configuration A Configuration B
No. of Samplers Distance No. of Samplers Distance
1 upwind 1 upwind
2 10 meters 2 20 meters
1 20 meters 1 30 meters
1 30 meters 1 40 meters*
1 40 meters* 2 50 meters*
2 50 meters*
* double samplers at these distances were alternated on different runs.
Emphasis during Post-Fair studies was on modeling of downwind aerosol concentrations. Thus,
38
-------
X Upwind (at a great distance)
-e e e e e-
SPRAYER LINE
-o ^-^—e e e-
i
/ Q | Wet Line Edge
5 m
f
X 10m
xlx 20m
X 1/2 maximum distance (usually 50 m)
LLJ
_l I
Q.
H
XIX maximum distance (usually 100 m)
Figure V.D-1.
AEROSOL SAMPLING CONFIGURATION FOR PRE-FAIR
39
-------
the configuration used for the 29 Post-Fair aerosol sampling runs (Figure V. D-2) required the placement of
samplers at greater distances (up to 600 meters) from the spray source.
During both Pre- and Post-Fair studies, detailed protocols were prepared to provide guidance to
field personnel. These included instructions by which the real-time on-site meteorological measurements
should be used to determine: specific sampler locations; criteria for initiating or aborting aerosol runs, and
whether collected samples should be shipped for analysis (good run) or discarded (bad run).
2. Dye Aerosol Runs
Dye aerosol runs were conducted during both Pre- and Post-Fair to determine the percent of
sprayed wastewater that left the spray field as aerosol. Thus, a close-in configuration, as shown in Figure V.
D-3 for the Post-Fair runs, was used for the eight AGI samplers used during Pre- and Post-Fair study periods
respectively. Again detailed instructions were prepared to guide field personnel on the placement of samplers
relative to wind direction and on the meteorological conditions necessary to initiate/accept as valid each run.
3. Quality Assurance Runs
Eight quality assurance runs were conducted (5 during Pre-Fair and 3 during Post-Fair studies). In
each, all functioning samplers were placed side-by-side (separated by one meter in Pre-Fair and three meters
in Post-Fair) at the same distance from the spray field. A randomized arrangement of samplers on both sides
of the center line permitted evaluation of bacterial analyses precision (both within- and between-samplers)
and a determination of systematic differences in the collection efficiency of the various high volume samplers.
4. Special Enteric Virus Aerosol Runs
During Pre-Fair studies, assays of the high-volume aerosol sampler collection fluid for enteric
virus were consistently negative. This was due to their relatively low concentration in the wastewater and the
inadequate sensitivity of the sampling procedure employed. Even at low concentrations, enteric virus aerosols
could represent a significant hazard to human health. Therefore, a special sampling protocol was designed for
Post-Fair studies to significantly improve the procedure's sensitivity and, thus, provide quantitative mea-
surements of enteric virus aerosol concentrations. These special virus aerosol runs were designed to character-
ize source strength by placing all available aerosol samplers (12 or 13 samplers) side-by-side at a 50-meter dis-
tance from the spray source. The samplers were then operated for many consecutive aerosol runs (four for the
first virus run and six for the second virus run). The 100 mL of collection fluid from each sampler, during
each run, was placed in a common container. The total volume was then transported to the laboratory where
it was concentrated and assayed to provide a single estimate of the enteric virus aerosol concentration. This
represented an approximately 50-fold increase in the sampling procedure's sensitivity. Since all samples had
to be pooled to obtain a single concentration estimate, the procedure could not provide an estimate of the
variability associated with this measurement. Two such sampling runs were conducted during Post-Fair stud-
ies.
E. Data Flow, Processing and Analysis Methods
1. Sample Identification and Labeling
To develop valid findings from this large-scale field sampling and analysis effort, it is essential that
the identity and integrity of the data be preserved. Accordingly, an integrated data system involving a unique
sample code, computer-generated sample labels, field data reporting forms, and analytical data reporting
forms was implemented. This system insured that each sample was uniquely identified (run number, sample
location, medium, analysis, sampling period) and that this identity was transmitted with the analytical values
from sample collection through processing, shipping , preparation, analysis, reporting, and statistical analy-
sis. This uniform accounting and tracking procedure also made it impossible for laboratory personnel to iden-
40
-------
50M
100M
STAKED LINE
• •
~300M
MEAN
WIND
DIRECTION
Q 15MIN
STAKED LINE
-600M
•
f
LEGEND
X HIGH VOLUME SAMPLER
0 STAKE
(§) ROTOROD SAMPLER
• FPSOURCE
•jf 2M TOWER-WIND
DIRECTION AND VELOCITY
O TEMPERATURE AND
REL. HUMIDITY
Figure V.D-2.
PRIMARY (MICROBIOLOGICAL) AEROSOL SAMPLING CONFIGURATION FOR POST-FAIR
-------
S
ffl
I
Cfl
WJ
3 9
o <*»
I
I
3
ya
s
S
-------
tify samples and, thus, introduce bias into their analyses. This was extremely important for the quality assur-
ance samples which were also submitted using this code system,
2. Data Forms and Reporting System
In conjunction with the sample labeling system, a series of forms was prepared and a reporting
procedure established to insure close coordination and control of the field and laboratory aspects of the
study. Field and laboratory personnel submitted completed forms on a routine basis to the project statisti-
cian, where they were reviewed for data reporting deficiencies and for sampling and analysis problems. This
provided a timely feedback mechanism to correct errors and to identify requirements to modify the study pro-
tocol.
3. Aerosol Data Processing
Ideally, the measured microbiological aerosol concentrations are presumed to accurately describe
the microorganism levels emanating from the source under study (e.g., the line of sprayers). However, in a
large-scale field program such as that conducted at Pleasanton, some measurements of microbiological aero-
sol concentrations are likely to reflect extraneous factors such as sampler/sample contamination or nonstudy
sources that dominate the study source effect. In addition, the high variability of quantum microbiological
measurements at levels close to the minimum detection limits of the methods used requires careful data
smoothing techniques to avoid logical inconsistencies. The extraneous factors and variability-induced logical
inconsistencies can have biasing and mathematically intractable consequences in data analyses such as the mi-
crobiological dispersion model development. Procedures developed and used to alleviate these data analysis
problem areas include:
(a) a procedure for standard data processing and event notation,
(b) a procedure for use/rejection of data based on evidence of sampler/sample contamination,
and
(c) data smoothing procedures used for application of the microbiological dispersion model.
4. Computational Techniques
The automated computational techniques which were used to analyze the extensive data generated
during this project were chosen as being the most appropriate from numerous options that were available.
SwRI currently maintains a special lease arrangement with the McDonnell Douglas (McAuto) computer fa-
cility in Huntington Beach, California for use of their CDC Cyber 70/74 equipment. This system is accessed
through a CDC remote batch terminal at the Institute's Computer Laboratory. In addition, a Hewlett Pack-
ard 9810A programmable calculator with a limited package of statistical routines was utilized when the de-
sired analysis was less involved and the quantity of data was sufficiently small to permit direct keyboard
entry. The major automated computational procedures used in this project are listed in Table V. E-l.
5. Statistical Approach
The specific objectives given in Section III. D. have been addressed by applying suitable statistical
methods to the study data to make appropriate inferences (i.e. findings). Statistical analyses relevant to many
of the objectives were performed upon obtaining the Pre-Fair data. In some cases, the Pre-Fair analysis pro-
vided adequate information and its methods and results are presented in Section VI. In other cases, a repeated
analysis encompassing both the Pre-Fair and the Post-Fair data was performed and is reported in Section VI.
The Pre-Fair data analyses suggested the propriety of developing the microbiological dispersion model. De-
velopment and evaluation of this model was the major emphasis of the Post-Fair data analysis.
43
-------
Table V.E-1
AUTOMATED COMPUTING PROCEDURES
Program
LABELS
DSTAT1
DSTAT2
CORREL
TPROB
ANOVA
PRESTO
CANCORR
CONDESCRIPTIVE
BMD08V
BIODCAY
MODEVAL
Source
SwRI
SwRI
HP
SwRI
SwRI
HP
SwRI
SPSS
SPSS
HMD
SwRI
SwRI
Type
New
Existing
Package
New
Existing
Package
Existing
Package
Package
Package
New
New
Computer
System
Cyber 74
HP9810A
HP9810A
HP9810A
HP9810A
HP9810A
Cyber 74
Cyber 74
Cyber 74
Cyber 74
Cyber 74
DG Eclipse
300
Usage
Generate sample labels.
Mean, standard deviation, and
coefficient of variation.
Calculate mean, standard devia-
tion, skewness, kurtosis; grouped
and ungrouped data.
Correlation coefficients
Percentiles of t-distribution.
Analysis of variance, one-way
and two-way without interaction.
Stepwise multiple linear regres-
sion; correlation coefficients.
Canonical correlation; correla-
tion coefficients,
Calculate mean, standard devia-
tion, skewness, kurtosis.
General analysis of variance
models.
Calculate estimates of microbio-
logical dispersion model parame-
ters I and A.
Evaluates the microbiological
dispersion model by comparing
its prediction P against the ob-
served value C-B.
44
-------
VI. RESULTS
A. Wastewater Characteristics
1. Chemical Data and Patterns
The results of water quality chemical analyses indicated that the Pleasanton wastewater was gener-
ally typical of an undisinfected, secondarily treated wastewater. Mean values for certain parameters, deter-
mined during Pre-Fair studies, were as follows: BOD - 18.7 mg/L, COD - 99.5 mg/L, TOC 33.0 mg/L, ph
8.4, hardness 235.2 mg/L, TSS - 33.0 mg/L, total phosphorus - 5.6 mg/L and nitrite, nitrate, ammonia and
organic nitrogen - 0.15 mg/L, 0.06 mg/L, 23.9 mg/L and 5.6 mg/L, respectively. Limited chemical analyses
continued during Post-Fair studies and revealed no major differences in wastewater quality.
A strong relationship was observed among TOC, COD, and BOD. The significance level of the
correlation between TOC and BOD was 0.006, and for the other pairs it was less than 0.001.
In general, no major wastewater quality differences were observed on a weekly or daily basis or
during the conduct of any individual aerosol runs. Thus, the chemical quality of the wastewater was not con-
sidered to have a significant influence over the variability of the microbiological aerosol levels measured.
These results do not imply that water quality has no effect on microorganism survival. Chemical
parameters may, in fact, have adverse (toxic) or beneficial (protection from desiccation) effects on the survival
of aerosolized microorganisms. Thus, water quality may have had an overall impact on the actual levels of
microbiological aerosols measured that could not be evaluated during this study. Since the water quality was
relatively consistent throughout the study period, it was not considered a reliable indicator/predictor of mi-
croorganism (especially pathogen) levels. Thus, the chemical quality of the wastewater was not considered a
relevant factor for inclusion in the microbiological dispersion model. A complete description of the wastewa-
ter chemical data and patterns is presented in Appendix E.
2. Microbiological Data and Patterns
a. Daily Composite Microbiological Da ta
Samples from the daily composite effluent sampler during Pre-Fair were sent to Pacific Envi-
ronmental Laboratory (PEL) for microbiological analyses of standard indicator microorganisms, including
total and fecal coliform and standard bacterial plate count, or the total aerobic bacteria plate count with stan-
dard methods medium. The concentration values obtained are shown in Table VI. A-l. Since the total col-
iform and fecal coliform analyses were performed using the membrane filter method, these values are re-
ported as MFC/100 mL (membrane filter count per 100 milliliter of effluent).
The total and fecal coliform values shown are the average of a minimum of three repetitions,
and often five or six. Standard bacterial plate count was generally performed in either duplicate or triplicate
and the average value is shown for these as well. There was only one sample which gave unusual values, sam-
ple 78 from June 12, where there was apparently some contamination of the sample. The laboratory was un-
able to complete the analysis for fecal coliform on sample 74 due to an equipment malfunction.
As can be seen from the table, the concentrations were fairly consistent during given inter-
vals within the sampling period. In addition, all three measurements exhibited similar indications of changes
in the effluent microbiological constituents.
Portions of the daily composite effluent samples were shipped to the UTSA-CART labo-
ratory for analyses for coliphage, selected pathogenic bacteria, and enteric viruses. The concentration values
for these daily effluent samples are presented in Table VI. A-2. Following the table is a list of footnotes. The
45
-------
Table VI.A-1.
DAILY COMPOSITE EFFLUENT CONCENTRATIONS OF
MICROBIOLOGICAL INDICATOR PARAMETERS
Sample
No.
Total
te Coliform
(MFC/100 mix 103)
1-2 5-2 1700
2 :
2300
3 4 2500
4 5 2200
5* 7 1350
32* i
750
33* 9 517
34* 10 832
35* 1
1 1200
36 11 913
37 12 255
98 13 817
71 14 958
1-3 15 1220
61 16 477
70 17 410
41 18 420
76 19 617
67 20 713
38 21 830
75 22 1200
64 23 1060
39 24 800
62 25 992
99 26 533
74 27 763
63 28 1450
40 29 1430
97 30 565
60 31 817
73 6-1 703
42 2 293
69 3 267
66 4 833
95 5 530
68 6 257
77 7 417
72 ,8
160
65 9 760
96 10 798
43 11 290
78J 12 2470
50 13 490
51 14 743
52 15 620
53 16 340
56 17 530
58 22 353
Fecal
Cobform
(MFC/ 100 ml X 103)
117
275
210
180
147
93.3
102.7
126
143
68
36.7
82.7
123
90
59.5
12.5
33.3
63
177
82.8
152
140
123
83.3
33.7
-t
134
220
53
76.7
57
24.3
24.3
59
35
13
59
22.3
85.7
80.4
31.3
263
70.3
93.2
79
82.3
63.3
64
Standaid Plate
Count
(No./lOOmlX 106)
130
250
100
110
76.7
77
107
91.7
157
180
57
237
120
247
110
34.3
75.7
133
115
75
393
218
85
57
101
13.5
133
90
62.7
96.7
133
102
35
107
63.5
157
11.3
36
19.3
15.7
11.7
>300
14
12.7
21.3
14.7
6.13
8.05
*Composite sampler inoperative, grab sample taken morning after spraying.
•(•Equipment
malfunction; laboratory unable to perform analysis.
^Probable contamination.
Note: 1700 (MFC/100 mlX 103)=1700X 103 MFC/100 ml = 1.7 X MFC/100 ml
46
-------
Table VI.A-2.
DAILY COMPOSITE EFFLUENT CONCENTRATIONS OF COLIPHAGE AND SELECTED
PATHOGENIC BACTERIA AND VIRUSES
Sample
Date
5-2-76
5-3-76
5-4-76
5-5-76
5-6-76
5-7-76
5-8-76
5-9-76
5-10-76
5-11-76
5-12-76
5-13-76
5-14-76
5-15-76
5-16-76
5-17-76
5-18-76
5-19-76
5-20-76
5-21-76
5-22-76
5-23-76
5-24-76
5-25-76
5-26-76
Coliphage
Count,
(PFU/1) X 103
68
370
CS
120
160
180
110
87
97
100
140
540
350
290
230
97
360
330
340
190
260
120
78
130
120
Bacteria
Klebsiella,
(CFU/100ml)X 103
25
39
58
41
28
36
28
52
55
72
66
52
55
28
Pseudomonas,
(CPU/ 100 ml) X 10'
CS
100
25
15
90
140
88
110
50
25
13
80
75
25
15
30
60
50
25
210
120
250
35
120
90
Streptococci,
(CPU/ 100 ml) X 10s
CS
>24t
CS
CS
CS
9.3f
CS
lit
lit
CS
2.3t
>24t
4.6|
23t
10
6.5
5.1
7.9
13
20
18
10.4
8.0
2.8
3.7
Clostridium
perfringens,
(MPN/100 ml) X 103
>24
11
11
>24
4.6
1.1
.46
>2.4
.24
.24
.093
11
9.3
9.3
4.6
4.6
4.6
4.6
7.5
>24
23
4.3
24
46
9.3
Viruses
3-Day
Plaques
(PFU/1)
9.9
4.6
340*
29
589*
18
25
26
<2.4
4.7
12
33
71
<1.8
20
21
19
15
39
18
9.3
1.3
CS
CS
5.3
5-Day
Plaques
(PFU/1)
TNTC
TNTC
TNTC
TNTC
TNTC
36
TNTC
28
4.7
9.8
16
31
90
15
43
43
26
24
39
18
9.3
3.3
CS
CS
7.2
High Sample
Temperature
Upon Lab
Receipt
10°C
rc
9°C
9°C
•Possible contamination.
fMPN.
-------
Table VI.A-2 (continued)
Sample
Date
5-27-76
5-28-76
5-29-76
5-30-76
5-31-76
6-1-76
6-2-76
6-3-76
6-4-76
6-5-76
6-6-76
6-7-76
6-8-76
6-9-76
6-10-76
6-11-76
6-12-76
6-13-76
6-14-76
6-15-76
6-16-76
6-17-76
Coliphage
Count,
(PFU/1) X 103
180
240
410
290
230
160
270
280
350
280
200
330
440
570
490
460
360
420
350
220
230
220
Bacteria
Klebsiella,
(CPU/ 100 ml) X 103
94
44
39
25
36
41
17
30
8.2
17
33
36
19
50
50
110
55
28
58
41
33
94
Pseudomonas,
(CPU/ 100 ml) X 103
30
1700
2900
450
1900
100
300
200
350
260
430
92
45
130
120
200
150
95
30
25
30
20
Streptococci,
(CPU/ 100 ml) X 103
6.8
9.6
14.6
2.5
10.3
5.7
1.7
2.3
6.6
4.3
4.2
3.4
1.1
5.7
12
24
3.0
6.4
17.1
9.2
5.5
4.0
PFU -Plaque forming units. MPN -Most probable number.
CPU -Colony forming units. TNTC -Too numerous to count.
Clostridium
perfringens,
(MPN/ 100 ml) X 103
4.3
4.3
4.3
2.3
7.5
9.3
9.3
4.3
2.3
2.3
2.3
4.3
15
9.3
9.3
4.3
4.3
9.3
9.3
9.3
4.3
9.3
Virus
3-Day
Plaques
(PFU/1)
15
4.7
4.7
18
<1.6
4
17
45
60
56
27
46
CS
5.9
<1 9
10
22
4.6
7.9
6.1
12
12
5-Day
Plaques
(PFU/1)
15
CS
6.8
18
8.8
66
23
50
86
87
CS
CS
CS
CS
<4.0
16
22
4.6
7.9
CS
14
14
High Sample
Temperature
Upon Lab
Receipt
CS —Contaminated sample.
M —Missing
-------
pathogenic bacteria selected for assay were Klebsiella, Pseudomonas, fecal streptococci, and Clostridium per-
fringens. The Klebsiella column of Table VI. A-2 consists of Klebsiella pneumoniae and Klebsiella ozaenae.
The Pseudomonas reported in Table VI. A-2 are fluorescent.
The streptococci column gives the fecal streptococci assayed according to Standard Methods
for the Examination of Water and Wastewater. The Klebsiella, Pseudomonas, and later streptococci assays
are reported as CPU/100 ml (colony forming units for 100 milliliters of effluent). The early streptococci va-
lues and the Clostridium perfringens values were obtained through analyses requiring most probable number
tables; these values are reported as MPN/100 ml (most probable number per 100 milliliters of effluent). Vi-
ruses in the effluent samples were plated on HeLa cell monolayers. Approximately 70% of each concentrated
sample was observed for virus plaques after three days; the remainder of each sample was observed after five
days. The 3- and 5-day virus counts in Table VI. A-2 are given in PFU/1 (plaque forming units per liter of
effluent). See Appendix D for a complete description of the microbiological analysis procedures performed
on the effluent samples.
The first daily composite sample sent for coliphage analysis was concentrated. However, its
liter was high enough (91 % efficiency) to establish that concentration was unnecessary for the coliphage anal-
yses of the remaining daily composite samples. Concentration was found to be necessary, and was conducted,
on all the effluent samples assayed for viruses. The recovery efficiency of the daily poliovirus reference sam-
ples ranged from 15 to 100%. The quartiles of the recovery efficiency distribution were 39% (first quartile),
50% (median), and 65% (third quartile). The corrected virus plaque count shown in Table VI. A-2 was ob-
tained by dividing the observed raw plaque count by the recovery efficiency for that day's reference sample.
Since quantitative effluent concentrations were obtained on nearly every sample, the data are suitable for sta-
tistical analyses.
As anticipated with such a large-scale systematic sampling and analysis protocol, there were
also a few problems. Some early bacteria and virus samples and several of the later virus samples were con-
taminated. Two or three of the early virus samples may have been contaminated with the reference poliovirus.
Four of the early sample shipments were received at temperatures considerably above the desired 4°C. The
elevated temperatures may have raised the bacterial pathogen levels and lowered the coliphage and virus levels
of these samples in comparison with the samples shipped at about 4°C. Streptococci levels were determined
by the MPN method for the first two weeks because the proper assay medium was unavailable.
Several special data symbols have been used in Table VI. A-2. When the sample assay was
negative, the result has been reported as <(less than) the detection limit. The virus and coliphage plaque-form-
ing units grow as they consume the host cells. When the individual virus plaques grew together, it was impos-
sible to determine how many plaque-forming units there were; such results were reported as TNTC (too nu-
merous to count).
It is important to note that a wide range of effluent concentration values was observed. The
coliphage, bacteria, and virus concentrations in Table VI. A-2 and the coliform and standard bacterial plate
count concentrations in Table VI. A-l all exhibit at least a one order of magnitude range of values. Many
approach a two orders of magnitude range. In fact, Pseudomonas and Clostridium perfringens vary over
more than two orders of magnitude.
b. Distributional Characteristics
Summary statistics were calculated for all wastewater microorganism concentration data.
The mean, standard deviation, skewness, and kurtosis statistics, given in Table VI. A-3, characterize the ef-
fluent sample distribution of each. The upper half of the table applies to the untransformed data. The arith-
metic standard deviations are large relative to the arithmetic means for each parameter, which implies the or-
ders of magnitude variation readily observed in Tables VI. A-l and VI. A-2. The skewness and kurtosis
49
-------
Table VI.A-3.
DISTRIBUTIONAL CHARACTERISTICS OF THE WEIGHTED DAILY AND LARGE—VOLUME
EFFLUENT SAMPLE CONCENTRATIONS OF THE INDICATOR AND PATHOGENIC
MICROBIOLOGICAL PARAMETERS
Untransformed Concentration
Tol.il Conform. (MFC/ 100 ml) X 10'
K-i-ilColilbrm, (MFC/IOOml) X 10'
Sid Plate Count, (No./lOO ml) X 10"
Coliphage (PFU/1) X 103
Klebsiella, (CFU/1 00 ml) X 103
Pseudomonas. (CFU/100 ml) X 103
Streptococci, (CFU/1 00 ml) X 103
Clostridmm perf . (MPN/100 ml) X 103
3-Day Virus Count (PFU/1)
5-Day Virus Count, (PFU/1)
Natural Log Transformed Concentrations
In (Total Coliform)
In (Fecal Coliform)
ln(Std. Plate Count)
In (Coliphage)
In (Klebsiella)
In (Pseudomonas)
In (Streptococci)
In (Clostridium pert')
In (3-Day Virus Count)
In (5-Day Virus Count)
No. of
Samples with
Numeric
Analysis Results
Effluent Sample Statistics
Mean
Standard
Deviation
Skewness
Kurtosis
P, Significance Levels of
One-Sided Tests of
Suitability
nf Normal
Distribution Model
Skewness
(H0: v£>0)
Kurtosis
(H0: 0, = 3)
(Arithmetic)
54
53
53
53
42
52
46
53
48
39
927.9
102.2
106.3
258.1
44.9
264.1
650.9
71.8
95.9
130.8
23.0
525.9
8.84 7.55
9.06 10.14
20.6
27.9
28.9
35.4
1.5
1.3
2.1
0.5
0.8
3.6
2.2
2.8
5.5
4.3
5.0
4.7
9.6
2.4
3.5
16.0
10.3
10.7
42.1
27.3
<.01
<.()!
«.01
.05
.01
«.01
«.01
«.01
«.01
«.01
<.01
.01
«.01
OK
OK
«.01
«.01
«.01
«.01
«.01
(Geometric)
54
53
53
53
42
52
46
53
48
39
748.5
79.5
69.9
223.9
38.8
104.8
1.24
1.33
1.68
1.17
1.19
2.18
6.73 1.39
5.39 2.02
12.11 1.81
17.29 1.61
0.02
-0.41
-0.57
-0.37
-0.88
0.61
-0.40
-1.15
-0.41
0.06
2.68
2.84
2.63
2.18
4.23
3.16
3.49
4.98
3.41
2.89
OK
OK
.04
OK
.01
.03
OK
«.01
OK
OK
OK
OK
OK
.05
.04
OK
OK
<.01
OK
OK
statistics test whether the distribution is normal. A skewed distribution has one tail that extends out farther
than the other tail. Kurtosis measures whether the distribution has a very sharp peak or a very broad, flat top.
The true normal distribution has a skewness parameter of 0 and a kurtosis statistic of 3. For each set of micro-
biological concentration data, tests of the null hypotheses that the data were normally distributed were con-
ducted. The significance levels of the results are presented in the two right-hand columns of Table VI. A-3.
All of the untransformed data exhibited positive skewness. Only the untransformed data for coliphage and pos-
sibly for Klebsiella had acceptable normal distribution kurtosis. This confirmed the requirement to transform
the effluent microbiological data to permit valid correlation analysis.
The summary statistics under the natural log transformation are shown in the lower half of
Table VI. A-3. To permit comparison with the untransformed data statistics, the geometric mean and the geo-
metric standard deviation are given in Table VI. A-3, rather than the In x mean and standard deviation. The
geometric mean has the same units as the arithmetic mean. The geometric mean values are lower than the
arithmetic mean because of the scale adjustment introduced by the logarithmic transformation.
The results of the skewness and kurtosis tests of the normal distribution null hypotheses are
also given in Table VI. A-3 for the natural log transformed data. For most microorganism groups, both the
skewness test and the kurtosis test indicate that the logarithmic transformed data have an acceptable normal
distribution (i.e., the untransformed data have a lognormal distribution). For each of the others the natural
50
-------
log data are more nearly normally distributed than are the untransformed data. Neither the natural log trans-
formation, the square root transformation, nor the untransformed data for Clostridium perfringens followed
a normal distribtuion. This may be due to the limited number of Clostridium perfringens values permitted in
the MPN index probability tables rather than an actual characteristic of the concentration levels. Except for
the effluent Clostridium perfringens data, each of the effluent microorganisms were generally considered to
follow a lognormal distribution.
c. Wastewater Analysis Variability
A precision study for the three indicator microorganism group analyses was conducted dur-
ing Pre-Fair in the same manner as for the chemical constituents. The replicate determinations from the daily
effluent runs were used to obtain estimates of the standard deviations associated with analytical repeatability.
The data used were transformed by the natural logarithms of the observed counts to achieve normality and
homogeneity of variance. The standard deviations were then pooled and the results exponentiated to obtain
the estimates of variance due to repeat analysis.
Ten samples were sent for analysis, five from each of two 500-ml grab samples (which were
split from a single 1 liter grab) to determine the shipping and analysis variance component. The results for
these samples and the standard deviations of the transformed data obtained are shown in Table VI. A-4. The
standard deviations for the 2 pond grab samples are pooled into a single estimate as above and exponentiated
to give an estimate of the percentage variation among the samples. The day-to-day variance estimate is taken
from the summary statistics portion of Table VI. A-3. The three variance components estimated for each mi-
croorganism group are presented in Table VI. A-5.
The total coliform data show no tendency to be affected by the shipping process, giving ap-
proximately equal values for both the repeatability and between-sample components. The fecal coliform anal-
ysis has a higher standard deviation for shipping and analysis than repeatability, but an F-test performed in
Table VI.A-4.
MICROBIOLOGICAL INDICATOR QUALITY ASSURANCE PRECISION STUDY-
ANALYTICAL RESULTS
Parameter
1
2
3
4
5
1
2
3
4
5
Total Coliform
Fecal Coliform
Standard
Plate Count
Sample (MFC/100ml x 10 ) (MFC/100ml x 10 ) (no/100 ml x 10 )
570
543
483
577
477
575
650
543
603
633
64.3
77.8
44.3
55.0
63.3
72.5
46.7
42.0
56.7
47.5
108
57
23.8
230
150
195
160
49.7
108
137
Coefficients
of variation
8%
24%
109%
51
-------
Table VI.A-5.
MICROBIOLOGICAL QUALITY ASSURANCE PRECISION STUDY—PRECISION ESTIMATES
(Percent Coefficient of Variation)
Parameter
Standard
Total Fecal Bacterial
Colif orm Coliform Plate Count
Variance Component
Repeatability 11 11 12
Shipping and Analysis 08 24 107
Day to Day 24 33 118
the transformed scale does not indicate a significant difference between the two components. The standard
bacterial plate counts are more variable. A variance component estimated from these data is 1.09, or 9-per-
cent variation from sample to sample induced by the shipping and handling, plus between-sample variability.
The precision portion of the quality assurance study on the analysis for coliphage and patho-
gens was conducted using replicate portions from two effluent grab samples. The coliphage was analyzed in
five replicates from each of the two samples and the pathogenic bacteria and viruses were analyzed in three
replicates from each sample. The data are used to estimate analytical variability, within replicates, and to de-
termine if there is a significant between-sample variance associated with these data. All analyses were per-
formed in the logarithmic scale on the basis of the distributional nature of the data.
The coliphage data are presented in Table VI. A-6, along with the variance components cal-
culated for the two components. The pooled within replicate -standard deviation is 0.13 in the transformed
data. Exponentiating gives 1.14, or 14% variability from sample to sample. The standard deviation between
samples is 0.061, which is smaller than the replication error estimate and hence insignificant. The implication
is that the observed variability is, in fact, replication error.
The data from the selected pathogenic bacteria analyses are presented in Table VI. A-7. As in
the coliphage data above, standard deviations are calculated for the results from each sample, then pooled to
obtain an overall estimate.
For Klebsiella, the pooled replication standard deviation is 0.26 in the transformed data,
which when exponentiated gives 1.30 in the original scale. This implies that there is 30-percent variability
from one analytical result to another from replicates of the same effluent sample. The difference between the
means of the two groups gives an estimated standard deviation of 0.36, and again the between-sample term is
insignificant.
The Pseudomonas data give a pooled standard deviation of 0.27 under the transformed
scale, which gives 1.31 or 31-percent variability in the original scale. This is to be compared with the variabil-
ity between the means of the two groups. This is estimated as 0.54, which is not significantly different. As a
result, the six values may be considered replicates in the usual sense.
The streptococci results from within each 500-mL grab sample yield a pooled replication
standard deviation of 0.13, and when exponentiated gives 1.14, or an estimated 14-percent variability between
the samples within a given pond grab sample. The mean values give a standard deviation under the transfor-
mation of 0.06, and once again is less than the replication error and insignificant.
52
-------
Table VI.A-6.
PRECISION QUALITY ASSURANCE STUDY FOR COLIPHAGE
Corrected Count Standard Deviation
Grab Sample (pfu/1 x i&) (InX)
6 4.5 0.06
7 4.6
8 5.3
9 4.8
10 4.7
11 4.3 0.17
12 3.6
13 4.4
14 5.7
15 4.9
The Clostridium perfringens results have a standard deviation of 0.21 in the transformed
scale for replication error in analyzing the same effluent sample. This corresponds in the original scale to a
value of 1.23, or 23-percent variability. The difference in the mean values is represented by a standard devia-
tion of 0.39, which again represents an insignificant difference.
In summary, for all the bacterial analyses, no variation between samples was detectable, and
the only discernable variability was due to analytical repeatability.
Table VI.A-7.
PRECISION QUALITY ASSURANCE STUDY FOR BACTERIA
Parameter
Fecal C.
Klebsiella Pseudomonas Streptococci Perfringens
Grab Sample (cfu/lOOml x 1Q4) (cfu/lOOml x 1Q2) (cfu/lOOml x 1Q3) (MPNxiQ3)
1 6 5.2 1100 9.2 4.3
7 3.8 1600 8.3 3.9
8 4.4 800 11 3.9
Std. Dev.
(Inx) 0.16 0.35 0.14 0.05
2 9 3.6 850 11 4.3
10 3.3 850 9.0 2.3
11 3.0 1100 9.7 2.3
Std. Dev.
(Inx) 0.09 0.15 0.10 0.36
53
-------
The results from the two virus analyses, 3-day count and 5-day count, are summarized in
Table VI. A-8. Sample 11 was contaminated and not used in the analyses. Sample 10, though high, was not
suggested to be contaminated and retained. The repeatability standard deviation for the natural logarithm of
the data is 0.60, which calculates to a value of 1.82 in the original scale. This suggests that over 80-percent
variability can be expected in replicate samples. The means represent a variability of only 0.16, and are much
closer than the sample values. Again the between-sample term is insignificant and all error can be assumed to
be replication error. This high between-sample variability is emphasized because it significantly impacts the
statistical reliability of mathematical modeling efforts for virus aerosol concentrations presented later in this
report.
For the 5-day count, the repeatability of the method is determined to be 0.55 in the trans-
formed scale and 1.73 in the original scale. The between-sample component is estimated as 0.14, and no sig-
nificant difference exists between these two values. Thus, for all the analyses, no differences could be found
between the two samples. A summary of the estimated replication error for coliphage and all the pathogenic
analyses is presented in Tafcle VI. A-9.
Table VI.A-8.
PRECISION QUALITY ASSURANCE STUDY FOR VIRUS
(Corrected Count, pfu/1)
Parameter
Grab Sample 3-Day Count 5-Day Count
1 6 24 35
7 11 15
8 18 25
Std.Dev.
(In x) 0.39 0.43
2 9 9.0 16
10 31 45
11* 852 926
Std.Dev.
(In x) 0.87 0.73
*contaminated sample
Table VI.A-9.
ESTIMATED REPLICATION ERROR FOR PATHOGENIC ANALYSES
Coefficient of
Parameter Variation, %
Coliphage 14
Klebsiella 30
Pseudomonas 31
Streptococci 14
Clostridium perfringens 23
3-Day Virus Count 82
5-Day Virus Count 73
54
-------
d. Equivalence of Composite and Pond Grab Samples
The daily composite effluent samples, taken during Pre-Fair, represented an average of the
effluent conditions in the pond near the pump over the time of spraying. At the time an aerosol run was made,
however, an additional composite effluent sample was taken from a spray head on the sprinkler line. These
represent the effluent quality, with respect to the microbiological indicators, over the 30-minute period of
aerosol sampling. These two determinations of the microbiological constituents of the effluent were com-
pared to indicate whether there was significant variability in these levels during a given day. This analysis was
also conducted to determine the requirement for continued effluent sampling during aerosol runs or whether
further studies could rely on microorganism concentration data from daily effluent composite samples.
The analysis was conducted in two parts. First, a correlation coefficient was calculated be-
tween the daily composite sample and the aerosol run composite sample and the significance of the correla-
tion determined. Second, a comparative t-test was conducted to determine if the two types of samples could
be estimating the same true mean. For these analyses, the results were transformed to normality using the nat-
ural logarithm in order to satisfy the assumptions of the statistical tests.
The comparative total coliform data are shown in Table VI. A-10 for the 25 aerosol runs
conducted. The summary statistics from the two analyses are presented at the bottom of the table. The corre-
lation coefficient for these data is estimated to be r = 0.73, and the t-test for significance gives a value of 5.13
with 23 degrees of freedom. These are significant at a level of less than 0.001, so that the daily and run sam-
ples can be said to be correlated. The t-test for equality of means is conducted by taking the difference be-
tween the two values for a given run and testing for a mean difference of zero. The t-statistic calculated for
this test is 0.54 with 24 degrees of freedom and is clearly not significant. This implies that both samples are
estimating the same mean level. Thus, the total coliform did not show a pattern of daily variability in these
data.
The paired values and results of these statistical analyses for the fecal coliform are shown in
Table VI. A-l 1. There are three missing values among these results when the laboratory was unable to per-
form the analyses due to equipment malfunction. The correlation coefficient for the remaining 22 pairs is esti-
mated as r = 0.40, with a t-statistic of 1.97 with 20 degrees of freedom. This has a significance level of approx-
imately 0.07, and, thus, is not significant at the 5-percent level. The t-test for equality gives a value of 0.50
with 21 degrees of freedom, which indicates that the two analyses are equivalent. The correlation analysis in-
dicates that an increase in one is not necessarily accompanied by an increase in the other, but overall they can
be said to be estimating the same level of fecal coliform in the effluent.
The standard bacterial plate count data are presented in Table VI. A-12. One value is missing
since the laboratory did not perform the analysis. The estimated correlation coefficient is r = 0.66, and the t-
statistic has a value of 4.17 with 22 degrees of freedom. This has a significance level of less than 0.001, and
they may thus be considered to be correlated.
The t-test for difference of means for the standard bacterial plate count gives a test statistic
of 2.96 with 23 degrees of freedom. This value is significant at the 0.01-percent level. By inspection, the aero-
sol effluent run samples can be seen to be estimating a higher mean level than the daily composite samples. It
was noted that the pipes carrying the water to the spray fields were not free of material, and as a result, the
water may be increasing in microbiological constituents, other than coliform, after it leaves the ponds.
The final indicator is the coliphage count data, presented in Table VI. A-13, for the twenty
ordinary aerosol runs and the two quality assurance runs. On the first run, neither the daily nor the run sam-
ple produced a valid result, and on run 34, no analytical results were obtained for the run sample. The esti-
mated correlation coefficient for the remaining 20 pairs of counts is 0.50, and the corresponding t-statistic is
calculated as 2.42 with 18 degrees of freedom. This is a significant value at the 0.05 level, and the values may
55
-------
be said to be correlated. The test for difference between means gives a test statistic of 0.68 with 19 degrees of
freedom, and is clearly non-significant. The conclusion, therefore, is that the two samples are providing esti-
mates of the same true mean levels.
In general, then, the results of the aerosol run effluent samples in the spray field were
Table VI.A-10.
COMPARATIVE TOTAL COLIFORM DATA—DAILY vs. RUN SAMPLE
Run Number
Date
Daily Sample
Spray Sample
(MFC/100 ml x 10 )
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16-19
20-23
24-25
26
27-30
31
32
33
34
35
Test
t-statistic
degrees of freedom
significance level
5-4
4-5
5-5
5-13
5-13
5-17
5-17
5-21
5-24
5-24
5-25
5-27
5-27
5-27
6-3
6-7
6-8
6-9
6-10
6-13
6-14
6-15
6-15
6-16
6-17
Correlation
(r = 0.73)
5.13
23
<0.001
2500
2200
2200
817
817
410
410
830
800
800
992
763
763
763
267
417
160
760
798
490
743
620
620
340
530
Equality
0.54
24
not significant
1480
2070
2140
690
720
930
700
950
1040
1280
1100
470
690
750
588
343
127
1100
758
267
970
265
170
550
350
56
-------
strongly related to those of daily composite samples from the effluent pond. The estimated correlation coeffi-
cients do not indicate a high degree of associativity, even though they are significant for the most part. The
amount of the explained variation, estimated by r2, has a maximum value of 54 percent for the parameters
studied, which is not particularly high. The more important result is that for all the parameters except stan-
dard bacterial plate count, the two results are estimating the same mean value. From these data it was con-
cluded that daily composite effluent samples were not necessary during Post-Fair studies.
Table VI.A-11.
COMPARATIVE FECAL COLIFORM DATA—DAILY vs. RUN SAMPLE
yivij: \-./ AVVJ mi x. iO
Run No. Date
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16-19
20-23
24-25
26
27-30
31
32
33
34
35
5-4
5
5
13
13
17
17
21
24
24
25
27
27
27
6-3
7
8
9
10
13
14
15
15
16
17
Test
t-statistic
degrees of freedom
significance level
* missing data
(MFC/100 ml x
Daily Sample
210
180
180
82.7
82.7
12.5
12.5
82.8
123
123
83.3
24.3
59
22.3
85.7
80.4
70.3
93.2
79
79
82.3
63.3
Correlation
(r=0.40)
1.97
20
0.07
Equality
)
Run Sample
150
186
174
97
180
75
110
80
81
125
124
#
*
57
59.5
14
177
75.3
27.7
137
24
34
81
45
0.50
21
not significant
57
-------
e. Relationship of Pathogen Levels to Indicator Organism Levels
An important Pre-Fair objective regarding the effluent sample analyses was to investigate
the relationships of the pathogenic organism (Pseudomonas, streptococci, C/ostr/cf/um perfringens, 3-day vi-
ruses, and 5-day viruses) effluent concentrations, to the indicator organism (total coliform, fecal coliform,
standard bacterial plate count, and coliphage) effluent concentrations. A common microbiological data anal-
ysis technique is to assume the existence of proportional relationships. If a pathogenic level was found to be
directly related to an indicator level in the Pleasanton effluent, then the indicator measurement and the identi-
fied relationship could substitute for assay of the pathogen in later phases of this study.
An effluent sample data base was constructed to investigate the potential pathogen-indicator
Table VI.A-12.
COMPARATIVE STANDARD BACTERIAL PLATE COUNT DATA
DAILY vs. RUN SAMPLE
Run Number
Date
Daily Run
Sample Sample
(No. /100 ml x 10°)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16-19
20-23
24-25
26
27-30
31
32
33
34
35
Test
5-4
5
5
13
13
17
17
21
24
24
25
27
27
27
6- 3
7
8
9
10
13
14
15
15
16
17
Correlation
(r = 0. 66)
t-statistic 4. 17
degrees of freedom 22
significance level ^0.01
^missing data
100
110
110
237
237
34.3
34.3
75
85
85
57
13.5
13.5
13.5
35
11.3
36
19.3
15.7
14
12.7
21.3
21.3
14.7
6. 13
Equality
2.96
23
•<0.01
110
120
#
101
167
76
46
70
104
145
35
34
26
73
45
14
36.7
44
34.7
12
16
20
95
54
76
58
-------
relationships. This data base consisted of the 47 daily effluent samples (the composites and the grab sample
substitutes when the composite sampler was inoperative) presented in Tables VI. A-1 and VI. A-2 and the
seven large-volume effluent grab samples taken for microbial characterization during Pre-Fair.
Because the data for the parameters given in Table VI. A-2 differ in validity and informatio-
nal content, a procedure for adjusting certain analysis results and weighting each result was developed and
applied to obtain the effluent data base. The analysis values that were missing or could not be quantified (the
Table VI.A-13.
COMPARATIVE COLIPHAGE DATA—DAILY vs. RUN SAMPLE
R un No.
Date
(pfu/1 x 103)
Daily Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
24
26
31
32
33
34
35
5-4
5
5
13
13
17
17
21
24
24
25
27
27
27
6-3
9
10
14
15
15
16
17
_ *
120
120
540
540
97
97
190
78
78
130
180
180
180
280
570
490
350
220
220
230
220
Run Sample
*
110
95
110
170
220
230
61
95
130
140
110
170
210
310
580
480
380
320
240
__
170
Test
t-statistic
degrees of freedom
significance level
* missing data
Correlation
(r = 0.50)
2.42
18
<0.05
Equality
0.68
19
not significant
59
-------
CS and TNTC designations in Tables VI. A-l and VI. A-2) were assigned zero weight, i.e., excluded from the
analysis. The extremely high 3-day virus values on the May 4 and May 6 daily samples were at first considered
to be possible contamination with the concentration efficiency reference poliovirus. With the perspective of
all the Pre-Fair effluent virus results, poliovirus contamination of these samples appears very likely. Since
these extremely high values would have a major effect on any virus relationship present in the effluent data,
these two virus values were also excluded from the analysis.
In the weighting procedure, the standard quantitative analysis values were given full weight.
The inferior quantitative analysis values were assigned half the full weight. These inferior values consisted of
results outside stated detection limits, results obtained by an analysis method inferior to the standard method
(MPN for streptococci), results probably affected by an elevated temperature during shipment, and results
reported as possible sample contamination. To quantify results outside the stated detection limits, analysis
results below the lower detection limit were reset at half the lower detection limit, and analysis results above
the upper detection limit were adjusted to be twice the upper detection limit. Unusually high 3- and 5-day
virus values were reported for the first large-volume pond grab sample. Because these virus values were not so
large as to be definite outliers, the likelihood of poliovirus contamination was less and these values were each
given half weight.
A correlation analysis of the natural log transformed microbiological effluent concentration
data was performed to seek relationships between pathogen-indicator parameter pairs. Of the 54 effluent
samples, there were no missing values for total coliform and only one missing value each for fecal coliform,
standard bacterial plate count, and coliphage. To facilitate the correlation, regression, and canonical correla-
tion calculations, the geometric means of the data on the other 53 samples were substituted for the missing
indicator value, and the substitute value was assigned half weight. A correlation coefficient was calculated for
each pathogen-indicator pair over all those effluent samples for which there was a pathogen concentration
value. The correlation coefficients obtained are presented in Table VI. A-14. The upper set of correlation co-
efficients in Table VI. A-14 are for the unweighted logarithmically transformed effluent data base. The lower
set of correlation coefficients is calculated from the weighted logarithmically transformed effluent data base.
The observation weight assigned each pathogen-indicator sample pair was the product of the pathogen sample
weight and the indicator sample weight, standardized so that 1.00 was the average observation weight. The
weighted correlation coefficients are considered more valid than the unweighted correlation coefficients, and
have been used to make the relationship inferences.
Inspection of the correlation coefficients in Table VI. A-14 shows very little correlation be-
tween the pathogen and indicator effluent concentrations. There are negative correlations (between virus and
coliform) as well as positive correlations (between some pathogenic bacteria and coliform). In contrast to the
low correlations shown in Table VI. A-14, the correlations among some of weighted log transformed effluent
indicator organisms were much higher: 0.879 for total coliform and fecal coliform, and 0.451 for total col-
iform and standard bacterial plate count.
The significance of the correlation coefficient between n pairs can be determined by testing
the null hypothesis of no correlation between the parameters against the two-sided alternative using a t-dis-
tributed statistic with n-2 degrees of freedom. This test is only valid when at least one of the pair of variates is
normally distributed. In Table VI. A-14, the significance levels of the correlation coefficients are presented
for those pairs of weighted natural log transformed parameters for which the correlation coefficient was sig-
nificant at the 0.05-percent level. Over the concentration ranges observed at Pleasanton, the only significant
effluent pathogen-indicator correlations were: 0.362 between streptococci and total coliform (P = .01); 0.354
between streptococci and fecal coliform (P = .05); and -0.355 between the 5-day virus plaques and fecal col-
iform. Because the distributional analysis showed at least one parameter in every correlated pair to be lognor-
mally distributed, the preceding correlation significance test is considered valid.
60
-------
Although there were some significant effluent pathogen-indicator correlations, the indicated
pairwise relationships were considered too weak to justify substitution of these relationships for future patho-
gen assays. However, for each of the pathogenic organisms that did have at least one significantly correlated
indicator organism, a stepwise multiple linear regression was conducted on all the natural log transformed
indicator organism data. The purpose was to see how well the best linear combination of the indicator orga-
nism concentration data might predict pathogenic concentrations. Weighted stepwise regression analyses were
performed using as dependent variables the natural log transformations of the streptococci, Clostridium perf-
ringens, and 5-day viruses effluent concentrations. In each regression, the potential regressor variables were
the natural log transformations of the total coliform, fecal coliform, standard bacterial plate count, and col-
iphage effluent concentrations. In the weighted regression analysis, a weight must be assigned to each obser-
vation (i.e., the set of indicator and pathogen results for a sample). The observation weighting procedure used
was to multiply the dependent parameter sample weight by a linear combination of the indicator parameter
sample weights and divide by the standardizing average observation weight. It is desirable to give more weight
to those samples for which the indicator organisms, more likely to be included in the regression equation, had
quantitative values. Accordingly, the significantly correlated indicator organism weights were doubled in
computing the indicator organism weight linear combination.
Table VI.A-14.
UNWEIGHTED AND WEIGHTED SAMPLE CORRELATIONS OF THE NATURAL LOG
TRANSFORMED EFFLUENT CONCENTRATIONS OF THE INDICATOR AND PATHOGENIC
MICROBIOLOGICAL PARAMETERS
Pathogenic Bacteria
Pseudo- Strepto-
monas cocci
No. of Daily and Large 52 46
Volume Effluent Samples
with Numeric Analysis
Results
Correlation Coefficients of
Unweighted Analyses-
Indicator Parameters
Total Co liform .169 .276
Fecal Coliform .091 .264
Std Plate Count .161 .081
Coliphage .192 -.009
Correlation Coefficients of
Weighted Analyses:
Indicator Parameters
Total Coliform .148 .362
Fecal Coliform .077 .354
Std Plate Count .163 .109
Coliphage .189 -.044
Two- Sided Significance
Level of the Significant
Weighted Analyses Corre-
lation Coefficients:
Indicator Parameters
Total Coliform - P=.01
Fecal Coliform - P=.02
Std Plate Count
Coliphage - -
Pathogenic
-i Viruses
Clostndium
perfrmgens Count Coun{
53 48 39
.332 -.240 -.228
.256 -.218 -.313
-.028 -.054 .166
.121 .146 .147
.269 -.238 -.266
.225 -.219 -.355
-.063 -.053 .126
.143 .124 .188
P=.05
P=.03
61
-------
To simplify presentation of the statistical analyses of the microbiological effluent and aero-
sol sample analyses, a consistent notation has been used. This notation is shown in Table VI. A-15.
The best equations generated through the weighted stepwise regression analysis are presented
in Table VI. A-16. The best streptococci regression equation involves only total coliforms. It is:
lnxST =0.498 In XTC + 2.136
This equation explains only 0.136 = 13.6% of the observed variation in the streptococci effluent concentra-
tion. Thus, while it is significantly better (P = .012) than no relationship, this streptococci-total coliform
relationship is not very strong. The best Clostridium perfringens regression equation contains the total col-
iform and standard bacterial plate count indicators:
In XCL = 0.658 In XTC — 0.263 In Xpc + 4.457
This regression equation has less predictive ability than the streptococci regression. It only accounts for
11.5% of the variation in the effluent sample concentrations of Clostridium perfringens. The negative coeffi-
cient of the standard bacterial plate count variable could suggest a negative correlation between Clostridium
perfringens and the non-coliform bacteria. However, with a t-statistic of -1.58, this coefficient is not signifi-
cant at the 0.1 level, and such an interpretation is unwarranted. The best regression equation for the 5-day
virus plaque effluent concentration is:
In XV5e = —0.622 In XFC +0.611 In Xcp + 0.302 In Xpc — 3.222
Table VI.A-15.
MICROBIOLOGICAL CONSTITUENT NOTATION
Microbiological Constituent Mnemonic Notation
Indicators:
Total Coliform TC
Fecal Coliform FC
Standard Plate Count PC
Coliphage CP
Pathogens:
Pseudomonas PS
Streptococci ST
Clostridium perfringens CL
Three-day virus plaques V3
Five-day Virus plaques V5
Aerosol Sample a
Effluent Sample e
Concentration X
62
-------
Table VI.A-16.
SUMMARY OF BEST MULTIPLE REGRESSION EQUATIONS FOR PREDICTING PATHOGEN
EFFLUENT CONCENTRATION FROM THE INDICATOR EFFLUENT CONCENTRATIONS
OS
Regressed
Effluent
Pathogen
Concentration
Variable
Ln Xgf
e
«-"k.
«.,.
Best Predictive Regression
Coefficient Significance Term
of Multiple of Indicator
Determination Regression Regression
(RZ) F- Ratio Variable
. 136 P=. 012 Ln XTc
I*
. U5 P=. 047 Ln X_.,
lue
LnX^
1*
. 275 P=. 010 Ln X_r
rue
Lnxcpe
LnXpce
I*
Equation
in Regression Equation
Regression
Coefficient
0.498
2.136
.0658
-0.263
4. 457
-0.622
0. 611
0.302
-3.222
Regression
Coefficient
t Statistic
2.63
2.50
-1.58
-3.12
2.17
2.12
*1 represents the constant term in the regression equation
-------
The equation explains 27.5% of the observed 5-day virus variation. It is significantly better (P= .01) than no
regression relationship. The 5-day virus level seems to depend somewhat on the non-fecal coliform bacteria
level and the coliphage level. However, the regression coefficient t-statistics are still too small to lend much
credence to such an interpretation.
The stepwise multiple linear regression equations given above are an inadequate basis for re-
lating the pathogen and indicator effluent concentrations over the microorganism concentration ranges ob-
served in the Pleasanton effluent. Considerably more than half of the log transformed pathogen concentra-
tion variability (i.e., pathogen percentage variation) cannot be accounted for by the variability of the
indicator organism concentrations.
A third method, canonical correlation, was also employed to seek those factors that the set
of pathogen measurements and the set of indicator measurements have in common over all the effluent sam-
ples. Canonical correlation might identify indicator relationships for some combination of pathogens that
could not be detected by the correlation and regression analyses, which only relate one pathogen at a time. As
in the correlation and regression analyses, natural log transformations of each microorganism's effluent con-
centration were employed to construct the analysis variables. Observation weighting was accomplished by
multiplying the indicator organism weight sum by the pathogenic organism weight sum and dividing by the
average observation weight. Only the 36 effluent samples, for which all nine pathogen and indicator orga-
nisms had values, were included in the initial canonical correlation analysis. The results of this analysis are
summarized in Table VI. A-17. One significant pair of canonical variables was identified. The canonical cor-
relation is 0.638. This pair of canonical variables share 0.407 = 40.7% of their variation. Wilk's lambda sta-
tistic is transformed into a chi-square statistic to determine that P = 0.041 is the statistical significance of this
canonical correlation. The coefficients of the canonical variables are shown in the lower part of Table VI. A-
17. The important components of the pathogen canonical variate are 5-day virus (positive coefficient) and
streptococci (negative coefficient). The important indicator canonical variate components are standard bacte-
rial plate count (positive coefficient) and total and fecal coliform (negative coefficients).
Variants of the preceding case were also analyzed by canonical correlation. The pathogenic
bacteria set was related to the indicator set, but there were no significant canonical correlations. The viruses
set was related to the indicator set. The one significant canonical variable pair is very similar to the significant
canonical variable pair displayed in Table VI. A-17. The variable coefficients have the same signs and nearly
the same magnitudes. Pairwise deletion of missing data was used to construct the basic correlation coefficient
matrix from all the observations for which both parameters in each pair had numerical values. The only
nearly significant (P = 0.07) canonical variable pair emerging from this analysis which used all the available
data was nearly identical to the significant canonical variable pair given in Table VI. A-17.
By comparing Table VI. A-17 with Table VI. A-16, it can be seen that the one significant
canonical variate pair is basically a linear combination of the regression equations for 5-day viruses and strep-
tococci. Substituting the regression equations into the expression 0.968 LnXY5 —0.422 LnXST yields a linear
combination of the natural log indicator parameter that agree in sign and compare well in magnitude with the
indicator set coefficients given in Table VI. A-17. Thus, canonical correlation analysis has not identified any
new pathogen-indicator relationships. In fact, it strongly suggests that, beyond the meager regression
relationships given in Table VI. A-16, there are no more substantive relationships among the pathogen and
indicator microorganism groups in the Pre-Fair effluent samples obtained at Pleasanton.
In summary, over the ranges of the effluent microbiological group concentrations obtained
during the Pre-Fair sampling at Pleasanton (from one order of magnitude for coliphage and total coliform to
well over two orders of magnitude for Pseudomonas and Clostridium perfringens), there are only the most
tenuous of relationships between some pathogenic parameters and some indicator parameters. These
64
-------
Pair
1
2
3
Eigenvalue
0.407
0.298
0.121
Canonical
Correlation
0.638
0.538
0.348
Wilk's
Lambda
0.354
0.597
0.840
relationships are certainly an insufficient basis for discontinuing the pathogenic analyses of the effluent sam-
ples in later phases of the study.
/. Microbial Characterization
A thorough characterization of the treated sewage effluent at the Pleasanton site was con-
ducted to identify the types of pathogenic bacteria and viruses. Few pathogenic organisms were isolated, de-
spite the detection of relatively high levels of indicator bacteria in effluent and aerosol samples. The analytical
methods employed in the present study were designed to provide definitive information on the types and ap-
proximate quantities of the bacterial population (including non-pathogens, opportunistic pathogens, and
overt pathogens) present in the sewage effluent.
The levels of the routinely-assessed microorganism groups are given in Table VI. A-18 for
Table VI.A-17.
CANONICAL CORRELATION OF THE PATHOGEN EFFLUENT CONCENTRATION SET WITH
THE INDICATOR EFFLUENT CONCENTRATION SET
Significance of
Canonical Correlation
P = 0.041
P = 0.191
P = 0.491
Coefficients of Significant Canonical Variables (P< 0.05)
Pair 1
Indicator Set
Ln XTC -0.623
e
LN Xpc -0.475
e
Ln XpC 0.679
e
•Ln Xcp 0.162
e
Pathogen Set
Ln XpS -0.045
e
Ln XST -0.422
e
Ln XCL -0.113
e
Ln Xy3 -0.313
e
Ln X.., 0.968
65
-------
Table VI.A-18.
EFFLUENT CONCENTRATIONS OF USUAL MICROBIOLOGICAL CONSTITUENTS IN
LARGE-VOLUME SAMPLES TAKEN FOR MICROBIAL CHARACTERIZATION
Sample
Date
4-27-76
5-5-76
5-13-76
5-1 1-76
5-24-76
61-76
6-22-76
Tolal
Conform
(MFC/100 ml) X 10'
670
3170
240
1300
2000
1100
473
Fecal
Coliform
< MFC/ 100 ml) X 10'
56
350
27
115
230
161
85
Standard
Plate Count
(No./IOfl ml) X 10*
M
100
40
61
200
61
II
Cnlipliagc
(PPU/OX 10'
210
500
260
490
80
510
190
Kkb
-------
the seven large-volume effluent samples taken for microbial characterization during Pre-Fair. These levels are
representative of those obtained for the daily composite effluent samples.
A summary of all other bacterial types identified in these large-volume effluent samples is
presented in Table VI. A-19. Data from the first large-volume sample and portions of the data from the sec-
ond and third large-volume samples could not be obtained. These samples were received before the various
differential and selective plating and diagnostic media required for the respective analyses were available.
Large-volume samples obtained, on 5-21, 5-24, 6-1, 6-22, and 11-29, 1976 were characterized for every bacte-
rial type or group listed, according to the procedures described in Appendix D, Methods and Materials. In
addition, three of the daily composite effluent samples from Pre-Fair (i.e. sample dates 6-15, 6-16, and 6-17,
1976) were subjected to the same rigorous analyses. This provides additional data for comparison of the bac-
terial populations in the two types of effluent samples. It should be noted that the data from the microbial
characterization of the aggregated large-volume aerosol sample (discussed as run Ml-36 in Section VLB.6)
are also presented for comparison.
The data show that the components of the bacterial populations of the large volume (grab)
and daily composite samples examined were qualitatively similar. Inspection of the quantitative data obtained
for the selected microbial parameters (Klebsiella, Pseudomonas, fecal streptococci, and Clostridium perfring-
ens) (Table VI.A-18) in the two types of effluent samples also suggests similarity. The approximate numbers
of other bacterial types isolated from both types of samples were not appreciably different, with the exception
of the isolation of Leptospira, Salmonella, and Shigella only from the large-volume samples. However, this
observation may be a reflection of the small number of samples examined.
The small number of Staphylococcus aureus isolated from the effluent samples was surpris-
ing. Four of the eight large-volume and all of the daily composite samples plated to Mannitol Salt Agar were
positive for this organism. Only the lowest dilution plates (0.1 mL) yielded numerous colonies, most of which
were Micrococcaceae. The maximal number of the latter which proved to be Staphylococcus aureus was 40
cfu/ml in the composite for 6-16. In contrast, large numbers of colonies grew on the selective medium for
Neisseria at 10'2 dilutions of the effluent samples. However, none of the representative colony types picked
proved to be Neisseria gonorrheas or Neisseria meningitidis. Enrichment for Leptospira was positive in four
of five large-volume samples but not the three daily composite samples examined. The negative results in the
latter may be the result of lack of sensitivity of the test system employed. Enrichment tubes showing turbidity
were examined by dark field microscopy. Failure to observe an organism with typical Leptospira morphology
and motility in one of at least 15 randomly-selected fields constituted a negative test. However, only a very
small fraction of an enrichment is examined when this is carried out microscopically, a fact complicated by
the wet mounts required for dark field. Thus, positive enrichment tubes with light growth of Leptospira may
not be detected. Of the positive samples, enrichment tubes from the large-volume sample of 5-24 yielded the
greatest number of "typical" Leptospira (40/15 microscopic fields).
The genus Mycobacterium includes species that range from saprophytes widely distributed in
soil and water to facultative and obligate intracellular parasites. Positive isolation of mycobacteria from ef-
fluent samples was expected. However, the large number of these organisms (approximately 10 CFU/mL) in
every large-volume and daily composite sample examined was surprising. Isolation and identification of these
organisms was facilitated by the treatment procedure which resulted in plates that were relatively free of other
organisms. None of the isolates of the various colony types that were carried through the identification
scheme (Figure III.B-2 in Appendix D) proved to be Mycobacterium tuberculosis. However, representatives
of the following groups or species were identified: Mycobacterium ulcerans, Mycobacterium gordanae, Ru-
nyon Group I, Runyon Group II, Runyon Group III, Mycobacterium marinum, Runyon Group IV. A single
confirmed Mycobacterium ulcerans was the only Mycobacterium isolated that is a pathogen of the same cat-
67
-------
89
~s rD
c-s
n o
X
I
r-o — •> K> r\j — ' en
ro i s> — ' u> !
i >-j i i i — i
— I CT>^J^J ^JC^
CTt CTl O~! CD
XXX XXXXXX X
XXX XXXXXX
x x x x
X X X X X
XXX XXXXXX X
XXX
Xxx Xxxxx
XXX XXXXX
XX XX
XXX XXXXX
XXX
XXX
X X X
XXXXX
oo z. :s:
QJ 3 rt- l/i
n> — " T ro
Clostridium
perfringens
Fecal Streptococci
StaphylococcUsaureus
Leptospira
Mycobacteria
Neisseria (patho-
genic species)
Fluorescent
Ps e udomonads
Other Oxidase Pos.
Glucose Oxidizers
Alcaliqenes and Other
Oxid. Pos. Glucose
Inactive
Aclnetobacter and
Other Oxid. Neg.
Klebsiella
Enterobacter
Serratia
Edwardsiella
Escherichia
Shiqella
Salmonella
Arizona
Citrobacter
Proteus
Providencia
ifersinia
Aeromonas and Other
Oxid. Pos. Fermenters
I
0
3
I
Ferme
3
rt
in
w
rt
n>
n
tu
n
rt
K
H-
B)
ceae
CA
O
•n
e
H ^
I—« •<
>
90
o
o
r
m
w
-------
egory as Mycobacterium tuberculosis^. Others range from usually pathogenic (Mycobacterium marinum) to
usually nonpathogenic.
The distribution of the various groups of Gram-negative nonfermenting organisms, the gen-
era of Enterobacteriaceae, and the oxidase positive fermenting organisms was determined by characterization
of colonies picked from moderately to highly selective enteric plating media (Tables VI. A-19 and VI. A-20).
A different pattern, with a better representation of the coliforms, would probably have been obtained by ex-
tensive isolation and characterization of colonies from the nonselective media EMB, ENDO, and MacCon-
keys. However, the possibility of enhancing the frequency of isolation of the major enteric pathogens was
desired. As illustrated in Table VI. A-20, 79 percent of the isolates from the direct platings on selective media
were oxidase positive. The majority of these organisms (49 percent) were fermenters which would have been
incorrectly identified as various genera of Enterobacteriaceae if the oxidase test had not been employed. In
contrast, 82 percent of isolates picked from the selective media following enrichment in GN, selenite, or tetra-
thionate broths were Enterobacteriaceae.
The distribution of the various species of Enterobacteriaceae observed in the large-volume
and daily composite effluent samples is summarized in Table VI. A-21. The data for the aggregated aerosol
sample are shown for comparison. Isolates of both Salmonella and Shigella were detected only in large-vol-
ume sample obtained on 6-22-76. When the data are taken as a whole, most of the species were represented in,
at least, one of the effluent samples. Table VI. A-22 is an analysis of the percentage of each species of Entero-
bacteriaceae isolated from the total of direct platings and enrichments of the large-volume samples. It should
be noted that the enrichment procedure for Yersinia enterocolitica (an organism increasingly implicated in
enterocolitis and mesenteric lymphadenitis) failed to yield a single isolate of this organism. The few colonies
which appeared on the plating media were predominantly oxidase-positive Gram-negative fermenters. How-
Table VI.A-20.
GROUPS OF ORGANISMS FROM DIRECT PLATINGS AND ENRICHMENTS FOR ENTERICS
Large-Volume Samples
% Isolates Direct Plating % Isolates From
Group on XLD, HER, BS, SS Enrichments
All Enterobacteriaceae 19 82
Fluorescent Pseudomonads 14 6
Non-Fermenting, Oxidase Positive 9 2
Glucose Oxidizers*
Alcaligenes and Other Oxidase 9 1
Positive Glucose Inactive
Acinetobacter and Other Oxidase 2 1
Negative Non-Fermenters
Aeromgnas and Other Oxidase 47 8
Positive Fermenters
* Other than fluorescent pseudomonads.
69
-------
OL
M i i i c n> : i i i
Z — J — ' — *- i T — 'TO — ' tn
2 Co ^-J Ch O i — ' ^) i
{^ t i i ~fc t/> --J i i --J
JJ1 en en o^
?s ^ x ^s x x x x
0 XX
H-
rt
H
(D
p. X X X X X X
V)
° X
(y XX XXXXX
M
X- XXX X
H-
3 .,
a x
O
* X X X
3 X
o
3
a>
a.
J xxx- xxxxx
n
rt
ro
a
X XX
X
XX XXX
X
X X
XXX" XXX
xxxx
X
X
on £ :u
o ai QJ o>
»A> 3 <-f in
<-i-T3 ro n-
(D — ' ~S fT)
m i
Klebsiella pneumoniae
Klebsiella ozaenae
Klebsiella rhinoschleroraatis
Enterobacter cloacae
Enterobacter aerogenes
Enterobacter hafniae
Acrogenic Ent. agglomerans
Anaerogenic Ent. agglomerans
,
Serratia marcescens
Serratia liquefaciens
Serratia rubidaea
Edwardsiella tarda
Escherichia coli
Escherichia coli A-D
Shigella (all species)
Salmonella (all species)
Arizona (Sal. arizonao)
Citrobacter freundii
Citrobacter diversus
Proteus vulgaris
Proteus mirabilis
Proteus morgann
Proteus rettgeri
Providencia alcalifaciens
Providencia stuartii
Yersinia enterocolitica
YerSinia pseudotuberculosis
W3
tn
o
tn
C/3
~
7
H
tn
b
^
H
3
90
o
>
CO
3 £
tn or
H ***
HM "^
2 r
w i
OKI
NH
O
W
O
c
W
to
o
o
r
Cfl
"B
>
-------
ever, a Yersinia pseudotuberculosis isolate and a Shigella isolate were identified from separate aerosol sam-
ples during a routine pick of colonies.
At the beginning of Post-Fair, a large volume sample was taken and analyzed by semi-quan-
titative procedures. The results of this microbiological screen are presented in Table VI. A-23.
g. Respiratory Virus
Five roller tubes showing viral cytopathology were subcultured successfully, indicating con-
firmation as viruses. The viruses from four of these five tubes were typed as ECHO 6. The virus in the other
tube remains unidentified, since typing was unsuccessful.
Thirty-five (35) roller tubes survived uncontaminated through primary subculture without
showing viral cytopathology. All subcultures were challenged with ECHO 11 for rubella. All showed typical
ECHO 11 cytopathology and were reported as negative for rubella. Fluids from all subcultures were inoc-
ulated into four 10-day embryonated chicken eggs, incubated 72 hours and the allantoic fluids harvested and
tested for hemagglutinins against crbc; all were negative. The control influenza strain gave 1:16 HA.
Table VI.A-22.
ANALYSIS OF ENTEROBACTERIACEAE
(Percent of Each Species Isolated From Total of Direct Platings and Enrichments-
Large Volume Samples of 5-21-76,5-24-76,6-1-76,6-22-76)
Excluding E. coli*
Percent
K. pneumoniae 9.2
K. ozaenae 1.5
K. rhinoscheromatis 0
!.• cloacae 11.4
E_. aerogenes 1. 5
E. hafniae 0.8
E_. agglomerans (aerogenic) 14.5
E_. agglomerans (anaerogenic) 3.1
S_. marcescens 0.8
§.• liquefaciens 2. 3
S_. rubidaea 0.8
]E. tarda 0
Shigella (all species) 3.1
Salmonella (all species) 0.8
Arizona
C. freundii
£. diversus
P_. vulgar is
P_. mirabilis
P_. morganii
—' rettgeri
P_. alcalifaciens
P. stuartii
]f • enterocolitica
Y_. pseudotuberculosis
Unidentified
Percent
16.0
1.5
3.8
16.8
3.1
2.3
0.8
0
0
0
6.1
*E. coli was the predominant representative of the Enterobacteriaceae. Values
are the percentage from direct platings on XLD, Hektoen, SS, and BS agars and from
enrichments in GN, tetrathionate, and selenite broths.
71
-------
Table VI. A-23.
SEMI-QUANTITATIVE MICROBIOLOGICAL SCREEN OF POST-FAIR SAMPLE (11-29-76)
Bacteria:
Arizona
Cltrobacter Freundll
EdwardsielTa
Enterobacter aerogenes
Enterobacter cloacae
Escherichia KgS*
Fecal Coliform
Fecal Streptococci
Klebsiella pneumoniae
Klebsiella ozaenae
Mycobacterla
Proteus
Providencia
Pseudomonas
Salmonella
Serratia
Shi gel la
Staphylococcus
Total col i form
Total Plate Count
Yerslnia
Of the total number of colonies which were
randomly picked and biochemically tested
for Enterobacteriaceaa:
Oxidae positive
Klebsiella ozaenae
Klebsiella pneumonia
Enterobacter cloacae
Enterobacter aerogenes
Escherichia H2ST '
Citrobacter freundii
No growth
Viruses:
Coliphage
Enteroviruses 3d»v
Sd"/,
Concentration Efficiency
Total Organic Carbon:
Filtered
Unfiltered
cfu/lOOml
<3.0xlOJ (NO)
1.0x105
<3.0x103 (NO)
l.OxlO4
\0xl06
"..0x105
.2x104
.7x103
.0x104
.0x10"
1.4x10"
0x103 (ND)
0x103 (ND)
2.0x104
3.0x103 (ND)
:3.0x!03 (ND)
3.0x103 (ND)
4.0x103
1.2x106
1.5x10?
* (ND)
46. 5*
6.9%
13.9%
13.9%
1.12
2.3%
5.8%
9.3%
4.8xl05 pfu/1
161 pfu/1
242 pfu/1
100%
33 mg/1
39 mg/1
Note: ND - none detected
* - nonquantitative procedure
B. Aerosol Run Data Characteristics
1. Meteorological and Sampling Conditions
a. Meteorological Conditions
Thirty-six microbiological aerosol runs were attempted during the Pre-Fair sampling period.
Five of the aerosol runs were quality assurance runs. Eleven aerosol runs had to be aborted because wind
shifts violated one or more of the aerosol run criteria specified in the aerosol sampling protocol (see Appendix
D). The 20 remaining aerosol sampling runs all met the protocol criteria. In addition to the microbiological
aerosol runs, seven dye aerosol runs were successfully completed during the Pre-Fair sampling period. The
environmental conditions during both the microbiological and dye aerosol runs appear in Table VI.B-1.
72
-------
Table VI.B-1.
AEROSOL RUN METEOROLOGICAL AND SOURCE DATA SUMMARY
a. Temperature and Relative Humidity
Aerosol Runs
Run
No.
Run
Date
Time
of
Run
Air Temperature, ° C
2m at
Effluent
Pond
2m at
Met.
Tower
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Relative Humidity , %
2m at
Effluent
Pond
2m at
Met.
Tower
Livermore
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Wastewater
Temperature,
°C
Dye Aerosol Runs
Pre-Fait
Dl-1
Dl-3
Dl-4
Dl-5
Dl-6
Dl-9
Dl-10
Post-Fair
D2-1
D2-2
D2-4
D2-5
D2-6
D2-7
02-8
D2-9
D2-10
D2-11
54-76
5-28-76
5-31-76
5-31-76
6-1-76
6-14-76
6-16-76
12-2-76
12-5-76
12-^-76
12-6-76
12-6-76
12-16-76
12-16-76
1-13-76
1-13-76
1-13-76
1355-1405
1555-1625
2030-2100
2126-2156
1805-1835
1740-1810
1640-1710
1435-1505
1253-1323
1345-1415
1520-1550
1523-1553
1303-1330
1347-1417
1312-1342
1400-1430
1835-1905
19.4
18.3
14.4
13.3
21.1
33.9
31.7
16.7
19.4
15.6
13.9
18.9
12.8
13.3
5.6
4.4
4.4
(19.4)
(18.3)
(14.4)
(13.3)
(21.1)
(33.9)
(31.7)
16.1
20.0
16.1
14.4
18.3
14.4
14.4
7.2
7.2
6.7
(19.9)
(19.0)
(15.8)
(14.9)
(21.3)
(31.8)
(30.0)
18.3
20.0
22.8
21.1
20.6
18.3
18.3
8.9
8.3
5.6
(18.3)
(17.5)
(14.7)
(13.9)
(19.5)
(28.9)
(27.3)
15.0
14.4
17.2
16.1
14.4
16.1
16.1
(8.3)
7.8
5.0
71
(70)
71
72
65
45
60
(74)
(70)
(74)
(77)
63
80
79
96
96
96
(39)
(22)
(40)
(76)
(24)
(5)
(19)
31
20
31
36
23
48
48
78
78
80
39
22
40
76
24
5
19
(31)
(20)
54
54
31
46
47
68
73
74
(54)
(37)
(55)
(91)
(39)
(20)
(34)
51
26
53
56
23
45
45
72
83
100
(59)
(44)
(60)
(92)
(46)
(29)
(41)
59
55
65
66
38
53
53
75
83
100
(21.7)
(23.8)
(22.7)
(22.4)
(23.5)
280
28.0
16.0
16.0
15.5
15.5
14.0
15.0
15.0
12.0
12.0
(1 1 5)
Microbiological Aerosol Runs
Prevail
MM
Ml-2
Ml-3
MM
Ml-5
Ml-6
Ml-7
Ml-8
MJj-9
54-77
5-5-77
5-5-77
5-13-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
1547-1617
1538-1608
1700-1730
1625-1655
1807-1837
1923-1953
2044-2114
1538-1608
1557-1627
21.1
18.9
18.3
35.0
32.2
13.9
10.0
23.3
17.2
(21.1)
(18.9)
(18.3)
(35.0)
(32.2)
(13.9)
(10.0)
(23.0)
(17.2)
(21.3)
(19.5)
(19.0)
(32.7)
(30.4)
(15.4)
(12.2)
(23.1)
(18.1)
(19.5)
(17.9)
(17.5)
(29.7)
(27.6)
(14.3)
(11.5)
(21.1)
(16.7)
72
72
71
52
52
65
75
65
76
(44)
(38)
(40)
(12)
(9)
(85)
(60)
(32)
(45)
44
38
40
12
9
85
60
32
45
(59)
(53)
(55)
(27)
(24)
(100)
(75)
(47)
(60)
(64)
(58)
(60)
(32)
(29)
(100)
(80)
(52)
(65)
(21.7)
(21.7)
(21.3)
(22.3)
(21.8)
(21.7)
(21.3)
(23.2)
(23.3)
NOTE: Values in parentheses have been substituted according to criteria described in the text. '
-------
Table VI.B-l.(confd)
Aerosol Runs
Run
No.
Run
Date
Time
of
Run
Air Temperature, ° C
2m at
Effluent
Pond
2m at
Met.
Tower
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Relative Humidity, %
2m at
Effluent
Pond
2m at
Met.
Tower
Livermore
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Wastewatet
Temperature,
°C
Microbiological Aerosol Runs (cont'd)
Pre-Faii (cont'd)
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
Ml-32
Ml-33
Ml-34
Ml-35
Post-Fair
M2-1
M2-2
M2-3
U2A
M2-5
M2-6
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
5-24-76
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-16-76
6-17-76
•-21-77
-25-77
-25-77
-30-77
-30-77
2-9-77
2-23-77
2-24-77
2-24-77
2-28-77
3-14-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-26-77
3-27-77
3-27-77
4-11-77
2032-2102
2214-2244
1717-1747
2040-2110
2150-2220
2117-2147
2010-2040
1420-1450
1635-1705
2140-2210
2242-2312
1600-1630
1415-1445
2012-2042
1421-1451
1603-1633
1353-1423
1835-1905
1403-1433
1513-1543
2003-2033
1625-1655
1555-1625
1616-1646
1620-1650
1344-1414
1315-1345
1634-1704
1330-1400
1518-1548
1515-1545
12.2
15.6
16.7
10.6
10.0
10.0
28.3
37.2
33.9
21.1
15.6
10.6
10.0
5.6
(8.7)
(8.5)
13.3
8.4
11.7
11.5
5.6
10.6
9.4
10.6
10.6
15.0
19.4
20.0
17.2
16.7
22.2
(12.2)
(15.6)
(16.7)
(10.6)
(10.0)
(10.0)
(28.3)
(37.2)
(33.9)
(21.1)
(15.6
13.3
12.1
8.9
9.7
9.5
15.0
9.4
13.3
13.5
7.8
11.4
10.0
11.6
18.9
16.7
20.6
19.5
17.8
16.1
21.7
(14.0)
(16.8)
(17.7)
(12.7)
(12.2)
(12.2)
(27.2)
(34.5)
(31.8)
(21.3)
(16.8)
12.8.
13.3
8.9
10.0
10.0
14.4
8.1
13.9
13.9
(8.6)
11.9
12.2
12.2
13.9
16.4
20.0
18.9
17.5
16.1
21.9
(13.1)
(15.5)
(16.3)
(11.9)
(11.5)
(11.5)
(24.8)
(31.3)
(28.9)
(19.5)
(15.5)
12.2
12.2
8.9
10.0
10.0
14.4
8.9
13.9
13.9
(8.3)
11 9
11.4
11.4
13.6
16.1
18.9
18.3
17.5
16.4
21.1
83
71
76
80
84
76
47
50
50
65
77
90
87
98
91
92
87
80
74
72
87
76
88
85
78
78
69
70
79
78
75
(86)
(37)
(46)
(75)
(80)
(64)
(5)
(6)
(20)
(45)
(65)
78
75
84
69
71
72
50
35
31
55
42
52
48
27
33
21
29
36
39
29
86
37
46
75
80
64
5
6
20
45
65
69
60
83
57
46
(72)
65
45
33
69
48
62
64
33
41
24
24
53
51
38
(100)
(52)
(61)
(90)
(95)
(79)
(20)
(21)
(35)
(60)
(80)
74
82
88
77
82
71
96
48
45
(70)
51
59
64
54
43
35
44
63
62
31
(100)
(57)
(66)
(95)
(100)
(84)
(25)
(26)
(40)
(65)
(85)
82
82
91
79
82
78
96
53
45
(75)
51
63
66
56
52
40
50
65
60
38
(22.0)
(21.6)
24.0
24.0
24.0
(22.6)
24.0
26.0
28.0
(23.7)
26.0
15.3
15.9
13.6
13.3
13.3
15.5
14.0
16.2
16.2
U b
17.5
17.3
17.3
17.6
17.6
17.9
18.0
17.8
18.0
19.8
-------
Table VI.B-l.(confd)
Aerosol Runs
Run
No.
Run
Date
Time
of
Run
Air Temperature, ° C
2m at
Effluent
Pond
2m at
Met.
Tower
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Relative Humidity , %
2m at
Effluent
Pond
2m at
Met.
Tower
Livermore
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Wastewater
Temperature,
°C
Microbiological A erosol Runs (con t 'd)
Post-Fair (c
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
mt'd)
4-12-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
1615-1645
2106-2135
2010-2040
1640-1710
2050-2120
1410-1440
1630-1700
1335-1405
1505-1535
22.8
8.9
(13.3)
21.7
15.6
(26.7)
(24.4)
21.7
18.9
23.3
10.0
13.3
22.8
13.3
26.7
24.4
22.5
19.5
21.9
8.1
108
21.7
11.4
24.4
23.3
20.3
18.9
21.4
8.3
11.4
20.8
11.7
23.9
22.5
19.4
17.5
69
91
(77)
68
68
(65)
(68)
73
74
21
52
37
16
25
9
15
30
23
28
82
55
16
21
20
22
36
32
17
63
65
29
43
23
26
39
36
19
80
66
33
44
28
29
45
45
19.8
19.5
18.0
19.9
19.4
19.6
19.9
19.7
19.6
Quality Assurance A erosol Runs
Pre-Fair
Ml-16-19
Ml -20-23
Ml-24-25
Ml-26
Ml-27-30
Post-Fair
M2-8-9
M2-27-28
Post-Fair
V2-I.1
V2-I.3
V2-1.4
V2-I.5
V2-I.6
V2-II.1
6-7-76
6-8-76
6-9-76
6-10-76
6-13-76
2-16-77
4-5-77
2104-2134
1750-1820
2259-2329
2212-2242
1707-1737
1805-1835
1635-1709
10.0
18.3
11.1
11.1
27.8
17.2
23.3
(10.0)
(18.3)
(11.1)
(11.1)
(27.8)
16.7
22.8
(12.2)
(19.0)
(13.1)
(13.1)
(268)
(18.1)
23.3
(11.5)
(17.5)
(12.3)
(12.3)
(24.4)
(16.7)
22.8
81
64
82
84
57
78
72
(74)
(40)
(70)
(64)
(9)
58
25
74
40
70
64
9
51
32
(89)
(55)
(85)
(79)
(24)
73
18
(94)
(60)
(90)
(84)
(29)
(76)
21
(23.0)
17.0
21.0
22.0
27.0
17.8
20.3
Virus Aerosol Runs
2-25-77
2-26-77
2-26-77
2-26-77
2-26-77
4-9-77
1753-1823
1505-1535
1545-1615
1624-1654
1703-1733
1450-1520
(11-1)
(16.7)
(16.7)
(15.6)
(14.4)
(16.1)
11.1
16.7
16.7
15.6
14.4
16.1
16.1
15.8
77
70
70
72
73
78
47
36
37
36
45
35
41
(62)
(51)
(52)
(51)
(60)
55
59
15.5
16.8
16.5
16.3
15.3
19.5
-------
Table VI.B-l.(cont'd)
Run
No.
Aerosol Runs
Run
Date
Time
of
Run
Air Temperature, ° C
2m at
Effluent
Pond
2m at
Met
Tower
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Relative Humidity. %
2m at
Effluent
Pond
2m at
Met.
Tower
Livermore
Just Upwind
of
Spray Line
Just Downwind
of (Within)
Spray Line
Wastewater
Temperature,
°C
Virus A erosol Runs (cont 'dj
Post-Fair (cont'd)
V2-II.2
V2-II.3
V2-I1.4
V2-1I.5
V2-II.6
4-9-77
4-9-77
4-9-77
4-9-77
4-9-77
1530-1600
1610-1640
1655-1725
1735-1805
1815-1845
(16.1)
U5.0)
(14.4)
(13.3)
(12.1)
16.1
15.0
14.4
13.3
12.1
15.8
15.0
14.2
13.9
11.7
155
14.4
13.6
13.3
11.7
78
78
79
80
81
35
38
38
40
44
41
41
41
41
41
54
56
60
59
66
59
63
61
61
66
19.3
19.2
19.2
19.2
19.0
-------
Table VI.B-1.
AEROSOL RUN METEOROLOGICAL AND SOURCE DATA SUMMARY
b. Wind Direction, Velocity, Stability and Solar Radiation
Aerosol Runs
Run
No.
Pre-Fair
DM
Dl-3
D14
Dl-5
Dl-6
Dl-9
Dl-10
Post-Fair
D2-1
D2-2
D2-4
D2-5
D2-6
D2-7
D2-8
D2-9
D2-10
D2-11
Pre-Fair
Ml-1
Ml-2
Ml-3
M14
Ml-5
Ml-6
Ml-7
Ml -8
Ml-9
Ml-10
Run
Date
Time of
Run
Mean Wind Direction,
deg
Relative
to
True North
Relative to
Perpendicular to
Spray Line
Wind Velocity, m/sec
10m at
Effluent
Pond
4m at
Met.
Tower
2m in
Spray
Field
Radiation
Cloud
Cover
in Eighths
Dye A erosol Runs
54-76
5-28-76
5-31-76
5-31-76
6-1-76
6-14-76
6-16-76
12-2-76
12-5-76
12-6-76
12-6-76
12-6-76
12-16-76
12-16-76
1-13-77
1-13-77
1-13-77
1355-1405
1555-1625
2030-2100
2126-2156
1805-1835
1740-1810
1640-1710
1435-1505
1253-1323
1345-1415
1520-1550
1523-1553
1303-1333
1347-1417
1312-1342
1400-1430
1835-1905
260
250
250
220
255
20
261
101
41
97
68
74
57
74
59
66
66
-40
-30
-30
0
-35
+20
-41
-61
_ I
-37
-28
-34
-17
-34
-19
-26
-26
8.9
6.0
2.3
1.8
2.1
2.3
6.7
1.1
4.1
4.3
4,0
3.8
2.9
3.6
3.7
4.5
2.9
(8.2)
(5.1)
(1.5)
(1.2)
(1.5)
(1.8)
(5.1)
2.0
4.1
2.8
3.1
3.3
2.8
2.5
2.8
3.6
2.1
(82)
(51)
(1.5)
(1.2)
(1.5)
(1.8)
(5.1)
1.1
3.6
1.8
2.2
3.1
1.8
1.6
1.8
3.1
1.1
4
<1
<1
<1
5
<1
<1
0
1
0
0
0
0
0
8
8
8
Cloud
Height,
m
Net
Radiation
Index
Solar
Radiation
W/m2
Wind Stability
Azimuth
St.Dev.
OA' (rad.)
Elevation
St.Dev.
aE'(rad.)
High
Haze
High
Haze
Haze
Haze
Haze
Haze
Low
Low
150
4
3
_2
-2
1
2
3
2
2
2
1
1
2
2
1
1
-1
(930)
(800)
(17)
(17)
(340)
(520)
(710)
290
430
410
105
170
(510)
(450)
180
126
35
0.24
0.18
0.48
0.38
0.45
0.42
0.18
0.23
0.24
0.30
0.21
0 10
0.27
0.28
0.26
0.16
025
0.052
0.065
0.051
0051
0.065
0.093
0.065
0.093
0.075
0.093
0.052
0.052
0.093
0.093
0.052
0.052
0051
Microbiological Aerosol Runs
54-77
5-5-77
5-5-77
5-13-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
1547-1617
1538-1608
1700-1730
1625-1655
1807-1837
1923-1953
2044-2114
1538-1608
1557-1627
2032-2102
290
260
275
265
240
175
145
275
251
215
-70
-40
-55
-45
-20
+45
+75
-55
-31
+5
5.1
4.7
3.0
5.9
2.5
5.7
5.4
3.2
5.4
5.5
(4.3)
(3.7)
(2.3)
(5.0)
(1.8)
(4.5)
(4.2)
(2.5)
(4.6)
(4.3)
4.3
3.7
2.3
5.0
1.8
4.5
4.2
2.5
4.6
4.3
2
3
4
0
0
0
0
0
1
1
Haze
Low
Low
3
3
2
3
2
1
-2
3
3
^
(730)
(750)
(570)
(700)
(370)
(112)
(17)
(840)
(750)
(17)
0.49
0.21
0.31
0.21
0.59
0.21
0.28
0.63
0.35
0.38
0.065
0.093
0.075
0.065
0.093
0.047
0.038
0.093
0.065
0.038
Pa squill
Stability
Class
3
3
7
7
3
2
3
3
3
3
4
4
3
3
4
4
5
3
2
3
3
2
4
5
2
3
5
-------
Table VI.B-l.(cont'd)
Aerosol Runs
Run
No.
Run
Date
Time of
Run
Mean Wind Direction,
deg
Relative
to
True North
Relative to
Perpendicular to
Spray Line
Wind Velocity, m/sec
10m at
Effluent
Pond
4m at
Met
Tower
2m in
Spray
Field
Radiation
Cloud
Cover
in Eighths
Cloud
Height,
m
Net
Radiation
Index
Solar
Radiation,
W/m2
Wind Stability
Azimuth
St. Dev
OA (rad.)
Elevation
St Dev
oE'(rad )
Pa squill
Stability
Class
Micro biological A erosol Runs (con t 'd)
Pre-Fair (cont'd)
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
Ml-32
Ml-33
Ml-34
Ml-35
Post-Fair
M2-1
M2-2
M2-3
M2-4
M2-5
M2-6
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-16-76
6-17-76
1-21-77
1-25-77
1-25-77
1-30-77
1-30-77
2-9-77
2-23-77
2-24-77
2-24-77
2-28-77
3-14-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-26-77
3-27-77
3-27-77
4-11-77
2214-2244
1717-1747
2040-2110
2150-2220
2117-2147
2010-2040
1420-1450
1635-1705
2140-2210
2242-2312
1600-1630
1415-1445
2012-2042
1421-1451
1603-1633
1853-1423
1835-1905
1403-1433
1513-1543
2003-2033
1625-1655
1555-1625
1616-1646
1620-1650
1344-1414
1315-1345
1634-1704
1330-1400
1518-1548
1515-1545
210
180
200
210
188
150
265
275
230
35
72
87
68
68
68
159
239
78
75
285
229
289
65
266
99
86
274
270
254
268
+10
+40
+20
+10
+32
+70
-45
-55
-10
+5
-32
-47
-28
-28
-28
+61
-19
+38
+35
-65
^9
-69
+25
-46
+59
+46
-54
-50
-34
-48
1 6
4.8
8.7
4.5
3.9
1.8
4.5
6.9
3.0
3.1
1.3
5.1
3.0
(2.3)
(2.3)
1.9
1 6
1 8
4.6
1.7
2.8
3.4
4.1
4.5
6.0
3.6
2.4
6.3
3.3
4.0
(1.1)
(3.7)
(6.8)
(33)
(1 1)
(0.5)
(38)
(5.6)
(0.9)
(1.1)
1.7
25
1.9
20
20
20
2.7
4.2
4.1
3.6
3.2
4.7
4.5
6.0
5.6
3.7
44
6.2
4.7
3.5
1.1
3.7
6.8
3.3
1.1
o'.s
3.8
5.6
0.9
1.1
1.3
3.1
2.0
2.0
2.0
1.3
1.8
3 1
34
2.2
1.6
3.6
4.9
4.5
4.9
27
3.4
5.4
1.8
3.1
0
1
1
<1
<1
<1
3
3
<1
-------
Table VI.B-l.(cont'd)
Aerosol Runs
Run
No.
Run
Date
Time of
Run
Mean Wind Direction,
deg
Relative
to
True North
Relative to
Perpendicular to
Spray Line
Wind Velocity, m/sec
10m at
Effluent
Pond
4m at
Met.
Tower
2m in
Spray
Field
Radiation
Cloud
Cover
in Eighths
Cloud
Height,
m
Net
Radiation
Index
Solar
Radiation.
W/m*
Wind Stability
Azimuth
St.Dev.
OA' (rad.)
Elevation
St.Dev.
°E'(rad.)
Pa squill
Stability
Class
Microbiological A erosol Runs (con t 'd)
Post-Fail (cont'd)
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
4-12-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
1615-1645
2106-2135
2010-2040
1640-1710
2050-2120
1410-1440
1630-1700
1335-1405
1505-1535
268
170
168
264
175
215
251
251
222
-48
+50
+52
-44
+45
+5
-31
-31
_2
7.8
4.0
(3.0)
5.4
4.0
(5.2)
(2.7)
7.2
5.5
6.9
1.5
2.2
7.2
1.9
4.8
4.1
6.9
3.4
5.6
0.9
(1.7)
2.9
1.1
4.2
2.2
6.0
3.6
2
<1
<1
<1
<1
4
3
8
8
6100
6100
4600
4600
2
-2
-2
2
-2
2
2
2
2
490
28
35
430
<70
850
450
490
290
0 16
0.23
0.23
0.17
0.54
0.29
0.19
0.16
0.73
0.065
0.051
0.051
0.047
0.051
0075
0075
0.065
0.075
4
7
6
4
6
3
3
4
3
Quality Assurance Aerosol Runs
Pre-Fair
Ml-16-19
Ml -20-23
Ml -24-25
Ml-26
Ml -27 -30
Post-Fair
M2-8-9
M2-18-21
M2-27-28
6-7-76
6-8-76
6-9-76
6-10-76
6-13-76
2-16-77
3-22-77
4-5-77
2104-2134
1750-1820
2259-2329
2212-2242
1707-1737
1805-1835
1821-1851
1639-1709
250
255
260
190
260
178
155
263
-30
-35
-40
+30
-40
+42
+65
-43
2.9
5.5
1.3
2.8
4 9
(1.4)
4.4
(2 1)
1.6
2.0
1.1
2.2
5.1
1.7
2.3
3.1
1.6
2.0
1.1
2.2
5.1
0.9
2.7
1.8
7
5
<7
<7
7
5
1
0
High
5500
-1
1
-1
-1
1
-1
-2
2
14
390
14
14
460
21
21
480
0.38
0.42
063
0.52
0.31
0.23
0.28
0.25
0.051
0047
0051
0.051
0.047
0.051
0.051
0.075
6
4
6
5
4
6
6
4
Virus Aerosol Runs
Post-Fair
V2-1.1
V2-1.3
V2-1.4
V2-I.5
V2-I.6
V2-H.1
2-25-77
2-26-77
2-26-77
2-26-77
2-26-77
4-9-77
1753-1823
1505-1535
1545-1615
1624-1654
1703-1733
1450-1520
253
281
274
270
260
266
-33
-61
-54
-50
-40
-46
(3.2)
(3.1)
(3.7)
(2.8)
(3.3)
(6.9)
5.0
2.7
33
2.5
3.0
6.2
26
2.5
3.0
2.3
27
5 7
8
1
2
3
3
3
High
High
High
High
High
1900
-1
1
1
1
1
3
35
570
350
290
84
710
0.14
0.33
0.26
0.19
0.19
0.23
0.038
0.065
0.052
0.065
0.052
0.065
4
4
4
4
4
4
-------
Table VI.B-1. (cont'd)
Aerosol Runs
Run
No
Run
Date
Run
Mean Wind Direction.
deg
Relative
to
True North
Relative to
Perpendicular to
Spray Line
Wind Velocity, m/'sec
1 Om at 4m at 2rn in
Effluent Met. Spray
Pond Towei Field
Radiation
Cloud
Cover
in Eighths
Cloud
Height,
m
Net
Radiation
Index
Solar
Radiation,
W/m2
Wind Stability
Azimuth
St. Dev.
aA'(rad.)
Elevation
St. Dev
Og'(rad.)
Pasquill
Stability
Class
Virus Aerosol Runs (cont'd)
Post-Fair (cont'd)
V2-II.2
V2-H.3
V2-II.4
V2-II.5
V2-II.6
4-9-77
4-9-77
4-9-77
4-9-77
4-9-77
1530-1600
1610-1640
1655-1725
1735-1805
1815-1845
273
273
169
268
268
-53
-53
+51
-48
-48
(7.3) 6.6 6.1
(80) 7.2 66
(6.9) 61 5.6
(6.1) 5.4 4.9
(57) 50 4.6
3
3
2
<1
<1
1900
1800
1800
1700
2
2
2
1
1
670
520
360
200
63
0.18
0.19
0 16
0.16
0.16
0.065
0047
0.065
0.047
0047
4
4
4
4
4
OO
o
-------
Fifty-two microbiological aerosol runs were attempted during the Post-Fair sampling period.
Three were quality assurance aerosol runs and two were special virus aerosol runs, which involved a total of
eleven 30-minute sampling periods. During these quality assurance and virus runs, all samplers were located
side-by-side at a distance of 50 meters from the wet-line edge of the spray line. Eighteen microbiological aero-
sol runs had to be aborted when meteorological conditions violated the aerosol run criteria. Twenty-nine of
the Post-Fair aerosol runs met the sampling protocol criteria. Ten dye aerosol runs were also successfully
completed during the Post-Fair sampling period. The environmental conditions which existed at the time of
the Post-Fair dye and microbiological aerosol runs also appear in Table VI.B-1.
The environmental conditions included in Table VI.B-1 represent measured, calculated, and
estimated values. In addition to the time and date, eight descriptive environmental measurements are in-
cluded. Five of these parameters were measured at several locations in and around the spray field, as indicated
by the table headings and as located on the site map, Figure V.A-1. Because the statistical analysis required
complete data sets, substitution rules were developed to replace missing values with the best available esti-
mate. The substituted values have been placed in parentheses.
The mean wind direction was obtained from the 10-meter level of the effluent pond station in
Pre-Fair and from the 4-meter level of the meteorological tower in Post-Fair. The direction in degrees is given
with respect to true north, as well as relative to a line perpendicular to the spray line used for sampling. Wind
velocity in meters per second is presented in Table VI.B-1 for three locations: (1) at the 10-meter level near the
effluent pond station, (2) at the 4-meter level at the meteorological tower, and (3) at the height of 3-meter in
the spray field. Both the wind direction and the velocity shown in Table VI.B-1 represent average values as
determined from strip charts that were recorded over the period of each microbiological aerosol run. Those
values of wind velocity measured in the field are considered the most accurate reflection of wind conditions
affecting the aerosolization of the sprayed wastewater. The wind velocity measured at the 10-meter effluent
pond tower is the least representative. Whenever a wind velocity was not available, it was calculated from the
most representative value which was available by the relationship:
i, \ P
where u2 and u, are the respective wind velocities at heights h2 and h,, and p is the wind profile exponent as
determined by H. E. Cramer Co. for each of the Pre- and Post-Fair runs. No adjustment for measuring loca-
tion differences was attempted other than the height adjustment.
Four parameters relating to solar radiation are included in Table VI.B-1. Solar radiation was
measured by a short-wave instrument on each Post-Fair run. The following regression equation, which has a
coefficient of multiple determination, r2 = 0.877, was developed from the Post-Fair run data and was used to
predict solar radiation for the Pre-Fair runs.
(sin(SA)xcc \
)
where R = solar radiation, W/m2
SA = solar altitude, deg.
cc = fractional cloud cover
h = cloud height factor, when cc is not equal to zero, where
h = 1 for low clouds
h = 2 for middle clouds
h = 3 for high clouds
h = 10 for haze
81
-------
Solar altitudes were calculated for each run at the Pleasanton location from the date of the run and the time
of day halfway through the run. Cloud cover factors were estimated for each run by the field sampling crew.
The standard deviations of the azimuth wind direction (OA) and of the elevation wind angle
(°E) were determined from the wind direction range, wind speed, and net radiation index, as described by H.
E. Cramer Co. ((24><25)). The H. E. Cramer meteorologists also estimated the Pasquill stability class for each
run from these data. The FP rotorod data was used to confirm the field measurements of wind direction mean
and standard deviation (OA) for the Post-Fair runs.
The air temperature for each Pre- and Post-Fair run is given for four locations in Table
VI.B-1. Although differences exist between the sets of air temperature values, those values measured at one
location tend to correlate well with values measured at another location. This fact was useful in developing
regression equations which allowed the calculation of missing temperature measurements. Missing values of
air temperatures upwind (TU) and downwind (TD) of the spray line were calculated from air temperatures
measured at the effluent pond (TP) using the following regression equations:
TU = 0.82 TP -I- 4.0 (r2 = 0.523)
TD = 0.73 TP -I- 4.2 (r2 = 0.885)
Effluent pond air temperatures were directly substituted for missing values of air temperature at the meteoro-
logical tower (TM).
Table VI.B-1 also contains relative humidity values for four locations in the vicinity of the
spray fields. In addition, relative humidity values were obtained from the Lawrence Laboratory at Livermore
which coincide with the day and time of each run. As was the case of air temperature, relative humidity mea-
sured at different locations within a general local area tend to differ in a predictable pattern and therefore can
be correlated. In order to predict missing values of relative humidity at the effluent pond (RHP) and down-
wind of the spray line (RHD), regression equations were developed which related those parameters with va-
lues of relative humidity which were measured at the meteorological tower (RHM) and upwind of the spray
line (RHU), respectively. The following are the regression equations which were used to predict missing va-
lues:
RHP = 0.427 RHM + 61 (r2 = 0.865)
RHD = 0.887 RHU + 11 (r2 = 0.925)
Values of relative humidity measured at the meteorological tower (RHM) and those values reported by the
Lawrence Laboratory at Livermore (RHL) were generally in agreement. Therefore, when one value was miss-
ing, the corresponding value at the other location was substituted. It was also determined that relative humid-
ity measured just upwind of the spray line averaged 15 percentage points higher than values reported for the
Livermore location; therefore, missing values for the upwind location were approximated by adding 15 per-
centage points to the corresponding Livermore value: RHU = RHL + 15.
Another measurement included in Table VI.B-1 is wastewater temperature, which was re-
corded at the spray line. Temperatures were observed to vary seasonally and by time of day. In order to calcu-
late missing values, a two-step procedure was used: (1) measured values of wastewater temperature were nor-
malized to the hottest part of the day (1-4 P.M.) and (2) a graph was drawn of the normalized temperature as
a function of the date on which it was measured. It was possible to construct a graph that spanned the annual
range since Pre-Fair runs were conducted in late spring and Post-Fair runs were conducted in the winter and
early spring. Missing values of wastewater temperature were then read from the graph for the corresponding
date of the run, and the value was corrected for the actual time of day during which the run was conducted.
82
-------
b. Spray Line and Sampler Configurations
The location and general configuration of the Pre- and Post-Fair microbiological, dye, and
quality assurance sampling runs and the Post-Fair virus trial runs are given in Tables VI.B-2 through Vl.B-5.
The spray fields and staked locations of the distant samplers are shown in Figure V.A-1. The line of spray
heads that was rotated daily from one setting to the next through the lettered spray field (A, B, C and D) in a
28-day cycle was used for most of the aerosol sampling. Another line of spray heads was rotated daily through
the numbered fields (0, 1, 2, 3 and 4). The side of the sprayer line on which the samplers were located during a
run was designated as wet (SW of the sprayer line; still wet from the preceding days' irrigation) or dry (NE of
the sprayer line).
For the first four weeks of the seven-week Pre-Fair sampling period, only seven high-volume
aerosol samplers were available for use in the microbiological aerosol sampling. One or more of these sam-
plers were often inoperative because of equipment failure or microbiological contamination. The sampling
protocol called for eight high-volume samplers on each run; therefore, when it was necessary, the second sam-
pler of the side-by-side sampler pairs was excluded. This is because the five quality assurance runs would pro-
vide better measurement variation estimates than could be obtained from the paired sampler data.
The sampler configurations for the Pre-Fair aerosol runs are shown in Table VI.B-3. The
standard Pre-Fair distance configuration of the sampler line was used on Runs 1 and 3. A paired sampler con-
figuration was employed on Runs 4 and 5, to assess the basic dye level variability at a given downwind dis-
tance due both to inherent variability in the sampled aerosol and to sampler variation. A modification of the
standard configuration was used on Runs 6 and 10 when it became apparent that sampling within 10 meters of
the wet-line edge yielded low dye concentrations. Configuration A was utilized for Run 9. While sprayer lines
were also operating in field 2 on Runs 3 to 6, this had no effect on the dye experiment; the dye was only in-
jected into the sprayer line directly upwind from the samplers. As shown in Table VI.B-3, the standard Post-
Fair dye sampler configuration was used for all of the Post-Fair dye aerosol runs.
For the Pre- and Post-Fair quality assurance aerosol runs and for the Post-Fair virus trial
runs, the samplers in each run were set side-by-side about 3 meters apart (1 meter apart in Pre-Fair) at a cer-
tain distance from the wet-line edge. In the Pre-Fair quality assurance runs, the sampler lines were set at dif-
ferent distances from the wet-line edge with the distances recorded in Table VI.B-4. For the Post-Fair quality
assurance aerosol runs and the virus trial runs, the sampler lines were set at 50 meters from the wet-line edge
as shown in Tables VI.B-4 and VI.B-5.
Concentration patterns and sampler locations for the Pre-Fair runs appear in Appendix B of
the H. E. Cramer Report #TR-76-303-03 (<26>). Concentration patterns and sampler configurations for the
Post-Fair runs appear in Appendices F and G of the H. E. Cramer Report #TR-77-309-01 (<27>).
For the Pre-Fair dye and microbiological aerosol runs, source strength profiles are included
in Appendix C of the H. E. Cramer Report #TR-76-303-03 (<28<). The individual spray head flow rates are tab-
ulated in Appendices B and C of the H. E. Cramer Report #TR-77-309-01 (<29>) for the Post-Fair dye, microbi-
ological and virus runs.
2. Sampled Concentration Data
a. Dye Runs
The wastewater and aerosol sample dye concentrations obtained on the Pre- and Post-Fair
dye runs are presented in Tables VI.B-6 and VI.B-7, respectively. These tables also contain the perpendicular
distance of each AGI sampler from the wet-line edge. For the source dye concentration, wastewater samples
were taken before and after a run, and the two values obtained were averaged to calculate the source dye con-
centration mean.
83
-------
Table VI.B-2.
SAMPLING CONDITIONS FOR MICROBIOLOGICAL AEROSOL RUNS*
Pre-Fair
Aerosol
Run
No.
Ml-1
Ml -2
Ml -3
M14
Ml -5
Ml -6
Ml -7
Ml-8
Ml-9
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml -15
Ml-31
Ml-32
Ml-33
Ml -34
Ml-35
Sprayer Line
Source
Sampled
Field
B
B
B
C
C
C
C
D
3
3
3
3
3
3
B
C
D
D
D
D
Setting
1
2
2
3
3
7
7
4
1
1
2
4
4
4
3
7
1
1
2
3
Location ot
Sampler Line
Side of
Sprayei
Line
Dry
Dry
Dn
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Wet
Position
(facing
spra> ei
line)
Left
Left
Lett
Lett
Lett
Right
Right
Left
Center
Center
Right
Rlghl
Center
Center
Right
Center
Left
Left
Right
Center
Sampler
Line
Configuration
Standard
Standard
Standard
Standard
Standard
Standard
Standard
Standard
Standard
Mod Std
Mod Std
Mod Std
Standard
Standard
Mod Std
A
A
B
B
A
Other Fields
Operating
Field-Setting
2-5
2-6
2-6
E-2
t-2
0-6 F-l
0-6 F-l
F-l
G-l
D-7.G-!
\-\ G-2
A-3.G-4
A-3
A-3, G-4
1 -1 . H-3
D-6. 1-6
D-7, 1-7
0-7,1-7
1-7
4-2. t-2
*See pages 26-28.
Post-Fair
Run
No,
M2-1
M2-2
M2-3
M2-t
M2-5
M2-6
M2-10
M2-1 1
M2-12
M2-13
M2-14
M2-15
MM6
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
Sprayer Line
Source
Sampled
Field
C
D
D
D
D
B
D
D
D
D
B
C
C
C
D
D
n
D
D
B
C
C
C
D
D
D
D
D
D
Setting
4
1
1
6
6
2
2
3
3
7
7
2
3
4
4
5
5
6
6
7
1
2
3
1
1
4
5
6
6
Location of
Sampler Line
Side of
Sprayer
Line
Wet
Wet
Wet
Wet
Wet
Dr\
Dr>
Wet
Wet
Dry
Dry
Dry
\\et
Dry
Wet
\\et
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Drv
Dry
Dry
Dry
Dry
Position
(facing
line)
Left
Left
Center
Center
Left
Right
Left
Right
Right
Left
Center
Left
Center
Left
Right
Right
Lett
Left
Left
Left
Left
Rich!
Right
Left
Right
Center
Left
Left
Left
Distant
Sampler
Lines
E
E
E
F
E
D
C.l.K
G,E
G.E
G,I,K
B
K.H
E
K.H
G.E
G.E
K,I
1,G,C
I.G.C
H.B.K
K.H.G
D
D
K.I.G.C
D
C,D
K.I.G
I.C
I.C
Near
Rotored
Lines
L.E
F
r
F
F
A
H.B
F.G
F.G
H.B
A
M.A
L
M
F.L
F.L
H.B
B.H
B.H
M,A
M,A
•\
A
H.G
B
B
H.G
H.G
H.G
Other
F iclds
Operating
1,1
4,1
4.1
4.E.F
4E,F
2,G
4,1
4
3 E.F
l.H
1 ,H,\VLF*
1,H,\VLF
l.l.WLF
4.E F.WLF
4,E.r,WLF
4,E F.WLF
4.E.F.WLE
4.E.F.WLF
l.H
l.H
1 H
l.H
4,1
4,1
4,E,F
4.E.F
4.E.F
4,E,F
Other Spray
Fields Contributing
to Aerosol
4
I
E,F
E.F
1
E.1
EF
4
'West Lake Field
-------
Aerosol dye concentrations versus downwind distance are plotted in Figure VI.B-1 for each
of the seven Pre-Fair dye runs. The mean downwind distance from the wet-line edge used as the abscissa in
these plots was obtained by dividing the perpendicular distance in Table VI.B-6 by the sine of the mean wind
angle 9.
Some interesting observations can be made by scanning Figure VI.B-1. The relationship of
aerosol dye concentration to downwind distance appears to differ substantially over the seven runs. On each
of dye Runs Dl-6 and Dl-9, there is an obvious decrease of dye concentration with increasing distance from
the wetline edge. It might be noted that these were the two dye runs located in the center sampler position. The
other dye Runs (Dl-1, Dl-3, Dl-4, Dl-5 and Dl-10) do not exhibit a consistent trend in the dye concentration
with increasing downwind sampler distance.
Table VI.B-3.
SAMPLING CONDITIONS FOR DYE AEROSOL RUNS
Pre-Fair
Dye
Run
No.
Dl-1
Dl-3
D14
Dl-5
1)1 -d
Dl-9
Dl-10
Sprayer Line
Source Sampled
Field
B
2
2
2
i
c
D
Setting
1
1
4
4
5
7
2
Location of
Sampler Line
Side of
Sprayer Line
Dry
Dry
Dry
Dry
Dry
Wei
Dry
Position (facing
sprayer line)
Left
Left
Left
Left
Co nl ci
Center
Left
Sampler
Line
Configuration
Standard
Standard
Paired
Paired
Mod Sid
A
Mod Std
Post-Fair
Run
No.
D2-1
D2-2
D2-4
D2-5
D2-6
D2-7
D2-8
D2-9
D2-10
D2-11
Sprayer Line
Source Sampled
Field
D
D
D
D
B
B
B
B
B
B
Setting
3
6
7
7
2
3
3
3
3
3
Location of
Sampler Line
Side of
Sprayer Line
Wet
Wet
Wet
Wet
Wet
Wet
Wet
Wet
Wet
Wet
Position (facing
sprayer line)
Center
Center
Center
Center
Center
Center
Center
Center
Center
Center
Sampler
Line
Configuration
Dye (50, 75, 100)
Dye (50, 75, 100)
Dye (50. 75, 100)
Dye(50,75, 100)
Dye (50, 75, 100)
Dye (50, 75, 100)
Dye (50, 75, 100)
Dye (50, 75, 100)
Dye (50, 75, 100)
Dye (50, 75,100)
85
-------
The plots for Runs 1, 3, 4 and 9 disclose one or more apparently erratic sampler points
within 50 meters downwind of the wet-line edge for which the sampled dye concentration differs appreciably
from the distance pattern suggested by the other sampler points. It appears that dye samplers should not be
located closer than 50 meters from the dye aerosol source.
b. Microbiological Runs
All of the adjusted and footnoted microbiological concentration data obtained on the 20
Pre-Fair aerosol runs and the 29 Post-Fair aerosol runs are presented in the tables of Appendix F. Each Ap-
pendix F table contains all of the analysis results (from every high-volume sample on all the runs for either
Pre-Fair or Post-Fair) for a single type of analysis. The tables are formatted according to the downwind sam-
pler distances from the wet-line edge that were used in the aerosol run sampler configuration. A list of foot-
notes, special data symbols, and microbiological units precedes the tables in Appendix F. When sample con-
tamination affecting the microorganism concentration result was inferred, the analysis result has been
underlined. The quality of the footnoted data must be evaluated taking the footnote(s) into account. Special
data symbols accompany or replace the usual data value in the indicated circumstances. The units identify the
type of assay procedure used to obtain the quantitative results.
A 30-minute composite wastewater sample was taken from a spray head in the sprayer line
during each aerosol run. This sample was analyzed for four microorganism groups in Pre-Fair and for all five
of the Post-Fair microorganisms. These results are shown in Appendix F, in the third column of the Pre-Fair
Table VI. B-4.
SAMPLING CONDITIONS FOR QUALITY ASSURANCE AEROSOL RUNS
Run
No.
Sprayer Line
Source
Field
Sampled
Setting
Location of
Sampler Line
Side of
Sprayer Line
PRE-FAIR
Ml-16-19
Ml-20-23
Ml-24-25
Ml-26
Ml-27-30
C
C
C
C
C
1
2
3
4
7
Dry
Dry
Dry
Dry
Dry
Position (facing
sprayer line)
Center
Center
Center
Center
Left
Distance
from
Wet Line
Edge, m
25
30
25
20
40
POST-FAIR
M2-8-9
M2-1 8-2 1
M2-27-28
C
D
B
2
1
1
Dry
Dry
Dry
Center
Right
Center
50
50
50
Table VI.B-5.
SAMPLING CONDITIONS FOR VIRUS AEROSOL RUNS
POST-FAIR
Run
No.
Sprayer Line
Source
Field
V2-I.3-6 D
V2-II.1-6 B
Sampled
Setting
Location of
Sampler Line
Side of
Sprayer Line
5 Dry
5 Dry
Position (facing
sprayer line)
Distance
from
Wet Line
Edge, m
Left 50
Left 50
86
-------
Table VI.B-6.
PRE-FAIR DYE AEROSOL RUN CONCENTRATION DATA
Dye
Run
No.
Dl-I
Wastewater Data
Source Per pi
Dye Dicta
Concentration Wet t
((IK/1)
Mean: 91.700
Before: 93,000
After: 90.333
Aerosol Sampler Data
rndicular Aerosol
nee from Dy*
ine Edge Concentration
(m) (ue/m3)
5 1.77
10 4.06
15 3.18
20 15.89
50 3.18
100 1.94
Dl-3
Mean: 85, 000
Before: 86,000
After: 84, 000
5 1.29
10 1.82
15 2.61
20 1.71
20 1.94
50 1.65
100 1.77
100 1.71
Dl-4
Dl-5
Mean: 96, 500
Before: 97.000
After: 96,000
Mean: 72, 000
Before: 76,000
After: 68, 000
10 2.88
10 2.35
20 2. 47
20 2. 00
30 1.35
30 0. 88
40 2. 77
40 2. 82
15 2.94
15 3.06
25 3.35
25 3.06
35 3.71
35 3.77
45 3. 24
87
-------
Table VI.B-6. (cont'd)
Wastewa
Dye Sou
Run Dy
No. Concen
(UK
Dl-6 Mean:
Before:
After:
Dl-9 Mean:
.Before:
After:
Dl-10 Mean:
Before:
After
ter Data Aerosol Sampler Data
rce Perpendicular
e Distance from
tration Wet Line Edge
tl) (m)
96,000 10
100,000 15
92, 000
20
20
30
50
100
100
17, 100 10
10
18,000
16,200 20
3,0
40
40
50
50
18,800 10
20,600 20
17,000 20
30
40
50
100
100
Aerosol
Dye
Concentration
(UE/m3)
9.89
9.77
7.47
9.12
8.06
4.35
2.77
5.28
2.29
1.82
1.00
1.71
1. 12
1.06
0.82
1.00
0.13
0. 12
0.13
0.18
0.18
0.23
0. 18
0.18
88
-------
Table VI.B-7.
POST-FAIR DYE AEROSOL RUN CONCENTRATION DATA
Dye
Run
Mo.
DZ-1
Wastewater Data
Source Perp<
Dye Dicta
Concentration Wet L
(UB/1)
Mean: 21,000
Before: 20,500
After: 2-1,500
Aerosol Sampler Data
;ndicular Aerosol
ace from Dye
ine Edge Concentration
(m) (ue/m3)
50 0. 53
50 0. 59
50 0. 53
SO 0. 47
75 0. 29
75 0. 26
100 0. 14
100 0.12
D2-2
Mean: 21,300
Before: 21,900
After: 20, 600
50 0.67
50 0.69
50 0.62
50 0. 58
75 0.41
75 0. 47
100 0.34
100 0.39
D2-4
Mean: 27, 400
Before: 27,200
After: 27,600
50 0. 56
50 0. 54
1
50 0. 46
50 0. 48
75 0. 29
75 0.29
100 0. 19
100 0.20
D2- 5
Mean: 26.100
Before: 26,800
After: 25, 400
50 0.62
50 0. 52
50 0. 54
50 0, 42
75 0.29
75 0.25
100 0.18
100 0. 16
89
-------
Table VI.B-7. (confd)
Dye
Run
No.
Wastewater Data
Source
Dye
Concentration
Aerosol Sampler Data
Perpendicular
Distance from
Wet Line Edge
(tn)
Aerosol
Dye
Concentration
(uK/m3)
D2-6
D2-7
D2-8
D2-9
Mean: 22.100
Before: 21.200
After: 23,000
Mean: 25, 600
Pefore: 18.600
After: 32. 500
Mean: 31.000
Before: 29. 500
After: 32, 500
Mean: 25. 500
Before: 23,900
After: 27, 000
50
50
50
50
75
75
100
100
50
50
50
50
75
75
100
100
50
50
50
50
75
75
100
100
50
50
50
50
75
75
100
100
0.49
0.65
0.38
0.45
0.16
0.12
0.07
0.15
0.62
0.35
0.29
0.48
0.26
0.28
0.19
0.16
0.65
0.34
0.51
0.53
0.39
0.24
0.26
0.13
0.33
0.32
0.36
0.47
0.17
0.19
0.12
0.12
90
-------
Table VI.B-7. (confd)
Dye
Run
No.
Waatewater Data
Source
Dye
Concentration
(ue/D
Aerosol Sampler Data
Perpendicular
Distance from
Wet Line Edge
(m)
Aerosol
Dye
Concentration
(ue/m3)
D2-10 Mean: 28.700 50
50
Before: Z8, 800
After: 28, 500 50
50
75
75
100
100
D2-11 Mean: 27,400 50
50
Before: 27, 500
After: 27.300 50
50
75
75
100
100
0.12
0.14
0.22
0.15
0.18
0.13
0.09
0.08
0.19
0.17
0.18
0.25
0.14
0.15
0.20
0.17
91
-------
Figure VI.B-1.
PLOTS OF AEROSOL DYE CONCENTRATION WITH DOWNWIND DISTANCE
Sampled Dye
Concert tr a t ion,
12-
10-
Run Dl-1
4-- O
2- .
-4-
"Too
46 6b 80 1<)0 ISO
Mean Downwind Distance From Wet Line Edge, m
ife"
92
-------
Figure VI.B-1. (continued)
Sampled Dye
Concentration,
l»g/m3
6--
Run Dl-3
2--
_L
40 60 80 ibO 120
Mean Downwind Distance From Hot Line Edge, m
140
Sampled Dye
Concentration.
yg/ra3
4--
Run Dl-4
2-
o
o
o
o
0
o
H-
"Tbo120
40 00 CO ibO 120
Mean Downwind Distance rrom Wet Line Edge, m
140
93
-------
Sampled Oye
Concentration,
A
4--
2--
Figure VI.B-1. (continued)
8
It"
-4-
45 6b 80
Mean Downwind Distance From Wet Line Edge, m
Run Dl-5
150140 Ifco
Sampled Dye
Concentration,
14 "|"
12--
10--
8--
6- -
4--
2--
Run Dl-6
406080Tbo120itc
Moan Downwind Distance Prom Wet Line Edge, m
94
-------
Sampled Dye
Concentration.
wg/m3
4- -
Figure VI.B-1. (continued)
Run Dl-9
2..
O
O
8
O
O
It-
12
Mean Downwind Distance Prom Net Line Edge, m
Sa.i-.plcd Dye
Concentration,
pg/o3
6--
4--
Run Dl-10
2-.
o . o
40 fib 80 160 120
Mean Downwind Distance From Met Line Edge, m
ifc
00
140
95
-------
tables for total coliform, fecal coliform, standard bacterial plate count, and coliphage, and on all of the Post-
Fair tables.
The upwind sampler was placed at an appropriate upwind sampler location based on wind
direction before the run. For Pre-Fair, there were three preselected locations in elevated wooded areas; the
distance of the upwind sampler from the configured downwind samplers ranged from 1400 meters to greater
than 3200 meters, with a median upwind distance of 1800 meters. During Post-Fair, there were five upwind
sampler stations whose elevation, terrain and land usage were similar to the downwind sampler positions. The
upwind sampler was also located closer to the configured downwind samplers and occasionally was located to
one side of the samplers depending on the wind direction.
Within 100 meters of the wet-line edge, the aerosol concentrations of standard bacterial plate
count, total coliform, fecal coliform, coliphage, fecal streptococci, Pseudomonas, and Clostridium perfring-
ens were generally quantifiable above the minimum detection limit. Because the Post-Fair wastewater concen-
trations of mycobacteria were lower than anticipated, mycobacteria were seldom found in the aerosol samples
on Runs M2-1 to M2-26 above the detection limit. However, in Runs M2-29 to M2-37, when the detection
limit was lowered by a factor of 30, mycobacteria were found in almost every sample. Hence, the aerosol con-
centrations of these eight microorganism groups warrant statistical analysis.
Fourteen biochemical confirmation tests for KlebsielJa were conducted on every mucoid col-
ony from the aerosol samples on the 20 Pre-Fair microbiological aerosol runs. Through this exhaustive effort,
Klebsiella was found in only four of the aerosol samples: Klebsiella ozaenae on Run Ml-8 at 5 meters or Run
Ml-10 at 50 meters and on Run Ml-14 at 50 meters; and Klebsiella pneumonias on Run 12 at 20 meters. After
confirmation testing for viruses, no positive three-day virus plaque counts were obtained from the Pre-Fair
aerosol run samples. Only two of these aerosol samples had positive five-day virus plaques: the 10 meter sam-
pler on Run Ml-9 and the upwind sampler on Run Ml-13. The discovery of Klebsiella and viruses in the aero-
sol is certainly important. However, the Klebsiella, three-day virus, and five-day virus aerosol concentrations
are not suitable microorganism groups for the statistical analyses.
An examination of the Pre-Fair and Post-Fair aerosol concentration values with distance
across a microbiological aerosol run in the quantitative aerosol data tables reveals a definite downwind dis-
tance pattern: a reduction in concentration with increasing downwind sampler distance. The microbiological
concentration reduction with distance is more pronounced than the dye concentration reduction with distance
exhibited in Figure VI.B-1. There clearly are microbiological decay factors in addition to the aerosol forma-
tion and dispersion factors. Also, it appears that microorganism die-off accumulates with distance from the
source or aerosol age. The microbiological aerosol levels on the nighttime and evening runs were generally
higher than on the afternoon runs. This suggests confirmation of the Phase I finding that solar radiation is a
very significant microbiological decay factor. This topic will be given a more rigorous statistical treatment in
Section VI.C.5.
As discussed in Appendix D, smoothing of the microbiological aerosol run data was nec-
essary to make it amenable for developing the microbiological dispersion model. The smoothing procedures
described in Appendix D (designating background and downwind samplers, computing the background con-
centration B, computing the downwind concentration Cd at sampler distance d, and excluding unusable runs)
were applied to the Appendix F data for the eight microorganism groups that warrant statistical analysis. The
resulting smoothed data for the microbiological aerosol runs are presented in Tables VI.B-8 through VI.B-15,
with each microorganism group appearing on a separate table.
c. Quality Assurance Runs
The purpose of the aerosol quality assurance runs in Pre-Fair and Post-Fair was to determine
whether there were systematic differences or biases among the high-volume aerosol samplers used in the aero-
96
-------
VO
-J
Table VI.B-8.
SMOOTHED STANDARD BACTERIAL PLATE COUNTS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run
Number
Ml-1
Ml -2
Ml -3
Ml -5
Ml -6
Ml -7
Ml -8
Ml -9
Ml -10
Ml -11
Ml -12
Ml-13
M1-15
Ml -31
Ml -32
Ml -33
Ml -34
Ml -35
M2-1
M2-3
M2-4
M2-6
M2-12
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-29
M2-31
M2-32
M2-34
M2-35
M2-38
Run
Date
5-4-76
5-5-76
5-5-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-16-76
6-17-76
1-21-77
1-25-77
1-30-77
2-9-77
2-24-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-26-77
3-27-77
4-11-77
4-13-77
4-14-77
4-19-77
4-22-77
4-24-77
Hastewater
Concentration
(No. /100ml)
no x io«
120 x 10'
no x io«
167 x 10s
76 x 10«
46 x 10s
70 x 10'
104 x 106
145 x 10'
35 x 10s
34 x 10s
26 x 10'
45 x 10'
16 x 10s
20 x 10'
95 x 10'
54 x 10'
76 x 10'
15 x 10'
113 x 10«
27 x 10«
7. Ox 10«
2.1x 10s
73 x 10«
105 x 10*
76 x 10'
2.2x 10'
20 x 10«
41 x 10'
26 x 10'
44 x 10«
67 x 10'
75 x 10s
15 x 10«
6.9x 10'
15 x 10«
Aprn«nl Concentrations (No./m3 of air)
Background
B
140
2000
680
1200
380
620
100
1450
1600
840
150
64
2100
220
570
680
1600
300
310
1560
650
85
90
180
38
130
47
39
48
53
370
365
380
930
145
210
Downwind Concentrations Crf at sampler ui stance a
5m 10m
9800 1400
6700 6000
4900
2600 2100
1700 1600
2000 2000
1000
6300 3000
4500
3000
850
3500
1500
5200 1000
20m
830
8600
5800
2000
3600
1700
3200
2100
2000
2580
3000
1200
5600 5000
1650
2900
3TJm 4um 50m
5500
1500
340 360
1800
2200 2000
2000 2200
750
790
3800 2000 2700
3800 6300
5300 3300
1600 1900 1300 1300
830
1820
1280
480 290
95
1400 330
1100 1300
950 720
60 72
66
190 190
310 270
1300 1900
1000
520
4800 2300
170
2100
10UD
6500
1300
1400
580
1400
830
4500 3800
92 500
500
380
150 83
2400
56
97 110
60
98 150
1700 3200
500 530
520
1900
500 430
870 1200
Distant
Sampler
Distant Distance
980 (200m)
430 (220m)
130 (290m)
360 (340m)
300 (435m)
-------
Table VI.B-9.
SMOOTHED TOTAL COLIFORM CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run Run
Number Date
Ml -2
Ml -3
Ml -4
Ml -5
Ml -6
Ml-7
Ml -8
Ml -9
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
Ml -32
Ml -33
Ml -34
Ml -35
vo
00 M2-1
M2-2
M2-3
M2-4
M2-5
M2-6
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-24
M2-25
M2-26
M2-29
H2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
5-5-76
5-5-76
5-13-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-16-76
6-17-76
1-21-77
1-25-77
1-25-77
1-30-77
1-30-77
2-9-77
2-23-77
2-24-77
2-24-77
2-28-77
3-14-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-27-77
3-27-77
4-11-77
4-12-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
Wastewater
Concentration
(MFC/1 00ml)
2070 x 103
2140 x 10»
690 x 103
720 x 103
930 x 10J
700 x 103
950 x 103
1040 x 103
1280 x 103
1100 x 103
470 x 103
690 x 103
750 x 103
588 x 103
970 x 103
265 x 103
170 x 103
550 x 103
350 x 103
60 x 103
280 x 103
450 x 103
260 x 103
340 x 103
220 x 10s
410 x 103
340 x 103
360 x 103
390 x 103
420 x 103
690 x 103
1100 x 103
1200 x 103
860 x 103
740 x 103
730 x 103
670 x 103
1600 x 103
640 x 103
450 x 103
1200 x 103
390 x 103
540 x 103
360 x 103
410 x 103
450 x 103
410 x 103
Background
B
0.15
0.15
0.07
0.07
0.1
0.1
0.1
0.15
0.5
0.1
0.1
0.15
0.1
0.1
0.05
0.03
0.08
0.08
0.1
0.025
0.02
0.025
0.01
0.02
0.05
0.1
0.02
0.02
0.05
0.01
0.01
0.02
0.1
0.02
0.05
0.01
0.017
0.01
0.01
0.05
0.01
0.02
0.45
0.025
0.02
0.01
0.01
Aerosol Concentrations (MFC/m3 of air)
Downwind Concentrations C. at Sampler Distance d
5m 10m
9.2 10.5
12.7 8.1
1.7 2.1
5.8 1.4
6.5 4.6
4.3 1.2
10.8 15.5
18.0 38
57
43
2.0
3.2
13.2
0.53 0.67
19.3 30.5
20m 30m
17.2 20.0
10.9 12.6
0.2
0.2
7.7
5.2
1.1
93 46
46 53
6.5
11.7
47
2.2
4.7 2.6
0.45
0.75
1.7 2.3 3.8 0.8
12.0 6.6
40m 50m
8.3
3.0
0.45
5.2
32
8.3
1.4
6.2
14.5
2.1
0.4 1.5
1.3
5.1 2.0 2.6
6.7 7.5
2.0 1.5
37 31
6.5 5.8
4.8 5.5
2.1 1.9
1.3 5.5
3.1 2.2
0.3 2.9
1.7 2.7
13.0 10.8
5.3 4.5
9.7 12.2
5.7 7.7
2.0 2.0
1.8 0.6
2.2 3.8
0.3 0.4
0.1
1.2 0.7
7.8 7.5
2.0 2.3
14.0 13.5
6.7
2.0
1.3
1.7 2.5
1.1 0.7
100m
5.1
0.5
0.6
0.5
4.4
4.9
0.15
0.15
4.3
4.1
0.9 2.0
0.4
0.4 1.6
15.2 10.5
2.0 1.3
2.6 3.1
0.2 0.2
4.7 5.3
0.9 1.3
1.3 0.2
2.8 3.3
1.1 1.3
1.6 1.1
4.3 1.8
5.7 2.4
0.4 0.2
0.1
0.35
0.025
2.0 0.6
4.8
1.5 1.5
0.2
2.0 1.7
0.1
0.25
0.025
Distant
Sampler
Distant Distance
2.6 (200m)
2.2 (200m)
0.3 (360m)
0.2 (290m)
0.2 (340m)
0.2 0.2 0.2(435/505/580m)
0.1 (655m)
-------
Table. VI.B-10.
SMOOTHED FECAL COLIFORM CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run
Number
Ml-2
Ml -3
Ml -5
Ml -6
Ml -7
Ml -8
Ml -9
Ml-10
Ml-11
Ml -15
Ml -31
Ml -32
Ml -34
Ml -35
Run
Date
5-5-76
5-5-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
6-3-76
6-14-76
6-15-76
6-16-76
6-17-76
Wastewater
Concentration
(MFC/lOOml)
186 x 10'
174 x 103
180 x 10s
75 x 10'
110 x 10'
80 x 10'
81 x 10'
125 x 10'
124 x 10'
57 x 10*
137 x 10'
24 x 10'
81 x 10'
45 x 10'
Background
B
0.02
0.05
0.02
0.075
0.075
0.02
0.075
0.3
0.06
0.02
0.05
0.02
0.075
0.05
5m
2.0
0.9
1.1
1.1
0.6
0.8
3.1
Aerosol
Concentrations (MFC/m3 of air)
Downwind Concentrations C. at Sampler Distance d
10m
2.1
1.7
0.7
0.3
0.75
2.3
9.5
6.6
1.8
2.9
0.15
4.9 4.4
20m
4.4
3.0
0
1
2
12
6
1
0
0
0.4 1
9
2.1
1.0
.1
.4
.0
.2
.2
.9
.6
.33
.0 0.6
.9
30m
3.8
1.3
0.2
0.6
0.25
1.2
40m 50m
1
0
1
5
0
0
0.2
0
0.6 0.4
.2
.075
.2
.6
.5
.5
0.2
.7
0.5
100m
O.'Z
0.25
0.2
0.5
0.1
-------
Table VI.B-11.
SMOOTHED COLIPHAGE CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run
Number
Ml-3
M1-5
Ml -6
Ml-7
Ml -8
Ml-9
Ml-10
Ml -11
Ml-12
Ml-13
Ml-14
Ml-15
Ml -31
Ml -32
Ml -33
Ml -35
M2-1
M2-2
M2-3
M2-4
M2-5
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
Run
Date
5-5-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-17-76
1-21-77
1-25-77
1-25-77
1-30-77
1-30-77
2-23-77
2-24-77
2-24-77
2-28-77
3-14-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-26-77
3-27-77
3-27-77
4-11-77
4-K-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
Wastewater
Concentration
(PFU/1)
95 x 10'
170 x 10'
220 x 10"
230 x 10s
61 x 10»
95 x 103
130 x 10'
140 x 103
110 x 103
170 x 10'
210 x 10s
310 x 10'
330 x 10s
320 x 10'
240 x 10'
170 x 10*
180 x 10'
94 x 10'
160 x 10'
120 x 10'
170 x 103
300 x 10'
390 x 103
230 x 10'
310 x 10s
190 x 103
120 x 103
140 x 10'
190 x 10'
170 x 10'
380 x 10'
700 x 10*
1200 x 10'
930 x 103
400 x 10'
180 x 10*
240 x 10'
240 x 10'
450 x 103
200 x 10'
1200 x 10'
780 x 103
480 x 10'
430 x 103
Background
B
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.075
0.1
0.025
0.0
0.0
0.0
0.0
0.0
0.025
0.075
0.0
0.0
0.0
0.025
0.0
0.0
0.033
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Aerosol Concentrations (PFU/m3 of air)
Downwind Concentrations Cri at Sampler Distance d
otn I Oin 20m ^Hm AOm cAm *""
— — — __ __ "••• *-""' jwiu fUlll DUIII
0.15 0.2
0.05
1.0 0.6 0.8
0.3 Q 025
0.7 0.4 o.l 0.3 n'n«i
11 01 17 no
1 • ' w.l I .£. no
2.0 3.7 1.1 1.4 2 6
2-3 0.8 1.4 1.1 o.'6
OQ n 7 n o n f.
•y u. / u.o o.2
^ •! 1.0
3-3 0 8
0.1 0.7 0.3 o!3
07 flfi n 7 n ni <*
• ' v • u u . / 0013
0.025 0.2 0.1 0.013
0.1 0.4 0.2 0 025
!-8 0.8 1.1 0.5 0.2 0.6
0.4 0.3
0.7 0.4
0.8 0.4
0.8 0.7
0.4 0.1
0.1 1.1
0.4 0.7
0.3 1.1
1.1 1.3
2.8 2.1
0.3 0.4
0.9 0.7
0.5 0.5
1.1 0.4
0.15
1.6 2.1
2.5 2.2
0.15
0.7 0.4
0.1 0.3
0.8 0.8
0.8 0.4
0.3 0.025
0.1 0.2
2.6 2.8
0.4 0.1
0.4 0.4
0.4 0.4
^ f\ *.'--'
100m
0.1
0.05
0.1
0.3
0.2
0.4
0.1
0.7
0.6
0.2 0.1
0.5 0.3
0.2
0.2
0.15
0.3
0.9 1.2
0.3 0.1
0.4 0.1
1.2 0.9
0.8 0.5
0.1 0.3
0.55
0.1 0.4
0.05
0.05
0.2 0.7
0.7 0.3
0.05
0.15
0.8 0.1
0.013
0.4
0.3 0.2
1.1 1.1
0.3 0.1
0.1 0.4
0.3 0.7
Distant
—„_ Sampler
Distant Distance
0.2 (200m)
0.2 (200m)
0.3 (220m)
0.1 (340m)
-------
Table VI.B-12.
SMOOTHED FECAL STREPTOCOCCI CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run
Number
Ml -6
Ml-8
Ml-9
Ml-10
Ml -11
Ml-12
Ml-13
Ml-14
Ml-15
Ml -31
Ml-33
Ml -34
Ml-35
M2-1
M2-2
M2-3
M2-4
M2-5
H2-6
H2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-24
M2-25
M2-29
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
Wastewater
Run Concentration
Date (CPU/ 100ml)
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
6-15-76
6-16-76
6-17-76
1-21-77
1-25-77
1-25-77
1-30-77
1-30-77
2-9-77
2-24-77
2-28-77
3-14-77
3-16-77
3-17-77
3-18-77
3-25-77
3-26-77
3-27-77
4-11-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
6.5
20
8.0
8.0
2.8
6.8
6.8
6.8
2.3
17.1
9.2
5.5
4.0
9.7
4.8
1.3
6.0
12.0
3.3
1.2
5.7
2.0
7.0
12.0
7.2
4.3
10.0
11.0
4.3
6.2
3.0
2.2
1.9
0.63
3.0
1.6
0.93
x
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
103
103
103
103
103
103
103
103
103
10'
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
103
Aerosol Concentrations (CFU/m3 of air)
Backg
B
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
0.
0.
0.
0.
0.
2.
0.
0.
0.
0.
0.
. Downwind Concentrations C, at Sampler Distance d
round d
5m 10m 20m 30m
43 0.8 3
15
15 6.0 1.6 0.7
15 4.2 4.0 2.0 2.7
6 3.1 5.7 2.7 2.7
1 15 30 25
3 1.3 1.7
1 14
3 2.1 3.0 2.8
5 3.3 1.9 3.3
1 0.5 0.3 0.7
6 0.8 1.8 4.3 5.1
3 4.7 4.4 2.7 2.4
4
3
0
2
3
1
01
1
0.03
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
01
01
01
03
03
025
025
77
16
01
3
01
17
01
1
40m 50m
0.25
4.7
1.5
13
8.5
90
1.8
3.2 2.1
0.55
1.2
3.3 6.5 2.4
1.0 1.0
1.4 0.4
3.3 7.7
1.0 0.7
1.3 1.3
0.4 0.4
0.025
0.4 2.1
0.3
0.1 0.2
0.8 0.8
0.6 0.3
0.2
0.2 0.4
0.7 0.9
0.85
0.35
0.05
0.4
0.2
0.3 0.3
0.6 0.6
4.9 5.1
100m
0.
3.
10
0.
26
2.3
1.8
0.9
8.3
0.
1.0
0.2
1.1
0.8
0.4
0.6
0.2
0.8
0.3
0.
0,
3.7
25
1
,6
1.
0.
0.
7.
,25
1.
0.
0.
0.
0.
0.
0.
0.
0.
,1
.175
1.
8
45
7
3
1
6
4
6
7
1
4
4
6
8
Distant
6.8
0.4
3.7
0.1
0.1
0.1
0.1
0.1
0.1
0.2
1.6
0.
0.
7.
3
025
0
0.2
0.2
Distant
Sampler
Distance
(200m)
(300m)
(360m)
(490m)
(280m)
(290m)
(420m)
(425m)
(435m)
-------
Table VI.B-13.
SMOOTHED PSEUDOMONAS CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
o
N>
Aerosol
Run
Number
Ml -4
Ml -5
Ml -6
Ml -7
Ml -8
Ml -9
Ml-10
Ml -11
Ml-13
Ml-15
Ml -31
Ml -32
Ml-33
Ml -34
Ml -35
Wastewater
Run Concentration
Date (CFU/lOOml)
5-13-76
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
6-3-76
6-14-76
6-15-76
6-15-76
6-16-76
6-17-76
80 x 10'
80 x 103
30 x 103
30 x 103
210 x 103
35 x 103
35 x 103
120 x 103
30 x 103
200 x 103
30 x 103
25 x 103
25 x 10s
30 x 10s
20 x 103
Background
B 5m
4 190
2 140
2 430
4
15 240
4 140
4
2
4
2
2
1 4
1
15
3.5
Aerosol Concentrations
(CFU/m3 of air)
Downwind Concentrations Cd at Sampler Distance d
10m
13
1300
750
290
26
570
250
20m 30m
33
330
4
380
235
150
83 68
165
17
9 8
4.5
91 76 180
210 100
40m 50m 100m
51
6 77
77
37
220 210
100 30
230 15
8 9
9 9
100
260 89 61
-------
Aerosol
Run Run
Number Date
Ml-5
Ml-6
Ml-7
Ml-8
Ml-9
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
5-13-76
5-17-76
5-17-76
5-21-76
5-24-76
5-24-76
5-25-76
5-27-76
5-27-76
5-27-76
6-3-76
6-14-76
Table VLB-14.
SMOOTHED CLOSTRJDIUM PERFRINGENS CONCENTRATIONS BY SAMPLER
DISTANCE FROM MICROBIOLOGICAL AEROSOL RUNS
Wastewater
Concentration
(MPN/lOOml)
11,000
4,600
4,600
48,000
24,000
24,000
4,600
4,300
4,300
4,300
4,300
9,300
Background
Herosoi concentrations ^nc«/m
or airj
Downwind Concentrations C. at Sampler Distance tf
B 5m 10m
0.06 5.
0.06 2.
0.06 1.
0.06 0.
0.06 1.
0.06
0.06
0.06
0.06
0.06
0.06
0.9
1 6.0
4 1.0
2
8
9 1.8
2.3
1.0
2.3
0.9
2.7
20m 30m 40m 50m
3.0
1.3
3.3 2.0
1.3
3.0 1.3 1.9
0.95
3.3 5.0
1.3 2.3
1.3
1.3 1.1
1.9
1.2 0.8
0.23
0.5
1.0
1.0
2.3
0.33
1.2 1.2
100m
5.9
2.3
4.4
0.23
0.23
0.45
Distant
Sampl er
Distant Distance
1.8 (200m)
Table VI.B-15.
SMOOTHED MYOBACTERIA CONCENTRATIONS BY SAMPLER DISTANCE FROM
MICROBIOLOGICAL AEROSOL RUNS
Aerosol
Run
Number
M2-15
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
Run
Date
3-16-77
4-11-77
4-12-77
4-13-77
4-14-77
4-19-77
4-19-77
4-22-77
4-23-77
4-24-77
4-24-77
Wastewater
Concentration
(CFU/IOOml)
11.0 x
1.0 x
10.0 x
5.0 x
7.0 x
7.5 x
7.6 x
1.0 x
6.0 x
7.3 x
5.0 x
103
103
103
103
103
103
103
10s
103
103
103
Heros
loi concentrations vcru/m or air)
Backa o nd Downwind Concentrations C, at Sampler Distance
B 5m 10m 20m 30m 40m
0.2
0.5
0.1
0.4
0.2
0.1
0.15
0.4
0.05
0.35
0.2
11
0,
2.
4,
1
1,
11
d
50m
1.
0,
0.
,3
.3
0
.3
.7
.0
11
.15
,15
.5
2,
1.
.4
2
0
1.
11
.3
.0
.7
.7
.3
100m
0,
1.
2,
1
1
1
0,
.3
.7
.0
.7
.0
.0
1
,65
0.7
1.0
1.0
i.o
0.7
1.0
.3
-------
sol sampling runs and to quantify the sources of variation associated with sampling the microorganism group
concentrations.
Tables VI.B-16 through VI.B-23 show the concentration data gathered for Pre-Fair and
Post-Fair. The three-day and five-day virus counts were below the detection limit for all samples and the CIos-
tridium perfringens levels were also undetected; tables for these groups are not included. Analysis for fecal
coliform, Klebsiella, and Pseudomonas was performed only in Pre-Fair, and the results are shown in Tables
VI.B-18, VI.B-21 and VI.B-25. Table VI.B-22 shows the results from the analysis of mycobacteria which was
performed only in Post-Fair. Analyses for standard bacterial plate count, total coliform, coliphage, and fecal
streptococci were performed in Pre-Fair and Post-Fair periods. The results from these analyses are in Tables
VI.B-16, VI.B-17, VI.B-19 and VI.B-20. Also, on the Post-Fair Runs M2-8-9 and M2-27-28, the samples were
received by the laboratory at a temperature higher than the specified 4°C.
The 100 mL samples from each sampler on the Pre-Fair and Post-Fair quality assurance runs
were divided into portions (usually 25 mL) in the field and then given run numbers. Some of these portions
were then divided into smaller parts, called aliquots by the lab. The run numbers and portion numbers are
shown in the columns labeled as such. After analysis the data were converted to the measured unit per m3 of
air by the method described in Appendix D using the sampling time and the flow rate. The footnotes used in
the quality assurance tables are given in Figure VI.B-2.
d. Virus Runs
Two special virus runs were conducted during the Post-Fair sampling period. Table VI.B-24
shows all of the concentration data gathered from these two runs. Besides being analyzed for enteroviruses,
the wastewater and aerosol samples were also checked for standard bacterial plate count, total coliform, fecal
streptococci, coliphage, and mycobacteria.
The enterovirus concentrations were about three orders of magnitude lower than coliphage,
so a large volume of air had to be sampled. This required the use of high-volume aerosol samplers and special
sample concentrating procedures. The special enterovirus runs established the ability to detect enteroviruses
in wastewater aerosols and provided a quantitative measure'of their concentration coming from a known
wastewater aerosol source.
The identifications of the confirmed enterovirus isolates obtained from the aerosol samples
on the Post-Fair virus aerosol runs are presented in Table VI.B-25. During Pre-Fair, the single confirmed
viral isolate from the 10-meter downwind sampler on Run Ml-9 and both confirmed isolates from the upwind
sampler on Run Ml-12 were identified as poliovirus 1. Since poliovirus 1 was used in the assay laboratory to
determine the efficiency of the sample concentration procedure, the Pre-Fair poliovirus 1 isolates may be the
result of laboratory contamination and thus are not reported in Table VI.B-25. On virus aerosol Run V2-I,
the single confirmed three-day viral isolate was identified as poliovirus 2; of the three confirmed five-day en-
terovirus isolates, two were identified as poliovirus 2. Of the seven confirmed five-day enterovirus isolates
from virus Run V2-II, one was identified as poliovirus 1 and another was identified as coxsackievirus B-3.
The process of identifying the remaining confirmed five-day enterovirus isolates from Runs V2-I and V2-II
was terminated before completion because of laboratory reorganization. The variety of enteroviruses identi-
fied from the aerosol samples of the virus aerosol runs increases the likelihood that the sprayed wastewater
was the enterovirus source.
104
-------
Table VI.B-16.
STANDARD BACTERIAL PLATE COUNTS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol Wastewater
Run Run Sample
No. Date (No./iOOml
PRE-FAIR
Ml-16-19 6-7-76 14 X 10*
Ml-20-23 6-8-76 37X106
Ml -2 4-25 6-9-76 44 X 106
Ml -26 6-10-76 35 X 10"
Ml-27-30 6-13-76 12 X 106
POST-FAIR
M2-18-21 3-22-77 88 X 106
Sampler Distance Aerosol
From Wet Between Sample
) Line Edge Samples Portion No
25 m 1m
16
17
18
30m 1m
20
21
22
23
25 in 1m
24
20m 1m
26
40m 1m
27
28
29
30
50m 3m
18
19
20
21
Aerosol Concentrations
from Samplers Aligned from Left to Right
(No /m' of air)
16500'* 8000 6000 5000 3600 4000 3600 3300 2700
7500'' 2900 1700 2000 2500 2000 1900 1600 1900
6700d 1700 1900 2500
2800 1700 1400 2300 1800 2400 1600 2100
2700 1200 1500 2100 2000 2400 1700 2200
1100 1900 2300 2000
1300 1300 2000 1600
2600 1500 1200 2000 1300 >89000k- 1600 1900
1300 1400 1500 >89000k 1400 27000 1500 1500
4300 930 2000 900 1600 2100 740
4000 980 2100 1900 1600 12200 740
6900 1700 3800 1200
11000 4100 1400 2400
>10000P 280 900 330 1800 5300 530 830 330 600 830 570
270 870 470 2100 5000 570 1200 370 430 730 570
>10000P 530 1100 370 430 10000 500 1200 430 670 1200 700
670 1300 300 570 9300 500 1400 430 670 1100 570
MOO
M1 830 1600 670 320 3700 970 1200 570 1100 2900 830
570 1700 630 320 3000 1000 1300 670 1000 2800 830
M* 730 1900 670 430 2800 1500 1500 970 930 8000 1600
800 1700 670 330 1500 1400 1100 1100 8300 1300
-------
Table VI.B-17.
TOTAL COLIFORM CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol Wastewater
Run Sample
No. (MFC/100 m
Pre-Faii
Ml-16-19 6-7-76 343 X 10'
Ml-20-23 6-8-76 127 X 10'
Ml -24-25 6-9-76 1100X103
Ml-26 6-10-76 758 X 10'
Ml-27-30 6-13-76 267 X 103
Post-Fail
M2-18-21 3-22-77 673 X 103
Sampler Distance Aerosol
From Wet Between Sample
) Line Edge Samples Portion No.
25 m 1m
16
17
18
30m 1m
20
21
22
23
25 m 1m
24
20m 1m
26
40m 1m
27
28
29
30
50m 3m
18
19
20
21
Aerosol Concentrations
from Samplers Aligned from Left to Right
(MFC/m3 of air)
13.2s1 9.2 7.8 9.8 10.7 4.3 6.3 12.9 4.9
7.2d 15.0 9.1 9.6 5.5 6.7 10.8 5.1 9.3
8.0d 7.6 5.8 7.8
1.2 1.2 4.2 3.9 0.6 1.2 2.3 3.0
1.2 1.0 1.2 1.7 0.6 2.4 1.5 1.1
12 3.8 1.2 1.9
1.9 3.1 3.0 1.6
19.5 7.5 14.7 9.6 12.1 22.0 16.1 13.4
6.1 9.4 8.9 11.2 8.6 13.0 84 13.8
<0.5 <0.4 0.9 <0.5 <0.6 <0.6 0.7
<0.5 1.0 <0.5 13 <0.6 0.8 <0.5
<0.5 <0.5 1.0 0.4
0.7 <0.5 <0.4 <0.5
8.0P 13.3 8.3 7.0 67 19.3 10.0 12.7 9.0 8.3 5.3 7.7
8.3P 14.7 8.7 5.3 4.7 8.0 7.3 7.3 5.3 4.0
10.0? 8.3 127 100 9.7 10.0 90 87 8.7 14.7 11.3 8.0
10.0 14.7 14.7 13.3 10.0 8.7 9.3 14.0 8.7
tA- 18.0 15.0 12.7 9.0 113 12.0 93 10.7 9.7 12.0 10.0
21.0 14.7 15.3 12.0 107 10.0 11.3 12.7 13.3 10.7
M- 28.0 23.0 210 77 22.0 29.0 220 22.0 22.0 25.0 15.7
27.0 60 19.3 27.0 21.0 12.7
-------
Table VI.B-18.
FECAL COLIFORM CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol Wastewater S
Run Sample Frc
No. (MFC/100 ml) Lin
Pre-Fair
ampler Distance Aerosol Aerosol Concentrations
m Wet Between Sample from Samplers Aligned from Left to Right
eEdge Samples Portion No. (MFC/m3 of air)
Ml-16-19 6-7-76 60 X 103 25m 1m
16 1.8-j 13 0.8 24 2.2 0.8 0.3 03 0.9
17 2.
-------
Table VI.B-19.
COLIPHAGE CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol Wastewater S
Run Run Sample Fr
No. Date (PFU/8) Lir
PRF.-FA1R
ampler Distance Aerosol Aerosol Concentrations
am Wet Between Sample from Samplers Aligned from Left to Right
le Edge Samples Portion No (PFU/m3 of air)
High Sample
Temperature
Upon Lab Receipt
Ml -24-25 6-9-76 25 m 1 m
24 4.3 2.8 20 20 2.6 2.4 23 4.1
Ml-26 6-10-76 20m 1 m
POST-FAIR
26 1.4 1.8 08 12 14 20 2.0 1.3
M2-8-9 2-16-77 170X103 50m 3m
8 0.3 0.5 03 02 59 1.51 10 05 05
9 0.1 03 07 2.1 0.3
8°C
8°C
M2-27-28 4-5-77 550 X 103 50m 3m
27 04 0.4 09 05 07 02 03 0.7 10 <0 2 0.4 07
28 0.3 0.3 01 0.4 03 0.7
9°C
9°C
Table VI.B-20.
FECAL STREPTOCOCCI CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol Wastewater S
Run Run Sample Frc
No. Date (CPU/ 100 ml) Lir
PRE-FAIR
ampler Distance Aerosol Aerosol Concentrations
>m Wet Between Sample tlom Samplers Aligned from Left to Right
e Edge Samples Portion No (CFU/m3 of air)
Hign Sample
Temperature
Upon Lab Receipt
Ml -24-25 6-9-76 25 m 1m
24 1.8 3.2 0.6 1.1 1.2 2.7 1.5 0.6
Ml-26 6-10-76 20m 1m
26 06 03 0.3 0.3 03 0.8 1.1 <0 3
POST-FAIR
M2-8-9 2-16-77 5.0 X 10s 50 m 3 m .
8 0.4 14 >330 0.7 06 Of.- 1.0 1.0 07 2.4
9 0.4 04 11' 07 17
M2-27-28 4-5-77 50 m 3m
27 01 <0.1 1 7 0.4 07 11 01 0.2 0.2 <0 1 <0.1 0.1
28 02 06 33 03 01 <0.1
8°C
8°C
9°C
9°C
-------
Table VI.B-21.
PSEUDOMONAS CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol
Run
No
Run
Date
Sampler Distance
From \\et
Line Fdge
PRE-FA1R
Ml -24-25 6-9-76 25m
Ml -26 6-10-76 20m
Between
Samples
Aerosol
Sample
Portion No
Aerosol Concent r.uion^
from Sample:* Aligned from Left to Richi
(CFU, m3 of air)
1m
24 86 6810141 30 1 1 8 95 89
1m
26 190 110 120 30 210 120 160 240
Table VI.B-22.
MYCOBACTERIA CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol
Run R
No. D,
POST-FAIR
Wastewater
un Sample
itc (CFU/ 100 ml)
M2-8-9 2-16-77 1 3 X 103
M2-27-28 4-5-77 2.0 X 103
Sampler Distanc
From Wet Betwi
Line Edge Samp
e Aerosol
en Sample
es Portion No
Aerosol Concentrations
trom Samplers Aligned Irom I elt to Fight
(CFU/mJ of air)
High S imple
Tempera! ure
Upon 1 ah Receipt
50m 3 m
8
9a
9b
-------
3. Nature of Aerosol Data
a. Distributional Characteristics
This section provides descriptive statistics which help characterize the distributions of the
data for each of the microorganism groups. The descriptive statistics are given for values at specified dis-
tances because the data analysis is concerned with determining if the spray line is a source of the microorga-
nisms and what effect distance has on the concentrations of the microorganisms.
All the aerosol concentration data in the Appendix F tables are used except underlined (indi-
cating contamination) values, Klebsiella, three-day enterovirus, and five-day enterovirus from the Pre-Fair
data and Runs 1-26 and 38 for mycobacteria from the Post-Fair data. The data from these groups are not
included because very few values are above the detection limit. All data used have been lognormally trans-
formed (InC).
Table VI.B-23.
KLEBSIELLA CONCENTRATIONS FROM QUALITY ASSURANCE AEROSOL RUNS
Aerosol
Run
No.
Run
Date
Sampler
From Wet
Line Edge
PRE-FAIR
Ml-24-25 6-9-76 25m
Ml-26 6-10-76 20m
Distance
Between
Samples
Aerosol
Sample
Portion No.
Aerosol Concentrations
from Samplers Aligned from Left to Right
(CFU/m3 of air)
1m
24 2.5 0.3 2.6 4.5 2.0 Mk 0.5 1.4
lm v
26 0.3 1.2 2.2 Mk 0.6 3.5 1.3 1.7
Table VI.B-24.
MICROBIOLOGICAL CONCENTRATIONS ON VIRUS AEROSOL RUNS
Virus Run „. . . . . .
Aerosol Date Mcrob.ological
Run No. (Time) Parameter
V2-I 2-26-77 Enterovirus-3-day count
(1505-1733) Enterovirus-5-day count
Standard Plate Count
Total Coliform
Fecal Streptococci
Coliphage
Mycobacteria
V2-II 4-9-77 Enterovirus-3-day count
(1450-1845) Enterovirus_5.day count
Standard Plate Count
Total Coliform
Fecal Streptococci
Coliphage
Mycobacteria
Concentration
Wastewater
Sample
23 PFU/fi
45 PFU/e
1.7 X 10' /100ml
Aerosol Sample
(50m)
0.0036 PFU/m3
0.011PFU/m3
10,000,000 lm
3
190 X 103 MFC/100 ml 4.6 MFC/m3
4.6 X 103 CFU/100 ml <0.1 CFU/m3
470 X 103 PFU/e
0.4 PFU/m3
1.3 X 103 CFU/100 ml <11 CFU/m3
250 PFU/e
330 PFU/B
1.9 X 10" /100ml
<0.0025 PFU/m3
0.01 7 PFU/m3
430,000 /m3
64 X 103 MFC/100 ml 28 MFC/m3
1.0 X 103 CFU/100 ml 0.2 CFU/m3
370 X 103 PFU/2
1.5 PFU/m3
1.2 X 103 CFU/100 ml <11 CFU/m3
110
-------
(S,nC), skewness, and kurtosis. Skewness indicates clustering of the data; if the statistic is negative, the data
are clustered to be right of the mean; if positive, clustering is to the left of the mean; if 0 (zero), the data are
clustered about the mean. Kurtosis indicates peakedness of the data; a positive value indicates a distribution
that is more peaked or narrower than the normal distribution; a negative value indicates a flatter distribution;
the normal distribution has zero kurtosis.
The results of the analyses are given in Table VI.B-26 for Pre-Fair and Post-Fair data. For
skewness, most of the values are between -1 and 1. These are within the acceptable range for a normal distri-
bution. Small sample sizes (<25) can have a value slightly greater than + 1 and be within the acceptable range.
The acceptable range of a normal distribution for kurtosis is about -1.5 to 1.5 for small sample sizes (for sam-
ple sizes greater than 50, the range is -1 to 1); most values fall within that range. Large skewness and kurtosis
values for several of the groups (such as total coliform at upwind) can be accounted for by the large number
of values below the detection limit. Since the log-transformed data usually have a distribution that is approxi-
mately normal, the microorganism aerosol concentration data can be considered approximately lognormally
distributed.
b. Relative Prevalence
The relative prevalence of the microorganism groups in the wastewater and aerosol samples
for Pre-Fair and Post-Fair is indicated by the geometric means or lower detection limit in Table VI.B-27 and
by the ratio of the group geometric mean to total coliform geometric mean in Table VI.B-28. The geometric
Figure VI. B-2
FOOTNOTES FOR UNUSUAL EVENTS
Footnotes
a MPN method used instead of normal assay method.
b Excessive sampler arcing observed; no data adjustment made.
c Data adjusted for extreme sampler arcing.
d Data adjusted for nonstandard volume of air sampled.
e Data adjusted for sampler power supply problems.
f Data adjusted for loss of BHI sampler fluid by means other than evaporation.
g Possible sample contamination; too close to spray line.
h Possible sample contamination; loss or foaming over of BHI.
i Possible sample contamination; equipment malfunction (pump/hose readjustment, open top, tubing
problems).
j Possible sample contamination; external sources (vehicle traffic, train dust, wind shift, insects, smoke
bomb, cattle too close, operator upwind).
k Possible sample contamination observed by analysis laboratory.
1 Probable sample contamination; external source (nearby truck or dirt road).
m Probable sample contamination; equipment malfunction (glass view on plate broken, top opened, tub-
ing broken, contaminated intake).
n Certain sample contamination; improper cleaning procedures (clorox residual left).
o Certain sample contamination; improper collection procedures used.
p Certain sample contamination; sample dropped.
q No sample collected because sampler developed problems during the run.
r Sample amount was insufficient for analysis.
s No analysis performed.
t No analysis result reported because laboratory observed sample contamination.
u Analysis results lost.
v Elevated sample temperature after shipment (8-9°).
Ill
-------
means for the aerosol concentrations are listed by distance. Values preceded by a less than (<) sign are the
detection limit for the parameter and indicate that the geometric mean of the parameter at that distance is
below the detection limit. The ratio of microorganism geometric mean to total coliform geometric mean does
not include the upwind values. Values preceded by a less than sign were calculated using the microorganism's
detection limit.
For the Pre-Fair downwind data, Pseudomonas are 15 times as prevalent as total coliform in
the aerosol. Fecal streptococci and Clostridium perfringens are between one-third and one-half as prevalent
as total coliform in the aerosol. Fecal coliform and coliphage are less prevalent than any of the above three.
The Klebsiella aerosol concentration detection limit is slightly below the total coliform geometric mean. For
the Post-Fair data, mycobacteria is about one and one-half times as prevalent as total coliform in the aerosol.
Coliphage and fecal streptococci are about one-third and one-half as prevalent, respectively, as total col-
iform.
c. Systematic Sampler Differences
The testing of the high volume microbiological samplers for collection efficiency at NBL
provides a basis for evaluating whether there is need for sampler correction factors to adjust for any collec-
tion efficiency bias. The data from the nine runs, as shown in Appendix D, are adjusted for the actual flow
Table VI.B-25.
IDENTIFICATION OF CONFIRMED ENTEROVIRUS ISOLATES FROM AEROSOL SAMPLES
Post-Fair
Aerosol Run Number V2-I V2-II
Run Date 2-26-77 4-9-77
Sampler Distance 50m 50m
Three-Day Enteroviruses
Confirmed Isolates 1 0
Poliovirus 1
Poliovirus 2 1
Poliovirus 3
Five-Day Enteroviruses
Confirmed Isolates 3 7
Poliovirus 1 '
Poliovirus 2 2
Poliovirus 3
Coxsackievirus B-3 ^
Not Identified* 1 5
* Identification process terminated prior to completion.
112
-------
Table VI.B-26.
DISTRIBUTIONAL CHARACTERISTICS OF THE NATURAL LOG TRANSFORMED
MICROORGANISM GROUP CONCENTRATIONS
PRE-FAIR
Upwind
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Pseudomonas
Fecal Streptococci
Clostridium perfringens
5-20 m Downwind
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Pseudomonas
Fecal Streptococci
Clostridium perfringens
30-50 m Downwind
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Pseudomonas
Fecal Streptococci
Clostridium perfringens
100-200 m Downwind
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coirphage
Pseudomonas
Fecal Slreptofocci
Clostridium ;x; rl 'rmnens
N
18
19
16
18
15
16
17
43
53
47
53
38
52
49
29
29
29
31
25
32
23
16
17
12
17
1 1
14
l<>
laC
6.692
-1.357
-1.657
-2. 557
2. 516
-0.720
-0.672
7.861
1.744
0.003
-1.069
4.270
0.365
0.376
7.355
0.889
-1.004
-1. 441
3.525
0. 528
-0.086
6.784
0. 17')
-1.437
-1.72';
3.766
0.676
0. 065
Slnc
1.360
0.691
0.474
0.531
1. 159
0.935
0.061
0.683
1.335
1. 250
1. 136
2. 102
1. 175
0.940
.0. 877
1.504
1.074
1. 186
1. 555
1.239
0.619
0.832
1.205
0, 505
0.752
1.6'H
1. 5 U>
0. K91
Skewness
-0.280
2.393
0.264
1.486
1.805
1.244
2.373
0.200
-0.243
0.055
0.070
0. 168
0.732
-0.050
-0.982
0.268
0. 803
0. 195
0. 165
0. 904
1. 167
0. 092
-0. 168
0. 095
0. OS6
0.824
0.65^
0. 471
Kurtosis
-1.031
6.658 2
-0.777
0. 545
2.324
1. 392
3.633 2
-1.031
-0. 134
-0.82<>
-0. 940
-1. 138
-0. 226
-1. 112
1. 439
-0.481
0. 038
-1.323
-1.411
1. 494
1.662
1.253
-1.636
-0. 07
-0.657
0. 169
-0. 731
-1.205
1 - 3 is subtracted
2 - The high skcwncss and kurtosis for these parameters c:an be attributed to the num-
ber of runs below the detection limit.
113
-------
rates from the pin-wheel anemometer test, and an average value calculated for each sampler tested. These are
shown in Table VLB-29.
The sample variance among these sample means is S2 = 11,606.9. For comparison, the nine
values for Sampler 7 are used to calculate an S2 of 5,657.0, and an F-test gives a ratio of 2.05 with 7 and 8 df.
Table VI.B-26. (confd)
POST-FAIR
Upwind
Gtd. Bacterial Plate Count
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
50 m Downwind
Std. Bacterial Plate Count
X
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
100 m Downwind
Std. Bacterial Plate Count
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
Distant Downwind
Std. Bacterial Plate Count
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
N
24
25
28
29
9
45
56
56
55
18
48
56
55
57
18
76
85
81
85
27
IraC
5.711
-2.104
-2.857
-1.871
-1.190
6.104
.912
-.722
-1.077
-.219
5.803
-.441
-1.479
-1.243
-.201
5.865
-2.072
-2.778
-1.863
-.710
Lnc
1.417
.698
.434
1.363
.964
1.308
1.374
1.014
1.360
1.022
1.239
1.499
1.041
1.417
.966
1.404
. 524
. 554
1.213
.932
Skewness
-.220
3.249
3.038
1.058
.898
-. 103
-.666
-.371
.083
-. 177
.215
.059
-.057
.438
-.468
-.216
2.832
2.694
.780
-. 105
Kurtosis
-1.003
8.896
7. 867
.142
-.592
-1. 153
.295
-.088
-.643
-1. 106
-.763
-1. 180
-1.205
-.437
-. 113
-.505
7.983
6.231
-.460
-1.414
2
2
2
2
1 - 3 is subtracted
2 - The high skcwness and kurtosis for those parameters can bo attributed to the
number of runs below the detection limit.
114
-------
Table VI.B-27.
GEOMETRIC MEANS AND RATIOS OF WASTEWATER AND AEROSOL CONCENTRATIONS
OF MICROORGANISM GROUPS
Geometric Mean
Wa«te\vater Cone. (I/ml)
Pre Fair Run*
4' ' ' '
Standard , Plate Count
Total Coliform
Fecal Coliform
Cotlphage
Fecal Streptococci
Pseudomonai
KlebsielU
CloYtridium perfringens
3-Day Enltrowrui
5-Day Enterovirui
Post Fair Runs
£*•'.'• -
Standard PUte Count
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
3-Day Enterovirui
5-Day Enterovirui
699,000
7.500
800
2ZO
67
1.050
390
54
0.
0.
158.000
4.900
Z90
34
46
0.
0.
01Z
017
076
12
Geometric Mean Aeroaol Concentrations
Aerofol/Waitcwater Ratio
of Geometric Mean Concentration
Upwind
805
<0.5
<0. 3
<0. 1
<0. 6
-------
Table VI.B-28.
RELATIVE PREVALENCE OF MICROORGANISM GROUPS
Ratio of Group Geometric Mean to
Total Coliform Geometric Mean
Pre Fair Runs
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Fecal Streptococci
Pseudomonas
Klebsiella
Clostridium perfringens
3-Day Enterovirus
5-Day Enterovirus
Wastewater
(#/ml)/(MFC/ml)
93.4
1.000
1.106
0.030
0. 009
0. 140
0.052
0.007
0.000002
0. 000002
Downwind Aerosol
(#/m3)/(MFC/m3)
530
1.000
0. 174
0.061
0.47
15.2
<0.78
0.359
< 0.035
<0. 035
Post Fair Runs
Std. Bacterial Plate Count
Total Coliform
Coliphage
Fecal Streptococci
Mycobacteria
3-Day Enterovirus
5-Day Enterovirus
32.6
1.000
0.060
0.0070
0.0094
0.000016
0. 000024
788
1.00
0.35
0.49
1.41
0.0008
0. 006
Table VI.B-29.
MEAN NORMALIZED FLAVOBACTERIUM COUNTS, ADJUSTED FOR FLOW RATE (CFU/1)
Sampler
1 5 6 7 8 9 10 11
288 544 503 297 270 423 475 432
116
-------
This is not a significant value; thus, there is no indication of a need for a sampler correction factor.
The aerosol data for indicator microorganisms from the Pre-Fair quality assurance aerosol
runs are used to further investigate the need for a correction factor for a sampler bias. An ANOVA is per-
formed for each type of data on each run with samplers and portions as factors. Only the PEL analyses are
used, because the additional factor of laboratories could interfere with the analysis. The F-ratios are summa-
rized in Table VI.B-30.
For the total coliform and fecal coliform data, no significant F-ratios are obtained. The stan-
dard bacterial plate count data provide two significant F-ratios, however, and these are investigated further.
From Run Ml-20-23 a means separation technique is used to determine which samplers are
different. From this, it is determined that Sampler 1 is higher than all the others, which can be considered
equivalent. On Run Ml-27-30, a similar analysis shows that Samplers 8 and 10 are high, relative to the others.
However, some of these high values were subsequently identified as affected by sampler contamination by the
usual contamination inference procedure, so that the significant F-ratios may be due to contaminated sam-
plers.
From the above, the conclusion drawn is that no systematic bias exists among the samplers.
In general, variation among samplers is no greater than can be explained by sample-to-sample variation, and
where a sampler term is significant, the usual contamination inference procedure often deletes the value prior
to data analysis. Thus, there is no evidence of a need for a bias correction factor.
4. Aerosol Measurement Precision from Quality Assurance Program
Inspection of the dye and microbiological aerosol concentration data in Appendix F shows
considerable variation, even between paired samplers. Special quality assurance aerosol runs (Tables VI.B-16
through VI.B-23) were conducted to more carefully determine the uncertainty in an aerosol concentration
measurement for each microorganism group. After the quality assurance runs, several aliquots were some-
times analyzed from up to four separate portion bottles each containing the BHI sample collection fluid from
one of many samplers placed side-by-side at the same distance from the wet-line edge. Therefore, it is of prac-
tical value to determine the total uncertainty in an aerosol concentration measurement (measurement varia-
tion), that fraction of measurement variation retained in the portions (portion variation), and that fraction of
the portion variation retained in the aliquots (aliquot variation). Measurement variation results from differ-
ing levels of the microorganism sampled by the samplers, variation in sampler operator procedures, unde-
tected sampler contamination, bottle variation, shipping variation, analytical technique of the laboratory
Table VI.B-30.
SUMMARY OF F-RATIOS
Aerosol
Run No.
16-19
Analysis
Parameter
Total Coliform
Fecal Coliform
Plate Count
20-23 Total Coliform
Fecal Coliform
Plate Count
27-30 Total Coliform
Fecal Coliform
Plate Count
F
0.96
0.79
1.78
2.34
0.75
7.9
0.41
1.20
3.10
df
f, | f,
3
i*
8
7
7
7
6
6
6
i:»
13
13
16
16
16
15
15
15
Significance
Level
OK
OK
OK
OK
OK
< 0.005
OK
OK
0.05
117
-------
doing the analysis, and random error. Because portions were taken from the same sample, but submitted to
the laboratory on the quality assurance runs as being different, portion variation is due to bottle variation,
shipping variation, analytical technique variation, and random error. Aliquots are taken from the same por-
tion during sample preparation by the laboratory; therefore, aliquot variation is due to random error in the
analytical measurement process.
The appropriate random effects model for the full nested design (aliquots k within portions j
with samplers i) occurring on some quality assurance runs is
Xljk = M + A, + BJ + £ljk (1)
Sampler i = 1, ..., a A, --^ N (0, OA)
Portion j = 1, ...,b, ES "^ N (0, OB)
Aliquot k = 1, ..., m £ljk ^ N (0, a)
Then in terms of the preceding definitions, the measurement variation variance is Var(Xljk) = a\ + o| + a2,
while the portion variation variance is oj- + a2, and the aliquot variation variance is a2.
On most quality assurance runs, one aliquot of each portion from each sampler was analyzed
and reported separately. The standard analysis of variance of this nested design would yield mean square esti-
mates of the "between sampler" and "within sampler" sources of variation. The within sampler mean square
does estimate the variance due to portion variation, o2 + a2. However, the between sampler mean square
must be suitably adjusted and combined to yield an estimate of the measurement variation, o\ + o| + a2.
The analysis of distributional characteristics in Section VLB.3 suggests that the aerosol con-
centration data for each microorganism group may be considered to follow a lognormal distribution. When
replicate measurements are made at each of several levels i to estimate the true values ^ for a log-normally
distributed variable, then the true coefficient of variation ft = o/^ for each grouped data set is the same value
P ((30)). Thus, a coefficient of variation analysis ((31>) is used to estimate and report each type of variation.
In a coefficient of variation analysis, the coefficient of variation is estimated over k groups
of sizeni(x,,x2...,xn)as
k k
£= 2 w^./k^ 2 w£.Si/Xj)/k (2)
i = 1 i = 1
where x( = sample mean of group i = Zx/n
s, = sample standard deviation of group i = [S. (x — x,)2 /(n — 1)]'/2
a- = bias correction factor = (2/n,y/2r(n/2)/r[(nl— l)/2]
A
ft = sample coefficient of variation of group i = a, s/x,
w, = group weight ) Zeigler ((33)) tabulates these correction factors, which approach 1.0 as
the sample size increases: C2 = 1.253, C3 = 1.128, C4= 1.058, Cg= 1.036, and C20 = 1.013. The weights w, are
computed ((34)) to be proportional to the inverse of the variance of the group estimator ft = OjSj/x",. In a coeffi-
cient of variation analysis, the estimates ft are ordered by increasing group mean x~j and tabulated to check the
assumption that ft is independent of ^,.
In the aerosol measurement precision analysis, most groups consisted only of a single pair of
concentration measurements. To facilitate presentation of the analysis, similar pairs were combined via equa-
tion (2) and tabulated as a single group. In such cases, the arithmetic mean was taken as the group mean for a
118
-------
quality assurance run group (when the true mean was presumably the same), and the geometric mean was
used as the group mean for basically unrelated pairs (e.g., all Post-Fair pairs at 50 meters).
a. Dye Aerosol Concentrations
(I) Dye Analysis Accuracy
To ensure that the dye measurements made on both the wastewater and the aerosol dye
samples were representative of the true concentration, a study was conducted using simulated samples at
known concentrations. The samples were prepared using a basic stock solution of dye and the medium in
which the actual samples would be analyzed, i.e., deionized water for the aerosol samples, and wastewater for
the run composite samples. Samples were randomized to ensure that the analyst did not know at which level
he was working.
For the aerosol simulation, a total of 18 sample portions were prepared, with 2 blanks
and 4 portions at each of 4 concentrations: 1.8, 18, 72 and 360 ng/1. The values obtained are shown in Table
VI.B-31. For each level, the mean and standard deviation are calculated and a 95-percent confidence interval
constructed. The measurements are said to be accurate if the actual value falls within the confidence limits
shown.
As can be seen, the measurements satisfy the accuracy criterion above at all levels.
However, in all cases there is slight positive bias. On a relative basis, this ranges from 1.6 to 4.2 percent of the
true value, and averages 2.5 percent. However, this is within the limits of the precision of the determination.
Similarly, 4 portions of wastewater were prepared at an actual level of 59,500 to rep-
resent the run composite samples taken. These were submitted to the same analysis as the samples above and
the results are summarized in Table VI.B-32. The actual concentration was unknown to the analyst.
As in the previous study, the confidence interval contains the actual value of the sam-
ple, and thus can be said to be accurate. The bias is calculated as before and is similar to that found in the
simulated aerosol samples. The possibility thus exists that there is a slight, 2 to 3-percent, high-side bias in the
measurements, but with only 4 portions analyzed at a given level, there is insufficient information to ascertain
this using ordinary parametric tests.
A chi-square test can be used to give further information, however. If the differences
are calculated between each determined value and the actual value, it is noted that of the 20 non-blank deter-
minations, 3 is equal to, 2 are less than, and the remaining 15 are greater than the actual value. For the deter-
Table VI.B-31.
DYE ACCURACY RESULT SUMMARY, SIMULATED AEROSOL SAMPLES
Sample
No.
1
2
3
4
Mean
Std Dev
95% ci
Bias, %
0
0.0
0.0
-
-
0.0
0.0
-
0.0
1.8
1.5
2.0
2.0
2.0
1.88
0.25
(1.48, 2.28)
4.2
Actual Level, Og/U
18 72
18.5
18.5
18.5
18.0
18.38
0.25
(17.98, 18.78)
2.1
74.0
72
.5
74.0
72
73
1
(71.49
1
.0
.13
.03
, 74.77)
.6
360
360
361
380
370
367.75
9.32
(352.92, 382.58)
2.2
119
-------
Table VI.B-32.
DYE ACCURACY RESULT SUMMARY, SIMULATED
EFFLUENT SAMPLES (pig/1)
Sample
No.
A
B
C
D
Mean
Std Dev
95% O
Biis.%
Actual Value, 59
60,000
60,800
58,000
60,000
59,700
1,194
,500
(57,800,61,600)
3.4
minations to be accurate, approximately equal numbers should fall above and below the true value, or 8.5,
discounting the equal values. A test of significance is used to determine the likelihood of obtaining a 2 and 15
split if the expected number of each were actually 8.5. The chi-square statistic has a value of 9.94 with 1 de-
gree of freedom, and this gives a significance level of about 0.002. Thus there is evidence that the bias is real,
though slight. The indicated 2-percent positive bias in both the wastewater and aerosol dye concentrations will
have no net effect on aerosolization efficiency estimates because of cancellation (see equation (4) of Section
VI.C).
(2) Dye Measurement Precision
A preliminary scan of the paired values in the dye run plots of Figure VI.B-1 suggests
that the magnitude of the paired value variability is approximately proportional to the average of the paired
dye levels. When the sample standard deviation is proportional to the sample mean, the coefficient of varia-
tion approach is an appropriate analysis methodology. Coefficients of variation were computed to estimate
the measurement variation from the paired samplers on the dye runs. Separate estimates were calculated for
the Pre-Fair pairs, the Post-Fair sets at 50 meters (after adjusting for diffusion concentration D), the Post-
Fair pairs at 75 meters, and the Post-Fair pairs at 100 meters. The portion coefficient of variation was esti-
mated from the aerosol simulation in the dye accuracy study (Table VI.B-31).
The dye measurement and portion coefficients of variation are presented in Table
VI.B-33. The measurement coefficient of variation estimates appear to be nearly constant over the aerosol
concentration range of 0.1 to 2.5 Mg/m3 . A weighted dye measurement coefficient of variation of 0.17 = 17
percent was obtained using equation (2). The measurement coefficient of variation is considerably larger than
the portion coefficient of variation (5 percent). Most of the dye measurement variation appears to be due to
sampling factors (and perhaps to localized aerosol concentration anomalies), rather than to analytical factors.
b. Microbiological Aerosol Concentrations
Because no techniques of preparing microbiological samples of a known stable concentra-
tion were available, the accuracy of microbiological aerosol measurements could not be determined. The log-
normally distributed nature of the aerosol concentration data for each microorganism group indicates that a
coefficient of variation analysis is warranted to estimate the precision of the aerosol measurement process.
The precision has been determined for each microorganism group in terms of measurement variation, portion
variation, and aliquot variation, when appropriate, for each quality assurance aerosol run and for suitable
groups of paired samples from the microbiological aerosol runs.
All of the applicable aerosol concentration data, except those values inferred to have been
120
-------
affected by sample contamination, were used in calculating the precision variation. The single exception was
that the calculation of measurement and portion variation for standard bacterial plate count and total col-
iform on quality assurance Run M2-18-21 was only based on portions 18 and 19. The data for this run (see
Tables VI.B-16 and VI.B-17) show a general increasing trend with portion number for most samplers. The
portions were refrigerated until analysis was begun. Total coliform analyses were prepared over a three-hour
period in the morning and standard bacterial plate count analyses were prepared over a two-hour period in the
afternoon. The portion bottles were removed from the refrigerator at the beginning of the analysis period and
prepared sequentially starting with portion 18. The increased levels in standard bacterial plate count and total
coliform for portions 20 and 21 from each sampler (often exceeding a factor of 2) is apparently due to the
lengthened exposure to room temperature prior to analysis. It should be noted that the preparation duration
on M2-18-21, in which 48 aerosol sample portions were analyzed, exceeded those of the other quality assur-
ance runs and any microbiological aerosol run (typically, eight aerosol sample portions).
The microbiological aerosol concentration precision results are presented in Tables VI.B-34
(standard bacterial plate count), VI.B-35 (total coliform), VI-B-36 (fecal coliform), VI.B-37 (fecal strepto-
cocci), VI-B-38 (coliphage), and VI.B-39 (pathogens - Pseudomonas, mycobacteria, Klebsiella, and Clostri-
dium perfringens). When the measured aerosol concentrations are sufficiently above the detection limit, the
measurement variation of a microorganism group tends to have a constant coefficient of variation, regardless
of the run group mean. This can be seen especially for total coliform above 3 MFC/m3 in Table VI.B-35 and
for fecal streptococci above 0.6 CFU/m3 in Table VI.B-37. Nearer the detection limit, the precision coeffi-
cients of variation tend to be higher because the sampling resembles a Poisson process with few events per
interval.
Weighted estimates of the precision coefficients of variation calculated using equation (2) are
provided on the bottom line of the precision tables. These weighted estimates indicate the extent of variation
of each type for the microorganism group over the aerosol concentration range examined. Thus, for example,
the precision coefficients of variation for total coliform in aerosols over the range from 0.4 to 14 MFC/m3
were 0.50 for measurement variation, 0.49 for portion variation, and 0.15 for aliquot variation. This means
that the standard deviation of a total coliform aerosol measurement due to all sources of uncertainty in the
measurement process is about 0.50 = 50 percent of the true aerosol concentration (which is best estimated by
Table VI.B-33.
DYE AEROSOL CONCENTRATION PRECISION
MEAN COEFFICIENT OF VARIATION
MEASUREMENT PORTION
RUN GROUP (ug/m5) VARIATION VARIATION
Post-Fair at 100m 0.17 0.21
Post-Fair at 75m 0.25 0.15
Post-Fair at 50m 0.44 0.18
Pre-Fair Pairs 2.5 0.15
Dye Accuracy Study 0.05
WEIGHTED ESTIMATE 0.17 0.05
121
-------
MEASUREMENT
VARIATION
0.54
0.46
0.60
0.06
0.31
0.29
0.27
0.98
0.56
0.50
PORTION
VARIATION
0.36
0.12
0.49
0.59
0.37
ALIQUOT
VARIATION
0.11
0.11
Table VI.B-34.
STANDARD BACTERIAL PLATE COUNT AEROSOL CONCENTRATION PRECISION
MEAN COEFFICIENT OF VARIATION
RUN GROUP (NO./m3)
Post-Fair at 100m 330
Post-Fair at 50m 620
QA.-M2-18-21 740
QA.-M1-26 1430
Pre-Fair Pairs 1700
QA:Ml-24-25 1730
QA:Ml--20-23 1890
QA.-M1-27-30 2690
QA:M1-16-19 3100
WEIGHTED ESTIMATE
the concentration value obtained). In a portion, the standard deviation of the uncertainty in the analyzed total
coliform value is about 0.49 = 49 percent of the value. In an aliquot, the standard deviation of the uncertainty
in the analyzed value is about 0.15 = 15 percent of its value.
The weighted estimates for the microorganism groups are presented in Table VI.B-40; the
groups are arranged by increasing measurement coefficient of variation. The microbiological aerosol mea-
surement variation for each of the microorganism groups is much greater than the dye aerosol measurement
variation. The microbiological portion variations are also much greater than the dye portion variation. Thus,
there is much less precision in the microorganism aerosol concentrations than in the dye aerosol concentra-
tions. Grouping of the microorganisms by measurement variation in Table VI.B-40 shows better precision for
total coliform and standard bacterial plate count than for fecal coliform, Pseudomonas, and Clostridium per-
fringens. Aerosol measurement of mycobacteria, fecal streptococci, Klebsiella, and coliphage has the least
precision. Microbiological aerosol measurement precision seems to improve as expected when the mea-
surements consistently exceed the detection limit and when the more routine measurements are made.
By making the assumption that measurement variation and portion variation were calculated
for the same distribution of aerosol concentrations, one can quantify the contributions of analytical sources
of variation and of field sampling sources of variation to the total measurement variation. Under this assump-
tion, the square of the measurement variation equals the sum of the squares of portion variation (representing
shipping and analytical sources) and of sampling variation (representing field sampling sources). The sam-
pling variation thus obtained is shown in Table VI.B-40. Comparing the relative magnitudes of sampling vari-
ation and portion variation shows that the dye measurement variation is mostly due to field sampling sources.
122
-------
Table VI.B-35.
TOTAL COLIFORM AEROSOL CONCENTRATION PRECISION
MEAN
COEFFICIENT OF VARIATION
RUN GROUP
QA:Ml-27-30
Post-Fair at 100m
QA.-M1-20-23
Pre-Fair Pairs
Post-Fair at 50m
QA:M1-16-19
QA.-M2-13-21
QA:Ml-26
QA.-M1-24-25
(MFC/m3)
0.4
1.0
1.9
2.5
3.0
8.5
9.9
9.9
14.4
MEASUREMENT
VARIATION
0.8S
0.65
0.55
0.57
0.37
0.37
0.34
0.27
0.35
PORTION t ALIQUOT
VARIATION VARIATION
0.84
0.37
0.38
0.35 0.15
WEIGHTED ESTIMATE
0.50
0.49
0.15
Table VI.B-36.
FECAL COLIFORM AEROSOL CONCENTRATION PRECISION
RUN GROUP
QA:Ml-20-23
Pre-Fair Pairs
QA:Ml-46-19
QA:Ml-26
QA:Ml-24-25
MEAN
(MFC/m3)
0.3
0.8
1.7
1.7
4.0
COEFFICIENT OF
MEASUREMENT
VARIATION
0.47
0.46
0.84
0.27
0.64
VARIATION
PORTION
VARIATION
0.49
0.83
WEIGHTED ESTIMATE
0.58
0.65
123
-------
Table VI.B-37.
FECAL STREPTOCOCCI AEROSOL CONCENTRATION PRECISION
RUN GROUP
QA:Ml-26
QA:M2-27-28
Post-Fair at 100m
Post-Fair at 50m
QA:M2-8-9
QA:Ml-24-25
Pre-Fair
WEIGHTED ESTIMATE
MEAN
(CFU/m3)
0.5
0.5
0.6
0.6
0.9
1.6
2.0
COEFFICIENT
MEASUREMENT
VARIATION
0.74
1.47
0.78
0.55
0.64
0.62
0.58
0.77
OF VARIATION
PORTION
VARIATION
0.81
0.52
0.67
Table VI.B-38.
COLIPHAGE AEROSOL CONCENTRATION PRECISION
RUN GROUP
Post-Fair at 100m
Pre-Fair Pairs
QA:M2-27-28
Post-Fair at 50m
QA:M2-8-9
QA:Ml-26
QA:Ml-24-25
MEAN
(PFU/m3)
0.3
0.3
0.5
0.5
1.0
1.5
2.8
COEFFICIENT
MEASUREMENT
VARIATION
0.88
0.87
0.63
0.56
1.32
0.29
0.33
OF VARIATION
PORTION
VARIATION
0.51
0.63
WEIGHTED ESTIMATE 0.73 0.56
124
-------
In contrast, for each microorganism group, the portion variation is larger than the sampling variation. Ship-
ping and analytical sources are apparently responsible for more of the microorganism aerosol measurement
variation than are the field sampling sources. For total coliform and standard bacterial plate count, aliquot
variation represents only a small part of the portion variation; for these microorganism groups, most of the
"shipping and analytical variation" is attributable to factors such as bottles, analysts, and day-to-day proce-
dural variations rather than to variations in repeated analyses. The differences in Table VI.B-40 between mea-
surement variation, portion variation, and the more frequently reported aliquot variation emphasize the need
for quality assurance runs and pairing of samplers if one is to accurately estimate aerosol measurement preci-
sion and assess its contributing factors.
Table VI.B-39.
PATHOGEN AEROSOL CONCENTRATION PRECISION
MICROORGANISM GROUP
RUN GROUP
MEAN
(CFU/nT)
COEFFICIENT OF VARIATION
MEASUREMENT PORTION
VARIATION VARIATION
PSEUDOMONAS
Pre-Fair Pairs 22
QA:Ml-24-25 79
QA:Ml-26 150
Pseudomonas Weighted Estimate
0.82
0.40
0.47
0.58
MYCOBACTERIA
QA.-M2-27-28 0.6
Post-Fair at 100m 1.0
Post-Fair at 50m 1.4
Mycobacteria Weighted Estimate
0.86
0.59
0.86
0.81
0.92
0.92
KLEBSIELLA
QA:Ml-26
QA:Ml-24-25
Klebsiella Weighted Estimate
1.6
2.0
0.73
0.76
0.74
CLOSTRIDIUM PERFRINGENS
Pre-Fair Pairs 1.5 MPN/m3
0.60
125
-------
Table VI. B-40.
AEROSOL CONCENTRATION PRECISION SUMMARY
Coefficient of Variation
MICROORGANISM GROUP
Dye
Total Colif orm
Standard Bacterial Plate Count
Fecal Coliform
Pseudomonas
Clostridium perfringens
Coliphage
Klebsiella
Fecal Streptococci
Mycobacteria
Total
(Measurement
Variation)
17%
50%
50%
58%
58%
60%
73%
74%
77%
81%
Sampling
(Sampling
Variation)
16%
9%
34%
Little
46%
39%
Analytical
(Portion
Variation)
5%
49%
37%
65%
56%
67%
Replication
(Aliquot
Variation)
15%
11%
5. Particle Size Distributions
The data available for particle size analysis come from Phase I of the program. A more extensive
study of particle size using six-stage Andersen samplers was planned for Phase II, but had to be deleted be-
cause of funding limitations. In the Phase I sampling, two-stage Andersen samplers were used to obtain the
samples, with the size ranges of >7.0 \im (Stage I) and 1.05-7.0 ^m (Stage II), which roughly correspond to
nonrespirable and respirable particles, respectively. While some information is obtained on the distribution, it
is less definitive than would have been available using the six-stage sampler. However, the purpose of the
aerosol runs made with these samplers in Phase I was not originally aimed to obtain particle size data. In some
instances, the purpose was to determine the correct times for the aerosol sampling in subsequent phases of the
program. On other runs, the Andersen sampler was paired with AGI and LEAP samplers to allow an evalua-
tion of the different means of obtaining aerosol samples. This breakout of the data by size is intended only to
give a general idea of the distribution and to provide comparative data with other studies.
Two collection media were used in the Phase I study: EMB agar for total coliform and Casitone
agar for total count. The summary of the total coliform results is shown in Table Vl.B-41, while for total
count the results are shown in Table VI.B-42. For those counts that fell either above or below the detection
limit, no percentage calculation was made. It should be noted that not all the runs made during Phase I are
summarized here. If the number of usable results was small, or in some cases nonexistent, the run was not
included in these summaries.
The particle size distributions for total coliform and total count are shown in Tables VI.B-43 and
VI.B-44, respectively. For purposes of summarizing the results, the upwind samples and those beyond 200
meters for total count and beyond 100 meters for total coliform are considered to represent background lev-
els. The results are summarized as percent respirable (1.05-7.0 ^m).
For total count, the percent of the total number of particles which can be considered to be respira-
ble range from a low of 14 percent to a high of 76 percent for the background. The median percent was 43.5
with a mean value of 43.6 percent, based on 20 values. The total count particle size distribution is broken
126
-------
down into two ranges for further study; close downwind (5 to 25 meters) and far downwind (50 to 200 me-
ters). The close downwind results gave a range from 65 percent respirable to 76 percent respirable, based on
four results. The median was 70 percent with a mean of 70.2 percent. The far downwind results had ten data
points, ranging from 20 percent to 74 percent respirable. The median for this data set was 43 percent, with a
Table VI.B-41.
AEROSOL PARTICLE SIZE DISTRIBUTION FROM TWO-STAGE ANDERSEN SAMPLERS-
TOTAL COLIFORM (No./irf)
Stage I Stage II
Run Distance (Non-Respirable) (Respirable)
7 100 1.4 x 102 3.9 x 102
100 1.1 x IQl 2.0 x 102
8 u 7.7 x 101 >2.4 x 102
20 1.4 x 102 6.0 x 102
20 1.3 x 102 6.4 x 102
75 7.1 x 101 4.9 x 102
75 2.8 x 101 5.6 x 101
75 3.5 x 101 1.4 x 101
390 «3.5 x 101 7.1 x 102
390 2.1 x 101 1.4 x 102
390 4.7 x 101 9.4 x 101
9 u 1.7 x 102 1.7 x 102
u 1.1 x 102 8.0 x 101
825 3.5 x 101 1.4 x 101
825 1.9 x 101 1.6 x K)l
825 3.5 x 101 1.9 x 101
10 500 <2.0 x 10° <2.0 x 10°
500 2.1 x 102 8.5 x 101
500 5.9 x ID*
1500 1.5 x 102 3.5 x 102
1500 — 4.9 x 101
1600 «?3.5 x 101 <3.5 x 101
1600 1.4 x 101 4.2 x 101
1600 2.0 x 10° 2.4 x 101
11 1 <7.1 x IQl <7.1 x 101
1 1.1 x 102 1.8 x 102
1 2.8 x 101 6.4 x 101
1 <2.0 x 10° 1.6 x 101
1600 <7.0 x 10° <7.0 x 10°
1600 2.0 x 10° 5.0 x 10°
Note: u = upwind.
127
-------
Table VI.B-42.
AEROSOL PARTICLE SIZE DISTRIBUTION FROM TWO-STAGE ANDERSEN SAMPLERS-
TOTAL COUNT
St.ge II
Run Distance tn (Non-Respirable) (Resiirable)
Distance m
u
5
25
50
100
150
u
10
50
100
200
500
700
u
20
100
200
600
1000
1600
u
40
100
450
800
1200
u
310
u
20
20
75
75
75
390
390
390
u
u
825
825
825
500
500
500
1500
1500
1600
1600
1600
1
1
1
1
1600
1600
Stage I
(Non-Respirable)
1.6 x 101
4.2 x 102
9.9 x 102
4.2 x 102
4.7 x 101
7.0 x 10°
4.7 x 101
6.4 x 10°
1.6 x 102
1.4 x 101
8.8 x 101
5.5 x 101
—
8.0 x 10°
8.0 x 101
1.9 x 102
1.1 x 102
1.2 x 101
>2.4 x 102
2.9 x 101
2.9 x 101
2.2 x 102
1.2 x 101
2.7 x 101
3.5 x 101
>2.4 x 102
1.4 x 102
>2.4 x 102
>2.4 x 102
>1.4 x 103
2.3 x 103
8.5 x 102
>4.7 x 102
4.2 x 102
6.7 x 102
>4.7 x 102
>1.4 x 103
>4.7 x 102
5.7 x 102
="4.7 x 102
>2.4 x 102
4.2 x 10'"-
9.9 x 102
2.9 x 102
7.8 x 102
>4.7 x 102
2.8 x 102
3.5 x 102
1.3 x 102
1.4 x 102
7.4 x 102
3.6 x 102
2.8 x 101
4.5 x 102
>4.7 x 102
5.0 x 10°
1.0 x 103
2.J x 103
2.6 x 102
1.2 x 101
-<7.0 x 10°
1.3 x IQl
1.2 x 102
2.1 x 102
1.5 x 101
5.9 x 101
8.0 x 101
6.0 x 10l
2.6 x 101
2.6 x 102
1.6 x 102
1.1 x 102
1.4 x IQl
1.6 x 102
5.3 x 101
1.6 x 1Q1
>4.7 x 101
3.5 x 101
3.5 x 101
3.6 x 101
1.2 x 102
1.8 x 102
>2.4 x 102
>2.4 x 102
>1.4 x 103
7.4 x 102
4.0 x 102
2.8 x 102
1.8 x 102
4.4 x 102
>4.7 x 102
8.8 x 102
>4.7 x 102
5.7 x 102
>4.7 x 102
>2.4 x 102
10 500 4.2 x 10'- <3.5 x 101
8.8 x 102
4.7 x 102
1.3 x 102
1.4 x 102
1.1 x 102
9.4 x 101
11 1 1.4 x 10* <7.1 x 101
2.8 x 102
1.8 x 102
4.2 x 101
2.1 x 102
>4.7 x 102
128
-------
mean of 42.1 percent. As can be seen, the far downwind results resembled the size distribution of the back-
ground samples for percent respirable, while the close downwind samples collected a higher percentage of res-
pirable (small) particles.
For total coliform, the background was represented by two upwind samples and ten samples taken
more than 100 meters from the spray line. The distances of these samples ranged from 390 to 1600 meters.
Percentage of the total particles in the respirable range for the background ranged from 29 percent to 92 per-
cent, with a median of 48 percent and a mean of 53.8 percent. The samples taken close to the spray line were
at distances of 20, 75 and 100 meters and no further division of these values was made. The percent respirable
at these distances went from a low of 29 percent to a high of 87 percent, based upon seven observations. The
Table VI.B-43.
AEROSOL PARTICLE SIZE DISTRIBUTION FROM TWO-STAGE ANDERSON SAMPLERS
—TOTAL COLIFORM (Percent)
Run
10
11
Distance
100
100
u
20
20
75
75
75
390
390
390
u
u
825
825
825
500
500
500
1500
1500
1600
1600
1600
1
1
1
1
1600
1600
,n
Stage I
(Non-Respirable)
26
36
19
17
13
33
71
60
48
50
58
71
54
65
71
30
25
8
38
30
29
Stage II
(Respirable)
74
64
81
83
87
67
29
40
52
50
42
29
46
35
29
70
75
92
62
70
71
129
-------
Table VI.B-44.
AEROSOL PARTICLE SIZE DISTRIBUTION FROM TWO-STAGE ANDERSEN SAMPLERS-
TOTAL COUNT
(Percent)
Stage I Stage II
Run Distance,m (Non-Resplrable) (Respirable)
1 u 76 24
5 30 70
25 30 70
50 74 26
100 80 20
150
2 u 78 22
10 35 65
50 43 57
100 48 52
200 60 40
500 41 59
700
3 u 24 76
20 24 76
100 54 46
200 50 50
600 46 54
1000
1600 35 65
4 u 64 36
40
100 26 74
450 44 56
800 49 51
1200
5 u 44 56
310
8 u —
20
20
75 76 24
75 68 32
75
390 70 30
390 60 40
390
825 50 50
825
825
10 500 —
500 53 47
500 38 62
1500 86 14
1500
1600 67 33
1600 76 24
1600 58 42
11 1
1 72 28
1 67 33
1 40 60
1600 68 32
1600
130
-------
median percent respirable was 74 percent, with a mean of 72.7 percent. In general, then, there was a higher
percentage of respirable total coliforms close to the spray line.
The results can be compared to other studies conducted at Ft. Huachuca, Arizona, and at Chi-
cago, Illinois. In each of these studies, the six-stage Andersen sampler was used, but the division according to
respirable (Stage 1) and nonrespirable (Stages 2-6) particles can be made. In the Ft. Huachuca study, the res-
pirable portion of the total viable particles accounted for 70-80 percent of the total particles for three of the
four runs. On the fourth run, only 50-60 percent of the particles were in the respirable range. These are similar
to the data obtained in this study.
At the Chicago site both standard bacterial plate count and total coliform data were obtained. For
standard bacterial plate count, the upwind sample gave a respirable portion of 35 percent, slightly below the
median for this study. The results showed median values of 67 percent for close downwind (10-20 meters) and
55 percent respirable for farther downwind (100 meters), similar to those for this study. No samples were ob-
tained beyond 200 meters in that study.
For total coliform, only at one distance were nonrespirable particles isolated at 97 meters down-
wind. However, the actual counts obtained at all distances fell below the recommended minimum of 30 parti-
cles and these data provide little comparative utility.
6. Aerosol Microbial Characterization
A special aerosol run, Ml-36, was made with eight samplers set side-by-side at 20 meters from the
wet-line edge to collect a high-volume aerosol sample for pathogen screen. Run Ml-36 was conducted on the
dry side of Field 3-1 at night (2137-2207) on June 21, 1976, at a temperature of 12°C, a relative humidity of 82
percent, and a wind velocity of 1.6 m/s. The BHI sample collection fluid from all samplers was aggregated
and analyzed both by the qualitative pathogen screen procedure used for the high-volume wastewater samples
and by the usual quantitative procedure. The aerosol microbial characterization results are presented as Run
36 in Tables VI.A-19 and VI.A-21 for comparison with the wastewater microbial characterization. The aero-
sol concentrations obtained from this sample are given in Table VI.B-45. The mycobacteria aerosol concen-
tration was surprisingly high on this run. The prevalence of mycobacteria in this aerosol sample and in the
wastewater microbial characterization sample during Pre-Fair was the basis for its selection as a microorga-
nism group to be routinely monitored in the Post-Fair sampling. In addition to the microorganism groups
listed in Table VI.B-45, the pathogen screen identified Proteus, Enterobacter, "other oxidase-negative gram-
negative nonfermenters," and "other oxidase positive fermenters." While the aerosol bacteria identified were
representative of those found in the wastewater microbial characterizations, fewer bacterial types were found
in the aerosol, as expected.
131
-------
Table VI.B-45.
AEROSOL CONCENTRATIONS FOR THE AEROSOL MICROBIAL CHARACTERIZATION RUN
MICROORGANISM GROUP AEROSOL CONCENTRATION
Std. Bacterial Plate Count 7800 /m3
Total Coliform 43 MFC/m3
Fecal Coliform 6.0 MFC/m3
o
Fecal Streptococci 2.3 CFU/m
Pseudomonas 240 CFU/m3
Mycobacteria 280 CFU/m3
Klebsiella 4.1 CFU/m3
Clostridium perfringens 4.3 MPN/m3
o
Staphylococcus aureus 9 CFU/m
C. Aerosol Data Analyses
1. Microbiological Dispersion Model
This section presents a microbiological dispersion model that may be applicable for prediction of
the viable aerosol concentration of any measurable microorganism group emanating from any sprayed waste-
water aerosol source. This microbiological dispersion model incorporates the microbiological aerosol charac-
teristics derived from the extensive aerosol sampling program of this study. Considering the imprecision
and cost of measuring microorganism aerosol concentrations from spray irrigation by field sampling, using
predictions of the microbiological dispersion model supplemented with minimal field sampling does appear to
be a preferable alternative to extensive field sampling when the sprayed wastewater does not contain residual
chlorine.
The microbiological dispersion model extends the standard air pollutant dispersion models*35*, and
predecessor microbiological models*36'37-38-39-40' by incorporating a site-specific parameter for aerosolization
efficiency and microorganism-specific parameters for microbiological die-off.
a. Model Deriva tion
Consider the aerosol concentration, Cg(r,d), of a microorganism group or dye, g, sampled
under the environmental conditions of aerosol run r at a perpendicular distance d downwind from the wet-line
edge of the spray line. This sampled aerosol concentration should equal the sum of the predicted microorga-
nism concentration, Pg (r,d), emanating from the spray line to that distance on the run, the existing back-
ground concentration, Bg (r), of the microorganism under the run conditions, and a random error, £g, with a
mean of zero to express the uncertainty of the relationship:
C = P + B + E (1)
132
-------
where: d = perpendicular distance from the spray wet-line edge, (m)
g = measured microorganism group or dye
r = environmental conditions during the aerosol run
C = C (r,d) = sampled aerosol concentration of microorganism g at distance d during
run r, (cfu/m3, colony forming units per cubic meter)
B = Bg (r) = sampled background aerosol concentration of microorganism g during run r
(upwind of the spray line), (cfu/m3)
P = Pg (r,d) = predicted aerosol concentration of microorganism g emanating from the
spray line to a distance d during run r obtained using the microbiological dispersion
model, (cfu/m3)
E = £g = random error with zero mean representing the measurement errors in C and B
and the prediction error in P, (cfu/m3).
The variance of E is the sum of the variances of C, B, and P under the reasonable assumption
of independence. The measured concentrations C and B have a relatively constant coefficient of variation, crg,
which was estimated in Section VI. B.4.b. as the measurement variation for each microorganism, g. Thus, the
variance of £ is:
Var(£g) = ag2(Cg2 + B£) + Var(Pg)
Since ag ranges from 0.50 for total coliform and standard bacterial plate count to 0.81 for mycobacteria, and
since aerosol model predictions may be expected to have coefficients of variation in the same range, the stan-
dard deviation of £g probably ranges between 70 and 110 percent of the aerosol concentration, C. Thus, equa-
tion (1) contains a substantial degree of uncertainty.
Figure VI.C-1 depicts the transport of aerosol downwind of a spray line to a position, x. Let:
a = a(r,d) = aerosol age at distance d during run r, (s)
u = mean wind velocity during run r, (m/s)
w = wet-line edge distance during run r, (m)
<(> = angle of the wind direction with the perpendicular to the line of the spray heads.
Then the aerosol age, a, is computed as
w + d (2)
Q = .
U COS<(>
At d = 50m, for example, a50 = (w + 50m)/(u cos <|>).
The following multiplication model for the predicted concentration, P, coming from the
spray line is postulated:
P = D x E x I x e*a (3)
where the model parameters are:
D = Dg (r,d) = physical diffusion model aerosol concentration of microorganism g emanat-
ing from the spray line to a distance d during run r, assuming all the sprayed wastewa-
ter (including its measured wastewater concentration of microorganism g) becomes
aerosol, and assuming no microbiological die-off, (cfu/m3)
E = E(r) = aerosolization efficiency factor: the fraction of the sprayed wastewater that is
aerosolized during run r, (0
-------
Wind
Direction
o
Wind Angle
a, Aerosol Age
at Downwind
Distance d
O Line of Spray Heads
w, Wet Line Edge Distance
Wet Line Edge
) d, Perpendicular Downwind Distance
x, Downwind Position
Figure VI.C-1.
SCHEMATIC OF AEROSOL TRANSPORT DOWNWIND OF A SPRAY LINE
134
-------
I = Ig(r) = microbiological impact factor: the proportion of the aerosolized microorga-
nisms of group g that remain viable immediately after aerosolization during run r,
(IX))
A = Ag (r) = microbiological age decay rate: rate at which the microorganisms of group g die
off with aerosol age during run r, (A<0).
The exponential microbiological viability decay factor, eAa, expresses the reasonable assumption that the run,
r, environmental conditions kill a constant proportion, A, of the remaining number of the viable microorga-
nisms of group g in the aerosol cloud with unit increase in the aerosol age, a.
Dispersion models generally*41 •42) contain a source strength term Q, in units of mass per unit
time, to express the pollutant emission rate at the source. For spray irrigation, Q is calculated as the product
of four factors: the wastewater concentration of the microorganism, the wastewater application rate, the
aerosolization efficiency (E), and the impact factor (1) for the microorganism. In the microbiological dispers-
ion model, the wastewater concentration and application rate factors are included in D, while E and I are
specified separately. This permits separate estimation of the E and 1 factors.
b. Effect of Each Model Factor
The effect of each factor of the microbiological dispersion model, P, is depicted schemati-
cally in Figure VI. C-2. The physical diffusion model calculates a very high microbiological aerosol concen-
tration, D, because it assumes that all of the sprayed wastewater is converted to aerosol and that all of the
microorganisms measured in the wastewater remain viable while entering and being transported in the aero-
solized state.
Actually, only a small fraction, E, of the sprayed wastewater becomes aerosolized. The aero-
solization efficiency factor, E, depends upon the type of spray equipment employed, the spray head pressure,
and such meteorological conditions, r, as the ambient wind velocity and air temperature. Typical values of E
are in the range of 0.001 to 0.01 (i.e., 0.1% to 1% of the sprayed wastewater). Thus, when the amount of
wastewater actually aerosolized is taken into account, a much lower microbiological aerosol concentration, D
x E, is obtained.
Further impact changes, I, in microbiological aerosol concentration occur during entry into
the aerosolized state. Originally it was assumed that these impact changes were reductions, varying in magni-
tude for different microorganism groups, g, and environmental conditions, r, that would result from the
shock of entry into the inhospitable aerosol environment and from possible collection inefficiency in the aero-
sol samplers. However, the estimates of I to be presented later consistently exceed 1.0 for the hardy microor-
ganisms (i.e., those microorganisms that survive best in the aerosol environment). As will be discussed later,
the impact factor, I, is actually an empirical catchall for various microorganism-specific initial effects. With
the impact factor contribution, the resulting aerosol concentration at the spray line is D x E x I.
Finally, there is gradual exponential die-off of microorganisms after aerosolization, e*a,
which is presumed to occur at a constant decay rate, A, with aerosol age. This microbiological die-off accumu-
lates as a result of continued exposure to hostile environmental factors, such as solar radiation and low rela-
tive humidity, throughout the travel of the aerosol cloud. Taking this exponential die-off factor into account,
the aerosol concentration curve for the complete predicted concentration, P = DxExix eia, is obtained.
Adding the background concentration, B, to the predicted concentration, P, yields the sampled aerosol con-
centration, C = P + B.
The microbiological dispersion model equations (1) and (3) were used to estimate values of
the model parameters, D, E, I, and A for each appropriate microorganism group from the aerosol sampling
135
-------
P»DxExIxe
\ a
Aerosol
Concentration
B
a50
Figure VI.C-2.
SCHEMATIC OF EFFECTS OF MODEL FACTORS
a,
Aerosol Age
136
-------
program. The estimation procedures used and the model parameter estimates obtained are presented in the
following three sections.
2. Diffusion Model Concentration D
The H. E. Cramer Company participated in the research effort and calculated the physical diffu-
sion model terms, D, for each microorganism group at each sampler location during each run(43,44) usmg their
Volume-Source Diffusion Models Program. D is calculated using the total volume of sprayed wastewater and
the measured wastewater concentration of each microorganism. Each sprayer along the sprayer line was mod-
eled individually as a separate source. The diffused concentration contributions from each sprayer that
reached a given location were then summed to compute D for that location.
a. Approach
Specific objectives of the H.E. Cramer Company work included:
• Calculations of the model concentrations, D, and the aerosolization effi-
ciency, E, of the wastewater spray system for the Rhodamine WT dye
tracer at each sampler during each dye aerosol run.
• Calculations of model concentrations, D, for each microorganism group
sampled at each sampler during each microbiological, quality assurance,
and special enterovirus aerosol run.
• Calculations of normalized concentration isopleth patterns downwind
from the spray lines for each of the runs.
The concentrations, D, were calculated by means of a diffusion model under the assump-
tions that all of the material is aerosolized and that no material is lost by decay, gravitational deposition or
other depletion processes (i.e., all aerosol particles are small enough to be treated as a gas having no settling
velocity over the aerosol ages sampled). The model calculations were made by using the Volume-Source Dif-
fusion Models Program developed for Dugway Proving Ground by the H.E. Cramer Company from the gen-
eralized models described by Cramer'45'.
A mathematical description of the Volume-Source Diffusion Models Program is presented in
Appendix A of the H.E. Cramer Company technical reports*46' 47>. In the model calculations, each spray head
was treated as an individual source and the model was used to calculate the composite concentration pattern
produced by the discharges from all spray heads in use during each aerosol run. Meteorological inputs used in
the diffusion model calculations were presented in Table VI. B-l of this report.
b. Source Inputs
The source geometry of the spray lines and the positions of the active spray lines relative to
the samplers were based on field measurements and observations. The spray heads were spaced about 9.1 me-
ters apart and were about 0.6 meters above the ground. The top of the wastewater spray cone from each head
was observed to be about 5 meters above the ground and the radius of the spray circle produced by each head
was about 9 meters.
Volume sources with the above dimensions and spacing were used to simulate the spray lines
in the model. The height and radius of the spray cone were divided by 2.15 to obtain the initial source dimen-
sions, ozR = 2.0 meters and oyR = 4.2 meters (the standard deviations of the vertical and the lateral aerosol
distributions at the source). The source height H was assumed to be 0.6 meters above ground. The distance
from the projected virtual point source over which rectilinear expansion was assumed was Xry = Xrz = 50 me-
ters. The separation distance between model volume sources was set equal to 9.1 meters, and the coordinates
of the spray lines and downwind sampling locations were entered on the model calculation grid.
Estimates of the source strengths (emission rates) of the dye and various microorganism
137
-------
groups contained in the spray were based on measurements of their wastewater concentrations during the run.
To obtain emission rates for individual spray heads, these concentrations were multiplied by a flow rate.
Flow-rate measurements were made for each spray head on the line directly upwind from the aerosol samplers
at least once each time the spray line was moved to a new spray field. During some of the microbiological and
virus aerosol runs, spray lines in other fields were operating and could have contributed to the measured con-
centrations at one or more of the samplers. Since flow-rate measurements were not made on these lines, the
flow-rate data for those lines on other runs in the Post-Fair and Pre-Fair programs were used to estimate the
flow rates for use in the model calculations.
The flow rate profiles used in the source strength calculations for the dye, microbiological
and virus runs are given in Appendix C of the Cramer Pre-Fair report*48' and Appendices B, C, and D of the
Cramer Post-Fair Reports'49'.
c. Calculation Procedure and Model Concentrations
The Volume Source Diffusion Model was used with the meteorological, FP rotorod, and
source inputs to produce concentration isopleth patterns for each trial. Concentrations were calculated for an
array of grid points located at distances up to 1000 meters downwind from the main spray line. As noted
above, grid points corresponding to the locations of the aerosol samplers were included in the array. Com-
puter plots of the calculated concentration isopleths for the dye and microbiological trial runs are presented in
Appendix B of the Cramer Pre-Fair report and in Appendices F and G of the Cramer Post-Fair report.
In this report, Figure VI. C-3 illustrates the dye concentration isopleths, D, obtained from
the diffusion model for a typical dye aerosol run. The normalized isopleths calculated from the diffusion
model are presented in Figures VI. C-4 and VI. C-5 for two of the microbiological aerosol runs. When these
normalized isopleths are multiplied by the wastewater concentration of a microorganism group, the diffusion
model microorganism aerosol concentration D is obtained.
Diffusion model concentrations, D, are presented and compared with the sampled concen-
trations C for each sampler on each run for each microorganism group in Appendix D of the Cramer Pre-Fair
report and in Appendix E of the Cramer Post-Fair report. While the results varied considerably between runs,
the diffusion model concentration D at 100 meters generally was from 50 to 80 percent of the model concen-
tration D at 50 meters.
The model calculation procedure for the two special virus runs was modified slightly to con-
form to the enterovirus concentration, which was based on pooling the sample from all samplers over four
and six successive 30-minute run sub-periods. The measured aerosol concentrations for the virus runs given in
Table VI. B-24 thus represent an average over the sub-periods as well as over the 13 sampling locations. A
similar averaging scheme was used to calculate the model concentrations. First, a set of normalized concentra-
tions was calculated for each sub-period using model inputs derived from meteorological measurements made
during the 30-minute sampling periods. The calculated concentrations for all samplers in all sub-trials were
then averaged to obtain an average normalized concentration for the virus trial. Average aerosol concentra-
tions for the enteroviruses and other microorganism groups were then obtained by multiplying the normalized
concentration by their corresponding wastewater sample concentrations.
3. Aerosolization Efficiency Factor E
a. Dye Run Aerosolization Efficiency Estimates
An aerosolization efficiency estimate was computed for each dye aerosol run. Since the dye
is not present in the background, and since it does not degrade significantly over the aerosol ages sampled, B
= 0, I = 1, and A = 0 for the dye runs. Thus, the aerosolization efficiency estimate at a dye run sampler loca-
tion derived from the model equations (1) and (3) is:
138
-------
25O
VO
-200 -150 -100
-50 0 50
DISTANCE (m)
100 ISO 200 250
Figure VI.C-3.
CONCENTRATION ISOPLETHS OF DIFFUSION MODEL D FOR A TYPICAL DYE RUN
-------
1000
i—7 i—71—T-TT 1—T-TT—i—i
O
-500
-250
250 500
DISTANCE (m)
750
1000
Figure VI.C-4.
NORMALIZED ISOPLETHS OF DIFFUSION MODEL D FOR A TYPICAL
MICROBIOLOGICAL AEROSOL RUN
-------
1000
750-
UJ
o
if)
o
500-
250-
-500 -250
250 500
DISTANCE (m)
750
1000
Figure VI.D-5.
NORMALIZED ISOPLETHS OF DIFFUSION MODEL D FOR A MICROBIOLOGICAL
AEROSOL RUN HAVING TWO SPRAY LINE CONTRIBUTIONS
-------
E = C/D
(4)
The H.E. Cramer Company computed aerosolization efficiency E values for each sampler
on each dye run using equation (4). These sampler values are presented in Table D-l of Appendix D of the
Cramer Pre-Fair report and in Table E-l of Appendix E of the Cramer Post-Fair report.
The geometric mean of the E values obtained at each sampler location beyond d = 5 meters
from the wet-line edge was taken as the aerosolization efficiency estimate E for the dye run. These dye run
estimates are presented in Table VI. C-l ./Note that the aerosolization efficiency on the Pre-Fair dye runs con-
ducted in late spring (May and June) were more variable and generally higher than the aerosolization effi-
ciency for the Post-Fair dye runs during December and January.
The distribution of the 17 dye run estimates of aerosolization efficiency is summarized in
Table VI. C-2. The median aerosolization efficiency value obtained at Pleasanton was 0.0033 (0.33%). There
was over an order of magnitude of variation in E values from the tenth percentile (0.09% to the ninetieth per-
centile (1.8%). Thus, the fraction of wastewater that was aerosolized varied considerably from one run to
another.
b. Microbiological R un Aerosolization Efficiency Predictions
To separate the effects of E and I during the microbiological aerosol runs conducted at
Pleasanton, it was necessary to estimate E independent of the microbiological aerosol data. Thus, a regression
equation was sought from the 17 dye runs to relate the E values occurring at Pleasanton to potentially relevant
meteorological and operating conditions. The conditions considered were wind velocity, air temperature, rel-
ative humidity, solar radiation, wastewater temperature, spray head pressure, and seasonal bias.
Table VI.C-1.
ESTIMATES AND PREDICTIONS OF DYE RUN
AEROSOLIZATION EFFICIENCY E
Dye Aerosol Runs
Aerosol
Run
Number
Aerosolization Efficiency E
Dye Run
Estimate
Regression
Prediction
Dl-1
Dl-3
Dl-4
Dl-5
Dl-6
Dl-9
Dl-10
D2-1
D2-2
D2-4
D2-5
D2-6
D2-7
D2-8
D2-9
D2-10
D2-11
.0272
.0067
.0019
.0038
.0067
.0160
.0059
.0041
.0062
.0033
.0023
.0022
.0021
.0025
.0014
.0009
.0008
.0172
.0072
.0022
.0021
.0040
.0110
.0168
.0030
.0047
.0031
.0023
.0035
.0027
.0026
.0013
.0012
.0011
142
-------
Table VI.C-2.
DISTRIBUTION OF AEROSOLIZATION EFFICIENCY VALUES E
Number of
Values
(Dye Runs)
E, Aerosolization Efficiency 17
at Pleasanton
Distribution of Values — Percentiles
10%
(Median)
25% 50%
75%
90%
0.0009 0.0019 0.0033 0.0064 0.018
The potential regressor variables considered for inclusion in the aerosolization efficiency re-
gression were:
KM — wind velocity at meteorological tower, 4 meter level
US — wind velocity in spray field, 3 meter level
P — spray head pressure, Pascals
PERIOD — 1 = Pre-Fair, 2 = Post-Fair
TM — air temperature at meteorological tower, 2 meter level
TP — air temperature at effluent ponds, 2 meter height
RHM — relative humidity at meteorological tower, 2 meter level
RHU — relative humidity upwind, 2 meter height
RHL — relative humidity at Lawrence Laboratory in Livermore
WT — wastewater temperature at end of spray line, °C
R — solar radiation at meteorological tower, W/m2
UMSQ — UM squared
USSQ — US squared
TMSQ — TM squared
TPSQ — TP squared
RSQ — R squared
RHMSQ — RHM squared
TMXRHM — product of TM and RHM
UMXP — product of UM and P
UMXTM — product of UM and TM
UMXR — product of UM and R
Most of these variables are environmental conditions whose values were given in Table
VI.B-1. The average spray head pressure variable was calculated from Darcy's formula using the flow rate that
was measured at the spray line for each run. Since the orifice diameters were the same for all spray heads, the
expression for spray head pressure reduced to:
P = 21.028 x (Avg Spray Head Flow Rate)2 Pascals
where the flow rate is in liters per minute. PERIOD is a dummy variable to take into account any unmeasured
systematic factors that differed between the late spring Pre-Fair dye runs and wintertime Post-Fair dye runs.
The dependent variable log,^ was utilized. Inspection of the dye run E values in Table VI.C-
1 suggests the uncertainty in an E estimate may be proportional to the estimated value. Thus, regression on
log10E should yield a relatively constant error term.
Stepwise regression yielded numerous candidate regression equations involving a similar set
of regressor variables, many of which were highly correlated with each other. Because its regressor variables
were less correlated, the following regression equation was selected to predict the aerosolization efficiency E
at Pleasanton over the wide range of meteorological conditions encountered:
log,0E = 0.031 t + 0.000096 u»r —3.10
(5)
143
-------
where t = air temperature, °C
u = wind velocity, m/s
r = solar radiation, W/m2
With a coefficient of multiple determination R2 = 0.801, this regression explains 80 percent of the observed
variation in log,0E. The standard error SE = .194 implies that the true aerosolization efficiency E will be be-
tween 0.64 E and 1 .56 E of the equation (5) estimate E for about two-thirds of the estimates.
Predicted values of aerosolization efficiency obtained using regression equation (5) are also
provided in Table Vl.C-1 for each dye run. The Pleasanton aerosolization efficiency predicted by regression
was usually within a factor of two of the dye run estimates.
Using equation (5), predictions of the aerosolization efficiency were also made for each mi-
crobiological aerosol run. These predictions are presented in Table VI.C-3.
4. Impact Factor I and Aerosol Viability Decay Rate A
a . Estima tion Procedures for landX
(1) Standard Estimation Procedure
The values of I and A for a microorganism group during a run were estimated jointly by
simple linear regression using the aerosol concentrations obtained from each of the samplers. Substituting the
equation (1) expression for P into equation (3), rearranging terms, and taking natural logarithms, one obtains
(6)
This equation has the form of a simple linear regression model
yd = b0 + b,ad + ed (7)
where values of the dependent variable, yd = In [(Cd — B)/(Dd • E)], and of the independent variable, ad=
aerosol age, can be computed as an observation pair (yd, ad) for each downwind sampler at distance, d, that
obtained a detectable microorganism concentration, Cd, above the background level, B. Since b0 = In I and b,
= A, the estimates of I and A are readily calculated from the coefficient estimates for b0 and b, resulting from
the simple linear regression on the n observation pairs for the microorganism and run:
1 = exp (b )
Q £ (8)
A. = bj
. Sadyd - £yd 2ad/n A _ A_
*>0 - y - b,a
The uncertainty in the equation (8) estimates of I and A for a microorganism group on a
run can be estimated from the deviation of the yd values about the linear regression equation (7) estimates yd.
Using the expressions for the standard errors of the regression coefficients<50)s.e(b0) and s.e(b,), and the usual
variance functional transformation'5 '>,
Var(y)* -H r ' Var(x)
E(x)
144
-------
the standard errors of the estimators I and A are
A
s.e.(I)
$ ,£
; I s.e.(b
A 5
i-yd)
- 2
(9)
A A
s.e.(X) = s.e.(b,) =
1
/2(ad-ap A/ n-2
The standard estimation procedure was to jointly estimate I, A, and their standard er-
rors for each microorganism group on'each run using equations (8) and (9). This procedure was applied if
Table VI.C-3.
REGRESSION PREDICTION OF AEROSOLIZATION EFFICIENCY E FOR THE
MICROBIOLOGICAL AEROSOL RUNS
Microbiological Aerosol Runs
Aerosol
Bun
Number
Regression
Prediction
E
Aerosol
Run
Number
Regression
Prediction
E
Ml-1
Ml -2
Ml-3
Ml -4
Ml-5
Ml -6
Ml-7
Ml-8
Ml-9
Ml-10
Ml-il
Ml-12
Ml-13
Ml-14
MI-IS
Ml-16-19
Ml-20-23
M1-Z4-25
Ml-26
Ml-27-30
Ml -31
Ml-32
Ml-33
Ml-34
Ml-35
.0072
.0057
.0039
.021
.0091
.0024
.0016
.0067
.0058
.0019
.0024
.0041
.0017
.0016
.0016
.0016
.0035
.0018
.0018
.0096
.0060
.024
.021
.0036
.0024
M2-1
M2-2
M2-3
M2-4
M2-5
M2-6
M2-8-9
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-18-21
M2-22
M2-23
M2-24
M2-25
M2-26
M2-27-28
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
V2-I
V2-II
.0018
,0021
.0012
.0017
.0015
.0023
.0027
.0015
.0034
.0030
.0012
,0018
.0028
.0024
.0025
.0019
.0066
.0065
,0050
,0092
,0054
.0058
,0074
,0085
,0015
,0021
,0074
0025
,0132
,0068
0079
0050
0028
0045
145
-------
there were more than n = 2 valid, observation pairs for the microorganism group at two or more distances on
the aerosol run, if the viability decay rate estimate from equation (8) was negative (A <0), and if each of the
observation pairs were considered to have roughly equivalent weight.
(2) Special Estimation Procedures
For the hardier microorganism groups, joint estimation of I and A by equations (8)
often yielded a positive value for the viability decay rate A. A positive A implies growth rather than die-off of
the microorganism with lengthening exposure to the hostile aerosol environment. Significant net growth in
aerosols over durations up to several minutes is highly unlikely. Thus, positive A values probably result from
the substantial uncertainty error in the model equations (1) and (3) in those situations when the actual decay
rate was too slight to be detected. This uncertainty error is attributable to measurement variation, localized
microbial aerosol sources, and estimation errors in D and E. Accordingly, positive A values obtained by equa-
tions (8) were considered indistinguishable from A = 0 and denoted as A = X
Substituting A = 0 in equation (6) yields a distinct value Id = (Cd — B)/(Dd • E) for each
downwind sampler. The geometric mean of these downwind sampler values was taken as the impact factor I
estimate when the A estimate was considered indistinguishable from zero:
A
A = X
I = Geometric Mean (Id) = exp (y) (10)
over all d
A A A
Assuming A = 0 were known, the standard error of the equation (10) I is s.e.(I) = I s/\/"n
A
where s2 = Z(yd —~y)2/(n — 1). Because A is actually indeterminate in such cases, the I standard error is
probably larger than this, and may be better reflected by the equation (9) estimator. Thus, when joint estima-
tion by equations (8) yield a positive A estimate, equations (10) were used to estimate I and A and the standard
error of 1 was estimated as
s.e.(I) = Max[is\/~n,Is.e.(b0)] (11)
When there were only n = 2 observation pairs per run, and when the observation pairs
did not have roughly equivalent weight, variants of the standard joint estimation procedure [equations (8) and
(9)] and the indeterminate decay rate estimation procedure [equations (10) and (11)] were used. For n = 2,
equation (6) represents n = 2 equations in two unknowns (I and A). I and A were estimated as the exact solution
of these simultaneous equations. When A was positive, equations (10) and (11) with n = 2 were used for the
estimation.
Because the sampling designs used spread the samplers farther apart at the greater dis-
tances, d, the observation at the most distant of the designated downwind samplers usually had the greatest
effect on the estimated viability decay rate, A. However, the concentration, Cd, at this sampler was occasion-
ally below the detection limit and, thus, arbitrarily close to the background concentration, B. On such a re-
gression, the most distant observation y value had considerably greater uncertainty than did the others and
should hence receive less weight. In these cases, a weighted regression estimation procedure was used<52>. The
weighted simple linear regression model is
146
-------
where the observation, d, weight, wd, was based on the number of CPUs (colony forming units) obtained dur-
ing analysis of the aerosol sample:
CPUs Found
0
1
2
3
Raw Weight
0.125
0.5
0.75
0.9
1.0
The raw weights were standardized so that Iwd = n. The estimation procedure can still be represented by equa-
A
tions (8) and (9) when A <0 and by equations (10) and (11) when the equations (8) A were positive, provided each
summation in these equations is replaced by a weighted summation [e.g., Iad yd becomes Iwdadyd, Iyd becomes
Iwdyd, and I(yd — yd)2 becomes Zwd(yd — yd)2].
b. Impact Factor I
The estimated impact factor I and its standard error s.e. (I) were obtained for each microor-
ganism group on each aerosol run by applying the appropriate estimation procedure to the data of Tables
VI.B-8 through VI.B-15. The run estimates of I and s.e.(I) are presented in Table VI.C-4.
Comparison of an individual impact factor estimate I against its corresponding standard
error s.e.(I) throughout Table VI.C-4 shows there is considerable uncertainty in the individual impact factor
estimates. The standard error is generally of the same order of magnitude as the impact factor estimate. Con-
sidering the sizable measurement variation in the aerosol concentrations C (cf. Section VI.B.4.b), the large
Table VI.C-4.
RUN ESTIMATES OF MICROORGANISM GROUP IMPACT FACTOR I AND STANDARD
ERROR SE(I)
a. Pre-Fair Runs
Aerosol
Run
Number
MM
Ml-2
Ml-3
Ml-4
Ml-S
Ml-6
Ml-7
Ml-8
Ml-9
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
Ml-32
Ml-33
Ml -34
Ml -35
Standard
Plate Count
I
2.0
.357
.25
.022
.34
.347
.23
.089
.20
.139
.21
1.48
.71
.21
.163
.18
SE(I)
4.3
.082
.13
.010
.13
.054
.28
.063
.15
.033
.13
.46
.14
.13
.077
.12
Total
Coliform
I
.046
.048
.028
.84
.153
.053
.13
.040
.59
1.08
.12
.252
.57
.102
.084
.0037
.013
1.1
1.06
SE(I)
.014
.014
.074
.29
.056
.038
.13
.021
.20
.86
.12
.042
.29
.023
.066
.0010
.012
1.8
.23
Fecal
Coliform
I
.126
.062
.16
.13
.073
.088
1.19
.54
.62
.091
.016
.048
2.9
SE(I)
.051
.095
.16
.10
.015
.023
.70
.35
.49
.047
.018
.035
1.9
Coliphage
I
.0077
.0006
1.18
.094
.62
.134
1.53
.277
.38
1.42
1.34
.17
.33
.031
.17
1.00
SE(I)
.0040
.0012
.30
.052
.33
.077
.87
.073
.13
.30
.56
.10
.53
.045
.47
.61
Fecal
Streptococci
I
3.2
.64
2.65
4.4
21.2
7
70
25.8
.26
.101
1.9
6.1
SE(I)
3.9
.45
.51
2.8
5.3
11
72
6.9
.12
.028
4.1
1.9
Pseudomonas
I
1.07
14
40
210
5.1
2.7
52
2.2
400
1.14
1.09
26.3
94
SE(I)
.78
14
92
1.2
3.7
28
1.0
610
.89
.19
9.5
62
Clostridium
Perfrineens
1.22
4.9
.34
.237
.62
.064
7.3
6.5
6.0
7.2
.736
SE(I)
.28
1.8
.19
.079
.31
.021
4.1
3.3
7.0
5.4
.049
147
-------
uncertainty in the impact factor estimates I is not surprising. The ratio of the estimate I to its standard error
^
s.e.(I) has the t distribution with n-2 degrees of freedom, where n is the number of observation pairs regressed
/\ *
for the microorganism on the run. Table VI.C-5 summarizes these t statistics, t = I/s.e.(I), by microorganism
group. For most microorganism groups, less than half of the impact factor estimates exceed twice their stan-
A
dard error. The exception is the total coliform I estimates, 62 percent of which exceed twice their standard
error. Hence, the individual impact factor estimates generally have low reliability. However, the majority of I
estimates (ranging from 63 percent for coliphage to 91 percent for Clostridium perfringens) do exceed their
standard error. Thus, aggregating the individual run I estimates for a microorganism group as an I-value dis-
tribution for the microorganism group should provide a good representation of the central tendency of I for
the microorganism group and a fair representation of its dispersion under different run conditions. The em-
pirical distribution of these impact values for each microorganism is presented in Table VI.C-6.
The microorganism groups are arranged in Table VI.C-6 according to the magnitude of their
median and quartile impact factor values. The microorganisms differ substantially with respect to their mid-
dle range of impact factor values. The "indicator" microorganism groups—total coliform, fecal coliform,
standard bacterial plate count, and coliphage—all had a similar middle range of impact factor values. The
Table VI.C-4.
RUN ESTIMATES OF MICROORGANISM GROUP IMPACT FACTOR I AND STANDARD
ERROR SE(I)
b. Post-Fair Runs
Aerosol
Run
Number
M2-1
M2-2
M2-3
M2-4
M2-5
M2-6
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
V2-I
V2-II
Standard
Plate Count
I
.35
.11
.8
.2
.102
.055
.25
.069
.19
.288
.18
.217
.209
.025
1.4
.053
.88
SE(I)
.14
.11
3.1
1.1
.085
.059
.16
.066
.10
.003
.13
.072
.072
.022
2.1
.055
.16
Total
Coliform
I
22.2
.08
1.05
.47
.225
.573
.148
.143
.032
.162
1.97
.143
.184
.175
.104
.20
.24
.30
.0110
.130
.0161
.0160
.32
.281
1.31
SE(I)
3.3
.11
.12
.19
.044
.046
.064
.067
.090
.053
.44
.053
.061
.052
.071
.34
.18
.13
.0054
.023
.0018
.0037
.154
.88
Coliphage
I
.163
1.2
.68
2.0
.17
.34
.34
.18
.91
2.8
.45
.72
.35
3.6
.016
.74
.62
.70
.061
.45
12
.018
.071
.0635
.014
.08
.101
SE(I)
.044
1.2
.81
.36
.10
.22
.38
.12
.16
1.4
.90
.19
.16
5.1
.038
.85
.51
.56
.071
.82
20
.017
.025
.0039
.015
.10
.027
Fecal
Streptococci
I
.8
.84
34
9.4
.97
'.52
1.7
3.0
2.3
.46
.71
1.31
.283
5.5
1.0
.12
.73
1.10
59
SE(I)
2.5
.61
14
6.2
.10
.49
2.4
2.4
1.1
.23
.21
.40
.088
1.9
1.2
.11
.68
.77
26
Mycobacteria
I
2.5
.14
.83
1.15
.76
31
.95
.81
SE(I)
8.7
.11
.76
.88
.53
24
.51
.19
Enteroviruses
3-Dav
I
44
<4.7
3 & 5 Dav
I
69
24
148
-------
inter-quartile ranges were (0.06—0.6) for total coliform, (0.07—0.6) for fecal coliform, (0.1—0.4) for stan-
dard bacterial plate count, and (0.09—0.9) for coliphage. It is important to note that all of the pathogenic
bacteria and virus groups had substantially higher middle ranges of I than did these indicator microorga-
nisms. The bacterial pathogen impact factor middle ranges extended from (0.8—2) for mycobacteria, (0.2—7)
for Clostridium perfringens, and (0.7—6) for fecal streptococci to (2—70) Pseudomonas. While only two im-
pact factor estimates were obtained for each enterovirus group, the enteroviruses appear to have a middle
range at least as high as any bacterial pathogen evaluated. The five-day enteroviruses (when polioviruses were
not suppressed) appear to have a middle range that is more than 100 times as high as any indicator microorga-
nism group.
Frequent impact factor values exceeding 1.0 were not expected for any microorganism
group. However, over half of the I estimates for enteroviruses, Pseudomonas, fecal streptococci, and Clostri-
dium perfringens exceeded 1.0. Possible explanations for the consistent occurrence of pathogen I values
above 1.0 are presented and discussed in Section VII.C.
The impact factor estimates I for most microorganism groups exhibited about two orders of
magnitude of variation in value between their tenth and ninetieth percentiles on the Pleasanton runs. The
standard bacterial plate count impact values were somewhat more consistent than those of the more specific
microorganism groups.
c. Viability Decay Rate A
^ ^
The individual viability decay rate estimates X and their standard errors s.e.(A) calculated for
Table VI.C-5.
RELIABILITY OF IMPACT FACTOR ESTIMATES I
^ No. of^ Percentage of Estimates Exceeding
(I, s.e.(I)) Two Std. Errors Std. Error
Microorganism Group Estimates t>2 t>l
Std. Bacterial Plate Count 33 39% 76%
Total Coliform 42 62% 86%
Fecal Coliform 13 23% 77%
Coliphage 43 30% 63%
Fecal Streptococci 31 45% 77%
Pseudomonas 12 33% 67%
Clostridium perfringens 11 45% 91%
Mycobacteria 8 13% 88%
Note: t = I/s.e.(I)
149
-------
Table VI.C-6.
DISTRIBUTIONS OF AEROSOL IMPACT FACTOR, I
(•pact Factor, I
Microorganism Croup
Fecal Col i form
Total Coliforn
Std. Bacterial Plate Count
Coliphagc
O Mycobacteria
Clostridium perfringens
Fecal Streptococci
I'seudomonas
3-Day Enteroviruses
S-Day Enteroviruses
Number of
Estimates
(Aerosol
Runs)
13
44
33
43
8
11
31
13
2
2
Distribution of Values - Percentiles
5\ lOt 2S4 401 SOt 60t 7St 90 1 »&t
.029 .068 .090 .13 .15 .58 2.0
.012 .016 .060 .13 .16 .23 .55 1.1 1.6
.021 .036 .11 .19 .21 .24 .35 1.2 1.7
.009 .017 .094 .18 .34 .52 .91 1.8 3.4
«*.13 .77 .82 .89 1.0 2.1 34
.085 .24 .71 1.2 S.I 6.5 7.3
.11 .27 .71 .97 1.7 2.7 6.1 32 64
1.1 1.7 (33%) 4.1 14 32 (67%) 73 320
<4.7 -10 44
24 ~40 68
~ x •- Interpolated or extrapolated value x.
-------
each microorganism and run from the data of Tables VI.B-8 through VI. B-15 are presented in Table VI.C-7.
A
The percentage of indeterminate viability decay rates (A. =X) in Table VI.C-7 is summarized in Table VI.C-8.
For all microorganism groups except total and fecal coliform, indeterminate decay rates were prevalent. We
A
interpret these indeterminate values A =X as reflecting a very slight (negative) decay rate which could not be
estimated because of the large uncertainties in the aerosol concentrations C and the background concentra-
tion B, relative to the limited range of sampled aerosol ages. Their prevalence suggests that many of the esti-
mated negative decay rates may also be indistinguishable from no decay (A = 0).
A A
For the estimated negative decay rates, their t statistics, t = A/s.e.(A), are summarized by mi-
croorganism group in Table VI.C-9. Only for total coliform and fecal coliform do more than half of the neg-
ative viability decay rate estimates exceed twice their standard error. However, over 60 percent of the esti-
mates exceed one standard error for all frequently estimated groups.
The distribution of the Table VI.C-7 viability decay rates for each microorganism group is
presented in Table VI.C-10. The indeterminate viability decay rates are indicated by anXin Table VI.C-10.
Question marks have been placed in parentheses after the smallest negative viability decay rates, since the
prevalence of X's suggests these values may also be indistinguishable from A = 0.
Based on their viability decay rates, the microorganism groups seem to fall into three catego-
ries. Total coliform and fecal coliform exhibit the most rapid decay and their decay rates could most fre-
quently be estimated. Furthermore, based on their t statistics, the viability decay rate estimates for total and
fecal coliform are also more reliable than for the other microorganism groups. Viability decay appeared to
occur on about 50 percent of the runs for coliphage, Clostridium perfringens and standard bacterial plate
county; their rates of decay were also slower than the coliform decay rates. Decay with aerosol age could sel-
dom be detected for mycobacteria, Pseudomonas, and fecal streptococci.
For all microorganism groups, the upper portion of the A distribution, which represents the
Table VI.C-7.
RUN ESTIMATES OF MICROORGANISM GROUP VIABILITY DECAY RATE A AND STANDARD
ERROR SE(A), s-i
a. Pre-Fair Runs
Aerosol
Run
Number
MM
Ml-2
Ml-3
M14
Ml-5
Ml-6
Ml-7
Ml-8
Ml-9
Ml-10
Ml-11
Ml-12
Ml-13
Ml-14
Ml-15
Ml-31
Ml-32
Ml-33
Ml-34
M1-3S
Standard
Plate Count
A.
-.19
X
-.014
-.018
X
X
-.042
-.134
-.165
-.0050
X
X
X
X
-.037
-.004
SE(X)
.13
.010
.030
.036
.070
.070
.0038
.037
.017
Total
Coliform
X
X
-.0316
-.33
-.383
-.062
-.003
-.089
-.025
-.094
-.077
-.093
X
-.025
X
-.0081
X
X
-.095
-.0497
SE(X)
.0066
.44
.023
.019
.013
.039
.032
.023
.014
.042
.013
.0028
.043
.0053
Fecal
Coliform
X
-.032
-.023
-.227
X
X
-.072
-.144
-.068
-.021
-.0073
X
X
-.063
SE(X)
.020
.070
.079
.017
.039
.017
.012
.0019
.015
Coliphage
X
X
X
-.108
X
-.051
X
-.065
-.0078
-.022
-.008
-.016
X
-.0037
-.15
-.19
-.028
SE(X)
.088
.021
.041
.0049
.016
.011
.011
.0078
.13
.25
.014
Fecal
Streptococci
X
X
-.081
X
-.022
X
-.039
X
X
X
X
X
X
SE(X)
.046
.017
.15
Pseudomonas
X
X
X
-.02
-.003
X
X
X
-.005
-.17
X
X
X
-.014
SE(X)
.12
.13
.015
Clostridium
Perfringens
X
X
X
X
-.118
-.039
X
-.048
X
-.020
-.018
-.0036
SE(X)
.047
.052
.037
.049
.017
.0002
X Viability decay rate estimate was positive.
151
-------
slower die-off rates, cannot be quantified based on the Pleasanton study. Table VI.C-10 suggests that —0.01s'
to —0.02s-' was the lowest viability decay rate which could be detected at Pleasanton. For the hardier and in-
frequently measured microorganisms, such as mycobacteria, Pseudomonas, and fecal streptococci, perhaps
—0.06s-' was the lowest detectable viability decay rate.
Because A is an exponential multiplier, e2Aa = (e*a)2; doubling the decay rate A. squares the ex-
ponential decay factor. Thus, when considering sizable aerosol ages, even differences in A of a factor of two
have a substantial effect. Using an average wind speed of 4 m/s, the median viability decay of total coliform
per 100 meters (25s) is exp(—.032s ' x 25s) = 0.45, slightly more than a two-fold reduction per 100 meters. The
quartile viability decay of total coliform per 100 meters is exp(—.094s-' x 25s) = 0.10, a ten-fold reduction per
100 meters. The corresponding decay rates for coliphage (a median 30 percent reduction per 100 meters and a
quartile 3-1/2 fold reduction per 100 meters) are substantially less. Consequently, Table VI.C-10 indicates that
age decay is considerably more rapid and prevalent for the coliforms than for coliphage, Clostridium perf-
ringens, and standard bacterial plate count. The pathogens fecal streptococci, Pseudomonas, and mycobacte-
ria seldom exhibited detectable die-off with aerosol age.
Table VI.C-7.
RUN ESTIMATES OF MICROORGANISM GROUP VIABILITY DECAY RATE A AND STANDARD
ERROR SE(A), g-i
b. Post-Fair Runs
Aerosol
Run
Number
M2-1
M2-2
M2-3
M24
M2-5
M2-6
M2-10
M2-11
M2-12
M2-13
M2-14
M2-15
M2-16
M2-17
M2-22
M2-23
M2-24
M2-25
M2-26
M2-29
M2-30
M2-31
M2-32
M2-33
M2-34
M2-35
M2-36
M2-37
M2-38
Standard
Plate Count
X
-.0044
X
-.022
X
X
-.006
X
-.007
X
-.1069
-.090
X
-.0185
X
-.010
X
-.0032
SE(\)
.0039
.040
.027
.033
.0004
.031
.0034
.024
.0027
Total
Colifoim
X
-.0757
-.014
-.0162
-.0253
-.0082
-.0393
X
-.033
-.02
X
-.1137
-.0197
-.0394
-.0379
-.083
-.133
-.158
-.250
X
-.0015
X
-.0077
-.17
-.214
-.237
SE(X)
.0026
.025
.0011
.0076
.0038
.0008
.017
.10
.0079
.0068
.0086
.0069
.022
.053
.036
.020
.0005
.0032
.039
.028
Coliphage
X
X
-.030
-.031
-.0397
X
X
-.048
X
X
-.055
-.019
X
X
-.116
-.008
-.061
-.119
-.089
X
-.011
-.091
X
X
-.0329
X
-.042
X
SE(\)
.020
.024
.0038
.041
.018
.038
.050
.070
.033
.035
.023
.020
.037
.0033
.081
Fecal
Streptococci
X
-.015
X
X
-.067
X
X
X
X
X
X
-.0062
X
X
-.174
X
X
X
,
X
-.0127
SE(X)
.048
.014
.0078
.017
.0068
Mycobacteria
X
-.009
X
X
X
X
-.134
X
X
SE(\)
.091
.042
X Viability decay rate estimate was positive.
152
-------
Table VI.C-8. .
PERCENTAGE OF INDETERMINATE VIABILITY DECAY RATE ESTIMATES (A - X)
Microorganism Group
Std. Bacterial Plate Count
Total Coliform
Fecal Coliform
Coliphage
Fecal Streptococci
Pseudomonas
Clostridium perfringens'
Mycobacteria
Number
of X
Estimates
33
44
13
43
31
13
11
8
Percentage of
Indeterminate
Estimates (X =
45%
:o%
31%
•10%
74%
62%
45%
75%
Table VI.C-9.
RELIABILITY OF NEGATIVE VIABILITY DECAY RATE ESTIMATES A < 0
Microorganism Group
Std. Bacterial Plate Count
Total Colifonn
Fecal Colifonn
Coliphage
Fecal Streptococci
Clostridium perfringens
No. of Negative Percentage of Negative Estimates Exceeding
(X,s.e.(X)) Two Std. Errors" " Std. Error
Estimates
Ifc
34
9
26
8
6
lt|>2
2%
9%
7%
1%
3%
3%
_T
61%
85%
89%
73%
75%
67%
Note: t = X/s.e.(X).
153
-------
Table VI.C-10.
DISTRIBUTIONS OF VIABILITY DECAY RATE k,
Viability Decay Rate *. *-»
Microorganism Croup
Total Col i form
Fecal Col i fora
Coliphage
Clost rid turn, ^erfrjngens
« Std. Bacterial Plate Count
Mycobacteria
Pseujomonas
Fecal Streptococci
Nunber of
Estimates
(Aerosol
Runs)
44
13
43
11
33
8
13
31
Distribution of
5% 10% 25%
-.31 -.23 -.094
-.19 -.070
-.14 -.11 -.051
-.10 -.039
-.17 -.12 -.020
(20%)
/••'-.IS -.027(7) -.009(7)
-.077 -.016(7) -.008(7)
-.12 -.060(7) -.014(7) -.006(7)
40%
-.050
-.045
-.029
-.019
-.006
X
X
X
50%
-.032
-.023
-.011
-.004(7)
-.004(7)
X
X
X
60%
-.020
-.016(7)
X
X
X
X
X
X
75%
-.004(7)
X
X
X
X
X
X
'X
90% 95%
X X
X X
X X
X
X X
X
X X
X -- Indeterminate viability decay rate.
/\/v -- Extrapolated value v.
v(?) -- Questionable value, perhaps indistinguishable from 0.
-------
S. Prediction Using the Microbiological Dispersion Model
Equation (3) and the data from Tables VI.C-6 and VI.C-10 define the microbiological dispersion
model. Considerations involved in using this model to predict the microbiological aerosol concentrations P
downwind of any spray irrigation site are discussed in this section.
a. Usage Considerations
(I) Assumptions
In developing and using the microbiological dispersion model, several assumptions are
made.
• The major physical and biological processes that affect mi-
croorganism aerosol levels emanating from a sprayed waste-
water source are adequately represented by the multiplicative
form of the microbiological dispersion model equation (3) out
to distances of 500 meters to 1000 meters (aerosol ages of 100
seconds to 500 seconds) from the spray location.
• The die-off of a microorganism group that occurs during
aerosolization and transport in the aerosol state is caused by
factors such as meteorological conditions that have the same
effect at any spray irrigation site. Thus, a given level of the
controlling factors will produce the same reduction in viable
aerosolized microorganisms (i.e., yield the same values of the
parameters I and A) at any spray site.
• The microbiological aerosol concentrations obtained in field
studies vary somewhat depending on the sampling, shipping,
and assay procedures employed. Since the distributions of the
microbiological parameters I and A. were derived from the
Pleasanton study, the concentrations P predicted by the
model assume the use of the methods of the Pleasanton study
(high volume electrostatic precipitator samplers, special and
standard assay methods, etc.).
(2) Procedure
Several steps are involved in the process of predicting microorganism aerosol concen-
trations P in the wastewater aerosol downwind from a spray irrigation site under specified meteorological
conditions using the microbiological dispersion model.
• Microorganism Wastewater Concentrations. Select the preva-
lent and relevant microorganism group(s) and determine typ-
ical microorganism concentrations in the wastewater.
• Diffusion Model Aerosol Concentration D. Apply an appro-
priate atmospheric dispersion model to project each microor-
ganism wastewater concentration to its corresponding dif-
fused aerosol concentration at the downwind location.
Validated atmospheric dispersion models of varying sophisti-
cation<50,54,55,56,57) are available to calculate D. Model input
generally includes the configuration of sprayers, spray trajec-
tory, wastewater spray rate, the microorganism wastewater
concentrations, pertinent meteorological conditions for the
case considered, the local topography, and the distance to the
downwind location. Usually the centerline (peak) concentra-
tion at the downwind distance is computed as D.
155
-------
• Aerosolization Efficiency E. Select an aerosolization effi-
ciency estimate for the case considered based on the type of
spray equipment, spray head pressure, and pertinent meteor-
ological factors. Once a median estimate of E is developed for
a spray site, equation (5) may be used to help select an E value
for the case based on the specified meteorological conditions.
• Impact Factor I and Viability Decay Rate A. For each micro-
organism group, select the proper I and A percentile values
from their distributions given in Tables IV.C-6 and IV.C-10.
Section V, Discussion, provides guidance in choosing the
proper percentile based on the relevant meteorological condi-
tions. Because of uncertainties in the estimation process, se-
lecting values of I and A outside their middle ranges (25th to
75th percentiles) is not recommended.
• Predicted Microorganism Aerosol Concentration P. The pre-
dicted microorganism concentration in the wastewater aero-
sol is then calculated from the estimates of D, E, I, and A
using equation (3).
b. Examples
(1) Pleasanton Residential Example
The microbiological dispersion model was used to obtain order-of-magnitude estimates
of the microorganism aerosol concentrations to which the residents are typically exposed in the Pleasanton
subdivision nearest the spray fields. The edge of this subdivision is located about 650 meters east and south-
east from the edges of the nearest spray fields.
The model input conditions, parameter values, and predicted aerosol concentrations
are presented in Table VI.C-11 for a typical summer nighttime case for total coliform, mycobacteria, and en-
teroviruses. Typical observed values were used for the wastewater concentrations and wastewater spray rate.
The H. E. Cramer Company Volume-Source Diffusion Models Program was used to calculate the centerline
concentration D for each microorganism group at the subdivision edge, taking into account the orientation of
the sprayer line when the subdivision would be downwind. A median aerosolization efficiency E = 0.0033 was
obtained from equation (5). As will be discussed later in the Discussion Section, the set of meteorological con-
ditions for this summer nighttime case tends to be associated with values of both I and A in the upper tails of
their distributions. Thus, the 60th percentile values of I in Table VI.C-6 and of A in Table VI.C-10 were se-
lected as typical values for the three microorganism groups. The centerline aerosol concentrations P predicted
at the subdivision edge are 0.01 mfc/m3 for total coliform, 0.09 cfu/m3 for mycobacteria, and 0.006 pfu/m3
forenteroviruses.
As a comparison, the aerosol concentrations of these microorganism groups at the sub-
division edge are also presented for a typical summer midday case in Table VI.C-12. Equation (5) yields an
aerosolization efficiency E = 0.016 for these meteorological conditions. Since the summer midday conditions
tend to give I and A values in the lower tails of their distributions, 40th percentile values for the microorga-
nisms were selected for this typical midday case. The predicted centerline aerosol concentrations were 0.001
mfc/m3 for total coliform, 0.06 cfu/m3 for mycobacteria, and 0.002 pfu/m3 for enteroviruses.
(2) Deer Creek Campsite Example
The microbiological dispersion model can also be applied at other spray irrigation sites.
For example, one can obtain order-of-magnitude estimates of the microorganism level extremes to which
156
-------
Table VI.C-1I.
PREDICTION OF TYPICAL NIGHTTIME MICROORGANISM AEROSOL LEVELS ENTERING
PLEASANTON RESIDENTIAL AREA
Model Input Conditions
Season and Time
Air Temperature
Relative Humidity
Solar Radiation
Wind Velocity
Stability Class
Mixing Height
Residential Distance
Residential Direction
Aerosol Age, a
Wastewater Spray Rate
Sumner Nighttime Case
summer night
20° C
70%
0 W/m2
2 m/s
E
30 m
650 m
E to SE
325 s
70 i/s
MICROORGANISM GROUP
Wastewater Concentration
Total Coliform Mycobacteria Enteroviruses
1 x 107 mfc/4 80,000 cfu/i 50 pfu/i
Model Parameter Values
D (Centerline)
E
I
. A. a
12,000 mfc/m3
0.0033
0.23
-0.020 s"1
0.0015
100 cfu/m3 0.06 pfu/m3
0.0033 0.0033
1.0 60
-0.004* s"1 -0.002* s
0.27 0.52
-1
Predicted Aerosol Concentration
P (Centerline)
0.01 mfc/m3
0.09 cfu/m3 0.006 pfu/m3
*Interpolated value between the quantified range of the model parameter and zero.
157
-------
Table VI.C-12.
PREDICTION OF TYPICAL MIDDAY MICROORGANISM AEROSOL LEVELS ENTERING
PLEASANTON RESIDENTIAL AREA
Model Input Conditions
Season and Time
Air Temperature
Relative Humidity
Solar Radiation
Wind Velocity
Stability Class
Mixing Height
Residential Distance
Residential Direction
Aerosol Age, a
Wastewater Spray Rate
Summer Midday Case
summer midday
30° C
40%
1000 W/m2
4 m/s
B
High
650 m
E to SE
162.5 s
70 i/s
MICROORGANISM GROUP
Wastewater Concentration
Total Coliform
1 x 107 mfc/Jt
Mycobacteria
80,000 cfu/i
Enteroviruses
50 pfu/JZ,
Model Parameter Values
D (Centerline)
E
I
A
„ Aa
1800 mfc/m3
0.016
0.13
-0.050 s'1
0.0003
15 cfu/m3 0.009 pfu/m3
0.016 0.016
0.82 30
-0.007*s"1 -0.004*s'1
0.32 0.52
Predicted Aerosol Concentration
P (Centerline)
0.001 mfc/m3
0.06 cfu/m3 0.002 pfu/m3
Interpolated value between the quantified range of the model parameter and zero.
158
-------
campers are exposed at an 80-KX) trailer campsite located from 700 to 900 meters northeast of the spray field
at Deer Creek Lake, Ohio.
Spray irrigation is accomplished at Deer Creek Lake with 96 Rainbird® impact spray-
ers arranged in an 8 x 12 grid in a 3-acre square field (sides of 150 meters). The U.S. Army conducted a field
sampling and assay program at the Deer Creek Lake site in the summer of 1976<58). Wastewater samples were
assayed routinely for indicator microorganisms and occasionally for pathogens. The microbiological aerosol
runs were performed using twelve Andersen samplers and two Litton Model M high volume samplers, with
samples only assayed for total aerobic bacteria. Four dye aerosol runs using fluorescein dye were also made.
Predictions of campsite microorganism aerosol level extremes emanating from the
spray field made by using the microbiological dispersion model are presented in Table VI.C-13 for a daytime
and a nighttime case. Typical values for the site were used for the wastewater concentrations of total coliform
and fecal streptococci and for the wastewater spray rate. Calculations based on applying the Volume-Source
Diffusion Models Program to the sampling data*59' were used to estimate the centerline concentration D at the
campsite edge and typical aerosolization efficiencies E for the daytime and nighttime cases. Considering the
meteorological conditions, the extreme values selected for I and A. were the 25th percentiles for the daytime
case and the 75th percentiles for the nighttime case. The predicted extreme daytime and nighttime aerosol
concentrations P at the campsite edge are, respectively, 2 x 10~9 and 0.05 cfu/tn3 for total coliform, and
0.0004 and 0.01 cfu/m3 for fecal streptococci.
6. Preliminary Evaluation of Distance and Solar Radiation Factors
Prior to developing the microbiological dispersion model, several preliminary analyses of the aero-
sol concentration data were performed to investigate some of the fundamental assumptions of the model.
These analysis results are presented in detail in Appendix G. The following synopsis indicates the purpose and
findings of each analysis.
a. Analysis of Variance
The significance of distance and solar radiation as factors affecting microbiological aerosol
levels was investigated by analysis of variance of the Post-Fair data. Except for the standard bacterial plate
count, aerosol levels varied significantly with sampler distance. The aerosol levels of standard bacterial plate
count, total coliform, and fecal streptococci were significantly reduced on runs made during high solar radia-
tion.
b. Source and Distance Analysis
The pairs of aerosol measurements from two sampler locations were compared using paired
comparison tests. Comparison of upwind versus close downwind concentrations demonstrated that the spray
line was a significant aerosol source of each of the microorganism groups monitored. Comparison of close
downwind versus distance downwind data only yielded significant aerosol concentration decreases with dis-
tance for total coliform, fecal coliform, coliphage, and fecal streptococci. Failure to find a significant de-
crease with distance for the other microorganism groups reflects the variability of their aerosol concentration
measurements and/or a small number of runs.
7. Preliminary Assessment of Factors Affecting Microbiological Aerosol Levels
A preliminary assessment was conducted after the Pre-Fair sampling of the environmental factors
affecting microbiological aerosol levels on the Pre-Fair aerosol runs. Stepwise multiple linear regression was
used to select, from a large list of plausible candidates, these environmental variables in the D, E, I, and A
categories which best fit the microorganism aerosol data. A full description of the analysis and its results is
given in Appendix G.
159
-------
Table VI.C-13.
PREDICTION OF MICROORGANISM AEROSOL LEVEL EXTREMES ENTERING
DEER CREEK LAKE CAMPSITE
Daytime Case
Nighttime Case
S
noaei input uonaicions
Season and Time
Air Temperature
Relative Humidity
Solar Radiation
Wind Velocity
Stability Class
Mixing Height
Residential Distance
Residential Direction
Aerosol Age, a
Wastewater Spray Rate
MICROORGANISM GROUP
Wasteuater Concentration
Model Parameter Values
D (Centerline)
E
I
A
eXa
Predicted Aerosol Concentration
summer midday
30° C
40Z
1,000 W/m2
4 m/a
B
high
700 m
NNE to ENE
175 8
30 1/8
Total Fecal
Coliform Streptococci
180,000 cfu/i 1,000 cfu/t
40 cfu/m3 0.2 cfu/m3
0.009 0.009
0.06 0.71
-0.094 s'1 -0.006 S"1
7 x 10'8 0.35
summer night
20° C
70Z
0 W/mz
2 a/8
E
30 a
700 m
NNE to ENE
350 8
30 i/s
Total Fecal
Coliform Streptococci
180,000 cfu/i 1,000 cfu/t
100 cfu/m3 0.6 cfu/n3
0.004 0.004
0.55 6.1
-0.004 s"1 -0.001* a"1
0.25 0.70
P (Centerline)
2 X 10~9 cfu/m3 0.0004 cfu/m3
0.05 cfu/m3 0.01 cfu/m3
Interpolated value between the quantified range of the model parameter and zero.
-------
Similar factors were identified as explaining much of the observed variation in the aerosol levels of
total coliform, fecal coliform, and coliphage. The diffusion model D was found to be important for all three
microorganisms. Temperature was indicated to have an important effect on wastewater aerosolization effi-
ciency, while low pond relative humidity and middle upwind relative humidity both had lethal impact effects.
Viability decay occurred primarily at middle relative humidities and at high temperature.
There is some evidence that the three pathogenic bacteria, Pseudomonas, fecal streptococci, and
Clostridium perfringens, may also be affected by similar environmental variables. All appear to be initially
reduced in viability by strong, dry winds. Desiccation also appears to play a role in their viability decay with
aerosol age or distance.
This preliminary assessment of environmental factors was important, both because it justified the
form of the microbiological dispersion model equation (3), and because it represents the most comprehensive
analysis to date of the effects that atmospheric and operating conditions have on the microorganism-specific
parameters I and A, and thereby on the aerosol concentration.
D. Evaluation of the Microbiological Dispersion Model
The predictive value of the microbiological dispersion model is determined by how well its prediction, P,
of the microorganism aerosol concentration from a spray source agrees with C-B, the measured concentration
corrected for background. In this section, model predictions are compared to the aerosol concentration data
from the Pleasanton runs, which were not used in the model development, and from field sampling programs
conducted at Deer Creek Lake, Ohio*60) and Fort Huachuca, Arizona'61-62'63'. Information is presented on the
accuracy and precision of the model predictions.
1. Evaluation Data
To evaluate the suitability of the model for a microorganism group, it is necessary to obtain many
pairs (C-B,P) of the net measured concentration, C-B, and the model-predicted concentration, P, for the mi-
croorganism group from various distances, runs and sites. Consistent procedures were used for calculating
values of C-B and P and for deciding whether to use these values in the model evaluation.
From the Pleasanton sampling program, the smoothed concentration values, C , not used in devel-
oping the microbiological dispersion model, were potentially usable in the model evaluation. These consisted
of a few values from the microbiological aerosol runs in the tables of Appendix F and all the quality assurance
run values (Tables VI.B-16 through VI.B-23). Nearly all the concentration values above background from the
Deer Creek Lake and Fort Huachuca sampling programs were potentially usable in the model evaluation. The
exceptions were several cases of presumed contamination in which very high outlier values relative to the sur-
rounding values were obtained, and several cases in which chlorinated effluent was sampled. At all sites, a line
of samplers, or at least a pair of samplers, was often located at the same distance from the spray source.
Sometimes a pair of samplers consisted of one high-volume aerosol sampler and one Andersen sampler. At
Ft. Huachuca, several downwind samplers were deployed along an arc at 10° to 40° intervals at the same dis-
tance from the single spray source. These situations were treated by averaging the concentrations from paired
high-volume and Andersen samplers, by averaging the concentrations along a line of samplers, and by averag-
ing the concentrations over an arc of samplers. Thus, no more than one net concentration value, C-B, was
used at a given sampler distance on a run. The measured concentration values from all sites were used for
model evaluation only if they exceeded both the run background (C-B > 0.5B) and the minimum detection
limit (C-B >DL).
Predicted values corresponding to the net measured concentration value were computed from
equation (3), P = D • E • I exp (A • a), using the procedures given in Section VI.C.S.a.2. Diffusion model
aerosol concentrations at each sampler position were calculated by H.E. Cramer Company, using their Vol-
161
-------
ume Source Diffusion Model.*64' Calculation of aerosolization efficiency from the dye runs at Deer Creek
Lake and Fort Huachuca, using equation (4) E = C/D, with Cramer's modeled D values, showed that the
aerosolization efficiencies of the Deer Creek Lake and Fort Huachuca spray systems were similar to Pleas-
anton. Consequently the aerosolization efficiency, E, during each microbiological run at Deer Creek Lake
and Fort Huachuca was computed from equation (5).
The solar radiation measurement, r, used in equation (5) should be recorded by a short-wave in-
strument of the vertical Eppley-type for consistency with the Belfort short-wave instrument used at Pleas-
anton. The incoming radiometer readings, in millivolts, from the Deer Creek Lake and Fort Huachuca re-
ports, were adjusted to equivalent short-wave values based on hourly cumulative vertical-Eppley readings
provided by Mr. Marmon of the Atmospheric Sciences Laboratory, White Sands, New Mexico. A percentile
of the I and A distributions was selected for each run following the guidance given in Section VII, below. Solar
radiation was used as the primary criterion in selecting the 25th, 40th, 50th, 60th or 75th percentile, with con-
sideration occasionally given to additional factors, such as run selectivity at Pleasanton and effluent chlorina-
tion at Fort Huachuca. In a few cases, no predicted value could be computed to correspond to a C-B value,
either because the run was not diffusion-modeled or because no wastewater concentration was measured.
The numbers of data pairs (C-B.P) for the model evaluation are presented in Table VI.D-1 by mi-
croorganism group and site. The first column gives the number of pairs for which a detectable measured value
and model prediction were obtained; these are the potential evaluation data. The number of these pairs, for
which the measured concentration is detectable (C-B > 0.5B and C > DL) is given in the second column; these
pairs were used to evaluate the precision of the model predictions. The number of these pairs for which the
predicted concentration would also be detectable (P > 0.5B and P > DL) is shown in the next column; these
were used to analyze the accuracy of the model predictions. Table VI.D-1 shows that standard bacterial plate
count, total coliform, and possibly coliphage are the only microorganism groups having enough data pairs to
perform an adequate model evaluation. For the other microorganism groups, nearly all of the usable data
pairs come from the Pleasanton sampling program.
It should be noted that the sampling and analytical methods used at Deer Creek Lake and Fort
Huachuca differed from those employed at Pleasanton. The aerosol samplers for the Deer Creek Lake and
Fort Huachuca runs were mainly six-stage Andersen samplers supplemented by one or two high-volume aero-
sol samplers, whereas only high-volume samplers were used at Pleasanton. Presumptive identification of total
coliforms using Endo agar or Endo broth was used at Deer Creek Lake and Fort Huachuca; at Pleasanton,
presumptive total coliforms on lactose broth were confirmed on brilliant green bile . At Fort Huachuca, col-
iphage measurements were obtained by seeding the wastewater with coliphage f2 and sampling this aerosol
with Andersen and high-volume samplers. Since laboratory studies indicated that the recovery efficiency of
airborne f2 from Andersen samplers was about 25 percent of that of liquid impaction samplers,*65' the field
coliphage f2 concentrations obtained with the Andersen samplers were multiplied by four. At Pleasanton, the
natural coliphage present in the wastewater were sampled and assayed. The preceding differences might affect
the model evaluation, since the I and A coefficients of the microbiological dispersion model are based on the
Pleasanton sampling and analytical methods.
Each data pair used in the model evaluation (i.e. those with detectable measured net concentra-
tions C-B for which totals are given in the second column of Table VI.D-1) is tabulated in Appendix H. The
tables of Appendix H are arranged by microorganism group, site, and analytical method. Displayed in the
Appendix H tables for each data pair are the aerosol run number, key meteorological parameters (tempera-
ture, relative humidity, solar radiation, and wind velocity), the selected percentile of the I and A distributions,
the measured aerosol concentration, C-B, the model-predicted aerosol concentration, P, and the orders of
magnitude of discrepancy statistic, OMD.
162
-------
In comparing the measured and predicted aerosol concentrations, proportional differences are
more relevant than absolute differences. For example, the discrepancies between the microorganism aerosol
concentration pair (10,000/m3 vs. 1,000/m3) and between the pair (10/m3 vs. 1/m3 )might be considered
Table VI.D-1.
NUMBER OF DATA PAIRS FOR MODEL EVALUATION
Number of Data Pairs
Microorganism Group
Site
Std. Bacterial Plate Count
Pleasanton
Deer Creek Lake
Ft. Huachuca
All Sites and Methods
Total Coliform
Pleasanton
Presumptive on Endo Agar:
Ft. Huachuca
Presumptive on Endo Broth:
Deer Creek Lake
Ft. Huachuca
All Endo Broth
All Sites and Methods
Fecal Coliform
Pleasanton
Ft. Huachuca
All Sites and Methods
Coliphage
Pleasanton
Ft. Huachuca (seeded f2)
All Sites and Methods
Fecal Streptococci
Pleasanton
Ft. Huachuca
All Sites and Methods
Pseudomonas
Run/Distance
Combinations
Measured and Modeled
9
60
50
119
8
30
10
13
23
61
6
4
10
5
16
21
10
8
18
Detectable* Measured
Concentrations
Number
7
54
37
98
8
24
8
9
17
49
5
2
7
5
15
20
8
2
10
Detectable*
Measured and
Predicted Concentrations
Number
6
45
34
85
8
17
0
7
7
32
5
1
6
5
15
20
8
0
8
Effective No
6
28.5
24.5
59
8
13.5
0
6.5
6.5
28
5
1
6
5
10.5
15.5
8
0
8
Pleasanton
Clostridium perfringens
Pleasanton
Mycobacteria
Pleasanton
* Above background (C-B>0.5B, P>0.5B) and detection limit (C> DL, P>DL),
and excluding presumably contaminated samples.
f Presumed equivalent number of independent observations.
163
-------
equivalent because both are "factor of 10" or "one order of magnitude" discrepancies. Hence, the model
evaluation is based on these two proportional difference statistics which are easily interpreted. The discrep-
ancy factor statistic, F, is defined as the larger of (C-B)/P and P/(C-B):
(( R P \
• 1 (14)
P C-B/
The orders of magnitude of discrepancy statistic is defined as:
OMD = lo
- (T)
(15)
These statistics are mathematically related since OMD = logl()F and F = 10l°MDl. Thus, three orders of
magnitude discrepancy (OMD = ± 3) is equivalent to a discrepancy factor of 1000 : F = 10-' -- 1000.
A scan of the Appendix H tables shows that the greater majority of model predictions differ from
the measured aerosol concentrations by less than an order of magnitude (—1 < OMD < 1). The accuracy and
precision of the model predictions are more fully characterized in the following sections.
Applicability of the model to chlorinated effluent aerosols can be inferred from the model predic-
tions for standard bacterial plate count on several runs at Fort Huachuca, in which the effluent was chlori-
nated prior to spraying. The results are presented in Table VI.D-2. Measurement of the aerosol concentra-
tions of standard bacterial plate count on the chlorinated runs is quite uncertain, because the low net
measured concentrations of nearly 0 to 307 m-' were less than the measured background concentrations of 19
to 170/m3 on each run, except Run 74-6. Also, the diffusion modeling of some of these runs could only pro-
vide approximate concentrations because the wind direction varied so widely. Yet the microbiological dispers-
ion model regularly predicted an aerosol concentration several orders of magnitude lower than the measured
value (OMD > + 1.5). The marked underprediction with chlorinated effluent contrasted with the relatively
Table VI.D-2
COMPARISON OF MODEL PREDICTIVE ABILITY OF CHLORINATED
AND UNCHLORINATED EFFLUENT
Aerosol
Run
Number
Wastewater
Chlorination (mg/1)
Total Free
Standard Bacterial Plate Count
Effluent Aerosol Cone (CFP/m)
(CFP/ml) C-B P
OMD
Average
Chlorination
(Ft. Huachuca)
74-5
74-6
75-6
75-5
75-15
6.3
0.8
6.0
6.5
6.0
1.5
0.0
0.7
0.8
0.6
50
220
88
87
50
2.3
30
29
5
0
0.08
0.05
0.06
0.05
0.4
+ 1.5
+ 2.8
+ 2.7
+ 2.0
7
Low
Chlorination
(Deer Creek Lake)
Unchlorinated
(Ft. Huachuca)
Unchlorinated
(Pleasanton)
45 pairs 0.1-0.4
34 pairs
29,000
310,000
6pairs <0.1 <0.05 610,000
+ 0.26
+ 0.14
+ 0.02
164
-------
good model predictions for low chlorination, and especially for unchlorinated effluents, are summarized in
Table VI.D-2. Thus, it appears that the microbiological dispersion model does not give valid predictions of
microbiological aerosol concentrations from highly chlorinated effluents.
An hypothesis is offered to explain model underprediction for chlorinated effluent aerosols. The
types of microorganisms that are most susceptible to die-off during aerosolization may be the same types that
are generally killed by chlorination of wastewater. Hence, little impact die-off would occur (I = 1) during
aerosolization of chlorinated effluent.
2. Accuracy of Model Predictions
The predictions of the model will intuitively be regarded as being accurate if the predicted microor-
ganism concentrations fall randomly above and below the net measured concentrations with no detectable
bias. It is assumed that each OMD calculated from equation (15) for a group of data pairs is sampled from the
same normal distribution with unknown population mean MOMD and variance OOMD. If the mean MOMD is not
significantly different from 0, the model predictions may be considered accurate in a statistical sense. The null
hypothesis of model prediction accuracy (MOMD = 0) can be tested against the two-sided alternative, MOMD ^ 0.
using the test statistic t = OMD/(SOMD/\/Ti) which has the t distribution with n-1 degrees of freedom, pro-
vided the OMD's are independent observations from the same normal distribution. Here OMD = (1 OMD)/n
and SOMD = I (OMD — OMD)2/(n — 1) are the sample mean and sample variance, respectively.
An analysis of model prediction accuracy using the t statistic was conducted for each microorga-
nism group at each site and for the total OMD observations for each microorganism group over all three sites.
To prevent an artificial selection bias, the accuracy analysis was based on those data pairs from the Appendix
H tables, for which both the net measured concentrations and the predicted concentrations were detectable.
A scan of the Appendix H data reveals that the OMD values of a microorganism group for the
different sampling distances on an aerosol run are related. The OMD values at the various sampler distances
on a run all tend to have similar values, especially for samplers that were located close to each other. Con-
sider, for example, the OMD values for standard bacterial plate count from the Deer Creek Lake study on
Run 15 [—0.25 (30m), —0.39 (50m)] and on Run 16 [0.84 (30m), 0.88 (50m), 0.78 (200m)]. Since the OMD
values at different distances on a run are similar, the OMD values obtained for a group are not independent
observations, as is assumed in calculating a probability value for the t statistic by the usual procedure. Be-
cause the OMD values for a run are related, in effect there are less than n independent observations in a group
of data pairs. The "effective number of independent observations" was calculated by assuming that runs with
OMD values at three sampler distances provided two independent observations, that runs with OMD values at
two distances provided one and one half independent observations, and that runs with a single OMD value
contributed one independent observation. Both the actual number of OMD values and the presumed effective
number of independent observations are presented in the last two columns of Table VI.D-1 for each group of
detectable measured and predicted concentrations on which an accuracy analysis was conducted. The t statis-
tic and its two-sided significance level (p) were computed for both the number of observations and the effec-
tive number of independent observations. Although similar values were obtained for both cases, the effective
number is regarded as more valid, and inferences are based on this procedure.
Provided the number of independent observations is correctly determined, the p value is the prob-
ability of getting a more discrepant t statistic than the value obtained for a random sample of equal size from
a population with zero mean (MOMD = 0). Thus, p indicates the accuracy of the model predictions relative to
the net measured aerosol concentration. A p value above 0.20 suggests little or no consistent bias, while a p
value below 0.05 suggests there is a consistent bias in the model predictions.
The results of the accuracy analyses for standard bacterial plate count, total coliform, fecal col-
165
-------
iform, coliphage, and the pathogenic microorganisms are presented respectively in Tables VI.D-3 through
VI.D-7. The geometric mean of the P/(C-B) ratios is tabulated to indicate the average discrepancy between
the predicted and measured values. The OMD provides another measure of this discrepancy. For groups in
which the number of detectable data pairs differed from the effective number of independent observations,
the t statistic and its two-sided p value are presented for both cases.
Table VI.D-3 shows that the accuracy of the model predictions for standard bacterial plate count
differ considerably among the study sites. At Pleasanton, there is no evidence of bias in the model predic-
tions. At Fort Huachuca, the model predictions tended to be 72 percent as large as the measured net concen-
trations, but there was sufficient variability that no consistent bias was detected (p = 0.16). At Deer Creek
Lake, the model predictions of standard bacterial plate count were consistently (p = 0.0016) less than the mea-
sured values, averaging 55 percent as large. Referring to Table VI.D-2, the slight model prediction underesti-
mation with low chlorination at Deer Creek Lake is consistent with the apparent two orders of magnitude
model underestimation for heavily chlorinated Fort Huachuca effluent. Thus, chlorination may be responsi-
ble for the slight underprediction at Deer Creek Lake. The summarized results over all three sites are similar
to those for Deer Creek Lake, which provided the majority of the data pairs. With p = 0.0012, the model
predictions of standard bacterial plate count were found to be consistently smaller than the measured values,
averaging 64 percent as large.
Table VI.D-3
ANALYSIS OF THE ACCURACY OF STANDARD BACTERIAL PLATE COUNT
MODEL PREDICTIONS
Sites All sites
Pleasanton Deer Creek Lake Ft. Huachuca and Methods
Number of Detectable*
Measured and Predicted 6 45 34 85
Concentration Pairs
n
Geometric Mean 0.95 0.55 0.72 0.64
C—B
OMD Mean, OMD 0.02 0.26 0.14 0.20
OMD Standard Deviation 0.56 0.40 0.48 0.45
t - OMD/(SOMD/V''n) 0.02 4.40 1.72 4.08
p value (two-sided) >0.90 0.00005 0.10 0.0001
Effective+ Number 6 28.5 24.5 59
t = OMD/(SOMD/vrn') 3.50 1.46 3.40
p value (two-sided) 0.0016 0.16 0.0012
Orders of Magnitude of Discrepancy, OMD = log,
*Above background (C—B > 0.5B, P > 0.5B) and detection limit (C > DL, P > DL), and excluding
presumably contaminated samples.
+ Presumed equivalent number of independent observations.
166
-------
The analysis given in Table VI.D-4 provides no evidence of a consistent bias in model predictions
of total coliform aerosol concentration. With unchlorinated wastewater aerosols assayed for confirmed total
coliforms at Pleasanton and for presumptive total coliforms at Fort Huachuca both on Endo agar and on
Endo broth, no consistent proportional differences were found between the measured values and the model
predictions. Thus, prediction of total coliform aerosol concentrations with the microbiological dispersion
model for unchlorinated wastewater aerosols apparently yields accurate estimates of the measured values that
could be obtained using a variety of aerosol sampler types and assay methods.
Accuracy analysis of fecal coliform model predictions in Table VI.D-5 shows that the model tends
to underestimate measured fecal coliform aerosol concentrations by about a factor of 1/0.40 = 2.5. The im-
pact factor distribution, I, for fecal coliform is based on only 13 Pre-Fair aerosol runs at Pleasanton, which
may be insufficient to characterize this distribution.
The accuracy analysis in Table VI.D-6 indicates that model predictions of total coliphage aerosol
concentrations tended to overpredict the measured values at Pleasanton by a factor of 2.4, while possibly un-
derpredicting seeded coliphage aerosol concentrations at Fort Huachuca at about half the measured value.
However, considering both sites together, there is no evidence of consistent prediction bias.
Table VI.D-4
ANALYSIS OF THE ACCURACY OF TOTAL COLIFORM MODEL PREDICTIONS
Number of Detectable*
Measured and Predicted
Concentration Pairs
Confirmed
Pleasanton
Assay Methods
Presumptive
Endo Agar Endo Broth
Ft. Huachuca Ft. Huachuca
17
All Sites
and Methods
32
C—B
" Geometric Mean
1.38
0.62
OMD Mean, OMD
OMD Standard Deviation
t = OMD/(SOMD/v^n)
p value (two sided)
Effective+ Number
t = OMD/(SOMD/v/V)
p value (two sided)
Orders of Magnitude of Discrepancy, OMD = log
-0.14
0.55
-0.71
0.50
8
0.21
0.51
1.67
0.13
13.5
1.49
0.16
/C— B\
0.52
0.28
0.63
1.17
0.30
6.5
1.13
0.30
0.73
0.14
0.55
1.39
0.18
28
1.30
0.21
*Above background (C—B > 0.5B, P > 0.5B) and detection limit (C 3* DL, P ^ DL), and exclud-
ing presumably contaminated samples.
+ Presumed equivalent number of independent observations.
167
-------
Analyses of the accuracy of model predictions at Pleasanton of fecal streptococci, Pseudomonas,
Clostridium perfringens, and mycobacteria are given in Table VI.D-7. The analyses suggest a tendency for the
model to overpredict fecal streptococci and Pseudomonas aerosol concentrations. However, many of the data
pairs available for model evaluation of the bacterial pathogens were from "left over" runs in which few de-
tectable measured aerosol concentrations were obtained. It appears that this bias in selecting data pairs has
produced biased (C-B,P) observations that may have confounded the accuracy analysis for each of these bac-
terial pathogens.
The accuracy of the model predictions over all methods and sites is summarized in Table VI.D-8
for each microorganism group. An assessment of the probable bias that results from using the microbiologi-
cal dispersion model to predict microorganism aerosol concentrations is given in the last column of Table
VI.D-8.
Model predictions tend to slightly underestimate the standard bacterial plate count and fecal col-
iform aerosol concentrations, giving predictions that average about half as large as the anticipated measured
values. Given the imprecision of measured aerosol concentrations of standard bacterial plate count and fecal
coliform (i.e. coefficients of variation of 50 and 58 percent respectively, were presented in Table VI.B-40),
and the variety of sampling methods, analytical methods, and effluent aerosols evaluated, the slight underes-
timation bias in model predictions of standard bacterial plate count and fecal coliform aerosol concentrations
is not considered to be of practical significance.
The accuracy analysis of the model predictions of the bacterial pathogens was based on few obser-
vations (C-B,P), some of which may not be valid for assessing accuracy. Thus, the accuracy of model predic-
tions of bacterial pathogen aerosol concentrations is still regarded as unknown.
Table VI. D-5.
ANALYSIS OF THE ACCURACY OF FECAL COLIFORM MODEL PREDICTIONS
Sites
Pleasanton Ft. Huachuca All Sites and Methods
Number of Detectable*
Measured and Predicted
Concentration Pairs
Geometric Mean 0.41 0.36 0.40
C—B
OMD Mean, OMD 0.39 0.44 0.40
OMD Standard Deviation 0.31 0.28
t = OMD/(SOMD/\An) 2.83 3.52
p value (two-sided) 0.048 0.017
Orders of Magnitude of Discrepancy, OMD = log,
•Above background (C—B ^ 0.5B, P > 0.5B) and detection limit (C > DL, P > DL), and
excluding presumably contaminated samples.
+ Presumed equivalent number of independent observations.
168
-------
Although varying results were obtained at different sites using different methods, the accuracy
analyses for all sites did not detect a consistent significant bias in the model predictions of the aerosol concen-
trations of total coliform and coliphage. While the overall accuracy analyses involved enough (C-B.P) obser-
vations to detect a substantial bias if it had existed, only slight site-specific or nonspecific biases were found.
Thus, for most model applications, the predictions of total coliform and coliphage aerosol concentrations
with the microbiological dispersion model appear to be sufficiently accurate estimates of the measured aero-
sol concentrations that would be obtained with various types of aerosol samplers and assay methods.
3. Precision of Model Predictions
The precision of the model predictions refers to how close the predicted aerosol concentrations
tend to be to the net measured aerosol concentrations regardless of which value tends to be larger. Thus, the
precision of a prediction relative to C-B, can be measured in terms of the discrepancy factor statistic F defined
in equation (14) or of the absolute value of the orders of magnitude of discrepancy statistic |OMD| .
An analysis of model prediction precision was conducted for each microorganism group at each
site and for all three sites. All the data pairs given in the Appendix H tables (i.e., all pairs for which the mea-
sured aerosol concentration was detectable) were considered valid measures of microbiological dispersion
model precision and thus were utilized in the precision analysis.
Table VI. D-6
ANALYSIS OF THE ACCURACY OF COLIPHAGE MODEL PREDICTIONS
Sites
Natural Seeded f2
Pleasanton Ft. Huachuca All Sites and Methods
Number of Detectable*
Measured and Predicted 5 15 20
Concentration Pairs
C—B
Geometric Mean 2.4 0.49 0.73
OMDMean, OMD -0.38 0.31 0.14
OMD Standard Deviation 0.10 0.48 0.52
t = OMD/(SOMD/v^n) -8.5 2.51 1.19
p value (two-sided) 0.0011 0.0025 0.25
Effective* Number 5 10.5 15.5
t = OMD/(SOMD/v^n') 2.10 1.05
p value (two-sided) 0.06 ~ 0.32
Orders of Magnitude of Discrepancy, OMD = log,
*Above background (C—B > 0.5B, P > 0.5B) and detection limit (C > DL, P ^ DL), and
excluding presumably contaminated samples.
+ Presumed equivalent number of independent observations.
169
-------
Table VI. D-7
ANALYSIS OF THE ACCURACY OF PATHOGENIC MICROORGANISM MODEL
PREDICTIONS AT PLEASANTON
Fecal Clostridium
Streptococci Pseudomonas Perfringens Mycobacteria
Number of Detectable*
Measured and Predicted 8424
Concentration Pairs
Geometric Mean
C—B
OMD Mean, OMD
OMD Standard Deviation
4.2
5.9
7.5
0.38
-0.62
0.36
-4.89
0.0018
-0.77
0.24
-6.53
0.007
-.088
0.81
-1.52
~0.35
0.43
0.53
1.70
0.19
t-OMD/(SOMD/Vn)
p-value (two-sided)
Orders of Magnitude of Discrepancy, OMD = log,0'
*Above background (C—B 3* 0.5B, P > 0.5B) and detection limit (C 3» DL, P > DL), and ex-
cluding presumably contaminated samples.
+ Presumed equivalent number of independent observations.
Table VI. D-8
SUMMARY OF MODEL PREDICTION ACCURACY FOR ALL SITES AND METHODS
Microorganism Group
Standard Bacterial
Plate Count
Total Colif orm
Fecal Coliform
Coliphage
Fecal Streptococci
Pseudomonas
Clostridium perfringens
Mycobacteria
Detectable
Measured and
Predicted
Pairs
85
32
6
20
8
4
2
4
Factor
P/(C-B)
Geometric
Mean
0.64
0.73
0.40
0.73
4.2
5.9
7.5
0.38
Mean of
Orders of
Magnitude of
Discrepancy
0.20
0.14
0.40
0.14
-0.62
-0.77
-0.88
0.43
P Value
(two-sided)
for Effective
No.ofObs.
0.0012
0.21
0.017
~ 0.32
0.0018
0.007
~0.35
0.19
Assessment of
Prediction Bias
slight underestimate
unbiased
slight underestimate
unbiased
unknown
unknown
unknown
unknown
170
-------
The results of the precision analyses for standard bacterial plate count, total coliform, fecal col-
iform, coliphage, fecal streptococci, and the pathogenic microorganism assayed only from Pleasanton are
presented respectively in Tables VI.D-9 through VI.D-14. The mean of IOMD I and the largest OMD value
summarize the prediction precision for a data group in terms of the orders of magnitude of the discrepancy.
The geometric means of the discrepancy factors F give the average size of the factor of discrepancy between
C-B and P. The distribution of the discrepancy factors is described by presenting the percentages of these fac-
tors below 2 (i.e., C-B and P within a factor of 2 of each other), below 5, and below 10.
The precision analysis of standard bacterial plate count model predictions is presented in Table
VI.D-9 separately for Pleasanton, Deer Creek Lake and Fort Huachuca, and together for all three sites. Satis-
factory and very similar precision was found in the standard bacterial plate count model predictions at all
three sites. The geometric mean of the discrepancy factors ranged only from 2.64 at Fort Huachuca to 2.89 at
Pleasanton, with an average discrepancy of a factor of 2.70 between P and C-B over all 98 data pairs at the
three sites. At each site, 43 percent of the (C-B,P) pair values differed from each other by less than a factor of
2. Ninety-five percent of the data pairs at the three sites had a discrepancy factor below 10 (i.e., less than a
one order of magnitude discrepancy). Thus, the microbiological dispersion model predicts aerosol concentra-
tions of standard bacterial plate count from unchlorinated and slightly chlorinated wastewater (< 0.4 mg/1
total chlorine) quite well. The model's predictive ability for standard bacterial plate count appears to be
equivalent at Pleasanton, Deer Creek Lake, and Fort Huachuca.
The precision analysis given in Table VI.D-10 for total coliform model predictions shows consider-
able differences in predictive precision from one site to another. The precision of model predictions of con-
firmed total coliforms at Pleasanton compares favorably with standard bacterial plate count precision at
Table VI. D-9
ANALYSIS OF THE PRECISION OF STANDARD BACTERIAL PLATE COUNT
MODEL PREDICTIONS
Sites
Pleasanton Deer Creek Lake Ft. Huachuca All Sites and Methods
(C-B.P) Pairs with
Detectable* Measured 7 54 37 98
Concentrations
Mean of |OMD| 0.46 0.43 0.42 0.43
Largest OMD +0.92 +1.15 +1.23 +1.23
Percentage of Discrepancy
Factors:
Below 2 43% 43% 43% 43%
Below5 57% 80% 76% 77%
Below 10 100% 93% 97% 95%
Geometric Mean of
Discrepancy Factors F 2.89 2.72 2.64 2 70
*Above background (C-B > 0.5B) and detection limit (C > DL) and excluding presumably contaminated sam-
ples.
171
-------
Table VI. D-10
ANALYSIS OF THE PRECISION OF TOTAL COLIFORM MODEL PREDICTIONS
Assay Methods
(C-B,P) Pairs with
Detectable* Measured
Concentrations
Mean of lOMDl
Largest OMD
Percentage of Discrepancy
Factors:
Below 2
Below 5
Below 10
Confirmed
Pleasanton
8
0.41
-1.18
50%
88%
88%
Presumptive
Endo Agar
Ft. Huachuca
24
0.72
+ 2.37
21%
67%
71%
Endo Broth
Ft. Huachuca
9
0.65
+ 1.43
33%
67%
67%
Deer Creek Lake
8
1.50
+ 2.09
0%
0%
13%
All Methods at
Unchlorinated
Sites
41
0.64
+ 2.37
29%
71%
73%
Geometric Mean of
Discrepancy Factors F
Total Chlorine in
Wastewater (mg/1)
2.55
5.2
none
4.5
none
32
0.1-0.4
4.4
* Above background (C-B > 0.5B) and detection limit (C > DL) and excluding presumably contaminated samples.
-------
Table VI. D-ll
ANALYSIS OF THE PRECISION OF FECAL COLIFORM MODEL PREDICTIONS
Sites
Pleasanton Ft. Huachuca All Sites and Methods
(C-B,P) Pairs with
Detectable* Measured
Concentrations
Mean of
Largest OMD
Percentage of Discrepancy
Factors:
Below 2
Below 5
Below 10
Geometric Mean of
Discrepancy Factors F
0.39
+ 0.88
60%
80%
100%
2.47
1.43
+ 1.97
0%
50%
50%
27
0.69
+ 1.97
43%
71%
86%
4.9
*Above background (C-B > 0.5B) and detection limit (C > DL) and excluding presumably
contaminated samples.
Table VI. D-12
ANALYSIS OF THE PRECISION OF COLIPHAGE MODEL PREDICTIONS
(C-B,P) Pairs with
Detectable* Measured
Mean of I OMD I
Largest OMD
Percentage of Discrepancy
Factors:
Below 2
Below 5
Below 10
Sites
Natural
Pleasanton
5
0.38
-0.52
Seeded f 2
Ft. Huachuca
15
0.48
+ 0.98
All Sites and Methods
20
0.46
+ 0.98
20%
100%
100%
27%
73%
100%
25%
80%
100%
Geometric Mean of
Discrepancy Factors F 2.42 3.04 2.87
*Above background (C-B > 0.5B) and detection limit (C > DL) and excluding presumably
contaminated samples.
173
-------
Pleasanton or any other site (e.g. an F geometric mean of 2.55 for total coliform at Pleasanton versus stan-
dard bacterial plate count values of 2.89 at Pleasanton and 2.70 at all sites). The precision of model predic-
tions of presumptive total coliforms at Fort Huachuca is considerably worse, for assays both on Endo agar (F
geometric mean = 5.2) and on Endo broth (F geometric mean = 4.5). At Deer Creek Lake, the model predic-
tions of presumptive total coliforms on Endo broth were consistent one-to-two orders of magnitude underes-
timates of the net measured aerosol concentrations; the geometric mean of the discrepancy factors was 32.
These site differences in model predictive ability may be related to the amount of chlorine in the sprayed wast-
ewater (shown at the bottom of Table VI.D-10). Since model underpredictions of about two orders of magni-
tude for standard bacterial plate count were observed on the runs at Fort Huachuca using highly chlorinated
wastewater (see Table VI.D-2), it is plausible to attribute model underpredictions of one-to-two orders of
magnitude for the fragile total coliforms at Deer Creek Lake to the slightly chlorinated wastewater (0.1-0.4
mg total chlorine per liter).
As with total coliform, the analyses of fecal coliform (Table VI.D-11), coliphage (Table VI.D-12),
and fecal streptococci (Table VI.D-13) show better precision for the model predictions at Pleasanton than at
Fort Huachuca, although the distinction for coliphage is slight. The model predictions at Pleasanton for col-
iphage (F geometric mean = 2.42) and fecal coliform (F geometric mean = 2.47) tended to give precise esti-
mates of the net measured aerosol concentrations. While subject to uncertainty due to the small number of
data pairs, the precision analyses of the pathogenic microorganism model predictions in Table VI.D-14 sug-
gest that model predictions of Clostridium perfringens and Pseudomonas have less precision than do predic-
tions at Pleasanton of the other microorganism groups.
Table VI. D-13
ANALYSIS OF THE PRECISION OF FECAL STREPTOCOCCI
MODEL PREDICTIONS
Sites
Pleasanton Ft. Huachuca All Sites and Methods
(C-B,P) Pairs with
Detectable* Measured 8 2 10
Concentrations
MeanoflOMDl 0.62 0.97 0.69
LargestOMD -1.33 +1.28 -1.33
Percentage of Discrepancy
Factors:
Below 2 25% 0% 20%
Below 5 63% 50% 60%
Below 10 88% 50% 80%
Geometric Mean of
Discrepancy Factors F 4.2 9.2 4.9
*
Above background (C-B > 0.5B) and detection limit (C > DL) and excluding presumably
contaminated samples.
174
-------
A summary of the precision of microbiological dispersion model predictions using all methods at
all sites (excluding the Deer Creek Lake predictions for total coliform because of the probable chlorination
effect) is presented in Table VI.D-15. Of the five microorganism groups evaluated at sampling sites in addi-
tion to Pleasanton, the most precise model predictions were obtained for standard bacterial plate count and
coliphage, for which the discrepancies averaged less than a factor of three. However, the geometric means of
the discrepancies averaged less than a factor of five for the other three microorganism groups (total coliform,
fecal coliform, and fecal streptococci). Considering the imprecision of microorganism aerosol measurements,
the predictions of the microbiological dispersion model may have sufficient precision to replace direct mea-
surement in many applications.
Table VI. D-14
ANALYSIS OF THE PRECISION OF PATHOGENIC MICROORGANISM
MODEL PREDICTIONS AT PLEASANTON
Clostridium
Pseudomonas Perfringens Mycobacteria
(C-B.P) Pairs with
Detectable* Measured 424
Concentrations
Meanof|OMD| 0.77 0.88 0.47
Largest OMD -1.01 -1.45 +1.01
Percentage of Discrepancy
Factors:
Below 2 0% 50% 50%
Below 5 25% 50% 75%
Below 10 75% 50% 75%
Geometric Mean of
Discrepancy Factors F 5.9 7.5 2.9
*Above background (C-B > 0.5B) and detection limit (C > DL) and excluding
presumably contaminated samples.
175
-------
Table VI. D-15
SUMMARY OF MODEL PREDICTION PRECISION FOR ALL SITES AND METHODS
(C-B.P) Pairs Geometric Percentage of Discrepancy
with Detectable Mean of All Factors F Below:
Measured Discrepancy
Microorganism Group Concentrations Factors F 2 5 10
Standard Bacterial Plate Count 98 2.7 43% 77% 95%
Coliphage 20 2.9 25% 80% 100%
Mycobacteria* 4 2.9 50% 75% 75%
Total Coliform+ 41 4.4 29% 71% 73%
Fecal Coliform 7 4.9 43% 71% 86%
Fecal Streptococci 10 4.9 20% 60% 80%
Pseudomonas* 4 5.9 0% 25% 75%
Clostridium Perfringens* 2 7.5 50% 50% 50%
+ Excludes Deer Creek Lake pairs because of probable clorination effect.
*A11 data pairs are from the Pleasanton sampling program.
176
-------
VII. DISCUSSION OF MICROBIOLOGICAL DISPERSION MODEL
A. Model Components
1. Aerosolization Efficiency E
It has been reported*66* that aerosolization efficiency depends upon spray nozzle type, spray arc
height, spray pressure, and wind velocity. Nozzle type, spray height, and spray pressures are operating condi-
tions that differ from one spray irrigation site to another. Thus, the aerosolization efficiency values given in
Table VI.C-2 can be considered characteristic only of the Pleasanton site, and not necessarily characteristic of
other spray irrigation sites.
The median aerosolization efficiency obtained for the Rainbird impact sprayers at Pleasanton
over the 17 runs during Phase II was 0.33 percent. This agrees very well with the aerosolization efficiencies
found for Rainbird impact sprayers at Ft. Huachuca, Arizona (median of 0.29 percent over three runs)<67>
at Deer Creek Lake, Ohio (median of 0.47 percent over four runs)*68', and at other sites(69). All of these studies
estimated aerosolization efficiency using water-soluble fluorescent dyes and diffusion modeling.
It is clear from Table VI.C-2 that an order-of-magnitude variation in aerosolization efficiency may
occur at a given site. The aerosolization efficiency regression equation (5) associates 80 percent of this varia-
tion with changes in meteorological conditions. Equation (5) indicates that the aerosolization efficiency at
Pleasanton increases with increasing air temperature, increasing wind velocity, and increasing solar radiation.
It agrees with the previously identified relationship of aerosolization efficiency to wind velocity*70 (perhaps
due to shearing forces), and implies that aerosolization efficiency is also influenced by other meteorological
factors that affect the evaporative capability of the air. Fairly high correlations typically occur among solar
radiation, air temperature, wind velocity, and relative humidity. Thus it is difficult to identify through regres-
sion the precise combination of these meteorological factors that affect evaporative capability. In summary,
an order-of-magnitude variation in aerosolization efficiency may occur at a given site, apparently as a result
of variation in atmospheric conditions that influence shearing forces, and evaporative capability.
The aerosolization efficiencies for the two dye runs made during Phase I at Pleasanton*71) were 0.6
percent and 0.5 percent. The equation (5) (p. 142) predictions of aerosolization efficiency for these runs'atmo-
spheric conditions were 1.1 percent and 0.9 percent. This verifies our inference (cf. Section VI.C.3) that equa-
tion (5) generally predicts the aerosolization efficiency at Pleasanton to within a factor of two.
Equation (5) may be insufficient to predict the aerosolization efficiency at spray irrigation sites
other than Pleasanton because aerosolization efficiency depends upon the operating conditions.*72' At some
sites such as Ft. Huachuca and Deer Creek Lake the operating conditions were similar enough to warrant use
of equation (5). However, at sites with different operating conditions, assuming that the effects of operating
conditions are relatively independent of the effects of meteorological conditions, reasonable aerosolization
efficiency estimates can still be obtained. A previous study of spray equipment and operating conditions*73',
or a limited dye aerosol sampling program at a site, can be utilized to estimate the site aerosolization effi-
ciency under one set of meteorological conditions. The constant in equation (5) can then be adjusted so that
aerosolization efficiencies for a site can be predicted under other meteorological conditions.
It should be noted that the solar radiation measurements at Pleasanton were made with a vertical
Belfort Pyrheliograph, which is a short wave instrument. Solar radiation measurements vary considerably de-
pending upon the type of recording instrument used. Thus in making aerosolization efficiency predictions
using equation (5), the solar radiation, r, used in the equation should be from a vertical short wave instrument
(e.g., vertical Eppley).
177
-------
2. Impact Factor I
The initial, supposedly rapid, microbiological die-off of wastewater aerosols is seldom quantified
separately from the aerosolization efficiency effect in the literature. Results reported by Sorber et a/(74) can be
converted to a net impact factor I =0.12, for total aerobic bacteria over eight aerosol runs using Andersen
samplers. Since this is approximately the 25th percentile of the standard bacterial plate count I distribution in
Table VI.C-6, satisfactory agreement is indicated.
The individual impact factor estimates for a microorganism group during a run contain consider-
able uncertainty. With the exception of total coliform, the majority of individual impact factor estimates were
less than twice their standard errors. Frequently, in fact, the standard error exceeded the impact factor esti-
mate. This uncertainty should be kept in mind when interpreting the empirical distributions of impact factors
given in Table VI.C-6. However, it is likely that both the central tendency and dispersion of the true I distri-
butions are well represented by the I estimate distributions.
The middle ranges (i.e., from the 25th to the 75th percentile) of the impact factor distributions
given in Table VI.C-6 suggest how well various microorganism groups survive the initial impact of aerosoliza-
tion. The viability of the putative wastewater indicator microorganism groups (fecal coliform, total coliform,
standard bacterial plate count, and coliphage) was substantially reduced through aerosol impact. Generally
only 6 to 60 percent of these microorganisms in the sprayed wastewater survived the initial seconds of aero-
solization.
The pathogenic bacteria and enteroviruses studied appear to survive aerosol formation and initial
contact with the atmospheric environment much better than the usual indicator organisms. The impact factor
estimates obtained at Pleasanton for Pseudomonas, fecal streptococci, Clostridium perfringens, mycobacte-
ria, and the enteroviruses were unexpectedly high, usually in excess of 1.0. A discussion is given in Section
VII.C.I of possible explanations for these large impact values, which superficially suggest survival above 100
percent.
The impact factor values for most of the microorganisms studied exhibited variation over several
orders of magnitude between their 10th and 90th percentiles. The literature generally implicates ultraviolet
solar radiation as a factor in microorganism die-off. There is also some evidence*75) that aerosolized bacteria
are reduced mainly at middle relative humidities (40-60 percent), and that high temperatures (>27°C) may
also reduce microbiological aerosols. Thus, variation in atmospheric conditions is suggested as a probable
cause of the variation in impact factor values within a microorganism group.
The preliminary analysis (see Appendix G), of the association of impact factor values with perti-
nent meteorological variables during the Pre-Fair runs, is also relevant. This analysis suggests that the impact
factors for total coliform, fecal coliform, and coliphage are reduced at low and middle relative humidities.
The total coliform and coliphage impact factors appear to be further reduced for the combination of high
solar radiation and high wind velocity, and for temperature difference between wastewater and air. The stan-
dard bacterial plate count impact factor may be lowered with strong solar radiation. Reductions in the impact
factors for the pathogenic bacteria (fecal streptococci, Pseudomonas, and Clostridium perfringens) seem to
occur primarily for the combination of high wind velocities with low relative humidities. Because there were
fairly high correlations among solar radiation, temperature, relative humidity, and wind velocity during Pre-
Fair runs, it is not clear whether the identified meteorological factor or a highly correlated alternative was
actually associated with an impact factor reduction.
As the model prediction examples illustrate, selection of an appropriate impact factor value is a
key step in making a model prediction for Pleasanton or elsewhere. To maximize predictive ability, selection
of the I value should be based on the relationship of impact factor estimates for a microorganism group in
Table VI.C-4 to the relevant atmospheric conditions occurring during their respective aerosol runs.
Such a direct analysis is needed to identify the nature of the relationship and to establish its
178
-------
strength. Unfortunately, the proper technique, a biased multiple regression analysis, was beyond the scope
and time frame of the current research effort. Hence, this analysis was not conducted.
Lacking this impact factor regression relationship, judicious application of the qualitative results
of the preliminary analysis of the Pre-Fair data must be substituted. These results provide some basis for pre-
dicting whether the impact factor for a microorganism group under a specific set of atmospheric conditions is
likely to be below, near, or above the median value obtained at Pleasanton. Thus, these associations of im-
pact factors with atmospheric conditions can be used in making microbiological dispersion model predictions
to select an appropriate percentile, presented in Table VI.C-6 as the microorganism I value. For making
model predictions, selecting I values below the 25th percentile or above the 75th percentile is not recom-
mended, both because of the uncertainty in the I estimates and because of the weak qualitative associations
with atmospheric conditions. Suppose, for example, that a prediction of the total coliform concentration is to
be made for a hot, sunny, dry, and windy summer afternoon, which might be the most hostile impact factor
case. The 40th percentile I = .13 might be used to obtain a typical total coliform aerosol concentration, while
the 25th percer.iile I = .06 might be appropriate for calculating a "best case" aerosol concentration. Conver-
sely, under very favorable atmospheric conditions, the 60th percentile I = .23 might yield a typical total col-
iform aerosol concentration, while the 75th percentile I = .55 might give a "worst case" aerosol concentra-
tion.
3. Viability Decay Rate
The measurement of a sizable reduction in the microbiological aerosol concentration over a large
aerosol age (distance) span downwind from the wastewater source is required to calculate an accurate viability
decay rate. In field studies, this requirement dictates the simultaneous operation of several high-volume aero-
sol samplers, which are expensive and difficult to operate<76). For this reason, viability decay rates based on
field studies have seldom been reported in the literature. A net viability decay rate equivalent to A = —0.06 s-'
was calculated for total aerobic bacteria over eight aerosol runs by Sorber, et a/<77>. Since this value is around
the 15th percentile of the standard bacterial plate count A distribution in Table VI.C-10, it suggests slightly
more rapid decay than do the viability decay rate values presented herein.
The individual viability decay rate estimates in Table VI.C-7 contain substantial uncertainty. Both
the relative magnitude of the standard errors of the A estimates and the frequency of non-negative decay rate
estimates (denoted as A =x) attest to this uncertainty in the individual estimates. These uncertainties, which
reflect state-of-the-art sampling and analytical limitations, should be recognized in interpreting the A distribu-
tions presented in Table VI.C-10.
For hardy microorganisms experiencing slow viability decay, very large sampling distances and
very low detection limits are needed to quantify the viability decay rate. Because of their relatively low waste-
water concentrations at Pleasanton, the hardy pathogenic bacteria were infrequently detected in the aerosol
samples taken far downwind of the spray line. For this reason, only values in the lower (rapid decay) portions
of the viability decay rate distributions of the pathogenic bacteria could be quantified in Table VI.C-10.
For all microorganism groups, the upper portion of the A distribution, which represents the slower
die-off rates, cannot be quantified based on the Pleasanton study. Table VI.C-10 suggests that — 0.01s'1 to
—0.02s"1 was the lowest viability decay rate which could be detected at Pleasanton. For the hardier and
partially analyzed microorganisms, such as mycobacteria, Pseudomonas, and fecal streptococci, perhaps
—0.06s"1 was the lowest detectable viability decay rate.
Values for the unquantified percentiles in Table VI.C-10 can be estimated by interpolating be-
tween the highest quantified percentile value and the logical upper limit, A = 0 (no viability decay). This inter-
polation procedure was employed in the prediction examples to estimate the 40th, 60th and 75th percentile A
values for mycobacteria, enteroviruses, and fecal streptococci in Tables VI.C-11, 12, and 13. Even at the siz-
179
-------
able aerosol ages of these examples, the exponential die-off factor, eAa, for these slowly decaying microorga-
nisms makes only a slight reduction in the predicted aerosol concentration. Thus, the added uncertainty in P,
introduced by using an interpolated value for the viability decay rate, is slight.
Perusal of Table VI.C-10 indicates that the indicator microorganism groups, especially total col-
iform and fecal coliform, experienced more consistent and rapid die-off with aerosol age than did the patho-
genic bacteria evaluated. Based on their viability decay rates, the microorganism groups seem to fall into three
categories. Total coliform and fecal coliform were similar in that they exhibited the most rapid decay and
their decay rates could most frequently be estimated. Furthermore, the viability decay rate estimates for total
and fecal coliform were also more reliable than for the other microorganism groups. Viability decay appeared
to occur on about 50 percent of the runs for a second category of microorganisms (coliphage, Clostridium
perfringens, and standard bacterial plate count); their rates of decay were also slower than the coliform decay
rates. Decay with aerosol age could seldom be detected for the third category (mycobacteria, Pseudomonas,
and fecal streptococci).
The viability decay rate distributions also reflect substantial variation in microbiological die-off
with aerosol age from one run to another. Die-off that increases with high solar radiation, low-to-middle rela-
tive humidity, and high temperature has often been suggested*78-79-80'. Thus, the different meteorological con-
ditions during runs are presumed to cause the variation in die-off rates. Our preliminary analysis of the asso-
ciation of viability decay rates with meteorological variables, based on the Pre-Fair Pleasanton data, confirms
these general relationships. For most microorganism groups, rapid viability decay does seem to be associated
with summer daytime atmospheric conditions as identified through meteorological variables such as high tem-
perature, middle relative humidities and high solar radiation. However, there is also much variation in the
viability decay rates of a microorganism group under apparently similar sets of atmospheric conditions.
Selecting a proper decay rate value is crucial to making an accurate model prediction at substantial
distances from the microbiological aerosol source. A direct analysis of the relationship of microorganism A
estimates to atmospheric conditions is needed, both to discriminate between the "rapid decay" and "insignif-
icant decay" sets of atmospheric conditions, and to develop a regression relationship to predict the decay rate
for the rapid decay conditions. Because of its complexity, such an analysis was beyond the scope and time
frame of the present research effort.
Lacking such discrimination and regression relationships, the only basis for selecting a decay rate A
for use in model prediction are the general findings discussed above. Consideration of general atmospheric
conditions on a daytime-nighttime scale is recommended as the basis for selecting a viability decay rate value
for any microorganism group from the Table VI.C-10 distributions. Because the actual relationships of atmo-
spheric conditions to decay rates are unknown, it is suggested that the decay rates used in predictive model
calculations be restricted to the middle range (i.e., from the 25th to the 75th percentiles). Perhaps the 40th
percentile might be taken as a typical summer daytime value, while the 60th percentile might reflect typical
nighttime conditions.
B. Validity of the Model and its Predictions
The preliminary assessment in Section VI.C.7 of factors affecting Pre-Fair microbiological aerosol lev-
els provided justification for establishing the multiplicative form of the microbiological dispersion model, as
given in equation (3). The reasonable physical interpretations that can be given to each model factor and pa-
rameter enhance the usefulness of the equation (3) form for expressing the microbiological dispersion model.
In using the microbiological dispersion model for predicting microorganism aerosol concentrations
from spray irrigation systems, three major assumptions are made: that the multiplicative form of the model is
valid; that microorganism die-off depends only on factors such as atmospheric conditions which are indepen-
dent of the spray site; and that the sampling, shipping, and assay methods of the Pleasanton study are em-
ployed. If these model assumptions are valid, the microbiological dispersion model should be applicable to
180
-------
any spray irrigation site over the range of meteorological conditions occurring during the Pleasanton aerosol
sampling runs. The extremes of meteorological conditions on the Pleasanton aerosol runs (see Table VI.B-1)
are summarized in Table VII.B-l.
Table VII.B-1
METEOROLOGICAL CONDITIONS OF PLEASANTON AEROSOL RUNS
Range of Values
Low High
Dry Aerosol Runs
Temperature, °C 7 34
Solar Radiation, W/m2 <20 930
Relative Humidity, % 5 80
Wind velocity, m/s 1.2 8.5
Microbiological Aerosol Runs
Temperature, °C 8 37
Solar Radiation, W/m2 <20 900
Relative Humidity, % 5 86
Wind Velocity, m/s 0.5 7.2
At the present state of model development, predictions of the model have an important shortcoming.
The procedure suggested above for selecting the I and A parameter percentiles as a function of atmospheric
conditions does not have adequate statistical justification. As discussed above, the existing Pleasanton data
have not been analyzed to determine the relationships of the individual I and X estimates for each microorga-
nism group to the aerosol run atmospheric conditions. Sophisticated regression techniques, such as biased re-
gression, appear necessary to elucidate such relationships. The complexity of this analysis precluded its con-
duct within the scope and time frame of this research effort. However, since this analysis is necessary to
achieve the full potential usefulness of the microbiological dispersion model, it is recommended as a priority
research area.
The effect of atmospheric conditions on an individual model parameter (D, E, I, or A) is sometimes
stronger that its net effect on the predicted concentration P. This happens because the atmospheric condition
effects on D, E, I, and A tend to vary in opposite directions that partially cancel out in the resulting prediction.
Comparison of the daytime and nighttime cases in Table VI.C-13 illustrates this characteristic. While the
summer midday atmospheric conditions reduce the microbiological model factors I and eia and the diffusion
concentration D well below their nighttime levels, they simultaneously elevate the aerosolization efficiency E.
Thus, when projecting the effect of a change in a single meteorological variable on the downwind microbiolo-
gical aerosol concentration, its influence on all of the model parameters should be taken into account.
The accuracy and precision of the predictions of the microbiological dispersion model were examined in
the model evaluation. Enough field data were available to provide a thorough model evaluation at realistic
field sampling distances (<200 meters from the source boundary) under a variety of sampling and analytical
procedures for standard bacterial plate count, and useful evaluations for total coliform and coliphage. How-
ever, the predictive ability of the model remains untested for fecal coliform, fecal streptococci, Pseudomonas,
Clostridium perfringens, and mycobacteria at sites other than Pleasanton.
The predictions of the microbiological dispersion model were quite accurate over the field data evalu-
ated. The predictions tended to be slightly less than the net measured aerosol concentrations, averaging 64
181
-------
percent of the net measured value for standard bacterial plate count, and 73 percent for both total coliform
and coliphage (see Table VI.D-8). No significant bias could be detected for total coliform and coliphage, but
the prediction underestimates for standard bacterial plate count and fecal coliform do indicate a slight bias in
their predictions.
The microbiological dispersion model predictions have satisfactory precision, considering the sizable
sampling and analytical uncertainty present in microorganism aerosol concentration determinations. The pre-
ponderance of the model predictions (e.g. 95 percent of standard bacterial plate count predictions, 73 percent
of total coliform predictions, and 100 percent of coliphage predictions) were within one order of magnitude
of the net measured values (see Table VII.D-15). Most model predictions (e.g., 77 percent for standard bacte-
rial plate count, 71 percent for total coliform, and 80 percent for coliphage) were within a factor of five of
their net measured value. The geometric mean of the discrepancy factors was less than three for standard bac-
terial plate count, coliphage, and mycobacteria and less than five for total coliform, fecal coliform, and fecal
streptococci.
It should be recognized that the model accuracy and precision statistics were based largely on field data
obtained within 50 meters of the edge of the wetted spray area. Over these distances, the viability decay model
factor eia has little effect on the computed model prediction P. Thus, the model evaluation has primarily eval-
uated the appropriateness of the E and I model parameters, with minimal attention given to A. Since the vi-
ability decay rates obtained at Pleasanton varied widely, it is to be expected that both the accuracy and preci-
sion of the model predictions will deteriorate with distance from the wastewater aerosol source.
When the wastewater that is aerosolized contains residual levels of total chlorine, the microbiological
dispersion model tends to underpredict the net measured aerosol concentration. Table VI.D-2 suggests under -
prediction of standard bacterial plate count aerosol concentrations by one and one half to three orders of
magnitude at a total residual chlorine concentration of 6 mg/1 in the sprayed wastewater. The Deer Creek
Lake wastewater that was sprayed contained total residual chlorine concentrations in the range of 0.1 mg/1 to
0.4 mg/1. The model predictions of presumptive total coliforms on Endo broth and of standard bacterial plate
count were both below the net measured aerosol concentrations at Deer Creek Lake, by one to two orders of
magnitude for total coliform (Table VI.D-10), and by nearly a factor of two for standard bacterial plate count
(Table VI.D-3). The degree of model underprediction for chlorinated wastewater aerosols appears to depend
both on the extent of residual chlorination and on the fragility of the microorganism group. Unless sufficient
data are available to adjust for this chlorination effect, the microbiological dispersion model should not be
used to predict microorganism aerosol concentrations of sprayed wastewater containing residual chlorine.
Considering the imprecision and cost of measuring microorganism aerosol concentrations by field sam-
pling, the predictions of the microbiological dispersion model do appear to be a preferable alternative when
the sprayed wastewater does not contain residual chlorine.
C. Microbiological Inferences Derived from the Model
1. Interpretation of Impact Factors Exceeding One
Frequent impact factor values exceeding 1.0 were not anticipated for any microorganism group.
Occasional I values above 1.0 are to be expected based on the large I standard errors presented in Table
VI.C-4. However, about 50 percent of the values obtained as impact factor estimates for fecal streptococci,
Clostridium perfringens and mycobacteria exceeded 1.0. Nearly all of the enterovirus and Pseudomonas im-
pact estimates exceeded this value. Thus, impact factors larger than one apparently characterize these patho-
gens under many atmospheric conditions with the measurement methodologies employed at Pleasanton.
The facile interpretation of the I values exceeding 1.0 is that substantial net growth rather than die-
off of the microorganisms occurred in the hostile aerosol environment during the initial seconds from aerosol
182
-------
formation to sampling at 50 meters downwind. However, this facile interpretation is inconsistent with known
microbiological behavior. Thus, one or several more subtle phenomena must be responsible for these high
observed I values.
Five possible explanations for the consistent pathogen I values above 1.0 are offered:
• Survival Hypotheses. After aerosol sample collection, microorganisms may have
survived in the supportive BHI collection fluid at undiminished concentrations over
the typical 16- to 24-hour holding time required for sample storage and shipment to
the analytical laboratory. However, the survival of human pathogens in the waste-
water samples which provide the baseline for comparison with the aerosol sampling
over the same holding time may have been reduced because the wastewater presents
a relatively hostile environment. The preliminary liquid collection media study con-
ducted in the laboratory showed that, when poliovirus 1 and f2 bacteriophage were
inoculated in wastewater, slight reductions in concentration occurred after 24 hours
at the holding temperature of 4°C. However, in the BHI medium, stable levels of
seeded pathogenic bacteria, poliovirus, and bacteriophage were maintained for seve-
ral days at 4°C. Thus, at least for coliphage and the enteroviruses, a slight elevation
(probably less than a factor of two) of the impact factor might be attributable to dif-
ferential survival in the wastewater and aerosol collection media.
• Masking Hypothesis. The assay procedures may have consistently underestimated
the wastewater concentrations of the pathogens due to the masking effect of chemi-
cal constituents or of the numerous other microorganisms present in the wastewater.
Conversely, the assay procedures may have more accurately estimated the pathogen
aerosol concentrations, because of the lower bacterial concentrations present in BHI
collection medium, the selective decimation of the masking bacteria in the aerosol
state, or the selective exclusion of inhibitory chemicals through aerosolization.
• Mechanical Splitting Hypothesis. Microorganisms that tend to exist in grouped form
in the wastewater (e.g., the long chains of fecal streptococci and the clumps of my-
cobacteria aggregates with wastewater solids) may have been mechanically split into
individual viable organisms prior to the aerosol assay. The potential for mechanical
splitting exists both during aerosolization, through impact and shear forces at the
spray head, during collection, through the rapid recirculation of the BHI fluid
within the high volume aerosol sampler.
• Regrowth Hypothesis. Regrowth of aerosol-sampled pathogenic microorganisms in
the supportive BHI medium could occur under opportune circumstances. Aerosol
sample regrowth would be much more rapid than the corresponding regrowth of
these microorganisms in the wastewater environment. An opportunity for regrowth
of pathogenic bacteria did occur occasionally when the sample shipping temperature
rose higher than the 4°C specified in the sampling protocol. However this hypothesis
is not valid for enteroviruses.
• Aerosolization Efficiency Hypothesis. The aerosolization efficiency E may have fre-
quently been underestimated. If the proportion of sprayed microorganisms that
were aerosolized exceeded the proportion of fluorescent dye aerosolized because the
dye is not an adequate physical model of microorganism aerosolization, E would
have consistently been underestimated. In addition, occasional underestimation of E
due to equation (5) uncertainty is also expected. Given the equation (5) standard
error of 0.194, the predicted E is expected to be low by at least a factor of two on 7
percent of the run estimates. Underestimation of E for either reason will result in a
compensating overestimation of the impact factor I values for all microorganism
groups for the affected runs.
The available evidence suggests that the phenomena represented by all five hypotheses may have
183
-------
had an effect on some of the impact factors calculated from the Pleasanton data. The survival hypothesis may
always be operative in wastewater assay, but the collection media study data indicate that its magnitude is
insufficient to alone account for the large impact factor estimates. The occurrence of masking in the wastewa-
ter assay is also quite probable; presumably the size of the masking effect could vary considerably from one
sample to another, depending on the concentrations of the masking microorganisms or chemicals that may be
present. Mechanical splitting is a likely mechanism of variable magnitude for the microorganisms that persist
in grouped form. Except for enteroviruses, the regrowth hypothesis is also plausible when the Pleasanton
sampling and shipping protocols were violated. However, all detected violations were carefully noted and
their occurrence was too infrequent to explain very many of the high I values. For example, the pathogen
assay laboratory received only 3 of the 55 microbiological run sample shipments at elevated temperatures
(8°C to 9°C). Occasional underestimation of E due to equation (5) uncertainty did occur, but also with insuf-
ficient frequency. Biased underestimation of E, because dye is an inadequate physical model of aerosoliza-
tion, is plausible, but unsubstantiated. In summary, the survival, masking, and mechanical splitting hypoth-
eses appear to be the most likely explanations of consistently obtaining impact factors above 1.0.
Regardless of which hypotheses are correct, the higher than expected impact factor estimates ap-
parently do reflect real phenomena associated with the current state-of-the-art for wastewater and aerosol
sampling and assay for pathogens. Therefore, to satisfy the third model assumption, the high impact factor
values presented in Table Vl.C-4 should be used without adjustment when predicting pathogen aerosol con-
centrations with the microbiological dispersion model.
2. Relative Aerosol Survival Hardiness of Microorganism Groups
Comparison of the microorganism groups, with respect to their impact factor values and their vi-
ability decay rates, provides an indication of the relative survivability of these groups through the wastewater
aerosolization process. A relative hardiness measure was constructed as the sum of separate rankings for ini-
tial survival (the median impact factor in Table VI.C-6) and for survival with age [the percentage of very low
(unquantified or below 0.01s'1)decay rates] for each microorganism. The resultant ranking of microorganism
groups is presented in Table VII.C-l, with microorganisms having a similar hardiness ranking being clustered
together.
It can be inferred from Table VII.C-l that the commonly used indicators of wastewater microor-
ganisms (total coliform, fecal coliform, coliphage, and standard bacterial plate count) do not survive waste-
water aerosolization nearly as well as do the pathogens studied. The evaluated pathogenic bacteria and enter-
oviruses both better survive the initial shock of aerosolization and more frequently resist aerosol age decay.
Therefore, the common "microbiological wastewater indicators", especially total coliform and fecal col-
iform, are actually very poor indicators of the pathogenic aerosol hazard posed by wastewater spray irriga-
tion. Fecal streptococci appear to be a more acceptable wastewater aerosol indicator. Fecal streptococci are
generally present in the wastewater, are readily assayed, and survive wastewater aerosolization well.
D. Model Applications
The microbiological dispersion model should prove valuable in many applications, especially after de-
velopment of a reliable procedure for selection of the I and A. parameters, and evaluation of model predictions
using this selection procedure. The Pleasanton sampling program has demonstrated that, with present micro-
biological aerosol sampling and assay methods, it is generally impractical to sample wastewater aerosols for
microorganisms beyond 100 or 200 meters from their source. Thus, a reliable modeling technique, such as the
microbiological dispersion model, is essential to estimate the level of human exposure to pathogens from
wastewater aerosols. The microbiological dispersion model could be used to calculate the pathogen exposure
levels of plant workers and neighboring residents at existing and candidate spray irrigation sites<81> as part of
184
-------
an evaluation of the potential public health risk. With limited aerosol sampling data, the microbiological dis-
persion model could also be used to calculate distant downwind concentrations emanating from other micro-
biological aerosol sources, such as the aeration basins of sewage treatment plants and cooling towers that
reuse municipal wastewater.
Table VII.C-1
AEROSOL SURVIVAL HARDINESS OF MICROORGANISM GROUPS
Initial Survival Survival with Age
Median Impact Percentage of Runs
Factor Value with Low Decay Rate
Microorganism Group (I) (A = Xor X > —0.01s ')
Total Coliform 0.16 32%
Fecal Coliform 0.13 38%
Coliphage 0.34 49%
Std. Bacterial Plate Count 0.21 67%
Clostridium perfringens 1-2 55%
Mycobacteria0.89 88%
Fecal Streptococci 1-7 77%
Pseudomonas 14 77%
-V*
Enteroviruses 40
*Based on only two special virus aerosol runs
185
-------
LIST OF REFERENCES
1. Turner, D. G., "Workbook of Atmospheric Dispersion Estimates," Publication No. AP-26, U. S. En-
vironmental Protection Agency, Research Triangle, NC (1970).
2. Sorber, C. A., H. T. Bausum, S. A. Schaub, and M. J. Small, "A Study of Bacterial Aerosols at a
Wastewater Irrigation Site," Jour. Water Poll. Control Fed., Vol. 48, 2367 (1976).
3. Cramer, H. E., J. R. Bjorklund, R. K. Dumbauld, J. E. Faulkner, and F. A. Record, Development of
Dosage Models and Concepts, Desert Test Center Report DTC-TR-72-609, U. S. Army, Ft. Douglas,
UT(1972).
4. Calder, K. L., A Numerical Analysis of the Protection Afforded by Buildings Against BW Aerosol At-
tack. BWL Tech. Study No. 2, U. S. Army Biological Laboratory, Fort Detrick, Maryland (1957).
5. Sorber, C. A., et al., op. cit.
6. Bausum, H. T., S. A. Schaub, M. J. Small, J. A. Highfill, and C. A. Sorber, Bacterial Aerosols Result-
ing from Spray Irrigation with Wastewater, Technical Report 7602, U. S. Army Medical Bioengineering
Research and Development Laboratory, Ft. Detrick, MD (1976).
7. Bausum, H. T., S. A. Schaub, and K. F. Kenyon, Viral and Bacterial Aerosols at a Wastewater Spray
Irrigation Site, Technical Report 7804, U.S. Army Medical Bioengineering Research and Development
Laboratory, Ft. Detrick, MD (1978).
8. "Clean Water Act of 1977" (P.L. 95-217, Sec. 12), U. S. Code, 91 Stat., 1959 (December 27, 1977).
9. Costle, D., "EPA Policy on Land Treatment of Municipal Wastewater," Letter to EPA Assistant and
Regional Administrators, (October 3, 1977).
10. Wellings, F. M., A. L. Lewis and C. W. Mountain, "Virus Studies in a Spray Irrigation Project," IFAS
Wastewater Workshop, 202, (1973).
11. Sorber, C. A. and K. J. Outer, "Health and Hygiene Aspects of Spray Irrigation, " AJPH, Vol. 65, 47
(1975).
12. Parson, D., C. Brownlee, D. Wetter, A. Maurer, E. Haughton, L. Kornder and M. Slezak, "Health
Aspects of Sewage Effluent Irrigation," Pollution Control Board, British Columbia Water Resources
Service Department of Lands, Forests, and Water Resources, Victoria, British Columbia (1975).
13. Katzenelson, E. and B. Teltch, "Dispersion of Enteric Bacteria by Spray Irrigation," Journal WPCF,
Vol.48, 710(1976).
14. Elliott, L. F. and J. R. Ellis, "Bacterial and Viral Pathogens Associated with Land Application of Or-
ganic Wastes," /. Environ. QuaL, Vol. 6, 245 (1977).
15. Ledbetter, J. O. and Randall, C. W., "Bacterial Emissions from Activated Sludge Units," Ind. Medi-
cine and Surgery, Vol. 34 No. 2, 130 (Feb. 1965).
16. Raynor, G. S. and J. V. Hayes, "Aerosol Production by Irrigation Equipment Used for Land Applica-
tion of Wastewater," Am. Ind. Hyg. Assoc. J., Vol. 37, 526 (Sept. 1976).
17. Sorber, C. A., H. T. Bausum, S. A. Schaub and M. J. Small, U. S. Army Medical Bioengineering Re-
search and Development Laboratory, Ft. Detrick, Maryland. Presented at 48th Annual Conference
Water Pollution Control Federation, Miami Beach, Florida (1975).
186
-------
18. Clark, C. S., E. J. Cleary, G. M. Schiff, C. C. Linnemann, Jr., J. P. Phair, and T. M. Briggs, "Disease
Risks of Occupational Exposure to Sewage," Journal of the Environmental Engineering Division,
EE2:375 (1976).
19. Katzenelson, E., I. Buium, and H. I. Shuval, "Risk of Communicable Disease Infection Associated
with Wastewater Irrigation in Agricultural Settlements," Science, Vol. 194, 944 (1976).
20. Johnson, D. E., D. E. Camann, J. W. Register, R. J. Prevost, J. B. Tillery, R. E. Thomas, J. M. Taylor
and J. M. Hosenfeld, "Health Implications of Sewage Treatment Facilities," EPA Contract 68-02-
1746, Final Report March 1978.
21. Akin, E. W., H. R. Pahren, W. Jakubowski, and J. B. Lucas, "Health Hazards Associated with Waste-
water Effluents and Sludges: Microbiological Considerations," Proceedings of the Conference on Risk
Assessment and Health Effects of Land Application of Municipal Wastewater and Sludges, University
of Texas at San Antonio (1978).
22. Johnson, D. E., J. W. Register, D. E. Camann, C. H. Millstein and J. L. Gulinson, Evaluation of
Health Effects Associated with the Application of Wastewater to Land, Phase I Report by Southwest
Research Institute to U. S. Army Medical Research and Development Command, Fort Detrick, MD
(December 1975).
23. Lennette, E. H., E. M. Spaulding, J. P. Truant, Manual of Clinical Microbiology, Second edition.
American Society for Microbiology, Washington, D. C. (1974).
24. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California, H. E. Cramer Co. Technical Report TR 76-303-03 to U. S. Army, Dugway
Proving Ground, UT (1976).
25. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California, H. E. Cramer Co. Technical Report TR 77-309-01 to Southwest Research Institute, San An-
tonio, TX (1977).
26. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California.
27. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California.
28. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California.
29. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California.
30. Hamil, H. F., and D. E. Camann, Collaborative Study of Method for the Determination of Nitrogen
Oxide Emissions from Stationary Sources (Fossil Fuel Fired Steam Generators), Southwest Research In-
stitute report for U. S. Environmental Protection Agency, Research Triangle Park, NC (1973).
31. Hamil H. F., D. E. Camann, and R. E. Thomas, Collaborative Study of Method for the Determination
of Sulfuric Acid Mist and Sulfur Dioxide Emissions from Stationary Sources, EPA-65014-75-003, U. S.
Environmental Protection Agency, Research Triangle Park, NC (1974).
32. Lindgren, B. W., Statistical Theory. Macmillan Co., New York (1962).
33. Zeigler, R. K., "Estimators of Coefficient of Variation Using K Samples," Technometrics, Vol. 15 No.
2, 409 (May 1973).
34. Hamil, H. F., et al., Collaborative Study of Method for the Determination of Sulfuric Acid Mist and
Sulfur Dioxide Emissions from Stationary Sources.
35. Turner, D. G., op. cit.
187
-------
36. Sorber, C. A., et al., "A Study of Bacterial Aerosols at a Wastewater Irrigation Site."
37. Cramer, H. E., et al., op. cit.
38. Calder, K. L., op. cit.
39. Dimmick, R. L. and R. J. Heckly, Theoretical Aspects of Microbial Survival. Pages 347-374 in R. L.
Dimmick and A. B. Akers, eds. An Introduction to Experimental Aerobiology. Wiley-Interscience,
New York (1969).
40. Lighthart, B. and D. S. Frisch, "Estimation of Viable Airborne Microbes Downwind from a Point
Source," Appl. Environ. MicrobioL, Vol 31(5), 700-704 (1976).
41. Turner, D. G., op. cit.
42. Cramer, H. E., etal., op. cit.
43. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California.
44. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California.
45. Cramer, H. E., etal., op. cit.
46. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California.
47. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California.
48. Anderson, A. J., Aerosolization Efficiencies and Viability Losses for Pre-Fair Wastewater Spray Trials
at Pleasanton, California.
49. Anderson, A. J., Aerosolization Efficiencies for the Post-Fair Wastewater Spray Trials at Pleasanton,
California.
50. Draper, N. R., and H. Smith, Applied Regression Analysis. J. Wiley & Sons, New York (1966), pp. 17-
21.
51. Armitage, P. Statistical Methods in Medical Research. J. Wiley & Sons, New York (1971), pp. 97-98.
52. Draper, N. R. and H. Smith, op. cit., pp. 77-81.
53. Turner, D. G., op. cit.
54. Guzewich, D. C., and W. J. B. Pringle, "Validation of the EPA-PTMTP Short Term Gaussian Dispers-
ion Model," /. Air Poll. Control Assoc., Vol. 27, 540 (June 1977).
55. Cramer, H. E., etal., op. cit.
56. Bjorklund, J. R., and R. K. Dumbauld, User's Instructions for the Volume Source Diffusion Models
Computer Program, H. E. Cramer Co. Technical Report TR 75-303-03 to U. S. Army, Dugway Proving
Ground, UT (1975).
57. Pasquill, F., "The Estimation of the Dispersion of Windborne Material," Meteor. Mag., Vol. 90, 1063,
33-49(1%!).
58. Bausum, H. T., B. Brockett, P. Schumacher, R. Bates, S. A. Schaub, and H. Mckin, Microbiological
Aerosols from a Field Source During Sprinkler Irrigation with Wastewater at Deer Creek Lake, Ohio,
joint technical report of U. S. Army Medical Bioengineering R & D Lab., Ft. Detrick, MD and U. S.
Army Cold Regions Res. and Eng. Lab., Hanover, NH, (in preparation).
188
-------
59. Dumbauld, R. E., Calculated Aerosolization Efficiencies for the Deer Creek Lake and the 1974 and
1975 Fort Huachuca Wastewater Spray Trials, H. E. Cramer Co. Technical Report TR 77-124-01 to U.
S. Army, Ft. Detrick, MD 1977).
60. Dumbauld, R. E., op. cit.
61. Sorber, C. A., et al, "A Study of Bacterial Aerosols at a Wastewater Irrigation Site," Jour. Water
Poll. Control Fed., Vol. 48, 2367 1976).
62. Bausum, H. T., et al., Bacterial Aerosols Resulting from Spray Irrigation with Wastewater, Technical
Report 7602.
63. Bausum, H. T. et al., Viral and Bacterial Aerosols at a Wastewater Spray Irrigation Site, Technical Re-
port 7804.
64. Dumbauld, R. E., op. cit.
65. Bausum, H. T., et al., Viral and Bacterial Aerosols at a Wastewater Spray Irrigation Site, Technical
Report 7804.
66. Raynor, G. S. and J. V. Hayes, op. cit.
67. Sorber, C. A., et al., "A Study of Bacterial Aerosols at a Wastewater Irrigation Site," Jour. Water
Poll. Control Fed., Vol. 48, 2367 (1976).
68. Dumbauld, R. E., op. cit.
69. Hennessy, J., G. Raynor, and M. Small, Aerosols and Land Deposition from Wastewater Spray Irriga-
tion, Final Comprehensive Report by Brookhaven National Laboratory to U. S. Army Medical Re-
search and Development Command, Washington, D. C. (July 1975).
70. Raynor, G. S. and J. V. Hayes, op. cit.
71. Johnson, D. E., et al. Evaluation of Health Effects Associated with the Application of Wastewater to
Land.
72. Raynor, G. S. and J. V. Hayes, op. cit.
73. Hennessy, J., etal., op. cit.
74. Sorber, C. A., et al., "A Study of Bacterial Aerosols at a Wastewater Irrigation Site."
75. Sorber, C. A., S. A. Schaub and H. T. Bausum, "An Assessment of a Potential Virus Hazard Associ-
ated with Spray Irrigation of Domestic Wastewaters" in Malina, J. F. and Sagik, B. P. (Eds.), Virus
Survival in Water and Wastewater Systems, Center for Research in Water Resources, The University of
Texas at Austin, Austin, TX (1974).
76. Johnson, D. E., D. E. Camann, C. A. Sorber, B. P. Sagik, and J. P. Glennon, "Aerosol Monitoring
and Microbial Organisms Near a Spray Irrigation Site," Risk Assessment and Health Effects of Land
Application of Municipal Wastewater and Sludges, Center for Applied Research and Technology, The
University of Texas at San Antonio, San Antonio, TX (1978).
77. Sorber, C. A., et. al., "A Study of Bacterial Aerosols at a Wastewater Irrigation Site."
78. Wells, W. F., et al., "The Environmental Control of Epidemic Contagion 1. An Epidemiologic Study
of Radiant Disinfection of Air in Day Schools," Amer. Jour. Hyg., Vol. 35, 97 (1942).
79. Watkins, H. M. S., L. J. Goldberg, E. F. Deig, and W. R. Leif, "Behavior of Colorado Tick Fever,
Vesicular Stomatitis, Neurovaccinia and Encephalomyocarditis Viruses in the Airborne State," Pro-
ceedings First International Symposium on Aerobiology, Oakland Naval Biological Laboratories, U. S.
Navy Supply Center, Oakland, CA (1965).
80. Sorber, C. A., et al., "An Assessment of a Potential Virus Hazard Associated with Spray Irrigation of
Domestic Wastewaters."
189
-------
81. Camann, D. E., C. A. Sorber, B. P. Sagik, J. P. Glennon, and D. E. Johnson, "A Model for Predicting
Pathogen Concentrations in Wastewater Aerosols," Risk Assessment and Health Effects of Land Ap-
plication of Municipal Wastewater and Sludges, Center for Applied Research and Technology, Univer-
sity of Texas at San Antonio, San Antonio, TX (1978).
82. Environmental Research Corporation, Operation and Maintenance Manual: LEAP Sampler, Model
3440. St. Paul, Minnesota (1975).
83. Litton Model M Large-Volume Air Sampler: Instruction Manual Report No. 3028. Minneapolis, Min-
nesota (1966).
84. Taras, M. J., A. E. Greenberg, R. D. Hoak, and M. C. Rand, Standard Methods for the Examination
of Water and Wastewater, 13th Edition. American Public Health Association, Washington, D. C.
(1971).
85. Rand, M. D., A. E. Greenberg, M. J. Taras, and M. A. Franson, Standard Methods for the Examina-
tion of Water and Wastewater, 14th Edition, American Public Health Association, Washington, D. C.
(1975).
86. Taras, M. J., et al., Standard Methods for the Examination of Water and Wastewater, 13th Edition.
87. Albertsson, Per-Ake, Partition of Cell Particles and Macromolecules, John Wiley & Sons, New York
(1960).
88. Grindrod, J. and D. O. Cliver, "A Polymer Two Phase System Adapted to Virus Detection," Archiv
fur diegesamte Virusforschung 31:365 (1970).
89. Grindrod, J. and D. O. Cliver, "Limitations of the Polymer Two Phase System for Detection of Vi-
ruses," Archiv fur diegesamte Virusforschung 28:337 (1969).
90. Johnson, D. E., J. W. Register, D. E. Camann, R. E. Thomas, J. L. Gulinson, B. P. Sagik, J. M. Tay-
lor and R. J. Prevost, Evaluation of Health Effects Associated with the Application of Wastewater to
Land, Phase II, Pre-Fair Report. Report by Southwest Research Institute to U. S. Army Medical Re-
search Institute and Development Command, Fort Detrick, MD (October 1976).
91. Snedecor, G. W. and Cochran, W. G., Statistical Methods, 6th Edition, Iowa State University Press
Ames, Iowa (1967).
92. Sorber, C. A., et al., "An Assessment of a Potential Virus Hazard Associated with Spray Irrigation of
Domestic Wastewaters."
93. Turner, D. G., op. cit., pp. 40-41.
190
-------
TECHNICAL REPORT DATA
(Please read Instructions on The reverse before completing)
1. REPORT NO.
EPA-600/1-80-015
I. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
The Evaluation of Microbiological Aerosols
Associated with the Application of Wastewater
to Land: Pleasanton, California
February 1980 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D.E. Johnson, D.E. Camann, J.W. Register, R.E. Thomas,
C.A. Sorber, M.N. Guentzel, J. Taylor, H.J. Harding
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southwest Research Institute
Post Office Drawer 28510
San Antonio, Texas 78284
10. PROGRAM ELEMENT NO.
1BA607
11. CONTRACT/GRANT NO.
DAMD 17-75-C-5072
IAG-D7-0701
12. SPONSORING AGENCY NAME AND ADDRESS
Health Effects Research Laboratory - Cinn, OH
Office of Research and Development
U.S- Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report 6/30/75-3/31/78
14. SPONSORING AGENCY CODE
EPA/600/10
15.SUPPLEMENTARY NOTES jhis study was conducted in cooperation with U.S. Army Medical
Research and Development Command, Fort Detrick, Frederick, Maryland 21701.
is.ABSTRACT y^ purp0se Of ^^ study was to determine the extent that individuals near
spray irrigation sites are exposed to microorganisms in wastewater aerosols. This
report reviews a monitoring effort of a spray irrigation site utilizing unchlorinated
secondarily-treated wastewater from biofiltration treatment processes. Objectives
included an in-depth pathogen screen of wastewater, establishing the relationship
between pathogen levels and traditional indicator organisms, monitoring microorganisms
in air within 600 meters of the spray source, and development/validation of a micro-
biological dispersion model for predicting aerosol pathogen concentrations. Effluent
was monitored for microbiological, chemical, and physical characteristics and extensive
microorganism and dye aerosol samples were collected (77 aerosol runs). Enteroviruses
were detected in air, but at a very low density. Conclusions: There is considerable
underestimation of pathogen aerosol levels when using traditional indicators to predict
human exposures. A microbiological dispersion model may be used with minimal monitoring
to estimate exposure. There is little correlation between wastewater levels of
traditional indicators and pathogens. Aerosols containing microorganisms are generated
by spray irrigation of wastewater; they do survive aerosolization and can be transported
to nearby populations. Until dose-response relationships are developed, neither the
levels of aerosolized microorganisms that constitute a hazard nor the degree of required
wastewater disinfection can be specified.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Waste water; enteroviruses; irrigation
aerosols; mathematical models; micro-
organisms; viruses; sampling.
Waste water aerosols,
land application; spray
irrigation; environmental
monitoring; dispersion
models; indicator micro-
organisms; pathogens;
aerosolization efficiency
57U
68G
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
207
20. SECURITY CLASS (Thispage}
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
191
,V U S GOVERNMENT PRINTING OfFICE 1980-657-146/5605
------- |